Acute respiratory illnesses (ARIs), including acute exacerbations of asthma, croup, pneumonia, and bronchiolitis, are among the most common illnesses in childhood.¹ Although most ARIs can be managed in the outpatient setting, hospitalization is common with respiratory illnesses accounting for >425,000 hospitalizations annually.¹ Pneumonia, asthma, and bronchiolitis each rank among the top five reasons for pediatric hospitalization in the United States.¹ Successful efforts to prevent or mitigate the severity of ARIs could have a major impact on child health.

Exposure to secondhand smoke (SHS) is a preventable risk factor for ARI in children, particularly when there is regular exposure in the home.² Chronic exposure to SHS impacts systemic inflammation by suppressing serum interferon-gamma,³ which can lead to increased susceptibility to viral and bacterial infections,⁴ and increasing Th2 (atopic) cytokine expression, which is associated with asthma.⁵ SHS exposure in children has also been linked to diminished lung function.⁶ As a result, SHS exposure is associated with increased ARI susceptibility and severity in children.^7-10

Much research has focused on the clinical impact of SHS exposure on respiratory health in children, but little is known about the impact on patient-reported outcomes, such as health-related quality of life (HRQOL). Patient-reported outcomes help provide a comprehensive evaluation of the effectiveness of healthcare delivery systems. These outcomes are increasingly used by health service researchers to better understand patient and caregiver perspectives.¹¹ Given the known associations between SHS exposure and ARI morbidity, we postulated that regular SHS exposure would also impact HRQOL in children. In this study, we assessed the relationship between SHS exposure and HRQOL within a large, multicenter, prospective cohort of children presenting to the emergency department (ED) and/or hospital with ARI.

Study Population

This study was nested within the Pediatric Respiratory Illness Measurement System (PRIMES) study, a prospective cohort study of children with ARI in the ED and inpatient settings at five tertiary care children’s hospitals within the Pediatric Research in Inpatient Settings Network in Colorado, Pennsylvania, Tennessee, Texas, and Washington. Eligible children were two weeks to 16 years of age hospitalized after presenting to the ED with a primary diagnosis of asthma, croup, bronchiolitis, or pneumonia between July 1, 2014 and June 30, 2016. Because of an anticipated low frequency of croup hospitalizations, we also included children presenting to the ED and then discharged to home with this diagnosis. Children were assigned to a PRIMES diagnosis group based on their final discharge diagnosis. If there was a discrepancy between admission and discharge diagnoses, the discharge diagnosis was used. If a child had more than one discharge diagnosis for a PRIMES condition (eg, acute asthma and pneumonia), we chose the PRIMES condition with the lowest total enrollments overall. If the final discharge diagnosis was not a PRIMES condition, the case was excluded from further analysis. Patients with immunodeficiency, cystic fibrosis, a history of prematurity <32 weeks, chronic neuromuscular disease, cardiovascular disease, pulmonary diseases (other than asthma), and moderate to severe developmental delay were also excluded. Children admitted to intensive care were eligible only if they were transferred to an acute care ward <72 hours following admission. A survey was administered at the time of enrollment that collected information on SHS exposure, HRQOL, healthcare utilization, and demographics. All study procedures were reviewed and approved by the institutional review boards at each of the participating hospitals.

To ascertain SHS exposure, we asked caregivers, “How many persons living in the child’s home smoke?” Responses were dichotomized into non-SHS exposed (0 smokers) and SHS exposed (≥1 smokers). Children with missing data on home SHS exposure were excluded.

Health-Related Quality of Life Outcomes

We estimated HRQOL using the Pediatric Quality of Life (PedsQL^TM) 4.0 Generic Core and Infant Scales. The PedsQL instruments are validated, population HRQOL measures that evaluate the physical, mental, emotional, and social functioning of children two to 18 years old based on self- or caregiver-proxy report.^12-15 These instruments have also shown responsiveness as well as construct and predictive validity in hospitalized children.¹¹ For this study, we focused on the PedsQL physical functioning subscale, which assesses for problems with physical activities (eg, sports activity or exercise, low energy, and hurts or aches) on a five-point Likert scale (never to almost always a problem). Scores range from 0 to 100 with higher scores indicating a better HRQOL. The reported minimal clinically important difference (MCID), defined as the smallest difference in which individuals would perceive a benefit or would necessitate a change in management, for this scale is 4.5 points.^16,17

Children >8 years old were invited to complete the self-report version of the PedsQL. For children <8 years old, and for older children who were unable to complete them, surveys were completed by a parent or legal guardian. Respondents were asked to assess perceptions of their (or their child’s) HRQOL during periods of baseline health (the child’s usual state of health in the month preceding the current illness) and during the acute illness (the child’s state of health at the time of admission) as SHS exposure may influence perceptions of general health and/or contribute to worse outcomes during periods of acute illness.

Covariates collected at the time of enrollment included sociodemographics (child age, gender, race/ethnicity, and caregiver education), and healthcare utilization (caregiver-reported patient visits to a healthcare provider in the preceding six months). Insurance status and level of medical complexity (using the Pediatric Medical Complexity Algorithm)¹⁸ were obtained using the Pediatric Hospital Information System database, an administrative database containing clinical and resource utilization data from >45 children’s hospitals in the United States including all of the PRIMES study hospitals.¹³

Analysis

Descriptive statistics included frequency (%) and mean (standard deviation). Bivariate comparisons according to SHS exposure status were analyzed using chi-squared tests for categorical variables and analysis of variance for continuous variables. Multivariable linear mixed regression models were used to examine associations between home SHS exposure and HRQOL for baseline health and during admission, overall and stratified by diagnosis. Covariates in each model included age, sex, race/ethnicity, caregiver education, and healthcare visits in the preceding six months. We also included a hospital random effect to account for clustering of patients within hospitals and used robust standard errors for inference.

In a secondary analysis to explore potential dose-response effects of SHS exposure, we examined associations between an ordinal exposure variable (0 smokers, 1 smoker, ≥2 smokers) and HRQOL for baseline health and during admission for the acute illness. Because of sample size limitations, diagnosis-specific analyses examining dose-response effects were not conducted.

Study Population

Of the 2,334 children enrolled in the PRIMES study, 25 (1%) respondents did not report on home SHS exposure and were excluded, yielding a final study population of 2,309 children, of whom 728 (32%) had reported home SHS exposure. The study population included 664 children with asthma (mean age seven years [3.5]; 38% with home SHS exposure), 740 with bronchiolitis (mean age 0.7 years [0.5]; 32% with home SHS exposure), 342 with croup (mean age 1.7 [1.1]; 25% with home SHS exposure), and 563 with pneumonia (mean age 4.4 [3.8]; 27% with home SHS exposure; Table 1). Compared with non-SHS-exposed children, those with home SHS exposure tend to be slightly older (3.9 vs 3.4 years, P = .01), more likely to be non-Hispanic Black (29% vs 19%, P < .001), to have a chronic condition (52% vs 41%, P < .001), to come from a household where caregiver(s) did not graduate from college (45% vs 29%, P < .001), and to have public insurance (73% vs 49%, P < .001).

Home SHS Exposure and Health-related Quality of Life

The overall mean HRQOL score for baseline health was 83 (15), with a range across diagnoses of 82 to 87. Compared with non-SHS-exposed children, children with home SHS exposure had a lower mean HRQOL score for baseline health (adjusted mean difference –3.04 [95% CI -4.34, –1.74]). In analyses stratified by diagnosis, baseline health scores were lower for SHS-exposed children for all four conditions, but differences were statistically significant only for bronchiolitis (adjusted mean difference –2.94 [–5.0, –0.89]) and pneumonia (adjusted mean value –4.13 [–6.82, –1.44]; Table 2); none of these differences met the MCID threshold.

The overall mean HRQOL score at the time of admission was 56 (23), with a range across diagnoses of 49 to 61, with lower scores noted among SHS-exposed children compared with non-SHS-exposed children (adjusted mean difference –2.16 [–4.22, –0.10]). Similar to scores representing baseline health, admission scores were lower across all four conditions for SHS-exposed children. Only children with croup, however, had significantly lower admission scores that also met the MCID threshold (adjusted mean difference –5.71 [–10.67, –0.75]; Table 2).

To assess for potential dose-response effects of SHS exposure on HRQOL, we stratified SHS-exposed children into those with one smoker in the home (n = 513) and those with ≥2 smokers in the home (n = 215). Compared with non-SHS-exposed children, both HRQOL scores (baseline health and admission) were lower for SHS-exposed children. Consistent with a dose-response association, scores were lowest for children with ≥2 smokers in the home, both at baseline health (adjusted mean difference –3.92 [–6.03, –1.81]) and on admission (adjusted mean difference –3.67 [–6.98, –0.36]; Table 3).

Within a multicenter cohort of 2,309 children hospitalized with ARI, we noted significantly lower HRQOL scores among children exposed to SHS in the home as compared with nonexposed children. Differences were greatest for children living with ≥2 smokers in the home. In analyses stratified by diagnosis, differences in baseline health HRQOL scores were greatest for children with bronchiolitis and pneumonia. Differences in acute illness scores were greatest for children with croup.¹⁶

Our study provides evidence for acute and chronic impacts of SHS on HRQOL in children hospitalized with ARI. Although several studies have linked SHS exposure to reduced HRQOL in adults,^19,20 few similar studies have been conducted in children. Nonetheless, a wealth of studies have documented the negative impact of SHS exposure on clinical outcomes among children with ARI.^8,10,21-23 Our findings that home SHS exposure was associated with reduced HRQOL among our cohort of children with ARI are therefore consistent with related findings in adults and children. The observation that the effects of SHS exposure on HRQOL were greatest among children living with ≥2 smokers provides further evidence of a potential causal link between regular SHS exposure and HRQOL.

Although the magnitude and significance of associations between SHS exposure and HRQOL varied for each of the four diagnoses for baseline health and the acute illness, it is important to note that the point estimates for the adjusted mean differences were uniformly lower for the SHS-exposed children in each subgroup. Even so, only acute illness scores for croup exceeded the MCID threshold.¹⁶ Croup is the only included condition of the upper airway and is characterized by laryngotracheal inflammation leading to the typical cough and, in moderate to severe cases, stridor. Given that chronic SHS exposure induces a proinflammatory state,³ it is possible that SHS-exposed children with croup had more severe illness compared with nonexposed children with croup resulting in lower HRQOL scores on admission. Further, perceived differences in illness severity and HRQOL may be more readily apparent in children with croup (eg, stridor at rest vs intermittent or no stridor) as compared with children with lower respiratory tract diseases.

Of the four included diagnoses, the link between SHS exposure and asthma outcomes has been most studied. Prior work has demonstrated more frequent and severe acute exacerbations, as well as worse long-term lung function among SHS-exposed children as compared with nonexposed children.^22-24 It was, therefore, surprising that our study failed to demonstrate associations between SHS exposure and HRQOL among children with asthma. Reasons for this finding are unclear. One hypothesis is that caregivers of SHS-exposed children with asthma may be more aware of the impacts of SHS exposure on respiratory health (through prior education) and, thus, more likely to modify their smoking behaviors, or for their children to be on daily asthma controller therapy. Alternatively, caregivers of children with asthma may be more likely to underreport home SHS exposure. Thirty-eight percent of children with asthma, however, were classified as SHS-exposed. This percentage was greater than the other three conditions studied (25%-32%), suggesting that differential bias in underreporting was minimal. Given that children with asthma were older, on average, than children with the other three conditions, it may also be that these children spent more time in smoke-free environments (eg, school).

Nearly one-third of children in our study were exposed to SHS in the home. This is similar to the prevalence of exposure in other studies conducted among hospitalized children^8,10,21,25 but higher than the national prevalence of home SHS exposure among children in the United States.²⁶ Thus, hospitalized children represent a particularly vulnerable population and an important target for interventions aiming to reduce exposure to SHS. Although longitudinal interventions are likely necessary to affect long-term success, hospitalization for ARI may serve as a powerful teachable moment to begin cessation efforts. Hospitalization also offers time beyond a typical primary care outpatient encounter to focus on cessation counseling and may be the only opportunity to engage in counseling activities for some families with limited time or access. Further, prior studies have demonstrated both the feasibility and the effectiveness of smoking cessation interventions in hospitalized children.^27-30 Unfortunately, however, SHS exposure is often not documented at the time of hospitalization, and many opportunities to intervene are missed.^25,31 Thus, there is a need for improved strategies to reliably identify and intervene on SHS-exposed children in the hospital setting.

These findings should be considered in the context of several limitations. The observational nature of our study raises the potential for confounding, specifically with regard to socioeconomic status, as this is associated with both SHS exposure and lower HRQOL. Our modeling approach attempted to control for several factors associated with socioeconomic status, including caregiver education and insurance coverage, but there is potential for residual confounding. No single question is sufficient to fully assess SHS exposure as the intensity of home SHS exposure likely varies widely, and some children may be exposed to SHS outside of the home environment.³² The home, however, is often the most likely source of regular SHS exposure,^33,34 especially among young children (our cohort’s mean age was 3.6 years). Misclassification of SHS exposure is also possible due to underreporting of smoking.^35,36 As a result, some children regularly exposed to SHS may have been misclassified as nonexposed, and the observed associations between SHS exposure and HRQOL may be underestimated. Confirming our study’s findings using objective assessments of SHS exposure, such as cotinine, are warranted. Given the young age of our cohort, the PedsQL surveys were completed by the parent or legal guardian only in >90% of the enrolled subjects, and caregiver perceptions may not accurately reflect the child’s perceptions. Prior work, however, has demonstrated the validity of parent-proxy reporting of the PedsQL, including correlation with child self-report.³⁷ In our study, correlation between child and caregiver reporting (when available) was also very good (r = 0.72, 95% CI 0.64, 0.77). It is also possible that the timing of the HRQOL assessments (on admission) may have biased perceptions of baseline HRQOL, although we anticipate any bias would likely be nondifferential between SHS-exposed and nonexposed children and across diagnoses.

Nearly one-third of children in our study were exposed to SHS exposure in the home, and SHS exposure was associated with lower HRQOL for baseline health and during acute illness, providing further evidence of the dangers of SHS. Much work is needed in order to eliminate the impact of SHS on child health and families of children hospitalized for respiratory illness should be considered a priority population for smoking cessation efforts.

Acknowledgment

The authors wish to acknowledge the efforts of PRIS-PRIMES study team. The authors also wish to thank the children and families who consented to be a part of the PRIMES study.

Disclosures

The authors have no conflicts of interest relevant to this article to disclose.

Funding

This study was supported by NIH-NHLBI 1R01HL121067 to RMS.

Acute respiratory illnesses (ARIs), including acute exacerbations of asthma, croup, pneumonia, and bronchiolitis, are among the most common illnesses in childhood.¹ Although most ARIs can be managed in the outpatient setting, hospitalization is common with respiratory illnesses accounting for >425,000 hospitalizations annually.¹ Pneumonia, asthma, and bronchiolitis each rank among the top five reasons for pediatric hospitalization in the United States.¹ Successful efforts to prevent or mitigate the severity of ARIs could have a major impact on child health.

Exposure to secondhand smoke (SHS) is a preventable risk factor for ARI in children, particularly when there is regular exposure in the home.² Chronic exposure to SHS impacts systemic inflammation by suppressing serum interferon-gamma,³ which can lead to increased susceptibility to viral and bacterial infections,⁴ and increasing Th2 (atopic) cytokine expression, which is associated with asthma.⁵ SHS exposure in children has also been linked to diminished lung function.⁶ As a result, SHS exposure is associated with increased ARI susceptibility and severity in children.^7-10

Much research has focused on the clinical impact of SHS exposure on respiratory health in children, but little is known about the impact on patient-reported outcomes, such as health-related quality of life (HRQOL). Patient-reported outcomes help provide a comprehensive evaluation of the effectiveness of healthcare delivery systems. These outcomes are increasingly used by health service researchers to better understand patient and caregiver perspectives.¹¹ Given the known associations between SHS exposure and ARI morbidity, we postulated that regular SHS exposure would also impact HRQOL in children. In this study, we assessed the relationship between SHS exposure and HRQOL within a large, multicenter, prospective cohort of children presenting to the emergency department (ED) and/or hospital with ARI.

Study Population

This study was nested within the Pediatric Respiratory Illness Measurement System (PRIMES) study, a prospective cohort study of children with ARI in the ED and inpatient settings at five tertiary care children’s hospitals within the Pediatric Research in Inpatient Settings Network in Colorado, Pennsylvania, Tennessee, Texas, and Washington. Eligible children were two weeks to 16 years of age hospitalized after presenting to the ED with a primary diagnosis of asthma, croup, bronchiolitis, or pneumonia between July 1, 2014 and June 30, 2016. Because of an anticipated low frequency of croup hospitalizations, we also included children presenting to the ED and then discharged to home with this diagnosis. Children were assigned to a PRIMES diagnosis group based on their final discharge diagnosis. If there was a discrepancy between admission and discharge diagnoses, the discharge diagnosis was used. If a child had more than one discharge diagnosis for a PRIMES condition (eg, acute asthma and pneumonia), we chose the PRIMES condition with the lowest total enrollments overall. If the final discharge diagnosis was not a PRIMES condition, the case was excluded from further analysis. Patients with immunodeficiency, cystic fibrosis, a history of prematurity <32 weeks, chronic neuromuscular disease, cardiovascular disease, pulmonary diseases (other than asthma), and moderate to severe developmental delay were also excluded. Children admitted to intensive care were eligible only if they were transferred to an acute care ward <72 hours following admission. A survey was administered at the time of enrollment that collected information on SHS exposure, HRQOL, healthcare utilization, and demographics. All study procedures were reviewed and approved by the institutional review boards at each of the participating hospitals.

To ascertain SHS exposure, we asked caregivers, “How many persons living in the child’s home smoke?” Responses were dichotomized into non-SHS exposed (0 smokers) and SHS exposed (≥1 smokers). Children with missing data on home SHS exposure were excluded.

Health-Related Quality of Life Outcomes

We estimated HRQOL using the Pediatric Quality of Life (PedsQL^TM) 4.0 Generic Core and Infant Scales. The PedsQL instruments are validated, population HRQOL measures that evaluate the physical, mental, emotional, and social functioning of children two to 18 years old based on self- or caregiver-proxy report.^12-15 These instruments have also shown responsiveness as well as construct and predictive validity in hospitalized children.¹¹ For this study, we focused on the PedsQL physical functioning subscale, which assesses for problems with physical activities (eg, sports activity or exercise, low energy, and hurts or aches) on a five-point Likert scale (never to almost always a problem). Scores range from 0 to 100 with higher scores indicating a better HRQOL. The reported minimal clinically important difference (MCID), defined as the smallest difference in which individuals would perceive a benefit or would necessitate a change in management, for this scale is 4.5 points.^16,17

Children >8 years old were invited to complete the self-report version of the PedsQL. For children <8 years old, and for older children who were unable to complete them, surveys were completed by a parent or legal guardian. Respondents were asked to assess perceptions of their (or their child’s) HRQOL during periods of baseline health (the child’s usual state of health in the month preceding the current illness) and during the acute illness (the child’s state of health at the time of admission) as SHS exposure may influence perceptions of general health and/or contribute to worse outcomes during periods of acute illness.

Covariates collected at the time of enrollment included sociodemographics (child age, gender, race/ethnicity, and caregiver education), and healthcare utilization (caregiver-reported patient visits to a healthcare provider in the preceding six months). Insurance status and level of medical complexity (using the Pediatric Medical Complexity Algorithm)¹⁸ were obtained using the Pediatric Hospital Information System database, an administrative database containing clinical and resource utilization data from >45 children’s hospitals in the United States including all of the PRIMES study hospitals.¹³

Analysis

Descriptive statistics included frequency (%) and mean (standard deviation). Bivariate comparisons according to SHS exposure status were analyzed using chi-squared tests for categorical variables and analysis of variance for continuous variables. Multivariable linear mixed regression models were used to examine associations between home SHS exposure and HRQOL for baseline health and during admission, overall and stratified by diagnosis. Covariates in each model included age, sex, race/ethnicity, caregiver education, and healthcare visits in the preceding six months. We also included a hospital random effect to account for clustering of patients within hospitals and used robust standard errors for inference.

In a secondary analysis to explore potential dose-response effects of SHS exposure, we examined associations between an ordinal exposure variable (0 smokers, 1 smoker, ≥2 smokers) and HRQOL for baseline health and during admission for the acute illness. Because of sample size limitations, diagnosis-specific analyses examining dose-response effects were not conducted.

Study Population

Of the 2,334 children enrolled in the PRIMES study, 25 (1%) respondents did not report on home SHS exposure and were excluded, yielding a final study population of 2,309 children, of whom 728 (32%) had reported home SHS exposure. The study population included 664 children with asthma (mean age seven years [3.5]; 38% with home SHS exposure), 740 with bronchiolitis (mean age 0.7 years [0.5]; 32% with home SHS exposure), 342 with croup (mean age 1.7 [1.1]; 25% with home SHS exposure), and 563 with pneumonia (mean age 4.4 [3.8]; 27% with home SHS exposure; Table 1). Compared with non-SHS-exposed children, those with home SHS exposure tend to be slightly older (3.9 vs 3.4 years, P = .01), more likely to be non-Hispanic Black (29% vs 19%, P < .001), to have a chronic condition (52% vs 41%, P < .001), to come from a household where caregiver(s) did not graduate from college (45% vs 29%, P < .001), and to have public insurance (73% vs 49%, P < .001).

Home SHS Exposure and Health-related Quality of Life

The overall mean HRQOL score for baseline health was 83 (15), with a range across diagnoses of 82 to 87. Compared with non-SHS-exposed children, children with home SHS exposure had a lower mean HRQOL score for baseline health (adjusted mean difference –3.04 [95% CI -4.34, –1.74]). In analyses stratified by diagnosis, baseline health scores were lower for SHS-exposed children for all four conditions, but differences were statistically significant only for bronchiolitis (adjusted mean difference –2.94 [–5.0, –0.89]) and pneumonia (adjusted mean value –4.13 [–6.82, –1.44]; Table 2); none of these differences met the MCID threshold.

The overall mean HRQOL score at the time of admission was 56 (23), with a range across diagnoses of 49 to 61, with lower scores noted among SHS-exposed children compared with non-SHS-exposed children (adjusted mean difference –2.16 [–4.22, –0.10]). Similar to scores representing baseline health, admission scores were lower across all four conditions for SHS-exposed children. Only children with croup, however, had significantly lower admission scores that also met the MCID threshold (adjusted mean difference –5.71 [–10.67, –0.75]; Table 2).

To assess for potential dose-response effects of SHS exposure on HRQOL, we stratified SHS-exposed children into those with one smoker in the home (n = 513) and those with ≥2 smokers in the home (n = 215). Compared with non-SHS-exposed children, both HRQOL scores (baseline health and admission) were lower for SHS-exposed children. Consistent with a dose-response association, scores were lowest for children with ≥2 smokers in the home, both at baseline health (adjusted mean difference –3.92 [–6.03, –1.81]) and on admission (adjusted mean difference –3.67 [–6.98, –0.36]; Table 3).

Within a multicenter cohort of 2,309 children hospitalized with ARI, we noted significantly lower HRQOL scores among children exposed to SHS in the home as compared with nonexposed children. Differences were greatest for children living with ≥2 smokers in the home. In analyses stratified by diagnosis, differences in baseline health HRQOL scores were greatest for children with bronchiolitis and pneumonia. Differences in acute illness scores were greatest for children with croup.¹⁶

Our study provides evidence for acute and chronic impacts of SHS on HRQOL in children hospitalized with ARI. Although several studies have linked SHS exposure to reduced HRQOL in adults,^19,20 few similar studies have been conducted in children. Nonetheless, a wealth of studies have documented the negative impact of SHS exposure on clinical outcomes among children with ARI.^8,10,21-23 Our findings that home SHS exposure was associated with reduced HRQOL among our cohort of children with ARI are therefore consistent with related findings in adults and children. The observation that the effects of SHS exposure on HRQOL were greatest among children living with ≥2 smokers provides further evidence of a potential causal link between regular SHS exposure and HRQOL.

Although the magnitude and significance of associations between SHS exposure and HRQOL varied for each of the four diagnoses for baseline health and the acute illness, it is important to note that the point estimates for the adjusted mean differences were uniformly lower for the SHS-exposed children in each subgroup. Even so, only acute illness scores for croup exceeded the MCID threshold.¹⁶ Croup is the only included condition of the upper airway and is characterized by laryngotracheal inflammation leading to the typical cough and, in moderate to severe cases, stridor. Given that chronic SHS exposure induces a proinflammatory state,³ it is possible that SHS-exposed children with croup had more severe illness compared with nonexposed children with croup resulting in lower HRQOL scores on admission. Further, perceived differences in illness severity and HRQOL may be more readily apparent in children with croup (eg, stridor at rest vs intermittent or no stridor) as compared with children with lower respiratory tract diseases.

Of the four included diagnoses, the link between SHS exposure and asthma outcomes has been most studied. Prior work has demonstrated more frequent and severe acute exacerbations, as well as worse long-term lung function among SHS-exposed children as compared with nonexposed children.^22-24 It was, therefore, surprising that our study failed to demonstrate associations between SHS exposure and HRQOL among children with asthma. Reasons for this finding are unclear. One hypothesis is that caregivers of SHS-exposed children with asthma may be more aware of the impacts of SHS exposure on respiratory health (through prior education) and, thus, more likely to modify their smoking behaviors, or for their children to be on daily asthma controller therapy. Alternatively, caregivers of children with asthma may be more likely to underreport home SHS exposure. Thirty-eight percent of children with asthma, however, were classified as SHS-exposed. This percentage was greater than the other three conditions studied (25%-32%), suggesting that differential bias in underreporting was minimal. Given that children with asthma were older, on average, than children with the other three conditions, it may also be that these children spent more time in smoke-free environments (eg, school).

Nearly one-third of children in our study were exposed to SHS in the home. This is similar to the prevalence of exposure in other studies conducted among hospitalized children^8,10,21,25 but higher than the national prevalence of home SHS exposure among children in the United States.²⁶ Thus, hospitalized children represent a particularly vulnerable population and an important target for interventions aiming to reduce exposure to SHS. Although longitudinal interventions are likely necessary to affect long-term success, hospitalization for ARI may serve as a powerful teachable moment to begin cessation efforts. Hospitalization also offers time beyond a typical primary care outpatient encounter to focus on cessation counseling and may be the only opportunity to engage in counseling activities for some families with limited time or access. Further, prior studies have demonstrated both the feasibility and the effectiveness of smoking cessation interventions in hospitalized children.^27-30 Unfortunately, however, SHS exposure is often not documented at the time of hospitalization, and many opportunities to intervene are missed.^25,31 Thus, there is a need for improved strategies to reliably identify and intervene on SHS-exposed children in the hospital setting.

These findings should be considered in the context of several limitations. The observational nature of our study raises the potential for confounding, specifically with regard to socioeconomic status, as this is associated with both SHS exposure and lower HRQOL. Our modeling approach attempted to control for several factors associated with socioeconomic status, including caregiver education and insurance coverage, but there is potential for residual confounding. No single question is sufficient to fully assess SHS exposure as the intensity of home SHS exposure likely varies widely, and some children may be exposed to SHS outside of the home environment.³² The home, however, is often the most likely source of regular SHS exposure,^33,34 especially among young children (our cohort’s mean age was 3.6 years). Misclassification of SHS exposure is also possible due to underreporting of smoking.^35,36 As a result, some children regularly exposed to SHS may have been misclassified as nonexposed, and the observed associations between SHS exposure and HRQOL may be underestimated. Confirming our study’s findings using objective assessments of SHS exposure, such as cotinine, are warranted. Given the young age of our cohort, the PedsQL surveys were completed by the parent or legal guardian only in >90% of the enrolled subjects, and caregiver perceptions may not accurately reflect the child’s perceptions. Prior work, however, has demonstrated the validity of parent-proxy reporting of the PedsQL, including correlation with child self-report.³⁷ In our study, correlation between child and caregiver reporting (when available) was also very good (r = 0.72, 95% CI 0.64, 0.77). It is also possible that the timing of the HRQOL assessments (on admission) may have biased perceptions of baseline HRQOL, although we anticipate any bias would likely be nondifferential between SHS-exposed and nonexposed children and across diagnoses.

Nearly one-third of children in our study were exposed to SHS exposure in the home, and SHS exposure was associated with lower HRQOL for baseline health and during acute illness, providing further evidence of the dangers of SHS. Much work is needed in order to eliminate the impact of SHS on child health and families of children hospitalized for respiratory illness should be considered a priority population for smoking cessation efforts.

Acknowledgment

The authors wish to acknowledge the efforts of PRIS-PRIMES study team. The authors also wish to thank the children and families who consented to be a part of the PRIMES study.

Disclosures

The authors have no conflicts of interest relevant to this article to disclose.

Funding

This study was supported by NIH-NHLBI 1R01HL121067 to RMS.

Acute respiratory illnesses (ARIs), including acute exacerbations of asthma, croup, pneumonia, and bronchiolitis, are among the most common illnesses in childhood.¹ Although most ARIs can be managed in the outpatient setting, hospitalization is common with respiratory illnesses accounting for >425,000 hospitalizations annually.¹ Pneumonia, asthma, and bronchiolitis each rank among the top five reasons for pediatric hospitalization in the United States.¹ Successful efforts to prevent or mitigate the severity of ARIs could have a major impact on child health.

Exposure to secondhand smoke (SHS) is a preventable risk factor for ARI in children, particularly when there is regular exposure in the home.² Chronic exposure to SHS impacts systemic inflammation by suppressing serum interferon-gamma,³ which can lead to increased susceptibility to viral and bacterial infections,⁴ and increasing Th2 (atopic) cytokine expression, which is associated with asthma.⁵ SHS exposure in children has also been linked to diminished lung function.⁶ As a result, SHS exposure is associated with increased ARI susceptibility and severity in children.^7-10

Much research has focused on the clinical impact of SHS exposure on respiratory health in children, but little is known about the impact on patient-reported outcomes, such as health-related quality of life (HRQOL). Patient-reported outcomes help provide a comprehensive evaluation of the effectiveness of healthcare delivery systems. These outcomes are increasingly used by health service researchers to better understand patient and caregiver perspectives.¹¹ Given the known associations between SHS exposure and ARI morbidity, we postulated that regular SHS exposure would also impact HRQOL in children. In this study, we assessed the relationship between SHS exposure and HRQOL within a large, multicenter, prospective cohort of children presenting to the emergency department (ED) and/or hospital with ARI.

Study Population

This study was nested within the Pediatric Respiratory Illness Measurement System (PRIMES) study, a prospective cohort study of children with ARI in the ED and inpatient settings at five tertiary care children’s hospitals within the Pediatric Research in Inpatient Settings Network in Colorado, Pennsylvania, Tennessee, Texas, and Washington. Eligible children were two weeks to 16 years of age hospitalized after presenting to the ED with a primary diagnosis of asthma, croup, bronchiolitis, or pneumonia between July 1, 2014 and June 30, 2016. Because of an anticipated low frequency of croup hospitalizations, we also included children presenting to the ED and then discharged to home with this diagnosis. Children were assigned to a PRIMES diagnosis group based on their final discharge diagnosis. If there was a discrepancy between admission and discharge diagnoses, the discharge diagnosis was used. If a child had more than one discharge diagnosis for a PRIMES condition (eg, acute asthma and pneumonia), we chose the PRIMES condition with the lowest total enrollments overall. If the final discharge diagnosis was not a PRIMES condition, the case was excluded from further analysis. Patients with immunodeficiency, cystic fibrosis, a history of prematurity <32 weeks, chronic neuromuscular disease, cardiovascular disease, pulmonary diseases (other than asthma), and moderate to severe developmental delay were also excluded. Children admitted to intensive care were eligible only if they were transferred to an acute care ward <72 hours following admission. A survey was administered at the time of enrollment that collected information on SHS exposure, HRQOL, healthcare utilization, and demographics. All study procedures were reviewed and approved by the institutional review boards at each of the participating hospitals.

To ascertain SHS exposure, we asked caregivers, “How many persons living in the child’s home smoke?” Responses were dichotomized into non-SHS exposed (0 smokers) and SHS exposed (≥1 smokers). Children with missing data on home SHS exposure were excluded.

Health-Related Quality of Life Outcomes

We estimated HRQOL using the Pediatric Quality of Life (PedsQL^TM) 4.0 Generic Core and Infant Scales. The PedsQL instruments are validated, population HRQOL measures that evaluate the physical, mental, emotional, and social functioning of children two to 18 years old based on self- or caregiver-proxy report.^12-15 These instruments have also shown responsiveness as well as construct and predictive validity in hospitalized children.¹¹ For this study, we focused on the PedsQL physical functioning subscale, which assesses for problems with physical activities (eg, sports activity or exercise, low energy, and hurts or aches) on a five-point Likert scale (never to almost always a problem). Scores range from 0 to 100 with higher scores indicating a better HRQOL. The reported minimal clinically important difference (MCID), defined as the smallest difference in which individuals would perceive a benefit or would necessitate a change in management, for this scale is 4.5 points.^16,17

Children >8 years old were invited to complete the self-report version of the PedsQL. For children <8 years old, and for older children who were unable to complete them, surveys were completed by a parent or legal guardian. Respondents were asked to assess perceptions of their (or their child’s) HRQOL during periods of baseline health (the child’s usual state of health in the month preceding the current illness) and during the acute illness (the child’s state of health at the time of admission) as SHS exposure may influence perceptions of general health and/or contribute to worse outcomes during periods of acute illness.

Covariates collected at the time of enrollment included sociodemographics (child age, gender, race/ethnicity, and caregiver education), and healthcare utilization (caregiver-reported patient visits to a healthcare provider in the preceding six months). Insurance status and level of medical complexity (using the Pediatric Medical Complexity Algorithm)¹⁸ were obtained using the Pediatric Hospital Information System database, an administrative database containing clinical and resource utilization data from >45 children’s hospitals in the United States including all of the PRIMES study hospitals.¹³

Analysis

Descriptive statistics included frequency (%) and mean (standard deviation). Bivariate comparisons according to SHS exposure status were analyzed using chi-squared tests for categorical variables and analysis of variance for continuous variables. Multivariable linear mixed regression models were used to examine associations between home SHS exposure and HRQOL for baseline health and during admission, overall and stratified by diagnosis. Covariates in each model included age, sex, race/ethnicity, caregiver education, and healthcare visits in the preceding six months. We also included a hospital random effect to account for clustering of patients within hospitals and used robust standard errors for inference.

In a secondary analysis to explore potential dose-response effects of SHS exposure, we examined associations between an ordinal exposure variable (0 smokers, 1 smoker, ≥2 smokers) and HRQOL for baseline health and during admission for the acute illness. Because of sample size limitations, diagnosis-specific analyses examining dose-response effects were not conducted.

Study Population

Of the 2,334 children enrolled in the PRIMES study, 25 (1%) respondents did not report on home SHS exposure and were excluded, yielding a final study population of 2,309 children, of whom 728 (32%) had reported home SHS exposure. The study population included 664 children with asthma (mean age seven years [3.5]; 38% with home SHS exposure), 740 with bronchiolitis (mean age 0.7 years [0.5]; 32% with home SHS exposure), 342 with croup (mean age 1.7 [1.1]; 25% with home SHS exposure), and 563 with pneumonia (mean age 4.4 [3.8]; 27% with home SHS exposure; Table 1). Compared with non-SHS-exposed children, those with home SHS exposure tend to be slightly older (3.9 vs 3.4 years, P = .01), more likely to be non-Hispanic Black (29% vs 19%, P < .001), to have a chronic condition (52% vs 41%, P < .001), to come from a household where caregiver(s) did not graduate from college (45% vs 29%, P < .001), and to have public insurance (73% vs 49%, P < .001).

Home SHS Exposure and Health-related Quality of Life

The overall mean HRQOL score for baseline health was 83 (15), with a range across diagnoses of 82 to 87. Compared with non-SHS-exposed children, children with home SHS exposure had a lower mean HRQOL score for baseline health (adjusted mean difference –3.04 [95% CI -4.34, –1.74]). In analyses stratified by diagnosis, baseline health scores were lower for SHS-exposed children for all four conditions, but differences were statistically significant only for bronchiolitis (adjusted mean difference –2.94 [–5.0, –0.89]) and pneumonia (adjusted mean value –4.13 [–6.82, –1.44]; Table 2); none of these differences met the MCID threshold.

The overall mean HRQOL score at the time of admission was 56 (23), with a range across diagnoses of 49 to 61, with lower scores noted among SHS-exposed children compared with non-SHS-exposed children (adjusted mean difference –2.16 [–4.22, –0.10]). Similar to scores representing baseline health, admission scores were lower across all four conditions for SHS-exposed children. Only children with croup, however, had significantly lower admission scores that also met the MCID threshold (adjusted mean difference –5.71 [–10.67, –0.75]; Table 2).

To assess for potential dose-response effects of SHS exposure on HRQOL, we stratified SHS-exposed children into those with one smoker in the home (n = 513) and those with ≥2 smokers in the home (n = 215). Compared with non-SHS-exposed children, both HRQOL scores (baseline health and admission) were lower for SHS-exposed children. Consistent with a dose-response association, scores were lowest for children with ≥2 smokers in the home, both at baseline health (adjusted mean difference –3.92 [–6.03, –1.81]) and on admission (adjusted mean difference –3.67 [–6.98, –0.36]; Table 3).

Within a multicenter cohort of 2,309 children hospitalized with ARI, we noted significantly lower HRQOL scores among children exposed to SHS in the home as compared with nonexposed children. Differences were greatest for children living with ≥2 smokers in the home. In analyses stratified by diagnosis, differences in baseline health HRQOL scores were greatest for children with bronchiolitis and pneumonia. Differences in acute illness scores were greatest for children with croup.¹⁶

Our study provides evidence for acute and chronic impacts of SHS on HRQOL in children hospitalized with ARI. Although several studies have linked SHS exposure to reduced HRQOL in adults,^19,20 few similar studies have been conducted in children. Nonetheless, a wealth of studies have documented the negative impact of SHS exposure on clinical outcomes among children with ARI.^8,10,21-23 Our findings that home SHS exposure was associated with reduced HRQOL among our cohort of children with ARI are therefore consistent with related findings in adults and children. The observation that the effects of SHS exposure on HRQOL were greatest among children living with ≥2 smokers provides further evidence of a potential causal link between regular SHS exposure and HRQOL.

Although the magnitude and significance of associations between SHS exposure and HRQOL varied for each of the four diagnoses for baseline health and the acute illness, it is important to note that the point estimates for the adjusted mean differences were uniformly lower for the SHS-exposed children in each subgroup. Even so, only acute illness scores for croup exceeded the MCID threshold.¹⁶ Croup is the only included condition of the upper airway and is characterized by laryngotracheal inflammation leading to the typical cough and, in moderate to severe cases, stridor. Given that chronic SHS exposure induces a proinflammatory state,³ it is possible that SHS-exposed children with croup had more severe illness compared with nonexposed children with croup resulting in lower HRQOL scores on admission. Further, perceived differences in illness severity and HRQOL may be more readily apparent in children with croup (eg, stridor at rest vs intermittent or no stridor) as compared with children with lower respiratory tract diseases.

Of the four included diagnoses, the link between SHS exposure and asthma outcomes has been most studied. Prior work has demonstrated more frequent and severe acute exacerbations, as well as worse long-term lung function among SHS-exposed children as compared with nonexposed children.^22-24 It was, therefore, surprising that our study failed to demonstrate associations between SHS exposure and HRQOL among children with asthma. Reasons for this finding are unclear. One hypothesis is that caregivers of SHS-exposed children with asthma may be more aware of the impacts of SHS exposure on respiratory health (through prior education) and, thus, more likely to modify their smoking behaviors, or for their children to be on daily asthma controller therapy. Alternatively, caregivers of children with asthma may be more likely to underreport home SHS exposure. Thirty-eight percent of children with asthma, however, were classified as SHS-exposed. This percentage was greater than the other three conditions studied (25%-32%), suggesting that differential bias in underreporting was minimal. Given that children with asthma were older, on average, than children with the other three conditions, it may also be that these children spent more time in smoke-free environments (eg, school).

Nearly one-third of children in our study were exposed to SHS in the home. This is similar to the prevalence of exposure in other studies conducted among hospitalized children^8,10,21,25 but higher than the national prevalence of home SHS exposure among children in the United States.²⁶ Thus, hospitalized children represent a particularly vulnerable population and an important target for interventions aiming to reduce exposure to SHS. Although longitudinal interventions are likely necessary to affect long-term success, hospitalization for ARI may serve as a powerful teachable moment to begin cessation efforts. Hospitalization also offers time beyond a typical primary care outpatient encounter to focus on cessation counseling and may be the only opportunity to engage in counseling activities for some families with limited time or access. Further, prior studies have demonstrated both the feasibility and the effectiveness of smoking cessation interventions in hospitalized children.^27-30 Unfortunately, however, SHS exposure is often not documented at the time of hospitalization, and many opportunities to intervene are missed.^25,31 Thus, there is a need for improved strategies to reliably identify and intervene on SHS-exposed children in the hospital setting.

These findings should be considered in the context of several limitations. The observational nature of our study raises the potential for confounding, specifically with regard to socioeconomic status, as this is associated with both SHS exposure and lower HRQOL. Our modeling approach attempted to control for several factors associated with socioeconomic status, including caregiver education and insurance coverage, but there is potential for residual confounding. No single question is sufficient to fully assess SHS exposure as the intensity of home SHS exposure likely varies widely, and some children may be exposed to SHS outside of the home environment.³² The home, however, is often the most likely source of regular SHS exposure,^33,34 especially among young children (our cohort’s mean age was 3.6 years). Misclassification of SHS exposure is also possible due to underreporting of smoking.^35,36 As a result, some children regularly exposed to SHS may have been misclassified as nonexposed, and the observed associations between SHS exposure and HRQOL may be underestimated. Confirming our study’s findings using objective assessments of SHS exposure, such as cotinine, are warranted. Given the young age of our cohort, the PedsQL surveys were completed by the parent or legal guardian only in >90% of the enrolled subjects, and caregiver perceptions may not accurately reflect the child’s perceptions. Prior work, however, has demonstrated the validity of parent-proxy reporting of the PedsQL, including correlation with child self-report.³⁷ In our study, correlation between child and caregiver reporting (when available) was also very good (r = 0.72, 95% CI 0.64, 0.77). It is also possible that the timing of the HRQOL assessments (on admission) may have biased perceptions of baseline HRQOL, although we anticipate any bias would likely be nondifferential between SHS-exposed and nonexposed children and across diagnoses.

Nearly one-third of children in our study were exposed to SHS exposure in the home, and SHS exposure was associated with lower HRQOL for baseline health and during acute illness, providing further evidence of the dangers of SHS. Much work is needed in order to eliminate the impact of SHS on child health and families of children hospitalized for respiratory illness should be considered a priority population for smoking cessation efforts.

Acknowledgment

The authors wish to acknowledge the efforts of PRIS-PRIMES study team. The authors also wish to thank the children and families who consented to be a part of the PRIMES study.

Disclosures

The authors have no conflicts of interest relevant to this article to disclose.

Funding

This study was supported by NIH-NHLBI 1R01HL121067 to RMS.

Internal Medicine (IM) residency graduates are expected to manage a wide range of acute clinical events.¹ Urgent and emergent inpatient situations require a broad knowledge base for rapid bedside diagnosis, yet the essential clinical skills required to manage acute clinical events pose a unique training challenge given the rarity and high-stakes nature of several such emergencies. For example, in three years of residency, a trainee may never have the opportunity to manage anaphylaxis, yet IM graduates must be able to recognize and quickly initiate proper lifesaving treatment for this relatively rare event² when it does occur.

In an era of work-hour limitations and heightened trainee supervision, residents perceive diminished familiarity with several clinical situations^3-5 and may feel unprepared to handle crisis events such as cardiac arrest.⁶ Given the sporadic nature of clinical medicine, many residents may not be exposed to certain acute inpatient clinical scenarios by the end of their training, a potentially critical education gap. To our knowledge, IM residents’ level of exposure to acute clinical events has not previously been studied. The aims of this study were to develop an instrument aimed at assessing IM residents’ exposure to hospital acute clinical events at a large academic medical center and to investigate the relationship between exposure and confidence in managing these events.

Survey Development

We reviewed the Massachusetts General Hospital (MGH) IM residency program curriculum (including simulation, conferences, and other didactics), the American Board of Internal Medicine certification requirements (primarily related to Advanced Cardiac Life Support [ACLS]), and the MGH inpatient rapid response events and gained input from the IM program leadership to develop a list of 50 acute clinical events that a graduating resident may be expected to manage independently (Box 1, Supplementary Appendix).^7-9 We then developed a survey assessing residents’ exposure to and confidence in managing such events. To classify the level of exposure, residents were asked to distinguish whether they had managed these events during a simulation session, inpatient as a part of a team, or inpatient independently. At our institution, IM postgraduate year 1 (PGY-1) interns manage a floor of patients overnight under a senior resident’s supervision, PGY-2 residents manage a team of several interns often without attending presence on ward rounds,¹⁰ and senior PGY-3 or -4 residents are expected to lead the hospital’s rapid response and code team and triage decompensating patients to the intensive care unit. Therefore, there are ample opportunities for IM residents to manage conditions independently (ie, in a direct leadership role) with attending supervision. House officers’ role in medical management, including calling appropriate subspecialty consultation, depends on the clinical condition; for example, a graduating senior resident would be expected to evaluate comprehensively a hypotensive patient and diagnose tension pneumothorax (while calling interventional pulmonary support for needle decompression and chest tube placement) and independently run an ACLS algorithm in the case of an unstable arrhythmia or cardiac arrest.

Residents were also asked to rate their perceived confidence in managing each condition independently on a five-point scale (ranging from “definitely cannot manage this condition independently” to “definitely can manage this condition independently”). We refined the survey instrument through a collaborative, iterative review process, including cognitive interviews and piloting with IM subspecialty fellows.

Participants and Data Collection

All IM residents at the Massachusetts General Hospital were invited to participate in the study. The study was conducted in May 2015 to reflect training throughout the prior academic year(s) and allow us to evaluate graduating residents’ exposures across all prior years of training. The instrument was administered anonymously via a web-based survey tool, Qualtrics (Provo, Utah). The study was approved as exempt by the Partners Institutional Review Board.

Data Analysis

Residents’ self-reported exposure to hospital acute events was classified into the following six ordinal categories: (1) never seen (have never seen the condition under any circumstances); (2) simulation alone (have managed the condition only during a mannequin-simulated patient case); (3) team alone (have managed the condition inpatient as a part of a team of providers, not in a primary leadership role); (4) team plus simulation; (5) independently (have managed the condition inpatient alone or in a primary leadership role); and (6) independently plus simulation. Residents’ self-reported exposure was examined for each postgraduate year (PGY) class both in aggregate and for each individual acute event. We sought to identify events that the majority of residents had managed independently (85% of residents or greater) and less common events that at least 15% of residents had never experienced.

We also examined residents’ self-reported confidence for each PGY class in aggregate and for each clinical acute scenario. Confidence was investigated in a dichotomized manner with a “definitely can” rating indicating “Confident” and with “probably can,” “neutral,” “probably cannot,” or “definitely cannot” ratings indicating “Not Confident” to manage the condition independently. Dichotomization thus allowed us to set a high bar for confidence, reflecting the self-perceived ability of the residents to manage the conditions as future independent physicians.

We used logistic regression models with the generalized estimating equations (GEE) approach to take into account the repeated measures of 50 clinical acute clinical events assessed for each resident. We compared the distribution of self-reported exposure and confidence among different PGY classes and examined the relationship between confidence and self-reported exposure stratified by level of training. We also assessed the independent effect of exposure on confidence controlling for level of training in a multivariable logistic regression model.

A total of 140 of 170 IM residents completed the survey (82% overall response rate: 72% of all PGY-1 residents, 86% of PGY-2 residents, and 89% of PGY-3/4 residents). In total, 41 PGY-1 residents (29% of respondents), 50 PGY-2 residents (36%), and 49 PGY-3 or PGY-4 residents (35%) participated. The majority of residents were in the Categorical IM training track (106 residents, 76% of respondents), whereas the remainder of respondents were in various subspecialty training tracks within our IM residency program, including Primary Care (14 residents, 10%), and four-year tracks, including Global Health (six residents, 4%), and Medicine-Pediatrics (14 residents, 10%).

Assessment of Exposure

Residents reported increasingly independent exposures as they progressed through residency training. PGY-1 residents on average had never seen 16.3% of the 50 acute events, whereas PGY-3/4 residents had never seen only 4.0% of the events (P < .0001). PGY-1 residents had managed 31.3% of events independently (or both independently and in simulation) as opposed to 71.7% of events for PGY-3/4 residents (P < .0001). Simulation alone accounted for a substantial proportion of exposures (16.4%) for PGY-1 residents, but this was significantly lower for PGY-2 or PGY-3/4 residents (P < .0001), who reported a greater percentage of exposures in nonsimulation clinical scenarios either independently or as a part of an inpatient team. There were no outlier residents who reported lower exposure compared with their PGY peers.

There was a wide spectrum of resident-reported exposures when individual acute events were examined (Table, full data in Supplementary Appendix Table 1). Events with the highest levels of exposure, which >85% of PGY-1 residents had managed independently, included alcohol withdrawal, chronic obstructive pulmonary disease exacerbation, rapid atrial fibrillation, agitated delirium, hypertensive urgency, and hyperkalemia. Events with the lowest levels of exposure, which at least 15% of graduating residents had never encountered in the hospital, included the following eight of 50 events (16%): torsades de pointes (51% of PGY-3/4 residents), acute mechanical valve failure (49%), tension pneumothorax (38.8%), use of emergency transcutaneous pacing (38.8%), elevated intracranial pressure (ICP)/herniation (24.5%), aortic dissection (22.4%), cord compression (16.3%), and use of emergency cardioversion (16.3%). Several PGY-3/4 residents had managed several of these events only in mannequin simulations, including torsades de pointes (41%), transcutaneous pacing (33%), and tension pneumothorax (24%).

Assessment of Confidence

Both levels of training and exposure to acute events were associated with increased confidence in managing such events. PGY-1 residents felt confident in managing 24.9% of acute events independently, compared to 48.4% of events for PGY-2 residents and 72.5% of events for PGY-3/4 residents (P < .0001). There was considerable variation in confidence among the individual acute events (Supplementary Appendix Table 2). A majority of graduating PGY-3/4 residents did not feel confident in managing the following 10 of the 50 events (20%): use of emergency cardioversion, aortic dissection, thrombotic thrombocytopenic purpura/hemolytic uremic syndrome (TTP/HUS), torsades de pointes, posterior reversible encephalopathy syndrome (PRES), intracranial hemorrhage, use of emergency transcutaneous pacing, tension pneumothorax, elevated ICP/herniation, and acute mechanical valve failure.

Residents’ self-reported confidence also correlated with level of exposure. There was a significant increase in resident confidence with increasingly independent exposure stratified by level of training (Figure; all with P < .0001). In the multivariable logistic regression model, increasing exposure correlated with increased resident confidence (P < .0001) while controlling for PGY year (P = .001).

We developed an instrument to assess resident exposure to and confidence in managing 50 inpatient acute clinical events. Both exposure and level of training were associated with increasing resident confidence. We identified specific events with low levels of exposure and confidence that could be targeted for educational interventions.

To our knowledge, this is the first study to examine IM residents’ exposure to and confidence in managing a wide range of inpatient acute clinical events. A primary goal of residency is to provide physicians-in-training graduated responsibility to prepare them for eventual independent practice. Although our survey confirmed that IM residents’ exposure and confidence significantly increased as they advanced through training (a not unexpected finding), our data also show that even after controlling for year in training, independent exposures significantly correlated with increased confidence. This speaks to the importance of preserving opportunities for residents to manage critical events in a supported manner, an admittedly challenging prospect given the oft-competing calls for supervision of and mentored feedback for trainees.¹¹

Despite identifying independent exposure as an important factor that impacts resident confidence, we found that there was still a substantial proportion of events (28.3%) that senior medical residents near the end of their training had not managed independently in a primary leadership role. Although our study was not designed to determine the reasons for this varied resident exposure, possible explanations may include the relative rarity of certain acute clinical events compared with others, or less likely the effect of duty hour limitations, attending supervision of trainees, or programmatic changes in resident leadership responsibilities. Whatever the cause, this finding uniquely identifies an area for improvement to prevent new attending physicians from feeling unprepared to manage potentially critical emergencies.

An important goal of our study was to develop an instrument that would enable training programs to identify their learning needs. Both program-wide and individual assessments of resident case exposure and confidence are essential for identifying such learning needs and areas for curricular development. Program-wide assessments can spur an important debate about program goals and requirements with respect to what scenarios residents must be able to manage competently by graduation.¹² In addition, such assessments can help individualize learning exposures based on a specific learner’s needs and career goals. The administration of our survey instrument required minimal resources, and the high response rate in our study suggests that other programs can implement our instrument to accomplish these goals.

Alternative methods, such as electronic learning portfolios (efolios), can be utilized to assess resident case exposure. In comparison to our survey instrument, efolios limit recall bias by utilizing case logs and have additional capabilities such as compiling evaluations and enabling trainees to set learning goals. However, there are considerable barriers to the effective use of efolios, including software cost, learner attitudes, and time constraints.¹³ Tools such as our end-of-year assessment offer an alternative method that limits these barriers.

Once educational growth opportunities have been identified through survey-based or other methods, residency programs must determine how to optimize curricula for the needs and career goals of their trainees. We found considerable overlap among conditions that graduating residents had both limited exposure to and low confidence in managing (eg, torsades de pointes, tension pneumothorax, and emergency cardioversion), which are logical topics for future curriculum development. We also identified a few conditions (including PRES, TTP/HUS, and intracranial hemorrhage) that graduating residents did not feel confident in managing despite a relatively higher reported level of exposure. Whether to focus specific educational interventions on the most rare or most commonly encountered acute clinical events is likely to be a topic of debate among individual training programs, but the results of our survey indicate that there is likely to be educational benefit to both strategies.

Residency programs can employ a variety of modalities to enhance learner exposure and confidence in managing clinical scenarios that are deemed important by the program, including didactics, simulation, and changes in program structure. There is a substantial literature on the use of dedicated curricula for crisis management and the use of simulation as a training tool for responding to acute clinical events in multiple specialties^14-24 and in nonmedical domains such as aviation.^25-27 Simulation has been shown to improve residents’ clinical skills and comfort level with some acute events^28-30 and may even be superior to traditional clinical medical education.³¹ In addition, programs can utilize targeted clinical experiences such as intensive care unit and subspecialty rotations^32,33 in an effort to customize educational interventions to fill identified gaps in learner exposure or confidence.

Our study has several limitations. First, we investigated a single large IM residency program at a quaternary academic medical center, and therefore, our findings may not be externally generalizable to all IM residencies or other medical specialties. Our unique peer-led simulation curriculum, including 16 PGY-1 and 8 PGY-2 cases chosen based on clinical rotations at Massachusetts General Hospital,⁷ likely impacted residents’ exposure to simulation that is specific to our institution. However, although specific inpatient acute events may vary among other institutions, our finding that graduating residents still reported gaps in their clinical experience is likely generalizable to other programs given the varied and unpredictable nature of ward medicine training. In addition, our survey tool was simple to administer and could be tailored to reflect the acute events and training needs relevant to other residency programs, specialties, and institutions. Second, the retrospective nature of our study may be subject to participants’ recall bias. We did not restrict our survey questions to urgent conditions managed only on IM hospital wards and some may have been experienced in the emergency room or intensive care units; however, these exposures are still relevant as key components of IM training. Third, our list of 50 acute clinical events was intentionally broad and included several conditions that require multidisciplinary subspecialist consultation, which could have impacted residents’ self-report of “independent” exposures. However, these scenarios are ones that hospitalists may independently recognize and stabilize, engaging appropriate specialists. Fourth, we were not able to validate residents’ self-reported exposures against other measures of the frequency of housestaff management of acute events (such as billing data or patient logs) as this information is not routinely collected. We also did not attempt to identify the reasons underlying the variation seen in resident exposure and confidence for individual acute events, but as a needs assessment, this was beyond the scope of our study. Finally, our assessment of resident confidence was subjective and we were not able to assess competence, with prior studies demonstrating conflicting results regarding the relationship between self-reported proficiency and observed competence.^34-36 Future studies are needed to investigate whether case exposure assessment leads to changes in residency curricula and whether such curricula increase resident confidence and competence in managing hospital acute clinical events.

We developed an easy-to-administer tool to assess IM residents’ exposure to and confidence in managing inpatient acute events. We found that both significantly increased as residents advanced through training, and self-reported confidence additionally correlated with level of exposure independent of PGY class. We identified several specific inpatient acute clinical events with low levels of resident exposure and confidence that can serve as targets for future IM residency curriculum development. Future studies assessing the impact of such curricula on resident confidence and competence are needed.

Disclosures

The authors declare no conflict of interest.

Internal Medicine (IM) residency graduates are expected to manage a wide range of acute clinical events.¹ Urgent and emergent inpatient situations require a broad knowledge base for rapid bedside diagnosis, yet the essential clinical skills required to manage acute clinical events pose a unique training challenge given the rarity and high-stakes nature of several such emergencies. For example, in three years of residency, a trainee may never have the opportunity to manage anaphylaxis, yet IM graduates must be able to recognize and quickly initiate proper lifesaving treatment for this relatively rare event² when it does occur.

In an era of work-hour limitations and heightened trainee supervision, residents perceive diminished familiarity with several clinical situations^3-5 and may feel unprepared to handle crisis events such as cardiac arrest.⁶ Given the sporadic nature of clinical medicine, many residents may not be exposed to certain acute inpatient clinical scenarios by the end of their training, a potentially critical education gap. To our knowledge, IM residents’ level of exposure to acute clinical events has not previously been studied. The aims of this study were to develop an instrument aimed at assessing IM residents’ exposure to hospital acute clinical events at a large academic medical center and to investigate the relationship between exposure and confidence in managing these events.

Survey Development

We reviewed the Massachusetts General Hospital (MGH) IM residency program curriculum (including simulation, conferences, and other didactics), the American Board of Internal Medicine certification requirements (primarily related to Advanced Cardiac Life Support [ACLS]), and the MGH inpatient rapid response events and gained input from the IM program leadership to develop a list of 50 acute clinical events that a graduating resident may be expected to manage independently (Box 1, Supplementary Appendix).^7-9 We then developed a survey assessing residents’ exposure to and confidence in managing such events. To classify the level of exposure, residents were asked to distinguish whether they had managed these events during a simulation session, inpatient as a part of a team, or inpatient independently. At our institution, IM postgraduate year 1 (PGY-1) interns manage a floor of patients overnight under a senior resident’s supervision, PGY-2 residents manage a team of several interns often without attending presence on ward rounds,¹⁰ and senior PGY-3 or -4 residents are expected to lead the hospital’s rapid response and code team and triage decompensating patients to the intensive care unit. Therefore, there are ample opportunities for IM residents to manage conditions independently (ie, in a direct leadership role) with attending supervision. House officers’ role in medical management, including calling appropriate subspecialty consultation, depends on the clinical condition; for example, a graduating senior resident would be expected to evaluate comprehensively a hypotensive patient and diagnose tension pneumothorax (while calling interventional pulmonary support for needle decompression and chest tube placement) and independently run an ACLS algorithm in the case of an unstable arrhythmia or cardiac arrest.

Residents were also asked to rate their perceived confidence in managing each condition independently on a five-point scale (ranging from “definitely cannot manage this condition independently” to “definitely can manage this condition independently”). We refined the survey instrument through a collaborative, iterative review process, including cognitive interviews and piloting with IM subspecialty fellows.

Participants and Data Collection

All IM residents at the Massachusetts General Hospital were invited to participate in the study. The study was conducted in May 2015 to reflect training throughout the prior academic year(s) and allow us to evaluate graduating residents’ exposures across all prior years of training. The instrument was administered anonymously via a web-based survey tool, Qualtrics (Provo, Utah). The study was approved as exempt by the Partners Institutional Review Board.

Data Analysis

Residents’ self-reported exposure to hospital acute events was classified into the following six ordinal categories: (1) never seen (have never seen the condition under any circumstances); (2) simulation alone (have managed the condition only during a mannequin-simulated patient case); (3) team alone (have managed the condition inpatient as a part of a team of providers, not in a primary leadership role); (4) team plus simulation; (5) independently (have managed the condition inpatient alone or in a primary leadership role); and (6) independently plus simulation. Residents’ self-reported exposure was examined for each postgraduate year (PGY) class both in aggregate and for each individual acute event. We sought to identify events that the majority of residents had managed independently (85% of residents or greater) and less common events that at least 15% of residents had never experienced.

We also examined residents’ self-reported confidence for each PGY class in aggregate and for each clinical acute scenario. Confidence was investigated in a dichotomized manner with a “definitely can” rating indicating “Confident” and with “probably can,” “neutral,” “probably cannot,” or “definitely cannot” ratings indicating “Not Confident” to manage the condition independently. Dichotomization thus allowed us to set a high bar for confidence, reflecting the self-perceived ability of the residents to manage the conditions as future independent physicians.

We used logistic regression models with the generalized estimating equations (GEE) approach to take into account the repeated measures of 50 clinical acute clinical events assessed for each resident. We compared the distribution of self-reported exposure and confidence among different PGY classes and examined the relationship between confidence and self-reported exposure stratified by level of training. We also assessed the independent effect of exposure on confidence controlling for level of training in a multivariable logistic regression model.

A total of 140 of 170 IM residents completed the survey (82% overall response rate: 72% of all PGY-1 residents, 86% of PGY-2 residents, and 89% of PGY-3/4 residents). In total, 41 PGY-1 residents (29% of respondents), 50 PGY-2 residents (36%), and 49 PGY-3 or PGY-4 residents (35%) participated. The majority of residents were in the Categorical IM training track (106 residents, 76% of respondents), whereas the remainder of respondents were in various subspecialty training tracks within our IM residency program, including Primary Care (14 residents, 10%), and four-year tracks, including Global Health (six residents, 4%), and Medicine-Pediatrics (14 residents, 10%).

Assessment of Exposure

Residents reported increasingly independent exposures as they progressed through residency training. PGY-1 residents on average had never seen 16.3% of the 50 acute events, whereas PGY-3/4 residents had never seen only 4.0% of the events (P < .0001). PGY-1 residents had managed 31.3% of events independently (or both independently and in simulation) as opposed to 71.7% of events for PGY-3/4 residents (P < .0001). Simulation alone accounted for a substantial proportion of exposures (16.4%) for PGY-1 residents, but this was significantly lower for PGY-2 or PGY-3/4 residents (P < .0001), who reported a greater percentage of exposures in nonsimulation clinical scenarios either independently or as a part of an inpatient team. There were no outlier residents who reported lower exposure compared with their PGY peers.

There was a wide spectrum of resident-reported exposures when individual acute events were examined (Table, full data in Supplementary Appendix Table 1). Events with the highest levels of exposure, which >85% of PGY-1 residents had managed independently, included alcohol withdrawal, chronic obstructive pulmonary disease exacerbation, rapid atrial fibrillation, agitated delirium, hypertensive urgency, and hyperkalemia. Events with the lowest levels of exposure, which at least 15% of graduating residents had never encountered in the hospital, included the following eight of 50 events (16%): torsades de pointes (51% of PGY-3/4 residents), acute mechanical valve failure (49%), tension pneumothorax (38.8%), use of emergency transcutaneous pacing (38.8%), elevated intracranial pressure (ICP)/herniation (24.5%), aortic dissection (22.4%), cord compression (16.3%), and use of emergency cardioversion (16.3%). Several PGY-3/4 residents had managed several of these events only in mannequin simulations, including torsades de pointes (41%), transcutaneous pacing (33%), and tension pneumothorax (24%).

Assessment of Confidence

Both levels of training and exposure to acute events were associated with increased confidence in managing such events. PGY-1 residents felt confident in managing 24.9% of acute events independently, compared to 48.4% of events for PGY-2 residents and 72.5% of events for PGY-3/4 residents (P < .0001). There was considerable variation in confidence among the individual acute events (Supplementary Appendix Table 2). A majority of graduating PGY-3/4 residents did not feel confident in managing the following 10 of the 50 events (20%): use of emergency cardioversion, aortic dissection, thrombotic thrombocytopenic purpura/hemolytic uremic syndrome (TTP/HUS), torsades de pointes, posterior reversible encephalopathy syndrome (PRES), intracranial hemorrhage, use of emergency transcutaneous pacing, tension pneumothorax, elevated ICP/herniation, and acute mechanical valve failure.

Residents’ self-reported confidence also correlated with level of exposure. There was a significant increase in resident confidence with increasingly independent exposure stratified by level of training (Figure; all with P < .0001). In the multivariable logistic regression model, increasing exposure correlated with increased resident confidence (P < .0001) while controlling for PGY year (P = .001).

We developed an instrument to assess resident exposure to and confidence in managing 50 inpatient acute clinical events. Both exposure and level of training were associated with increasing resident confidence. We identified specific events with low levels of exposure and confidence that could be targeted for educational interventions.

To our knowledge, this is the first study to examine IM residents’ exposure to and confidence in managing a wide range of inpatient acute clinical events. A primary goal of residency is to provide physicians-in-training graduated responsibility to prepare them for eventual independent practice. Although our survey confirmed that IM residents’ exposure and confidence significantly increased as they advanced through training (a not unexpected finding), our data also show that even after controlling for year in training, independent exposures significantly correlated with increased confidence. This speaks to the importance of preserving opportunities for residents to manage critical events in a supported manner, an admittedly challenging prospect given the oft-competing calls for supervision of and mentored feedback for trainees.¹¹

Despite identifying independent exposure as an important factor that impacts resident confidence, we found that there was still a substantial proportion of events (28.3%) that senior medical residents near the end of their training had not managed independently in a primary leadership role. Although our study was not designed to determine the reasons for this varied resident exposure, possible explanations may include the relative rarity of certain acute clinical events compared with others, or less likely the effect of duty hour limitations, attending supervision of trainees, or programmatic changes in resident leadership responsibilities. Whatever the cause, this finding uniquely identifies an area for improvement to prevent new attending physicians from feeling unprepared to manage potentially critical emergencies.

An important goal of our study was to develop an instrument that would enable training programs to identify their learning needs. Both program-wide and individual assessments of resident case exposure and confidence are essential for identifying such learning needs and areas for curricular development. Program-wide assessments can spur an important debate about program goals and requirements with respect to what scenarios residents must be able to manage competently by graduation.¹² In addition, such assessments can help individualize learning exposures based on a specific learner’s needs and career goals. The administration of our survey instrument required minimal resources, and the high response rate in our study suggests that other programs can implement our instrument to accomplish these goals.

Alternative methods, such as electronic learning portfolios (efolios), can be utilized to assess resident case exposure. In comparison to our survey instrument, efolios limit recall bias by utilizing case logs and have additional capabilities such as compiling evaluations and enabling trainees to set learning goals. However, there are considerable barriers to the effective use of efolios, including software cost, learner attitudes, and time constraints.¹³ Tools such as our end-of-year assessment offer an alternative method that limits these barriers.

Once educational growth opportunities have been identified through survey-based or other methods, residency programs must determine how to optimize curricula for the needs and career goals of their trainees. We found considerable overlap among conditions that graduating residents had both limited exposure to and low confidence in managing (eg, torsades de pointes, tension pneumothorax, and emergency cardioversion), which are logical topics for future curriculum development. We also identified a few conditions (including PRES, TTP/HUS, and intracranial hemorrhage) that graduating residents did not feel confident in managing despite a relatively higher reported level of exposure. Whether to focus specific educational interventions on the most rare or most commonly encountered acute clinical events is likely to be a topic of debate among individual training programs, but the results of our survey indicate that there is likely to be educational benefit to both strategies.

Residency programs can employ a variety of modalities to enhance learner exposure and confidence in managing clinical scenarios that are deemed important by the program, including didactics, simulation, and changes in program structure. There is a substantial literature on the use of dedicated curricula for crisis management and the use of simulation as a training tool for responding to acute clinical events in multiple specialties^14-24 and in nonmedical domains such as aviation.^25-27 Simulation has been shown to improve residents’ clinical skills and comfort level with some acute events^28-30 and may even be superior to traditional clinical medical education.³¹ In addition, programs can utilize targeted clinical experiences such as intensive care unit and subspecialty rotations^32,33 in an effort to customize educational interventions to fill identified gaps in learner exposure or confidence.

Our study has several limitations. First, we investigated a single large IM residency program at a quaternary academic medical center, and therefore, our findings may not be externally generalizable to all IM residencies or other medical specialties. Our unique peer-led simulation curriculum, including 16 PGY-1 and 8 PGY-2 cases chosen based on clinical rotations at Massachusetts General Hospital,⁷ likely impacted residents’ exposure to simulation that is specific to our institution. However, although specific inpatient acute events may vary among other institutions, our finding that graduating residents still reported gaps in their clinical experience is likely generalizable to other programs given the varied and unpredictable nature of ward medicine training. In addition, our survey tool was simple to administer and could be tailored to reflect the acute events and training needs relevant to other residency programs, specialties, and institutions. Second, the retrospective nature of our study may be subject to participants’ recall bias. We did not restrict our survey questions to urgent conditions managed only on IM hospital wards and some may have been experienced in the emergency room or intensive care units; however, these exposures are still relevant as key components of IM training. Third, our list of 50 acute clinical events was intentionally broad and included several conditions that require multidisciplinary subspecialist consultation, which could have impacted residents’ self-report of “independent” exposures. However, these scenarios are ones that hospitalists may independently recognize and stabilize, engaging appropriate specialists. Fourth, we were not able to validate residents’ self-reported exposures against other measures of the frequency of housestaff management of acute events (such as billing data or patient logs) as this information is not routinely collected. We also did not attempt to identify the reasons underlying the variation seen in resident exposure and confidence for individual acute events, but as a needs assessment, this was beyond the scope of our study. Finally, our assessment of resident confidence was subjective and we were not able to assess competence, with prior studies demonstrating conflicting results regarding the relationship between self-reported proficiency and observed competence.^34-36 Future studies are needed to investigate whether case exposure assessment leads to changes in residency curricula and whether such curricula increase resident confidence and competence in managing hospital acute clinical events.

We developed an easy-to-administer tool to assess IM residents’ exposure to and confidence in managing inpatient acute events. We found that both significantly increased as residents advanced through training, and self-reported confidence additionally correlated with level of exposure independent of PGY class. We identified several specific inpatient acute clinical events with low levels of resident exposure and confidence that can serve as targets for future IM residency curriculum development. Future studies assessing the impact of such curricula on resident confidence and competence are needed.

Disclosures

The authors declare no conflict of interest.

Internal Medicine (IM) residency graduates are expected to manage a wide range of acute clinical events.¹ Urgent and emergent inpatient situations require a broad knowledge base for rapid bedside diagnosis, yet the essential clinical skills required to manage acute clinical events pose a unique training challenge given the rarity and high-stakes nature of several such emergencies. For example, in three years of residency, a trainee may never have the opportunity to manage anaphylaxis, yet IM graduates must be able to recognize and quickly initiate proper lifesaving treatment for this relatively rare event² when it does occur.

In an era of work-hour limitations and heightened trainee supervision, residents perceive diminished familiarity with several clinical situations^3-5 and may feel unprepared to handle crisis events such as cardiac arrest.⁶ Given the sporadic nature of clinical medicine, many residents may not be exposed to certain acute inpatient clinical scenarios by the end of their training, a potentially critical education gap. To our knowledge, IM residents’ level of exposure to acute clinical events has not previously been studied. The aims of this study were to develop an instrument aimed at assessing IM residents’ exposure to hospital acute clinical events at a large academic medical center and to investigate the relationship between exposure and confidence in managing these events.

Survey Development

We reviewed the Massachusetts General Hospital (MGH) IM residency program curriculum (including simulation, conferences, and other didactics), the American Board of Internal Medicine certification requirements (primarily related to Advanced Cardiac Life Support [ACLS]), and the MGH inpatient rapid response events and gained input from the IM program leadership to develop a list of 50 acute clinical events that a graduating resident may be expected to manage independently (Box 1, Supplementary Appendix).^7-9 We then developed a survey assessing residents’ exposure to and confidence in managing such events. To classify the level of exposure, residents were asked to distinguish whether they had managed these events during a simulation session, inpatient as a part of a team, or inpatient independently. At our institution, IM postgraduate year 1 (PGY-1) interns manage a floor of patients overnight under a senior resident’s supervision, PGY-2 residents manage a team of several interns often without attending presence on ward rounds,¹⁰ and senior PGY-3 or -4 residents are expected to lead the hospital’s rapid response and code team and triage decompensating patients to the intensive care unit. Therefore, there are ample opportunities for IM residents to manage conditions independently (ie, in a direct leadership role) with attending supervision. House officers’ role in medical management, including calling appropriate subspecialty consultation, depends on the clinical condition; for example, a graduating senior resident would be expected to evaluate comprehensively a hypotensive patient and diagnose tension pneumothorax (while calling interventional pulmonary support for needle decompression and chest tube placement) and independently run an ACLS algorithm in the case of an unstable arrhythmia or cardiac arrest.

Residents were also asked to rate their perceived confidence in managing each condition independently on a five-point scale (ranging from “definitely cannot manage this condition independently” to “definitely can manage this condition independently”). We refined the survey instrument through a collaborative, iterative review process, including cognitive interviews and piloting with IM subspecialty fellows.

Participants and Data Collection

All IM residents at the Massachusetts General Hospital were invited to participate in the study. The study was conducted in May 2015 to reflect training throughout the prior academic year(s) and allow us to evaluate graduating residents’ exposures across all prior years of training. The instrument was administered anonymously via a web-based survey tool, Qualtrics (Provo, Utah). The study was approved as exempt by the Partners Institutional Review Board.

Data Analysis

Residents’ self-reported exposure to hospital acute events was classified into the following six ordinal categories: (1) never seen (have never seen the condition under any circumstances); (2) simulation alone (have managed the condition only during a mannequin-simulated patient case); (3) team alone (have managed the condition inpatient as a part of a team of providers, not in a primary leadership role); (4) team plus simulation; (5) independently (have managed the condition inpatient alone or in a primary leadership role); and (6) independently plus simulation. Residents’ self-reported exposure was examined for each postgraduate year (PGY) class both in aggregate and for each individual acute event. We sought to identify events that the majority of residents had managed independently (85% of residents or greater) and less common events that at least 15% of residents had never experienced.

We also examined residents’ self-reported confidence for each PGY class in aggregate and for each clinical acute scenario. Confidence was investigated in a dichotomized manner with a “definitely can” rating indicating “Confident” and with “probably can,” “neutral,” “probably cannot,” or “definitely cannot” ratings indicating “Not Confident” to manage the condition independently. Dichotomization thus allowed us to set a high bar for confidence, reflecting the self-perceived ability of the residents to manage the conditions as future independent physicians.

We used logistic regression models with the generalized estimating equations (GEE) approach to take into account the repeated measures of 50 clinical acute clinical events assessed for each resident. We compared the distribution of self-reported exposure and confidence among different PGY classes and examined the relationship between confidence and self-reported exposure stratified by level of training. We also assessed the independent effect of exposure on confidence controlling for level of training in a multivariable logistic regression model.

A total of 140 of 170 IM residents completed the survey (82% overall response rate: 72% of all PGY-1 residents, 86% of PGY-2 residents, and 89% of PGY-3/4 residents). In total, 41 PGY-1 residents (29% of respondents), 50 PGY-2 residents (36%), and 49 PGY-3 or PGY-4 residents (35%) participated. The majority of residents were in the Categorical IM training track (106 residents, 76% of respondents), whereas the remainder of respondents were in various subspecialty training tracks within our IM residency program, including Primary Care (14 residents, 10%), and four-year tracks, including Global Health (six residents, 4%), and Medicine-Pediatrics (14 residents, 10%).

Assessment of Exposure

Residents reported increasingly independent exposures as they progressed through residency training. PGY-1 residents on average had never seen 16.3% of the 50 acute events, whereas PGY-3/4 residents had never seen only 4.0% of the events (P < .0001). PGY-1 residents had managed 31.3% of events independently (or both independently and in simulation) as opposed to 71.7% of events for PGY-3/4 residents (P < .0001). Simulation alone accounted for a substantial proportion of exposures (16.4%) for PGY-1 residents, but this was significantly lower for PGY-2 or PGY-3/4 residents (P < .0001), who reported a greater percentage of exposures in nonsimulation clinical scenarios either independently or as a part of an inpatient team. There were no outlier residents who reported lower exposure compared with their PGY peers.

There was a wide spectrum of resident-reported exposures when individual acute events were examined (Table, full data in Supplementary Appendix Table 1). Events with the highest levels of exposure, which >85% of PGY-1 residents had managed independently, included alcohol withdrawal, chronic obstructive pulmonary disease exacerbation, rapid atrial fibrillation, agitated delirium, hypertensive urgency, and hyperkalemia. Events with the lowest levels of exposure, which at least 15% of graduating residents had never encountered in the hospital, included the following eight of 50 events (16%): torsades de pointes (51% of PGY-3/4 residents), acute mechanical valve failure (49%), tension pneumothorax (38.8%), use of emergency transcutaneous pacing (38.8%), elevated intracranial pressure (ICP)/herniation (24.5%), aortic dissection (22.4%), cord compression (16.3%), and use of emergency cardioversion (16.3%). Several PGY-3/4 residents had managed several of these events only in mannequin simulations, including torsades de pointes (41%), transcutaneous pacing (33%), and tension pneumothorax (24%).

Assessment of Confidence

Both levels of training and exposure to acute events were associated with increased confidence in managing such events. PGY-1 residents felt confident in managing 24.9% of acute events independently, compared to 48.4% of events for PGY-2 residents and 72.5% of events for PGY-3/4 residents (P < .0001). There was considerable variation in confidence among the individual acute events (Supplementary Appendix Table 2). A majority of graduating PGY-3/4 residents did not feel confident in managing the following 10 of the 50 events (20%): use of emergency cardioversion, aortic dissection, thrombotic thrombocytopenic purpura/hemolytic uremic syndrome (TTP/HUS), torsades de pointes, posterior reversible encephalopathy syndrome (PRES), intracranial hemorrhage, use of emergency transcutaneous pacing, tension pneumothorax, elevated ICP/herniation, and acute mechanical valve failure.

Residents’ self-reported confidence also correlated with level of exposure. There was a significant increase in resident confidence with increasingly independent exposure stratified by level of training (Figure; all with P < .0001). In the multivariable logistic regression model, increasing exposure correlated with increased resident confidence (P < .0001) while controlling for PGY year (P = .001).

We developed an instrument to assess resident exposure to and confidence in managing 50 inpatient acute clinical events. Both exposure and level of training were associated with increasing resident confidence. We identified specific events with low levels of exposure and confidence that could be targeted for educational interventions.

To our knowledge, this is the first study to examine IM residents’ exposure to and confidence in managing a wide range of inpatient acute clinical events. A primary goal of residency is to provide physicians-in-training graduated responsibility to prepare them for eventual independent practice. Although our survey confirmed that IM residents’ exposure and confidence significantly increased as they advanced through training (a not unexpected finding), our data also show that even after controlling for year in training, independent exposures significantly correlated with increased confidence. This speaks to the importance of preserving opportunities for residents to manage critical events in a supported manner, an admittedly challenging prospect given the oft-competing calls for supervision of and mentored feedback for trainees.¹¹

Despite identifying independent exposure as an important factor that impacts resident confidence, we found that there was still a substantial proportion of events (28.3%) that senior medical residents near the end of their training had not managed independently in a primary leadership role. Although our study was not designed to determine the reasons for this varied resident exposure, possible explanations may include the relative rarity of certain acute clinical events compared with others, or less likely the effect of duty hour limitations, attending supervision of trainees, or programmatic changes in resident leadership responsibilities. Whatever the cause, this finding uniquely identifies an area for improvement to prevent new attending physicians from feeling unprepared to manage potentially critical emergencies.

An important goal of our study was to develop an instrument that would enable training programs to identify their learning needs. Both program-wide and individual assessments of resident case exposure and confidence are essential for identifying such learning needs and areas for curricular development. Program-wide assessments can spur an important debate about program goals and requirements with respect to what scenarios residents must be able to manage competently by graduation.¹² In addition, such assessments can help individualize learning exposures based on a specific learner’s needs and career goals. The administration of our survey instrument required minimal resources, and the high response rate in our study suggests that other programs can implement our instrument to accomplish these goals.

Alternative methods, such as electronic learning portfolios (efolios), can be utilized to assess resident case exposure. In comparison to our survey instrument, efolios limit recall bias by utilizing case logs and have additional capabilities such as compiling evaluations and enabling trainees to set learning goals. However, there are considerable barriers to the effective use of efolios, including software cost, learner attitudes, and time constraints.¹³ Tools such as our end-of-year assessment offer an alternative method that limits these barriers.

Once educational growth opportunities have been identified through survey-based or other methods, residency programs must determine how to optimize curricula for the needs and career goals of their trainees. We found considerable overlap among conditions that graduating residents had both limited exposure to and low confidence in managing (eg, torsades de pointes, tension pneumothorax, and emergency cardioversion), which are logical topics for future curriculum development. We also identified a few conditions (including PRES, TTP/HUS, and intracranial hemorrhage) that graduating residents did not feel confident in managing despite a relatively higher reported level of exposure. Whether to focus specific educational interventions on the most rare or most commonly encountered acute clinical events is likely to be a topic of debate among individual training programs, but the results of our survey indicate that there is likely to be educational benefit to both strategies.

Residency programs can employ a variety of modalities to enhance learner exposure and confidence in managing clinical scenarios that are deemed important by the program, including didactics, simulation, and changes in program structure. There is a substantial literature on the use of dedicated curricula for crisis management and the use of simulation as a training tool for responding to acute clinical events in multiple specialties^14-24 and in nonmedical domains such as aviation.^25-27 Simulation has been shown to improve residents’ clinical skills and comfort level with some acute events^28-30 and may even be superior to traditional clinical medical education.³¹ In addition, programs can utilize targeted clinical experiences such as intensive care unit and subspecialty rotations^32,33 in an effort to customize educational interventions to fill identified gaps in learner exposure or confidence.

Our study has several limitations. First, we investigated a single large IM residency program at a quaternary academic medical center, and therefore, our findings may not be externally generalizable to all IM residencies or other medical specialties. Our unique peer-led simulation curriculum, including 16 PGY-1 and 8 PGY-2 cases chosen based on clinical rotations at Massachusetts General Hospital,⁷ likely impacted residents’ exposure to simulation that is specific to our institution. However, although specific inpatient acute events may vary among other institutions, our finding that graduating residents still reported gaps in their clinical experience is likely generalizable to other programs given the varied and unpredictable nature of ward medicine training. In addition, our survey tool was simple to administer and could be tailored to reflect the acute events and training needs relevant to other residency programs, specialties, and institutions. Second, the retrospective nature of our study may be subject to participants’ recall bias. We did not restrict our survey questions to urgent conditions managed only on IM hospital wards and some may have been experienced in the emergency room or intensive care units; however, these exposures are still relevant as key components of IM training. Third, our list of 50 acute clinical events was intentionally broad and included several conditions that require multidisciplinary subspecialist consultation, which could have impacted residents’ self-report of “independent” exposures. However, these scenarios are ones that hospitalists may independently recognize and stabilize, engaging appropriate specialists. Fourth, we were not able to validate residents’ self-reported exposures against other measures of the frequency of housestaff management of acute events (such as billing data or patient logs) as this information is not routinely collected. We also did not attempt to identify the reasons underlying the variation seen in resident exposure and confidence for individual acute events, but as a needs assessment, this was beyond the scope of our study. Finally, our assessment of resident confidence was subjective and we were not able to assess competence, with prior studies demonstrating conflicting results regarding the relationship between self-reported proficiency and observed competence.^34-36 Future studies are needed to investigate whether case exposure assessment leads to changes in residency curricula and whether such curricula increase resident confidence and competence in managing hospital acute clinical events.

We developed an easy-to-administer tool to assess IM residents’ exposure to and confidence in managing inpatient acute events. We found that both significantly increased as residents advanced through training, and self-reported confidence additionally correlated with level of exposure independent of PGY class. We identified several specific inpatient acute clinical events with low levels of resident exposure and confidence that can serve as targets for future IM residency curriculum development. Future studies assessing the impact of such curricula on resident confidence and competence are needed.

Disclosures

The authors declare no conflict of interest.

Graduate medical education (GME) is heavily reliant on experiential learning. Most of a resident’s time is spent in progressively independent delivery of patient care, which is associated with decreasing supervision. Attainment and demonstration of competence in patient care is the goal and responsibility of GME training programs. What happens, then, if the medicine resident never has the experience necessary to enable experiential learning? What if she never “sees one,” let alone “does one”?

In this month’s Journal of Hospital Medicine, Sclafani et al¹ examine how exposure to urgent clinical situations impacts residents’ confidence in managing these ward emergencies. They astutely reveal the idiosyncratic nature of residency training and consequent gaps created when an educational delivery model predicated on experience lacks certain experiences. How can a resident without certain key experiences be ready for independent practice?

The ACGME’s Next Accreditation System is intended to ensure that residents are prepared for independent practice. The educational outcomes that learners must attain are comprised of six core competencies, with milestones intended to operationalize the measurement and reporting of learner progression toward competence.^2,3 It is challenging to apply general competencies to assessment of day to day clinical activities. This challenge led to the development of 16 Entrustable Professional Activities (EPAs). These allow the direct observation of concrete clinical activities that could then infer the attainment (or not) of multiple competencies. Ideally, EPAs are paired with and mapped to curricular milestones which describe a learner’s trajectory within the framework of competencies and determine if a resident is prepared for independent practice.^4,5

In Sclafani et al.¹ the authors characterize resident exposure to, and confidence in, 50 urgent clinical situations. Both level of training and exposure were associated with increased confidence. However, the most important finding of this paper is the wide variation of resident exposures and confidence with respect to specific urgent clinical events. At least 15% of graduating residents had never seen 16% of the 50 emergency events, and a majority of graduating residents did not feel confident managing 20% of the 50 events, highlighting the idiosyncratic nature of GME training.¹ Of course, while certain entities on the list of clinical emergencies were not identified as final diagnoses, it is possible they were still considered in the process of caring for patients in different situations.

Several factors account for the idiosyncratic nature of medical training, including the rarity of certain clinical events, seasonal variation in conditions, and other variables (ie, learner elective choices). In addition, the scheduling of most residency programs is based on patient care needs instead of individual trainees’ educational needs. Other areas of medicine have attempted to standardize experience and ensure specific exposure and/or competence using strategies such as surgical case logs and case-based certifying examinations. There are very important recently described projects in undergraduate medical education aimed at using longitudinal assessment of EPAs in multiple contexts to make entrustment decisions.⁶ However, Internal Medicine residencies do not routinely employ these strategies.

It must be noted that Sclafani et al. surveyed residents from only one site, and examined only self-reported exposure and confidence, not competence. The relationship between confidence and competence is notoriously problematic⁷ and there is a risk of familiarity creating an illusion of knowledge and/or competence. Ultimately, a competency-based medical system is intended to be dynamic, adaptive, and contextual. Despite the extensive competency-based framework in place to track the development of physicians, data about the contexts in which competency is demonstrated are lacking. There is no reason to think that the key gaps identified in Sclafani et al are unique to their institution.

Given the ultimate goal of developing curricula that prepare residents for independent practice coupled with robust systems of assessment that ensure they are ready to do so, educators must implement strategies to identify and alleviate the idiosyncrasy of the resident experience. The survey tool in the present work could be used as a needs assessment and would require minimal resources, but is limited by recall bias, illusion of knowledge, and lack of data regarding actual competence. Other potential strategies include case logs or e-folios, although these tools are also limited by the understanding that familiarity and exposure do not necessarily engender competence.

One potential strategy suggested by Warm et al. is the addition of the “Observable Practice Activities” (OPA), “a collection of learning objectives/activities that must be observed in daily practice in order to form entrustment decisions.”⁸ The intention is to more granularly define what residents actually do and then map these activities to the established competency-based framework. Using these observable activities as an assessment unit may allow for identification of individual experience gaps, thereby improving the dynamicity and adaptiveness of GME training. Certainly, there are very real concerns about further complicating an already complex and abstract system and using a reductionist approach to define the activities of a profession. However, the findings of Sclafani et al with respect to the wide range of resident experience elucidates the need for continued study and innovation regarding the manner in which the medical education community determines our trainees are prepared for independent practice.

Disclosures

The authors have nothing to disclose.

Graduate medical education (GME) is heavily reliant on experiential learning. Most of a resident’s time is spent in progressively independent delivery of patient care, which is associated with decreasing supervision. Attainment and demonstration of competence in patient care is the goal and responsibility of GME training programs. What happens, then, if the medicine resident never has the experience necessary to enable experiential learning? What if she never “sees one,” let alone “does one”?

In this month’s Journal of Hospital Medicine, Sclafani et al¹ examine how exposure to urgent clinical situations impacts residents’ confidence in managing these ward emergencies. They astutely reveal the idiosyncratic nature of residency training and consequent gaps created when an educational delivery model predicated on experience lacks certain experiences. How can a resident without certain key experiences be ready for independent practice?

The ACGME’s Next Accreditation System is intended to ensure that residents are prepared for independent practice. The educational outcomes that learners must attain are comprised of six core competencies, with milestones intended to operationalize the measurement and reporting of learner progression toward competence.^2,3 It is challenging to apply general competencies to assessment of day to day clinical activities. This challenge led to the development of 16 Entrustable Professional Activities (EPAs). These allow the direct observation of concrete clinical activities that could then infer the attainment (or not) of multiple competencies. Ideally, EPAs are paired with and mapped to curricular milestones which describe a learner’s trajectory within the framework of competencies and determine if a resident is prepared for independent practice.^4,5

In Sclafani et al.¹ the authors characterize resident exposure to, and confidence in, 50 urgent clinical situations. Both level of training and exposure were associated with increased confidence. However, the most important finding of this paper is the wide variation of resident exposures and confidence with respect to specific urgent clinical events. At least 15% of graduating residents had never seen 16% of the 50 emergency events, and a majority of graduating residents did not feel confident managing 20% of the 50 events, highlighting the idiosyncratic nature of GME training.¹ Of course, while certain entities on the list of clinical emergencies were not identified as final diagnoses, it is possible they were still considered in the process of caring for patients in different situations.

Several factors account for the idiosyncratic nature of medical training, including the rarity of certain clinical events, seasonal variation in conditions, and other variables (ie, learner elective choices). In addition, the scheduling of most residency programs is based on patient care needs instead of individual trainees’ educational needs. Other areas of medicine have attempted to standardize experience and ensure specific exposure and/or competence using strategies such as surgical case logs and case-based certifying examinations. There are very important recently described projects in undergraduate medical education aimed at using longitudinal assessment of EPAs in multiple contexts to make entrustment decisions.⁶ However, Internal Medicine residencies do not routinely employ these strategies.

It must be noted that Sclafani et al. surveyed residents from only one site, and examined only self-reported exposure and confidence, not competence. The relationship between confidence and competence is notoriously problematic⁷ and there is a risk of familiarity creating an illusion of knowledge and/or competence. Ultimately, a competency-based medical system is intended to be dynamic, adaptive, and contextual. Despite the extensive competency-based framework in place to track the development of physicians, data about the contexts in which competency is demonstrated are lacking. There is no reason to think that the key gaps identified in Sclafani et al are unique to their institution.

Given the ultimate goal of developing curricula that prepare residents for independent practice coupled with robust systems of assessment that ensure they are ready to do so, educators must implement strategies to identify and alleviate the idiosyncrasy of the resident experience. The survey tool in the present work could be used as a needs assessment and would require minimal resources, but is limited by recall bias, illusion of knowledge, and lack of data regarding actual competence. Other potential strategies include case logs or e-folios, although these tools are also limited by the understanding that familiarity and exposure do not necessarily engender competence.

One potential strategy suggested by Warm et al. is the addition of the “Observable Practice Activities” (OPA), “a collection of learning objectives/activities that must be observed in daily practice in order to form entrustment decisions.”⁸ The intention is to more granularly define what residents actually do and then map these activities to the established competency-based framework. Using these observable activities as an assessment unit may allow for identification of individual experience gaps, thereby improving the dynamicity and adaptiveness of GME training. Certainly, there are very real concerns about further complicating an already complex and abstract system and using a reductionist approach to define the activities of a profession. However, the findings of Sclafani et al with respect to the wide range of resident experience elucidates the need for continued study and innovation regarding the manner in which the medical education community determines our trainees are prepared for independent practice.

Disclosures

The authors have nothing to disclose.

Graduate medical education (GME) is heavily reliant on experiential learning. Most of a resident’s time is spent in progressively independent delivery of patient care, which is associated with decreasing supervision. Attainment and demonstration of competence in patient care is the goal and responsibility of GME training programs. What happens, then, if the medicine resident never has the experience necessary to enable experiential learning? What if she never “sees one,” let alone “does one”?

In this month’s Journal of Hospital Medicine, Sclafani et al¹ examine how exposure to urgent clinical situations impacts residents’ confidence in managing these ward emergencies. They astutely reveal the idiosyncratic nature of residency training and consequent gaps created when an educational delivery model predicated on experience lacks certain experiences. How can a resident without certain key experiences be ready for independent practice?

The ACGME’s Next Accreditation System is intended to ensure that residents are prepared for independent practice. The educational outcomes that learners must attain are comprised of six core competencies, with milestones intended to operationalize the measurement and reporting of learner progression toward competence.^2,3 It is challenging to apply general competencies to assessment of day to day clinical activities. This challenge led to the development of 16 Entrustable Professional Activities (EPAs). These allow the direct observation of concrete clinical activities that could then infer the attainment (or not) of multiple competencies. Ideally, EPAs are paired with and mapped to curricular milestones which describe a learner’s trajectory within the framework of competencies and determine if a resident is prepared for independent practice.^4,5

In Sclafani et al.¹ the authors characterize resident exposure to, and confidence in, 50 urgent clinical situations. Both level of training and exposure were associated with increased confidence. However, the most important finding of this paper is the wide variation of resident exposures and confidence with respect to specific urgent clinical events. At least 15% of graduating residents had never seen 16% of the 50 emergency events, and a majority of graduating residents did not feel confident managing 20% of the 50 events, highlighting the idiosyncratic nature of GME training.¹ Of course, while certain entities on the list of clinical emergencies were not identified as final diagnoses, it is possible they were still considered in the process of caring for patients in different situations.

Several factors account for the idiosyncratic nature of medical training, including the rarity of certain clinical events, seasonal variation in conditions, and other variables (ie, learner elective choices). In addition, the scheduling of most residency programs is based on patient care needs instead of individual trainees’ educational needs. Other areas of medicine have attempted to standardize experience and ensure specific exposure and/or competence using strategies such as surgical case logs and case-based certifying examinations. There are very important recently described projects in undergraduate medical education aimed at using longitudinal assessment of EPAs in multiple contexts to make entrustment decisions.⁶ However, Internal Medicine residencies do not routinely employ these strategies.

It must be noted that Sclafani et al. surveyed residents from only one site, and examined only self-reported exposure and confidence, not competence. The relationship between confidence and competence is notoriously problematic⁷ and there is a risk of familiarity creating an illusion of knowledge and/or competence. Ultimately, a competency-based medical system is intended to be dynamic, adaptive, and contextual. Despite the extensive competency-based framework in place to track the development of physicians, data about the contexts in which competency is demonstrated are lacking. There is no reason to think that the key gaps identified in Sclafani et al are unique to their institution.

Given the ultimate goal of developing curricula that prepare residents for independent practice coupled with robust systems of assessment that ensure they are ready to do so, educators must implement strategies to identify and alleviate the idiosyncrasy of the resident experience. The survey tool in the present work could be used as a needs assessment and would require minimal resources, but is limited by recall bias, illusion of knowledge, and lack of data regarding actual competence. Other potential strategies include case logs or e-folios, although these tools are also limited by the understanding that familiarity and exposure do not necessarily engender competence.

One potential strategy suggested by Warm et al. is the addition of the “Observable Practice Activities” (OPA), “a collection of learning objectives/activities that must be observed in daily practice in order to form entrustment decisions.”⁸ The intention is to more granularly define what residents actually do and then map these activities to the established competency-based framework. Using these observable activities as an assessment unit may allow for identification of individual experience gaps, thereby improving the dynamicity and adaptiveness of GME training. Certainly, there are very real concerns about further complicating an already complex and abstract system and using a reductionist approach to define the activities of a profession. However, the findings of Sclafani et al with respect to the wide range of resident experience elucidates the need for continued study and innovation regarding the manner in which the medical education community determines our trainees are prepared for independent practice.

Disclosures

The authors have nothing to disclose.

Hip fracture is a common reason for unexpected, urgent inpatient surgery in older patients. In 2005, the incidence of hip fracture was 369.0 and 793.5 per 100,000 in men and women respectively.¹ These numbers declined over the preceding decade, potentially as a result of bisphosphonate use. Age- and risk-adjusted 30-day mortality rates for men and women in 2005 were approximately 10% and 5%, respectively.

Evidence suggests that timely surgical repair of hip fractures improves outcomes, although the optimal timing is controversial. Guidelines from the American College of Surgeons Committee on Trauma from 2015 recommend surgical intervention within 48 hours for geriatric hip fracures.² A 2008 systematic review found that operative delay beyond 48 hours was associated with a 41% increase in 30-day all-cause mortality and a 32% increase in one-year all-cause mortality.³ Recent evidence suggests that the rate of complications begins to increase with delays beyond 24 hours.⁴

There has been a focus over the past decade on overuse of preoperative testing for low- and intermediate-risk surgeries.^5-7 Beginning in 2012, the American Board of Internal Medicine initiated the Choosing Wisely^® campaign in which numerous societies issued recommendations on reducing utilization of various diagnostic tests, a number of which have focused on preoperative tests. Two groups—the American Society of Anesthesiologists (ASA) and the American Society of Echocardiography (ASE)— issued specific recommendations on preoperative cardiac testing.⁸ In February 2013, the ASE recommended avoiding preoperative echocardiograms in patients without a history or symptoms of heart disease. In October 2013, the ASA recommended against transthoracic echocardiogram (TTE), transesophageal echocardiogram (TEE), or stress testing for low- or intermediate-risk noncardiac surgery for patients with stable cardiac disease.

Finally, in 2014, the American College of Cardiology (ACC)/American Heart Association (AHA) issued updated perioperative guidelines for patients undergoing noncardiac surgeries.⁹ They recommended preoperative stress testing only in a small subset of cases (patients with an elevated perioperative risk of major adverse cardiac event, a poor or unknown functional capacity, or those in whom stress testing would impact perioperative care).

Given the high cost of preoperative cardiac testing, the potential for delays in care that can adversely impact outcomes, and the recent recommendations, we sought to characterize the rates of inpatient preoperative cardiac testing prior to hip fracture surgery in recent years and to see whether recent recommendations to curb use of these tests were temporally associated with changing rates.

Overview

We utilized two datasets—the Healthcare Cost and Utilization Project (HCUP) State Inpatient Databases (SID) and the American Hospital Association (AHA) Annual Survey—to characterize preoperative cardiac testing. SID data from Maryland, New Jersey, and Washington State from 2011 through September 2015 were used (the ICD coding system changed from ICD9 to ICD10 on October 1). This was combined with AHA data for these years. We included all hospitalizations with a primary ICD9 procedure code for hip fracture repair—78.55, 78.65, 79.05, 79.15, 79.25, 79.35, 79.45, 79.55, 79.65, 79.75, 79.85, and 79.95. We excluded all observations that involved an interhospital transfer. This study was exempt from institutional review board approval.

Measurement and Outcomes

We summarized demographic data for the hospitalizations that met the inclusion criteria as well as the associated hospitals. The primary outcome was the percentage of patients undergoing TTE, stress test, and cardiac catheterization during a hospitalization with a primary procedure code of hip fracture repair. Random effects logistic regression models for each type of diagnostic test were developed to determine the factors that might impact test utilization. In addition to running each test as a separate model, we also performed an analysis in which the outcome was performance of any of these three cardiac tests. Random effects were used to account for clustering of testing within hospitals. Variables included time (3-month intervals), state, age (continuous variable), gender, length of stay, payer (Medicare/Medicaid/private insurance/self-pay/other), hospital teaching status (major teaching/minor teaching/nonteaching), hospital size according to number of beds (continuous variable), and mortality score. Major teaching hospitals are defined as members of the Council of Teaching Hospitals. Minor teaching hospitals are defined as (1) those with one or more postgraduate training programs recognized by the American Council on Graduate Medical Education, (2) those with a medical school affiliation reported to the American Medical Association, or (3) those with an internship or residency approved by the American Osteopathic Association.

The SID has a specific binary indicator variable for each of the three diagnostic tests we evaluated. The use of the diagnostic test is evaluated through both UB-92 revenue codes and ICD9 procedure codes, with the presence of either leading to the indicator variable being positive.¹⁰ Finally, we performed a sensitivity analysis to evaluate the significance of changing utilization trends by interrupted time series analysis. A level of 0.05 was used to determine statistical significance. Analyses were done in STATA 15 (College Station, Texas).

The dataset included 75,144 hospitalizations with a primary procedure code of hip fracture over the study period (Table). The number of hospitalizations per year was fairly consistent over the study period in each state, although there were fewer hospitalizations for 2015 as this included only January through September. The mean age was 72.8 years, and 67% were female. The primary payer was Medicare for 71.7% of hospitalizations. Hospitalizations occurred at 181 hospitals, the plurality of which (42.9%) were minor teaching hospitals. The proportions of hospitalizations that included a TTE, stress test, and cardiac catheterization were 12.6%, 1.1%, and 0.5%, respectively. Overall, 13.5% of patients underwent any cardiac testing.

There was a statistically significantly lower rate of stress tests (odds ratio [OR], 0.32; 95% CI, 0.19-0.54) and cardiac catheterizations (OR, 0.46; 95% CI, 0.27-0.79) in Washington than in Maryland and New Jersey. Female gender was associated with significantly lower adjusted ORs for stress tests (OR, 0.74; 95% CI, 0.63-0.86) and cardiac catheterizations (OR, 0.73; 95% CI, 0.59-0.91), and increasing age was associated with higher adjusted ORs for each test (TTE, OR, 1.033; 95% CI, 1.031-1.035; stress tests, OR, 1.007; 95% CI, 1.001-1.013; cardiac catheterizations, OR, 1.011; 95% CI, 1.003-1.019). Private insurance was associated with a lower likelihood of stress tests (OR, 0.65; 95% CI, 0.50-0.85) and cardiac catheterizations (OR, 0.67; 95% CI,0.46-0.98), and self-pay was associated with a lower likelihood of TTE (OR, 0.76; 95% CI, 0.61-0.95) and stress test (OR, 0.43; 95% CI, 0.21-0.90), all compared with Medicare.

Larger hospitals were associated with a greater likelihood of cardiac catheterizations (OR, 1.18; 95% CI, 1.03-1.36) and a lower likelihood of TTE (OR, 0.89; 95% CI, 0.82-0.96). An unweighted average of these tests between 2011 and October 2015 showed a modest increase in TTEs and a modest decrease in stress tests and cardiac catheterizations (Figure). A multivariable random effects regression for use of TTEs revealed a significantly increasing trend from 2011 to 2014 (OR, 1.04, P < .0001), but the decreasing trend for 2015 was not statistically significant when analyzed according to quarters or months (for which data from only New Jersey and Washington are available).

There has been limited research into how often preoperative cardiac testing occurs in the inpatient setting. Our aim was to study its prevalence prior to hip fracture surgery during a time period when multiple recommendations had been issued to limit its use. We found rates of ischemic testing (stress tests and cardiac catheterizations) to be appropriately, and perhaps surprisingly, low. Our results on ischemic testing rates are consistent with previous studies, which have focused on the outpatient setting where much of the preoperative workup for nonurgent surgeries occurs. The rate of TTEs was higher than in previous studies of the outpatient preoperative setting, although it is unclear what an optimal rate of TTEs is.

A recent study examining outpatient preoperative stress tests within the 30 days before cataract surgeries, knee arthroscopies, or shoulder arthroscopies found a rate of 2.1% for Medicare fee-for-service patients in 2009 with little regional variation.¹¹ Another evaluation using 2009 Medicare claims data found rates of preoperative TTEs and stress tests to be 0.8% and 0.7%, respectively.¹² They included TTEs and stress tests performed within 30 days of a low- or intermediate-risk surgery. A study analyzing the rate of preoperative TTEs between 2009 and 2014 found that rates varied from 2.0% to 3.4% for commercially insured patients aged 50-64 years and Medicare-advantage patients, respectively, in 2009.¹³ These rates decreased by 7.0% and 12.6% from 2009 to 2014. These studies, like ours, suggest that preoperative cardiac testing has not been a major source of wasteful spending. One explanation for the higher rate of TTEs we observed in the inpatient setting might be that primary care physicians in the outpatient setting are more likely to have historical cardiac testing results compared with physicians in a hospital.

We found that the rate of stress testing and cardiac catheterization in Washington was significantly lower than that in Maryland and New Jersey. This is consistent with a number of measures of healthcare utilization – total Medicare reimbursement in the last six months of life, mean number of hospital days in the last six months of life, and healthcare intensity index—for all of which Washington was below the national mean and Maryland and New Jersey were above it.¹⁴

Finally, we found evidence of a lower rate of preoperative stress tests and cardiac catheterizations for women despite controlling for age and mortality score. Of course, we did not control directly for cardiovascular comorbidities; as a result, there could be residual confounding. However, these results are consistent with previous findings of gender bias in both pharmacologic management of coronary artery disease (CAD)¹⁵ and diagnostic testing for suspected CAD.¹⁶

We focused on hospitalizations with a primary procedure code to surgically treat hip fracture. We are unable to tell if the cardiac testing of these patients had occurred before or after the procedure. However, we suspect that the vast majority were completed for preoperative evaluation. It is likely that a small subset were done to diagnose and manage cardiac complications that either accompanied the hip fracture or occurred postoperatively. Another limitation is that we cannot determine if a patient had one of these tests recently in the emergency department or as an outpatient.

We also chose to include only patients who actually had hip fracture surgery. It is possible that the testing rate is higher for all patients admitted for hip fracture and that some of these patients did not have surgery because of abnormal cardiac testing. However, we suspect that this is a very small fraction given the high degree of morbidity and mortality associated with untreated hip fracture.

We found a low rate of preoperative cardiac testing in patients hospitalized for hip fracture surgery both in the years before and after the issuance of recommendations intended to curb its use. Although it is reassuring that the volume of low-value testing is lower than we expected, these findings highlight the importance of targeting utilization improvement efforts toward low-value tests and procedures that are more heavily used, since further curbing the use of infrequently utilized tests and procedures will have only a modest impact on overall healthcare expenditure. Our findings highlight the necessity that professional organizations ensure that they focus on true areas of inappropriate utilization. These are the areas in which improvements will have a major impact on healthcare spending. Further research should aim to quantify unwarranted cardiac testing for other inpatient surgeries that are less urgent, as the urgency of hip fracture repair may be driving the relatively low utilization of inpatient cardiac testing.

Disclosures

The authors have nothing to disclose.

Funding

This project was supported by the Johns Hopkins Hospitalist Scholars Fund and the Johns Hopkins School of Medicine Biostatistics, Epidemiology and Data Management (BEAD) Core.

Hip fracture is a common reason for unexpected, urgent inpatient surgery in older patients. In 2005, the incidence of hip fracture was 369.0 and 793.5 per 100,000 in men and women respectively.¹ These numbers declined over the preceding decade, potentially as a result of bisphosphonate use. Age- and risk-adjusted 30-day mortality rates for men and women in 2005 were approximately 10% and 5%, respectively.

Evidence suggests that timely surgical repair of hip fractures improves outcomes, although the optimal timing is controversial. Guidelines from the American College of Surgeons Committee on Trauma from 2015 recommend surgical intervention within 48 hours for geriatric hip fracures.² A 2008 systematic review found that operative delay beyond 48 hours was associated with a 41% increase in 30-day all-cause mortality and a 32% increase in one-year all-cause mortality.³ Recent evidence suggests that the rate of complications begins to increase with delays beyond 24 hours.⁴

There has been a focus over the past decade on overuse of preoperative testing for low- and intermediate-risk surgeries.^5-7 Beginning in 2012, the American Board of Internal Medicine initiated the Choosing Wisely^® campaign in which numerous societies issued recommendations on reducing utilization of various diagnostic tests, a number of which have focused on preoperative tests. Two groups—the American Society of Anesthesiologists (ASA) and the American Society of Echocardiography (ASE)— issued specific recommendations on preoperative cardiac testing.⁸ In February 2013, the ASE recommended avoiding preoperative echocardiograms in patients without a history or symptoms of heart disease. In October 2013, the ASA recommended against transthoracic echocardiogram (TTE), transesophageal echocardiogram (TEE), or stress testing for low- or intermediate-risk noncardiac surgery for patients with stable cardiac disease.

Finally, in 2014, the American College of Cardiology (ACC)/American Heart Association (AHA) issued updated perioperative guidelines for patients undergoing noncardiac surgeries.⁹ They recommended preoperative stress testing only in a small subset of cases (patients with an elevated perioperative risk of major adverse cardiac event, a poor or unknown functional capacity, or those in whom stress testing would impact perioperative care).

Given the high cost of preoperative cardiac testing, the potential for delays in care that can adversely impact outcomes, and the recent recommendations, we sought to characterize the rates of inpatient preoperative cardiac testing prior to hip fracture surgery in recent years and to see whether recent recommendations to curb use of these tests were temporally associated with changing rates.

Overview

We utilized two datasets—the Healthcare Cost and Utilization Project (HCUP) State Inpatient Databases (SID) and the American Hospital Association (AHA) Annual Survey—to characterize preoperative cardiac testing. SID data from Maryland, New Jersey, and Washington State from 2011 through September 2015 were used (the ICD coding system changed from ICD9 to ICD10 on October 1). This was combined with AHA data for these years. We included all hospitalizations with a primary ICD9 procedure code for hip fracture repair—78.55, 78.65, 79.05, 79.15, 79.25, 79.35, 79.45, 79.55, 79.65, 79.75, 79.85, and 79.95. We excluded all observations that involved an interhospital transfer. This study was exempt from institutional review board approval.

Measurement and Outcomes

We summarized demographic data for the hospitalizations that met the inclusion criteria as well as the associated hospitals. The primary outcome was the percentage of patients undergoing TTE, stress test, and cardiac catheterization during a hospitalization with a primary procedure code of hip fracture repair. Random effects logistic regression models for each type of diagnostic test were developed to determine the factors that might impact test utilization. In addition to running each test as a separate model, we also performed an analysis in which the outcome was performance of any of these three cardiac tests. Random effects were used to account for clustering of testing within hospitals. Variables included time (3-month intervals), state, age (continuous variable), gender, length of stay, payer (Medicare/Medicaid/private insurance/self-pay/other), hospital teaching status (major teaching/minor teaching/nonteaching), hospital size according to number of beds (continuous variable), and mortality score. Major teaching hospitals are defined as members of the Council of Teaching Hospitals. Minor teaching hospitals are defined as (1) those with one or more postgraduate training programs recognized by the American Council on Graduate Medical Education, (2) those with a medical school affiliation reported to the American Medical Association, or (3) those with an internship or residency approved by the American Osteopathic Association.

The SID has a specific binary indicator variable for each of the three diagnostic tests we evaluated. The use of the diagnostic test is evaluated through both UB-92 revenue codes and ICD9 procedure codes, with the presence of either leading to the indicator variable being positive.¹⁰ Finally, we performed a sensitivity analysis to evaluate the significance of changing utilization trends by interrupted time series analysis. A level of 0.05 was used to determine statistical significance. Analyses were done in STATA 15 (College Station, Texas).

The dataset included 75,144 hospitalizations with a primary procedure code of hip fracture over the study period (Table). The number of hospitalizations per year was fairly consistent over the study period in each state, although there were fewer hospitalizations for 2015 as this included only January through September. The mean age was 72.8 years, and 67% were female. The primary payer was Medicare for 71.7% of hospitalizations. Hospitalizations occurred at 181 hospitals, the plurality of which (42.9%) were minor teaching hospitals. The proportions of hospitalizations that included a TTE, stress test, and cardiac catheterization were 12.6%, 1.1%, and 0.5%, respectively. Overall, 13.5% of patients underwent any cardiac testing.

There was a statistically significantly lower rate of stress tests (odds ratio [OR], 0.32; 95% CI, 0.19-0.54) and cardiac catheterizations (OR, 0.46; 95% CI, 0.27-0.79) in Washington than in Maryland and New Jersey. Female gender was associated with significantly lower adjusted ORs for stress tests (OR, 0.74; 95% CI, 0.63-0.86) and cardiac catheterizations (OR, 0.73; 95% CI, 0.59-0.91), and increasing age was associated with higher adjusted ORs for each test (TTE, OR, 1.033; 95% CI, 1.031-1.035; stress tests, OR, 1.007; 95% CI, 1.001-1.013; cardiac catheterizations, OR, 1.011; 95% CI, 1.003-1.019). Private insurance was associated with a lower likelihood of stress tests (OR, 0.65; 95% CI, 0.50-0.85) and cardiac catheterizations (OR, 0.67; 95% CI,0.46-0.98), and self-pay was associated with a lower likelihood of TTE (OR, 0.76; 95% CI, 0.61-0.95) and stress test (OR, 0.43; 95% CI, 0.21-0.90), all compared with Medicare.

Larger hospitals were associated with a greater likelihood of cardiac catheterizations (OR, 1.18; 95% CI, 1.03-1.36) and a lower likelihood of TTE (OR, 0.89; 95% CI, 0.82-0.96). An unweighted average of these tests between 2011 and October 2015 showed a modest increase in TTEs and a modest decrease in stress tests and cardiac catheterizations (Figure). A multivariable random effects regression for use of TTEs revealed a significantly increasing trend from 2011 to 2014 (OR, 1.04, P < .0001), but the decreasing trend for 2015 was not statistically significant when analyzed according to quarters or months (for which data from only New Jersey and Washington are available).

There has been limited research into how often preoperative cardiac testing occurs in the inpatient setting. Our aim was to study its prevalence prior to hip fracture surgery during a time period when multiple recommendations had been issued to limit its use. We found rates of ischemic testing (stress tests and cardiac catheterizations) to be appropriately, and perhaps surprisingly, low. Our results on ischemic testing rates are consistent with previous studies, which have focused on the outpatient setting where much of the preoperative workup for nonurgent surgeries occurs. The rate of TTEs was higher than in previous studies of the outpatient preoperative setting, although it is unclear what an optimal rate of TTEs is.

A recent study examining outpatient preoperative stress tests within the 30 days before cataract surgeries, knee arthroscopies, or shoulder arthroscopies found a rate of 2.1% for Medicare fee-for-service patients in 2009 with little regional variation.¹¹ Another evaluation using 2009 Medicare claims data found rates of preoperative TTEs and stress tests to be 0.8% and 0.7%, respectively.¹² They included TTEs and stress tests performed within 30 days of a low- or intermediate-risk surgery. A study analyzing the rate of preoperative TTEs between 2009 and 2014 found that rates varied from 2.0% to 3.4% for commercially insured patients aged 50-64 years and Medicare-advantage patients, respectively, in 2009.¹³ These rates decreased by 7.0% and 12.6% from 2009 to 2014. These studies, like ours, suggest that preoperative cardiac testing has not been a major source of wasteful spending. One explanation for the higher rate of TTEs we observed in the inpatient setting might be that primary care physicians in the outpatient setting are more likely to have historical cardiac testing results compared with physicians in a hospital.

We found that the rate of stress testing and cardiac catheterization in Washington was significantly lower than that in Maryland and New Jersey. This is consistent with a number of measures of healthcare utilization – total Medicare reimbursement in the last six months of life, mean number of hospital days in the last six months of life, and healthcare intensity index—for all of which Washington was below the national mean and Maryland and New Jersey were above it.¹⁴

Finally, we found evidence of a lower rate of preoperative stress tests and cardiac catheterizations for women despite controlling for age and mortality score. Of course, we did not control directly for cardiovascular comorbidities; as a result, there could be residual confounding. However, these results are consistent with previous findings of gender bias in both pharmacologic management of coronary artery disease (CAD)¹⁵ and diagnostic testing for suspected CAD.¹⁶

We focused on hospitalizations with a primary procedure code to surgically treat hip fracture. We are unable to tell if the cardiac testing of these patients had occurred before or after the procedure. However, we suspect that the vast majority were completed for preoperative evaluation. It is likely that a small subset were done to diagnose and manage cardiac complications that either accompanied the hip fracture or occurred postoperatively. Another limitation is that we cannot determine if a patient had one of these tests recently in the emergency department or as an outpatient.

We also chose to include only patients who actually had hip fracture surgery. It is possible that the testing rate is higher for all patients admitted for hip fracture and that some of these patients did not have surgery because of abnormal cardiac testing. However, we suspect that this is a very small fraction given the high degree of morbidity and mortality associated with untreated hip fracture.

We found a low rate of preoperative cardiac testing in patients hospitalized for hip fracture surgery both in the years before and after the issuance of recommendations intended to curb its use. Although it is reassuring that the volume of low-value testing is lower than we expected, these findings highlight the importance of targeting utilization improvement efforts toward low-value tests and procedures that are more heavily used, since further curbing the use of infrequently utilized tests and procedures will have only a modest impact on overall healthcare expenditure. Our findings highlight the necessity that professional organizations ensure that they focus on true areas of inappropriate utilization. These are the areas in which improvements will have a major impact on healthcare spending. Further research should aim to quantify unwarranted cardiac testing for other inpatient surgeries that are less urgent, as the urgency of hip fracture repair may be driving the relatively low utilization of inpatient cardiac testing.

Disclosures

The authors have nothing to disclose.

Funding

This project was supported by the Johns Hopkins Hospitalist Scholars Fund and the Johns Hopkins School of Medicine Biostatistics, Epidemiology and Data Management (BEAD) Core.

Hip fracture is a common reason for unexpected, urgent inpatient surgery in older patients. In 2005, the incidence of hip fracture was 369.0 and 793.5 per 100,000 in men and women respectively.¹ These numbers declined over the preceding decade, potentially as a result of bisphosphonate use. Age- and risk-adjusted 30-day mortality rates for men and women in 2005 were approximately 10% and 5%, respectively.

Evidence suggests that timely surgical repair of hip fractures improves outcomes, although the optimal timing is controversial. Guidelines from the American College of Surgeons Committee on Trauma from 2015 recommend surgical intervention within 48 hours for geriatric hip fracures.² A 2008 systematic review found that operative delay beyond 48 hours was associated with a 41% increase in 30-day all-cause mortality and a 32% increase in one-year all-cause mortality.³ Recent evidence suggests that the rate of complications begins to increase with delays beyond 24 hours.⁴

There has been a focus over the past decade on overuse of preoperative testing for low- and intermediate-risk surgeries.^5-7 Beginning in 2012, the American Board of Internal Medicine initiated the Choosing Wisely^® campaign in which numerous societies issued recommendations on reducing utilization of various diagnostic tests, a number of which have focused on preoperative tests. Two groups—the American Society of Anesthesiologists (ASA) and the American Society of Echocardiography (ASE)— issued specific recommendations on preoperative cardiac testing.⁸ In February 2013, the ASE recommended avoiding preoperative echocardiograms in patients without a history or symptoms of heart disease. In October 2013, the ASA recommended against transthoracic echocardiogram (TTE), transesophageal echocardiogram (TEE), or stress testing for low- or intermediate-risk noncardiac surgery for patients with stable cardiac disease.

Finally, in 2014, the American College of Cardiology (ACC)/American Heart Association (AHA) issued updated perioperative guidelines for patients undergoing noncardiac surgeries.⁹ They recommended preoperative stress testing only in a small subset of cases (patients with an elevated perioperative risk of major adverse cardiac event, a poor or unknown functional capacity, or those in whom stress testing would impact perioperative care).

Given the high cost of preoperative cardiac testing, the potential for delays in care that can adversely impact outcomes, and the recent recommendations, we sought to characterize the rates of inpatient preoperative cardiac testing prior to hip fracture surgery in recent years and to see whether recent recommendations to curb use of these tests were temporally associated with changing rates.

Overview

We utilized two datasets—the Healthcare Cost and Utilization Project (HCUP) State Inpatient Databases (SID) and the American Hospital Association (AHA) Annual Survey—to characterize preoperative cardiac testing. SID data from Maryland, New Jersey, and Washington State from 2011 through September 2015 were used (the ICD coding system changed from ICD9 to ICD10 on October 1). This was combined with AHA data for these years. We included all hospitalizations with a primary ICD9 procedure code for hip fracture repair—78.55, 78.65, 79.05, 79.15, 79.25, 79.35, 79.45, 79.55, 79.65, 79.75, 79.85, and 79.95. We excluded all observations that involved an interhospital transfer. This study was exempt from institutional review board approval.

Measurement and Outcomes

We summarized demographic data for the hospitalizations that met the inclusion criteria as well as the associated hospitals. The primary outcome was the percentage of patients undergoing TTE, stress test, and cardiac catheterization during a hospitalization with a primary procedure code of hip fracture repair. Random effects logistic regression models for each type of diagnostic test were developed to determine the factors that might impact test utilization. In addition to running each test as a separate model, we also performed an analysis in which the outcome was performance of any of these three cardiac tests. Random effects were used to account for clustering of testing within hospitals. Variables included time (3-month intervals), state, age (continuous variable), gender, length of stay, payer (Medicare/Medicaid/private insurance/self-pay/other), hospital teaching status (major teaching/minor teaching/nonteaching), hospital size according to number of beds (continuous variable), and mortality score. Major teaching hospitals are defined as members of the Council of Teaching Hospitals. Minor teaching hospitals are defined as (1) those with one or more postgraduate training programs recognized by the American Council on Graduate Medical Education, (2) those with a medical school affiliation reported to the American Medical Association, or (3) those with an internship or residency approved by the American Osteopathic Association.

The SID has a specific binary indicator variable for each of the three diagnostic tests we evaluated. The use of the diagnostic test is evaluated through both UB-92 revenue codes and ICD9 procedure codes, with the presence of either leading to the indicator variable being positive.¹⁰ Finally, we performed a sensitivity analysis to evaluate the significance of changing utilization trends by interrupted time series analysis. A level of 0.05 was used to determine statistical significance. Analyses were done in STATA 15 (College Station, Texas).

The dataset included 75,144 hospitalizations with a primary procedure code of hip fracture over the study period (Table). The number of hospitalizations per year was fairly consistent over the study period in each state, although there were fewer hospitalizations for 2015 as this included only January through September. The mean age was 72.8 years, and 67% were female. The primary payer was Medicare for 71.7% of hospitalizations. Hospitalizations occurred at 181 hospitals, the plurality of which (42.9%) were minor teaching hospitals. The proportions of hospitalizations that included a TTE, stress test, and cardiac catheterization were 12.6%, 1.1%, and 0.5%, respectively. Overall, 13.5% of patients underwent any cardiac testing.

There was a statistically significantly lower rate of stress tests (odds ratio [OR], 0.32; 95% CI, 0.19-0.54) and cardiac catheterizations (OR, 0.46; 95% CI, 0.27-0.79) in Washington than in Maryland and New Jersey. Female gender was associated with significantly lower adjusted ORs for stress tests (OR, 0.74; 95% CI, 0.63-0.86) and cardiac catheterizations (OR, 0.73; 95% CI, 0.59-0.91), and increasing age was associated with higher adjusted ORs for each test (TTE, OR, 1.033; 95% CI, 1.031-1.035; stress tests, OR, 1.007; 95% CI, 1.001-1.013; cardiac catheterizations, OR, 1.011; 95% CI, 1.003-1.019). Private insurance was associated with a lower likelihood of stress tests (OR, 0.65; 95% CI, 0.50-0.85) and cardiac catheterizations (OR, 0.67; 95% CI,0.46-0.98), and self-pay was associated with a lower likelihood of TTE (OR, 0.76; 95% CI, 0.61-0.95) and stress test (OR, 0.43; 95% CI, 0.21-0.90), all compared with Medicare.

Larger hospitals were associated with a greater likelihood of cardiac catheterizations (OR, 1.18; 95% CI, 1.03-1.36) and a lower likelihood of TTE (OR, 0.89; 95% CI, 0.82-0.96). An unweighted average of these tests between 2011 and October 2015 showed a modest increase in TTEs and a modest decrease in stress tests and cardiac catheterizations (Figure). A multivariable random effects regression for use of TTEs revealed a significantly increasing trend from 2011 to 2014 (OR, 1.04, P < .0001), but the decreasing trend for 2015 was not statistically significant when analyzed according to quarters or months (for which data from only New Jersey and Washington are available).

There has been limited research into how often preoperative cardiac testing occurs in the inpatient setting. Our aim was to study its prevalence prior to hip fracture surgery during a time period when multiple recommendations had been issued to limit its use. We found rates of ischemic testing (stress tests and cardiac catheterizations) to be appropriately, and perhaps surprisingly, low. Our results on ischemic testing rates are consistent with previous studies, which have focused on the outpatient setting where much of the preoperative workup for nonurgent surgeries occurs. The rate of TTEs was higher than in previous studies of the outpatient preoperative setting, although it is unclear what an optimal rate of TTEs is.

A recent study examining outpatient preoperative stress tests within the 30 days before cataract surgeries, knee arthroscopies, or shoulder arthroscopies found a rate of 2.1% for Medicare fee-for-service patients in 2009 with little regional variation.¹¹ Another evaluation using 2009 Medicare claims data found rates of preoperative TTEs and stress tests to be 0.8% and 0.7%, respectively.¹² They included TTEs and stress tests performed within 30 days of a low- or intermediate-risk surgery. A study analyzing the rate of preoperative TTEs between 2009 and 2014 found that rates varied from 2.0% to 3.4% for commercially insured patients aged 50-64 years and Medicare-advantage patients, respectively, in 2009.¹³ These rates decreased by 7.0% and 12.6% from 2009 to 2014. These studies, like ours, suggest that preoperative cardiac testing has not been a major source of wasteful spending. One explanation for the higher rate of TTEs we observed in the inpatient setting might be that primary care physicians in the outpatient setting are more likely to have historical cardiac testing results compared with physicians in a hospital.

We found that the rate of stress testing and cardiac catheterization in Washington was significantly lower than that in Maryland and New Jersey. This is consistent with a number of measures of healthcare utilization – total Medicare reimbursement in the last six months of life, mean number of hospital days in the last six months of life, and healthcare intensity index—for all of which Washington was below the national mean and Maryland and New Jersey were above it.¹⁴

Finally, we found evidence of a lower rate of preoperative stress tests and cardiac catheterizations for women despite controlling for age and mortality score. Of course, we did not control directly for cardiovascular comorbidities; as a result, there could be residual confounding. However, these results are consistent with previous findings of gender bias in both pharmacologic management of coronary artery disease (CAD)¹⁵ and diagnostic testing for suspected CAD.¹⁶

We focused on hospitalizations with a primary procedure code to surgically treat hip fracture. We are unable to tell if the cardiac testing of these patients had occurred before or after the procedure. However, we suspect that the vast majority were completed for preoperative evaluation. It is likely that a small subset were done to diagnose and manage cardiac complications that either accompanied the hip fracture or occurred postoperatively. Another limitation is that we cannot determine if a patient had one of these tests recently in the emergency department or as an outpatient.

We also chose to include only patients who actually had hip fracture surgery. It is possible that the testing rate is higher for all patients admitted for hip fracture and that some of these patients did not have surgery because of abnormal cardiac testing. However, we suspect that this is a very small fraction given the high degree of morbidity and mortality associated with untreated hip fracture.

We found a low rate of preoperative cardiac testing in patients hospitalized for hip fracture surgery both in the years before and after the issuance of recommendations intended to curb its use. Although it is reassuring that the volume of low-value testing is lower than we expected, these findings highlight the importance of targeting utilization improvement efforts toward low-value tests and procedures that are more heavily used, since further curbing the use of infrequently utilized tests and procedures will have only a modest impact on overall healthcare expenditure. Our findings highlight the necessity that professional organizations ensure that they focus on true areas of inappropriate utilization. These are the areas in which improvements will have a major impact on healthcare spending. Further research should aim to quantify unwarranted cardiac testing for other inpatient surgeries that are less urgent, as the urgency of hip fracture repair may be driving the relatively low utilization of inpatient cardiac testing.

Disclosures

The authors have nothing to disclose.

Funding

This project was supported by the Johns Hopkins Hospitalist Scholars Fund and the Johns Hopkins School of Medicine Biostatistics, Epidemiology and Data Management (BEAD) Core.

Advance care planning (ACP) is the process wherein patients, in discussions with their healthcare providers, family members, and other loved ones, make individual decisions about their future healthcare or prepare proxies to guide future medical treatment decisions.^1,2 In 2016, the Centers for Medicare and Medicaid Services (CMS) began paying providers for ACP by using billing codes 99497 (first 30 min of ACP) and 99498 (additional 30 min of ACP). According to the CMS, during the first year after the billing codes were introduced, 22,864 providers billed for ACP conversations with 574,621 patients.³ While all adults are eligible, common triggers for ACP include advanced age, serious illness, and functional status changes that confer an increased risk of dying. We explored the early uptake of the ACP billing code in a large national physician practice that provided mandatory education in use of the ACP billing code, offered a small financial incentive for ACP documentation, and primed physicians to reflect on the patient’s risk of dying in the next year at the time of hospital admission.

We analyzed ACP billing for hospitalized adults aged 65 years or above and who were managed by a large national physician practice that employs acute care providers in hospital medicine, emergency medicine and critical care between January 1, 2017 and March 31, 2017. This practice employs approximately 2,500 hospital-based physicians in 250 community hospitals in 38 states. They collect data through handheld and desktop information technology (IT) tools to facilitate coding, billing, and compliance by hospitalists. Hospitalists receive mandatory web-based training in compliance with CMS ACP billing and templated ACP documentation. Additionally, they receive web-based training in serious illness communication skills during the first two years of employment. The training includes didactic content regarding steps for collaborative decision making, words to use during the encounter, and videos of simulated patient encounters demonstrating best practices. Hospitalists also receive a small financial incentive ($20) for each properly documented ACP conversation that meets CMS criteria for ACP code payment.

Beginning in 2017, hospitalists were required to answer the validated Surprise Question⁴ (SQ; “Would you be surprised if the patient died in the next year?”) for all admitted patients aged 65 years and older. The SQ is useful because it is intuitive and not burdensome for physicians to answer. Moreover, it is predictive of mortality. The pooled prognostic characteristics of the SQ across multiple populations for predicting the outcome of death at 6 months to 18 months include a sensitivity of 67.0% (95% confidence interval [CI] 55.7%-76.7%), a specificity of 80.2% (95% CI 73.3%-85.6%), a positive likelihood ratio of 3.4 (95% CI 2.8–4.1), a negative likelihood ratio of 0.41 (95% CI 0.32-0.54), a positive predictive value of 37.1% (95% CI 30.2%-44.6%), and a negative predictive value of 93.1% (95% CI 91.0%-94.8%).⁵ The SQ primed the admitting physician and triggered an “EoL” (end-of-life) icon next to the patient’s name on the hospitalists’ handheld electronic patient census.

We summarized ACP billing rates and used mixed-effects regression to estimate adjusted ACP rates accounting for patient covariates and clustering at the provider and hospital level. Patient covariates included age; answer to the SQ [“yes,” “no,” or “missing”]); and the presence or absence of seven comorbidities: dementia, heart failure, chronic obstructive pulmonary disease, renal failure, liver failure, metastatic cancer, and nonmetastatic cancer. We quantified the magnitude of provider and hospital variation in ACP rates by using the intraclass correlation coefficient (ICC).

In the first quarter of 2017, hospitalists admitted 113,612 patients aged 65 years and older. Hospitalists were prompted to answer the SQ for 73,731 (65%) of the patients. They were not prompted to answer the SQ for 39,881 (35%) of the patients (ie, missing data for the SQ). Reasons for not prompting include delayed implementation at a site and the patient not being admitted to the hospital (eg, managed on observation status). When prompted, hospitalists answered “no” to the SQ for 41,276/73,731 (56%) of admissions.

Only 6,146/113,612 (5.4%) of all admissions involved a billed ACP conversation. Rates were highest among SQ-prompted/answer “no” cases (8.3%) compared with SQ-prompted/answer “yes” cases (4.1%) and non-SQ-prompted cases (3.5%), with all pairwise differences being statistically significant (P values “yes” vs “no” = .0079, “yes” vs not prompted = .0043, “no” vs not prompted < .0001; see Table 1).

In this large national hospital-based physician practice group, the rates of ACP among acute care patients 65 years of age and older were very low despite the use of education and IT- and incentive-based strategies to encourage ACP conversations among seriously ill older adults. Priming physicians to reflect on the patient’s risk of dying at the time of admission was associated with the doubling of ACP rates.

Despite some lawmakers’ concerns that the ACP billing code may be overused and therefore become a financial burden to the Medicare program⁶, we find the very low use of ACP billing in a population for whom having goals of care conversations is critical—seriously ill older adults who the physician would not be surprised if they died in the next year. This gap is significant because these ACP conversations, when they did occur, were associated with a comfort-focused trajectory, including a more than four-fold increase in hospice referral at discharge.

Causal inference is limited because of the observational nature of the study. While we hypothesize that priming the physicians to reflect on prognosis activated them to prioritize ACP, based on a prior scenario-based randomized trial,⁷ illness severity likely drives ACP conversations. Specifically, patients on observation status (who had missing SQ data) and those for whom the physician answered “yes” to the SQ are less sick than other patients. Additional decision-making heuristics in addition to mortality risk may influence ACP conversations, as suggested by the independent influence of diagnoses, such as dementia or cancer, on ACP. Notably, however, the large amounts of unexplained variation at the physician and the hospital levels exceed the amounts explained by any individual observed patient factor.

Other key limitations of this study include the use of ACP billing as a primary outcome rather than observed and documented ACP conversations and the lack of information on the quality of ACP conversations. These findings reflect the uptake of ACP billing rates soon after the code was introduced. ACP billing rates have likely increased since the first quarter of 2017. Future work should explore diffusion and variation in physician-specific use over time. Finally, despite the nationwide sample, findings may not be generalizable to hospitalists who have not received training and financial incentives for ACP billing.

This study reinforces the possibility that variation in ACP conversations may contribute to variation in end-of-life treatment intensity between providers.^8-10 Low ACP rates among even those with high hospitalist-predicted mortality risk and considerable between-provider variation underscore the need for quality improvement interventions to increase hospital-based ACP.

Acknowledgments

The authors thank Jared Wasserman, Maxwell Bessler, Devon Zoller MD, Mark Rudolph MD, Kristi Franz, and Weiping Zhou for their research assistance.

Disclosures

The authors have nothing to disclose.

Funding

National Institute on Aging award P01 AG019783

Advance care planning (ACP) is the process wherein patients, in discussions with their healthcare providers, family members, and other loved ones, make individual decisions about their future healthcare or prepare proxies to guide future medical treatment decisions.^1,2 In 2016, the Centers for Medicare and Medicaid Services (CMS) began paying providers for ACP by using billing codes 99497 (first 30 min of ACP) and 99498 (additional 30 min of ACP). According to the CMS, during the first year after the billing codes were introduced, 22,864 providers billed for ACP conversations with 574,621 patients.³ While all adults are eligible, common triggers for ACP include advanced age, serious illness, and functional status changes that confer an increased risk of dying. We explored the early uptake of the ACP billing code in a large national physician practice that provided mandatory education in use of the ACP billing code, offered a small financial incentive for ACP documentation, and primed physicians to reflect on the patient’s risk of dying in the next year at the time of hospital admission.

We analyzed ACP billing for hospitalized adults aged 65 years or above and who were managed by a large national physician practice that employs acute care providers in hospital medicine, emergency medicine and critical care between January 1, 2017 and March 31, 2017. This practice employs approximately 2,500 hospital-based physicians in 250 community hospitals in 38 states. They collect data through handheld and desktop information technology (IT) tools to facilitate coding, billing, and compliance by hospitalists. Hospitalists receive mandatory web-based training in compliance with CMS ACP billing and templated ACP documentation. Additionally, they receive web-based training in serious illness communication skills during the first two years of employment. The training includes didactic content regarding steps for collaborative decision making, words to use during the encounter, and videos of simulated patient encounters demonstrating best practices. Hospitalists also receive a small financial incentive ($20) for each properly documented ACP conversation that meets CMS criteria for ACP code payment.

Beginning in 2017, hospitalists were required to answer the validated Surprise Question⁴ (SQ; “Would you be surprised if the patient died in the next year?”) for all admitted patients aged 65 years and older. The SQ is useful because it is intuitive and not burdensome for physicians to answer. Moreover, it is predictive of mortality. The pooled prognostic characteristics of the SQ across multiple populations for predicting the outcome of death at 6 months to 18 months include a sensitivity of 67.0% (95% confidence interval [CI] 55.7%-76.7%), a specificity of 80.2% (95% CI 73.3%-85.6%), a positive likelihood ratio of 3.4 (95% CI 2.8–4.1), a negative likelihood ratio of 0.41 (95% CI 0.32-0.54), a positive predictive value of 37.1% (95% CI 30.2%-44.6%), and a negative predictive value of 93.1% (95% CI 91.0%-94.8%).⁵ The SQ primed the admitting physician and triggered an “EoL” (end-of-life) icon next to the patient’s name on the hospitalists’ handheld electronic patient census.

We summarized ACP billing rates and used mixed-effects regression to estimate adjusted ACP rates accounting for patient covariates and clustering at the provider and hospital level. Patient covariates included age; answer to the SQ [“yes,” “no,” or “missing”]); and the presence or absence of seven comorbidities: dementia, heart failure, chronic obstructive pulmonary disease, renal failure, liver failure, metastatic cancer, and nonmetastatic cancer. We quantified the magnitude of provider and hospital variation in ACP rates by using the intraclass correlation coefficient (ICC).

In the first quarter of 2017, hospitalists admitted 113,612 patients aged 65 years and older. Hospitalists were prompted to answer the SQ for 73,731 (65%) of the patients. They were not prompted to answer the SQ for 39,881 (35%) of the patients (ie, missing data for the SQ). Reasons for not prompting include delayed implementation at a site and the patient not being admitted to the hospital (eg, managed on observation status). When prompted, hospitalists answered “no” to the SQ for 41,276/73,731 (56%) of admissions.

Only 6,146/113,612 (5.4%) of all admissions involved a billed ACP conversation. Rates were highest among SQ-prompted/answer “no” cases (8.3%) compared with SQ-prompted/answer “yes” cases (4.1%) and non-SQ-prompted cases (3.5%), with all pairwise differences being statistically significant (P values “yes” vs “no” = .0079, “yes” vs not prompted = .0043, “no” vs not prompted < .0001; see Table 1).

In this large national hospital-based physician practice group, the rates of ACP among acute care patients 65 years of age and older were very low despite the use of education and IT- and incentive-based strategies to encourage ACP conversations among seriously ill older adults. Priming physicians to reflect on the patient’s risk of dying at the time of admission was associated with the doubling of ACP rates.

Despite some lawmakers’ concerns that the ACP billing code may be overused and therefore become a financial burden to the Medicare program⁶, we find the very low use of ACP billing in a population for whom having goals of care conversations is critical—seriously ill older adults who the physician would not be surprised if they died in the next year. This gap is significant because these ACP conversations, when they did occur, were associated with a comfort-focused trajectory, including a more than four-fold increase in hospice referral at discharge.

Causal inference is limited because of the observational nature of the study. While we hypothesize that priming the physicians to reflect on prognosis activated them to prioritize ACP, based on a prior scenario-based randomized trial,⁷ illness severity likely drives ACP conversations. Specifically, patients on observation status (who had missing SQ data) and those for whom the physician answered “yes” to the SQ are less sick than other patients. Additional decision-making heuristics in addition to mortality risk may influence ACP conversations, as suggested by the independent influence of diagnoses, such as dementia or cancer, on ACP. Notably, however, the large amounts of unexplained variation at the physician and the hospital levels exceed the amounts explained by any individual observed patient factor.

Other key limitations of this study include the use of ACP billing as a primary outcome rather than observed and documented ACP conversations and the lack of information on the quality of ACP conversations. These findings reflect the uptake of ACP billing rates soon after the code was introduced. ACP billing rates have likely increased since the first quarter of 2017. Future work should explore diffusion and variation in physician-specific use over time. Finally, despite the nationwide sample, findings may not be generalizable to hospitalists who have not received training and financial incentives for ACP billing.

This study reinforces the possibility that variation in ACP conversations may contribute to variation in end-of-life treatment intensity between providers.^8-10 Low ACP rates among even those with high hospitalist-predicted mortality risk and considerable between-provider variation underscore the need for quality improvement interventions to increase hospital-based ACP.

Acknowledgments

The authors thank Jared Wasserman, Maxwell Bessler, Devon Zoller MD, Mark Rudolph MD, Kristi Franz, and Weiping Zhou for their research assistance.

Disclosures

The authors have nothing to disclose.

Funding

National Institute on Aging award P01 AG019783

Advance care planning (ACP) is the process wherein patients, in discussions with their healthcare providers, family members, and other loved ones, make individual decisions about their future healthcare or prepare proxies to guide future medical treatment decisions.^1,2 In 2016, the Centers for Medicare and Medicaid Services (CMS) began paying providers for ACP by using billing codes 99497 (first 30 min of ACP) and 99498 (additional 30 min of ACP). According to the CMS, during the first year after the billing codes were introduced, 22,864 providers billed for ACP conversations with 574,621 patients.³ While all adults are eligible, common triggers for ACP include advanced age, serious illness, and functional status changes that confer an increased risk of dying. We explored the early uptake of the ACP billing code in a large national physician practice that provided mandatory education in use of the ACP billing code, offered a small financial incentive for ACP documentation, and primed physicians to reflect on the patient’s risk of dying in the next year at the time of hospital admission.

We analyzed ACP billing for hospitalized adults aged 65 years or above and who were managed by a large national physician practice that employs acute care providers in hospital medicine, emergency medicine and critical care between January 1, 2017 and March 31, 2017. This practice employs approximately 2,500 hospital-based physicians in 250 community hospitals in 38 states. They collect data through handheld and desktop information technology (IT) tools to facilitate coding, billing, and compliance by hospitalists. Hospitalists receive mandatory web-based training in compliance with CMS ACP billing and templated ACP documentation. Additionally, they receive web-based training in serious illness communication skills during the first two years of employment. The training includes didactic content regarding steps for collaborative decision making, words to use during the encounter, and videos of simulated patient encounters demonstrating best practices. Hospitalists also receive a small financial incentive ($20) for each properly documented ACP conversation that meets CMS criteria for ACP code payment.

Beginning in 2017, hospitalists were required to answer the validated Surprise Question⁴ (SQ; “Would you be surprised if the patient died in the next year?”) for all admitted patients aged 65 years and older. The SQ is useful because it is intuitive and not burdensome for physicians to answer. Moreover, it is predictive of mortality. The pooled prognostic characteristics of the SQ across multiple populations for predicting the outcome of death at 6 months to 18 months include a sensitivity of 67.0% (95% confidence interval [CI] 55.7%-76.7%), a specificity of 80.2% (95% CI 73.3%-85.6%), a positive likelihood ratio of 3.4 (95% CI 2.8–4.1), a negative likelihood ratio of 0.41 (95% CI 0.32-0.54), a positive predictive value of 37.1% (95% CI 30.2%-44.6%), and a negative predictive value of 93.1% (95% CI 91.0%-94.8%).⁵ The SQ primed the admitting physician and triggered an “EoL” (end-of-life) icon next to the patient’s name on the hospitalists’ handheld electronic patient census.

We summarized ACP billing rates and used mixed-effects regression to estimate adjusted ACP rates accounting for patient covariates and clustering at the provider and hospital level. Patient covariates included age; answer to the SQ [“yes,” “no,” or “missing”]); and the presence or absence of seven comorbidities: dementia, heart failure, chronic obstructive pulmonary disease, renal failure, liver failure, metastatic cancer, and nonmetastatic cancer. We quantified the magnitude of provider and hospital variation in ACP rates by using the intraclass correlation coefficient (ICC).

In the first quarter of 2017, hospitalists admitted 113,612 patients aged 65 years and older. Hospitalists were prompted to answer the SQ for 73,731 (65%) of the patients. They were not prompted to answer the SQ for 39,881 (35%) of the patients (ie, missing data for the SQ). Reasons for not prompting include delayed implementation at a site and the patient not being admitted to the hospital (eg, managed on observation status). When prompted, hospitalists answered “no” to the SQ for 41,276/73,731 (56%) of admissions.

Only 6,146/113,612 (5.4%) of all admissions involved a billed ACP conversation. Rates were highest among SQ-prompted/answer “no” cases (8.3%) compared with SQ-prompted/answer “yes” cases (4.1%) and non-SQ-prompted cases (3.5%), with all pairwise differences being statistically significant (P values “yes” vs “no” = .0079, “yes” vs not prompted = .0043, “no” vs not prompted < .0001; see Table 1).

In this large national hospital-based physician practice group, the rates of ACP among acute care patients 65 years of age and older were very low despite the use of education and IT- and incentive-based strategies to encourage ACP conversations among seriously ill older adults. Priming physicians to reflect on the patient’s risk of dying at the time of admission was associated with the doubling of ACP rates.

Despite some lawmakers’ concerns that the ACP billing code may be overused and therefore become a financial burden to the Medicare program⁶, we find the very low use of ACP billing in a population for whom having goals of care conversations is critical—seriously ill older adults who the physician would not be surprised if they died in the next year. This gap is significant because these ACP conversations, when they did occur, were associated with a comfort-focused trajectory, including a more than four-fold increase in hospice referral at discharge.

Causal inference is limited because of the observational nature of the study. While we hypothesize that priming the physicians to reflect on prognosis activated them to prioritize ACP, based on a prior scenario-based randomized trial,⁷ illness severity likely drives ACP conversations. Specifically, patients on observation status (who had missing SQ data) and those for whom the physician answered “yes” to the SQ are less sick than other patients. Additional decision-making heuristics in addition to mortality risk may influence ACP conversations, as suggested by the independent influence of diagnoses, such as dementia or cancer, on ACP. Notably, however, the large amounts of unexplained variation at the physician and the hospital levels exceed the amounts explained by any individual observed patient factor.

Other key limitations of this study include the use of ACP billing as a primary outcome rather than observed and documented ACP conversations and the lack of information on the quality of ACP conversations. These findings reflect the uptake of ACP billing rates soon after the code was introduced. ACP billing rates have likely increased since the first quarter of 2017. Future work should explore diffusion and variation in physician-specific use over time. Finally, despite the nationwide sample, findings may not be generalizable to hospitalists who have not received training and financial incentives for ACP billing.

This study reinforces the possibility that variation in ACP conversations may contribute to variation in end-of-life treatment intensity between providers.^8-10 Low ACP rates among even those with high hospitalist-predicted mortality risk and considerable between-provider variation underscore the need for quality improvement interventions to increase hospital-based ACP.

Acknowledgments

The authors thank Jared Wasserman, Maxwell Bessler, Devon Zoller MD, Mark Rudolph MD, Kristi Franz, and Weiping Zhou for their research assistance.

Disclosures

The authors have nothing to disclose.

Funding

National Institute on Aging award P01 AG019783

Acute kidney injury (AKI) is a common complication in hospitalized patients, affecting one in five inpatients^1,2 and more than half of patients in intensive care units (ICU).³ The incidence of AKI appears to be increasing over time.⁴ Potential contributing factors include an aging population, rising prevalence of comorbid conditions such as heart failure and chronic kidney disease (CKD), using nephrotoxic agents, and increasing complexity of surgical procedures.^5,6 AKI during a hospital stay is associated with a two to 10-fold increased risk of inhospital mortality,^1,2,7-10 longer hospital length of stay,^7,10 higher risk for hospital readmissions,¹¹ and higher healthcare costs.⁷ Patients who survive an episode of AKI have a higher risk for CKD and dialysis-dependence,⁹ even after an episode of reversible AKI.¹² Despite its clinical importance, several areas of controversy remain regarding the management of AKI and, in particular, the optimal timing of renal replacement therapy (RRT) in patients with AKI. The purpose of this manuscript is to review the approaches to diagnosis and management of AKI in hospitalized patients. We also review recent evidence regarding the timing of dialysis in patients with AKI. This journal recently reviewed the differential diagnosis and diagnostic evaluation of AKI, which is not covered here.¹³

AKI refers to an acute change in kidney function characterized by an increase in serum creatinine and/or a reduction in urine output. It is a clinical syndrome caused by a broad range of etiologies and may be related to primary kidney pathology and/or systemic illness. Until 2004, there was no standard definition for AKI and over 30 different definitions were found in the literature, which resulted in wide variation in the reported incidence and outcomes of AKI and made it challenging to apply an evidence-based approach to patient care. In 2004, the Risk, Injury, Failure, Loss, and End-stage kidney disease (RIFLE)¹⁴ criteria for AKI were proposed, which were modified to the Acute Kidney Injury Network (AKIN)¹⁵ criteria in 2007 (Table 1). Multiple studies show that the RIFLE and AKIN criteria for AKI are associated with higher mortality^1,2,8,10 and increased risk for requiring RRT.^1,10

International clinical practice guidelines for AKI were released by Kidney Disease: Improving Global Outcomes (KDIGO) in 2012, which included a standardized definition of AKI that was adapted from the previously validated RIFLE and AKIN definitions.¹⁶ Patients are considered to have AKI when the serum creatinine rises by as little as 0.3 mg/dL. It is notable that when the baseline serum creatinine is high, there is more inherent variability in the serum creatinine measurement; thus, patients with CKD have a higher risk of being misclassified as having AKI.¹⁷ Although the KDIGO definition for AKI is commonly used in research settings, components of this definition have not been well validated, and it is not widely used in clinical practice. Other renal professional societies still recommend an individualized approach to the diagnosis of AKI, taking into account other factors such as trajectories in kidney function, fluid balance, electrolyte abnormalities, comorbid conditions, and clinical context.^18,19 While we endorse the KDIGO approach to the categorization of AKI severity, in practice, a more patient-centered approach is generally required to guide the optimal approach to determining the etiology of AKI and guiding management.

All patients with AKI should have close monitoring of their serum creatinine and urine output. Noninvasive diagnostic studies (urine microscopy, postvoid residual, and renal ultrasound) should be considered based on the clinical scenario. General management strategies include treatment of the reversible causes of AKI and optimization of volume status, hemodynamics, and nutritional status (Table 2).

Reversible Causes of Acute Kidney Injury

The first step in the treatment of AKI is to identify and treat readily reversible causes of AKI such as volume depletion, hypotension, infection, and urinary obstruction. Nephrotoxins should be avoided and all medications should be reviewed and adjusted for kidney function, particularly those that may affect mental status. Avoid opiates with noxious or active metabolites, including meperidine and morphine. Instead, hydromorphone, fentanyl, and methadone are preferred in patients with AKI. Other commonly used medications that require dose adjustment include gabapentin, baclofen, metoclopramide, H2 antagonists, many commonly prescribed antibiotics (penicillins, most cephalosporins, carbapenems, quinolones, and sulfa drugs), many hypoglycemic agents, and insulin. For patients on RRT, dosing is dependent on dialysis modality. Consultation with a hospital pharmacist is recommended when RRT modalities are initiated or changed.

Intravenous Fluids

Patients with AKI should have their volume status assessed and receive adequate resuscitation with intravenous fluids to promote renal perfusion. However, the optimal type and volume of fluid to give in AKI remains controversial. Colloid-containing solutions are theoretically confined to the intravascular space and should pose a lower risk for pulmonary edema compared with crystalloids. However, these solutions are costly, are not associated with any meaningful benefit,^20-22 and may even be associated with potential harm.^22-27

The most commonly used colloid worldwide is hydroxyethyl starch (HES). Its potential adverse effects include anaphylactoid reactions, coagulopathy, and AKI. HES is cleared by the kidneys and can cause osmotic nephrosis, a form of AKI characterized by vacuole formation and proximal renal tubular damage.²⁸ Randomized controlled trials have shown an increased risk of AKI, RRT use, and mortality in critically ill patients who were resuscitated with HES.^22,26,27 HES is not currently recommended in patients who are critically ill or have impaired kidney function and sepsis guidelines advise against its use.²⁹

In the United States, albumin is the most common colloid-containing solution used for intravascular volume resuscitation. Albumin has been shown to be safe for volume resuscitation in critically ill patients,²⁰ but there is no proven advantage to using albumin over saline with respect to mortality, length of hospital stay, duration of mechanical ventilation, duration of RRT, or number of organ systems failure.^20,21 Furthermore, albumin may be harmful in certain patient populations. In patients with traumatic brain injury, albumin resuscitation is associated with higher mean intracranial pressures²³ and long-term mortality.²⁴ In a retrospective study of patients undergoing cardiac surgery, albumin administration was associated with more than twice the risk of AKI compared with crystalloids.²⁵ In contrast, in patients with cirrhosis, intravenous albumin lowers the rate of AKI when administered in the setting of a large volume paracentesis³⁰ or spontaneous bacterial peritonitis.³¹ Outside of these narrow settings, current evidence does not support the use of intravenous albumin to prevent AKI and we would not endorse the use of intravenous albumin as a part of the treatment paradigm for established AKI.

Many renal and critical care guidelines recommend initial fluid resuscitation with isotonic crystalloids except in specific circumstances (ie, hemorrhagic shock), with consideration of albumin in select cases (ie, severe sepsis or cirrhosis).^16,18,19,29 That stated, the optimal type of crystalloid solution that should be used in resuscitation remains unclear. Because of its low cost, normal (0.9%) saline is the most commonly used solution, but it can result in hyperchloremic metabolic acidosis, which can cause renal vasoconstriction and may be associated with mortality in critically ill patients.³² A prospective study found that administration of chloride-liberal fluids (including normal saline) to critically ill patients was associated with nearly twice the risk of AKI and RRT use compared with chloride-restrictive fluids,³³ but a subsequent trial found no difference in AKI or mortality among patients receiving saline versus a balanced crystalloid (Plasma-Lyte 148).³⁴ A recent pair of large, randomized control trials compared outcomes in patients at a single center who were resuscitated with normal saline versus balanced crystalloid solutions (Lactated Ringer’s or Plasma-Lyte A).^35,36 In critically ill patients, the use of balanced crystalloid solutions was associated with a lower risk of the composite outcome of mortality, new RRT, or persistent kidney impairment, but there were no differences in any of the individual components of the composite outcome.³⁵ In noncritically ill patients, there were no differences in the number of hospital-free days based on the type of crystalloid solution used.³⁶ In the absence of compelling evidence for using balanced crystalloid solutions, we continue to use normal saline for initial fluid resuscitation, but to avoid severe hyperchloremia and acidosis, we will consider switching to a balanced solution (Lactated Ringer’s, Plasma-Lyte, or Normosol) for large volume resuscitation (>2 L), particularly in critically ill patients.

Diuretics

As above, volume status is a key component in the management of patients with AKI. In patients with AKI and hypervolemia, loop diuretics are often given prior to the initiation of RRT. Loop diuretics act on the sodium-potassium-chloride cotransporters in the thick ascending limb of the loop of Henle to increase urinary losses of these ions and urine volume. Loop diuretics are dose-dependent, and often, higher doses are needed (eg, furosemide 100 mg intravenous dose) in patients with AKI, since the diuretic effect depends on the proximal tubular secretion of the drug into the urine. The role of diuretics in AKI is controversial and some observational data suggest an increased mortality risk with diuretic use in patients with AKI.³⁷ In critically ill patients with acute lung injury, diuretic use improved survival, which was attributed to better control of volume overload.³⁸ But, a meta-analysis of 11 randomized controlled trials failed to demonstrate that diuretics directly improved survival or recovery of AKI.³⁹ Moreover, randomized controlled trials found that diuretics given to a patient with AKI requiring RRT did not improve recovery of kidney function.^40,41 The KDIGO guidelines recommend that diuretics should not be routinely used for AKI except in the management of volume overload.¹⁶

Nutritional Targets in Acute Kidney Injury

Critically ill patients have high protein catabolic rates, which put them at increased risk for malnutrition, which in turn is associated with mortality. Patients who receive continuous RRT (CRRT) may lose 5-10 g of protein and 10-15 g of amino acids daily, and these patients may have protein requirements that are twice the usual recommended daily protein intake.¹⁶ But excess protein administration can result in high urea generation and azotemia unrelated to the patient’s kidney function. Blood urea nitrogen may also be disproportionately elevated in conditions where tubular reabsorption of urea is increased, such as in volume depletion, diuretic use, corticosteroid use, and gastrointestinal bleeding. Interpretation of blood urea nitrogen results must be made in the appropriate clinical context, with recognition that azotemia alone may not be a good surrogate marker of the patient’s underlying kidney function. We recommend dietary consultation in critically ill patients with AKI to ensure that adequate, but not excessive, protein is administered.

In patients with AKI, RRT is initiated for control of volume overload, electrolyte abnormalities, acidemia, or uremic symptoms or complications that are refractory to medical management (Table 3). In a nonoliguric patient, fluid and electrolyte abnormalities can oftentimes be managed medically. Patients with oligoanuria (generally defined as urine output less than 400 mL/day or <20 mL/hour), however, require nephrology evaluation for consideration of RRT. Early nephrology consultation (within 48 hours of AKI diagnosis) may be associated with lower dialysis dependence and mortality in critically ill patients with AKI.⁴² The decision to initiate dialysis is individualized based on the patient’s comorbid conditions, urine output, and trajectory of kidney function.

Timing of Renal Replacement Therapy

The optimal timing of dialysis initiation in patients with AKI is not known. Theoretically, earlier initiation of dialysis could allow for better volume and electrolyte control and prevent the development of more serious complications of kidney failure such as uremic seizures, encephalopathy, and pericarditis. However, RRT is associated with its own risks and earlier initiation may expose the patient to unnecessary procedures and complications that might delay renal recovery. A meta-analysis of predominantly observational data found that earlier initiation of RRT in AKI was associated with lower 28-day mortality, greater renal recovery, decreased duration of RRT, and decreased ICU length of stay.⁴³ Subsequently, two prospective trials reported conflicting results regarding associations between dialysis timing and outcomes in patients with severe AKI (Table 4).^44,45

The Early vs Late Initiation of Renal Replacement Therapy in Critically Ill Patients with Acute Kidney Injury (ELAIN) was a prospective, single-center randomized trial in Germany of 231 critically ill, predominantly surgical ICU patients (about half postcardiac surgery) with at least KDIGO stage 2 AKI.⁴⁴ Patients were randomized to early (within eight hours of developing KDIGO stage 2 AKI) or delayed (within 12 hours of developing KDIGO stage 3 AKI) RRT initiation; patients in the early RRT group initiated dialysis on average 20 hours earlier than the patients in the late group. All patients were treated with continuous venovenous hemodiafiltration. Early RRT initiation was associated with a 34% lower risk of mortality at 90 days, shorter hospital length of stay, and shorter RRT duration compared with delayed RRT initiation. There was no difference between groups in dialysis dependence at 90 days, but there was a lower risk of dialysis dependence at one year.⁴⁶The Artificial Kidney Initiation in Kidney Injury Study (AKIKI)⁴⁵ was a prospective, multicenter randomized trial in France that compared early versus delayed strategies of RRT initiation in 620 critically ill, mostly medical ICU patients with severe AKI (KDIGO stage 3). The median time between randomization and RRT initiation was two hours for the early and 57 hours for the delayed strategy groups. There were no differences between groups in length of hospital or ICU stay, vasopressor use, dialysis dependence, or 60-day survival. The early strategy group had a higher incidence of catheter-related bloodstream infections (10% vs 5%) and hypophosphatemia (22% vs 15%) compared with that of the delayed strategy group. Patients in the delayed strategy group regained normal urine output sooner than in the early strategy group. Approximately half of the patients in the delayed strategy group avoided RRT altogether. The authors of AKIKI concluded that there was no benefit to the early strategy of RRT in critically ill patients with severe AKI, and a delayed strategy of RRT initiation may avoid unnecessary RRT and reduce catheter-related infectious complications.

How can we interpret these discrepant results? Although ELAIN found a benefit to earlier RRT initiation in AKI, it has limited generalizability to medical ICU patients, who have higher mortality and whose outcomes might be less affected by dialysis timing. Patients in ELAIN had a high prevalence of congestive heart failure and CKD; it is possible that select patient populations may derive greater benefit from earlier RRT initiation. Although both ELAIN and AKIKI used the standardized criteria for RRT initiation, neither study could incorporate important clinical factors such as trajectory of kidney function, comorbid conditions, or symptoms, which play a significant role in the decision-making process in real-world clinical practice. Additional large-scale, multicenter trials are needed to guide the timing of RRT in critically ill patients with AKI. The Initiation of Dialysis Early Versus Delayed in the ICU (IDEAL-ICU)⁴⁷ and Standard versus Accelerated Initiation of RRT in Acute Kidney Injury (STARRT-AKI)⁴⁸ studies are currently underway and hope to provide clearer guidance regarding the optimal timing of RRT initiation in AKI (Table 4). Until further evidence is available, experts recommend taking into consideration the trajectory of kidney disease, concurrent organ dysfunction, and expected need for fluid and solute control when making decisions regarding RRT initiation in AKI.¹⁶

When RRT is required in patients with AKI, the dialysis modality is often determined by local availability. CRRT and sustained low-efficiency dialysis (SLED) are thought to be better tolerated than intermittent hemodialysis in hemodynamically unstable patients, although a randomized controlled trial could not demonstrate a survival difference between these modalities.⁴⁹ In general, in settings where CRRT or SLED is available, these modalities are favored for patients with hemodynamic instability, but practice patterns vary widely.

Among hospitalized patients, AKI is common and associated with a higher risk of mortality. Although serum creatinine and urine output criteria are used to define AKI, other clinical factors (comorbid conditions, volume status, and trajectory of kidney function decline) can inform the assessment and management of patients with AKI. General strategies for AKI management include treatment of reversible conditions, optimization of volume status, hemodynamics, and nutritional status. The optimal timing of RRT in critically ill patients with AKI is not known, with unclear mortality benefit of earlier dialysis initiation. Two large-scale randomized controlled trials regarding early versus delayed dialysis timing in AKI are currently underway and will hopefully provide clarity in the near future.

Disclosures

Dr. Yu and Dr. Kamal have nothing to disclose. Dr. Chertow is an advisor to DURECT Corporation.

Acute kidney injury (AKI) is a common complication in hospitalized patients, affecting one in five inpatients^1,2 and more than half of patients in intensive care units (ICU).³ The incidence of AKI appears to be increasing over time.⁴ Potential contributing factors include an aging population, rising prevalence of comorbid conditions such as heart failure and chronic kidney disease (CKD), using nephrotoxic agents, and increasing complexity of surgical procedures.^5,6 AKI during a hospital stay is associated with a two to 10-fold increased risk of inhospital mortality,^1,2,7-10 longer hospital length of stay,^7,10 higher risk for hospital readmissions,¹¹ and higher healthcare costs.⁷ Patients who survive an episode of AKI have a higher risk for CKD and dialysis-dependence,⁹ even after an episode of reversible AKI.¹² Despite its clinical importance, several areas of controversy remain regarding the management of AKI and, in particular, the optimal timing of renal replacement therapy (RRT) in patients with AKI. The purpose of this manuscript is to review the approaches to diagnosis and management of AKI in hospitalized patients. We also review recent evidence regarding the timing of dialysis in patients with AKI. This journal recently reviewed the differential diagnosis and diagnostic evaluation of AKI, which is not covered here.¹³

AKI refers to an acute change in kidney function characterized by an increase in serum creatinine and/or a reduction in urine output. It is a clinical syndrome caused by a broad range of etiologies and may be related to primary kidney pathology and/or systemic illness. Until 2004, there was no standard definition for AKI and over 30 different definitions were found in the literature, which resulted in wide variation in the reported incidence and outcomes of AKI and made it challenging to apply an evidence-based approach to patient care. In 2004, the Risk, Injury, Failure, Loss, and End-stage kidney disease (RIFLE)¹⁴ criteria for AKI were proposed, which were modified to the Acute Kidney Injury Network (AKIN)¹⁵ criteria in 2007 (Table 1). Multiple studies show that the RIFLE and AKIN criteria for AKI are associated with higher mortality^1,2,8,10 and increased risk for requiring RRT.^1,10

International clinical practice guidelines for AKI were released by Kidney Disease: Improving Global Outcomes (KDIGO) in 2012, which included a standardized definition of AKI that was adapted from the previously validated RIFLE and AKIN definitions.¹⁶ Patients are considered to have AKI when the serum creatinine rises by as little as 0.3 mg/dL. It is notable that when the baseline serum creatinine is high, there is more inherent variability in the serum creatinine measurement; thus, patients with CKD have a higher risk of being misclassified as having AKI.¹⁷ Although the KDIGO definition for AKI is commonly used in research settings, components of this definition have not been well validated, and it is not widely used in clinical practice. Other renal professional societies still recommend an individualized approach to the diagnosis of AKI, taking into account other factors such as trajectories in kidney function, fluid balance, electrolyte abnormalities, comorbid conditions, and clinical context.^18,19 While we endorse the KDIGO approach to the categorization of AKI severity, in practice, a more patient-centered approach is generally required to guide the optimal approach to determining the etiology of AKI and guiding management.

All patients with AKI should have close monitoring of their serum creatinine and urine output. Noninvasive diagnostic studies (urine microscopy, postvoid residual, and renal ultrasound) should be considered based on the clinical scenario. General management strategies include treatment of the reversible causes of AKI and optimization of volume status, hemodynamics, and nutritional status (Table 2).

Reversible Causes of Acute Kidney Injury

The first step in the treatment of AKI is to identify and treat readily reversible causes of AKI such as volume depletion, hypotension, infection, and urinary obstruction. Nephrotoxins should be avoided and all medications should be reviewed and adjusted for kidney function, particularly those that may affect mental status. Avoid opiates with noxious or active metabolites, including meperidine and morphine. Instead, hydromorphone, fentanyl, and methadone are preferred in patients with AKI. Other commonly used medications that require dose adjustment include gabapentin, baclofen, metoclopramide, H2 antagonists, many commonly prescribed antibiotics (penicillins, most cephalosporins, carbapenems, quinolones, and sulfa drugs), many hypoglycemic agents, and insulin. For patients on RRT, dosing is dependent on dialysis modality. Consultation with a hospital pharmacist is recommended when RRT modalities are initiated or changed.

Intravenous Fluids

Patients with AKI should have their volume status assessed and receive adequate resuscitation with intravenous fluids to promote renal perfusion. However, the optimal type and volume of fluid to give in AKI remains controversial. Colloid-containing solutions are theoretically confined to the intravascular space and should pose a lower risk for pulmonary edema compared with crystalloids. However, these solutions are costly, are not associated with any meaningful benefit,^20-22 and may even be associated with potential harm.^22-27

The most commonly used colloid worldwide is hydroxyethyl starch (HES). Its potential adverse effects include anaphylactoid reactions, coagulopathy, and AKI. HES is cleared by the kidneys and can cause osmotic nephrosis, a form of AKI characterized by vacuole formation and proximal renal tubular damage.²⁸ Randomized controlled trials have shown an increased risk of AKI, RRT use, and mortality in critically ill patients who were resuscitated with HES.^22,26,27 HES is not currently recommended in patients who are critically ill or have impaired kidney function and sepsis guidelines advise against its use.²⁹

In the United States, albumin is the most common colloid-containing solution used for intravascular volume resuscitation. Albumin has been shown to be safe for volume resuscitation in critically ill patients,²⁰ but there is no proven advantage to using albumin over saline with respect to mortality, length of hospital stay, duration of mechanical ventilation, duration of RRT, or number of organ systems failure.^20,21 Furthermore, albumin may be harmful in certain patient populations. In patients with traumatic brain injury, albumin resuscitation is associated with higher mean intracranial pressures²³ and long-term mortality.²⁴ In a retrospective study of patients undergoing cardiac surgery, albumin administration was associated with more than twice the risk of AKI compared with crystalloids.²⁵ In contrast, in patients with cirrhosis, intravenous albumin lowers the rate of AKI when administered in the setting of a large volume paracentesis³⁰ or spontaneous bacterial peritonitis.³¹ Outside of these narrow settings, current evidence does not support the use of intravenous albumin to prevent AKI and we would not endorse the use of intravenous albumin as a part of the treatment paradigm for established AKI.

Many renal and critical care guidelines recommend initial fluid resuscitation with isotonic crystalloids except in specific circumstances (ie, hemorrhagic shock), with consideration of albumin in select cases (ie, severe sepsis or cirrhosis).^16,18,19,29 That stated, the optimal type of crystalloid solution that should be used in resuscitation remains unclear. Because of its low cost, normal (0.9%) saline is the most commonly used solution, but it can result in hyperchloremic metabolic acidosis, which can cause renal vasoconstriction and may be associated with mortality in critically ill patients.³² A prospective study found that administration of chloride-liberal fluids (including normal saline) to critically ill patients was associated with nearly twice the risk of AKI and RRT use compared with chloride-restrictive fluids,³³ but a subsequent trial found no difference in AKI or mortality among patients receiving saline versus a balanced crystalloid (Plasma-Lyte 148).³⁴ A recent pair of large, randomized control trials compared outcomes in patients at a single center who were resuscitated with normal saline versus balanced crystalloid solutions (Lactated Ringer’s or Plasma-Lyte A).^35,36 In critically ill patients, the use of balanced crystalloid solutions was associated with a lower risk of the composite outcome of mortality, new RRT, or persistent kidney impairment, but there were no differences in any of the individual components of the composite outcome.³⁵ In noncritically ill patients, there were no differences in the number of hospital-free days based on the type of crystalloid solution used.³⁶ In the absence of compelling evidence for using balanced crystalloid solutions, we continue to use normal saline for initial fluid resuscitation, but to avoid severe hyperchloremia and acidosis, we will consider switching to a balanced solution (Lactated Ringer’s, Plasma-Lyte, or Normosol) for large volume resuscitation (>2 L), particularly in critically ill patients.

Diuretics

As above, volume status is a key component in the management of patients with AKI. In patients with AKI and hypervolemia, loop diuretics are often given prior to the initiation of RRT. Loop diuretics act on the sodium-potassium-chloride cotransporters in the thick ascending limb of the loop of Henle to increase urinary losses of these ions and urine volume. Loop diuretics are dose-dependent, and often, higher doses are needed (eg, furosemide 100 mg intravenous dose) in patients with AKI, since the diuretic effect depends on the proximal tubular secretion of the drug into the urine. The role of diuretics in AKI is controversial and some observational data suggest an increased mortality risk with diuretic use in patients with AKI.³⁷ In critically ill patients with acute lung injury, diuretic use improved survival, which was attributed to better control of volume overload.³⁸ But, a meta-analysis of 11 randomized controlled trials failed to demonstrate that diuretics directly improved survival or recovery of AKI.³⁹ Moreover, randomized controlled trials found that diuretics given to a patient with AKI requiring RRT did not improve recovery of kidney function.^40,41 The KDIGO guidelines recommend that diuretics should not be routinely used for AKI except in the management of volume overload.¹⁶

Nutritional Targets in Acute Kidney Injury

Critically ill patients have high protein catabolic rates, which put them at increased risk for malnutrition, which in turn is associated with mortality. Patients who receive continuous RRT (CRRT) may lose 5-10 g of protein and 10-15 g of amino acids daily, and these patients may have protein requirements that are twice the usual recommended daily protein intake.¹⁶ But excess protein administration can result in high urea generation and azotemia unrelated to the patient’s kidney function. Blood urea nitrogen may also be disproportionately elevated in conditions where tubular reabsorption of urea is increased, such as in volume depletion, diuretic use, corticosteroid use, and gastrointestinal bleeding. Interpretation of blood urea nitrogen results must be made in the appropriate clinical context, with recognition that azotemia alone may not be a good surrogate marker of the patient’s underlying kidney function. We recommend dietary consultation in critically ill patients with AKI to ensure that adequate, but not excessive, protein is administered.

In patients with AKI, RRT is initiated for control of volume overload, electrolyte abnormalities, acidemia, or uremic symptoms or complications that are refractory to medical management (Table 3). In a nonoliguric patient, fluid and electrolyte abnormalities can oftentimes be managed medically. Patients with oligoanuria (generally defined as urine output less than 400 mL/day or <20 mL/hour), however, require nephrology evaluation for consideration of RRT. Early nephrology consultation (within 48 hours of AKI diagnosis) may be associated with lower dialysis dependence and mortality in critically ill patients with AKI.⁴² The decision to initiate dialysis is individualized based on the patient’s comorbid conditions, urine output, and trajectory of kidney function.

Timing of Renal Replacement Therapy

The optimal timing of dialysis initiation in patients with AKI is not known. Theoretically, earlier initiation of dialysis could allow for better volume and electrolyte control and prevent the development of more serious complications of kidney failure such as uremic seizures, encephalopathy, and pericarditis. However, RRT is associated with its own risks and earlier initiation may expose the patient to unnecessary procedures and complications that might delay renal recovery. A meta-analysis of predominantly observational data found that earlier initiation of RRT in AKI was associated with lower 28-day mortality, greater renal recovery, decreased duration of RRT, and decreased ICU length of stay.⁴³ Subsequently, two prospective trials reported conflicting results regarding associations between dialysis timing and outcomes in patients with severe AKI (Table 4).^44,45

The Early vs Late Initiation of Renal Replacement Therapy in Critically Ill Patients with Acute Kidney Injury (ELAIN) was a prospective, single-center randomized trial in Germany of 231 critically ill, predominantly surgical ICU patients (about half postcardiac surgery) with at least KDIGO stage 2 AKI.⁴⁴ Patients were randomized to early (within eight hours of developing KDIGO stage 2 AKI) or delayed (within 12 hours of developing KDIGO stage 3 AKI) RRT initiation; patients in the early RRT group initiated dialysis on average 20 hours earlier than the patients in the late group. All patients were treated with continuous venovenous hemodiafiltration. Early RRT initiation was associated with a 34% lower risk of mortality at 90 days, shorter hospital length of stay, and shorter RRT duration compared with delayed RRT initiation. There was no difference between groups in dialysis dependence at 90 days, but there was a lower risk of dialysis dependence at one year.⁴⁶The Artificial Kidney Initiation in Kidney Injury Study (AKIKI)⁴⁵ was a prospective, multicenter randomized trial in France that compared early versus delayed strategies of RRT initiation in 620 critically ill, mostly medical ICU patients with severe AKI (KDIGO stage 3). The median time between randomization and RRT initiation was two hours for the early and 57 hours for the delayed strategy groups. There were no differences between groups in length of hospital or ICU stay, vasopressor use, dialysis dependence, or 60-day survival. The early strategy group had a higher incidence of catheter-related bloodstream infections (10% vs 5%) and hypophosphatemia (22% vs 15%) compared with that of the delayed strategy group. Patients in the delayed strategy group regained normal urine output sooner than in the early strategy group. Approximately half of the patients in the delayed strategy group avoided RRT altogether. The authors of AKIKI concluded that there was no benefit to the early strategy of RRT in critically ill patients with severe AKI, and a delayed strategy of RRT initiation may avoid unnecessary RRT and reduce catheter-related infectious complications.

How can we interpret these discrepant results? Although ELAIN found a benefit to earlier RRT initiation in AKI, it has limited generalizability to medical ICU patients, who have higher mortality and whose outcomes might be less affected by dialysis timing. Patients in ELAIN had a high prevalence of congestive heart failure and CKD; it is possible that select patient populations may derive greater benefit from earlier RRT initiation. Although both ELAIN and AKIKI used the standardized criteria for RRT initiation, neither study could incorporate important clinical factors such as trajectory of kidney function, comorbid conditions, or symptoms, which play a significant role in the decision-making process in real-world clinical practice. Additional large-scale, multicenter trials are needed to guide the timing of RRT in critically ill patients with AKI. The Initiation of Dialysis Early Versus Delayed in the ICU (IDEAL-ICU)⁴⁷ and Standard versus Accelerated Initiation of RRT in Acute Kidney Injury (STARRT-AKI)⁴⁸ studies are currently underway and hope to provide clearer guidance regarding the optimal timing of RRT initiation in AKI (Table 4). Until further evidence is available, experts recommend taking into consideration the trajectory of kidney disease, concurrent organ dysfunction, and expected need for fluid and solute control when making decisions regarding RRT initiation in AKI.¹⁶

When RRT is required in patients with AKI, the dialysis modality is often determined by local availability. CRRT and sustained low-efficiency dialysis (SLED) are thought to be better tolerated than intermittent hemodialysis in hemodynamically unstable patients, although a randomized controlled trial could not demonstrate a survival difference between these modalities.⁴⁹ In general, in settings where CRRT or SLED is available, these modalities are favored for patients with hemodynamic instability, but practice patterns vary widely.

Among hospitalized patients, AKI is common and associated with a higher risk of mortality. Although serum creatinine and urine output criteria are used to define AKI, other clinical factors (comorbid conditions, volume status, and trajectory of kidney function decline) can inform the assessment and management of patients with AKI. General strategies for AKI management include treatment of reversible conditions, optimization of volume status, hemodynamics, and nutritional status. The optimal timing of RRT in critically ill patients with AKI is not known, with unclear mortality benefit of earlier dialysis initiation. Two large-scale randomized controlled trials regarding early versus delayed dialysis timing in AKI are currently underway and will hopefully provide clarity in the near future.

Disclosures

Dr. Yu and Dr. Kamal have nothing to disclose. Dr. Chertow is an advisor to DURECT Corporation.

Acute kidney injury (AKI) is a common complication in hospitalized patients, affecting one in five inpatients^1,2 and more than half of patients in intensive care units (ICU).³ The incidence of AKI appears to be increasing over time.⁴ Potential contributing factors include an aging population, rising prevalence of comorbid conditions such as heart failure and chronic kidney disease (CKD), using nephrotoxic agents, and increasing complexity of surgical procedures.^5,6 AKI during a hospital stay is associated with a two to 10-fold increased risk of inhospital mortality,^1,2,7-10 longer hospital length of stay,^7,10 higher risk for hospital readmissions,¹¹ and higher healthcare costs.⁷ Patients who survive an episode of AKI have a higher risk for CKD and dialysis-dependence,⁹ even after an episode of reversible AKI.¹² Despite its clinical importance, several areas of controversy remain regarding the management of AKI and, in particular, the optimal timing of renal replacement therapy (RRT) in patients with AKI. The purpose of this manuscript is to review the approaches to diagnosis and management of AKI in hospitalized patients. We also review recent evidence regarding the timing of dialysis in patients with AKI. This journal recently reviewed the differential diagnosis and diagnostic evaluation of AKI, which is not covered here.¹³

AKI refers to an acute change in kidney function characterized by an increase in serum creatinine and/or a reduction in urine output. It is a clinical syndrome caused by a broad range of etiologies and may be related to primary kidney pathology and/or systemic illness. Until 2004, there was no standard definition for AKI and over 30 different definitions were found in the literature, which resulted in wide variation in the reported incidence and outcomes of AKI and made it challenging to apply an evidence-based approach to patient care. In 2004, the Risk, Injury, Failure, Loss, and End-stage kidney disease (RIFLE)¹⁴ criteria for AKI were proposed, which were modified to the Acute Kidney Injury Network (AKIN)¹⁵ criteria in 2007 (Table 1). Multiple studies show that the RIFLE and AKIN criteria for AKI are associated with higher mortality^1,2,8,10 and increased risk for requiring RRT.^1,10

International clinical practice guidelines for AKI were released by Kidney Disease: Improving Global Outcomes (KDIGO) in 2012, which included a standardized definition of AKI that was adapted from the previously validated RIFLE and AKIN definitions.¹⁶ Patients are considered to have AKI when the serum creatinine rises by as little as 0.3 mg/dL. It is notable that when the baseline serum creatinine is high, there is more inherent variability in the serum creatinine measurement; thus, patients with CKD have a higher risk of being misclassified as having AKI.¹⁷ Although the KDIGO definition for AKI is commonly used in research settings, components of this definition have not been well validated, and it is not widely used in clinical practice. Other renal professional societies still recommend an individualized approach to the diagnosis of AKI, taking into account other factors such as trajectories in kidney function, fluid balance, electrolyte abnormalities, comorbid conditions, and clinical context.^18,19 While we endorse the KDIGO approach to the categorization of AKI severity, in practice, a more patient-centered approach is generally required to guide the optimal approach to determining the etiology of AKI and guiding management.

All patients with AKI should have close monitoring of their serum creatinine and urine output. Noninvasive diagnostic studies (urine microscopy, postvoid residual, and renal ultrasound) should be considered based on the clinical scenario. General management strategies include treatment of the reversible causes of AKI and optimization of volume status, hemodynamics, and nutritional status (Table 2).

Reversible Causes of Acute Kidney Injury

The first step in the treatment of AKI is to identify and treat readily reversible causes of AKI such as volume depletion, hypotension, infection, and urinary obstruction. Nephrotoxins should be avoided and all medications should be reviewed and adjusted for kidney function, particularly those that may affect mental status. Avoid opiates with noxious or active metabolites, including meperidine and morphine. Instead, hydromorphone, fentanyl, and methadone are preferred in patients with AKI. Other commonly used medications that require dose adjustment include gabapentin, baclofen, metoclopramide, H2 antagonists, many commonly prescribed antibiotics (penicillins, most cephalosporins, carbapenems, quinolones, and sulfa drugs), many hypoglycemic agents, and insulin. For patients on RRT, dosing is dependent on dialysis modality. Consultation with a hospital pharmacist is recommended when RRT modalities are initiated or changed.

Intravenous Fluids

Patients with AKI should have their volume status assessed and receive adequate resuscitation with intravenous fluids to promote renal perfusion. However, the optimal type and volume of fluid to give in AKI remains controversial. Colloid-containing solutions are theoretically confined to the intravascular space and should pose a lower risk for pulmonary edema compared with crystalloids. However, these solutions are costly, are not associated with any meaningful benefit,^20-22 and may even be associated with potential harm.^22-27

The most commonly used colloid worldwide is hydroxyethyl starch (HES). Its potential adverse effects include anaphylactoid reactions, coagulopathy, and AKI. HES is cleared by the kidneys and can cause osmotic nephrosis, a form of AKI characterized by vacuole formation and proximal renal tubular damage.²⁸ Randomized controlled trials have shown an increased risk of AKI, RRT use, and mortality in critically ill patients who were resuscitated with HES.^22,26,27 HES is not currently recommended in patients who are critically ill or have impaired kidney function and sepsis guidelines advise against its use.²⁹

In the United States, albumin is the most common colloid-containing solution used for intravascular volume resuscitation. Albumin has been shown to be safe for volume resuscitation in critically ill patients,²⁰ but there is no proven advantage to using albumin over saline with respect to mortality, length of hospital stay, duration of mechanical ventilation, duration of RRT, or number of organ systems failure.^20,21 Furthermore, albumin may be harmful in certain patient populations. In patients with traumatic brain injury, albumin resuscitation is associated with higher mean intracranial pressures²³ and long-term mortality.²⁴ In a retrospective study of patients undergoing cardiac surgery, albumin administration was associated with more than twice the risk of AKI compared with crystalloids.²⁵ In contrast, in patients with cirrhosis, intravenous albumin lowers the rate of AKI when administered in the setting of a large volume paracentesis³⁰ or spontaneous bacterial peritonitis.³¹ Outside of these narrow settings, current evidence does not support the use of intravenous albumin to prevent AKI and we would not endorse the use of intravenous albumin as a part of the treatment paradigm for established AKI.

Many renal and critical care guidelines recommend initial fluid resuscitation with isotonic crystalloids except in specific circumstances (ie, hemorrhagic shock), with consideration of albumin in select cases (ie, severe sepsis or cirrhosis).^16,18,19,29 That stated, the optimal type of crystalloid solution that should be used in resuscitation remains unclear. Because of its low cost, normal (0.9%) saline is the most commonly used solution, but it can result in hyperchloremic metabolic acidosis, which can cause renal vasoconstriction and may be associated with mortality in critically ill patients.³² A prospective study found that administration of chloride-liberal fluids (including normal saline) to critically ill patients was associated with nearly twice the risk of AKI and RRT use compared with chloride-restrictive fluids,³³ but a subsequent trial found no difference in AKI or mortality among patients receiving saline versus a balanced crystalloid (Plasma-Lyte 148).³⁴ A recent pair of large, randomized control trials compared outcomes in patients at a single center who were resuscitated with normal saline versus balanced crystalloid solutions (Lactated Ringer’s or Plasma-Lyte A).^35,36 In critically ill patients, the use of balanced crystalloid solutions was associated with a lower risk of the composite outcome of mortality, new RRT, or persistent kidney impairment, but there were no differences in any of the individual components of the composite outcome.³⁵ In noncritically ill patients, there were no differences in the number of hospital-free days based on the type of crystalloid solution used.³⁶ In the absence of compelling evidence for using balanced crystalloid solutions, we continue to use normal saline for initial fluid resuscitation, but to avoid severe hyperchloremia and acidosis, we will consider switching to a balanced solution (Lactated Ringer’s, Plasma-Lyte, or Normosol) for large volume resuscitation (>2 L), particularly in critically ill patients.

Diuretics

As above, volume status is a key component in the management of patients with AKI. In patients with AKI and hypervolemia, loop diuretics are often given prior to the initiation of RRT. Loop diuretics act on the sodium-potassium-chloride cotransporters in the thick ascending limb of the loop of Henle to increase urinary losses of these ions and urine volume. Loop diuretics are dose-dependent, and often, higher doses are needed (eg, furosemide 100 mg intravenous dose) in patients with AKI, since the diuretic effect depends on the proximal tubular secretion of the drug into the urine. The role of diuretics in AKI is controversial and some observational data suggest an increased mortality risk with diuretic use in patients with AKI.³⁷ In critically ill patients with acute lung injury, diuretic use improved survival, which was attributed to better control of volume overload.³⁸ But, a meta-analysis of 11 randomized controlled trials failed to demonstrate that diuretics directly improved survival or recovery of AKI.³⁹ Moreover, randomized controlled trials found that diuretics given to a patient with AKI requiring RRT did not improve recovery of kidney function.^40,41 The KDIGO guidelines recommend that diuretics should not be routinely used for AKI except in the management of volume overload.¹⁶

Nutritional Targets in Acute Kidney Injury

Critically ill patients have high protein catabolic rates, which put them at increased risk for malnutrition, which in turn is associated with mortality. Patients who receive continuous RRT (CRRT) may lose 5-10 g of protein and 10-15 g of amino acids daily, and these patients may have protein requirements that are twice the usual recommended daily protein intake.¹⁶ But excess protein administration can result in high urea generation and azotemia unrelated to the patient’s kidney function. Blood urea nitrogen may also be disproportionately elevated in conditions where tubular reabsorption of urea is increased, such as in volume depletion, diuretic use, corticosteroid use, and gastrointestinal bleeding. Interpretation of blood urea nitrogen results must be made in the appropriate clinical context, with recognition that azotemia alone may not be a good surrogate marker of the patient’s underlying kidney function. We recommend dietary consultation in critically ill patients with AKI to ensure that adequate, but not excessive, protein is administered.

In patients with AKI, RRT is initiated for control of volume overload, electrolyte abnormalities, acidemia, or uremic symptoms or complications that are refractory to medical management (Table 3). In a nonoliguric patient, fluid and electrolyte abnormalities can oftentimes be managed medically. Patients with oligoanuria (generally defined as urine output less than 400 mL/day or <20 mL/hour), however, require nephrology evaluation for consideration of RRT. Early nephrology consultation (within 48 hours of AKI diagnosis) may be associated with lower dialysis dependence and mortality in critically ill patients with AKI.⁴² The decision to initiate dialysis is individualized based on the patient’s comorbid conditions, urine output, and trajectory of kidney function.

Timing of Renal Replacement Therapy

The optimal timing of dialysis initiation in patients with AKI is not known. Theoretically, earlier initiation of dialysis could allow for better volume and electrolyte control and prevent the development of more serious complications of kidney failure such as uremic seizures, encephalopathy, and pericarditis. However, RRT is associated with its own risks and earlier initiation may expose the patient to unnecessary procedures and complications that might delay renal recovery. A meta-analysis of predominantly observational data found that earlier initiation of RRT in AKI was associated with lower 28-day mortality, greater renal recovery, decreased duration of RRT, and decreased ICU length of stay.⁴³ Subsequently, two prospective trials reported conflicting results regarding associations between dialysis timing and outcomes in patients with severe AKI (Table 4).^44,45

The Early vs Late Initiation of Renal Replacement Therapy in Critically Ill Patients with Acute Kidney Injury (ELAIN) was a prospective, single-center randomized trial in Germany of 231 critically ill, predominantly surgical ICU patients (about half postcardiac surgery) with at least KDIGO stage 2 AKI.⁴⁴ Patients were randomized to early (within eight hours of developing KDIGO stage 2 AKI) or delayed (within 12 hours of developing KDIGO stage 3 AKI) RRT initiation; patients in the early RRT group initiated dialysis on average 20 hours earlier than the patients in the late group. All patients were treated with continuous venovenous hemodiafiltration. Early RRT initiation was associated with a 34% lower risk of mortality at 90 days, shorter hospital length of stay, and shorter RRT duration compared with delayed RRT initiation. There was no difference between groups in dialysis dependence at 90 days, but there was a lower risk of dialysis dependence at one year.⁴⁶The Artificial Kidney Initiation in Kidney Injury Study (AKIKI)⁴⁵ was a prospective, multicenter randomized trial in France that compared early versus delayed strategies of RRT initiation in 620 critically ill, mostly medical ICU patients with severe AKI (KDIGO stage 3). The median time between randomization and RRT initiation was two hours for the early and 57 hours for the delayed strategy groups. There were no differences between groups in length of hospital or ICU stay, vasopressor use, dialysis dependence, or 60-day survival. The early strategy group had a higher incidence of catheter-related bloodstream infections (10% vs 5%) and hypophosphatemia (22% vs 15%) compared with that of the delayed strategy group. Patients in the delayed strategy group regained normal urine output sooner than in the early strategy group. Approximately half of the patients in the delayed strategy group avoided RRT altogether. The authors of AKIKI concluded that there was no benefit to the early strategy of RRT in critically ill patients with severe AKI, and a delayed strategy of RRT initiation may avoid unnecessary RRT and reduce catheter-related infectious complications.

How can we interpret these discrepant results? Although ELAIN found a benefit to earlier RRT initiation in AKI, it has limited generalizability to medical ICU patients, who have higher mortality and whose outcomes might be less affected by dialysis timing. Patients in ELAIN had a high prevalence of congestive heart failure and CKD; it is possible that select patient populations may derive greater benefit from earlier RRT initiation. Although both ELAIN and AKIKI used the standardized criteria for RRT initiation, neither study could incorporate important clinical factors such as trajectory of kidney function, comorbid conditions, or symptoms, which play a significant role in the decision-making process in real-world clinical practice. Additional large-scale, multicenter trials are needed to guide the timing of RRT in critically ill patients with AKI. The Initiation of Dialysis Early Versus Delayed in the ICU (IDEAL-ICU)⁴⁷ and Standard versus Accelerated Initiation of RRT in Acute Kidney Injury (STARRT-AKI)⁴⁸ studies are currently underway and hope to provide clearer guidance regarding the optimal timing of RRT initiation in AKI (Table 4). Until further evidence is available, experts recommend taking into consideration the trajectory of kidney disease, concurrent organ dysfunction, and expected need for fluid and solute control when making decisions regarding RRT initiation in AKI.¹⁶

When RRT is required in patients with AKI, the dialysis modality is often determined by local availability. CRRT and sustained low-efficiency dialysis (SLED) are thought to be better tolerated than intermittent hemodialysis in hemodynamically unstable patients, although a randomized controlled trial could not demonstrate a survival difference between these modalities.⁴⁹ In general, in settings where CRRT or SLED is available, these modalities are favored for patients with hemodynamic instability, but practice patterns vary widely.

Among hospitalized patients, AKI is common and associated with a higher risk of mortality. Although serum creatinine and urine output criteria are used to define AKI, other clinical factors (comorbid conditions, volume status, and trajectory of kidney function decline) can inform the assessment and management of patients with AKI. General strategies for AKI management include treatment of reversible conditions, optimization of volume status, hemodynamics, and nutritional status. The optimal timing of RRT in critically ill patients with AKI is not known, with unclear mortality benefit of earlier dialysis initiation. Two large-scale randomized controlled trials regarding early versus delayed dialysis timing in AKI are currently underway and will hopefully provide clarity in the near future.

Disclosures

Dr. Yu and Dr. Kamal have nothing to disclose. Dr. Chertow is an advisor to DURECT Corporation.

The “Things We Do for No Reason” series reviews practices which have become common parts of hospital care but which may provide little value to our patients. Practices reviewed in the TWDFNR series do not represent “black and white” conclusions or clinical practice standards, but are meant as a starting place for research and active discussions among hospitalists and patients. We invite you to be part of that discussion. https://www.choosingwisely.org/

A 34-year-old man is admitted for a complicated urinary tract infection related to a chronic in-dwelling Foley catheter. The patient suffered a spinal cord injury at the C4/C5 level as a result of a motor vehicle accident 10 years ago and is confined to a motorized wheelchair. He is an engineer and lives independently but has caregivers. His body mass index (BMI) is 18.5 kg/m², and he reports his weight has been stable. He has slight muscle atrophy of the biceps, triceps, interosseous muscles, and quadriceps. The patient reports that he eats well, has no chronic conditions, and has not had any gastrointestinal symptoms (eg, anorexia, nausea, diarrhea) over the last six months. You consider whether to order a serum prealbumin test to assess for possible malnutrition.

The presence of malnutrition in hospitalized patients is widely recognized as an independent predictor of hospital mortality.¹ According to the American Society for Parenteral and Enteral Nutrition (ASPEN), malnutrition is defined as “an acute, subacute or chronic state of nutrition, in which varying degrees of overnutrition or undernutrition with or without inflammatory activity have led to a change in body composition and diminished function.”² In one large European study, patients screening positive for being at risk of malnutrition had a 12-fold increase in hospital mortality.¹

Inpatient malnutrition is remarkably underdocumented. Studies using chart reviews have found a prevalence of malnutrition in hospitalized patients of between 20% and 50%, and only 3% of hospital discharges are associated with a diagnostic code for malnutrition.^3–5 Appropriate diagnosis and documentation of malnutrition is important given the profound prognostic and management implications of a malnutrition diagnosis. Appropriate documentation benefits health systems as malnutrition documentation increases expected mortality, thereby improving the observed-to-expected mortality ratio.

Serum prealbumin testing is widely available and frequently ordered in the inpatient setting. In a query we performed of the large aggregate Cerner Electronic Health Record database, HealthFacts, which includes data from inpatient encounters for approximately 700 United States hospitals, prealbumin tests were ordered 129,152 times in 2015. This activity corresponds to estimated total charges of $2,562,375 based on the 2015 clinical laboratory fee schedule.⁶

Prealbumin is synthesized in the liver and released into circulation prior to excretion by the kidneys and gastrointestinal tract. Prealbumin transports thyroxine, triiodothyronine, and holo-retinol binding protein and, as a result, is also known as transthyretin.⁷ It was first proposed as a nutritional marker in 1972 with the publication of a study that showed low levels of prealbumin in 40 children with kwashiorkor that improved with intensive dietary supplementation.⁸ The shorter half-life of prealbumin (2.5 days) as compared with other identified nutritional markers, such as albumin, indicate that it would be suitable for detecting rapid changes in nutritional status.

Prealbumin Is Not Specific

An ideal nutritional marker should be specific enough that changes in this marker reflect changes in nutritional status.⁹ While there are many systemic factors that affect nutritional markers, such as prealbumin (Table 1), the acute phase response triggered by inflammation is the most significant confounder in the acutely ill hospitalized patient.⁹ This response to infection, stress, and malignancy leads to an increase in proinflammatory cytokines, increased liver synthesis of inflammatory proteins, such as C-reactive protein (CRP), and increased vascular permeability. Prealbumin is a negative acute phase reactant that decreases in concentration during the stress response due to slowed synthesis and extravasation.⁹ In a study of 24 patients with severe sepsis and trauma, levels of prealbumin inversely correlated with CRP, a reflection of the stress response, and returned to normal when CRP levels normalized. Neither prealbumin nor CRP, however, correlated with total body protein changes.¹⁰ Unfortunately, many studies supporting the use of prealbumin as a nutritional marker do not address the role of the acute phase response in their results. These studies include the original report on prealbumin in kwashiorkor, a condition known to be associated with a high rate of infectious diseases that can trigger the acute phase response.⁹ A consensus statement from the Academy of Nutrition and Dietetics (AND) and ASPEN noted that prealbumin is an indicator of inflammation and lacks the specificity to diagnose malnutrition.¹¹

Prealbumin Is Not Sensitive

A sensitive laboratory test for malnutrition should allow for detection of malnutrition at an early stage.⁹ However, patients who demonstrate severe malnutrition without a coexisting inflammatory state do not consistently show low levels of prealbumin. In a systematic review of 20 studies in nondiseased malnourished patients, only two studies, both of which assessed patients with anorexia nervosa, had a mean prealbumin below normal (<20 mg/dL), and this finding corresponded to patient populations with mean BMIs less than 12 kg/m². More importantly, normal prealbumin levels were seen in groups of patients with a mean BMI as low as 12.9 kg/m².¹² Analysis by AND found insufficient evidence to support a correlation between prealbumin and weight loss in anorexia nervosa, calorie restricted diets, or starvation.¹³ The data suggest that prealbumin lacks sufficient sensitivity to consistently detect cases of malnutrition easily diagnosed by history and/or physical exam.

Prealbumin Is Not Consistently Responsive to Nutritional Interventions

An accurate marker for malnutrition should improve when nutritional intervention results in adequate nutritional intake.⁹ While some studies have shown improvements in prealbumin in the setting of a nutritional intervention, many of these works are subject to the same limitations related to specificity and lack of control for concurrent inflammatory processes. In a retrospective study, prealbumin increased significantly in 102 patients receiving TPN for one week. Unfortunately, patients with renal or hepatic disease were excluded, and the role of inflammation was not assessed.¹⁴ Institutionalized patients with Alzheimer’s disease and normal CRP levels showed a statistically significant increase in weight gain, arm muscle circumference, and triceps skin-fold thickness following a nutritional program without a notable change in prealbumin.¹⁵ In a study assessing the relationship of prealbumin, CRP, and nutritional intake, critically ill populations receiving less than or greater than 60% of their estimated caloric needs showed no significant difference in prealbumin. In fact, prealbumin levels were only correlated with CRP levels.¹⁶ This finding argues against the routine use of prealbumin for nutrition monitoring in the acutely ill hospitalized patient.

Prealbumin Is Not Consistently Correlated with Health Outcomes

Even if prealbumin increased consistently in response to nutritional intervention, whether this change corresponds to an improvement in clinical outcomes has yet to be demonstrated.⁹ In 2005, Koretz reviewed 99 clinical trials and concluded that even when changes in nutritional markers are seen with nutritional support, the “changes in nutritional markers do not predict clinical outcomes.”¹⁷

Given the lack of a suitable biologic assay to identify malnutrition, dieticians and clinicians must rely on other means to assess malnutrition. Professional societies, including ASPEN and the European Society for Clinical Nutrition and Metabolism, have proposed different guidelines for the screening and assessment of malnutrition (Table 2).^11,18 In 2016, these organizations, along with the Latin American Federation of Nutritional Therapy, Clinical Nutrition, and Metabolism and the Parenteral and Enteral Nutrition Society of Asia, formed The Global Leadership Initiative on Malnutrition (GLIM). In 2017, the GLIM taskforce agreed on clinically relevant diagnostic variables for the screening and assessment of malnutrition, including reduced food intake (anorexia), nonvolitional weight loss, (reduced) lean mass, status of disease burden and inflammation, and low body mass index or underweight status.¹⁹

• Do not use prealbumin to screen for or diagnose malnutrition.
• Consult with local dietitians to ensure that your institutional approach is in agreement with consensus recommendations.

In revisiting the case above, the patient does not have clear evidence of malnutrition based on his history (stable weight and good reported nutritional intake), although he does have a low BMI of 18.5 kg/m². Rather than prealbumin testing, which would likely be low secondary to the acute phase response, he would better benefit from a nutrition-focused history and physical exam.

The uncertainties faced by clinicians in diagnosing malnutrition cannot readily be resolved by relying on a solitary laboratory marker (eg, prealbumin) or a stand-alone assessment protocol. The data obtained reflect the need for multidisciplinary teams of dieticians and clinicians to contextualize each patient’s medical history and ensure that the selected metrics are used appropriately to aid in diagnosis and documentation. We advocate that clinicians not routinely use prealbumin to screen for, confirm the diagnosis of, or assess the severity of malnutrition in the hospitalized patient.

Do you think this is a low-value practice? Is this truly a “Thing We Do for No Reason?” Share what you do in your practice and join in the conversation online by retweeting it on Twitter (#TWDFNR) and liking it on Facebook. We invite you to propose ideas for other “Things We Do for No Reason” topics by emailing [email protected].

Disclosures

The authors have nothing to disclose.

The “Things We Do for No Reason” series reviews practices which have become common parts of hospital care but which may provide little value to our patients. Practices reviewed in the TWDFNR series do not represent “black and white” conclusions or clinical practice standards, but are meant as a starting place for research and active discussions among hospitalists and patients. We invite you to be part of that discussion. https://www.choosingwisely.org/

A 34-year-old man is admitted for a complicated urinary tract infection related to a chronic in-dwelling Foley catheter. The patient suffered a spinal cord injury at the C4/C5 level as a result of a motor vehicle accident 10 years ago and is confined to a motorized wheelchair. He is an engineer and lives independently but has caregivers. His body mass index (BMI) is 18.5 kg/m², and he reports his weight has been stable. He has slight muscle atrophy of the biceps, triceps, interosseous muscles, and quadriceps. The patient reports that he eats well, has no chronic conditions, and has not had any gastrointestinal symptoms (eg, anorexia, nausea, diarrhea) over the last six months. You consider whether to order a serum prealbumin test to assess for possible malnutrition.

The presence of malnutrition in hospitalized patients is widely recognized as an independent predictor of hospital mortality.¹ According to the American Society for Parenteral and Enteral Nutrition (ASPEN), malnutrition is defined as “an acute, subacute or chronic state of nutrition, in which varying degrees of overnutrition or undernutrition with or without inflammatory activity have led to a change in body composition and diminished function.”² In one large European study, patients screening positive for being at risk of malnutrition had a 12-fold increase in hospital mortality.¹

Inpatient malnutrition is remarkably underdocumented. Studies using chart reviews have found a prevalence of malnutrition in hospitalized patients of between 20% and 50%, and only 3% of hospital discharges are associated with a diagnostic code for malnutrition.^3–5 Appropriate diagnosis and documentation of malnutrition is important given the profound prognostic and management implications of a malnutrition diagnosis. Appropriate documentation benefits health systems as malnutrition documentation increases expected mortality, thereby improving the observed-to-expected mortality ratio.

Serum prealbumin testing is widely available and frequently ordered in the inpatient setting. In a query we performed of the large aggregate Cerner Electronic Health Record database, HealthFacts, which includes data from inpatient encounters for approximately 700 United States hospitals, prealbumin tests were ordered 129,152 times in 2015. This activity corresponds to estimated total charges of $2,562,375 based on the 2015 clinical laboratory fee schedule.⁶

Prealbumin is synthesized in the liver and released into circulation prior to excretion by the kidneys and gastrointestinal tract. Prealbumin transports thyroxine, triiodothyronine, and holo-retinol binding protein and, as a result, is also known as transthyretin.⁷ It was first proposed as a nutritional marker in 1972 with the publication of a study that showed low levels of prealbumin in 40 children with kwashiorkor that improved with intensive dietary supplementation.⁸ The shorter half-life of prealbumin (2.5 days) as compared with other identified nutritional markers, such as albumin, indicate that it would be suitable for detecting rapid changes in nutritional status.

Prealbumin Is Not Specific

An ideal nutritional marker should be specific enough that changes in this marker reflect changes in nutritional status.⁹ While there are many systemic factors that affect nutritional markers, such as prealbumin (Table 1), the acute phase response triggered by inflammation is the most significant confounder in the acutely ill hospitalized patient.⁹ This response to infection, stress, and malignancy leads to an increase in proinflammatory cytokines, increased liver synthesis of inflammatory proteins, such as C-reactive protein (CRP), and increased vascular permeability. Prealbumin is a negative acute phase reactant that decreases in concentration during the stress response due to slowed synthesis and extravasation.⁹ In a study of 24 patients with severe sepsis and trauma, levels of prealbumin inversely correlated with CRP, a reflection of the stress response, and returned to normal when CRP levels normalized. Neither prealbumin nor CRP, however, correlated with total body protein changes.¹⁰ Unfortunately, many studies supporting the use of prealbumin as a nutritional marker do not address the role of the acute phase response in their results. These studies include the original report on prealbumin in kwashiorkor, a condition known to be associated with a high rate of infectious diseases that can trigger the acute phase response.⁹ A consensus statement from the Academy of Nutrition and Dietetics (AND) and ASPEN noted that prealbumin is an indicator of inflammation and lacks the specificity to diagnose malnutrition.¹¹

Prealbumin Is Not Sensitive

A sensitive laboratory test for malnutrition should allow for detection of malnutrition at an early stage.⁹ However, patients who demonstrate severe malnutrition without a coexisting inflammatory state do not consistently show low levels of prealbumin. In a systematic review of 20 studies in nondiseased malnourished patients, only two studies, both of which assessed patients with anorexia nervosa, had a mean prealbumin below normal (<20 mg/dL), and this finding corresponded to patient populations with mean BMIs less than 12 kg/m². More importantly, normal prealbumin levels were seen in groups of patients with a mean BMI as low as 12.9 kg/m².¹² Analysis by AND found insufficient evidence to support a correlation between prealbumin and weight loss in anorexia nervosa, calorie restricted diets, or starvation.¹³ The data suggest that prealbumin lacks sufficient sensitivity to consistently detect cases of malnutrition easily diagnosed by history and/or physical exam.

Prealbumin Is Not Consistently Responsive to Nutritional Interventions

An accurate marker for malnutrition should improve when nutritional intervention results in adequate nutritional intake.⁹ While some studies have shown improvements in prealbumin in the setting of a nutritional intervention, many of these works are subject to the same limitations related to specificity and lack of control for concurrent inflammatory processes. In a retrospective study, prealbumin increased significantly in 102 patients receiving TPN for one week. Unfortunately, patients with renal or hepatic disease were excluded, and the role of inflammation was not assessed.¹⁴ Institutionalized patients with Alzheimer’s disease and normal CRP levels showed a statistically significant increase in weight gain, arm muscle circumference, and triceps skin-fold thickness following a nutritional program without a notable change in prealbumin.¹⁵ In a study assessing the relationship of prealbumin, CRP, and nutritional intake, critically ill populations receiving less than or greater than 60% of their estimated caloric needs showed no significant difference in prealbumin. In fact, prealbumin levels were only correlated with CRP levels.¹⁶ This finding argues against the routine use of prealbumin for nutrition monitoring in the acutely ill hospitalized patient.

Prealbumin Is Not Consistently Correlated with Health Outcomes

Even if prealbumin increased consistently in response to nutritional intervention, whether this change corresponds to an improvement in clinical outcomes has yet to be demonstrated.⁹ In 2005, Koretz reviewed 99 clinical trials and concluded that even when changes in nutritional markers are seen with nutritional support, the “changes in nutritional markers do not predict clinical outcomes.”¹⁷

Given the lack of a suitable biologic assay to identify malnutrition, dieticians and clinicians must rely on other means to assess malnutrition. Professional societies, including ASPEN and the European Society for Clinical Nutrition and Metabolism, have proposed different guidelines for the screening and assessment of malnutrition (Table 2).^11,18 In 2016, these organizations, along with the Latin American Federation of Nutritional Therapy, Clinical Nutrition, and Metabolism and the Parenteral and Enteral Nutrition Society of Asia, formed The Global Leadership Initiative on Malnutrition (GLIM). In 2017, the GLIM taskforce agreed on clinically relevant diagnostic variables for the screening and assessment of malnutrition, including reduced food intake (anorexia), nonvolitional weight loss, (reduced) lean mass, status of disease burden and inflammation, and low body mass index or underweight status.¹⁹

• Do not use prealbumin to screen for or diagnose malnutrition.
• Consult with local dietitians to ensure that your institutional approach is in agreement with consensus recommendations.

In revisiting the case above, the patient does not have clear evidence of malnutrition based on his history (stable weight and good reported nutritional intake), although he does have a low BMI of 18.5 kg/m². Rather than prealbumin testing, which would likely be low secondary to the acute phase response, he would better benefit from a nutrition-focused history and physical exam.

The uncertainties faced by clinicians in diagnosing malnutrition cannot readily be resolved by relying on a solitary laboratory marker (eg, prealbumin) or a stand-alone assessment protocol. The data obtained reflect the need for multidisciplinary teams of dieticians and clinicians to contextualize each patient’s medical history and ensure that the selected metrics are used appropriately to aid in diagnosis and documentation. We advocate that clinicians not routinely use prealbumin to screen for, confirm the diagnosis of, or assess the severity of malnutrition in the hospitalized patient.

Do you think this is a low-value practice? Is this truly a “Thing We Do for No Reason?” Share what you do in your practice and join in the conversation online by retweeting it on Twitter (#TWDFNR) and liking it on Facebook. We invite you to propose ideas for other “Things We Do for No Reason” topics by emailing [email protected].

Disclosures

The authors have nothing to disclose.

The “Things We Do for No Reason” series reviews practices which have become common parts of hospital care but which may provide little value to our patients. Practices reviewed in the TWDFNR series do not represent “black and white” conclusions or clinical practice standards, but are meant as a starting place for research and active discussions among hospitalists and patients. We invite you to be part of that discussion. https://www.choosingwisely.org/

A 34-year-old man is admitted for a complicated urinary tract infection related to a chronic in-dwelling Foley catheter. The patient suffered a spinal cord injury at the C4/C5 level as a result of a motor vehicle accident 10 years ago and is confined to a motorized wheelchair. He is an engineer and lives independently but has caregivers. His body mass index (BMI) is 18.5 kg/m², and he reports his weight has been stable. He has slight muscle atrophy of the biceps, triceps, interosseous muscles, and quadriceps. The patient reports that he eats well, has no chronic conditions, and has not had any gastrointestinal symptoms (eg, anorexia, nausea, diarrhea) over the last six months. You consider whether to order a serum prealbumin test to assess for possible malnutrition.

The presence of malnutrition in hospitalized patients is widely recognized as an independent predictor of hospital mortality.¹ According to the American Society for Parenteral and Enteral Nutrition (ASPEN), malnutrition is defined as “an acute, subacute or chronic state of nutrition, in which varying degrees of overnutrition or undernutrition with or without inflammatory activity have led to a change in body composition and diminished function.”² In one large European study, patients screening positive for being at risk of malnutrition had a 12-fold increase in hospital mortality.¹

Inpatient malnutrition is remarkably underdocumented. Studies using chart reviews have found a prevalence of malnutrition in hospitalized patients of between 20% and 50%, and only 3% of hospital discharges are associated with a diagnostic code for malnutrition.^3–5 Appropriate diagnosis and documentation of malnutrition is important given the profound prognostic and management implications of a malnutrition diagnosis. Appropriate documentation benefits health systems as malnutrition documentation increases expected mortality, thereby improving the observed-to-expected mortality ratio.

Serum prealbumin testing is widely available and frequently ordered in the inpatient setting. In a query we performed of the large aggregate Cerner Electronic Health Record database, HealthFacts, which includes data from inpatient encounters for approximately 700 United States hospitals, prealbumin tests were ordered 129,152 times in 2015. This activity corresponds to estimated total charges of $2,562,375 based on the 2015 clinical laboratory fee schedule.⁶

Prealbumin is synthesized in the liver and released into circulation prior to excretion by the kidneys and gastrointestinal tract. Prealbumin transports thyroxine, triiodothyronine, and holo-retinol binding protein and, as a result, is also known as transthyretin.⁷ It was first proposed as a nutritional marker in 1972 with the publication of a study that showed low levels of prealbumin in 40 children with kwashiorkor that improved with intensive dietary supplementation.⁸ The shorter half-life of prealbumin (2.5 days) as compared with other identified nutritional markers, such as albumin, indicate that it would be suitable for detecting rapid changes in nutritional status.

Prealbumin Is Not Specific

An ideal nutritional marker should be specific enough that changes in this marker reflect changes in nutritional status.⁹ While there are many systemic factors that affect nutritional markers, such as prealbumin (Table 1), the acute phase response triggered by inflammation is the most significant confounder in the acutely ill hospitalized patient.⁹ This response to infection, stress, and malignancy leads to an increase in proinflammatory cytokines, increased liver synthesis of inflammatory proteins, such as C-reactive protein (CRP), and increased vascular permeability. Prealbumin is a negative acute phase reactant that decreases in concentration during the stress response due to slowed synthesis and extravasation.⁹ In a study of 24 patients with severe sepsis and trauma, levels of prealbumin inversely correlated with CRP, a reflection of the stress response, and returned to normal when CRP levels normalized. Neither prealbumin nor CRP, however, correlated with total body protein changes.¹⁰ Unfortunately, many studies supporting the use of prealbumin as a nutritional marker do not address the role of the acute phase response in their results. These studies include the original report on prealbumin in kwashiorkor, a condition known to be associated with a high rate of infectious diseases that can trigger the acute phase response.⁹ A consensus statement from the Academy of Nutrition and Dietetics (AND) and ASPEN noted that prealbumin is an indicator of inflammation and lacks the specificity to diagnose malnutrition.¹¹

Prealbumin Is Not Sensitive

A sensitive laboratory test for malnutrition should allow for detection of malnutrition at an early stage.⁹ However, patients who demonstrate severe malnutrition without a coexisting inflammatory state do not consistently show low levels of prealbumin. In a systematic review of 20 studies in nondiseased malnourished patients, only two studies, both of which assessed patients with anorexia nervosa, had a mean prealbumin below normal (<20 mg/dL), and this finding corresponded to patient populations with mean BMIs less than 12 kg/m². More importantly, normal prealbumin levels were seen in groups of patients with a mean BMI as low as 12.9 kg/m².¹² Analysis by AND found insufficient evidence to support a correlation between prealbumin and weight loss in anorexia nervosa, calorie restricted diets, or starvation.¹³ The data suggest that prealbumin lacks sufficient sensitivity to consistently detect cases of malnutrition easily diagnosed by history and/or physical exam.

Prealbumin Is Not Consistently Responsive to Nutritional Interventions

An accurate marker for malnutrition should improve when nutritional intervention results in adequate nutritional intake.⁹ While some studies have shown improvements in prealbumin in the setting of a nutritional intervention, many of these works are subject to the same limitations related to specificity and lack of control for concurrent inflammatory processes. In a retrospective study, prealbumin increased significantly in 102 patients receiving TPN for one week. Unfortunately, patients with renal or hepatic disease were excluded, and the role of inflammation was not assessed.¹⁴ Institutionalized patients with Alzheimer’s disease and normal CRP levels showed a statistically significant increase in weight gain, arm muscle circumference, and triceps skin-fold thickness following a nutritional program without a notable change in prealbumin.¹⁵ In a study assessing the relationship of prealbumin, CRP, and nutritional intake, critically ill populations receiving less than or greater than 60% of their estimated caloric needs showed no significant difference in prealbumin. In fact, prealbumin levels were only correlated with CRP levels.¹⁶ This finding argues against the routine use of prealbumin for nutrition monitoring in the acutely ill hospitalized patient.

Prealbumin Is Not Consistently Correlated with Health Outcomes

Even if prealbumin increased consistently in response to nutritional intervention, whether this change corresponds to an improvement in clinical outcomes has yet to be demonstrated.⁹ In 2005, Koretz reviewed 99 clinical trials and concluded that even when changes in nutritional markers are seen with nutritional support, the “changes in nutritional markers do not predict clinical outcomes.”¹⁷

Given the lack of a suitable biologic assay to identify malnutrition, dieticians and clinicians must rely on other means to assess malnutrition. Professional societies, including ASPEN and the European Society for Clinical Nutrition and Metabolism, have proposed different guidelines for the screening and assessment of malnutrition (Table 2).^11,18 In 2016, these organizations, along with the Latin American Federation of Nutritional Therapy, Clinical Nutrition, and Metabolism and the Parenteral and Enteral Nutrition Society of Asia, formed The Global Leadership Initiative on Malnutrition (GLIM). In 2017, the GLIM taskforce agreed on clinically relevant diagnostic variables for the screening and assessment of malnutrition, including reduced food intake (anorexia), nonvolitional weight loss, (reduced) lean mass, status of disease burden and inflammation, and low body mass index or underweight status.¹⁹

• Do not use prealbumin to screen for or diagnose malnutrition.
• Consult with local dietitians to ensure that your institutional approach is in agreement with consensus recommendations.

In revisiting the case above, the patient does not have clear evidence of malnutrition based on his history (stable weight and good reported nutritional intake), although he does have a low BMI of 18.5 kg/m². Rather than prealbumin testing, which would likely be low secondary to the acute phase response, he would better benefit from a nutrition-focused history and physical exam.

The uncertainties faced by clinicians in diagnosing malnutrition cannot readily be resolved by relying on a solitary laboratory marker (eg, prealbumin) or a stand-alone assessment protocol. The data obtained reflect the need for multidisciplinary teams of dieticians and clinicians to contextualize each patient’s medical history and ensure that the selected metrics are used appropriately to aid in diagnosis and documentation. We advocate that clinicians not routinely use prealbumin to screen for, confirm the diagnosis of, or assess the severity of malnutrition in the hospitalized patient.

Do you think this is a low-value practice? Is this truly a “Thing We Do for No Reason?” Share what you do in your practice and join in the conversation online by retweeting it on Twitter (#TWDFNR) and liking it on Facebook. We invite you to propose ideas for other “Things We Do for No Reason” topics by emailing [email protected].

Disclosures

The authors have nothing to disclose.

Lacy and colleagues identify an important “Thing We Do For No Reason”—prealbumin testing to diagnose malnutrition in hospitalized patients.¹ They highlight the frequency and costs of ordering prealbumin tests although prealbumin is neither specific nor sensitive as a “marker of nutritional status,” shows no response to nutritional interventions, and has not been shown to correlate with clinical outcomes. We strongly support their analysis. A core problem in the process of nutrition assessment underlies this meaningless and costly practice. The term “malnutrition” is perfectly ambiguous. In one common usage, the term means that “markers of nutritional status” are abnormal. This usage allows a circular reasoning process where prealbumin is defined as a marker of nutritional status, and people with low prealbumin are then diagnosed as malnourished.

The term is also used to mean a condition where evidence shows better patient outcomes when improved nutrition is provided. Distinguishing between these two meanings is essential, as numerous patients with inflammatory illness will present abnormal “markers” when good evidence shows that they cannot benefit from nutritional support.

For example, a patient with advanced untreated human immunodeficiency virus (HIV) is likely to be considered malnourished because all of her “markers of nutritional status” are abnormal. She barely eats, has lost weight, and has low anthropometric, immunologic, and serologic measures, poor functional status, extreme vulnerability, and very poor prognosis. In this way she resembles a person in a famine situation. However, the patient is not malnourished in the sense that improved nutrient intake will lead to better patient outcomes. A Cochrane review of “nutritional interventions for reducing morbidity and mortality in people with HIV” found “no evidence that such supplementation translates into reductions in disease progression or HIV‐related complications, such as opportunistic infections or death.”² The patient is dying of an inflammatory, cachectic illness. The same is true in managing patients with advanced cancer or several other serious illnesses.

Low prealbumin measures are associated with poor outcomes, which are then attributed to “malnutrition.” However, as Lacy and colleagues argue, prealbumin is a negative acute phase reactant and is thus a marker of the inflammatory effects of sickness/injury; it also responds variably to nutritional support. Citing Koretz, they note that “even when changes in nutritional markers are seen with nutritional support, the ‘changes in nutritional markers do not predict clinical outcomes.’”^1,3 We know of no evidence from randomized controlled trials that prealbumin measurements help identify patients who can benefit from nutrition support.

By contrast, we and our colleagues have shown that in people who barely eat but show no inflammatory disease, eg, prison hunger-strikers and patients with anorexia nervosa, prealbumin level remains normal down to a body mass index below 13. The same is generally true for albumin.⁴ These measures fail to identify “malnutrition” in people who are starving.

Despite the complete lack of clinical trial evidence of benefit, prealbumin is widely used as an indicator of malnutrition. The National Institutes of Health’s Medline Plus website for the general public lists low prealbumin levels as a possible sign of malnutrition, for example, and advises that the prealbumin test may be used to “find out if you are getting enough nutrients, especially protein, in your diet” and to “check to see if you are getting enough nutrition if you are in the hospital.”⁵ Unjustified assertions such as these contribute to the dramatic overuse of nutritional interventions.

However, as a rule, things do occur for a reason. Using the term “prealbumin” conjures a certain relationship, perhaps as a precursor, to albumin, a venerable (but valueless) “marker of nutrition status.” In fact, the term refers only to a difference in electrophoretic mobility (prealbumin migrates faster). If prealbumin were called it by its proper name, transthyretin, it would probably have languished in obscurity among serum proteins until, in recent years, drug suppression of transthyretin synthesis has been shown to benefit patients with hereditary transthyretin amyloidosis.⁶ Using a name that references albumin, this protein has found the limelight as a marker of nutritional status.

The close similarity in appearance between starvation and wasting illness enables the strong, largely evidence-free⁷ emphasis on nutrition support. Many families and individuals suffer when a loved one loses weight. As a prominent reminder of serious illness, this wasted appearance can be painful to bear. Several caregivers may fear that they will be judged as neglectful by outside observers. Other individuals also wish to maintain their body weight for social reasons (as weight loss may be interpreted as a sign of illness, especially HIV). Nutrition maintains a special status in various contexts during the care of sick patients, and the drive to provide food to individuals who appear undernourished seems fundamental in humans.

A third reason for the frivolous, widespread overdiagnosis of “malnutrition” is that it leads directly to favorable consequences for the multibillion-dollar nutritional support industry. A consistent rational approach to the use of nutritional support products for sick people would lead to multibillion-dollar harm for that industry. For now, however, no self-respecting clinician could fail to provide nutritional support to a patient diagnosed as “malnourished” regardless of evidence.

The consistent rational approach in caring for patients is to search for good evidence of benefit before initiating a treatment course. Although sending blood tests for “nutritional markers” to diagnose nutritional needs may be easier and more popular, we caution against such over-simplification. Using prealbumin as a marker for malnutrition could lead to overlooking potentially treatable inflammatory or infectious illness. On the other hand, the use of prealbumin could also lead to unnecessary and potentially dangerous treatments, such as feeding tube placement and/or total parental nutrition. Thus, with one small amendment, we fully support Lacy and colleagues’ conclusion that prealbumin testing to identify malnutrition in hospitalized patients is a “Thing We Do For No (good) Reason.”

Disclosures

Drs. Lee and Finucane declare no financial conflicts of interest. Dr. Finucane discloses that he serves the pharmacy committee of an insurance company.

Lacy and colleagues identify an important “Thing We Do For No Reason”—prealbumin testing to diagnose malnutrition in hospitalized patients.¹ They highlight the frequency and costs of ordering prealbumin tests although prealbumin is neither specific nor sensitive as a “marker of nutritional status,” shows no response to nutritional interventions, and has not been shown to correlate with clinical outcomes. We strongly support their analysis. A core problem in the process of nutrition assessment underlies this meaningless and costly practice. The term “malnutrition” is perfectly ambiguous. In one common usage, the term means that “markers of nutritional status” are abnormal. This usage allows a circular reasoning process where prealbumin is defined as a marker of nutritional status, and people with low prealbumin are then diagnosed as malnourished.

The term is also used to mean a condition where evidence shows better patient outcomes when improved nutrition is provided. Distinguishing between these two meanings is essential, as numerous patients with inflammatory illness will present abnormal “markers” when good evidence shows that they cannot benefit from nutritional support.

For example, a patient with advanced untreated human immunodeficiency virus (HIV) is likely to be considered malnourished because all of her “markers of nutritional status” are abnormal. She barely eats, has lost weight, and has low anthropometric, immunologic, and serologic measures, poor functional status, extreme vulnerability, and very poor prognosis. In this way she resembles a person in a famine situation. However, the patient is not malnourished in the sense that improved nutrient intake will lead to better patient outcomes. A Cochrane review of “nutritional interventions for reducing morbidity and mortality in people with HIV” found “no evidence that such supplementation translates into reductions in disease progression or HIV‐related complications, such as opportunistic infections or death.”² The patient is dying of an inflammatory, cachectic illness. The same is true in managing patients with advanced cancer or several other serious illnesses.

Low prealbumin measures are associated with poor outcomes, which are then attributed to “malnutrition.” However, as Lacy and colleagues argue, prealbumin is a negative acute phase reactant and is thus a marker of the inflammatory effects of sickness/injury; it also responds variably to nutritional support. Citing Koretz, they note that “even when changes in nutritional markers are seen with nutritional support, the ‘changes in nutritional markers do not predict clinical outcomes.’”^1,3 We know of no evidence from randomized controlled trials that prealbumin measurements help identify patients who can benefit from nutrition support.

By contrast, we and our colleagues have shown that in people who barely eat but show no inflammatory disease, eg, prison hunger-strikers and patients with anorexia nervosa, prealbumin level remains normal down to a body mass index below 13. The same is generally true for albumin.⁴ These measures fail to identify “malnutrition” in people who are starving.

Despite the complete lack of clinical trial evidence of benefit, prealbumin is widely used as an indicator of malnutrition. The National Institutes of Health’s Medline Plus website for the general public lists low prealbumin levels as a possible sign of malnutrition, for example, and advises that the prealbumin test may be used to “find out if you are getting enough nutrients, especially protein, in your diet” and to “check to see if you are getting enough nutrition if you are in the hospital.”⁵ Unjustified assertions such as these contribute to the dramatic overuse of nutritional interventions.

However, as a rule, things do occur for a reason. Using the term “prealbumin” conjures a certain relationship, perhaps as a precursor, to albumin, a venerable (but valueless) “marker of nutrition status.” In fact, the term refers only to a difference in electrophoretic mobility (prealbumin migrates faster). If prealbumin were called it by its proper name, transthyretin, it would probably have languished in obscurity among serum proteins until, in recent years, drug suppression of transthyretin synthesis has been shown to benefit patients with hereditary transthyretin amyloidosis.⁶ Using a name that references albumin, this protein has found the limelight as a marker of nutritional status.

The close similarity in appearance between starvation and wasting illness enables the strong, largely evidence-free⁷ emphasis on nutrition support. Many families and individuals suffer when a loved one loses weight. As a prominent reminder of serious illness, this wasted appearance can be painful to bear. Several caregivers may fear that they will be judged as neglectful by outside observers. Other individuals also wish to maintain their body weight for social reasons (as weight loss may be interpreted as a sign of illness, especially HIV). Nutrition maintains a special status in various contexts during the care of sick patients, and the drive to provide food to individuals who appear undernourished seems fundamental in humans.

A third reason for the frivolous, widespread overdiagnosis of “malnutrition” is that it leads directly to favorable consequences for the multibillion-dollar nutritional support industry. A consistent rational approach to the use of nutritional support products for sick people would lead to multibillion-dollar harm for that industry. For now, however, no self-respecting clinician could fail to provide nutritional support to a patient diagnosed as “malnourished” regardless of evidence.

The consistent rational approach in caring for patients is to search for good evidence of benefit before initiating a treatment course. Although sending blood tests for “nutritional markers” to diagnose nutritional needs may be easier and more popular, we caution against such over-simplification. Using prealbumin as a marker for malnutrition could lead to overlooking potentially treatable inflammatory or infectious illness. On the other hand, the use of prealbumin could also lead to unnecessary and potentially dangerous treatments, such as feeding tube placement and/or total parental nutrition. Thus, with one small amendment, we fully support Lacy and colleagues’ conclusion that prealbumin testing to identify malnutrition in hospitalized patients is a “Thing We Do For No (good) Reason.”

Disclosures

Drs. Lee and Finucane declare no financial conflicts of interest. Dr. Finucane discloses that he serves the pharmacy committee of an insurance company.

Lacy and colleagues identify an important “Thing We Do For No Reason”—prealbumin testing to diagnose malnutrition in hospitalized patients.¹ They highlight the frequency and costs of ordering prealbumin tests although prealbumin is neither specific nor sensitive as a “marker of nutritional status,” shows no response to nutritional interventions, and has not been shown to correlate with clinical outcomes. We strongly support their analysis. A core problem in the process of nutrition assessment underlies this meaningless and costly practice. The term “malnutrition” is perfectly ambiguous. In one common usage, the term means that “markers of nutritional status” are abnormal. This usage allows a circular reasoning process where prealbumin is defined as a marker of nutritional status, and people with low prealbumin are then diagnosed as malnourished.

The term is also used to mean a condition where evidence shows better patient outcomes when improved nutrition is provided. Distinguishing between these two meanings is essential, as numerous patients with inflammatory illness will present abnormal “markers” when good evidence shows that they cannot benefit from nutritional support.

For example, a patient with advanced untreated human immunodeficiency virus (HIV) is likely to be considered malnourished because all of her “markers of nutritional status” are abnormal. She barely eats, has lost weight, and has low anthropometric, immunologic, and serologic measures, poor functional status, extreme vulnerability, and very poor prognosis. In this way she resembles a person in a famine situation. However, the patient is not malnourished in the sense that improved nutrient intake will lead to better patient outcomes. A Cochrane review of “nutritional interventions for reducing morbidity and mortality in people with HIV” found “no evidence that such supplementation translates into reductions in disease progression or HIV‐related complications, such as opportunistic infections or death.”² The patient is dying of an inflammatory, cachectic illness. The same is true in managing patients with advanced cancer or several other serious illnesses.

Low prealbumin measures are associated with poor outcomes, which are then attributed to “malnutrition.” However, as Lacy and colleagues argue, prealbumin is a negative acute phase reactant and is thus a marker of the inflammatory effects of sickness/injury; it also responds variably to nutritional support. Citing Koretz, they note that “even when changes in nutritional markers are seen with nutritional support, the ‘changes in nutritional markers do not predict clinical outcomes.’”^1,3 We know of no evidence from randomized controlled trials that prealbumin measurements help identify patients who can benefit from nutrition support.

By contrast, we and our colleagues have shown that in people who barely eat but show no inflammatory disease, eg, prison hunger-strikers and patients with anorexia nervosa, prealbumin level remains normal down to a body mass index below 13. The same is generally true for albumin.⁴ These measures fail to identify “malnutrition” in people who are starving.

Despite the complete lack of clinical trial evidence of benefit, prealbumin is widely used as an indicator of malnutrition. The National Institutes of Health’s Medline Plus website for the general public lists low prealbumin levels as a possible sign of malnutrition, for example, and advises that the prealbumin test may be used to “find out if you are getting enough nutrients, especially protein, in your diet” and to “check to see if you are getting enough nutrition if you are in the hospital.”⁵ Unjustified assertions such as these contribute to the dramatic overuse of nutritional interventions.

However, as a rule, things do occur for a reason. Using the term “prealbumin” conjures a certain relationship, perhaps as a precursor, to albumin, a venerable (but valueless) “marker of nutrition status.” In fact, the term refers only to a difference in electrophoretic mobility (prealbumin migrates faster). If prealbumin were called it by its proper name, transthyretin, it would probably have languished in obscurity among serum proteins until, in recent years, drug suppression of transthyretin synthesis has been shown to benefit patients with hereditary transthyretin amyloidosis.⁶ Using a name that references albumin, this protein has found the limelight as a marker of nutritional status.

The close similarity in appearance between starvation and wasting illness enables the strong, largely evidence-free⁷ emphasis on nutrition support. Many families and individuals suffer when a loved one loses weight. As a prominent reminder of serious illness, this wasted appearance can be painful to bear. Several caregivers may fear that they will be judged as neglectful by outside observers. Other individuals also wish to maintain their body weight for social reasons (as weight loss may be interpreted as a sign of illness, especially HIV). Nutrition maintains a special status in various contexts during the care of sick patients, and the drive to provide food to individuals who appear undernourished seems fundamental in humans.

A third reason for the frivolous, widespread overdiagnosis of “malnutrition” is that it leads directly to favorable consequences for the multibillion-dollar nutritional support industry. A consistent rational approach to the use of nutritional support products for sick people would lead to multibillion-dollar harm for that industry. For now, however, no self-respecting clinician could fail to provide nutritional support to a patient diagnosed as “malnourished” regardless of evidence.

The consistent rational approach in caring for patients is to search for good evidence of benefit before initiating a treatment course. Although sending blood tests for “nutritional markers” to diagnose nutritional needs may be easier and more popular, we caution against such over-simplification. Using prealbumin as a marker for malnutrition could lead to overlooking potentially treatable inflammatory or infectious illness. On the other hand, the use of prealbumin could also lead to unnecessary and potentially dangerous treatments, such as feeding tube placement and/or total parental nutrition. Thus, with one small amendment, we fully support Lacy and colleagues’ conclusion that prealbumin testing to identify malnutrition in hospitalized patients is a “Thing We Do For No (good) Reason.”

Disclosures

Drs. Lee and Finucane declare no financial conflicts of interest. Dr. Finucane discloses that he serves the pharmacy committee of an insurance company.

Inspired by the ABIM Foundation’s Choosing Wisely^® campaign, the “Things We Do for No Reason” (TWDFNR) series reviews practices that have become common parts of hospital care but may provide little value to our patients. Practices reviewed in the TWDFNR series do not represent “black and white” conclusions or clinical practice standards but are meant as a starting place for research and active discussions among hospitalists and patients. We invite you to be part of that discussion.

A 28 year-old woman presents to the emergency department with acute onset bilateral chest pain and dyspnea. She has a respiratory rate of 28, a heart rate of 106, blood pressure of 110/65 mm Hg, and pulse oximetry of 92% saturation on room air. She has no history of cardiac or pulmonary disease and no personal history of venous thromboembolism. She takes an estrogen-containing oral contraceptive. On examination, she has no jugular venous distention, normal cardiac tones without murmur, and no lower extremity swelling. D-dimer is elevated at 3.4 mg/L (normal < 0.5 mg/L), and she undergoes computed tomography (CT) of the chest, which demonstrates acute segmental pulmonary emboli (PE) in the right upper and middle lobes as well as multiple bilateral subsegmental PEs. The CT suggests right ventricular dysfunction (RVD), and her troponin T is 0.06 ng/mL (normal < 0.01 ng/mL). Bilateral lower extremity venous Doppler ultrasonography demonstrates no acute thrombus.

Acute pulmonary embolism (PE) accounts for more than 300,000 inpatient admissions annually in the United States.¹ The vast majority of patients with acute PE who receive adequate anticoagulation will have favorable outcomes.^2,3 In the past two decades, for example, mortality has decreased significantly among patients admitted with acute PE,² with 30-day all-cause mortality falling to approximately 5%.³ The risk-adjusted rate of recurrent venous thromboembolism (VTE) within 30 days has concomitantly dropped below 1%.³

Acute PE severity was previously classified as massive or high risk, submassive or intermediate risk, and low risk.⁴ Massive PE was defined by RVD and persistent hypotension or shock requiring vasopressors. ⁴ Intermediate-risk or submassive PE typically referred to normotensive patients with RVD and/or myocardial necrosis (eg, elevated troponin).^4,5 Low-risk PEs had neither hemodynamic instability nor RVD. This classification scheme, however, has fallen out of favor as PE severity exists on a risk spectrum.⁶ Instead, recent guidelines from the European Society of Cardiology and the American College of Chest Physicians recommend first parsing PE severity by the presence or absence of hypotension (Figure 1).^6,7 Risk assessment can be subsequently enhanced by validated clinical risk prediction scores, imaging-based assessment of RVD, and cardiac biomarker testing.⁶

Echocardiography is a common method for evaluating RVD, and echocardiographic RVD confers an increased risk of adverse outcomes in PE.^10-12 In the earliest meta-analysis to evaluate this association, Sanchez et al. combined data from five studies that included 623 patients from emergency room and inpatient settings. They found that echocardiographic RVD conferred an unadjusted relative risk for short-term mortality of 2.53 (95%CI 1.17-5.50).¹² A subsequent meta-analysis by Cho et al. pooled data from both prospective and retrospective cohorts to examine short-term mortality in a total of 3,283 hemodynamically stable patients with PE, of whom 1,223 (37.3%) had RVD diagnosed by echocardiogram.¹⁰ In this population, RVD was associated with an odds ratio of 2.29 (95%CI 1.61-3.26) for short-term death. Thus, echocardiography could be viewed as a risk stratification tool, even in hemodynamically stable PE.

For most hemodynamically stable patients, echocardiographic findings will not enhance prognostication and/or have a therapeutic impact. The following four reasons explain why echocardiography adds little value to the care of these patients.

First, phenotypic expression of RVD varies from asymptomatic, despite abnormalities on diagnostic testing, to obstructive shock. Unfortunately, available prognostic models classify echocardiographic RVD in a binary fashion (present/absent)^4,7,10 whereas RVD exists on a continuum. Consequently, RVD is commonly found in acute PE^8,10,11 and has been identified in more than half of patients hospitalized with PE referred for echocardiography.⁸ Existing data do not allow clinicians to judge the clinical impact of the severity of echocardiographic RVD,⁸ and only the phenotypic expression of refractory hypotension has clear therapeutic implications.^6,7

Second, while echocardiographic RVD is associated with short-term mortality,^10-12 absolute rates of adverse outcomes are quite low when RVD is identified. For example, in a study merging multiple prospective cohorts, Becattini et al. demonstrated that RVD diagnosed by echocardiography or CT occurred in 41% of hospitalized patients stratified to low-risk PE by the simplified Pulmonary Embolism Severity Index (sPESI).⁸ For these patients, the 30-day mortality was 1.2%,⁸ which approximates the expected mortality from a low-risk sPESI score alone (1.1%).¹³ Even among intermediate-risk acute PE patients with RVD and/or elevated troponin enrolled in thrombolysis trials, the overall risk of death at 30 days was approximately 2%-3%, irrespective of the treatment arm.^5,14,15

Third, RVD identified by echocardiography does not inform or enhance prognostication as compared with cardiac biomarker testing. In a meta-analysis by Sanchez et al., echocardiographic RVD predicted death with a risk ratio of 2.53 (95% CI 1.17-5.50).¹² However, both elevated cardiac troponin and brain natriuretic peptide indicated a significantly worse outcome than imaging findings, with risk ratios of 8.3 (95% CI 3.6-19.3) and 9.5 (95% CI 3.2-28.6), respectively.¹³ More recently, Jiménez derived and validated a multivariable risk prediction model for stable PE.¹¹ In their data, echocardiographic RVD had an unadjusted odds ratio of 2.62 (95% CI 1.54-4.45) for predicting a 30-day complicated course. After multivariable adjustment that included sPESI scores, lower extremity ultrasound results, and cardiac biomarker testing, these odds became insignificant.¹¹ In other words, identifying echocardiographic RVD did not improve prognostication in hemodynamically stable PE patients when other commonly available variables were used.

Finally, in hemodynamically stable patients, echocardiographic RVD might create patient anxiety and cause harm. In a recent retrospective cohort study of 64,037 stable patients with PE, exposure to echocardiography was associated with a five-fold increase in likelihood of having received thrombolysis without any significant differences in risk-adjusted mortality.¹⁶ These data suggest that when faced with an abnormal echocardiogram, clinicians and patients may opt for more aggressive, time-sensitive therapies. Basing thrombolysis decisions on echocardiographic RVD potentially subjects patients to harm without decreasing mortality.^5,14,15 For example, the PEITHO study, which was the largest randomized trial evaluating thrombolysis in intermediate-risk acute PE, enrolled 1,006 patients and demonstrated that treating 29 intermediate-risk patients with thrombolysis prevented one case of hemodynamic decompensation.⁵ These benefits were counterbalanced by a number needed to harm of 14 to cause stroke or major bleeding. Ominous echocardiographic findings may also bias clinicians toward more intensive monitoring. Rates of echocardiogram utilization in hemodynamically stable PE are linked to higher rates of ICU admission and longer hospital stays without significant impact on patient outcomes.¹⁶

Echocardiography should be used to exclude other causes of hypotension in patients with presumed PE-related shock^7,9 and to improve clinicians’ confidence prescribing systemic thrombolytics in the face of hemodynamic instability.^6,7 Otherwise, echocardiography should be reserved for highly selected intermediate-risk patients with acute PE. Among patients with intermediate-risk PE, those most likely to decompensate or die typically satisfy all of the following conditions: (1) highest-risk PESI or sPESI scores, (2) elevated natriuretic peptides, (3) elevated troponin, and (4) proximal deep vein thrombosis (DVT) on lower extremity ultrasound.^11,13 In such patients, the echocardiogram may reveal a critical “tipping point,” such as a right atrial or ventricular thrombus-in-transit, that may warrant more intensive monitoring and multidisciplinary input into the most appropriate treatment plan.

Echocardiography could aid therapeutic decisions when the benefits from thrombolysis may outweigh the risks, such as for patients with minimal physiologic reserve and/or a low risk of major bleeding complications. Prognostic models like sPESI utilize binary variables, such as the presence/absence of chronic cardiopulmonary disease or oxygen saturation above/below 90%. Clearly, these variables exist on a spectrum; intuitively, patients with severe comorbidities and more alarming vital signs have a higher risk of death or decompensation than predicted by sPESI. Analogously, echocardiographic findings of RVD also encompass a spectrum. Because prognostic models and clinical trials cannot guide decisions for each individual patient, clinicians could justify using echocardiography to “fine tune” prognostication and to provide a personalized approach for carefully selected patients.

Clinicians should use a risk prediction model for all hemodynamically stable patients with confirmed PE.^6,7 Validated risk calculators include the sPESI,^6,7,14 which relies exclusively on the patient’s history and vital signs, and the eStiMaTe© tool (www.peprognosis.org), which enhances prognostication from sPESI by incorporating troponin, natriuretic peptide, and lower- extremity Doppler results. ¹¹ For patients with symptoms or physical signs of RVD, chest CT and cardiac biomarkers (ie, troponin and/or natriuretic peptides) are sufficient for prognostication.^11,14 In intermediate-risk patients with the highest risk for decompensation based on risk prediction scores, the echocardiogram should represent a part of a comprehensive clinical evaluation, not the sole criterion for intensive monitoring and aggressive treatment.

• Clinicians should use a validated tool, such as the sPESI, for initial risk stratification of hemodynamically stable patients with acute pulmonary embolism.
• Hemodynamically unstable patients with confirmed or suspected acute PE may benefit from early echocardiography to confirm RVD as the cause of shock.^6,7,9
• The majority of normotensive adults with acute PE should not undergo echocardiography. To identify the patients at the greatest risk for decompensation, clinicians may consider using the eStiMaTe© tool (www.peprognosis.org), which augments risk stratification afforded by sPESI.
• For hemodynamically stable patients with PE who have already undergone echocardiography, clinicians should avoid being biased by the finding of RVD, particularly if other prognostic markers are reassuring.

In evaluating the patient described earlier, echocardiography has no clear prognostic implications. Her admission sPESI score equals zero, predicting a 30-day mortality of 1.1%. Including her lower extremity ultrasound and troponin T results into the eStiMaTe© calculator (www.peprognosis.org) surprisingly predicts an even lower rate of 30-day mortality (0.4%) and low risk of a complicated course (2.4%). Assessing for RVD on echocardiography may increase her risk of unnecessary and potentially injurious interventions.

Do you think this is a low-value practice? Is this truly a “Thing We Do for No Reason?” Share what you do in your practice and join in the conversation online by retweeting it on Twitter (#TWDFNR) and liking it on Facebook. We invite you to propose ideas for other “Things We Do for No Reason” topics by emailing[email protected].

Disclosures

The authors have no conflicts of interest relevant to this article.

Inspired by the ABIM Foundation’s Choosing Wisely^® campaign, the “Things We Do for No Reason” (TWDFNR) series reviews practices that have become common parts of hospital care but may provide little value to our patients. Practices reviewed in the TWDFNR series do not represent “black and white” conclusions or clinical practice standards but are meant as a starting place for research and active discussions among hospitalists and patients. We invite you to be part of that discussion.

A 28 year-old woman presents to the emergency department with acute onset bilateral chest pain and dyspnea. She has a respiratory rate of 28, a heart rate of 106, blood pressure of 110/65 mm Hg, and pulse oximetry of 92% saturation on room air. She has no history of cardiac or pulmonary disease and no personal history of venous thromboembolism. She takes an estrogen-containing oral contraceptive. On examination, she has no jugular venous distention, normal cardiac tones without murmur, and no lower extremity swelling. D-dimer is elevated at 3.4 mg/L (normal < 0.5 mg/L), and she undergoes computed tomography (CT) of the chest, which demonstrates acute segmental pulmonary emboli (PE) in the right upper and middle lobes as well as multiple bilateral subsegmental PEs. The CT suggests right ventricular dysfunction (RVD), and her troponin T is 0.06 ng/mL (normal < 0.01 ng/mL). Bilateral lower extremity venous Doppler ultrasonography demonstrates no acute thrombus.

Acute pulmonary embolism (PE) accounts for more than 300,000 inpatient admissions annually in the United States.¹ The vast majority of patients with acute PE who receive adequate anticoagulation will have favorable outcomes.^2,3 In the past two decades, for example, mortality has decreased significantly among patients admitted with acute PE,² with 30-day all-cause mortality falling to approximately 5%.³ The risk-adjusted rate of recurrent venous thromboembolism (VTE) within 30 days has concomitantly dropped below 1%.³

Acute PE severity was previously classified as massive or high risk, submassive or intermediate risk, and low risk.⁴ Massive PE was defined by RVD and persistent hypotension or shock requiring vasopressors. ⁴ Intermediate-risk or submassive PE typically referred to normotensive patients with RVD and/or myocardial necrosis (eg, elevated troponin).^4,5 Low-risk PEs had neither hemodynamic instability nor RVD. This classification scheme, however, has fallen out of favor as PE severity exists on a risk spectrum.⁶ Instead, recent guidelines from the European Society of Cardiology and the American College of Chest Physicians recommend first parsing PE severity by the presence or absence of hypotension (Figure 1).^6,7 Risk assessment can be subsequently enhanced by validated clinical risk prediction scores, imaging-based assessment of RVD, and cardiac biomarker testing.⁶

Echocardiography is a common method for evaluating RVD, and echocardiographic RVD confers an increased risk of adverse outcomes in PE.^10-12 In the earliest meta-analysis to evaluate this association, Sanchez et al. combined data from five studies that included 623 patients from emergency room and inpatient settings. They found that echocardiographic RVD conferred an unadjusted relative risk for short-term mortality of 2.53 (95%CI 1.17-5.50).¹² A subsequent meta-analysis by Cho et al. pooled data from both prospective and retrospective cohorts to examine short-term mortality in a total of 3,283 hemodynamically stable patients with PE, of whom 1,223 (37.3%) had RVD diagnosed by echocardiogram.¹⁰ In this population, RVD was associated with an odds ratio of 2.29 (95%CI 1.61-3.26) for short-term death. Thus, echocardiography could be viewed as a risk stratification tool, even in hemodynamically stable PE.

For most hemodynamically stable patients, echocardiographic findings will not enhance prognostication and/or have a therapeutic impact. The following four reasons explain why echocardiography adds little value to the care of these patients.

First, phenotypic expression of RVD varies from asymptomatic, despite abnormalities on diagnostic testing, to obstructive shock. Unfortunately, available prognostic models classify echocardiographic RVD in a binary fashion (present/absent)^4,7,10 whereas RVD exists on a continuum. Consequently, RVD is commonly found in acute PE^8,10,11 and has been identified in more than half of patients hospitalized with PE referred for echocardiography.⁸ Existing data do not allow clinicians to judge the clinical impact of the severity of echocardiographic RVD,⁸ and only the phenotypic expression of refractory hypotension has clear therapeutic implications.^6,7

Second, while echocardiographic RVD is associated with short-term mortality,^10-12 absolute rates of adverse outcomes are quite low when RVD is identified. For example, in a study merging multiple prospective cohorts, Becattini et al. demonstrated that RVD diagnosed by echocardiography or CT occurred in 41% of hospitalized patients stratified to low-risk PE by the simplified Pulmonary Embolism Severity Index (sPESI).⁸ For these patients, the 30-day mortality was 1.2%,⁸ which approximates the expected mortality from a low-risk sPESI score alone (1.1%).¹³ Even among intermediate-risk acute PE patients with RVD and/or elevated troponin enrolled in thrombolysis trials, the overall risk of death at 30 days was approximately 2%-3%, irrespective of the treatment arm.^5,14,15

Third, RVD identified by echocardiography does not inform or enhance prognostication as compared with cardiac biomarker testing. In a meta-analysis by Sanchez et al., echocardiographic RVD predicted death with a risk ratio of 2.53 (95% CI 1.17-5.50).¹² However, both elevated cardiac troponin and brain natriuretic peptide indicated a significantly worse outcome than imaging findings, with risk ratios of 8.3 (95% CI 3.6-19.3) and 9.5 (95% CI 3.2-28.6), respectively.¹³ More recently, Jiménez derived and validated a multivariable risk prediction model for stable PE.¹¹ In their data, echocardiographic RVD had an unadjusted odds ratio of 2.62 (95% CI 1.54-4.45) for predicting a 30-day complicated course. After multivariable adjustment that included sPESI scores, lower extremity ultrasound results, and cardiac biomarker testing, these odds became insignificant.¹¹ In other words, identifying echocardiographic RVD did not improve prognostication in hemodynamically stable PE patients when other commonly available variables were used.

Finally, in hemodynamically stable patients, echocardiographic RVD might create patient anxiety and cause harm. In a recent retrospective cohort study of 64,037 stable patients with PE, exposure to echocardiography was associated with a five-fold increase in likelihood of having received thrombolysis without any significant differences in risk-adjusted mortality.¹⁶ These data suggest that when faced with an abnormal echocardiogram, clinicians and patients may opt for more aggressive, time-sensitive therapies. Basing thrombolysis decisions on echocardiographic RVD potentially subjects patients to harm without decreasing mortality.^5,14,15 For example, the PEITHO study, which was the largest randomized trial evaluating thrombolysis in intermediate-risk acute PE, enrolled 1,006 patients and demonstrated that treating 29 intermediate-risk patients with thrombolysis prevented one case of hemodynamic decompensation.⁵ These benefits were counterbalanced by a number needed to harm of 14 to cause stroke or major bleeding. Ominous echocardiographic findings may also bias clinicians toward more intensive monitoring. Rates of echocardiogram utilization in hemodynamically stable PE are linked to higher rates of ICU admission and longer hospital stays without significant impact on patient outcomes.¹⁶

Echocardiography should be used to exclude other causes of hypotension in patients with presumed PE-related shock^7,9 and to improve clinicians’ confidence prescribing systemic thrombolytics in the face of hemodynamic instability.^6,7 Otherwise, echocardiography should be reserved for highly selected intermediate-risk patients with acute PE. Among patients with intermediate-risk PE, those most likely to decompensate or die typically satisfy all of the following conditions: (1) highest-risk PESI or sPESI scores, (2) elevated natriuretic peptides, (3) elevated troponin, and (4) proximal deep vein thrombosis (DVT) on lower extremity ultrasound.^11,13 In such patients, the echocardiogram may reveal a critical “tipping point,” such as a right atrial or ventricular thrombus-in-transit, that may warrant more intensive monitoring and multidisciplinary input into the most appropriate treatment plan.

Echocardiography could aid therapeutic decisions when the benefits from thrombolysis may outweigh the risks, such as for patients with minimal physiologic reserve and/or a low risk of major bleeding complications. Prognostic models like sPESI utilize binary variables, such as the presence/absence of chronic cardiopulmonary disease or oxygen saturation above/below 90%. Clearly, these variables exist on a spectrum; intuitively, patients with severe comorbidities and more alarming vital signs have a higher risk of death or decompensation than predicted by sPESI. Analogously, echocardiographic findings of RVD also encompass a spectrum. Because prognostic models and clinical trials cannot guide decisions for each individual patient, clinicians could justify using echocardiography to “fine tune” prognostication and to provide a personalized approach for carefully selected patients.

Clinicians should use a risk prediction model for all hemodynamically stable patients with confirmed PE.^6,7 Validated risk calculators include the sPESI,^6,7,14 which relies exclusively on the patient’s history and vital signs, and the eStiMaTe© tool (www.peprognosis.org), which enhances prognostication from sPESI by incorporating troponin, natriuretic peptide, and lower- extremity Doppler results. ¹¹ For patients with symptoms or physical signs of RVD, chest CT and cardiac biomarkers (ie, troponin and/or natriuretic peptides) are sufficient for prognostication.^11,14 In intermediate-risk patients with the highest risk for decompensation based on risk prediction scores, the echocardiogram should represent a part of a comprehensive clinical evaluation, not the sole criterion for intensive monitoring and aggressive treatment.

• Clinicians should use a validated tool, such as the sPESI, for initial risk stratification of hemodynamically stable patients with acute pulmonary embolism.
• Hemodynamically unstable patients with confirmed or suspected acute PE may benefit from early echocardiography to confirm RVD as the cause of shock.^6,7,9
• The majority of normotensive adults with acute PE should not undergo echocardiography. To identify the patients at the greatest risk for decompensation, clinicians may consider using the eStiMaTe© tool (www.peprognosis.org), which augments risk stratification afforded by sPESI.
• For hemodynamically stable patients with PE who have already undergone echocardiography, clinicians should avoid being biased by the finding of RVD, particularly if other prognostic markers are reassuring.

In evaluating the patient described earlier, echocardiography has no clear prognostic implications. Her admission sPESI score equals zero, predicting a 30-day mortality of 1.1%. Including her lower extremity ultrasound and troponin T results into the eStiMaTe© calculator (www.peprognosis.org) surprisingly predicts an even lower rate of 30-day mortality (0.4%) and low risk of a complicated course (2.4%). Assessing for RVD on echocardiography may increase her risk of unnecessary and potentially injurious interventions.

Do you think this is a low-value practice? Is this truly a “Thing We Do for No Reason?” Share what you do in your practice and join in the conversation online by retweeting it on Twitter (#TWDFNR) and liking it on Facebook. We invite you to propose ideas for other “Things We Do for No Reason” topics by emailing[email protected].

Disclosures

The authors have no conflicts of interest relevant to this article.

Inspired by the ABIM Foundation’s Choosing Wisely^® campaign, the “Things We Do for No Reason” (TWDFNR) series reviews practices that have become common parts of hospital care but may provide little value to our patients. Practices reviewed in the TWDFNR series do not represent “black and white” conclusions or clinical practice standards but are meant as a starting place for research and active discussions among hospitalists and patients. We invite you to be part of that discussion.

A 28 year-old woman presents to the emergency department with acute onset bilateral chest pain and dyspnea. She has a respiratory rate of 28, a heart rate of 106, blood pressure of 110/65 mm Hg, and pulse oximetry of 92% saturation on room air. She has no history of cardiac or pulmonary disease and no personal history of venous thromboembolism. She takes an estrogen-containing oral contraceptive. On examination, she has no jugular venous distention, normal cardiac tones without murmur, and no lower extremity swelling. D-dimer is elevated at 3.4 mg/L (normal < 0.5 mg/L), and she undergoes computed tomography (CT) of the chest, which demonstrates acute segmental pulmonary emboli (PE) in the right upper and middle lobes as well as multiple bilateral subsegmental PEs. The CT suggests right ventricular dysfunction (RVD), and her troponin T is 0.06 ng/mL (normal < 0.01 ng/mL). Bilateral lower extremity venous Doppler ultrasonography demonstrates no acute thrombus.

Acute pulmonary embolism (PE) accounts for more than 300,000 inpatient admissions annually in the United States.¹ The vast majority of patients with acute PE who receive adequate anticoagulation will have favorable outcomes.^2,3 In the past two decades, for example, mortality has decreased significantly among patients admitted with acute PE,² with 30-day all-cause mortality falling to approximately 5%.³ The risk-adjusted rate of recurrent venous thromboembolism (VTE) within 30 days has concomitantly dropped below 1%.³

Acute PE severity was previously classified as massive or high risk, submassive or intermediate risk, and low risk.⁴ Massive PE was defined by RVD and persistent hypotension or shock requiring vasopressors. ⁴ Intermediate-risk or submassive PE typically referred to normotensive patients with RVD and/or myocardial necrosis (eg, elevated troponin).^4,5 Low-risk PEs had neither hemodynamic instability nor RVD. This classification scheme, however, has fallen out of favor as PE severity exists on a risk spectrum.⁶ Instead, recent guidelines from the European Society of Cardiology and the American College of Chest Physicians recommend first parsing PE severity by the presence or absence of hypotension (Figure 1).^6,7 Risk assessment can be subsequently enhanced by validated clinical risk prediction scores, imaging-based assessment of RVD, and cardiac biomarker testing.⁶

Echocardiography is a common method for evaluating RVD, and echocardiographic RVD confers an increased risk of adverse outcomes in PE.^10-12 In the earliest meta-analysis to evaluate this association, Sanchez et al. combined data from five studies that included 623 patients from emergency room and inpatient settings. They found that echocardiographic RVD conferred an unadjusted relative risk for short-term mortality of 2.53 (95%CI 1.17-5.50).¹² A subsequent meta-analysis by Cho et al. pooled data from both prospective and retrospective cohorts to examine short-term mortality in a total of 3,283 hemodynamically stable patients with PE, of whom 1,223 (37.3%) had RVD diagnosed by echocardiogram.¹⁰ In this population, RVD was associated with an odds ratio of 2.29 (95%CI 1.61-3.26) for short-term death. Thus, echocardiography could be viewed as a risk stratification tool, even in hemodynamically stable PE.

For most hemodynamically stable patients, echocardiographic findings will not enhance prognostication and/or have a therapeutic impact. The following four reasons explain why echocardiography adds little value to the care of these patients.

First, phenotypic expression of RVD varies from asymptomatic, despite abnormalities on diagnostic testing, to obstructive shock. Unfortunately, available prognostic models classify echocardiographic RVD in a binary fashion (present/absent)^4,7,10 whereas RVD exists on a continuum. Consequently, RVD is commonly found in acute PE^8,10,11 and has been identified in more than half of patients hospitalized with PE referred for echocardiography.⁸ Existing data do not allow clinicians to judge the clinical impact of the severity of echocardiographic RVD,⁸ and only the phenotypic expression of refractory hypotension has clear therapeutic implications.^6,7

Second, while echocardiographic RVD is associated with short-term mortality,^10-12 absolute rates of adverse outcomes are quite low when RVD is identified. For example, in a study merging multiple prospective cohorts, Becattini et al. demonstrated that RVD diagnosed by echocardiography or CT occurred in 41% of hospitalized patients stratified to low-risk PE by the simplified Pulmonary Embolism Severity Index (sPESI).⁸ For these patients, the 30-day mortality was 1.2%,⁸ which approximates the expected mortality from a low-risk sPESI score alone (1.1%).¹³ Even among intermediate-risk acute PE patients with RVD and/or elevated troponin enrolled in thrombolysis trials, the overall risk of death at 30 days was approximately 2%-3%, irrespective of the treatment arm.^5,14,15

Third, RVD identified by echocardiography does not inform or enhance prognostication as compared with cardiac biomarker testing. In a meta-analysis by Sanchez et al., echocardiographic RVD predicted death with a risk ratio of 2.53 (95% CI 1.17-5.50).¹² However, both elevated cardiac troponin and brain natriuretic peptide indicated a significantly worse outcome than imaging findings, with risk ratios of 8.3 (95% CI 3.6-19.3) and 9.5 (95% CI 3.2-28.6), respectively.¹³ More recently, Jiménez derived and validated a multivariable risk prediction model for stable PE.¹¹ In their data, echocardiographic RVD had an unadjusted odds ratio of 2.62 (95% CI 1.54-4.45) for predicting a 30-day complicated course. After multivariable adjustment that included sPESI scores, lower extremity ultrasound results, and cardiac biomarker testing, these odds became insignificant.¹¹ In other words, identifying echocardiographic RVD did not improve prognostication in hemodynamically stable PE patients when other commonly available variables were used.

Finally, in hemodynamically stable patients, echocardiographic RVD might create patient anxiety and cause harm. In a recent retrospective cohort study of 64,037 stable patients with PE, exposure to echocardiography was associated with a five-fold increase in likelihood of having received thrombolysis without any significant differences in risk-adjusted mortality.¹⁶ These data suggest that when faced with an abnormal echocardiogram, clinicians and patients may opt for more aggressive, time-sensitive therapies. Basing thrombolysis decisions on echocardiographic RVD potentially subjects patients to harm without decreasing mortality.^5,14,15 For example, the PEITHO study, which was the largest randomized trial evaluating thrombolysis in intermediate-risk acute PE, enrolled 1,006 patients and demonstrated that treating 29 intermediate-risk patients with thrombolysis prevented one case of hemodynamic decompensation.⁵ These benefits were counterbalanced by a number needed to harm of 14 to cause stroke or major bleeding. Ominous echocardiographic findings may also bias clinicians toward more intensive monitoring. Rates of echocardiogram utilization in hemodynamically stable PE are linked to higher rates of ICU admission and longer hospital stays without significant impact on patient outcomes.¹⁶

Echocardiography should be used to exclude other causes of hypotension in patients with presumed PE-related shock^7,9 and to improve clinicians’ confidence prescribing systemic thrombolytics in the face of hemodynamic instability.^6,7 Otherwise, echocardiography should be reserved for highly selected intermediate-risk patients with acute PE. Among patients with intermediate-risk PE, those most likely to decompensate or die typically satisfy all of the following conditions: (1) highest-risk PESI or sPESI scores, (2) elevated natriuretic peptides, (3) elevated troponin, and (4) proximal deep vein thrombosis (DVT) on lower extremity ultrasound.^11,13 In such patients, the echocardiogram may reveal a critical “tipping point,” such as a right atrial or ventricular thrombus-in-transit, that may warrant more intensive monitoring and multidisciplinary input into the most appropriate treatment plan.

Echocardiography could aid therapeutic decisions when the benefits from thrombolysis may outweigh the risks, such as for patients with minimal physiologic reserve and/or a low risk of major bleeding complications. Prognostic models like sPESI utilize binary variables, such as the presence/absence of chronic cardiopulmonary disease or oxygen saturation above/below 90%. Clearly, these variables exist on a spectrum; intuitively, patients with severe comorbidities and more alarming vital signs have a higher risk of death or decompensation than predicted by sPESI. Analogously, echocardiographic findings of RVD also encompass a spectrum. Because prognostic models and clinical trials cannot guide decisions for each individual patient, clinicians could justify using echocardiography to “fine tune” prognostication and to provide a personalized approach for carefully selected patients.

Clinicians should use a risk prediction model for all hemodynamically stable patients with confirmed PE.^6,7 Validated risk calculators include the sPESI,^6,7,14 which relies exclusively on the patient’s history and vital signs, and the eStiMaTe© tool (www.peprognosis.org), which enhances prognostication from sPESI by incorporating troponin, natriuretic peptide, and lower- extremity Doppler results. ¹¹ For patients with symptoms or physical signs of RVD, chest CT and cardiac biomarkers (ie, troponin and/or natriuretic peptides) are sufficient for prognostication.^11,14 In intermediate-risk patients with the highest risk for decompensation based on risk prediction scores, the echocardiogram should represent a part of a comprehensive clinical evaluation, not the sole criterion for intensive monitoring and aggressive treatment.

• Clinicians should use a validated tool, such as the sPESI, for initial risk stratification of hemodynamically stable patients with acute pulmonary embolism.
• Hemodynamically unstable patients with confirmed or suspected acute PE may benefit from early echocardiography to confirm RVD as the cause of shock.^6,7,9
• The majority of normotensive adults with acute PE should not undergo echocardiography. To identify the patients at the greatest risk for decompensation, clinicians may consider using the eStiMaTe© tool (www.peprognosis.org), which augments risk stratification afforded by sPESI.
• For hemodynamically stable patients with PE who have already undergone echocardiography, clinicians should avoid being biased by the finding of RVD, particularly if other prognostic markers are reassuring.

In evaluating the patient described earlier, echocardiography has no clear prognostic implications. Her admission sPESI score equals zero, predicting a 30-day mortality of 1.1%. Including her lower extremity ultrasound and troponin T results into the eStiMaTe© calculator (www.peprognosis.org) surprisingly predicts an even lower rate of 30-day mortality (0.4%) and low risk of a complicated course (2.4%). Assessing for RVD on echocardiography may increase her risk of unnecessary and potentially injurious interventions.

Do you think this is a low-value practice? Is this truly a “Thing We Do for No Reason?” Share what you do in your practice and join in the conversation online by retweeting it on Twitter (#TWDFNR) and liking it on Facebook. We invite you to propose ideas for other “Things We Do for No Reason” topics by emailing[email protected].

Disclosures

The authors have no conflicts of interest relevant to this article.

A 65-year-old man was transferred to a tertiary academic medical center with one week of progressive shortness of breath, dry cough, and fevers. He reported no weight loss or night sweats but had experienced mild right upper quadrant pain and anorexia for the preceding three weeks. Several years had passed since he had consulted a physician, and he did not take any medications. He immigrated to the United States from Mexico four decades prior. He traveled back frequently to visit his family, most recently one month before his presentation. He worked as a farming supervisor in the Central Valley of California. He smoked tobacco and had a 30 pack-year history. He drank alcohol occasionally and denied any drug use.

Causes of subacute cough and dyspnea include bronchitis, pneumonia, heart failure, and asthma. Pneumonia and heart failure might cause right upper quadrant pain from diaphragmatic irritation and hepatic congestion, respectively. Metastatic cancer or infection may lead to synchronous pulmonary and hepatic involvement. The patient is at increased risk of lung cancer, given his extensive smoking history.

The patient’s place of residence in the Southwestern United States places him at risk of respiratory illness from coccidioidomycosis. His exact involvement with animals and their products should be further explored. For example, consumption of unpasteurized milk might result in pneumonia, hepatitis, or both from M. bovis, Brucella species, or C. burnetii. His travel to Mexico prompts consideration of tuberculosis, histoplasmosis, and paracoccidiomycosis as causes of respiratory and possible hepatic illness.

Two weeks prior, the patient had initially presented to another hospital with one week of intermittent right upper quadrant pain unrelated to eating. An abdominal ultrasound and hepatobiliary iminodiacetic acid (HIDA) scan were normal. Computed tomography (CT) of the chest, abdomen, and pelvis with contrast demonstrated a left upper lobe lung mass measuring 5.5 × 4.4 × 3.7 cm³ and scattered right-sided pulmonary nodules (Figure 1). He underwent CT-guided biopsy of the mass and was discharged with a presumed diagnosis of primary pulmonary malignancy with plans for outpatient follow-up.

Over the next four days, the patient developed progressive dyspnea with cough and subjective fevers. The patient was readmitted with a diagnosis of postobstructive pneumonia and acute kidney injury (creatinine increased from 0.7 mg/dL to 2.9 mg/dL between admissions), and this finding was attributed to contrast-induced nephropathy from his recent CT scan. He was treated with vancomycin and piperacillin/tazobactam for two days but wished to transfer to a tertiary care hospital for a second opinion.

Postobstructive pneumonia, pulmonary embolism, and pleural effusion are common causes of dyspnea in patients with lung cancer. The patient’s travel and occupational history, lung nodules, acute renal insufficiency, and rapidly progressive respiratory symptoms prompt consideration for radiographic mimickers of lung cancer. Tuberculosis might present as a lung mass (pulmonary tuberculoma) during primary infection or reactivation. Noninfectious causes of pulmonary masses and nodules include metastatic cancer (eg, colon cancer), sarcoidosis, IgG4-related disease, and granulomatous polyangiitis (GPA).

Contrast-induced nephropathy is unusual in patients with normal renal function. More probable explanations include hypovolemia or acute tubular necrosis (ATN) from underlying inflammation. The patient’s CT-negative right upper quadrant pain may be a distinct process or represent another facet of a disseminated illness such as hepatic infiltration from lymphoma.

Upon arrival, the patient’s temperature was 38°C, heart rate (HR) 107 beats per minute, blood pressure (BP) 159/89 mm Hg, respiratory rate 25 breaths per minute, and oxygen saturation 92% on 2 L of oxygen per minute. He showed no signs of distress. Mild scleral icterus was noted. The cardiac exam was normal. Auscultation revealed scattered wheezes and crackles in the left upper lobe. Mild right upper quadrant tenderness without hepatosplenomegaly was noted on the abdominal exam. The patient’s lower extremities exhibited bilateral trace edema. No rash was observed, and his neurologic exam was normal.

The white blood cell (WBC) count was 28,300 per cubic millimeter (87% neutrophils, 3.6% lymphocytes, and 0.03% eosinophils), hemoglobin 11.1 g per deciliter, and platelet count 789,000 per cubic millimeter. Sodium was 127 mmol per liter, potassium 4.6 mmol per liter, chloride 101 mmol per liter, bicarbonate 13 mmol per liter, blood urea nitrogen 60 mg per deciliter, and creatinine 3.4 mg per deciliter. Aspartate aminotransferase and alanine aminotransferase levels were normal. Alkaline phosphatase was 283 units per liter (normal range, 31-95), and total bilirubin was 4.5 mg per deciliter (normal range, 0.2­-1.3) with a direct bilirubin of 2.7 mg per deciliter. Urinalysis demonstrated urine protein of 30 mg/dL, specific gravity of 1.013, negative nitrites, 10­-21 white cells per high-powered field (normal, < 5), and 21­-50 red cells per high-powered field (normal, < 3). Urine microscopy revealed muddy brown casts but no cellular casts or dysmorphic red cells. A chest radiograph (CXR) showed patchy consolidations in the bilateral upper lobes and left lower lobe along with Kerley B lines, a small left pleural effusion, and thickened right horizontal fissure; the left upper lobe mass was re-demonstrated. Vancomycin, piperacillin-tazobactam, and azithromycin were administered.

At this point, the most likely source of sepsis is multifocal pneumonia. The patient is at risk for S. aureus and P. aeruginosa given his recent hospitalization. A severe form of leptospirosis (Weil’s disease) is associated with pulmonary disease, hyperbilirubinemia, and renal failure. Repeat abdominal imaging is necessary to evaluate for cholangitis given the patient’s right upper quadrant pain, fever, and jaundice. It would also help categorize his cholestatic pattern of liver injury as intrahepatic or extrahepatic (eg, stricture). An infiltrative disease such as sarcoidosis may cause both intrahepatic cholestasis and parenchymal lung disease, although the pleural pathology is uncommon.

His normal cardiac exam does not exclude cardiogenic pulmonary edema, a common cause of interstitial edema and pleural effusion. In this setting of systemic inflammation (neutrophilia, thrombocytosis, and hypoalbuminemia), the thickened right horizontal fissure and interlobular septa might represent an infiltrative process, such as lymphangitic carcinomatosis, lymphoma, or sarcoidosis.

Muddy brown casts are characteristic of ATN. The patient’s risk factors for ATN include sepsis and previously administered iodinated contrast. Fluid retention from oliguric renal failure is likely contributing to his hyponatremia and lower extremity edema. Pathology isolated to the tubules, however, would not cause hematuria and pyuria and suggests glomerular or interstitial disease. The lack of cellular casts on a single urinary specimen does not eliminate the likelihood of either disease. Hematuria and diffuse parenchymal lung disease prompt consideration of pulmonary-renal syndromes, such as anti-glomerular basement membrane disease, GPA, and systemic lupus erythematosus, which can all be triggered by infection.

On the night of transfer, the patient experienced acute respiratory distress. Heart rate was 130 beats per minute, BP 170/95 mm Hg, respiratory rate 40 breaths per minute, and oxygen saturation 88% on six liters of supplemental oxygen by nasal cannula. His arterial blood gas demonstrated a pH of 7.23, PaCO₂ of 32 mm Hg, and PaO₂ of 65 mm Hg. He was emergently intubated for progressive hypoxemic respiratory failure. A small amount of blood was noted in the endotracheal tube. A noncontrast CT of the chest demonstrated multifocal airspace opacities and bilateral pleural effusions. The previously noted left upper lobe mass was unchanged.

Rapid respiratory decline and diffuse alveolar disease commonly result from aspiration, flash pulmonary edema, and acute respiratory distress syndrome (ARDS). Necrotizing pneumonia (eg, S. aureus) and trauma during intubation are possible causes of blood in his endotracheal tube. However, in the setting of multifocal airspace opacity, renal insufficiency, hematuria, and rapid respiratory decline, the blood might represent diffuse alveolar hemorrhage (DAH). Bronchoscopy with bronchioalveolar lavage to evaluate for pulmonary edema, infection, and hemorrhage would be indicated.

The patient subsequently developed oliguria, requiring continuous renal replacement therapy. An echocardiogram demonstrated impaired left ventricular relaxation and a reduced ejection fraction of 45% without segmental wall motion abnormalities or valvular disease, and a right ventricular systolic pressure of 36 mm Hg. Over the next 12 hours, his respiratory status improved, and he was extubated to 15 L per minute of supplemental oxygen by high-flow nasal cannula (HFNC).

The pathology report of the lung biopsy from the other hospital disclosed chronic inflammation and fibrosis with ill-defined areas of necrosis and myxoid degeneration surrounded by nuclear palisading suggestive of granulomatous inflammation. Staining for acid-fast bacilli (AFB) and fungal organisms was negative.

The rapid pulmonary recovery is inconsistent with multifocal pneumonia or ARDS. Flash pulmonary edema might result in sudden hypoxemic respiratory failure that resolves with positive pressure ventilation and ultrafiltration. However, this condition would not explain the biopsy results. Granulomatous lung pathology often results from mycobacterial or fungal disease. Tuberculosis and fungal pneumonia are not excluded with negative staining alone. However, neither would cause self-limited respiratory failure. Histologic evidence of necrosis lessens the likelihood of sarcoidosis, which rarely causes fulminant pulmonary disease. Lymphoma can result in granulomatous inflammation but would not cause transient pulmonary disease. GPA, a cause of necrotizing granulomatous lung disease, might result in a lung mass and worsened hypoxemia through DAH.

The patient continued to require 15 L of oxygen per minute by HFNC. He had persistent bilateral perihilar alveolar and interstitial opacities on CXR. Repeat WBC count was 29,200 per cubic millimeter, hemoglobin 7.8 g per deciliter, and platelets 656,000 per cubic millimeter. The C-reactive protein was 300 mg per L (normal range, <6.3) and erythrocyte sedimentation rate 100 mm per hour (normal range, <10). Legionella urinary antigen, serum immunodiffusion for Coccidiodes imitus, human immunodeficiency virus antibody, respiratory viral panel, and beta-D glucan were negative. Rare acid-fast bacilli were visualized in one out of three concentrated AFB sputum smears. He was started on empiric antituberculous therapy with rifampin, isoniazid, pyrazinamide, and ethambutol.

The sputum sample is suggestive of pulmonary tuberculosis. The salient features of this case include systemic inflammation, pulmonary nodules and mass, necrotizing granulomatous lung pathology, renal insufficiency, and hematuria. Disseminated tuberculosis might explain all these findings. However, a positive AFB smear may signal the presence of a nontuberculous mycobacteria, which is less likely to cause this clinical syndrome.

M. tuberculosis complex polymerase chain reaction (MTB PCR) assay returned negative for M. tuberculosis. Antiproteinase 3 antibody was 1,930 units (normal range, <20). Antimyeloperoxidase and antiglomerular basement membrane antibodies were negative.

Tuberculosis and GPA share several overlapping features, such as necrotizing lung pathology and less commonly antineutrophil cytoplasmic autoantibody (ANCA)-associated antibodies. However, the lung mass, acute renal and respiratory failure, hematuria, and the degree of anti-proteinase 3 level elevation are highly suggestive of GPA. The negative MTB PCR raises the possibility that a nontuberculous mycobacterium was detected on the sputum smear. Nevertheless, continued treatment until finalization of culture results is appropriate given that tuberculosis is endemic in Mexico.

The patient’s presenting features of right upper quadrant tenderness, jaundice, and cholestatic hepatitis remain poorly explained by either of these diagnoses.  Neither tuberculosis nor GPA commonly presents with accompanying hepatic involvement, though both have been occasionally described as causing hepatitis. As the greatest concern in this patient remains his progressive renal failure and accompanying pulmonary hemorrhage, a renal biopsy to assess for glomerulonephritis associated with GPA is warranted before further investigation into the cause of his cholestatic hepatitis.

A core renal biopsy demonstrated pauci-immune focal crescentic and necrotizing glomerulonephritis with mixed tubulointerstitial inflammation (Figure 2). In conjunction with the pulmonary syndrome and positive antiproteinase 3 serology, a diagnosis of granulomatosis with polyangiitis was made. The patient was treated with pulse dose steroids, rituximab, and plasma exchange. Two weeks later, the sputum mycobacterial culture returned positive for Mycobacterium llatzerense and anti-tuberculous treatment was discontinued.

Over the following weeks, the patient improved and was transitioned off dialysis prior to hospital discharge. By six months later, he had resolution of his hemoptysis, shortness of breath, liver biochemical test abnormalities, and significant improvement in his renal function. Repeat sputum mycobacterial cultures were negative.

A 65-year-old man from Mexico with a significant smoking history presented with an apical lung mass and cough, prioritizing tuberculosis and pulmonary malignancy. As the case unfolded, renal failure, multifocal lung opacities, conflicting tuberculosis test results, positive anti-proteinase 3 antibody, and ultimately a renal biopsy led to the diagnosis of granulomatosis with polyangiitis (GPA).

The correct interpretation of occasionally conflicting mycobacterial testing is crucial. Mycobacterial cultures remain the gold standard for diagnosing tuberculosis. However, results take weeks to return. Rapid tests include acid-fast bacilli (AFB) smear microscopy and nucleic acid-amplification tests (NAAT) of sputum or bronchoalveolar samples.¹ When three sputum smears are performed, the sensitivity of AFB smear microscopy for tuberculosis in immunocompetent hosts is 70%.¹ The AFB smear does not distinguish between different mycobacterial organisms. Thus, a positive result must be interpreted with the relative prevalence of tuberculosis and nontuberculous mycobacteria (NTM) in mind. The addition of NAAT-based assays has allowed for enhanced sensitivity and specificity in the diagnosis of tuberculosis, such that a negative NAAT in a patient with a positive AFB smear strongly argues for the presence of a NTM.^2-4

NTM are widely prevalent environmental microbes, with over 140 species described, and careful consideration is required to determine if an isolate is pathogenic.⁵ Given their ubiquitous nature, a high rate of asymptomatic respiratory and cutaneous colonization occurs. Correspondingly, the diagnosis of NTM disease requires multiple positive cultures or pathologic features on tissue biopsy, compatible clinical findings, and diligent exclusion of other causes.⁵ A retrospective study of all NTM isolates in Oregon from 2005­-2006 revealed that only 47% of patients met the guideline criteria for having symptomatic NTM disease.⁶ In our case, the patient’s sputum grew M. llatzerense, an aerobic, nonfermenting mycobacterium found in water sources that has only infrequently been implicated as a human pathogen.^7,8 Subsequent AFB sputum cultures were negative, and serial imaging showed resolution of the pulmonary findings without additional antimycobacterial therapy, suggesting that this organism was not responsible for the disease process.

Along with microscopic polyangiitis (MPA) and eosinophilic granulomatosis with polyangiitis (EGPA), GPA is an antineutrophil cytoplasmic autoantibody (ANCA)-associated vasculitis that predominantly affects small to medium sized vessels. Although it can occur at any age, GPA most commonly afflicts older adults, with men and women being diagnosed at roughly equal rates.⁹ GPA is a multisystem disease with a wide array of clinical manifestations. The most frequently involved sites of disease are the respiratory tract and kidneys, although virtually any organ can be affected. Sino-nasal disease, such as destructive sinusitis, or ear involvement are nearly universal. Lower respiratory manifestations occur in 60% of patients, but are highly diverse and reflect the inherent difficulty in diagnosing this condition.^9-11 Additionally, GPA is a frequent cause of the pulmonary-renal syndromes, with glomerulonephritis occurring in 80% of patients.⁹

The diagnosis of GPA in this case was delayed, in part, due to features suggestive of malignancy and pulmonary tuberculosis. While sino-nasal disease was not noted during this hospitalization, the patient had many different respiratory manifestations, including a dominant pulmonary mass, diffuse nodules, and hypoxemic respiratory failure due to suspected diffuse alveolar hemorrhage (DAH), all of which have been reported in GPA.¹² Dysmorphic red cells and red blood cell casts are not sensitive for renal involvement in GPA; their absence does not exclude the possibility of an ANCA-associated vasculitis.¹³ Hematuria and rapid progression to oliguric renal failure are characteristic of a vasculitic process and should sway clinicians away from a working diagnosis of ATN.

The diagnosis of GPA involves the synthesis of clinical data, radiographic findings, serologic testing, and histopathology. ANCA testing is an essential step in the diagnosis of GPA but has limitations. Patients with GPA more commonly have ANCAs targeting the enzyme proteinase-3 (PR3-ANCA), with MPA being more closely associated with myeloperoxidase (MPO-ANCA), although cross-reactivity and antibody-negative disease can occur.¹⁴ Although 90% of patients with GPA with multiorgan involvement will have a positive ANCA, a negative test is more common in localized upper airway disease, where only 50% have a positive ANCA.¹⁵ A number of drugs, medications, infections, and nonvasculitic autoimmune diseases have been associated with positive ANCA serologies in the absence of systemic vasculitis.^14,16,17 As such, pathologic demonstration of vasculitis is necessary for establishing the diagnosis. Typical sites for biopsy include the kidneys and lungs.⁹

This case illustrates how clinicians often find themselves at a diagnostic crossroads—being forced to choose which clinical elements to prioritize. At various points, our patient’s presentation could have been framed as “a man from a Tb-endemic country with hemoptysis and an apical opacity,” “an elderly man with extensive smoking history and lung mass,” or “a patient with elevated inflammatory markers and pulmonary-renal syndrome”. In such situations, it is incumbent on the clinician to evaluate how well a given problem representation encompasses or highlights the salient features of a case. As with painting or photography, an essential aspect of appreciating the whole picture involves carefully selecting the right frame.

• The diagnosis of tuberculosis relies on smear microscopy, nucleic-acid amplification testing (NAAT), and cultures. A positive AFB smear with negative NAAT suggests the presence of a nontuberculous mycobacteria (NTM).
• NTM are common environmental organisms and often exist as nonpathogenic colonizers.⁶ The diagnosis of NTM disease requires exclusion of other causes and careful clinical, microbiologic, and radiographic correlation.
• Granulomatosis with polyangiitis is a multisystem disease often involving the respiratory track and kidney. Pulmonary disease can present with airway involvement, parenchymal nodules, opacities, pleural findings, and diffuse alveolar hemorrhage.¹²

Disclosures

Drs. Minter, Geha, Boslett, Chung, and Ramani have no disclosures. Dr. Manesh is supported by the Jeremiah A. Barondess Fellowship in the Clinical Transaction of the New York Academy of Medicine, in collaboration with the Accreditation Council for Graduate Medical Education (ACGME).

A 65-year-old man was transferred to a tertiary academic medical center with one week of progressive shortness of breath, dry cough, and fevers. He reported no weight loss or night sweats but had experienced mild right upper quadrant pain and anorexia for the preceding three weeks. Several years had passed since he had consulted a physician, and he did not take any medications. He immigrated to the United States from Mexico four decades prior. He traveled back frequently to visit his family, most recently one month before his presentation. He worked as a farming supervisor in the Central Valley of California. He smoked tobacco and had a 30 pack-year history. He drank alcohol occasionally and denied any drug use.

Causes of subacute cough and dyspnea include bronchitis, pneumonia, heart failure, and asthma. Pneumonia and heart failure might cause right upper quadrant pain from diaphragmatic irritation and hepatic congestion, respectively. Metastatic cancer or infection may lead to synchronous pulmonary and hepatic involvement. The patient is at increased risk of lung cancer, given his extensive smoking history.

The patient’s place of residence in the Southwestern United States places him at risk of respiratory illness from coccidioidomycosis. His exact involvement with animals and their products should be further explored. For example, consumption of unpasteurized milk might result in pneumonia, hepatitis, or both from M. bovis, Brucella species, or C. burnetii. His travel to Mexico prompts consideration of tuberculosis, histoplasmosis, and paracoccidiomycosis as causes of respiratory and possible hepatic illness.

Two weeks prior, the patient had initially presented to another hospital with one week of intermittent right upper quadrant pain unrelated to eating. An abdominal ultrasound and hepatobiliary iminodiacetic acid (HIDA) scan were normal. Computed tomography (CT) of the chest, abdomen, and pelvis with contrast demonstrated a left upper lobe lung mass measuring 5.5 × 4.4 × 3.7 cm³ and scattered right-sided pulmonary nodules (Figure 1). He underwent CT-guided biopsy of the mass and was discharged with a presumed diagnosis of primary pulmonary malignancy with plans for outpatient follow-up.

Over the next four days, the patient developed progressive dyspnea with cough and subjective fevers. The patient was readmitted with a diagnosis of postobstructive pneumonia and acute kidney injury (creatinine increased from 0.7 mg/dL to 2.9 mg/dL between admissions), and this finding was attributed to contrast-induced nephropathy from his recent CT scan. He was treated with vancomycin and piperacillin/tazobactam for two days but wished to transfer to a tertiary care hospital for a second opinion.

Postobstructive pneumonia, pulmonary embolism, and pleural effusion are common causes of dyspnea in patients with lung cancer. The patient’s travel and occupational history, lung nodules, acute renal insufficiency, and rapidly progressive respiratory symptoms prompt consideration for radiographic mimickers of lung cancer. Tuberculosis might present as a lung mass (pulmonary tuberculoma) during primary infection or reactivation. Noninfectious causes of pulmonary masses and nodules include metastatic cancer (eg, colon cancer), sarcoidosis, IgG4-related disease, and granulomatous polyangiitis (GPA).

Contrast-induced nephropathy is unusual in patients with normal renal function. More probable explanations include hypovolemia or acute tubular necrosis (ATN) from underlying inflammation. The patient’s CT-negative right upper quadrant pain may be a distinct process or represent another facet of a disseminated illness such as hepatic infiltration from lymphoma.

Upon arrival, the patient’s temperature was 38°C, heart rate (HR) 107 beats per minute, blood pressure (BP) 159/89 mm Hg, respiratory rate 25 breaths per minute, and oxygen saturation 92% on 2 L of oxygen per minute. He showed no signs of distress. Mild scleral icterus was noted. The cardiac exam was normal. Auscultation revealed scattered wheezes and crackles in the left upper lobe. Mild right upper quadrant tenderness without hepatosplenomegaly was noted on the abdominal exam. The patient’s lower extremities exhibited bilateral trace edema. No rash was observed, and his neurologic exam was normal.

The white blood cell (WBC) count was 28,300 per cubic millimeter (87% neutrophils, 3.6% lymphocytes, and 0.03% eosinophils), hemoglobin 11.1 g per deciliter, and platelet count 789,000 per cubic millimeter. Sodium was 127 mmol per liter, potassium 4.6 mmol per liter, chloride 101 mmol per liter, bicarbonate 13 mmol per liter, blood urea nitrogen 60 mg per deciliter, and creatinine 3.4 mg per deciliter. Aspartate aminotransferase and alanine aminotransferase levels were normal. Alkaline phosphatase was 283 units per liter (normal range, 31-95), and total bilirubin was 4.5 mg per deciliter (normal range, 0.2­-1.3) with a direct bilirubin of 2.7 mg per deciliter. Urinalysis demonstrated urine protein of 30 mg/dL, specific gravity of 1.013, negative nitrites, 10­-21 white cells per high-powered field (normal, < 5), and 21­-50 red cells per high-powered field (normal, < 3). Urine microscopy revealed muddy brown casts but no cellular casts or dysmorphic red cells. A chest radiograph (CXR) showed patchy consolidations in the bilateral upper lobes and left lower lobe along with Kerley B lines, a small left pleural effusion, and thickened right horizontal fissure; the left upper lobe mass was re-demonstrated. Vancomycin, piperacillin-tazobactam, and azithromycin were administered.

At this point, the most likely source of sepsis is multifocal pneumonia. The patient is at risk for S. aureus and P. aeruginosa given his recent hospitalization. A severe form of leptospirosis (Weil’s disease) is associated with pulmonary disease, hyperbilirubinemia, and renal failure. Repeat abdominal imaging is necessary to evaluate for cholangitis given the patient’s right upper quadrant pain, fever, and jaundice. It would also help categorize his cholestatic pattern of liver injury as intrahepatic or extrahepatic (eg, stricture). An infiltrative disease such as sarcoidosis may cause both intrahepatic cholestasis and parenchymal lung disease, although the pleural pathology is uncommon.

His normal cardiac exam does not exclude cardiogenic pulmonary edema, a common cause of interstitial edema and pleural effusion. In this setting of systemic inflammation (neutrophilia, thrombocytosis, and hypoalbuminemia), the thickened right horizontal fissure and interlobular septa might represent an infiltrative process, such as lymphangitic carcinomatosis, lymphoma, or sarcoidosis.

Muddy brown casts are characteristic of ATN. The patient’s risk factors for ATN include sepsis and previously administered iodinated contrast. Fluid retention from oliguric renal failure is likely contributing to his hyponatremia and lower extremity edema. Pathology isolated to the tubules, however, would not cause hematuria and pyuria and suggests glomerular or interstitial disease. The lack of cellular casts on a single urinary specimen does not eliminate the likelihood of either disease. Hematuria and diffuse parenchymal lung disease prompt consideration of pulmonary-renal syndromes, such as anti-glomerular basement membrane disease, GPA, and systemic lupus erythematosus, which can all be triggered by infection.

On the night of transfer, the patient experienced acute respiratory distress. Heart rate was 130 beats per minute, BP 170/95 mm Hg, respiratory rate 40 breaths per minute, and oxygen saturation 88% on six liters of supplemental oxygen by nasal cannula. His arterial blood gas demonstrated a pH of 7.23, PaCO₂ of 32 mm Hg, and PaO₂ of 65 mm Hg. He was emergently intubated for progressive hypoxemic respiratory failure. A small amount of blood was noted in the endotracheal tube. A noncontrast CT of the chest demonstrated multifocal airspace opacities and bilateral pleural effusions. The previously noted left upper lobe mass was unchanged.

Rapid respiratory decline and diffuse alveolar disease commonly result from aspiration, flash pulmonary edema, and acute respiratory distress syndrome (ARDS). Necrotizing pneumonia (eg, S. aureus) and trauma during intubation are possible causes of blood in his endotracheal tube. However, in the setting of multifocal airspace opacity, renal insufficiency, hematuria, and rapid respiratory decline, the blood might represent diffuse alveolar hemorrhage (DAH). Bronchoscopy with bronchioalveolar lavage to evaluate for pulmonary edema, infection, and hemorrhage would be indicated.

The patient subsequently developed oliguria, requiring continuous renal replacement therapy. An echocardiogram demonstrated impaired left ventricular relaxation and a reduced ejection fraction of 45% without segmental wall motion abnormalities or valvular disease, and a right ventricular systolic pressure of 36 mm Hg. Over the next 12 hours, his respiratory status improved, and he was extubated to 15 L per minute of supplemental oxygen by high-flow nasal cannula (HFNC).

The pathology report of the lung biopsy from the other hospital disclosed chronic inflammation and fibrosis with ill-defined areas of necrosis and myxoid degeneration surrounded by nuclear palisading suggestive of granulomatous inflammation. Staining for acid-fast bacilli (AFB) and fungal organisms was negative.

The rapid pulmonary recovery is inconsistent with multifocal pneumonia or ARDS. Flash pulmonary edema might result in sudden hypoxemic respiratory failure that resolves with positive pressure ventilation and ultrafiltration. However, this condition would not explain the biopsy results. Granulomatous lung pathology often results from mycobacterial or fungal disease. Tuberculosis and fungal pneumonia are not excluded with negative staining alone. However, neither would cause self-limited respiratory failure. Histologic evidence of necrosis lessens the likelihood of sarcoidosis, which rarely causes fulminant pulmonary disease. Lymphoma can result in granulomatous inflammation but would not cause transient pulmonary disease. GPA, a cause of necrotizing granulomatous lung disease, might result in a lung mass and worsened hypoxemia through DAH.

The patient continued to require 15 L of oxygen per minute by HFNC. He had persistent bilateral perihilar alveolar and interstitial opacities on CXR. Repeat WBC count was 29,200 per cubic millimeter, hemoglobin 7.8 g per deciliter, and platelets 656,000 per cubic millimeter. The C-reactive protein was 300 mg per L (normal range, <6.3) and erythrocyte sedimentation rate 100 mm per hour (normal range, <10). Legionella urinary antigen, serum immunodiffusion for Coccidiodes imitus, human immunodeficiency virus antibody, respiratory viral panel, and beta-D glucan were negative. Rare acid-fast bacilli were visualized in one out of three concentrated AFB sputum smears. He was started on empiric antituberculous therapy with rifampin, isoniazid, pyrazinamide, and ethambutol.

The sputum sample is suggestive of pulmonary tuberculosis. The salient features of this case include systemic inflammation, pulmonary nodules and mass, necrotizing granulomatous lung pathology, renal insufficiency, and hematuria. Disseminated tuberculosis might explain all these findings. However, a positive AFB smear may signal the presence of a nontuberculous mycobacteria, which is less likely to cause this clinical syndrome.

M. tuberculosis complex polymerase chain reaction (MTB PCR) assay returned negative for M. tuberculosis. Antiproteinase 3 antibody was 1,930 units (normal range, <20). Antimyeloperoxidase and antiglomerular basement membrane antibodies were negative.

Tuberculosis and GPA share several overlapping features, such as necrotizing lung pathology and less commonly antineutrophil cytoplasmic autoantibody (ANCA)-associated antibodies. However, the lung mass, acute renal and respiratory failure, hematuria, and the degree of anti-proteinase 3 level elevation are highly suggestive of GPA. The negative MTB PCR raises the possibility that a nontuberculous mycobacterium was detected on the sputum smear. Nevertheless, continued treatment until finalization of culture results is appropriate given that tuberculosis is endemic in Mexico.

The patient’s presenting features of right upper quadrant tenderness, jaundice, and cholestatic hepatitis remain poorly explained by either of these diagnoses.  Neither tuberculosis nor GPA commonly presents with accompanying hepatic involvement, though both have been occasionally described as causing hepatitis. As the greatest concern in this patient remains his progressive renal failure and accompanying pulmonary hemorrhage, a renal biopsy to assess for glomerulonephritis associated with GPA is warranted before further investigation into the cause of his cholestatic hepatitis.

A core renal biopsy demonstrated pauci-immune focal crescentic and necrotizing glomerulonephritis with mixed tubulointerstitial inflammation (Figure 2). In conjunction with the pulmonary syndrome and positive antiproteinase 3 serology, a diagnosis of granulomatosis with polyangiitis was made. The patient was treated with pulse dose steroids, rituximab, and plasma exchange. Two weeks later, the sputum mycobacterial culture returned positive for Mycobacterium llatzerense and anti-tuberculous treatment was discontinued.

Over the following weeks, the patient improved and was transitioned off dialysis prior to hospital discharge. By six months later, he had resolution of his hemoptysis, shortness of breath, liver biochemical test abnormalities, and significant improvement in his renal function. Repeat sputum mycobacterial cultures were negative.

A 65-year-old man from Mexico with a significant smoking history presented with an apical lung mass and cough, prioritizing tuberculosis and pulmonary malignancy. As the case unfolded, renal failure, multifocal lung opacities, conflicting tuberculosis test results, positive anti-proteinase 3 antibody, and ultimately a renal biopsy led to the diagnosis of granulomatosis with polyangiitis (GPA).

The correct interpretation of occasionally conflicting mycobacterial testing is crucial. Mycobacterial cultures remain the gold standard for diagnosing tuberculosis. However, results take weeks to return. Rapid tests include acid-fast bacilli (AFB) smear microscopy and nucleic acid-amplification tests (NAAT) of sputum or bronchoalveolar samples.¹ When three sputum smears are performed, the sensitivity of AFB smear microscopy for tuberculosis in immunocompetent hosts is 70%.¹ The AFB smear does not distinguish between different mycobacterial organisms. Thus, a positive result must be interpreted with the relative prevalence of tuberculosis and nontuberculous mycobacteria (NTM) in mind. The addition of NAAT-based assays has allowed for enhanced sensitivity and specificity in the diagnosis of tuberculosis, such that a negative NAAT in a patient with a positive AFB smear strongly argues for the presence of a NTM.^2-4

NTM are widely prevalent environmental microbes, with over 140 species described, and careful consideration is required to determine if an isolate is pathogenic.⁵ Given their ubiquitous nature, a high rate of asymptomatic respiratory and cutaneous colonization occurs. Correspondingly, the diagnosis of NTM disease requires multiple positive cultures or pathologic features on tissue biopsy, compatible clinical findings, and diligent exclusion of other causes.⁵ A retrospective study of all NTM isolates in Oregon from 2005­-2006 revealed that only 47% of patients met the guideline criteria for having symptomatic NTM disease.⁶ In our case, the patient’s sputum grew M. llatzerense, an aerobic, nonfermenting mycobacterium found in water sources that has only infrequently been implicated as a human pathogen.^7,8 Subsequent AFB sputum cultures were negative, and serial imaging showed resolution of the pulmonary findings without additional antimycobacterial therapy, suggesting that this organism was not responsible for the disease process.

Along with microscopic polyangiitis (MPA) and eosinophilic granulomatosis with polyangiitis (EGPA), GPA is an antineutrophil cytoplasmic autoantibody (ANCA)-associated vasculitis that predominantly affects small to medium sized vessels. Although it can occur at any age, GPA most commonly afflicts older adults, with men and women being diagnosed at roughly equal rates.⁹ GPA is a multisystem disease with a wide array of clinical manifestations. The most frequently involved sites of disease are the respiratory tract and kidneys, although virtually any organ can be affected. Sino-nasal disease, such as destructive sinusitis, or ear involvement are nearly universal. Lower respiratory manifestations occur in 60% of patients, but are highly diverse and reflect the inherent difficulty in diagnosing this condition.^9-11 Additionally, GPA is a frequent cause of the pulmonary-renal syndromes, with glomerulonephritis occurring in 80% of patients.⁹

The diagnosis of GPA in this case was delayed, in part, due to features suggestive of malignancy and pulmonary tuberculosis. While sino-nasal disease was not noted during this hospitalization, the patient had many different respiratory manifestations, including a dominant pulmonary mass, diffuse nodules, and hypoxemic respiratory failure due to suspected diffuse alveolar hemorrhage (DAH), all of which have been reported in GPA.¹² Dysmorphic red cells and red blood cell casts are not sensitive for renal involvement in GPA; their absence does not exclude the possibility of an ANCA-associated vasculitis.¹³ Hematuria and rapid progression to oliguric renal failure are characteristic of a vasculitic process and should sway clinicians away from a working diagnosis of ATN.

The diagnosis of GPA involves the synthesis of clinical data, radiographic findings, serologic testing, and histopathology. ANCA testing is an essential step in the diagnosis of GPA but has limitations. Patients with GPA more commonly have ANCAs targeting the enzyme proteinase-3 (PR3-ANCA), with MPA being more closely associated with myeloperoxidase (MPO-ANCA), although cross-reactivity and antibody-negative disease can occur.¹⁴ Although 90% of patients with GPA with multiorgan involvement will have a positive ANCA, a negative test is more common in localized upper airway disease, where only 50% have a positive ANCA.¹⁵ A number of drugs, medications, infections, and nonvasculitic autoimmune diseases have been associated with positive ANCA serologies in the absence of systemic vasculitis.^14,16,17 As such, pathologic demonstration of vasculitis is necessary for establishing the diagnosis. Typical sites for biopsy include the kidneys and lungs.⁹

This case illustrates how clinicians often find themselves at a diagnostic crossroads—being forced to choose which clinical elements to prioritize. At various points, our patient’s presentation could have been framed as “a man from a Tb-endemic country with hemoptysis and an apical opacity,” “an elderly man with extensive smoking history and lung mass,” or “a patient with elevated inflammatory markers and pulmonary-renal syndrome”. In such situations, it is incumbent on the clinician to evaluate how well a given problem representation encompasses or highlights the salient features of a case. As with painting or photography, an essential aspect of appreciating the whole picture involves carefully selecting the right frame.

• The diagnosis of tuberculosis relies on smear microscopy, nucleic-acid amplification testing (NAAT), and cultures. A positive AFB smear with negative NAAT suggests the presence of a nontuberculous mycobacteria (NTM).
• NTM are common environmental organisms and often exist as nonpathogenic colonizers.⁶ The diagnosis of NTM disease requires exclusion of other causes and careful clinical, microbiologic, and radiographic correlation.
• Granulomatosis with polyangiitis is a multisystem disease often involving the respiratory track and kidney. Pulmonary disease can present with airway involvement, parenchymal nodules, opacities, pleural findings, and diffuse alveolar hemorrhage.¹²

Disclosures

Drs. Minter, Geha, Boslett, Chung, and Ramani have no disclosures. Dr. Manesh is supported by the Jeremiah A. Barondess Fellowship in the Clinical Transaction of the New York Academy of Medicine, in collaboration with the Accreditation Council for Graduate Medical Education (ACGME).

A 65-year-old man was transferred to a tertiary academic medical center with one week of progressive shortness of breath, dry cough, and fevers. He reported no weight loss or night sweats but had experienced mild right upper quadrant pain and anorexia for the preceding three weeks. Several years had passed since he had consulted a physician, and he did not take any medications. He immigrated to the United States from Mexico four decades prior. He traveled back frequently to visit his family, most recently one month before his presentation. He worked as a farming supervisor in the Central Valley of California. He smoked tobacco and had a 30 pack-year history. He drank alcohol occasionally and denied any drug use.

Causes of subacute cough and dyspnea include bronchitis, pneumonia, heart failure, and asthma. Pneumonia and heart failure might cause right upper quadrant pain from diaphragmatic irritation and hepatic congestion, respectively. Metastatic cancer or infection may lead to synchronous pulmonary and hepatic involvement. The patient is at increased risk of lung cancer, given his extensive smoking history.

The patient’s place of residence in the Southwestern United States places him at risk of respiratory illness from coccidioidomycosis. His exact involvement with animals and their products should be further explored. For example, consumption of unpasteurized milk might result in pneumonia, hepatitis, or both from M. bovis, Brucella species, or C. burnetii. His travel to Mexico prompts consideration of tuberculosis, histoplasmosis, and paracoccidiomycosis as causes of respiratory and possible hepatic illness.

Two weeks prior, the patient had initially presented to another hospital with one week of intermittent right upper quadrant pain unrelated to eating. An abdominal ultrasound and hepatobiliary iminodiacetic acid (HIDA) scan were normal. Computed tomography (CT) of the chest, abdomen, and pelvis with contrast demonstrated a left upper lobe lung mass measuring 5.5 × 4.4 × 3.7 cm³ and scattered right-sided pulmonary nodules (Figure 1). He underwent CT-guided biopsy of the mass and was discharged with a presumed diagnosis of primary pulmonary malignancy with plans for outpatient follow-up.

Over the next four days, the patient developed progressive dyspnea with cough and subjective fevers. The patient was readmitted with a diagnosis of postobstructive pneumonia and acute kidney injury (creatinine increased from 0.7 mg/dL to 2.9 mg/dL between admissions), and this finding was attributed to contrast-induced nephropathy from his recent CT scan. He was treated with vancomycin and piperacillin/tazobactam for two days but wished to transfer to a tertiary care hospital for a second opinion.

Postobstructive pneumonia, pulmonary embolism, and pleural effusion are common causes of dyspnea in patients with lung cancer. The patient’s travel and occupational history, lung nodules, acute renal insufficiency, and rapidly progressive respiratory symptoms prompt consideration for radiographic mimickers of lung cancer. Tuberculosis might present as a lung mass (pulmonary tuberculoma) during primary infection or reactivation. Noninfectious causes of pulmonary masses and nodules include metastatic cancer (eg, colon cancer), sarcoidosis, IgG4-related disease, and granulomatous polyangiitis (GPA).

Contrast-induced nephropathy is unusual in patients with normal renal function. More probable explanations include hypovolemia or acute tubular necrosis (ATN) from underlying inflammation. The patient’s CT-negative right upper quadrant pain may be a distinct process or represent another facet of a disseminated illness such as hepatic infiltration from lymphoma.

Upon arrival, the patient’s temperature was 38°C, heart rate (HR) 107 beats per minute, blood pressure (BP) 159/89 mm Hg, respiratory rate 25 breaths per minute, and oxygen saturation 92% on 2 L of oxygen per minute. He showed no signs of distress. Mild scleral icterus was noted. The cardiac exam was normal. Auscultation revealed scattered wheezes and crackles in the left upper lobe. Mild right upper quadrant tenderness without hepatosplenomegaly was noted on the abdominal exam. The patient’s lower extremities exhibited bilateral trace edema. No rash was observed, and his neurologic exam was normal.

The white blood cell (WBC) count was 28,300 per cubic millimeter (87% neutrophils, 3.6% lymphocytes, and 0.03% eosinophils), hemoglobin 11.1 g per deciliter, and platelet count 789,000 per cubic millimeter. Sodium was 127 mmol per liter, potassium 4.6 mmol per liter, chloride 101 mmol per liter, bicarbonate 13 mmol per liter, blood urea nitrogen 60 mg per deciliter, and creatinine 3.4 mg per deciliter. Aspartate aminotransferase and alanine aminotransferase levels were normal. Alkaline phosphatase was 283 units per liter (normal range, 31-95), and total bilirubin was 4.5 mg per deciliter (normal range, 0.2­-1.3) with a direct bilirubin of 2.7 mg per deciliter. Urinalysis demonstrated urine protein of 30 mg/dL, specific gravity of 1.013, negative nitrites, 10­-21 white cells per high-powered field (normal, < 5), and 21­-50 red cells per high-powered field (normal, < 3). Urine microscopy revealed muddy brown casts but no cellular casts or dysmorphic red cells. A chest radiograph (CXR) showed patchy consolidations in the bilateral upper lobes and left lower lobe along with Kerley B lines, a small left pleural effusion, and thickened right horizontal fissure; the left upper lobe mass was re-demonstrated. Vancomycin, piperacillin-tazobactam, and azithromycin were administered.

At this point, the most likely source of sepsis is multifocal pneumonia. The patient is at risk for S. aureus and P. aeruginosa given his recent hospitalization. A severe form of leptospirosis (Weil’s disease) is associated with pulmonary disease, hyperbilirubinemia, and renal failure. Repeat abdominal imaging is necessary to evaluate for cholangitis given the patient’s right upper quadrant pain, fever, and jaundice. It would also help categorize his cholestatic pattern of liver injury as intrahepatic or extrahepatic (eg, stricture). An infiltrative disease such as sarcoidosis may cause both intrahepatic cholestasis and parenchymal lung disease, although the pleural pathology is uncommon.

His normal cardiac exam does not exclude cardiogenic pulmonary edema, a common cause of interstitial edema and pleural effusion. In this setting of systemic inflammation (neutrophilia, thrombocytosis, and hypoalbuminemia), the thickened right horizontal fissure and interlobular septa might represent an infiltrative process, such as lymphangitic carcinomatosis, lymphoma, or sarcoidosis.

Muddy brown casts are characteristic of ATN. The patient’s risk factors for ATN include sepsis and previously administered iodinated contrast. Fluid retention from oliguric renal failure is likely contributing to his hyponatremia and lower extremity edema. Pathology isolated to the tubules, however, would not cause hematuria and pyuria and suggests glomerular or interstitial disease. The lack of cellular casts on a single urinary specimen does not eliminate the likelihood of either disease. Hematuria and diffuse parenchymal lung disease prompt consideration of pulmonary-renal syndromes, such as anti-glomerular basement membrane disease, GPA, and systemic lupus erythematosus, which can all be triggered by infection.

On the night of transfer, the patient experienced acute respiratory distress. Heart rate was 130 beats per minute, BP 170/95 mm Hg, respiratory rate 40 breaths per minute, and oxygen saturation 88% on six liters of supplemental oxygen by nasal cannula. His arterial blood gas demonstrated a pH of 7.23, PaCO₂ of 32 mm Hg, and PaO₂ of 65 mm Hg. He was emergently intubated for progressive hypoxemic respiratory failure. A small amount of blood was noted in the endotracheal tube. A noncontrast CT of the chest demonstrated multifocal airspace opacities and bilateral pleural effusions. The previously noted left upper lobe mass was unchanged.

Rapid respiratory decline and diffuse alveolar disease commonly result from aspiration, flash pulmonary edema, and acute respiratory distress syndrome (ARDS). Necrotizing pneumonia (eg, S. aureus) and trauma during intubation are possible causes of blood in his endotracheal tube. However, in the setting of multifocal airspace opacity, renal insufficiency, hematuria, and rapid respiratory decline, the blood might represent diffuse alveolar hemorrhage (DAH). Bronchoscopy with bronchioalveolar lavage to evaluate for pulmonary edema, infection, and hemorrhage would be indicated.

The patient subsequently developed oliguria, requiring continuous renal replacement therapy. An echocardiogram demonstrated impaired left ventricular relaxation and a reduced ejection fraction of 45% without segmental wall motion abnormalities or valvular disease, and a right ventricular systolic pressure of 36 mm Hg. Over the next 12 hours, his respiratory status improved, and he was extubated to 15 L per minute of supplemental oxygen by high-flow nasal cannula (HFNC).

The pathology report of the lung biopsy from the other hospital disclosed chronic inflammation and fibrosis with ill-defined areas of necrosis and myxoid degeneration surrounded by nuclear palisading suggestive of granulomatous inflammation. Staining for acid-fast bacilli (AFB) and fungal organisms was negative.

The rapid pulmonary recovery is inconsistent with multifocal pneumonia or ARDS. Flash pulmonary edema might result in sudden hypoxemic respiratory failure that resolves with positive pressure ventilation and ultrafiltration. However, this condition would not explain the biopsy results. Granulomatous lung pathology often results from mycobacterial or fungal disease. Tuberculosis and fungal pneumonia are not excluded with negative staining alone. However, neither would cause self-limited respiratory failure. Histologic evidence of necrosis lessens the likelihood of sarcoidosis, which rarely causes fulminant pulmonary disease. Lymphoma can result in granulomatous inflammation but would not cause transient pulmonary disease. GPA, a cause of necrotizing granulomatous lung disease, might result in a lung mass and worsened hypoxemia through DAH.

The patient continued to require 15 L of oxygen per minute by HFNC. He had persistent bilateral perihilar alveolar and interstitial opacities on CXR. Repeat WBC count was 29,200 per cubic millimeter, hemoglobin 7.8 g per deciliter, and platelets 656,000 per cubic millimeter. The C-reactive protein was 300 mg per L (normal range, <6.3) and erythrocyte sedimentation rate 100 mm per hour (normal range, <10). Legionella urinary antigen, serum immunodiffusion for Coccidiodes imitus, human immunodeficiency virus antibody, respiratory viral panel, and beta-D glucan were negative. Rare acid-fast bacilli were visualized in one out of three concentrated AFB sputum smears. He was started on empiric antituberculous therapy with rifampin, isoniazid, pyrazinamide, and ethambutol.

The sputum sample is suggestive of pulmonary tuberculosis. The salient features of this case include systemic inflammation, pulmonary nodules and mass, necrotizing granulomatous lung pathology, renal insufficiency, and hematuria. Disseminated tuberculosis might explain all these findings. However, a positive AFB smear may signal the presence of a nontuberculous mycobacteria, which is less likely to cause this clinical syndrome.

M. tuberculosis complex polymerase chain reaction (MTB PCR) assay returned negative for M. tuberculosis. Antiproteinase 3 antibody was 1,930 units (normal range, <20). Antimyeloperoxidase and antiglomerular basement membrane antibodies were negative.

Tuberculosis and GPA share several overlapping features, such as necrotizing lung pathology and less commonly antineutrophil cytoplasmic autoantibody (ANCA)-associated antibodies. However, the lung mass, acute renal and respiratory failure, hematuria, and the degree of anti-proteinase 3 level elevation are highly suggestive of GPA. The negative MTB PCR raises the possibility that a nontuberculous mycobacterium was detected on the sputum smear. Nevertheless, continued treatment until finalization of culture results is appropriate given that tuberculosis is endemic in Mexico.

The patient’s presenting features of right upper quadrant tenderness, jaundice, and cholestatic hepatitis remain poorly explained by either of these diagnoses.  Neither tuberculosis nor GPA commonly presents with accompanying hepatic involvement, though both have been occasionally described as causing hepatitis. As the greatest concern in this patient remains his progressive renal failure and accompanying pulmonary hemorrhage, a renal biopsy to assess for glomerulonephritis associated with GPA is warranted before further investigation into the cause of his cholestatic hepatitis.

A core renal biopsy demonstrated pauci-immune focal crescentic and necrotizing glomerulonephritis with mixed tubulointerstitial inflammation (Figure 2). In conjunction with the pulmonary syndrome and positive antiproteinase 3 serology, a diagnosis of granulomatosis with polyangiitis was made. The patient was treated with pulse dose steroids, rituximab, and plasma exchange. Two weeks later, the sputum mycobacterial culture returned positive for Mycobacterium llatzerense and anti-tuberculous treatment was discontinued.

Over the following weeks, the patient improved and was transitioned off dialysis prior to hospital discharge. By six months later, he had resolution of his hemoptysis, shortness of breath, liver biochemical test abnormalities, and significant improvement in his renal function. Repeat sputum mycobacterial cultures were negative.

A 65-year-old man from Mexico with a significant smoking history presented with an apical lung mass and cough, prioritizing tuberculosis and pulmonary malignancy. As the case unfolded, renal failure, multifocal lung opacities, conflicting tuberculosis test results, positive anti-proteinase 3 antibody, and ultimately a renal biopsy led to the diagnosis of granulomatosis with polyangiitis (GPA).

The correct interpretation of occasionally conflicting mycobacterial testing is crucial. Mycobacterial cultures remain the gold standard for diagnosing tuberculosis. However, results take weeks to return. Rapid tests include acid-fast bacilli (AFB) smear microscopy and nucleic acid-amplification tests (NAAT) of sputum or bronchoalveolar samples.¹ When three sputum smears are performed, the sensitivity of AFB smear microscopy for tuberculosis in immunocompetent hosts is 70%.¹ The AFB smear does not distinguish between different mycobacterial organisms. Thus, a positive result must be interpreted with the relative prevalence of tuberculosis and nontuberculous mycobacteria (NTM) in mind. The addition of NAAT-based assays has allowed for enhanced sensitivity and specificity in the diagnosis of tuberculosis, such that a negative NAAT in a patient with a positive AFB smear strongly argues for the presence of a NTM.^2-4

NTM are widely prevalent environmental microbes, with over 140 species described, and careful consideration is required to determine if an isolate is pathogenic.⁵ Given their ubiquitous nature, a high rate of asymptomatic respiratory and cutaneous colonization occurs. Correspondingly, the diagnosis of NTM disease requires multiple positive cultures or pathologic features on tissue biopsy, compatible clinical findings, and diligent exclusion of other causes.⁵ A retrospective study of all NTM isolates in Oregon from 2005­-2006 revealed that only 47% of patients met the guideline criteria for having symptomatic NTM disease.⁶ In our case, the patient’s sputum grew M. llatzerense, an aerobic, nonfermenting mycobacterium found in water sources that has only infrequently been implicated as a human pathogen.^7,8 Subsequent AFB sputum cultures were negative, and serial imaging showed resolution of the pulmonary findings without additional antimycobacterial therapy, suggesting that this organism was not responsible for the disease process.

Along with microscopic polyangiitis (MPA) and eosinophilic granulomatosis with polyangiitis (EGPA), GPA is an antineutrophil cytoplasmic autoantibody (ANCA)-associated vasculitis that predominantly affects small to medium sized vessels. Although it can occur at any age, GPA most commonly afflicts older adults, with men and women being diagnosed at roughly equal rates.⁹ GPA is a multisystem disease with a wide array of clinical manifestations. The most frequently involved sites of disease are the respiratory tract and kidneys, although virtually any organ can be affected. Sino-nasal disease, such as destructive sinusitis, or ear involvement are nearly universal. Lower respiratory manifestations occur in 60% of patients, but are highly diverse and reflect the inherent difficulty in diagnosing this condition.^9-11 Additionally, GPA is a frequent cause of the pulmonary-renal syndromes, with glomerulonephritis occurring in 80% of patients.⁹

The diagnosis of GPA in this case was delayed, in part, due to features suggestive of malignancy and pulmonary tuberculosis. While sino-nasal disease was not noted during this hospitalization, the patient had many different respiratory manifestations, including a dominant pulmonary mass, diffuse nodules, and hypoxemic respiratory failure due to suspected diffuse alveolar hemorrhage (DAH), all of which have been reported in GPA.¹² Dysmorphic red cells and red blood cell casts are not sensitive for renal involvement in GPA; their absence does not exclude the possibility of an ANCA-associated vasculitis.¹³ Hematuria and rapid progression to oliguric renal failure are characteristic of a vasculitic process and should sway clinicians away from a working diagnosis of ATN.

The diagnosis of GPA involves the synthesis of clinical data, radiographic findings, serologic testing, and histopathology. ANCA testing is an essential step in the diagnosis of GPA but has limitations. Patients with GPA more commonly have ANCAs targeting the enzyme proteinase-3 (PR3-ANCA), with MPA being more closely associated with myeloperoxidase (MPO-ANCA), although cross-reactivity and antibody-negative disease can occur.¹⁴ Although 90% of patients with GPA with multiorgan involvement will have a positive ANCA, a negative test is more common in localized upper airway disease, where only 50% have a positive ANCA.¹⁵ A number of drugs, medications, infections, and nonvasculitic autoimmune diseases have been associated with positive ANCA serologies in the absence of systemic vasculitis.^14,16,17 As such, pathologic demonstration of vasculitis is necessary for establishing the diagnosis. Typical sites for biopsy include the kidneys and lungs.⁹

This case illustrates how clinicians often find themselves at a diagnostic crossroads—being forced to choose which clinical elements to prioritize. At various points, our patient’s presentation could have been framed as “a man from a Tb-endemic country with hemoptysis and an apical opacity,” “an elderly man with extensive smoking history and lung mass,” or “a patient with elevated inflammatory markers and pulmonary-renal syndrome”. In such situations, it is incumbent on the clinician to evaluate how well a given problem representation encompasses or highlights the salient features of a case. As with painting or photography, an essential aspect of appreciating the whole picture involves carefully selecting the right frame.

• The diagnosis of tuberculosis relies on smear microscopy, nucleic-acid amplification testing (NAAT), and cultures. A positive AFB smear with negative NAAT suggests the presence of a nontuberculous mycobacteria (NTM).
• NTM are common environmental organisms and often exist as nonpathogenic colonizers.⁶ The diagnosis of NTM disease requires exclusion of other causes and careful clinical, microbiologic, and radiographic correlation.
• Granulomatosis with polyangiitis is a multisystem disease often involving the respiratory track and kidney. Pulmonary disease can present with airway involvement, parenchymal nodules, opacities, pleural findings, and diffuse alveolar hemorrhage.¹²

Disclosures

Drs. Minter, Geha, Boslett, Chung, and Ramani have no disclosures. Dr. Manesh is supported by the Jeremiah A. Barondess Fellowship in the Clinical Transaction of the New York Academy of Medicine, in collaboration with the Accreditation Council for Graduate Medical Education (ACGME).