Original Research

ABSTRACT

Osteoarthritis (OA) of the knee is a top cause of disability among the elderly. Total knee replacement (TKR) has been available as an effective and definite surgical method to treat severe OA of the knee. However, TKR is a significant procedure with potential risk for serious complications and high costs. Alternative lower risk therapies that can delay or obviate TKR are valuable to those who are poor candidates for surgery or wish to avoid TKR as long as possible. Given the chondroprotective effects of hyaluronic acid (HA) injections, they are a safe and effective treatment to improve pain, function, and longevity of the knee. Thus, HA features the potential to delay or obviate TKR.

We aim to study the safety and effectiveness of repeated courses of HA on the time to TKR over a 3-year period using data from a large US health plan administrative claims database.

Retrospective analyses were conducted by identifying knee OA patients during the selection period (2007-2010). The follow-up period was 36 months, post-index date of initial HA injection. Procedural outcomes and adverse events of interest were tabulated and analyzed. A Cox proportional hazards model was used to model the risk of TKR.

A total of 50,389 patients who received HA for treatment of knee OA and met the study inclusion criteria were analyzed. Successive courses of HA showed a good safety profile and led to high proportions of patients without TKR 3 years after treatment initiation. Multivariate statistical modeling showed that multiple courses of HA injections significantly decreased the rates of TKR (95.0% without TKR for ≥5 courses vs 71.6% without TKR for 1 course; hazard ratio, 0.138; P < .0001).

Repeated courses of treatment with HA are safe and are associated with the delay of TKR for up to 3 years. Additional research is needed to evaluate the effect of repeated HA courses on delaying TKR beyond a 3-year time horizon.

Continue to: Osteoarthritis (OA) of the knee...

Osteoarthritis (OA) of the knee has emerged as one of the main causes of disability in the United States. Although no currently known cure of OA can reverse the progression of the disease, total knee replacement (TKR) is an effective and definitive treatment. However, TKR is an invasive procedure with potential risk for serious complications, and it has imposed high costs on the US healthcare system, with expenses accounting for hospital expenditures of TKR estimated at $28.5 billion in 2009.¹Alternative low-risk therapies that can delay or obviate TKR are valuable to a number of patients, especially the poor candidates for surgery or those who wish to avoid TKR.

Intra-articular (IA) hyaluronic acid (HA) injections have been available as a safe and effective treatment option to alleviate pain and to improve joint functions.² Results of randomized double-blind controlled clinical trials have demonstrated the pain-relieving effect of IA HA injections.^3-5 Furthermore, a recent network meta-analysis comparing various pharmacologic interventions for knee OA has confirmed the efficacy of IA HA injections, which outperformed other interventions when compared with oral placebos.^6,7 IA therapies are more effective than oral therapies for knee OA pain, with IA HA injections demonstrating the most pain reduction, potentially due to the benefit associated with needle injection and aspiration. Recent experimental studies have also suggested that IA HA may provide cartilage protection, reduce inflammation, and boost the viscosity of synovial fluid;⁸ IA HA may also exert therapeutic effects by inhibiting bone formation in OA patients.^9,10 HA possesses the potential to delay or obviate TKR. Previous research with a case series review of patients in an orthopedic specialty practice reported that the use of IA HA injections in patients with grade IV OA delayed TKR substantially.¹¹ One study analyzed retrospective medical claims data from a single private insurer and discovered potential evidence for the modest benefit of IA HA injections in delaying TKR.¹²

More detailed research work on a large sample of patients with knee OA and the requirement of TKR as a condition for inclusion using US administrative claims data has demonstrated the TKR-delaying effects of IA HA injections in comparison with a control group without claims for IA HA injections.^13,14 This study also uses real-world US administrative data but utilizes a different approach by starting with a sample of patients with knee OA and evidence of IA HA injections and then assessing the effect of repeated courses of HA treatment on the delay of TKR, without TKR as a mandatory condition for inclusion. All patients with knee OA within the time window were included, regardless of the need for TKR compared with previous studies which only considered patients who ultimately received TKR. Safety information and effectiveness information were examined to achieve a balanced risk-benefit assessment. We also analyzed how multiple courses of HA treatment and other potentially relevant covariates at baseline affected the risk of receiving TKR in a multivariate survival model. We aimed to achieve a realistic assessment of the clinical utility of HA injections in delaying TKR in a real-world setting using both safety and effectiveness data.

METHODS

DATA SOURCE

A retrospective cohort observational study using IMS Health’s PharMetrics Plus Health Plan Claims Database was conducted by identifying knee OA patients with claims indicating initiation of HA injection at an index date during the selection period (July 1, 2007 to June 30, 2010). All common HA agents in the US market during this period (Euflexxa, Hyalgan, Orthovisc, Supartz, and Synvisc) were selected via the corresponding J-codes and pooled for investigation of HA class effects. The follow-up period was 36 months, post-index date of the initial HA injection. Outcomes were measured, and adverse events were identified during this period. The time window for identification of adverse events was within 2 weeks from any injection during the course of therapy (evidence of an emergency room visit and/or physician office visit with requisite code). The data during the 12-month pre-index baseline period from the claims database was used to obtain information about baseline patient characteristics, such as age, gender, type of coverage, physician specialty, Charlson Comorbidity Index (CCI), major comorbidities, and major medications of interest commonly used among patients with knee OA.

STUDY SAMPLE SELECTION

The eligible patients required an outpatient claim indicating the initiation of HA injection. The date of the first claim for the patient within the selection window was defined as their index date. Patients had to be ≥18 years of age in the year of their index date. They had to present at least 1 clinical knee OA diagnosis at any point in the 12-month pre-index period (including the index date), and only patients who were continuously enrolled from 12 months pre-index to 36 months post-index date were evaluated. Among these patients (approximately 1.4 million), the following were excluded to minimize complications in data analysis and interpretation: patients with evidence of any HA use in the pre-index period; patients with evidence of a different kind of HA index medication in the post-index period; patients with evidence of TKR within 30 days of the index event during the post-index period; patients with evidence of 2 different kinds of HA index medications on the index date; and patients with evidence of diagnosis of hip OA, fibromyalgia, rheumatoid arthritis, lupus, or gout during the pre-index period.

Five patient cohorts were defined according to the number of courses of IA HA injections over the entire post-index period.

Continue to: Statistical analysis...

STATISTICAL ANALYSIS

All statistical analyses were performed using SAS version 9.2 (SAS Institute Inc.). Descriptive statistics such as means, standard deviations, medians, and 25% and 75% percentiles (Q1 and Q3, respectively) were provided for the continuous variables. Numbers and percentages were provided for the categorical variables. For statistical testing, Student’s t-tests were applied for the continuous variables and chi-square tests for the categorical variables. All the statistical tests were two-tailed. The sample sizes in this database study are remarkably large, such that differences that are not clinically important could still be statistically significant at the conventional alpha level of 0.05. Thus, we applied a more stringent requirement of the alpha level of 0.0001 to identify highly statistically significant results. The number and percentage of patients within each cohort with at least 1 instance of an adverse event of interest (those adverse events commonly expected for patients who receive IA injections for knee OA) were assessed. Times to TKR during the 36-month post-index period were analyzed and compared among different cohorts. Any patients who had not undergone TKR by the end of the post-index period were considered censored at 36 months. The Kaplan-Meier method was employed to model survival curves with time to TKR data, and log-rank tests were used to compare survival curves among different cohorts. A Cox proportional hazards model (PHM) was used to model the risk of TKR with a pre-specified set of covariates adjusted for baseline attributes, such as age, gender, comorbidities, and pre-index healthcare costs. Hazard ratios with 95% confidence intervals were used to examine the measures of event risk.

RESULTS

PATIENT CHARACTERISTICS

Applying study selection criteria to the claims database yielded 50,389 patients (Figure 1), providing an ample sample size for the statistical analysis. Only patients with evidence of knee OA and use of HA injections (the index medication of interest) were selected, regardless of whether they received TKR during the post-index period. The requirement for a knee OA diagnosis during the 12-month pre-index period resulted in the significant attrition of patients, with 584,956 patients being excluded. Among the 50,389 patients who received HA for treatment of knee OA, 36,260 (72.0%) received a single course of treatment, 8709 (17.3%) received 2 courses, 3179 (6.3%) received 3 courses, 1354 (2.7%) received 4 courses, and 887 (1.8%) received ≥5 courses of treatment.

Comparison of baseline characteristics among the 5 IA HA cohorts showed the fairly similar baseline characteristics of all cohorts (Table 1). Geographic region, physician specialty, and opioid use showed differences among the cohorts. Cohorts with ≥5 HA courses presented lower proportions of patients from Southern US states, patients seeing orthopedic surgeons, and patients using opioids than cohorts with fewer HA courses.

PROCEDURES OF INTEREST

An analysis of the procedures patients received after HA treatment initiation showed that higher numbers of HA treatment courses resulted in lower proportions of patients receiving TKR within 3 years after HA treatment initiation (Table 2). With an increasing number of HA treatment courses, the proportion of patients with TKR within 3 years post-index consistently decreased from 28.4% (for 1 HA course) to 5.0% (for ≥5 HA courses), with all differences being highly statistically significant (P < .0001). Similarly, partial knee replacement exhibited a similar trend, with the proportion of patients decreasing from 3.3% (for 1 HA course) to 0.8% (for ≥5 HA courses; P < .0001). Among the patients with TKR within 3 years post-index, increasing numbers of treatment courses correlated with increasing time to TKR, with a mean of 375.6 days (for 1 HA course) rising to a mean of 971.5 days (for ≥5 HA courses; P < .0001). On the other hand, patients with multiple courses of HA treatment were more likely to undergo radiologic examinations of the knee, arthrocenteses, and image-guided injections than patients with only a single course of HA treatment (P < .0001).

ADVERSE EVENTS

Arthralgia and joint pain in the knee were the most commonly recorded adverse events (Table 3). More courses of HA treatment were associated with higher rates of adverse events. Overall, the reported adverse events profile of repeated courses of HA treatment consisted of mostly common and mild adverse events and displayed no safety concern for patients with knee OA that was followed-up for 3 years. The causality of these adverse events directly related to HA injections vs a specific disease state cannot be determined from an administrative claims data set.

TIME TO TKR

Successive courses of HA led to high proportions of patients without TKR 3 years after HA treatment initiation. This result is evident in the Kaplan-Meier survival curves of time to TKR for different HA cohorts (Figure 2), with log-rank tests of multiple courses vs a single course of HA (P < .0001) showing highly statistically significance. Tabulation of proportions of patients without TKR by various time points showed that increasing numbers of HA treatment courses correlated with higher proportions of patients without TKR at almost all time points (Table 4); within 3 years post-index, 71.6% of patients in the 1 HA course cohort exhibited no TKR, whereas 95.0% of patients in ≥5 HA courses cohort presented no TKR. We also performed a multivariate Cox PHM (Table 5) to account for baseline characteristics of different HA cohorts with covariates when estimating the risks of receiving TKR. The results of the Cox PHM showed that multiple courses of HA treatment significantly decreased the risk of TKR (hazard ratio, 0.138 for ≥5 HA courses vs 1 HA course; P < .0001). Inspection of other highly significant covariates showed that being older, living in the Midwest region of the US (vs the Northeast), receiving pre-index corticosteroids, having an orthopedic surgeon as a treating physician (vs a general practitioner, a rheumatologist, or a physical medicine and rehabilitation specialist), experiencing hypertension or hyperlipidemia, and higher pre-index total healthcare costs were associated with an increased risk of TKR (all P < .0001). Vascular disease and high CCI scores were associated with a decreased risk of TKR (P < .0001).

Continue to: Discussion...

DISCUSSION

This study demonstrated that multiple courses of HA treatment can delay the need for surgery for up to 3 years, with risk for both TKR and partial knee replacement decreasing in a dose-dependent manner. The potentially confounding effect of differences in baseline characteristics that could influence patients’ propensity to receive TKR in a database study was controlled by performing a multivariate analysis with covariate adjustment. The TKR-delaying effect of HA injection was more prominent in cohorts with a high number of HA treatment courses: 19 out of 20 patients in the cohort of ≥5 HA courses were free of TKR at the end of the 3-year post-index period. Such a high proportion of patients avoiding TKR with repeated courses of HA suggests that some patients may be able to successfully delay TKR well beyond the 3-year time span. This finding is counter-evidence to the frequently made assumption¹⁵ that all patients with knee OA will eventually progress to a state of disability, making TKR inevitable. The patients with end-stage radiographic knee OA can also benefit from IA HA injections for an extended period of time;¹⁶ the latest evidence indicates that nonoperative management can improve symptoms irrespective of radiographic disease severity, implying that TKR needs not to be the only therapeutic option for patients with end-stage radiographic knee OA.¹⁷ This finding suggests that HA treatment should be considered an important clinical treatment option for patients with knee OA.

Although the incidence rates of certain adverse events, such as arthralgia/joint pain, are sizable, these temporary adverse events commonly occur among patients who receive IA injections for knee OA; most of these events may simply include symptoms of the remaining underlying knee OA. These results are consistent with those of previous literature reporting the safety of repeated treatment with IA HA injections in a prospective clinical trial¹⁸ and demonstrating that repeated courses of HA treatment pose no greater safety risk than a single course of HA treatment.

Multivariate modeling outcomes of factors influencing risk of receiving TKR are broadly consistent with the generally accepted notions that different levels of disease severity and patients’ willingness to consider TKR at baseline influence the likelihood and timing of receiving TKR.^19,20 Age and obesity are common risk factors for progression of OA. Orthopedic surgeons are more likely to recommend surgery than non-surgeons. The pre-index use of corticosteroids and high pre-index healthcare costs could be associated with more severe symptoms at baseline. Patients with vascular disease or severe comorbidities, as evidenced by high CCI scores, make poor candidates for major elective surgeries such as TKR. These results are intuitive and validate the clinical insights of this study. Moreover, inclusion of these covariates in the analysis model allows for indirect adjustment of the most important prognostic factors for TKR at baseline, permitting proper statistical comparison of the results for different cohort groups.

Recently, the efficacy of HA injections for OA patients has become the subject of debate when the American Academy of Orthopaedic Surgeons (AAOS) revised its clinical practice guideline, recommending against the use of HA.²¹ The AAOS’ findings differ from those of other clinical societies, such as the American College of Rheumatology²² and the European League Against Rheumatism,²³ which provide no strong recommendation against the use of HA injections. The announcement of the new guideline by AAOS caused concern among clinicians and payers who had valued IA HA injections as a means to control knee OA pain before patients progress to TKR;²⁴ on the other hand, the demand for nonoperative treatment of knee OA remains high. Utilization rates of TKR have increased dramatically, and surgeries are now performed on younger patients with increasing burden on the healthcare system,^25,26 in spite of the fact that as high as a third of TKR surgeries may have been performed in inappropriate patients.²⁷ Part of the confusion surrounding clinical utility of HA stems from the fact that up until recently, relatively little research looked into the practical benefits of HA in actual clinical practice. Analyses of databases such as registries are now gaining attention to overcome that problem. Examination of large administrative databases maintained by commercial payers offers the benefit of probing realistically the safety and efficacy of treatments in actual clinical environments in a very large number of patients with heterogeneous backgrounds. Recently, the Agency for Healthcare Research and Quality’s Technology Assessment Program in the US called for such studies to determine whether HA injections can delay progression to TKR.²⁸ The results of this study and several others^11,13,14,16 suggest that use of HA to treat OA of the knee is associated with the delay of TKR, supporting the utility of HA in clinical practice and the healthcare system. Potential clinical benefits of delaying TKR may include the reduced risk of aseptic loosening if younger patients can wait for TKR or more time to allow the modification of risk factors in patients who will ultimately undergo TKR.

LIMITATIONS

Follow-up period was limited to 3 years post-index date because longer follow-up data were not available at the time of the study design. If an incorrect adverse event or OA diagnosis was listed in the medical record, or if the medical record was incomplete, then patients might have been misclassified, resulting in selection bias. The claims dataset includes no uninsured and Medicare patients, as the population in the database consisted primarily of commercially-insured patients in the US. Therefore, the results are most generalizable to other commercially-insured patients in the US. Generalizability to other populations may not be assured if they differ in their accessibility to physician services or prescriptions from the patients in this study. Other treatments such as the nonsteroidal anti-inflammatory drugs used by patients were not included within the pre-specified statistical model because their potential effects were assumed to be short-lived and much less than those of corticosteroid. Including these treatments would overload the statistical model with too many covariates, leading to potential computational instability. The database used provides no information on systemic factors, including plan limits on medication use, that could affect care. Given the large and diverse nature of the healthcare plans in the database. However, these factors should not have materially affected our study results. The claims database also lacks direct indicators of OA disease severity, such as Kellgren-Lawrence scores or patient-reported outcomes, including pain and function questionnaire scores. Our multivariate analysis indirectly makes up for this deficiency by considering other baseline characteristics or clinical indicators that may be correlated with information unavailable in a claims database. Patients who opt to undergo repeated courses of HA treatment may be more inclined to avoid surgery or may naturally experience OA disease progression more slowly, making them potentially different from patients who select to undergo surgery earlier without repeated courses of HA treatment. This condition may introduce a bias that causes difficulty in proving the causality between repeated HA use and delay of TKR.

CONCLUSION

Analysis of the knee OA patient data from a real-world database showed that repeated courses of treatment with HA are safe and are associated with the delay of TKR for up to 3 years. Additional research is needed to evaluate the effects of repeated HA courses on delaying TKR beyond a 3-year period.

ABSTRACT

Osteoarthritis (OA) of the knee is a top cause of disability among the elderly. Total knee replacement (TKR) has been available as an effective and definite surgical method to treat severe OA of the knee. However, TKR is a significant procedure with potential risk for serious complications and high costs. Alternative lower risk therapies that can delay or obviate TKR are valuable to those who are poor candidates for surgery or wish to avoid TKR as long as possible. Given the chondroprotective effects of hyaluronic acid (HA) injections, they are a safe and effective treatment to improve pain, function, and longevity of the knee. Thus, HA features the potential to delay or obviate TKR.

We aim to study the safety and effectiveness of repeated courses of HA on the time to TKR over a 3-year period using data from a large US health plan administrative claims database.

Retrospective analyses were conducted by identifying knee OA patients during the selection period (2007-2010). The follow-up period was 36 months, post-index date of initial HA injection. Procedural outcomes and adverse events of interest were tabulated and analyzed. A Cox proportional hazards model was used to model the risk of TKR.

A total of 50,389 patients who received HA for treatment of knee OA and met the study inclusion criteria were analyzed. Successive courses of HA showed a good safety profile and led to high proportions of patients without TKR 3 years after treatment initiation. Multivariate statistical modeling showed that multiple courses of HA injections significantly decreased the rates of TKR (95.0% without TKR for ≥5 courses vs 71.6% without TKR for 1 course; hazard ratio, 0.138; P < .0001).

Repeated courses of treatment with HA are safe and are associated with the delay of TKR for up to 3 years. Additional research is needed to evaluate the effect of repeated HA courses on delaying TKR beyond a 3-year time horizon.

Continue to: Osteoarthritis (OA) of the knee...

Osteoarthritis (OA) of the knee has emerged as one of the main causes of disability in the United States. Although no currently known cure of OA can reverse the progression of the disease, total knee replacement (TKR) is an effective and definitive treatment. However, TKR is an invasive procedure with potential risk for serious complications, and it has imposed high costs on the US healthcare system, with expenses accounting for hospital expenditures of TKR estimated at $28.5 billion in 2009.¹Alternative low-risk therapies that can delay or obviate TKR are valuable to a number of patients, especially the poor candidates for surgery or those who wish to avoid TKR.

Intra-articular (IA) hyaluronic acid (HA) injections have been available as a safe and effective treatment option to alleviate pain and to improve joint functions.² Results of randomized double-blind controlled clinical trials have demonstrated the pain-relieving effect of IA HA injections.^3-5 Furthermore, a recent network meta-analysis comparing various pharmacologic interventions for knee OA has confirmed the efficacy of IA HA injections, which outperformed other interventions when compared with oral placebos.^6,7 IA therapies are more effective than oral therapies for knee OA pain, with IA HA injections demonstrating the most pain reduction, potentially due to the benefit associated with needle injection and aspiration. Recent experimental studies have also suggested that IA HA may provide cartilage protection, reduce inflammation, and boost the viscosity of synovial fluid;⁸ IA HA may also exert therapeutic effects by inhibiting bone formation in OA patients.^9,10 HA possesses the potential to delay or obviate TKR. Previous research with a case series review of patients in an orthopedic specialty practice reported that the use of IA HA injections in patients with grade IV OA delayed TKR substantially.¹¹ One study analyzed retrospective medical claims data from a single private insurer and discovered potential evidence for the modest benefit of IA HA injections in delaying TKR.¹²

More detailed research work on a large sample of patients with knee OA and the requirement of TKR as a condition for inclusion using US administrative claims data has demonstrated the TKR-delaying effects of IA HA injections in comparison with a control group without claims for IA HA injections.^13,14 This study also uses real-world US administrative data but utilizes a different approach by starting with a sample of patients with knee OA and evidence of IA HA injections and then assessing the effect of repeated courses of HA treatment on the delay of TKR, without TKR as a mandatory condition for inclusion. All patients with knee OA within the time window were included, regardless of the need for TKR compared with previous studies which only considered patients who ultimately received TKR. Safety information and effectiveness information were examined to achieve a balanced risk-benefit assessment. We also analyzed how multiple courses of HA treatment and other potentially relevant covariates at baseline affected the risk of receiving TKR in a multivariate survival model. We aimed to achieve a realistic assessment of the clinical utility of HA injections in delaying TKR in a real-world setting using both safety and effectiveness data.

METHODS

DATA SOURCE

A retrospective cohort observational study using IMS Health’s PharMetrics Plus Health Plan Claims Database was conducted by identifying knee OA patients with claims indicating initiation of HA injection at an index date during the selection period (July 1, 2007 to June 30, 2010). All common HA agents in the US market during this period (Euflexxa, Hyalgan, Orthovisc, Supartz, and Synvisc) were selected via the corresponding J-codes and pooled for investigation of HA class effects. The follow-up period was 36 months, post-index date of the initial HA injection. Outcomes were measured, and adverse events were identified during this period. The time window for identification of adverse events was within 2 weeks from any injection during the course of therapy (evidence of an emergency room visit and/or physician office visit with requisite code). The data during the 12-month pre-index baseline period from the claims database was used to obtain information about baseline patient characteristics, such as age, gender, type of coverage, physician specialty, Charlson Comorbidity Index (CCI), major comorbidities, and major medications of interest commonly used among patients with knee OA.

STUDY SAMPLE SELECTION

The eligible patients required an outpatient claim indicating the initiation of HA injection. The date of the first claim for the patient within the selection window was defined as their index date. Patients had to be ≥18 years of age in the year of their index date. They had to present at least 1 clinical knee OA diagnosis at any point in the 12-month pre-index period (including the index date), and only patients who were continuously enrolled from 12 months pre-index to 36 months post-index date were evaluated. Among these patients (approximately 1.4 million), the following were excluded to minimize complications in data analysis and interpretation: patients with evidence of any HA use in the pre-index period; patients with evidence of a different kind of HA index medication in the post-index period; patients with evidence of TKR within 30 days of the index event during the post-index period; patients with evidence of 2 different kinds of HA index medications on the index date; and patients with evidence of diagnosis of hip OA, fibromyalgia, rheumatoid arthritis, lupus, or gout during the pre-index period.

Five patient cohorts were defined according to the number of courses of IA HA injections over the entire post-index period.

Continue to: Statistical analysis...

STATISTICAL ANALYSIS

All statistical analyses were performed using SAS version 9.2 (SAS Institute Inc.). Descriptive statistics such as means, standard deviations, medians, and 25% and 75% percentiles (Q1 and Q3, respectively) were provided for the continuous variables. Numbers and percentages were provided for the categorical variables. For statistical testing, Student’s t-tests were applied for the continuous variables and chi-square tests for the categorical variables. All the statistical tests were two-tailed. The sample sizes in this database study are remarkably large, such that differences that are not clinically important could still be statistically significant at the conventional alpha level of 0.05. Thus, we applied a more stringent requirement of the alpha level of 0.0001 to identify highly statistically significant results. The number and percentage of patients within each cohort with at least 1 instance of an adverse event of interest (those adverse events commonly expected for patients who receive IA injections for knee OA) were assessed. Times to TKR during the 36-month post-index period were analyzed and compared among different cohorts. Any patients who had not undergone TKR by the end of the post-index period were considered censored at 36 months. The Kaplan-Meier method was employed to model survival curves with time to TKR data, and log-rank tests were used to compare survival curves among different cohorts. A Cox proportional hazards model (PHM) was used to model the risk of TKR with a pre-specified set of covariates adjusted for baseline attributes, such as age, gender, comorbidities, and pre-index healthcare costs. Hazard ratios with 95% confidence intervals were used to examine the measures of event risk.

RESULTS

PATIENT CHARACTERISTICS

Applying study selection criteria to the claims database yielded 50,389 patients (Figure 1), providing an ample sample size for the statistical analysis. Only patients with evidence of knee OA and use of HA injections (the index medication of interest) were selected, regardless of whether they received TKR during the post-index period. The requirement for a knee OA diagnosis during the 12-month pre-index period resulted in the significant attrition of patients, with 584,956 patients being excluded. Among the 50,389 patients who received HA for treatment of knee OA, 36,260 (72.0%) received a single course of treatment, 8709 (17.3%) received 2 courses, 3179 (6.3%) received 3 courses, 1354 (2.7%) received 4 courses, and 887 (1.8%) received ≥5 courses of treatment.

Comparison of baseline characteristics among the 5 IA HA cohorts showed the fairly similar baseline characteristics of all cohorts (Table 1). Geographic region, physician specialty, and opioid use showed differences among the cohorts. Cohorts with ≥5 HA courses presented lower proportions of patients from Southern US states, patients seeing orthopedic surgeons, and patients using opioids than cohorts with fewer HA courses.

PROCEDURES OF INTEREST

An analysis of the procedures patients received after HA treatment initiation showed that higher numbers of HA treatment courses resulted in lower proportions of patients receiving TKR within 3 years after HA treatment initiation (Table 2). With an increasing number of HA treatment courses, the proportion of patients with TKR within 3 years post-index consistently decreased from 28.4% (for 1 HA course) to 5.0% (for ≥5 HA courses), with all differences being highly statistically significant (P < .0001). Similarly, partial knee replacement exhibited a similar trend, with the proportion of patients decreasing from 3.3% (for 1 HA course) to 0.8% (for ≥5 HA courses; P < .0001). Among the patients with TKR within 3 years post-index, increasing numbers of treatment courses correlated with increasing time to TKR, with a mean of 375.6 days (for 1 HA course) rising to a mean of 971.5 days (for ≥5 HA courses; P < .0001). On the other hand, patients with multiple courses of HA treatment were more likely to undergo radiologic examinations of the knee, arthrocenteses, and image-guided injections than patients with only a single course of HA treatment (P < .0001).

ADVERSE EVENTS

Arthralgia and joint pain in the knee were the most commonly recorded adverse events (Table 3). More courses of HA treatment were associated with higher rates of adverse events. Overall, the reported adverse events profile of repeated courses of HA treatment consisted of mostly common and mild adverse events and displayed no safety concern for patients with knee OA that was followed-up for 3 years. The causality of these adverse events directly related to HA injections vs a specific disease state cannot be determined from an administrative claims data set.

TIME TO TKR

Successive courses of HA led to high proportions of patients without TKR 3 years after HA treatment initiation. This result is evident in the Kaplan-Meier survival curves of time to TKR for different HA cohorts (Figure 2), with log-rank tests of multiple courses vs a single course of HA (P < .0001) showing highly statistically significance. Tabulation of proportions of patients without TKR by various time points showed that increasing numbers of HA treatment courses correlated with higher proportions of patients without TKR at almost all time points (Table 4); within 3 years post-index, 71.6% of patients in the 1 HA course cohort exhibited no TKR, whereas 95.0% of patients in ≥5 HA courses cohort presented no TKR. We also performed a multivariate Cox PHM (Table 5) to account for baseline characteristics of different HA cohorts with covariates when estimating the risks of receiving TKR. The results of the Cox PHM showed that multiple courses of HA treatment significantly decreased the risk of TKR (hazard ratio, 0.138 for ≥5 HA courses vs 1 HA course; P < .0001). Inspection of other highly significant covariates showed that being older, living in the Midwest region of the US (vs the Northeast), receiving pre-index corticosteroids, having an orthopedic surgeon as a treating physician (vs a general practitioner, a rheumatologist, or a physical medicine and rehabilitation specialist), experiencing hypertension or hyperlipidemia, and higher pre-index total healthcare costs were associated with an increased risk of TKR (all P < .0001). Vascular disease and high CCI scores were associated with a decreased risk of TKR (P < .0001).

Continue to: Discussion...

DISCUSSION

This study demonstrated that multiple courses of HA treatment can delay the need for surgery for up to 3 years, with risk for both TKR and partial knee replacement decreasing in a dose-dependent manner. The potentially confounding effect of differences in baseline characteristics that could influence patients’ propensity to receive TKR in a database study was controlled by performing a multivariate analysis with covariate adjustment. The TKR-delaying effect of HA injection was more prominent in cohorts with a high number of HA treatment courses: 19 out of 20 patients in the cohort of ≥5 HA courses were free of TKR at the end of the 3-year post-index period. Such a high proportion of patients avoiding TKR with repeated courses of HA suggests that some patients may be able to successfully delay TKR well beyond the 3-year time span. This finding is counter-evidence to the frequently made assumption¹⁵ that all patients with knee OA will eventually progress to a state of disability, making TKR inevitable. The patients with end-stage radiographic knee OA can also benefit from IA HA injections for an extended period of time;¹⁶ the latest evidence indicates that nonoperative management can improve symptoms irrespective of radiographic disease severity, implying that TKR needs not to be the only therapeutic option for patients with end-stage radiographic knee OA.¹⁷ This finding suggests that HA treatment should be considered an important clinical treatment option for patients with knee OA.

Although the incidence rates of certain adverse events, such as arthralgia/joint pain, are sizable, these temporary adverse events commonly occur among patients who receive IA injections for knee OA; most of these events may simply include symptoms of the remaining underlying knee OA. These results are consistent with those of previous literature reporting the safety of repeated treatment with IA HA injections in a prospective clinical trial¹⁸ and demonstrating that repeated courses of HA treatment pose no greater safety risk than a single course of HA treatment.

Multivariate modeling outcomes of factors influencing risk of receiving TKR are broadly consistent with the generally accepted notions that different levels of disease severity and patients’ willingness to consider TKR at baseline influence the likelihood and timing of receiving TKR.^19,20 Age and obesity are common risk factors for progression of OA. Orthopedic surgeons are more likely to recommend surgery than non-surgeons. The pre-index use of corticosteroids and high pre-index healthcare costs could be associated with more severe symptoms at baseline. Patients with vascular disease or severe comorbidities, as evidenced by high CCI scores, make poor candidates for major elective surgeries such as TKR. These results are intuitive and validate the clinical insights of this study. Moreover, inclusion of these covariates in the analysis model allows for indirect adjustment of the most important prognostic factors for TKR at baseline, permitting proper statistical comparison of the results for different cohort groups.

Recently, the efficacy of HA injections for OA patients has become the subject of debate when the American Academy of Orthopaedic Surgeons (AAOS) revised its clinical practice guideline, recommending against the use of HA.²¹ The AAOS’ findings differ from those of other clinical societies, such as the American College of Rheumatology²² and the European League Against Rheumatism,²³ which provide no strong recommendation against the use of HA injections. The announcement of the new guideline by AAOS caused concern among clinicians and payers who had valued IA HA injections as a means to control knee OA pain before patients progress to TKR;²⁴ on the other hand, the demand for nonoperative treatment of knee OA remains high. Utilization rates of TKR have increased dramatically, and surgeries are now performed on younger patients with increasing burden on the healthcare system,^25,26 in spite of the fact that as high as a third of TKR surgeries may have been performed in inappropriate patients.²⁷ Part of the confusion surrounding clinical utility of HA stems from the fact that up until recently, relatively little research looked into the practical benefits of HA in actual clinical practice. Analyses of databases such as registries are now gaining attention to overcome that problem. Examination of large administrative databases maintained by commercial payers offers the benefit of probing realistically the safety and efficacy of treatments in actual clinical environments in a very large number of patients with heterogeneous backgrounds. Recently, the Agency for Healthcare Research and Quality’s Technology Assessment Program in the US called for such studies to determine whether HA injections can delay progression to TKR.²⁸ The results of this study and several others^11,13,14,16 suggest that use of HA to treat OA of the knee is associated with the delay of TKR, supporting the utility of HA in clinical practice and the healthcare system. Potential clinical benefits of delaying TKR may include the reduced risk of aseptic loosening if younger patients can wait for TKR or more time to allow the modification of risk factors in patients who will ultimately undergo TKR.

LIMITATIONS

Follow-up period was limited to 3 years post-index date because longer follow-up data were not available at the time of the study design. If an incorrect adverse event or OA diagnosis was listed in the medical record, or if the medical record was incomplete, then patients might have been misclassified, resulting in selection bias. The claims dataset includes no uninsured and Medicare patients, as the population in the database consisted primarily of commercially-insured patients in the US. Therefore, the results are most generalizable to other commercially-insured patients in the US. Generalizability to other populations may not be assured if they differ in their accessibility to physician services or prescriptions from the patients in this study. Other treatments such as the nonsteroidal anti-inflammatory drugs used by patients were not included within the pre-specified statistical model because their potential effects were assumed to be short-lived and much less than those of corticosteroid. Including these treatments would overload the statistical model with too many covariates, leading to potential computational instability. The database used provides no information on systemic factors, including plan limits on medication use, that could affect care. Given the large and diverse nature of the healthcare plans in the database. However, these factors should not have materially affected our study results. The claims database also lacks direct indicators of OA disease severity, such as Kellgren-Lawrence scores or patient-reported outcomes, including pain and function questionnaire scores. Our multivariate analysis indirectly makes up for this deficiency by considering other baseline characteristics or clinical indicators that may be correlated with information unavailable in a claims database. Patients who opt to undergo repeated courses of HA treatment may be more inclined to avoid surgery or may naturally experience OA disease progression more slowly, making them potentially different from patients who select to undergo surgery earlier without repeated courses of HA treatment. This condition may introduce a bias that causes difficulty in proving the causality between repeated HA use and delay of TKR.

CONCLUSION

Analysis of the knee OA patient data from a real-world database showed that repeated courses of treatment with HA are safe and are associated with the delay of TKR for up to 3 years. Additional research is needed to evaluate the effects of repeated HA courses on delaying TKR beyond a 3-year period.

ABSTRACT

Osteoarthritis (OA) of the knee is a top cause of disability among the elderly. Total knee replacement (TKR) has been available as an effective and definite surgical method to treat severe OA of the knee. However, TKR is a significant procedure with potential risk for serious complications and high costs. Alternative lower risk therapies that can delay or obviate TKR are valuable to those who are poor candidates for surgery or wish to avoid TKR as long as possible. Given the chondroprotective effects of hyaluronic acid (HA) injections, they are a safe and effective treatment to improve pain, function, and longevity of the knee. Thus, HA features the potential to delay or obviate TKR.

We aim to study the safety and effectiveness of repeated courses of HA on the time to TKR over a 3-year period using data from a large US health plan administrative claims database.

Retrospective analyses were conducted by identifying knee OA patients during the selection period (2007-2010). The follow-up period was 36 months, post-index date of initial HA injection. Procedural outcomes and adverse events of interest were tabulated and analyzed. A Cox proportional hazards model was used to model the risk of TKR.

A total of 50,389 patients who received HA for treatment of knee OA and met the study inclusion criteria were analyzed. Successive courses of HA showed a good safety profile and led to high proportions of patients without TKR 3 years after treatment initiation. Multivariate statistical modeling showed that multiple courses of HA injections significantly decreased the rates of TKR (95.0% without TKR for ≥5 courses vs 71.6% without TKR for 1 course; hazard ratio, 0.138; P < .0001).

Repeated courses of treatment with HA are safe and are associated with the delay of TKR for up to 3 years. Additional research is needed to evaluate the effect of repeated HA courses on delaying TKR beyond a 3-year time horizon.

Continue to: Osteoarthritis (OA) of the knee...

Osteoarthritis (OA) of the knee has emerged as one of the main causes of disability in the United States. Although no currently known cure of OA can reverse the progression of the disease, total knee replacement (TKR) is an effective and definitive treatment. However, TKR is an invasive procedure with potential risk for serious complications, and it has imposed high costs on the US healthcare system, with expenses accounting for hospital expenditures of TKR estimated at $28.5 billion in 2009.¹Alternative low-risk therapies that can delay or obviate TKR are valuable to a number of patients, especially the poor candidates for surgery or those who wish to avoid TKR.

Intra-articular (IA) hyaluronic acid (HA) injections have been available as a safe and effective treatment option to alleviate pain and to improve joint functions.² Results of randomized double-blind controlled clinical trials have demonstrated the pain-relieving effect of IA HA injections.^3-5 Furthermore, a recent network meta-analysis comparing various pharmacologic interventions for knee OA has confirmed the efficacy of IA HA injections, which outperformed other interventions when compared with oral placebos.^6,7 IA therapies are more effective than oral therapies for knee OA pain, with IA HA injections demonstrating the most pain reduction, potentially due to the benefit associated with needle injection and aspiration. Recent experimental studies have also suggested that IA HA may provide cartilage protection, reduce inflammation, and boost the viscosity of synovial fluid;⁸ IA HA may also exert therapeutic effects by inhibiting bone formation in OA patients.^9,10 HA possesses the potential to delay or obviate TKR. Previous research with a case series review of patients in an orthopedic specialty practice reported that the use of IA HA injections in patients with grade IV OA delayed TKR substantially.¹¹ One study analyzed retrospective medical claims data from a single private insurer and discovered potential evidence for the modest benefit of IA HA injections in delaying TKR.¹²

More detailed research work on a large sample of patients with knee OA and the requirement of TKR as a condition for inclusion using US administrative claims data has demonstrated the TKR-delaying effects of IA HA injections in comparison with a control group without claims for IA HA injections.^13,14 This study also uses real-world US administrative data but utilizes a different approach by starting with a sample of patients with knee OA and evidence of IA HA injections and then assessing the effect of repeated courses of HA treatment on the delay of TKR, without TKR as a mandatory condition for inclusion. All patients with knee OA within the time window were included, regardless of the need for TKR compared with previous studies which only considered patients who ultimately received TKR. Safety information and effectiveness information were examined to achieve a balanced risk-benefit assessment. We also analyzed how multiple courses of HA treatment and other potentially relevant covariates at baseline affected the risk of receiving TKR in a multivariate survival model. We aimed to achieve a realistic assessment of the clinical utility of HA injections in delaying TKR in a real-world setting using both safety and effectiveness data.

METHODS

DATA SOURCE

A retrospective cohort observational study using IMS Health’s PharMetrics Plus Health Plan Claims Database was conducted by identifying knee OA patients with claims indicating initiation of HA injection at an index date during the selection period (July 1, 2007 to June 30, 2010). All common HA agents in the US market during this period (Euflexxa, Hyalgan, Orthovisc, Supartz, and Synvisc) were selected via the corresponding J-codes and pooled for investigation of HA class effects. The follow-up period was 36 months, post-index date of the initial HA injection. Outcomes were measured, and adverse events were identified during this period. The time window for identification of adverse events was within 2 weeks from any injection during the course of therapy (evidence of an emergency room visit and/or physician office visit with requisite code). The data during the 12-month pre-index baseline period from the claims database was used to obtain information about baseline patient characteristics, such as age, gender, type of coverage, physician specialty, Charlson Comorbidity Index (CCI), major comorbidities, and major medications of interest commonly used among patients with knee OA.

STUDY SAMPLE SELECTION

The eligible patients required an outpatient claim indicating the initiation of HA injection. The date of the first claim for the patient within the selection window was defined as their index date. Patients had to be ≥18 years of age in the year of their index date. They had to present at least 1 clinical knee OA diagnosis at any point in the 12-month pre-index period (including the index date), and only patients who were continuously enrolled from 12 months pre-index to 36 months post-index date were evaluated. Among these patients (approximately 1.4 million), the following were excluded to minimize complications in data analysis and interpretation: patients with evidence of any HA use in the pre-index period; patients with evidence of a different kind of HA index medication in the post-index period; patients with evidence of TKR within 30 days of the index event during the post-index period; patients with evidence of 2 different kinds of HA index medications on the index date; and patients with evidence of diagnosis of hip OA, fibromyalgia, rheumatoid arthritis, lupus, or gout during the pre-index period.

Five patient cohorts were defined according to the number of courses of IA HA injections over the entire post-index period.

Continue to: Statistical analysis...

STATISTICAL ANALYSIS

All statistical analyses were performed using SAS version 9.2 (SAS Institute Inc.). Descriptive statistics such as means, standard deviations, medians, and 25% and 75% percentiles (Q1 and Q3, respectively) were provided for the continuous variables. Numbers and percentages were provided for the categorical variables. For statistical testing, Student’s t-tests were applied for the continuous variables and chi-square tests for the categorical variables. All the statistical tests were two-tailed. The sample sizes in this database study are remarkably large, such that differences that are not clinically important could still be statistically significant at the conventional alpha level of 0.05. Thus, we applied a more stringent requirement of the alpha level of 0.0001 to identify highly statistically significant results. The number and percentage of patients within each cohort with at least 1 instance of an adverse event of interest (those adverse events commonly expected for patients who receive IA injections for knee OA) were assessed. Times to TKR during the 36-month post-index period were analyzed and compared among different cohorts. Any patients who had not undergone TKR by the end of the post-index period were considered censored at 36 months. The Kaplan-Meier method was employed to model survival curves with time to TKR data, and log-rank tests were used to compare survival curves among different cohorts. A Cox proportional hazards model (PHM) was used to model the risk of TKR with a pre-specified set of covariates adjusted for baseline attributes, such as age, gender, comorbidities, and pre-index healthcare costs. Hazard ratios with 95% confidence intervals were used to examine the measures of event risk.

RESULTS

PATIENT CHARACTERISTICS

Applying study selection criteria to the claims database yielded 50,389 patients (Figure 1), providing an ample sample size for the statistical analysis. Only patients with evidence of knee OA and use of HA injections (the index medication of interest) were selected, regardless of whether they received TKR during the post-index period. The requirement for a knee OA diagnosis during the 12-month pre-index period resulted in the significant attrition of patients, with 584,956 patients being excluded. Among the 50,389 patients who received HA for treatment of knee OA, 36,260 (72.0%) received a single course of treatment, 8709 (17.3%) received 2 courses, 3179 (6.3%) received 3 courses, 1354 (2.7%) received 4 courses, and 887 (1.8%) received ≥5 courses of treatment.

Comparison of baseline characteristics among the 5 IA HA cohorts showed the fairly similar baseline characteristics of all cohorts (Table 1). Geographic region, physician specialty, and opioid use showed differences among the cohorts. Cohorts with ≥5 HA courses presented lower proportions of patients from Southern US states, patients seeing orthopedic surgeons, and patients using opioids than cohorts with fewer HA courses.

PROCEDURES OF INTEREST

An analysis of the procedures patients received after HA treatment initiation showed that higher numbers of HA treatment courses resulted in lower proportions of patients receiving TKR within 3 years after HA treatment initiation (Table 2). With an increasing number of HA treatment courses, the proportion of patients with TKR within 3 years post-index consistently decreased from 28.4% (for 1 HA course) to 5.0% (for ≥5 HA courses), with all differences being highly statistically significant (P < .0001). Similarly, partial knee replacement exhibited a similar trend, with the proportion of patients decreasing from 3.3% (for 1 HA course) to 0.8% (for ≥5 HA courses; P < .0001). Among the patients with TKR within 3 years post-index, increasing numbers of treatment courses correlated with increasing time to TKR, with a mean of 375.6 days (for 1 HA course) rising to a mean of 971.5 days (for ≥5 HA courses; P < .0001). On the other hand, patients with multiple courses of HA treatment were more likely to undergo radiologic examinations of the knee, arthrocenteses, and image-guided injections than patients with only a single course of HA treatment (P < .0001).

ADVERSE EVENTS

Arthralgia and joint pain in the knee were the most commonly recorded adverse events (Table 3). More courses of HA treatment were associated with higher rates of adverse events. Overall, the reported adverse events profile of repeated courses of HA treatment consisted of mostly common and mild adverse events and displayed no safety concern for patients with knee OA that was followed-up for 3 years. The causality of these adverse events directly related to HA injections vs a specific disease state cannot be determined from an administrative claims data set.

TIME TO TKR

Successive courses of HA led to high proportions of patients without TKR 3 years after HA treatment initiation. This result is evident in the Kaplan-Meier survival curves of time to TKR for different HA cohorts (Figure 2), with log-rank tests of multiple courses vs a single course of HA (P < .0001) showing highly statistically significance. Tabulation of proportions of patients without TKR by various time points showed that increasing numbers of HA treatment courses correlated with higher proportions of patients without TKR at almost all time points (Table 4); within 3 years post-index, 71.6% of patients in the 1 HA course cohort exhibited no TKR, whereas 95.0% of patients in ≥5 HA courses cohort presented no TKR. We also performed a multivariate Cox PHM (Table 5) to account for baseline characteristics of different HA cohorts with covariates when estimating the risks of receiving TKR. The results of the Cox PHM showed that multiple courses of HA treatment significantly decreased the risk of TKR (hazard ratio, 0.138 for ≥5 HA courses vs 1 HA course; P < .0001). Inspection of other highly significant covariates showed that being older, living in the Midwest region of the US (vs the Northeast), receiving pre-index corticosteroids, having an orthopedic surgeon as a treating physician (vs a general practitioner, a rheumatologist, or a physical medicine and rehabilitation specialist), experiencing hypertension or hyperlipidemia, and higher pre-index total healthcare costs were associated with an increased risk of TKR (all P < .0001). Vascular disease and high CCI scores were associated with a decreased risk of TKR (P < .0001).

Continue to: Discussion...

DISCUSSION

This study demonstrated that multiple courses of HA treatment can delay the need for surgery for up to 3 years, with risk for both TKR and partial knee replacement decreasing in a dose-dependent manner. The potentially confounding effect of differences in baseline characteristics that could influence patients’ propensity to receive TKR in a database study was controlled by performing a multivariate analysis with covariate adjustment. The TKR-delaying effect of HA injection was more prominent in cohorts with a high number of HA treatment courses: 19 out of 20 patients in the cohort of ≥5 HA courses were free of TKR at the end of the 3-year post-index period. Such a high proportion of patients avoiding TKR with repeated courses of HA suggests that some patients may be able to successfully delay TKR well beyond the 3-year time span. This finding is counter-evidence to the frequently made assumption¹⁵ that all patients with knee OA will eventually progress to a state of disability, making TKR inevitable. The patients with end-stage radiographic knee OA can also benefit from IA HA injections for an extended period of time;¹⁶ the latest evidence indicates that nonoperative management can improve symptoms irrespective of radiographic disease severity, implying that TKR needs not to be the only therapeutic option for patients with end-stage radiographic knee OA.¹⁷ This finding suggests that HA treatment should be considered an important clinical treatment option for patients with knee OA.

Although the incidence rates of certain adverse events, such as arthralgia/joint pain, are sizable, these temporary adverse events commonly occur among patients who receive IA injections for knee OA; most of these events may simply include symptoms of the remaining underlying knee OA. These results are consistent with those of previous literature reporting the safety of repeated treatment with IA HA injections in a prospective clinical trial¹⁸ and demonstrating that repeated courses of HA treatment pose no greater safety risk than a single course of HA treatment.

Multivariate modeling outcomes of factors influencing risk of receiving TKR are broadly consistent with the generally accepted notions that different levels of disease severity and patients’ willingness to consider TKR at baseline influence the likelihood and timing of receiving TKR.^19,20 Age and obesity are common risk factors for progression of OA. Orthopedic surgeons are more likely to recommend surgery than non-surgeons. The pre-index use of corticosteroids and high pre-index healthcare costs could be associated with more severe symptoms at baseline. Patients with vascular disease or severe comorbidities, as evidenced by high CCI scores, make poor candidates for major elective surgeries such as TKR. These results are intuitive and validate the clinical insights of this study. Moreover, inclusion of these covariates in the analysis model allows for indirect adjustment of the most important prognostic factors for TKR at baseline, permitting proper statistical comparison of the results for different cohort groups.

Recently, the efficacy of HA injections for OA patients has become the subject of debate when the American Academy of Orthopaedic Surgeons (AAOS) revised its clinical practice guideline, recommending against the use of HA.²¹ The AAOS’ findings differ from those of other clinical societies, such as the American College of Rheumatology²² and the European League Against Rheumatism,²³ which provide no strong recommendation against the use of HA injections. The announcement of the new guideline by AAOS caused concern among clinicians and payers who had valued IA HA injections as a means to control knee OA pain before patients progress to TKR;²⁴ on the other hand, the demand for nonoperative treatment of knee OA remains high. Utilization rates of TKR have increased dramatically, and surgeries are now performed on younger patients with increasing burden on the healthcare system,^25,26 in spite of the fact that as high as a third of TKR surgeries may have been performed in inappropriate patients.²⁷ Part of the confusion surrounding clinical utility of HA stems from the fact that up until recently, relatively little research looked into the practical benefits of HA in actual clinical practice. Analyses of databases such as registries are now gaining attention to overcome that problem. Examination of large administrative databases maintained by commercial payers offers the benefit of probing realistically the safety and efficacy of treatments in actual clinical environments in a very large number of patients with heterogeneous backgrounds. Recently, the Agency for Healthcare Research and Quality’s Technology Assessment Program in the US called for such studies to determine whether HA injections can delay progression to TKR.²⁸ The results of this study and several others^11,13,14,16 suggest that use of HA to treat OA of the knee is associated with the delay of TKR, supporting the utility of HA in clinical practice and the healthcare system. Potential clinical benefits of delaying TKR may include the reduced risk of aseptic loosening if younger patients can wait for TKR or more time to allow the modification of risk factors in patients who will ultimately undergo TKR.

LIMITATIONS

Follow-up period was limited to 3 years post-index date because longer follow-up data were not available at the time of the study design. If an incorrect adverse event or OA diagnosis was listed in the medical record, or if the medical record was incomplete, then patients might have been misclassified, resulting in selection bias. The claims dataset includes no uninsured and Medicare patients, as the population in the database consisted primarily of commercially-insured patients in the US. Therefore, the results are most generalizable to other commercially-insured patients in the US. Generalizability to other populations may not be assured if they differ in their accessibility to physician services or prescriptions from the patients in this study. Other treatments such as the nonsteroidal anti-inflammatory drugs used by patients were not included within the pre-specified statistical model because their potential effects were assumed to be short-lived and much less than those of corticosteroid. Including these treatments would overload the statistical model with too many covariates, leading to potential computational instability. The database used provides no information on systemic factors, including plan limits on medication use, that could affect care. Given the large and diverse nature of the healthcare plans in the database. However, these factors should not have materially affected our study results. The claims database also lacks direct indicators of OA disease severity, such as Kellgren-Lawrence scores or patient-reported outcomes, including pain and function questionnaire scores. Our multivariate analysis indirectly makes up for this deficiency by considering other baseline characteristics or clinical indicators that may be correlated with information unavailable in a claims database. Patients who opt to undergo repeated courses of HA treatment may be more inclined to avoid surgery or may naturally experience OA disease progression more slowly, making them potentially different from patients who select to undergo surgery earlier without repeated courses of HA treatment. This condition may introduce a bias that causes difficulty in proving the causality between repeated HA use and delay of TKR.

CONCLUSION

Analysis of the knee OA patient data from a real-world database showed that repeated courses of treatment with HA are safe and are associated with the delay of TKR for up to 3 years. Additional research is needed to evaluate the effects of repeated HA courses on delaying TKR beyond a 3-year period.

ABSTRACT

Despite advancements in surgical technique and understanding of throwing mechanics, controversy persists regarding the treatment of grade III acromioclavicular (AC) joint separations, particularly in throwing athletes. Twenty-eight major league baseball (MLB) orthopedic team physicians were surveyed to determine their definitive management of a grade III AC separation in the dominant arm of a professional baseball pitcher and their experience treating AC joint separations in starting pitchers and position players. Return-to-play outcomes were also evaluated. Twenty (71.4%) team physicians recommended nonoperative intervention compared to 8 (28.6%) who would have operated acutely. Eighteen (64.3%) team physicians had treated at least 1 professional pitcher with a grade III AC separation; 51 (77.3%) pitchers had been treated nonoperatively compared to 15 (22.7%) operatively. No difference was observed in the proportion of pitchers who returned to the same level of play (P = .54), had full, unrestricted range of motion (P = .23), or had full pain relief (P = .19) between the operatively and nonoperatively treated MLB pitchers. The majority (53.6%) of physicians would not include an injection if the injury was treated nonoperatively. Open coracoclavicular reconstruction (65.2%) was preferred for operative cases; 66.7% of surgeons would also include distal clavicle excision as an adjunct procedure. About 90% of physicians would return pitchers to throwing >12 weeks after surgery compared to after 4 to 6 weeks in nonoperatively treated cases. In conclusion, MLB team physicians preferred nonoperative management for an acute grade III AC joint separation in professional pitchers. If operative intervention is required, ligament reconstruction with adjunct distal clavicle excision were the most commonly performed procedures.

Continue to: Despite advancements in surgucal technique...

Despite advancements in surgical technique and improved understanding of the physiology of throwing mechanics, controversy persists regarding the preferred treatment for grade III acromioclavicular (AC) joint separations.^1-6 Nonsurgical management has demonstrated return to prior function with fewer complications.⁷ However, there is a growing body of evidence demonstrating that surgical intervention is associated with more favorable outcomes⁸ and should be considered in patients who place high functional demands on their shoulders.⁹

The reported results on professional athletes in the literature remain ambivalent. Multiple small case reports/series have reported successful nonoperative treatment of elite athletes.^10-12 Not surprisingly, McFarland and colleagues¹³ reported in 1997 that 69% of major league baseball (MLB) team physicians preferred nonoperative treatment for a theoretical starting pitcher sustaining a grade III AC separation 1 week prior to the start of the season. In contrast, reports of an inability to throw at a pre-injury level are equally commonplace.^14,15 Nevertheless, all of these studies were limited to small cohorts, as the incidence of grade III AC separations in elite throwing athletes is relatively uncommon.^13,16

In this study, we re-evaluated the study performed by McFarland and colleagues¹³ in 1997 by surveying all active MLB team orthopedic surgeons. We asked them how they would treat a grade III AC separation in a starting professional baseball pitcher. The physicians were also asked about their personal experience evaluating outcomes in these elite athletes. Given our improved understanding of the anatomy, pathophysiology, and surgical techniques for treating grade III AC separations, we hypothesize that more MLB team physicians would favor operative intervention treatment in professional baseball pitchers, as their vocation places higher demands on their shoulders.

MATERIALS AND METHODS

A questionnaire (Appendix A) was distributed to the team physicians of all 30 MLB teams. In addition to surgeon demographics, including age, years in practice, and years of taking care of an MLB team, the initial section of the questionnaire asked orthopedic surgeons how they would treat a theoretical starting pitcher who sustained a grade III AC joint separation of the dominant throwing arm 1 week prior to the start of the season. Physicians who preferred nonoperative treatment were asked whether they would use an injection (and what type), as well as when they would allow the pitcher to start a progressive interval throwing program. Physicians who preferred operative treatment were asked to rank their indications for operating, what procedure they would use (eg, open vs arthroscopic or coracoclavicular ligament repair vs reconstruction), and whether the surgical intervention would include distal clavicle excision. Both groups of physicians were also asked if their preferred treatment would change if the injury were to occur at the end of the season.

The second portion of the questionnaire asked surgeons about their experience treating AC joint separations in both starting pitchers and position players, as well as to describe the long-term outcomes of their preferred treatment, including time to return to full clearance for pitching, whether their patients returned to their prior level of play, and whether these patients had full pain relief. Finally, physicians were asked if any of the nonoperatively treated players ultimately crossed over and required operative intervention.

Continue to: Statistics...

STATISTICS

Descriptive statistics were used for continuous variables, and frequencies were used for categorical variables. Linear regression was performed to determine the correlation between the physician’s training or experience in treating AC joint separations and their recommended treatment. Fischer’s exact test/chi-square analysis was used to compare categorical variables. All tests were conducted using 2-sided hypothesis testing with statistical significance set at P < .05. All statistical analyses were conducted with SPSS 21.0 software (IBM Corporation).

RESULTS

A total of 28 MLB team physicians completed the questionnaires from 18 of the 30 MLB teams. The average age of the responders was 50.5 years (range, 34-60 years), with an average of 18.2 years in practice (range, 2-30 years) and 10.8 years (range, 1-24 years) taking care of their current professional baseball team. About 82% of the team physicians completed a sports medicine fellowship. On average, physicians saw 16.6 (range, 5-50) grade III or higher AC joint separations per year, and operated on 4.6 (range, 0-10) per year.

Nonoperative treatment was the preferred treatment for a grade III AC joint separation in a starting professional baseball pitcher for the majority of team physicians (20/28). No correlation was observed between the physician’s age (P = .881), years in practice (P = .915), years taking care of their professional team (P = .989), percentage of practice focused on shoulders (P = .986), number of AC joint injuries seen (P = .325), or number of surgeries performed per year (P = .807) with the team physician’s preferred treatment. Compared to the proportion reported originally by McFarland and colleagues¹³ in 1997 (69%), there was no difference in the proportion of team physicians that recommended nonoperative treatment (P = 1).

If treating this injury nonoperatively, 46.4% of physicians would also use an injection, with orthobiologics (eg, platelet-rich plasma) as the most popular choice (Table 1). No consensus was provided on the timeframe to return pitchers back to a progressive interval throwing program; however, 46.67% of physicians would return pitchers 4 to 6 weeks after a nonoperatively treated injury, while 35.7% would return pitchers 7 to 12 weeks after the initial injury.

Table 1. Treatment Preferences of Grade III AC Separation by MLB Team Physicians

^aRespondents were allowed to choose more than 1 treatment in each category. ^bChosen as an adjunct treatment.

Abbreviations: AC, acromioclavicular; MLB, major league baseball.

Most physicians (64.3%) cited functional limitations as the most important reason for indicating operative treatment, followed by pain (21.4%), and a deformity (14.3%). About 65% preferred open coracoclavicular ligament reconstruction. No physician recommended the Weaver-Dunn procedure or use of hardware (eg, hook plate). Of those who preferred an operative intervention, 66.7% would also include a distal clavicle excision, which is significantly higher than the proportion reported by McFarland and colleagues¹³ (23%, P = .0170). About 90% of physicians would return pitchers to play >12 weeks after operative treatment.

Continue to: If the injury occurred at the end ...

If the injury occurred at the end of the season, 7 of the 20 orthopedists (35%) who recommended nonoperative treatment said they would change to an operative intervention. Eighteen of 28 responders would have the same algorithm for MLB position players. Team physicians were less likely to recommend operative intervention in position players due to less demand on the arm and increased ability to accommodate the injury by altering their throwing mechanics.

Eighteen (64%) of the team physicians had treated at least 1 professional pitcher with a grade III AC separation in his dominant arm, and 11 (39.3%) had treated >1. Collectively, team physicians had treated 15 professional pitchers operatively, and 51 nonoperatively; only 3 patients converted to operative intervention after a failed nonoperative treatment.

Of the pitchers treated operatively, 93.3% (14) of pitchers returned to their prior level of pitching. The 1 patient who failed to return to the same level of pitching retired instead of returning to play. About 80% (12) of the pitchers had full pain relief, and 93.3% (14) had full range of motion (ROM). The pitcher who failed to regain full ROM also had a concomitant rotator cuff repair. The only complication reported from an operative intervention was a pitcher who sustained a coracoid fracture 10 months postoperatively while throwing 100 mph. Of the pitchers treated nonoperatively, 96% returned to their prior level of pitching, 92.2% (47) had full complete pain relief when throwing, and 100% had full ROM. No differences were observed between the proportion of pitchers who returned to their prior level of pitching, regained full ROM, or had full pain relief in the operative and nonoperative groups (Table 2).

Table 2. Outcomes of Treatment of Grade III AC Separation in 58 Professional Baseball Players

Abbreviations: AC, acromioclavicular; ROM, range of motion.

DISCUSSION 

Controversy persists regarding the optimal management of acute grade III AC separations, with the current available evidence potentially suggesting better cosmetic and radiological results but no definite differences in clinical results.^1-6,17,18 In the absence of formal clinical practice guidelines, surgeons rely on their own experience or defer to the anecdotal experience of experts in the field. Our initial hypothesis was false in this survey of MLB team physicians taking care of overhead throwing athletes at the highest level. Our results demonstrate that despite improved techniques and an increased understanding of the pathophysiology of AC joint separations, conservative management is still the preferred treatment for acute grade III AC joint separations in professional baseball pitchers. The proportion of team physicians recommending nonoperative treatment in our series was essentially equivalent to the results reported by McFarland and colleagues¹³ in 1997, suggesting that the pendulum continues to favor conservative management initially. This status quo likely reflects both the dearth of literature suggesting a substantial benefit of acute operative repair, as well as the ability to accommodate with conservative measures after most grade III AC injuries, even at the highest level of athletic competition.

These results are also consistent with trends from the last few decades. In the 1970s, the overwhelming preference for treating an acute complete AC joint separation was surgical repair, with Powers and Bach¹⁰ reporting in a 1974 survey of 163 chairmen of orthopedic programs around the country that 91.5% advocated surgical treatment. However, surgical preference had reversed by the 1990s. Of the 187 chairmen and 59 team physicians surveyed by Cox¹⁹ in 1992, 72% and 86% respectively preferred nonoperative treatment in a theoretical 21-year-old athlete with a grade III AC separation. Nissen and Chatterjee²⁰ reported in 2007 on a survey of all American Orthopaedic Society for Sports Medicine surgeons (N = 577) and Accreditation Council for Graduate Medical Education orthopedic program residency directors (N = 87) that >80% of responders preferred conservative measures for this acute injury. The reversal of trends has also been corroborated by recent multicenter trials demonstrating no difference in clinical outcomes between operative and nonoperative treatment of high grade AC joint dislocations, albeit these patients were not all high level overhead throwing athletes.^17,18

Continue to: The trends in surgical interventions are notable...

The trends in surgical interventions are notable within the smaller subset of patients who are indicated for operative repair. Use of hardware and primary ligament repair, while popular in the surveys conducted in the 1970s¹⁰ and 1990s¹³ and even present in Nissen and Chatterjee’s²⁰ 2007 survey, were noticeably absent from our survey results, with the majority of respondents preferring open coracoclavicular ligament reconstruction. The role of distal clavicle excision has also expanded, from 23% of team physicians recommending it in 1997¹³ to 57% to 59% in Nissen and Chatterjee’s²⁰ 2007 survey, to 66.7% in our series. This trend is not surprising as several recent cadaveric biomechanical studies have demonstrated that not only do peak graft forces not increase significantly,²¹ the anterior-posterior and superior-inferior motion at the AC joint following ligament reconstruction is maintained despite resection of the lateral clavicle.²² Additionally, primary distal clavicle excision may prevent the development of post-traumatic arthritis at the AC joint and osteolysis of the distal clavicle as a possible pain generator in the future.²³ However, some respondents cautioned against performing a concomitant distal clavicle excision, as some biomechanical data demonstrate that resecting the distal clavicle may lead to increased horizontal translation at the AC joint despite intact superior and posterior AC capsules.²⁴ Professional baseball pitchers may also be more lax and thus prone to more instability. Primary repair or reconstruction may not always lead to complete pre-injury stability in these individuals. This subtle unrecognized instability is hard to diagnosis and may be a persistent source of pain; thus, adding a distal clavicle excision may actually exacerbate the instability.

The nuanced indications for operative intervention, such as the presence of associated lesions were not captured by our survey.²⁵ While most team physicians cited functional limitations as their most common reason for offering surgery, several MLB orthopedic surgeons also commented on evaluating the stability of the AC joint after a grade III injury, akin to the consensus statement from the International Society of Arthroscopy, Knee Surgery and Orthopaedic Sports Medicine (ISAKOS) Upper Extremity Committee²⁶ in 2014 that diversified the Rockwood Grade III AC joint separation into its IIIA and IIIB classifications. The ISAKOS recommendations include initial conservative management and a second evaluation (both clinical and radiographic) for grade III lesions 3 to 6 weeks after the injury. However, as professional baseball is an incredibly profitable sport with an annual revenue approaching $9.5 billion²⁷ and pitching salaries up to $32.5 million in 2015, serious financial considerations must be given to players who wish to avoid undergoing delayed surgery.

This study has shortcomings typical of expert opinion papers. The retrospective nature of this study places the data at risk of recall bias. Objective data (eg, terminal ROM, pain relief, and return to play) were obtained from a retrospective chart review; however, no standard documentation or collection method was used given the number of surgeons involved and, thus, conclusions based on treatment outcomes are imperfect. Another major weakness of this survey is the relatively small number of patients and respondents. An a priori power analysis was not available, as this was a retrospective review. A comparative trial will be necessary to definitively support one treatment over another. Assuming a 95% return to play in the nonoperatively treated group, approximately 300 patients would be needed in a prospective 2-armed study with 80% power to detect a 10% reduction in the incidence of return to play using an alpha level of 0.05 and assuming no loss to follow-up. This sample size would be difficult to achieve in this patient population.

However, compared to past series,¹³ the number of professional baseball players treated by the collective experience of these MLB team physicians is the largest reported to date. As suggested above, the rarity of this condition in elite athletes precludes the ability to have matched controls to definitively determine the optimal treatment, which may explain the lack of difference in the return to play, ROM, and pain relief outcomes. Instead, we can only extrapolate based on the collective anecdotal experience of the MLB team physicians.

CONCLUSION

Despite advances in surgical technique and understanding of throwing mechanics, the majority of MLB team physicians preferred nonoperative management for an acute grade III AC joint separation in a professional baseball pitcher. Open coracoclavicular ligament reconstruction was preferred for those who preferred operative intervention. An increasing number of orthopedic surgeons now consider a distal clavicle excision as an adjunct procedure.

This paper will be judged for the Resident Writer’s Award.

ABSTRACT

Despite advancements in surgical technique and understanding of throwing mechanics, controversy persists regarding the treatment of grade III acromioclavicular (AC) joint separations, particularly in throwing athletes. Twenty-eight major league baseball (MLB) orthopedic team physicians were surveyed to determine their definitive management of a grade III AC separation in the dominant arm of a professional baseball pitcher and their experience treating AC joint separations in starting pitchers and position players. Return-to-play outcomes were also evaluated. Twenty (71.4%) team physicians recommended nonoperative intervention compared to 8 (28.6%) who would have operated acutely. Eighteen (64.3%) team physicians had treated at least 1 professional pitcher with a grade III AC separation; 51 (77.3%) pitchers had been treated nonoperatively compared to 15 (22.7%) operatively. No difference was observed in the proportion of pitchers who returned to the same level of play (P = .54), had full, unrestricted range of motion (P = .23), or had full pain relief (P = .19) between the operatively and nonoperatively treated MLB pitchers. The majority (53.6%) of physicians would not include an injection if the injury was treated nonoperatively. Open coracoclavicular reconstruction (65.2%) was preferred for operative cases; 66.7% of surgeons would also include distal clavicle excision as an adjunct procedure. About 90% of physicians would return pitchers to throwing >12 weeks after surgery compared to after 4 to 6 weeks in nonoperatively treated cases. In conclusion, MLB team physicians preferred nonoperative management for an acute grade III AC joint separation in professional pitchers. If operative intervention is required, ligament reconstruction with adjunct distal clavicle excision were the most commonly performed procedures.

Continue to: Despite advancements in surgucal technique...

Despite advancements in surgical technique and improved understanding of the physiology of throwing mechanics, controversy persists regarding the preferred treatment for grade III acromioclavicular (AC) joint separations.^1-6 Nonsurgical management has demonstrated return to prior function with fewer complications.⁷ However, there is a growing body of evidence demonstrating that surgical intervention is associated with more favorable outcomes⁸ and should be considered in patients who place high functional demands on their shoulders.⁹

The reported results on professional athletes in the literature remain ambivalent. Multiple small case reports/series have reported successful nonoperative treatment of elite athletes.^10-12 Not surprisingly, McFarland and colleagues¹³ reported in 1997 that 69% of major league baseball (MLB) team physicians preferred nonoperative treatment for a theoretical starting pitcher sustaining a grade III AC separation 1 week prior to the start of the season. In contrast, reports of an inability to throw at a pre-injury level are equally commonplace.^14,15 Nevertheless, all of these studies were limited to small cohorts, as the incidence of grade III AC separations in elite throwing athletes is relatively uncommon.^13,16

In this study, we re-evaluated the study performed by McFarland and colleagues¹³ in 1997 by surveying all active MLB team orthopedic surgeons. We asked them how they would treat a grade III AC separation in a starting professional baseball pitcher. The physicians were also asked about their personal experience evaluating outcomes in these elite athletes. Given our improved understanding of the anatomy, pathophysiology, and surgical techniques for treating grade III AC separations, we hypothesize that more MLB team physicians would favor operative intervention treatment in professional baseball pitchers, as their vocation places higher demands on their shoulders.

MATERIALS AND METHODS

A questionnaire (Appendix A) was distributed to the team physicians of all 30 MLB teams. In addition to surgeon demographics, including age, years in practice, and years of taking care of an MLB team, the initial section of the questionnaire asked orthopedic surgeons how they would treat a theoretical starting pitcher who sustained a grade III AC joint separation of the dominant throwing arm 1 week prior to the start of the season. Physicians who preferred nonoperative treatment were asked whether they would use an injection (and what type), as well as when they would allow the pitcher to start a progressive interval throwing program. Physicians who preferred operative treatment were asked to rank their indications for operating, what procedure they would use (eg, open vs arthroscopic or coracoclavicular ligament repair vs reconstruction), and whether the surgical intervention would include distal clavicle excision. Both groups of physicians were also asked if their preferred treatment would change if the injury were to occur at the end of the season.

The second portion of the questionnaire asked surgeons about their experience treating AC joint separations in both starting pitchers and position players, as well as to describe the long-term outcomes of their preferred treatment, including time to return to full clearance for pitching, whether their patients returned to their prior level of play, and whether these patients had full pain relief. Finally, physicians were asked if any of the nonoperatively treated players ultimately crossed over and required operative intervention.

Continue to: Statistics...

STATISTICS

Descriptive statistics were used for continuous variables, and frequencies were used for categorical variables. Linear regression was performed to determine the correlation between the physician’s training or experience in treating AC joint separations and their recommended treatment. Fischer’s exact test/chi-square analysis was used to compare categorical variables. All tests were conducted using 2-sided hypothesis testing with statistical significance set at P < .05. All statistical analyses were conducted with SPSS 21.0 software (IBM Corporation).

RESULTS

A total of 28 MLB team physicians completed the questionnaires from 18 of the 30 MLB teams. The average age of the responders was 50.5 years (range, 34-60 years), with an average of 18.2 years in practice (range, 2-30 years) and 10.8 years (range, 1-24 years) taking care of their current professional baseball team. About 82% of the team physicians completed a sports medicine fellowship. On average, physicians saw 16.6 (range, 5-50) grade III or higher AC joint separations per year, and operated on 4.6 (range, 0-10) per year.

Nonoperative treatment was the preferred treatment for a grade III AC joint separation in a starting professional baseball pitcher for the majority of team physicians (20/28). No correlation was observed between the physician’s age (P = .881), years in practice (P = .915), years taking care of their professional team (P = .989), percentage of practice focused on shoulders (P = .986), number of AC joint injuries seen (P = .325), or number of surgeries performed per year (P = .807) with the team physician’s preferred treatment. Compared to the proportion reported originally by McFarland and colleagues¹³ in 1997 (69%), there was no difference in the proportion of team physicians that recommended nonoperative treatment (P = 1).

If treating this injury nonoperatively, 46.4% of physicians would also use an injection, with orthobiologics (eg, platelet-rich plasma) as the most popular choice (Table 1). No consensus was provided on the timeframe to return pitchers back to a progressive interval throwing program; however, 46.67% of physicians would return pitchers 4 to 6 weeks after a nonoperatively treated injury, while 35.7% would return pitchers 7 to 12 weeks after the initial injury.

Table 1. Treatment Preferences of Grade III AC Separation by MLB Team Physicians

^aRespondents were allowed to choose more than 1 treatment in each category. ^bChosen as an adjunct treatment.

Abbreviations: AC, acromioclavicular; MLB, major league baseball.

Most physicians (64.3%) cited functional limitations as the most important reason for indicating operative treatment, followed by pain (21.4%), and a deformity (14.3%). About 65% preferred open coracoclavicular ligament reconstruction. No physician recommended the Weaver-Dunn procedure or use of hardware (eg, hook plate). Of those who preferred an operative intervention, 66.7% would also include a distal clavicle excision, which is significantly higher than the proportion reported by McFarland and colleagues¹³ (23%, P = .0170). About 90% of physicians would return pitchers to play >12 weeks after operative treatment.

Continue to: If the injury occurred at the end ...

If the injury occurred at the end of the season, 7 of the 20 orthopedists (35%) who recommended nonoperative treatment said they would change to an operative intervention. Eighteen of 28 responders would have the same algorithm for MLB position players. Team physicians were less likely to recommend operative intervention in position players due to less demand on the arm and increased ability to accommodate the injury by altering their throwing mechanics.

Eighteen (64%) of the team physicians had treated at least 1 professional pitcher with a grade III AC separation in his dominant arm, and 11 (39.3%) had treated >1. Collectively, team physicians had treated 15 professional pitchers operatively, and 51 nonoperatively; only 3 patients converted to operative intervention after a failed nonoperative treatment.

Of the pitchers treated operatively, 93.3% (14) of pitchers returned to their prior level of pitching. The 1 patient who failed to return to the same level of pitching retired instead of returning to play. About 80% (12) of the pitchers had full pain relief, and 93.3% (14) had full range of motion (ROM). The pitcher who failed to regain full ROM also had a concomitant rotator cuff repair. The only complication reported from an operative intervention was a pitcher who sustained a coracoid fracture 10 months postoperatively while throwing 100 mph. Of the pitchers treated nonoperatively, 96% returned to their prior level of pitching, 92.2% (47) had full complete pain relief when throwing, and 100% had full ROM. No differences were observed between the proportion of pitchers who returned to their prior level of pitching, regained full ROM, or had full pain relief in the operative and nonoperative groups (Table 2).

Table 2. Outcomes of Treatment of Grade III AC Separation in 58 Professional Baseball Players

Abbreviations: AC, acromioclavicular; ROM, range of motion.

DISCUSSION 

Controversy persists regarding the optimal management of acute grade III AC separations, with the current available evidence potentially suggesting better cosmetic and radiological results but no definite differences in clinical results.^1-6,17,18 In the absence of formal clinical practice guidelines, surgeons rely on their own experience or defer to the anecdotal experience of experts in the field. Our initial hypothesis was false in this survey of MLB team physicians taking care of overhead throwing athletes at the highest level. Our results demonstrate that despite improved techniques and an increased understanding of the pathophysiology of AC joint separations, conservative management is still the preferred treatment for acute grade III AC joint separations in professional baseball pitchers. The proportion of team physicians recommending nonoperative treatment in our series was essentially equivalent to the results reported by McFarland and colleagues¹³ in 1997, suggesting that the pendulum continues to favor conservative management initially. This status quo likely reflects both the dearth of literature suggesting a substantial benefit of acute operative repair, as well as the ability to accommodate with conservative measures after most grade III AC injuries, even at the highest level of athletic competition.

These results are also consistent with trends from the last few decades. In the 1970s, the overwhelming preference for treating an acute complete AC joint separation was surgical repair, with Powers and Bach¹⁰ reporting in a 1974 survey of 163 chairmen of orthopedic programs around the country that 91.5% advocated surgical treatment. However, surgical preference had reversed by the 1990s. Of the 187 chairmen and 59 team physicians surveyed by Cox¹⁹ in 1992, 72% and 86% respectively preferred nonoperative treatment in a theoretical 21-year-old athlete with a grade III AC separation. Nissen and Chatterjee²⁰ reported in 2007 on a survey of all American Orthopaedic Society for Sports Medicine surgeons (N = 577) and Accreditation Council for Graduate Medical Education orthopedic program residency directors (N = 87) that >80% of responders preferred conservative measures for this acute injury. The reversal of trends has also been corroborated by recent multicenter trials demonstrating no difference in clinical outcomes between operative and nonoperative treatment of high grade AC joint dislocations, albeit these patients were not all high level overhead throwing athletes.^17,18

Continue to: The trends in surgical interventions are notable...

The trends in surgical interventions are notable within the smaller subset of patients who are indicated for operative repair. Use of hardware and primary ligament repair, while popular in the surveys conducted in the 1970s¹⁰ and 1990s¹³ and even present in Nissen and Chatterjee’s²⁰ 2007 survey, were noticeably absent from our survey results, with the majority of respondents preferring open coracoclavicular ligament reconstruction. The role of distal clavicle excision has also expanded, from 23% of team physicians recommending it in 1997¹³ to 57% to 59% in Nissen and Chatterjee’s²⁰ 2007 survey, to 66.7% in our series. This trend is not surprising as several recent cadaveric biomechanical studies have demonstrated that not only do peak graft forces not increase significantly,²¹ the anterior-posterior and superior-inferior motion at the AC joint following ligament reconstruction is maintained despite resection of the lateral clavicle.²² Additionally, primary distal clavicle excision may prevent the development of post-traumatic arthritis at the AC joint and osteolysis of the distal clavicle as a possible pain generator in the future.²³ However, some respondents cautioned against performing a concomitant distal clavicle excision, as some biomechanical data demonstrate that resecting the distal clavicle may lead to increased horizontal translation at the AC joint despite intact superior and posterior AC capsules.²⁴ Professional baseball pitchers may also be more lax and thus prone to more instability. Primary repair or reconstruction may not always lead to complete pre-injury stability in these individuals. This subtle unrecognized instability is hard to diagnosis and may be a persistent source of pain; thus, adding a distal clavicle excision may actually exacerbate the instability.

The nuanced indications for operative intervention, such as the presence of associated lesions were not captured by our survey.²⁵ While most team physicians cited functional limitations as their most common reason for offering surgery, several MLB orthopedic surgeons also commented on evaluating the stability of the AC joint after a grade III injury, akin to the consensus statement from the International Society of Arthroscopy, Knee Surgery and Orthopaedic Sports Medicine (ISAKOS) Upper Extremity Committee²⁶ in 2014 that diversified the Rockwood Grade III AC joint separation into its IIIA and IIIB classifications. The ISAKOS recommendations include initial conservative management and a second evaluation (both clinical and radiographic) for grade III lesions 3 to 6 weeks after the injury. However, as professional baseball is an incredibly profitable sport with an annual revenue approaching $9.5 billion²⁷ and pitching salaries up to $32.5 million in 2015, serious financial considerations must be given to players who wish to avoid undergoing delayed surgery.

This study has shortcomings typical of expert opinion papers. The retrospective nature of this study places the data at risk of recall bias. Objective data (eg, terminal ROM, pain relief, and return to play) were obtained from a retrospective chart review; however, no standard documentation or collection method was used given the number of surgeons involved and, thus, conclusions based on treatment outcomes are imperfect. Another major weakness of this survey is the relatively small number of patients and respondents. An a priori power analysis was not available, as this was a retrospective review. A comparative trial will be necessary to definitively support one treatment over another. Assuming a 95% return to play in the nonoperatively treated group, approximately 300 patients would be needed in a prospective 2-armed study with 80% power to detect a 10% reduction in the incidence of return to play using an alpha level of 0.05 and assuming no loss to follow-up. This sample size would be difficult to achieve in this patient population.

However, compared to past series,¹³ the number of professional baseball players treated by the collective experience of these MLB team physicians is the largest reported to date. As suggested above, the rarity of this condition in elite athletes precludes the ability to have matched controls to definitively determine the optimal treatment, which may explain the lack of difference in the return to play, ROM, and pain relief outcomes. Instead, we can only extrapolate based on the collective anecdotal experience of the MLB team physicians.

CONCLUSION

Despite advances in surgical technique and understanding of throwing mechanics, the majority of MLB team physicians preferred nonoperative management for an acute grade III AC joint separation in a professional baseball pitcher. Open coracoclavicular ligament reconstruction was preferred for those who preferred operative intervention. An increasing number of orthopedic surgeons now consider a distal clavicle excision as an adjunct procedure.

This paper will be judged for the Resident Writer’s Award.

ABSTRACT

Despite advancements in surgical technique and understanding of throwing mechanics, controversy persists regarding the treatment of grade III acromioclavicular (AC) joint separations, particularly in throwing athletes. Twenty-eight major league baseball (MLB) orthopedic team physicians were surveyed to determine their definitive management of a grade III AC separation in the dominant arm of a professional baseball pitcher and their experience treating AC joint separations in starting pitchers and position players. Return-to-play outcomes were also evaluated. Twenty (71.4%) team physicians recommended nonoperative intervention compared to 8 (28.6%) who would have operated acutely. Eighteen (64.3%) team physicians had treated at least 1 professional pitcher with a grade III AC separation; 51 (77.3%) pitchers had been treated nonoperatively compared to 15 (22.7%) operatively. No difference was observed in the proportion of pitchers who returned to the same level of play (P = .54), had full, unrestricted range of motion (P = .23), or had full pain relief (P = .19) between the operatively and nonoperatively treated MLB pitchers. The majority (53.6%) of physicians would not include an injection if the injury was treated nonoperatively. Open coracoclavicular reconstruction (65.2%) was preferred for operative cases; 66.7% of surgeons would also include distal clavicle excision as an adjunct procedure. About 90% of physicians would return pitchers to throwing >12 weeks after surgery compared to after 4 to 6 weeks in nonoperatively treated cases. In conclusion, MLB team physicians preferred nonoperative management for an acute grade III AC joint separation in professional pitchers. If operative intervention is required, ligament reconstruction with adjunct distal clavicle excision were the most commonly performed procedures.

Continue to: Despite advancements in surgucal technique...

Despite advancements in surgical technique and improved understanding of the physiology of throwing mechanics, controversy persists regarding the preferred treatment for grade III acromioclavicular (AC) joint separations.^1-6 Nonsurgical management has demonstrated return to prior function with fewer complications.⁷ However, there is a growing body of evidence demonstrating that surgical intervention is associated with more favorable outcomes⁸ and should be considered in patients who place high functional demands on their shoulders.⁹

The reported results on professional athletes in the literature remain ambivalent. Multiple small case reports/series have reported successful nonoperative treatment of elite athletes.^10-12 Not surprisingly, McFarland and colleagues¹³ reported in 1997 that 69% of major league baseball (MLB) team physicians preferred nonoperative treatment for a theoretical starting pitcher sustaining a grade III AC separation 1 week prior to the start of the season. In contrast, reports of an inability to throw at a pre-injury level are equally commonplace.^14,15 Nevertheless, all of these studies were limited to small cohorts, as the incidence of grade III AC separations in elite throwing athletes is relatively uncommon.^13,16

In this study, we re-evaluated the study performed by McFarland and colleagues¹³ in 1997 by surveying all active MLB team orthopedic surgeons. We asked them how they would treat a grade III AC separation in a starting professional baseball pitcher. The physicians were also asked about their personal experience evaluating outcomes in these elite athletes. Given our improved understanding of the anatomy, pathophysiology, and surgical techniques for treating grade III AC separations, we hypothesize that more MLB team physicians would favor operative intervention treatment in professional baseball pitchers, as their vocation places higher demands on their shoulders.

MATERIALS AND METHODS

A questionnaire (Appendix A) was distributed to the team physicians of all 30 MLB teams. In addition to surgeon demographics, including age, years in practice, and years of taking care of an MLB team, the initial section of the questionnaire asked orthopedic surgeons how they would treat a theoretical starting pitcher who sustained a grade III AC joint separation of the dominant throwing arm 1 week prior to the start of the season. Physicians who preferred nonoperative treatment were asked whether they would use an injection (and what type), as well as when they would allow the pitcher to start a progressive interval throwing program. Physicians who preferred operative treatment were asked to rank their indications for operating, what procedure they would use (eg, open vs arthroscopic or coracoclavicular ligament repair vs reconstruction), and whether the surgical intervention would include distal clavicle excision. Both groups of physicians were also asked if their preferred treatment would change if the injury were to occur at the end of the season.

The second portion of the questionnaire asked surgeons about their experience treating AC joint separations in both starting pitchers and position players, as well as to describe the long-term outcomes of their preferred treatment, including time to return to full clearance for pitching, whether their patients returned to their prior level of play, and whether these patients had full pain relief. Finally, physicians were asked if any of the nonoperatively treated players ultimately crossed over and required operative intervention.

Continue to: Statistics...

STATISTICS

Descriptive statistics were used for continuous variables, and frequencies were used for categorical variables. Linear regression was performed to determine the correlation between the physician’s training or experience in treating AC joint separations and their recommended treatment. Fischer’s exact test/chi-square analysis was used to compare categorical variables. All tests were conducted using 2-sided hypothesis testing with statistical significance set at P < .05. All statistical analyses were conducted with SPSS 21.0 software (IBM Corporation).

RESULTS

A total of 28 MLB team physicians completed the questionnaires from 18 of the 30 MLB teams. The average age of the responders was 50.5 years (range, 34-60 years), with an average of 18.2 years in practice (range, 2-30 years) and 10.8 years (range, 1-24 years) taking care of their current professional baseball team. About 82% of the team physicians completed a sports medicine fellowship. On average, physicians saw 16.6 (range, 5-50) grade III or higher AC joint separations per year, and operated on 4.6 (range, 0-10) per year.

Nonoperative treatment was the preferred treatment for a grade III AC joint separation in a starting professional baseball pitcher for the majority of team physicians (20/28). No correlation was observed between the physician’s age (P = .881), years in practice (P = .915), years taking care of their professional team (P = .989), percentage of practice focused on shoulders (P = .986), number of AC joint injuries seen (P = .325), or number of surgeries performed per year (P = .807) with the team physician’s preferred treatment. Compared to the proportion reported originally by McFarland and colleagues¹³ in 1997 (69%), there was no difference in the proportion of team physicians that recommended nonoperative treatment (P = 1).

If treating this injury nonoperatively, 46.4% of physicians would also use an injection, with orthobiologics (eg, platelet-rich plasma) as the most popular choice (Table 1). No consensus was provided on the timeframe to return pitchers back to a progressive interval throwing program; however, 46.67% of physicians would return pitchers 4 to 6 weeks after a nonoperatively treated injury, while 35.7% would return pitchers 7 to 12 weeks after the initial injury.

Table 1. Treatment Preferences of Grade III AC Separation by MLB Team Physicians

^aRespondents were allowed to choose more than 1 treatment in each category. ^bChosen as an adjunct treatment.

Abbreviations: AC, acromioclavicular; MLB, major league baseball.

Most physicians (64.3%) cited functional limitations as the most important reason for indicating operative treatment, followed by pain (21.4%), and a deformity (14.3%). About 65% preferred open coracoclavicular ligament reconstruction. No physician recommended the Weaver-Dunn procedure or use of hardware (eg, hook plate). Of those who preferred an operative intervention, 66.7% would also include a distal clavicle excision, which is significantly higher than the proportion reported by McFarland and colleagues¹³ (23%, P = .0170). About 90% of physicians would return pitchers to play >12 weeks after operative treatment.

Continue to: If the injury occurred at the end ...

If the injury occurred at the end of the season, 7 of the 20 orthopedists (35%) who recommended nonoperative treatment said they would change to an operative intervention. Eighteen of 28 responders would have the same algorithm for MLB position players. Team physicians were less likely to recommend operative intervention in position players due to less demand on the arm and increased ability to accommodate the injury by altering their throwing mechanics.

Eighteen (64%) of the team physicians had treated at least 1 professional pitcher with a grade III AC separation in his dominant arm, and 11 (39.3%) had treated >1. Collectively, team physicians had treated 15 professional pitchers operatively, and 51 nonoperatively; only 3 patients converted to operative intervention after a failed nonoperative treatment.

Of the pitchers treated operatively, 93.3% (14) of pitchers returned to their prior level of pitching. The 1 patient who failed to return to the same level of pitching retired instead of returning to play. About 80% (12) of the pitchers had full pain relief, and 93.3% (14) had full range of motion (ROM). The pitcher who failed to regain full ROM also had a concomitant rotator cuff repair. The only complication reported from an operative intervention was a pitcher who sustained a coracoid fracture 10 months postoperatively while throwing 100 mph. Of the pitchers treated nonoperatively, 96% returned to their prior level of pitching, 92.2% (47) had full complete pain relief when throwing, and 100% had full ROM. No differences were observed between the proportion of pitchers who returned to their prior level of pitching, regained full ROM, or had full pain relief in the operative and nonoperative groups (Table 2).

Table 2. Outcomes of Treatment of Grade III AC Separation in 58 Professional Baseball Players

Abbreviations: AC, acromioclavicular; ROM, range of motion.

DISCUSSION 

Controversy persists regarding the optimal management of acute grade III AC separations, with the current available evidence potentially suggesting better cosmetic and radiological results but no definite differences in clinical results.^1-6,17,18 In the absence of formal clinical practice guidelines, surgeons rely on their own experience or defer to the anecdotal experience of experts in the field. Our initial hypothesis was false in this survey of MLB team physicians taking care of overhead throwing athletes at the highest level. Our results demonstrate that despite improved techniques and an increased understanding of the pathophysiology of AC joint separations, conservative management is still the preferred treatment for acute grade III AC joint separations in professional baseball pitchers. The proportion of team physicians recommending nonoperative treatment in our series was essentially equivalent to the results reported by McFarland and colleagues¹³ in 1997, suggesting that the pendulum continues to favor conservative management initially. This status quo likely reflects both the dearth of literature suggesting a substantial benefit of acute operative repair, as well as the ability to accommodate with conservative measures after most grade III AC injuries, even at the highest level of athletic competition.

These results are also consistent with trends from the last few decades. In the 1970s, the overwhelming preference for treating an acute complete AC joint separation was surgical repair, with Powers and Bach¹⁰ reporting in a 1974 survey of 163 chairmen of orthopedic programs around the country that 91.5% advocated surgical treatment. However, surgical preference had reversed by the 1990s. Of the 187 chairmen and 59 team physicians surveyed by Cox¹⁹ in 1992, 72% and 86% respectively preferred nonoperative treatment in a theoretical 21-year-old athlete with a grade III AC separation. Nissen and Chatterjee²⁰ reported in 2007 on a survey of all American Orthopaedic Society for Sports Medicine surgeons (N = 577) and Accreditation Council for Graduate Medical Education orthopedic program residency directors (N = 87) that >80% of responders preferred conservative measures for this acute injury. The reversal of trends has also been corroborated by recent multicenter trials demonstrating no difference in clinical outcomes between operative and nonoperative treatment of high grade AC joint dislocations, albeit these patients were not all high level overhead throwing athletes.^17,18

Continue to: The trends in surgical interventions are notable...

The trends in surgical interventions are notable within the smaller subset of patients who are indicated for operative repair. Use of hardware and primary ligament repair, while popular in the surveys conducted in the 1970s¹⁰ and 1990s¹³ and even present in Nissen and Chatterjee’s²⁰ 2007 survey, were noticeably absent from our survey results, with the majority of respondents preferring open coracoclavicular ligament reconstruction. The role of distal clavicle excision has also expanded, from 23% of team physicians recommending it in 1997¹³ to 57% to 59% in Nissen and Chatterjee’s²⁰ 2007 survey, to 66.7% in our series. This trend is not surprising as several recent cadaveric biomechanical studies have demonstrated that not only do peak graft forces not increase significantly,²¹ the anterior-posterior and superior-inferior motion at the AC joint following ligament reconstruction is maintained despite resection of the lateral clavicle.²² Additionally, primary distal clavicle excision may prevent the development of post-traumatic arthritis at the AC joint and osteolysis of the distal clavicle as a possible pain generator in the future.²³ However, some respondents cautioned against performing a concomitant distal clavicle excision, as some biomechanical data demonstrate that resecting the distal clavicle may lead to increased horizontal translation at the AC joint despite intact superior and posterior AC capsules.²⁴ Professional baseball pitchers may also be more lax and thus prone to more instability. Primary repair or reconstruction may not always lead to complete pre-injury stability in these individuals. This subtle unrecognized instability is hard to diagnosis and may be a persistent source of pain; thus, adding a distal clavicle excision may actually exacerbate the instability.

The nuanced indications for operative intervention, such as the presence of associated lesions were not captured by our survey.²⁵ While most team physicians cited functional limitations as their most common reason for offering surgery, several MLB orthopedic surgeons also commented on evaluating the stability of the AC joint after a grade III injury, akin to the consensus statement from the International Society of Arthroscopy, Knee Surgery and Orthopaedic Sports Medicine (ISAKOS) Upper Extremity Committee²⁶ in 2014 that diversified the Rockwood Grade III AC joint separation into its IIIA and IIIB classifications. The ISAKOS recommendations include initial conservative management and a second evaluation (both clinical and radiographic) for grade III lesions 3 to 6 weeks after the injury. However, as professional baseball is an incredibly profitable sport with an annual revenue approaching $9.5 billion²⁷ and pitching salaries up to $32.5 million in 2015, serious financial considerations must be given to players who wish to avoid undergoing delayed surgery.

This study has shortcomings typical of expert opinion papers. The retrospective nature of this study places the data at risk of recall bias. Objective data (eg, terminal ROM, pain relief, and return to play) were obtained from a retrospective chart review; however, no standard documentation or collection method was used given the number of surgeons involved and, thus, conclusions based on treatment outcomes are imperfect. Another major weakness of this survey is the relatively small number of patients and respondents. An a priori power analysis was not available, as this was a retrospective review. A comparative trial will be necessary to definitively support one treatment over another. Assuming a 95% return to play in the nonoperatively treated group, approximately 300 patients would be needed in a prospective 2-armed study with 80% power to detect a 10% reduction in the incidence of return to play using an alpha level of 0.05 and assuming no loss to follow-up. This sample size would be difficult to achieve in this patient population.

However, compared to past series,¹³ the number of professional baseball players treated by the collective experience of these MLB team physicians is the largest reported to date. As suggested above, the rarity of this condition in elite athletes precludes the ability to have matched controls to definitively determine the optimal treatment, which may explain the lack of difference in the return to play, ROM, and pain relief outcomes. Instead, we can only extrapolate based on the collective anecdotal experience of the MLB team physicians.

CONCLUSION

Despite advances in surgical technique and understanding of throwing mechanics, the majority of MLB team physicians preferred nonoperative management for an acute grade III AC joint separation in a professional baseball pitcher. Open coracoclavicular ligament reconstruction was preferred for those who preferred operative intervention. An increasing number of orthopedic surgeons now consider a distal clavicle excision as an adjunct procedure.

This paper will be judged for the Resident Writer’s Award.

ABSTRACT

The traditional technique for patella fracture fixation utilizes prominent hardware. Prominent hardware use, however, results in a high rate of reoperation for symptomatic implant removal. This biomechanical study evaluates the effectiveness of a novel patella fixation technique that minimizes implant prominence.

Patellar transverse osteotomies were created in 13 pairs of cadaveric knees. Paired knees were assigned to either standard fixation (SF) using cannulated partially threaded screws and stainless steel wire tension band, or buried fixation (BF) using headless compression screws with a No. 2 FiberWire tension band and a No. 5 FiberWire cerclage suture. Quadriceps tendons were cyclically loaded to full extension followed by load to failure. The gap across the fracture site, stiffness, and load to failure were measured.

The differences in stiffness and load to failure between the 2 groups were not statistically significant. During cyclic loading, significantly greater gapping was observed across the fracture site in the BF group compared with SF group (P < .05).

Both constructs failed under loads that exceeded typical loads experienced during the postoperative rehabilitation period. Nevertheless, the BF technique demonstrated larger gap formation and a reduced load to failure than the SF technique. Further clinical studies are therefore underway to determine whether the use of constructs with decreased stability but increased patient comfort could improve clinical outcomes and reduce reoperation rates.

Continue to: Patella fractures are common...

Patella fractures are common injuries that can cause considerable disability to the knee extensor apparatus.^1-3 Transverse patella fractures are the most common fracture pattern associated with patella fractures.{Harrell, 2003 #3}² Given that the patella plays a crucial role in knee extensor biomechanics, its proper integrity is vital for physiological knee motion and ambulation.⁴ Traditionally, patella fractures with >2 mm of displacement have been managed with cannulated screws or Kirschner wires (K-wires) and a stainless-steel wire tension band.^5-9 The goal in the treatment of patellar fractures is to reduce fracture fragments accurately and to minimize additional insults to the articular cartilage.¹⁰

Despite advances in surgical protocols and acceptable radiographic outcomes, functional impairment remains common after the treatment of patella fractures. Functional impairment includes knee pain, screw head pain, implant removal, wire breakage, and patella baja.¹ The need for implant removal is one of the most common complications following the open reduction internal fixation of patella fractures.^2,11 The subcutaneous and exposed nature of the patella in conjunction with soft tissue irritation resulting from standard fixation (SF) predisposes the patient toward prominence and discomfort with the retained implant. Although nonunion rates are low, the rate of implant removal can reach as high as 52%.^2,10-12 To overcome some of these complications, we designed a novel buried fixation (BF) method for the treatment of transverse fractures. Our method minimizes the amount of exposed implant to improve patient comfort and potentially reduce the need for future implant removal. These effects are achieved by using headless compression screws and nonabsorbable sutures to attenuate the soft tissue irritation associated with traditional fixation.¹³ While our novel technique has demonstrated improved clinical results, it has not been tested biomechanically against a traditional fixation technique. Therefore, this study aims to evaluate and compare the structural integrity of our novel BF technique with that of the standard technique that uses cannulated screws and wire tension band. We hypothesized that the stability provided by our technique would be similar to that provided by SF for transverse patella fractures.

MATERIALS AND METHODS

SPECIMEN PREPARATION

Thirteen matched pairs of fresh-frozen human cadaveric knees were obtained from a Cedars-Sinai approved tissue bank. Specimens were cut midfemur and were intact to the foot. Legs with major structural bony or ligamentous abnormalities, extensor mechanism disruption, or septic knees were excluded from testing. To assess the bone quality of each specimen prior to testing, dual-energy X-ray absorptiometry was performed using a GE Lunar iDXA scanner (GE Healthcare). Specimens were stored at −30°C and thawed at room temperature for 24 hours prior to biomechanical testing.

A midline anterior approach to the patella was performed, and the extensor retinaculum, quadriceps tendon, and patellar tendon were exposed. A digital caliper was used to measure the craniocaudal and mediolateral dimensions of the patella, and a transverse osteotomy (Arbeitsgemeinschaft für Osteosynthesefragen/Orthopaedic Trauma Association [AO/OTA] type 34-C1) was created at the midway point between superior and inferior poles by using an oscillating saw. The retinaculum was then incised to the level of the midaxial line of the femur. One leg from each matched pair was allocated to the SF group, and the other was allocated to the BF group. Left and right legs were alternately assigned to each group to ensure that laterality was balanced between the 2 groups.

SURGICAL TECHNIQUE

The repair of the specimens in the SF group involved the use of 2 parallel 4.0-mm partially threaded cannulated screws (Acumed) and an 18-gauge monofilament steel wire (Ethicon) in a figure-eight tension band (Figure 1A). The repair of the specimens in the BF group involved the use of 2 parallel standard Acutrak headless compression screws (Acumed), a No. 2 FiberWire (Arthrex) in a figure-eight tension band, and a No. 5 FiberWire (Arthrex) was applied as cerclage around the patella (Figure 1B).

Continue to: Mechanical testing...

MECHANICAL TESTING

Mechanical testing was performed on a biaxial 370.02 Bionix Testing System (MTS Systems Corp.). The femur was rigidly and horizontally secured to a custom-built test frame, and the lower leg was left free to move. The quadriceps tendon was secured in a freeze clamp and was attached to the MTS actuator for loading via a pulley system such that raising the actuator was translated into a simulated quadriceps extensor force.

A differential variable reluctance transducer (DVRT) (Lord MicroStrain) was placed across the osteotomy site to measure the distraction, or gap, across the fracture line. The minimum load to full extension for each specimen was then determined under a slow, controlled increase in load until the leg was in a fully extended position. Any distraction across the fracture line during the initial loading phase was determined by using digital calipers. The specimen was then subjected to a preconditioning phase with 10 cycles from 0 N to full extension under the previously determined load, which was applied at the rate of 5 N/s. Meanwhile, displacement across the fracture site was recorded via the DVRT. Following the preconditioning phase, each specimen was then tested to failure in displacement control at the rate of 1.5 mm/s. Failure was defined as implant failure (screw pullout) or DVRT gapping across the osteotomy site >3 mm.^10,14

Outcome measures included stiffness (N/mm), which was calculated as the slope of the linear change in load from full extension to failure vs DVRT displacement during the final loading phase; failure load (N); gapping (mm) across the osteotomy site at each cycle during the preconditioning phase; and failure mode (pullout vs >3.0 mm gap).

STATISTICAL ANALYSIS

An a priori power analysis revealed that 13 knees per group would be required to obtain an α of 0.05 and a power of 0.80. This calculation was based on a 20% difference in fracture displacement calculated by using the standard deviation and mean previously reported for cannulated screws with nonabsorbable sutures.¹⁴

Means and standard deviations for all dependent outcome measures were computed and compared across the independent measure of fixation type (BF vs SF) through repeated measures Analysis of variance (ANOVA-GLM, SAS 9.3, SAS Institute, Inc.) after controlling for bone mineral density (BMD), gender, and age. Multivariate repeated-measures ANOVA with Tukey's studentized range was applied to cyclic gap data. The mode of failure was compared across fixation type (BF vs SF) for matched data using McNemar’s test. Intracorrelations were computed and examined over all data and separately on the basis of screw fixation type (BF vs SF). All tests were considered statistically significant when P < .05.

Continue to: Results...

RESULTS

Specimen donors were 46% (6/13) male with an average age of 78.5 years (±13.77; range, 56-91 years) and 54% (7/13) female with an average age of 76.57 years (±14.37; range, 59-102 years). Average BMD was significantly lower in female (0.71 ± 0.18) than in male specimens (1.15 ± 0.33) (P < .05).

The average load to full extension across all specimens was 272 N (±54; range, 160-360 N) and was well balanced across matched pairs (270 ± 56 N for BF and 273 ± 54 N for SF). Of the 13 BF specimens, 4 experienced distraction across the fracture line during the determination of the minimum load to full extension. This initial pretest gap was measured with digital calipers (average, 1.5 mm; range, 0.90-1.85 mm) and added as an offset to the respective DVRT displacement data recorded during testing.

The total number of specimens included in the displacement data calculations decreased from 13 to 11 per group because DVRT data were not recorded during cyclic loading for 1 specimen and were considered unreliable in another. The maximum displacement measured across the fracture site during cyclic loading was significantly higher in the BF (0.94 ± 1.21) group than in the SF group (0.19 ± 0.26) as shown in the Table. The average slope of the gap per cycle for each specimen was calculated and compared between the BF and SF groups. The BF group demonstrated a significantly greater increase in gap per cycle than the SF group (Figure 2). Stiffness during load to failure was calculated for all but 1 specimen that did not display any measurable displacement during the final loading cycle. The average final stiffness and failure load between the BF and SF groups were not significantly different (Table). An equal number of specimens in both groups failed through gapping (6/13) and pullout (7/13).

Table. Means and Standard Deviations of the Main Outcome Measures

^aIndicates statistical significance (P < .05).

Abbreviation: NS, not significant.

Failure load was significantly positively correlated with BMD (R = 0.62, P < .001) when all specimens were grouped together. When analyzed separately, the SF group was significantly correlated with BMD (P < .01), whereas the BF group had a marginally significant correlation (P = .06) with BMD (Figure 3). In both groups, BMD was positively correlated with stiffness and negatively correlated with gapping. Neither of these trends, however, was significant.

Continue to: Discussion...

DISCUSSION

We proposed a novel BF technique for the treatment of noncomminuted transverse patella fractures. Our technique utilizes headless cannulated compression screws and nonabsorbable suture tension bands. We then biomechanically compared our proposed technique with an established fixation technique that uses partially threaded cannulated screws and stainless steel wire tension bands. We hypothesized that the mechanical response of the BF technique to cyclic and failure loading would be similar to that of the SF technique. Our results demonstrate a significant increase in gap formation across the fracture site among knees and an overall reduced load to failure in the BF group (Figure 2). Whether these inferior results manifest clinically is not yet established. Both constructs could withstand forces that are typically experienced during the postoperative period. Given the high rate of symptomatic implant removal associated with the traditional technique, the low-profile buried technique might be an attractive alternative that provides increased patient comfort but may require an extended period of postoperative protection against bony ingrowths.

Patellar fixation constructs that use a combination of cannulated screws and a wire tension band provide the best resistance to patella fracture displacement when compared with screws or wires alone.^4,15 Although this combination is biomechanically favorable, the steel wire often causes the painful irritation of the surrounding soft tissues and can break or migrate, thus increasing the rates of implant removal surgery to as high as 52%.^4,10,12,15 We developed our novel BF technique, which uses headless compression screws and a No. 2 FiberWire tension band, to address the high rates of reoperation and patient dissatisfaction associated with the SF technique.

Headless compression screws have been successfully used in the reduction and fixation of scaphoid fractures and sesamoid fractures.^16,17 The pull-out strengths of these screws are comparable with those of other commonly used screws, such as Twinfix and Herbert-Whipple screws.¹⁶ Similarly, the strength of a No. 5 FiberWire is comparable with that of an 18-gauge stainless-steel wire.^14,18 Several studies have also obtained good outcomes with nonmetallic constructs that use nonabsorbable sutures alone.^19,20 In this study, we utilized a No. 2 FiberWire as the tension band. The use of the No. 2 FiberWire facilitated threading through headless cannulated screws and created a low-profile knot. However, the use of thin FiberWire, despite a No. 5 FiberWire cerclage, likely contributed to the increase in distraction across the fracture.

The highest patellofemoral joint reaction force during level walking is approximately 35 kg (half body weight), which is equivalent to 350 N.^15,21,22 This force is similar to the average cyclic load used in this experiment (272 ± 54 N). Gapping increased in the BF group but did not reach the defined failure value of 3 mm, and the ultimate load to failure was relatively high across both groups (SF, 1123 N; BF, 973 N). These results suggest that both fixation methods can withstand the typical patellofemoral joint forces that are experienced during the postoperative period.⁴ In addition, in a clinical setting, patients are placed in hinged knee braces for at least 2 weeks to limit their flexion angle and to allow for healing and bony ingrowth. Postoperative knee-brace protection presumably increases the overall strength of the fixation.

The number of specimens (n = 26) evaluated in this study was greater than that used in other biomechanical patella fracture studies.¹⁴ Furthermore, none of our specimens were reused. Our study design was further strengthened given that fellowship-trained trauma surgeons performed all surgical procedures. Finally, the data collection and analysis of numerous clinically relevant factors, such as BMD, age, and cyclical loading, contributed to the comprehensive description of each technique with respect to patient-specific criteria.

Similar to all cadaveric studies, our data only represent the immediate postoperative condition and does not represent any healing that would occur during postoperative rehabilitation. Postoperative knee-brace protection and bone healing across the fracture site would likely strengthen both constructs in a clinical setting. In addition, the average age of our specimens is 77.5 years, and therefore does not best represent the age range (20-50 years) of the typical adult population affected by patella fractures.^3,23,24 Finally, postsurgical reduction was confirmed through visual inspection and not through fluoroscopy as in a clinical setting. Radiographic images were obtained after each experiment only to confirm screw placement post facto (Figures 4A, 4B).

CONCLUSION

This study demonstrates the utility of a novel BF technique. Nevertheless, the proposed technique exhibited increased gapping and a lower load to failure than the current gold standard. The significance of these inferior results in clinical and functional settings has not been established. The proposed BF technique may be an appealing alternative to the SF technique given its low profile and potential to reduce the rates of future implant removal. Further studies on the long-term outcomes of patients treated through the BF technique are currently under way and will ultimately determine the utility of the proposed construct.

This paper will be judged for the Resident Writer’s Award.

ABSTRACT

The traditional technique for patella fracture fixation utilizes prominent hardware. Prominent hardware use, however, results in a high rate of reoperation for symptomatic implant removal. This biomechanical study evaluates the effectiveness of a novel patella fixation technique that minimizes implant prominence.

Patellar transverse osteotomies were created in 13 pairs of cadaveric knees. Paired knees were assigned to either standard fixation (SF) using cannulated partially threaded screws and stainless steel wire tension band, or buried fixation (BF) using headless compression screws with a No. 2 FiberWire tension band and a No. 5 FiberWire cerclage suture. Quadriceps tendons were cyclically loaded to full extension followed by load to failure. The gap across the fracture site, stiffness, and load to failure were measured.

The differences in stiffness and load to failure between the 2 groups were not statistically significant. During cyclic loading, significantly greater gapping was observed across the fracture site in the BF group compared with SF group (P < .05).

Both constructs failed under loads that exceeded typical loads experienced during the postoperative rehabilitation period. Nevertheless, the BF technique demonstrated larger gap formation and a reduced load to failure than the SF technique. Further clinical studies are therefore underway to determine whether the use of constructs with decreased stability but increased patient comfort could improve clinical outcomes and reduce reoperation rates.

Continue to: Patella fractures are common...

Patella fractures are common injuries that can cause considerable disability to the knee extensor apparatus.^1-3 Transverse patella fractures are the most common fracture pattern associated with patella fractures.{Harrell, 2003 #3}² Given that the patella plays a crucial role in knee extensor biomechanics, its proper integrity is vital for physiological knee motion and ambulation.⁴ Traditionally, patella fractures with >2 mm of displacement have been managed with cannulated screws or Kirschner wires (K-wires) and a stainless-steel wire tension band.^5-9 The goal in the treatment of patellar fractures is to reduce fracture fragments accurately and to minimize additional insults to the articular cartilage.¹⁰

Despite advances in surgical protocols and acceptable radiographic outcomes, functional impairment remains common after the treatment of patella fractures. Functional impairment includes knee pain, screw head pain, implant removal, wire breakage, and patella baja.¹ The need for implant removal is one of the most common complications following the open reduction internal fixation of patella fractures.^2,11 The subcutaneous and exposed nature of the patella in conjunction with soft tissue irritation resulting from standard fixation (SF) predisposes the patient toward prominence and discomfort with the retained implant. Although nonunion rates are low, the rate of implant removal can reach as high as 52%.^2,10-12 To overcome some of these complications, we designed a novel buried fixation (BF) method for the treatment of transverse fractures. Our method minimizes the amount of exposed implant to improve patient comfort and potentially reduce the need for future implant removal. These effects are achieved by using headless compression screws and nonabsorbable sutures to attenuate the soft tissue irritation associated with traditional fixation.¹³ While our novel technique has demonstrated improved clinical results, it has not been tested biomechanically against a traditional fixation technique. Therefore, this study aims to evaluate and compare the structural integrity of our novel BF technique with that of the standard technique that uses cannulated screws and wire tension band. We hypothesized that the stability provided by our technique would be similar to that provided by SF for transverse patella fractures.

MATERIALS AND METHODS

SPECIMEN PREPARATION

Thirteen matched pairs of fresh-frozen human cadaveric knees were obtained from a Cedars-Sinai approved tissue bank. Specimens were cut midfemur and were intact to the foot. Legs with major structural bony or ligamentous abnormalities, extensor mechanism disruption, or septic knees were excluded from testing. To assess the bone quality of each specimen prior to testing, dual-energy X-ray absorptiometry was performed using a GE Lunar iDXA scanner (GE Healthcare). Specimens were stored at −30°C and thawed at room temperature for 24 hours prior to biomechanical testing.

A midline anterior approach to the patella was performed, and the extensor retinaculum, quadriceps tendon, and patellar tendon were exposed. A digital caliper was used to measure the craniocaudal and mediolateral dimensions of the patella, and a transverse osteotomy (Arbeitsgemeinschaft für Osteosynthesefragen/Orthopaedic Trauma Association [AO/OTA] type 34-C1) was created at the midway point between superior and inferior poles by using an oscillating saw. The retinaculum was then incised to the level of the midaxial line of the femur. One leg from each matched pair was allocated to the SF group, and the other was allocated to the BF group. Left and right legs were alternately assigned to each group to ensure that laterality was balanced between the 2 groups.

SURGICAL TECHNIQUE

The repair of the specimens in the SF group involved the use of 2 parallel 4.0-mm partially threaded cannulated screws (Acumed) and an 18-gauge monofilament steel wire (Ethicon) in a figure-eight tension band (Figure 1A). The repair of the specimens in the BF group involved the use of 2 parallel standard Acutrak headless compression screws (Acumed), a No. 2 FiberWire (Arthrex) in a figure-eight tension band, and a No. 5 FiberWire (Arthrex) was applied as cerclage around the patella (Figure 1B).

Continue to: Mechanical testing...

MECHANICAL TESTING

Mechanical testing was performed on a biaxial 370.02 Bionix Testing System (MTS Systems Corp.). The femur was rigidly and horizontally secured to a custom-built test frame, and the lower leg was left free to move. The quadriceps tendon was secured in a freeze clamp and was attached to the MTS actuator for loading via a pulley system such that raising the actuator was translated into a simulated quadriceps extensor force.

A differential variable reluctance transducer (DVRT) (Lord MicroStrain) was placed across the osteotomy site to measure the distraction, or gap, across the fracture line. The minimum load to full extension for each specimen was then determined under a slow, controlled increase in load until the leg was in a fully extended position. Any distraction across the fracture line during the initial loading phase was determined by using digital calipers. The specimen was then subjected to a preconditioning phase with 10 cycles from 0 N to full extension under the previously determined load, which was applied at the rate of 5 N/s. Meanwhile, displacement across the fracture site was recorded via the DVRT. Following the preconditioning phase, each specimen was then tested to failure in displacement control at the rate of 1.5 mm/s. Failure was defined as implant failure (screw pullout) or DVRT gapping across the osteotomy site >3 mm.^10,14

Outcome measures included stiffness (N/mm), which was calculated as the slope of the linear change in load from full extension to failure vs DVRT displacement during the final loading phase; failure load (N); gapping (mm) across the osteotomy site at each cycle during the preconditioning phase; and failure mode (pullout vs >3.0 mm gap).

STATISTICAL ANALYSIS

An a priori power analysis revealed that 13 knees per group would be required to obtain an α of 0.05 and a power of 0.80. This calculation was based on a 20% difference in fracture displacement calculated by using the standard deviation and mean previously reported for cannulated screws with nonabsorbable sutures.¹⁴

Means and standard deviations for all dependent outcome measures were computed and compared across the independent measure of fixation type (BF vs SF) through repeated measures Analysis of variance (ANOVA-GLM, SAS 9.3, SAS Institute, Inc.) after controlling for bone mineral density (BMD), gender, and age. Multivariate repeated-measures ANOVA with Tukey's studentized range was applied to cyclic gap data. The mode of failure was compared across fixation type (BF vs SF) for matched data using McNemar’s test. Intracorrelations were computed and examined over all data and separately on the basis of screw fixation type (BF vs SF). All tests were considered statistically significant when P < .05.

Continue to: Results...

RESULTS

Specimen donors were 46% (6/13) male with an average age of 78.5 years (±13.77; range, 56-91 years) and 54% (7/13) female with an average age of 76.57 years (±14.37; range, 59-102 years). Average BMD was significantly lower in female (0.71 ± 0.18) than in male specimens (1.15 ± 0.33) (P < .05).

The average load to full extension across all specimens was 272 N (±54; range, 160-360 N) and was well balanced across matched pairs (270 ± 56 N for BF and 273 ± 54 N for SF). Of the 13 BF specimens, 4 experienced distraction across the fracture line during the determination of the minimum load to full extension. This initial pretest gap was measured with digital calipers (average, 1.5 mm; range, 0.90-1.85 mm) and added as an offset to the respective DVRT displacement data recorded during testing.

The total number of specimens included in the displacement data calculations decreased from 13 to 11 per group because DVRT data were not recorded during cyclic loading for 1 specimen and were considered unreliable in another. The maximum displacement measured across the fracture site during cyclic loading was significantly higher in the BF (0.94 ± 1.21) group than in the SF group (0.19 ± 0.26) as shown in the Table. The average slope of the gap per cycle for each specimen was calculated and compared between the BF and SF groups. The BF group demonstrated a significantly greater increase in gap per cycle than the SF group (Figure 2). Stiffness during load to failure was calculated for all but 1 specimen that did not display any measurable displacement during the final loading cycle. The average final stiffness and failure load between the BF and SF groups were not significantly different (Table). An equal number of specimens in both groups failed through gapping (6/13) and pullout (7/13).

Table. Means and Standard Deviations of the Main Outcome Measures

^aIndicates statistical significance (P < .05).

Abbreviation: NS, not significant.

Failure load was significantly positively correlated with BMD (R = 0.62, P < .001) when all specimens were grouped together. When analyzed separately, the SF group was significantly correlated with BMD (P < .01), whereas the BF group had a marginally significant correlation (P = .06) with BMD (Figure 3). In both groups, BMD was positively correlated with stiffness and negatively correlated with gapping. Neither of these trends, however, was significant.

Continue to: Discussion...

DISCUSSION

We proposed a novel BF technique for the treatment of noncomminuted transverse patella fractures. Our technique utilizes headless cannulated compression screws and nonabsorbable suture tension bands. We then biomechanically compared our proposed technique with an established fixation technique that uses partially threaded cannulated screws and stainless steel wire tension bands. We hypothesized that the mechanical response of the BF technique to cyclic and failure loading would be similar to that of the SF technique. Our results demonstrate a significant increase in gap formation across the fracture site among knees and an overall reduced load to failure in the BF group (Figure 2). Whether these inferior results manifest clinically is not yet established. Both constructs could withstand forces that are typically experienced during the postoperative period. Given the high rate of symptomatic implant removal associated with the traditional technique, the low-profile buried technique might be an attractive alternative that provides increased patient comfort but may require an extended period of postoperative protection against bony ingrowths.

Patellar fixation constructs that use a combination of cannulated screws and a wire tension band provide the best resistance to patella fracture displacement when compared with screws or wires alone.^4,15 Although this combination is biomechanically favorable, the steel wire often causes the painful irritation of the surrounding soft tissues and can break or migrate, thus increasing the rates of implant removal surgery to as high as 52%.^4,10,12,15 We developed our novel BF technique, which uses headless compression screws and a No. 2 FiberWire tension band, to address the high rates of reoperation and patient dissatisfaction associated with the SF technique.

Headless compression screws have been successfully used in the reduction and fixation of scaphoid fractures and sesamoid fractures.^16,17 The pull-out strengths of these screws are comparable with those of other commonly used screws, such as Twinfix and Herbert-Whipple screws.¹⁶ Similarly, the strength of a No. 5 FiberWire is comparable with that of an 18-gauge stainless-steel wire.^14,18 Several studies have also obtained good outcomes with nonmetallic constructs that use nonabsorbable sutures alone.^19,20 In this study, we utilized a No. 2 FiberWire as the tension band. The use of the No. 2 FiberWire facilitated threading through headless cannulated screws and created a low-profile knot. However, the use of thin FiberWire, despite a No. 5 FiberWire cerclage, likely contributed to the increase in distraction across the fracture.

The highest patellofemoral joint reaction force during level walking is approximately 35 kg (half body weight), which is equivalent to 350 N.^15,21,22 This force is similar to the average cyclic load used in this experiment (272 ± 54 N). Gapping increased in the BF group but did not reach the defined failure value of 3 mm, and the ultimate load to failure was relatively high across both groups (SF, 1123 N; BF, 973 N). These results suggest that both fixation methods can withstand the typical patellofemoral joint forces that are experienced during the postoperative period.⁴ In addition, in a clinical setting, patients are placed in hinged knee braces for at least 2 weeks to limit their flexion angle and to allow for healing and bony ingrowth. Postoperative knee-brace protection presumably increases the overall strength of the fixation.

The number of specimens (n = 26) evaluated in this study was greater than that used in other biomechanical patella fracture studies.¹⁴ Furthermore, none of our specimens were reused. Our study design was further strengthened given that fellowship-trained trauma surgeons performed all surgical procedures. Finally, the data collection and analysis of numerous clinically relevant factors, such as BMD, age, and cyclical loading, contributed to the comprehensive description of each technique with respect to patient-specific criteria.

Similar to all cadaveric studies, our data only represent the immediate postoperative condition and does not represent any healing that would occur during postoperative rehabilitation. Postoperative knee-brace protection and bone healing across the fracture site would likely strengthen both constructs in a clinical setting. In addition, the average age of our specimens is 77.5 years, and therefore does not best represent the age range (20-50 years) of the typical adult population affected by patella fractures.^3,23,24 Finally, postsurgical reduction was confirmed through visual inspection and not through fluoroscopy as in a clinical setting. Radiographic images were obtained after each experiment only to confirm screw placement post facto (Figures 4A, 4B).

CONCLUSION

This study demonstrates the utility of a novel BF technique. Nevertheless, the proposed technique exhibited increased gapping and a lower load to failure than the current gold standard. The significance of these inferior results in clinical and functional settings has not been established. The proposed BF technique may be an appealing alternative to the SF technique given its low profile and potential to reduce the rates of future implant removal. Further studies on the long-term outcomes of patients treated through the BF technique are currently under way and will ultimately determine the utility of the proposed construct.

This paper will be judged for the Resident Writer’s Award.

ABSTRACT

The traditional technique for patella fracture fixation utilizes prominent hardware. Prominent hardware use, however, results in a high rate of reoperation for symptomatic implant removal. This biomechanical study evaluates the effectiveness of a novel patella fixation technique that minimizes implant prominence.

Patellar transverse osteotomies were created in 13 pairs of cadaveric knees. Paired knees were assigned to either standard fixation (SF) using cannulated partially threaded screws and stainless steel wire tension band, or buried fixation (BF) using headless compression screws with a No. 2 FiberWire tension band and a No. 5 FiberWire cerclage suture. Quadriceps tendons were cyclically loaded to full extension followed by load to failure. The gap across the fracture site, stiffness, and load to failure were measured.

The differences in stiffness and load to failure between the 2 groups were not statistically significant. During cyclic loading, significantly greater gapping was observed across the fracture site in the BF group compared with SF group (P < .05).

Both constructs failed under loads that exceeded typical loads experienced during the postoperative rehabilitation period. Nevertheless, the BF technique demonstrated larger gap formation and a reduced load to failure than the SF technique. Further clinical studies are therefore underway to determine whether the use of constructs with decreased stability but increased patient comfort could improve clinical outcomes and reduce reoperation rates.

Continue to: Patella fractures are common...

Patella fractures are common injuries that can cause considerable disability to the knee extensor apparatus.^1-3 Transverse patella fractures are the most common fracture pattern associated with patella fractures.{Harrell, 2003 #3}² Given that the patella plays a crucial role in knee extensor biomechanics, its proper integrity is vital for physiological knee motion and ambulation.⁴ Traditionally, patella fractures with >2 mm of displacement have been managed with cannulated screws or Kirschner wires (K-wires) and a stainless-steel wire tension band.^5-9 The goal in the treatment of patellar fractures is to reduce fracture fragments accurately and to minimize additional insults to the articular cartilage.¹⁰

Despite advances in surgical protocols and acceptable radiographic outcomes, functional impairment remains common after the treatment of patella fractures. Functional impairment includes knee pain, screw head pain, implant removal, wire breakage, and patella baja.¹ The need for implant removal is one of the most common complications following the open reduction internal fixation of patella fractures.^2,11 The subcutaneous and exposed nature of the patella in conjunction with soft tissue irritation resulting from standard fixation (SF) predisposes the patient toward prominence and discomfort with the retained implant. Although nonunion rates are low, the rate of implant removal can reach as high as 52%.^2,10-12 To overcome some of these complications, we designed a novel buried fixation (BF) method for the treatment of transverse fractures. Our method minimizes the amount of exposed implant to improve patient comfort and potentially reduce the need for future implant removal. These effects are achieved by using headless compression screws and nonabsorbable sutures to attenuate the soft tissue irritation associated with traditional fixation.¹³ While our novel technique has demonstrated improved clinical results, it has not been tested biomechanically against a traditional fixation technique. Therefore, this study aims to evaluate and compare the structural integrity of our novel BF technique with that of the standard technique that uses cannulated screws and wire tension band. We hypothesized that the stability provided by our technique would be similar to that provided by SF for transverse patella fractures.

MATERIALS AND METHODS

SPECIMEN PREPARATION

Thirteen matched pairs of fresh-frozen human cadaveric knees were obtained from a Cedars-Sinai approved tissue bank. Specimens were cut midfemur and were intact to the foot. Legs with major structural bony or ligamentous abnormalities, extensor mechanism disruption, or septic knees were excluded from testing. To assess the bone quality of each specimen prior to testing, dual-energy X-ray absorptiometry was performed using a GE Lunar iDXA scanner (GE Healthcare). Specimens were stored at −30°C and thawed at room temperature for 24 hours prior to biomechanical testing.

A midline anterior approach to the patella was performed, and the extensor retinaculum, quadriceps tendon, and patellar tendon were exposed. A digital caliper was used to measure the craniocaudal and mediolateral dimensions of the patella, and a transverse osteotomy (Arbeitsgemeinschaft für Osteosynthesefragen/Orthopaedic Trauma Association [AO/OTA] type 34-C1) was created at the midway point between superior and inferior poles by using an oscillating saw. The retinaculum was then incised to the level of the midaxial line of the femur. One leg from each matched pair was allocated to the SF group, and the other was allocated to the BF group. Left and right legs were alternately assigned to each group to ensure that laterality was balanced between the 2 groups.

SURGICAL TECHNIQUE

The repair of the specimens in the SF group involved the use of 2 parallel 4.0-mm partially threaded cannulated screws (Acumed) and an 18-gauge monofilament steel wire (Ethicon) in a figure-eight tension band (Figure 1A). The repair of the specimens in the BF group involved the use of 2 parallel standard Acutrak headless compression screws (Acumed), a No. 2 FiberWire (Arthrex) in a figure-eight tension band, and a No. 5 FiberWire (Arthrex) was applied as cerclage around the patella (Figure 1B).

Continue to: Mechanical testing...

MECHANICAL TESTING

Mechanical testing was performed on a biaxial 370.02 Bionix Testing System (MTS Systems Corp.). The femur was rigidly and horizontally secured to a custom-built test frame, and the lower leg was left free to move. The quadriceps tendon was secured in a freeze clamp and was attached to the MTS actuator for loading via a pulley system such that raising the actuator was translated into a simulated quadriceps extensor force.

A differential variable reluctance transducer (DVRT) (Lord MicroStrain) was placed across the osteotomy site to measure the distraction, or gap, across the fracture line. The minimum load to full extension for each specimen was then determined under a slow, controlled increase in load until the leg was in a fully extended position. Any distraction across the fracture line during the initial loading phase was determined by using digital calipers. The specimen was then subjected to a preconditioning phase with 10 cycles from 0 N to full extension under the previously determined load, which was applied at the rate of 5 N/s. Meanwhile, displacement across the fracture site was recorded via the DVRT. Following the preconditioning phase, each specimen was then tested to failure in displacement control at the rate of 1.5 mm/s. Failure was defined as implant failure (screw pullout) or DVRT gapping across the osteotomy site >3 mm.^10,14

Outcome measures included stiffness (N/mm), which was calculated as the slope of the linear change in load from full extension to failure vs DVRT displacement during the final loading phase; failure load (N); gapping (mm) across the osteotomy site at each cycle during the preconditioning phase; and failure mode (pullout vs >3.0 mm gap).

STATISTICAL ANALYSIS

An a priori power analysis revealed that 13 knees per group would be required to obtain an α of 0.05 and a power of 0.80. This calculation was based on a 20% difference in fracture displacement calculated by using the standard deviation and mean previously reported for cannulated screws with nonabsorbable sutures.¹⁴

Means and standard deviations for all dependent outcome measures were computed and compared across the independent measure of fixation type (BF vs SF) through repeated measures Analysis of variance (ANOVA-GLM, SAS 9.3, SAS Institute, Inc.) after controlling for bone mineral density (BMD), gender, and age. Multivariate repeated-measures ANOVA with Tukey's studentized range was applied to cyclic gap data. The mode of failure was compared across fixation type (BF vs SF) for matched data using McNemar’s test. Intracorrelations were computed and examined over all data and separately on the basis of screw fixation type (BF vs SF). All tests were considered statistically significant when P < .05.

Continue to: Results...

RESULTS

Specimen donors were 46% (6/13) male with an average age of 78.5 years (±13.77; range, 56-91 years) and 54% (7/13) female with an average age of 76.57 years (±14.37; range, 59-102 years). Average BMD was significantly lower in female (0.71 ± 0.18) than in male specimens (1.15 ± 0.33) (P < .05).

The average load to full extension across all specimens was 272 N (±54; range, 160-360 N) and was well balanced across matched pairs (270 ± 56 N for BF and 273 ± 54 N for SF). Of the 13 BF specimens, 4 experienced distraction across the fracture line during the determination of the minimum load to full extension. This initial pretest gap was measured with digital calipers (average, 1.5 mm; range, 0.90-1.85 mm) and added as an offset to the respective DVRT displacement data recorded during testing.

The total number of specimens included in the displacement data calculations decreased from 13 to 11 per group because DVRT data were not recorded during cyclic loading for 1 specimen and were considered unreliable in another. The maximum displacement measured across the fracture site during cyclic loading was significantly higher in the BF (0.94 ± 1.21) group than in the SF group (0.19 ± 0.26) as shown in the Table. The average slope of the gap per cycle for each specimen was calculated and compared between the BF and SF groups. The BF group demonstrated a significantly greater increase in gap per cycle than the SF group (Figure 2). Stiffness during load to failure was calculated for all but 1 specimen that did not display any measurable displacement during the final loading cycle. The average final stiffness and failure load between the BF and SF groups were not significantly different (Table). An equal number of specimens in both groups failed through gapping (6/13) and pullout (7/13).

Table. Means and Standard Deviations of the Main Outcome Measures

^aIndicates statistical significance (P < .05).

Abbreviation: NS, not significant.

Failure load was significantly positively correlated with BMD (R = 0.62, P < .001) when all specimens were grouped together. When analyzed separately, the SF group was significantly correlated with BMD (P < .01), whereas the BF group had a marginally significant correlation (P = .06) with BMD (Figure 3). In both groups, BMD was positively correlated with stiffness and negatively correlated with gapping. Neither of these trends, however, was significant.

Continue to: Discussion...

DISCUSSION

We proposed a novel BF technique for the treatment of noncomminuted transverse patella fractures. Our technique utilizes headless cannulated compression screws and nonabsorbable suture tension bands. We then biomechanically compared our proposed technique with an established fixation technique that uses partially threaded cannulated screws and stainless steel wire tension bands. We hypothesized that the mechanical response of the BF technique to cyclic and failure loading would be similar to that of the SF technique. Our results demonstrate a significant increase in gap formation across the fracture site among knees and an overall reduced load to failure in the BF group (Figure 2). Whether these inferior results manifest clinically is not yet established. Both constructs could withstand forces that are typically experienced during the postoperative period. Given the high rate of symptomatic implant removal associated with the traditional technique, the low-profile buried technique might be an attractive alternative that provides increased patient comfort but may require an extended period of postoperative protection against bony ingrowths.

Patellar fixation constructs that use a combination of cannulated screws and a wire tension band provide the best resistance to patella fracture displacement when compared with screws or wires alone.^4,15 Although this combination is biomechanically favorable, the steel wire often causes the painful irritation of the surrounding soft tissues and can break or migrate, thus increasing the rates of implant removal surgery to as high as 52%.^4,10,12,15 We developed our novel BF technique, which uses headless compression screws and a No. 2 FiberWire tension band, to address the high rates of reoperation and patient dissatisfaction associated with the SF technique.

Headless compression screws have been successfully used in the reduction and fixation of scaphoid fractures and sesamoid fractures.^16,17 The pull-out strengths of these screws are comparable with those of other commonly used screws, such as Twinfix and Herbert-Whipple screws.¹⁶ Similarly, the strength of a No. 5 FiberWire is comparable with that of an 18-gauge stainless-steel wire.^14,18 Several studies have also obtained good outcomes with nonmetallic constructs that use nonabsorbable sutures alone.^19,20 In this study, we utilized a No. 2 FiberWire as the tension band. The use of the No. 2 FiberWire facilitated threading through headless cannulated screws and created a low-profile knot. However, the use of thin FiberWire, despite a No. 5 FiberWire cerclage, likely contributed to the increase in distraction across the fracture.

The highest patellofemoral joint reaction force during level walking is approximately 35 kg (half body weight), which is equivalent to 350 N.^15,21,22 This force is similar to the average cyclic load used in this experiment (272 ± 54 N). Gapping increased in the BF group but did not reach the defined failure value of 3 mm, and the ultimate load to failure was relatively high across both groups (SF, 1123 N; BF, 973 N). These results suggest that both fixation methods can withstand the typical patellofemoral joint forces that are experienced during the postoperative period.⁴ In addition, in a clinical setting, patients are placed in hinged knee braces for at least 2 weeks to limit their flexion angle and to allow for healing and bony ingrowth. Postoperative knee-brace protection presumably increases the overall strength of the fixation.

The number of specimens (n = 26) evaluated in this study was greater than that used in other biomechanical patella fracture studies.¹⁴ Furthermore, none of our specimens were reused. Our study design was further strengthened given that fellowship-trained trauma surgeons performed all surgical procedures. Finally, the data collection and analysis of numerous clinically relevant factors, such as BMD, age, and cyclical loading, contributed to the comprehensive description of each technique with respect to patient-specific criteria.

Similar to all cadaveric studies, our data only represent the immediate postoperative condition and does not represent any healing that would occur during postoperative rehabilitation. Postoperative knee-brace protection and bone healing across the fracture site would likely strengthen both constructs in a clinical setting. In addition, the average age of our specimens is 77.5 years, and therefore does not best represent the age range (20-50 years) of the typical adult population affected by patella fractures.^3,23,24 Finally, postsurgical reduction was confirmed through visual inspection and not through fluoroscopy as in a clinical setting. Radiographic images were obtained after each experiment only to confirm screw placement post facto (Figures 4A, 4B).

CONCLUSION

This study demonstrates the utility of a novel BF technique. Nevertheless, the proposed technique exhibited increased gapping and a lower load to failure than the current gold standard. The significance of these inferior results in clinical and functional settings has not been established. The proposed BF technique may be an appealing alternative to the SF technique given its low profile and potential to reduce the rates of future implant removal. Further studies on the long-term outcomes of patients treated through the BF technique are currently under way and will ultimately determine the utility of the proposed construct.

This paper will be judged for the Resident Writer’s Award.

Social media has become a prominent source of medical information for patients, including those with dermatologic conditions.^1,2 Physicians, patients, and pharmaceutical companies can use social media platforms to communicate with each other and share knowledge and advertisements related to conditions. Social media can influence patients’ perceptions of their disease and serve as a modality to acquire medical treatments.³ Furthermore, social media posts from illicit pharmacies can result in patients buying harmful medications without physician oversight.^4,5 Examination of the content and sources of social media posts related to acne may be useful in determining those who are primarily utilizing social media and for what purpose. The goal of this systematic review was to identify sources of acne-related social media posts to determine communication trends to gain a better understanding of the potential impact social media may have on patient care.

Methods

Social media posts were identified (May 2008 to May 2016) using the search terms acne and treatment across all social media platforms available through a commercial social media data aggregating software (Crimson Hexagon). Information from relevant posts was extracted and compiled into a spreadsheet that included the content, post date, social media platform, and hyperlink. To further analyze the data, the first 100 posts on acne treatment from May 2008 to May 2016 were selected and manually classified by the following types of communication: (1) patient-to-patient (eg, testimonies of patients’ medical experiences); (2) professional-to-patient (eg, clinical knowledge or experience provided by a medical provider and/or cited article in reference to relevant treatments); (3) pharmaceutical company–to-patient (eg, information from reputable drug manufacturers regarding drug activity and adverse effects); (4) illicit pharmacy–to-patient (eg, pharmacies with advertisements calling patients to buy a drug online or offering discrete shipping without a prescription)^4,5; or (5) other-to-patient (eg, posts that did not contain enough detail to be classified).

Results

Hundreds of thousands of social media posts discussing acne treatment were identified over the 8-year study period (Figure 1). The social media data aggregator extracted posts from various blogs, website comment sections, and online forums, as well as major social media platforms (ie, Facebook, Twitter, Google+, Tumblr). The first 100 posts selected for further analysis included 0 from 2008, 6 from 2009, 36 from 2010, 15 from 2011, 7 from 2012, 8 from 2013, 12 from 2014, 11 from 2015, and 5 from 2016. From this sample, 65 posts were considered to have an illicit source; conversely, 18 posts were from patients and 7 posts were from pharmaceutical companies (Figure 2).

Comment

This study demonstrated that discussion of acne treatment is prevalent in social media. Although our research underrepresents the social media interest in specific acne treatments, as only posts mentioning the terms acne and treatment were evaluated to gain insights into how social media platforms are being used by individuals with cutaneous disease. As such, even with this potential underrepresentation, our study demonstrated a high incidence of illicit marketing of prescription acne medications across multiple social media platforms (Figure 2). The sale of dermatologic pharmaceuticals (eg, isotretinoin) without a prescription is recognized by the US Government as a problem that is rapidly growing.^4,5 Illicit pharmacies pose as legitimate pharmacies that can provide prescription medications to consumers without a prescription.^5,6 The fact that these illicit pharmacy–to-patient posts were the most abundant in our study may speak to their relative success on social media platforms in encouraging patients to purchase prescription medications without physician oversight. These findings should concern health care providers, as the procurement of prescription medications without a prescription may put patients at risk.

Social media has become a prominent source of medical information for patients, including those with dermatologic conditions.^1,2 Physicians, patients, and pharmaceutical companies can use social media platforms to communicate with each other and share knowledge and advertisements related to conditions. Social media can influence patients’ perceptions of their disease and serve as a modality to acquire medical treatments.³ Furthermore, social media posts from illicit pharmacies can result in patients buying harmful medications without physician oversight.^4,5 Examination of the content and sources of social media posts related to acne may be useful in determining those who are primarily utilizing social media and for what purpose. The goal of this systematic review was to identify sources of acne-related social media posts to determine communication trends to gain a better understanding of the potential impact social media may have on patient care.

Methods

Social media posts were identified (May 2008 to May 2016) using the search terms acne and treatment across all social media platforms available through a commercial social media data aggregating software (Crimson Hexagon). Information from relevant posts was extracted and compiled into a spreadsheet that included the content, post date, social media platform, and hyperlink. To further analyze the data, the first 100 posts on acne treatment from May 2008 to May 2016 were selected and manually classified by the following types of communication: (1) patient-to-patient (eg, testimonies of patients’ medical experiences); (2) professional-to-patient (eg, clinical knowledge or experience provided by a medical provider and/or cited article in reference to relevant treatments); (3) pharmaceutical company–to-patient (eg, information from reputable drug manufacturers regarding drug activity and adverse effects); (4) illicit pharmacy–to-patient (eg, pharmacies with advertisements calling patients to buy a drug online or offering discrete shipping without a prescription)^4,5; or (5) other-to-patient (eg, posts that did not contain enough detail to be classified).

Results

Hundreds of thousands of social media posts discussing acne treatment were identified over the 8-year study period (Figure 1). The social media data aggregator extracted posts from various blogs, website comment sections, and online forums, as well as major social media platforms (ie, Facebook, Twitter, Google+, Tumblr). The first 100 posts selected for further analysis included 0 from 2008, 6 from 2009, 36 from 2010, 15 from 2011, 7 from 2012, 8 from 2013, 12 from 2014, 11 from 2015, and 5 from 2016. From this sample, 65 posts were considered to have an illicit source; conversely, 18 posts were from patients and 7 posts were from pharmaceutical companies (Figure 2).

Comment

This study demonstrated that discussion of acne treatment is prevalent in social media. Although our research underrepresents the social media interest in specific acne treatments, as only posts mentioning the terms acne and treatment were evaluated to gain insights into how social media platforms are being used by individuals with cutaneous disease. As such, even with this potential underrepresentation, our study demonstrated a high incidence of illicit marketing of prescription acne medications across multiple social media platforms (Figure 2). The sale of dermatologic pharmaceuticals (eg, isotretinoin) without a prescription is recognized by the US Government as a problem that is rapidly growing.^4,5 Illicit pharmacies pose as legitimate pharmacies that can provide prescription medications to consumers without a prescription.^5,6 The fact that these illicit pharmacy–to-patient posts were the most abundant in our study may speak to their relative success on social media platforms in encouraging patients to purchase prescription medications without physician oversight. These findings should concern health care providers, as the procurement of prescription medications without a prescription may put patients at risk.

Social media has become a prominent source of medical information for patients, including those with dermatologic conditions.^1,2 Physicians, patients, and pharmaceutical companies can use social media platforms to communicate with each other and share knowledge and advertisements related to conditions. Social media can influence patients’ perceptions of their disease and serve as a modality to acquire medical treatments.³ Furthermore, social media posts from illicit pharmacies can result in patients buying harmful medications without physician oversight.^4,5 Examination of the content and sources of social media posts related to acne may be useful in determining those who are primarily utilizing social media and for what purpose. The goal of this systematic review was to identify sources of acne-related social media posts to determine communication trends to gain a better understanding of the potential impact social media may have on patient care.

Methods

Social media posts were identified (May 2008 to May 2016) using the search terms acne and treatment across all social media platforms available through a commercial social media data aggregating software (Crimson Hexagon). Information from relevant posts was extracted and compiled into a spreadsheet that included the content, post date, social media platform, and hyperlink. To further analyze the data, the first 100 posts on acne treatment from May 2008 to May 2016 were selected and manually classified by the following types of communication: (1) patient-to-patient (eg, testimonies of patients’ medical experiences); (2) professional-to-patient (eg, clinical knowledge or experience provided by a medical provider and/or cited article in reference to relevant treatments); (3) pharmaceutical company–to-patient (eg, information from reputable drug manufacturers regarding drug activity and adverse effects); (4) illicit pharmacy–to-patient (eg, pharmacies with advertisements calling patients to buy a drug online or offering discrete shipping without a prescription)^4,5; or (5) other-to-patient (eg, posts that did not contain enough detail to be classified).

Results

Hundreds of thousands of social media posts discussing acne treatment were identified over the 8-year study period (Figure 1). The social media data aggregator extracted posts from various blogs, website comment sections, and online forums, as well as major social media platforms (ie, Facebook, Twitter, Google+, Tumblr). The first 100 posts selected for further analysis included 0 from 2008, 6 from 2009, 36 from 2010, 15 from 2011, 7 from 2012, 8 from 2013, 12 from 2014, 11 from 2015, and 5 from 2016. From this sample, 65 posts were considered to have an illicit source; conversely, 18 posts were from patients and 7 posts were from pharmaceutical companies (Figure 2).

Comment

This study demonstrated that discussion of acne treatment is prevalent in social media. Although our research underrepresents the social media interest in specific acne treatments, as only posts mentioning the terms acne and treatment were evaluated to gain insights into how social media platforms are being used by individuals with cutaneous disease. As such, even with this potential underrepresentation, our study demonstrated a high incidence of illicit marketing of prescription acne medications across multiple social media platforms (Figure 2). The sale of dermatologic pharmaceuticals (eg, isotretinoin) without a prescription is recognized by the US Government as a problem that is rapidly growing.^4,5 Illicit pharmacies pose as legitimate pharmacies that can provide prescription medications to consumers without a prescription.^5,6 The fact that these illicit pharmacy–to-patient posts were the most abundant in our study may speak to their relative success on social media platforms in encouraging patients to purchase prescription medications without physician oversight. These findings should concern health care providers, as the procurement of prescription medications without a prescription may put patients at risk.

Rosacea is a chronic skin condition that can be classified into 4 subtypes: erythematotelangiectatic, papulopustular, phymatous, and ocular. Erythematotelangiectatic rosacea is characterized by redness of the face and excessive blushing. Papulopustular rosacea is a more severe form of disease that is characterized by papules and pustules of the central face. If left untreated, these subtypes may progress to phymatous rosacea, which is characterized by skin thickening, fibrosis, and cosmetic disfigurement. Ocular rosacea is characterized by redness and irritation of the eyes.¹ Rosacea patients often are burdened with embarrassment, social anxiety, and psychiatric comorbidities.

The Patient Health Questionnaire 9 (PHQ-9) is a validated and reliable self-administered tool for diagnosis of depression and designation of depression severity. This instrument could prove useful in screening for depression in rosacea patients given the high incidence of psychiatric comorbidities in this patient population.² The PHQ-9 consists of 9 questions that assess for criteria used to define depressive disorders in the Diagnostic and Statistical Manual of Mental Disorders (Fourth Edition).³ The questionnaire is brief, easy to administer, and has 88% specificity and sensitivity.⁴

Other studies have evaluated the relationship between rosacea and psychiatric illness, but the PHQ-9 was not used as a screening tool.^7,8 Rosacea patients are at increased risk for having psychiatrist-diagnosed depression.⁵ In one assessment, a positive correlation between rosacea and psychiatric illness was noted using the Dermatology Life Quality Index, the rejection scale of the Questionnaire on Experience with Skin Complaints, and the German version of the Hospital Anxiety and Depression Scale.⁶ Interpretation of Rosacea Quality of Life and Dermatology Life Quality Index scores indicated that rosacea has a negative impact on quality of life.⁷

The purpose of this study was to examine the relationship between self-assessed rosacea severity scores and level of depression using the validated rosacea self-assessment tool and the PHQ-9 questionnaire, respectively.

Methods

Study Population
Study participants were adult patients from the Wake Forest Baptist Medical Center (Winston-Salem, North Carolina) dermatology clinic from January 2011 to December 2014 who had received a diagnosis of rosacea (International Classification of Diseases, Ninth Revision [ICD-9] code 695.3) from a Wake Forest dermatologist. Institutional review board approval was obtained prior to initiation of the study. Data collection occurred from October 2014 through February 2015. A total of 478 patients met criteria for participation in the study and were identified from the Wake Forest Baptist Hospital Transitional Data Warehouse and the electronic medical record. Because rosacea typically is not diagnosed in children and the data measures are not validated in children, this demographic group was excluded from participation.

Of 478 eligible patients who were invited to participate via mail or telephone, 46 completed the rosacea self-assessment tool and PHQ-9 survey in person. A total of 432 patients were mailed a presurvey recruitment letter notifying them that they would be receiving a survey in the mail unless they contacted the study team to decline participation. An email address and telephone number for the study team was provided. Twenty patients declined to participate in the study; surveys were then mailed to the remaining 412 patients. Sixteen of the mailed surveys were returned by the post office due to an incorrect address. A total of 195 surveys (149 through mail and 46 in person) were completed and analyzed. All survey respondents completed the validated rosacea self-assessment tool (Figure 1); of them, 183 completed the PHQ-9. Participants in this study received compensation for travel expenses and time.

Results

There is a direct relationship between rosacea severity and depression when comparing across the following levels of depression: nondepressed, minimal depression symptoms, minor depression, major depression (moderate), and major depression (severe)(P=.006; F=5.18; N=183)(Figures 2 and 3). There was no statistically significant difference in rosacea severity between the moderate and severe major depression groups.

Most patients reported they were nondepressed (68.9%). As measured by the PHQ-9, 31.1% of patients experienced some level of depression: 21.9% reported minimal depression symptoms, 7.1% reported minor depression, 1.1% reported major depression (moderate), and 1.1% reported major depression (severe)(Table).

Comment

There is a direct relationship between rosacea severity and level of depression. In our study, nearly one-third of patients reported some degree of depression. The reason for this correlation may be due to disease stigmatization and decreased quality of life due to the somatic symptoms of rosacea. Our study reinforced the results of other studies evaluating the psychosocial impact of rosacea.^8,9 Depression is associated with poor treatment adherence and poor outcomes in rosacea patients; therefore, depression may serve as an important outcome measure.¹⁰ The psychosocial impact of rosacea can be severe, but with disease improvement, there often is an improvement in the patient’s psychosocial status.⁷

There are several limitations to our study. The study population consisted of patients at a university dermatology clinic who may not be representative of patients in the general population; however, our hospital system does not require referral and provides care to a large percentage of the surrounding community.

Clinical implementation of the validated rosacea self-assessment tool and PHQ-9 may have several benefits. Patient-assessed rosacea severity and psychosocial impact obtained via use of these tools would provide physicians with information to fine-tune rosacea treatment regimens. Patients with the greatest social impact may require a more aggressive treatment approach. Early detection of depression in the rosacea population is important in informing treatment strategy and improving outcomes. Physicians should pay close attention to signs of depression in rosacea patients and determine if psychiatric treatment or referral for psychiatric evaluation is indicated. The correlation between rosacea and depression underscores the importance of treating this highly impactful disease; however, the low number of responders from the major depression (moderate) subgroup prevented us from making any strong conclusion about this specific subgroup.

Rosacea is a chronic skin condition that can be classified into 4 subtypes: erythematotelangiectatic, papulopustular, phymatous, and ocular. Erythematotelangiectatic rosacea is characterized by redness of the face and excessive blushing. Papulopustular rosacea is a more severe form of disease that is characterized by papules and pustules of the central face. If left untreated, these subtypes may progress to phymatous rosacea, which is characterized by skin thickening, fibrosis, and cosmetic disfigurement. Ocular rosacea is characterized by redness and irritation of the eyes.¹ Rosacea patients often are burdened with embarrassment, social anxiety, and psychiatric comorbidities.

The Patient Health Questionnaire 9 (PHQ-9) is a validated and reliable self-administered tool for diagnosis of depression and designation of depression severity. This instrument could prove useful in screening for depression in rosacea patients given the high incidence of psychiatric comorbidities in this patient population.² The PHQ-9 consists of 9 questions that assess for criteria used to define depressive disorders in the Diagnostic and Statistical Manual of Mental Disorders (Fourth Edition).³ The questionnaire is brief, easy to administer, and has 88% specificity and sensitivity.⁴

Other studies have evaluated the relationship between rosacea and psychiatric illness, but the PHQ-9 was not used as a screening tool.^7,8 Rosacea patients are at increased risk for having psychiatrist-diagnosed depression.⁵ In one assessment, a positive correlation between rosacea and psychiatric illness was noted using the Dermatology Life Quality Index, the rejection scale of the Questionnaire on Experience with Skin Complaints, and the German version of the Hospital Anxiety and Depression Scale.⁶ Interpretation of Rosacea Quality of Life and Dermatology Life Quality Index scores indicated that rosacea has a negative impact on quality of life.⁷

The purpose of this study was to examine the relationship between self-assessed rosacea severity scores and level of depression using the validated rosacea self-assessment tool and the PHQ-9 questionnaire, respectively.

Methods

Study Population
Study participants were adult patients from the Wake Forest Baptist Medical Center (Winston-Salem, North Carolina) dermatology clinic from January 2011 to December 2014 who had received a diagnosis of rosacea (International Classification of Diseases, Ninth Revision [ICD-9] code 695.3) from a Wake Forest dermatologist. Institutional review board approval was obtained prior to initiation of the study. Data collection occurred from October 2014 through February 2015. A total of 478 patients met criteria for participation in the study and were identified from the Wake Forest Baptist Hospital Transitional Data Warehouse and the electronic medical record. Because rosacea typically is not diagnosed in children and the data measures are not validated in children, this demographic group was excluded from participation.

Of 478 eligible patients who were invited to participate via mail or telephone, 46 completed the rosacea self-assessment tool and PHQ-9 survey in person. A total of 432 patients were mailed a presurvey recruitment letter notifying them that they would be receiving a survey in the mail unless they contacted the study team to decline participation. An email address and telephone number for the study team was provided. Twenty patients declined to participate in the study; surveys were then mailed to the remaining 412 patients. Sixteen of the mailed surveys were returned by the post office due to an incorrect address. A total of 195 surveys (149 through mail and 46 in person) were completed and analyzed. All survey respondents completed the validated rosacea self-assessment tool (Figure 1); of them, 183 completed the PHQ-9. Participants in this study received compensation for travel expenses and time.

Results

There is a direct relationship between rosacea severity and depression when comparing across the following levels of depression: nondepressed, minimal depression symptoms, minor depression, major depression (moderate), and major depression (severe)(P=.006; F=5.18; N=183)(Figures 2 and 3). There was no statistically significant difference in rosacea severity between the moderate and severe major depression groups.

Most patients reported they were nondepressed (68.9%). As measured by the PHQ-9, 31.1% of patients experienced some level of depression: 21.9% reported minimal depression symptoms, 7.1% reported minor depression, 1.1% reported major depression (moderate), and 1.1% reported major depression (severe)(Table).

Comment

There is a direct relationship between rosacea severity and level of depression. In our study, nearly one-third of patients reported some degree of depression. The reason for this correlation may be due to disease stigmatization and decreased quality of life due to the somatic symptoms of rosacea. Our study reinforced the results of other studies evaluating the psychosocial impact of rosacea.^8,9 Depression is associated with poor treatment adherence and poor outcomes in rosacea patients; therefore, depression may serve as an important outcome measure.¹⁰ The psychosocial impact of rosacea can be severe, but with disease improvement, there often is an improvement in the patient’s psychosocial status.⁷

There are several limitations to our study. The study population consisted of patients at a university dermatology clinic who may not be representative of patients in the general population; however, our hospital system does not require referral and provides care to a large percentage of the surrounding community.

Clinical implementation of the validated rosacea self-assessment tool and PHQ-9 may have several benefits. Patient-assessed rosacea severity and psychosocial impact obtained via use of these tools would provide physicians with information to fine-tune rosacea treatment regimens. Patients with the greatest social impact may require a more aggressive treatment approach. Early detection of depression in the rosacea population is important in informing treatment strategy and improving outcomes. Physicians should pay close attention to signs of depression in rosacea patients and determine if psychiatric treatment or referral for psychiatric evaluation is indicated. The correlation between rosacea and depression underscores the importance of treating this highly impactful disease; however, the low number of responders from the major depression (moderate) subgroup prevented us from making any strong conclusion about this specific subgroup.

Rosacea is a chronic skin condition that can be classified into 4 subtypes: erythematotelangiectatic, papulopustular, phymatous, and ocular. Erythematotelangiectatic rosacea is characterized by redness of the face and excessive blushing. Papulopustular rosacea is a more severe form of disease that is characterized by papules and pustules of the central face. If left untreated, these subtypes may progress to phymatous rosacea, which is characterized by skin thickening, fibrosis, and cosmetic disfigurement. Ocular rosacea is characterized by redness and irritation of the eyes.¹ Rosacea patients often are burdened with embarrassment, social anxiety, and psychiatric comorbidities.

The Patient Health Questionnaire 9 (PHQ-9) is a validated and reliable self-administered tool for diagnosis of depression and designation of depression severity. This instrument could prove useful in screening for depression in rosacea patients given the high incidence of psychiatric comorbidities in this patient population.² The PHQ-9 consists of 9 questions that assess for criteria used to define depressive disorders in the Diagnostic and Statistical Manual of Mental Disorders (Fourth Edition).³ The questionnaire is brief, easy to administer, and has 88% specificity and sensitivity.⁴

Other studies have evaluated the relationship between rosacea and psychiatric illness, but the PHQ-9 was not used as a screening tool.^7,8 Rosacea patients are at increased risk for having psychiatrist-diagnosed depression.⁵ In one assessment, a positive correlation between rosacea and psychiatric illness was noted using the Dermatology Life Quality Index, the rejection scale of the Questionnaire on Experience with Skin Complaints, and the German version of the Hospital Anxiety and Depression Scale.⁶ Interpretation of Rosacea Quality of Life and Dermatology Life Quality Index scores indicated that rosacea has a negative impact on quality of life.⁷

The purpose of this study was to examine the relationship between self-assessed rosacea severity scores and level of depression using the validated rosacea self-assessment tool and the PHQ-9 questionnaire, respectively.

Methods

Study Population
Study participants were adult patients from the Wake Forest Baptist Medical Center (Winston-Salem, North Carolina) dermatology clinic from January 2011 to December 2014 who had received a diagnosis of rosacea (International Classification of Diseases, Ninth Revision [ICD-9] code 695.3) from a Wake Forest dermatologist. Institutional review board approval was obtained prior to initiation of the study. Data collection occurred from October 2014 through February 2015. A total of 478 patients met criteria for participation in the study and were identified from the Wake Forest Baptist Hospital Transitional Data Warehouse and the electronic medical record. Because rosacea typically is not diagnosed in children and the data measures are not validated in children, this demographic group was excluded from participation.

Of 478 eligible patients who were invited to participate via mail or telephone, 46 completed the rosacea self-assessment tool and PHQ-9 survey in person. A total of 432 patients were mailed a presurvey recruitment letter notifying them that they would be receiving a survey in the mail unless they contacted the study team to decline participation. An email address and telephone number for the study team was provided. Twenty patients declined to participate in the study; surveys were then mailed to the remaining 412 patients. Sixteen of the mailed surveys were returned by the post office due to an incorrect address. A total of 195 surveys (149 through mail and 46 in person) were completed and analyzed. All survey respondents completed the validated rosacea self-assessment tool (Figure 1); of them, 183 completed the PHQ-9. Participants in this study received compensation for travel expenses and time.

Results

There is a direct relationship between rosacea severity and depression when comparing across the following levels of depression: nondepressed, minimal depression symptoms, minor depression, major depression (moderate), and major depression (severe)(P=.006; F=5.18; N=183)(Figures 2 and 3). There was no statistically significant difference in rosacea severity between the moderate and severe major depression groups.

Most patients reported they were nondepressed (68.9%). As measured by the PHQ-9, 31.1% of patients experienced some level of depression: 21.9% reported minimal depression symptoms, 7.1% reported minor depression, 1.1% reported major depression (moderate), and 1.1% reported major depression (severe)(Table).

Comment

There is a direct relationship between rosacea severity and level of depression. In our study, nearly one-third of patients reported some degree of depression. The reason for this correlation may be due to disease stigmatization and decreased quality of life due to the somatic symptoms of rosacea. Our study reinforced the results of other studies evaluating the psychosocial impact of rosacea.^8,9 Depression is associated with poor treatment adherence and poor outcomes in rosacea patients; therefore, depression may serve as an important outcome measure.¹⁰ The psychosocial impact of rosacea can be severe, but with disease improvement, there often is an improvement in the patient’s psychosocial status.⁷

There are several limitations to our study. The study population consisted of patients at a university dermatology clinic who may not be representative of patients in the general population; however, our hospital system does not require referral and provides care to a large percentage of the surrounding community.

Clinical implementation of the validated rosacea self-assessment tool and PHQ-9 may have several benefits. Patient-assessed rosacea severity and psychosocial impact obtained via use of these tools would provide physicians with information to fine-tune rosacea treatment regimens. Patients with the greatest social impact may require a more aggressive treatment approach. Early detection of depression in the rosacea population is important in informing treatment strategy and improving outcomes. Physicians should pay close attention to signs of depression in rosacea patients and determine if psychiatric treatment or referral for psychiatric evaluation is indicated. The correlation between rosacea and depression underscores the importance of treating this highly impactful disease; however, the low number of responders from the major depression (moderate) subgroup prevented us from making any strong conclusion about this specific subgroup.

ABSTRACT

An increasing interest focuses on the rates and risk factors for hospital readmission. However, little is known regarding the readmission following total shoulder arthroplasty (TSA). This study aims to determine the rates, risk factors, and reasons for hospital readmission following primary TSA. Patients undergoing TSA (anatomic or reverse) as part of the American College of Surgeons National Surgical Quality Improvement Program in 2011 to 2013 were identified. The rate of unplanned readmission to the hospital within 30 postoperative days was characterized. Using multivariate regression, demographic and comorbidity factors were tested for independent association with readmission. Finally, the reasons for readmission were characterized. A total of 3627 patients were identified. Among the admitted patients, 93 (2.56%) were readmitted within 30 days of surgery. The independent risk factors for readmission included old age (for age 60-69 years, relative risk [RR] = 1.6; for age 70-79 years, RR = 2.3; for age ≥80 years, RR = 23.1; P = .042), male sex (RR = 1.6, P = .025), anemia (RR = 1.9, P = .005), and dependent functional status (RR = 2.8, P = .012). The reasons for readmission were available for 84 of the 93 readmitted patients. The most common reasons for readmission comprised pneumonia (14 cases, 16.7%), dislocation (7 cases, 8.3%), pulmonary embolism (7 cases, 8.3%), and surgical site infection (6 cases, 7.1%). Unplanned readmission occurs following about 1 in 40 cases of TSA. The most common causes of readmission include pneumonia, dislocation, pulmonary embolism, and surgical site infection. Patients with old age, male sex, anemia, and dependent functional status are at higher risk for readmission and should be counseled and monitored accordingly.

Continue to: Total shoulder arthroplasty...

Total shoulder arthroplasty (TSA) is performed with increasing frequency in the United States and is considered to be cost-effective.^1-4 Following the procedure, patients generally achieve shoulder function and pain relief.^5-8 Despite the success of the procedure, the growing literature on TSA has also reported rates of complications between 3.6% and 25% of the treated patients.^9-16

In recent years, an increasing interest has focused on the rates and risk factors for unplanned hospital readmissions; these variables may not only reflect the quality of patient care but also result in considerable costs to the healthcare system. For instance, among Medicare patients, readmissions within 30 days of discharge occur in almost 20% of cases, costing $17.4 billion per year.¹⁷ Readmission rates increasingly factor into hospital performance metrics and reimbursement, including the Hospital Readmissions Reduction Program of the Patient Protection and Affordable Care Act that reduces Centers for Medicare and Medicaid Services payments to hospitals with high 30-day readmission rates.¹⁸

To date, only a few studies have evaluated readmission following TSA, with 30- to 90-day readmission rates ranging from 4.5% to 7.3%.^19-23 These studies comprised single institution series^20,22 and analyses of administrative databases.^19,21,23 Most studies have shown that readmission occurs more often for medical than surgical reasons, with surgical reasons most commonly including infection and dislocation.^19-23 However, only limited analyses have been conducted regarding risk factors for readmission.^21,23 To date and to our knowledge, no study has investigated reasons for readmission following TSA using nationwide data.

This study aims to determine the rates, risk factors, and reasons for hospital readmission following primary TSA in the United States using the American College of Surgeons National Surgical Quality Improvement Program (NSQIP) database.

METHODS

DATA SOURCE

The NSQIP database was utilized to address the study purpose. NSQIP is a nationwide prospective surgical registry established by the American College of Surgeons and reports data from academic and community hospitals across the United States.²⁴ Patients undertaking surgery at these centers are followed by the surgical clinical reviewers at the participating NSQIP sites prospectively for 30 days following the procedure to record complications including readmission. Preoperative and surgical data, such as demographics, medical comorbid diseases, and operative time, are also included. Previous studies have analyzed the complications of various orthopedic surgeries using the NSQIP data.^14,16,25-30

DATA COLLECTION

We retrospectively identified from NSQIP the patients who underwent primary TSA (anatomic or reverse) in 2013 to 2014. The timeframe 2013 to 2014 was used because NSQIP only began recording reasons for readmission in 2013. The inclusion criteria were as follows: Current Procedural Terminology (CPT) code for TSA (23472); preoperative diagnosis according to the International Classification of Diseases, Ninth Revision (ICD-9) codes 714.0, 715.11, 715.31, 715.91, 715.21, 715.89, 716.xx 718.xx, 719.xx, 726.x, 727.xx, and 733.41 (where x is a wild card digit); and no missing demographic, comorbidity, or outcome data. Anatomic and reverse TSA were analyzed together because they share the same CPT code, and the NSQIP database prevents searching by the ICD-9 procedure code.

The rate of unplanned readmission to the hospital within 30 postoperative days was characterized. The reasons for readmission in this 30-day period were only available in 2013 and were determined using the ICD-9 diagnosis codes. Patient demographics were recorded for use in identifying potential risk factors for readmission; the demographic data included sex, age, smoking status, body mass index (BMI), and comorbidities, including end-stage renal disease, dyspnea on exertion, congestive heart failure, diabetes mellitus, hypertension, and chronic obstructive pulmonary disease (COPD).

Continue to: Statistical analysis...

STATISTICAL ANALYSIS

Statistical analyses were performed using Stata version 13.1 (StataCorp). First, using bivariate and multivariate regression, demographic and comorbidity factors were tested for independent association with readmission to the hospital within 30 days of surgery. Second, among the readmitted patients, the reasons for readmission were tabulated. Of note, the reasons for readmission were only documented for the procedures performed in 2013. All tests were 2-tailed and conducted at an α level of 0.05.

RESTULTS

A total of 3627 TSA patients were identified. The mean age (± standard deviation) was 69.4 ± 9.5 years, 55.8% of patients were female, and mean BMI was 30.1 ± 7.0 years. Table 1 provides the additional demographic data. Of the 3627 included patients, 93 (2.56%) were readmitted within 30 days of surgery. The 95% confidence interval for the estimated rate of readmission reached 2.05% to 3.08%.

Table 1. Patient Population

Abbreviation: COPD, chronic obstructive pulmonary disease.

In the bivariate analyses (Table 2), the following factors were positively associated readmission: older age (60-69 years, relative risk [RR] = 1.6; 70-79 years, RR = 2.2; ≥80 years, RR = 3.3; P = .011), dependent functional status (RR = 2.9, P = .008), and anemia (RR = 2.2, P < .001).

Table 2. Bivariate Analysis of Risk Factors for Readmission

Abbreviations: CI, confidence interval; COPD, chronic obstructive pulmonary disease; RR, relative risk.

In the multivariate analyses (Table 3), the following factors were independent risk factors for readmission: older age (60-69 years, RR = 1.6; 70-79 years, RR = 2.3; ≥80 years, RR = 3.1; P =.027), male sex (RR = 1.6, P = .025), anemia (RR = 1.9, P = .005), and dependent functional status (RR = 2.8, P = .012). Interestingly, readmission showed no independent association with diabetes, dyspnea on exertion, BMI, COPD, hypertension, or current smoking status (P > .05 for each).

Table 3. Independent Risk Factors for Readmission on Multivariate Analysis

Abbreviations: CI, confidence interval; COPD, chronic obstructive pulmonary disease; RR, relative risk.

Continue to: Table 4...

The reasons for readmission were available for 84 of the 93 readmitted patients. The most common reasons for readmission included pneumonia (14 cases, 16.7%), dislocation (7 cases, 8.3%), pulmonary embolism (7 cases, 8.3%), and surgical site infection (6 cases, 7.1%) (Table 4).

Table 4. Reasons for Readmission

DISCUSSION

Our analysis of 3042 TSAs from the NSQIP database suggests that unplanned readmission to the hospital occurs following about 1 in 40 cases of TSA. The study also suggests that the most common reasons for readmission encompass pneumonia, dislocation, pulmonary embolism, and surgical site infection. Old age, male sex, anemia, and dependent functional status serve as risk factors for readmission, and patients with such factors should be counseled and monitored accordingly.

In recent years, an increasing emphasis has centered on reducing rates of hospital readmission, with programs such as the Hospital Readmissions Reduction Program of the Affordable Care Act cutting reimbursements for hospitals with high 30-day readmission rates.^17,18 To date, only a few studies have evaluated the reasons for readmission and readmission rates for TSA.^19-23 Initial reports consisted of single-institution TSA registry reviews. For example, Mahoney and colleagues²⁰ retrospectively evaluated shoulder arthroplasty procedures at their institution to document the readmission rates, finding a 5.9% readmission rate at 30 days. Readmission occurred more frequently in the first 30 days following discharge than in the 30- to 90-day period, with the most common reasons for readmission including medical complications, infection, and dislocation. Streubel and colleagues²² evaluated reoperation rates from their institution’s TSA registry, finding a 0.6% reoperation rate for primary TSA at 30 days and 1.5% for revision TSA. Instability and infection were the most common indications for reoperation. Our findings confirm these single-institution results and demonstrate their application to a nationwide sample of TSA, not just to high-volume academic centers. We similarly observed that dislocation, surgical site infection, and medical complications (mostly pneumonia and pulmonary embolism) were common causes of readmission, and that the 30-day readmission rate was about 1 in 40.

Several authors have since used statewide databases to analyze and determine risk factors for readmission following TSA. Lyman and colleagues¹⁹ used the New York State Database to show that higher hospital TSA surgical volume was associated with a lower rate of readmission when age and comorbidities were controlled for in a multivariate model. Old age was also associated with an increased readmission rate in their multivariate analysis, but comorbidities (as measured by the Charlson comorbidity index) presented a nonsignificant associative trend. These authors opted not to determine specific causes of readmission. Schairer and colleagues²¹ used State Inpatient Databases from 7 states, finding a 90-day readmission rate of 7.3%, 82% of which were due to medical complications and 18% of which were due to surgical complications (mostly infection and dislocation). Their multivariate regression revealed that male sex, reverse TSA, Medicaid insurance, patients discharged to inpatient rehabilitation or nursing facilities, medical comorbidities, and low-volume TSA hospitals were associated with readmission. Zhang and colleagues²³ used the same source to show that the 90-day readmission rate reached 14% for surgically treated proximal humerus fractures and higher for patients who underwent open reduction internal fixation, were female, were African American, were discharged to a nursing facility, possessed Medicaid insurance, or experienced medical comorbidities. Most recently, Basques and colleagues³¹ analyzed 1505 TSA cases from 2011 and 2012 in the NSQIP database, finding a 3.3% rate of readmission, with heart disease and hypertension as risk factors for readmission. Although the limitations of the NSQIP database prevented us from analyzing surgeon and hospital TSA volume or reverse vs anatomic TSA, our results confirm that the findings from statewide database studies apply to the United States nationwide NSQIP database. Old patient age, male sex, and medical comorbidities (anemia and dependent functional status) are independent risk factors for TSA readmission. We identified pneumonia, dislocation, pulmonary embolism, and surgical site infection as the most common reasons for readmission.

This study features several limitations that should be considered when interpreting the results. Anatomic and reverse TSA share a CPT code and were not separated using NSQIP data. A number of studies have reported that reverse TSA may place patients at higher risk for readmission;^20,21 however, confounding by other patient factors could play a role in this finding. The 30-day timeframe for readmission is another potential limitation; however, this timeframe is frequently used in other studies and is the relevant timeframe for the reduced reimbursement penalties from the Hospital Readmissions Reduction Program of the Affordable Care Act.¹⁸ Furthermore, the NSQIP database contains no information on surgeon or hospital TSA volume, which is a result of safeguards for patient and provider privacy. Additionally, readmission data were only available for 2011 to 2013, with causes of readmission only present in 2013. Although provided with such current information, we cannot analyze readmission trends over time, such as in response to the Affordable Care Act of 2010. Finally, although NSQIP surgical clinical reviewers strive to identify readmissions to other hospitals during their reviews of outpatient medical records, proportions of these readmissions are possibly missed. Therefore, our 30-day readmission rate may slightly underestimate the true rate.

Despite these limitations, the NSQIP database offers a unique opportunity to examine risk factors and reasons for readmission following TSA. The prior literature on readmission following TSA stemmed either from limited samples or administrative data, which feature known limitations.³² By utilizing a large, prospective, non-administrative, nationwide sample, our findings are probably both more reliable and generalizable to the country as a whole.

CONCLUSION

Unplanned readmission occurs following about 1 in 40 cases of TSA. The most common causes of readmission include pneumonia, dislocation, pulmonary embolism, and surgical site infection. Patients with old age, male sex, anemia, and dependent functional status are at a higher risk for readmission and should be counseled and monitored accordingly.

This paper will be judged for the Resident Writer’s Award.

ABSTRACT

An increasing interest focuses on the rates and risk factors for hospital readmission. However, little is known regarding the readmission following total shoulder arthroplasty (TSA). This study aims to determine the rates, risk factors, and reasons for hospital readmission following primary TSA. Patients undergoing TSA (anatomic or reverse) as part of the American College of Surgeons National Surgical Quality Improvement Program in 2011 to 2013 were identified. The rate of unplanned readmission to the hospital within 30 postoperative days was characterized. Using multivariate regression, demographic and comorbidity factors were tested for independent association with readmission. Finally, the reasons for readmission were characterized. A total of 3627 patients were identified. Among the admitted patients, 93 (2.56%) were readmitted within 30 days of surgery. The independent risk factors for readmission included old age (for age 60-69 years, relative risk [RR] = 1.6; for age 70-79 years, RR = 2.3; for age ≥80 years, RR = 23.1; P = .042), male sex (RR = 1.6, P = .025), anemia (RR = 1.9, P = .005), and dependent functional status (RR = 2.8, P = .012). The reasons for readmission were available for 84 of the 93 readmitted patients. The most common reasons for readmission comprised pneumonia (14 cases, 16.7%), dislocation (7 cases, 8.3%), pulmonary embolism (7 cases, 8.3%), and surgical site infection (6 cases, 7.1%). Unplanned readmission occurs following about 1 in 40 cases of TSA. The most common causes of readmission include pneumonia, dislocation, pulmonary embolism, and surgical site infection. Patients with old age, male sex, anemia, and dependent functional status are at higher risk for readmission and should be counseled and monitored accordingly.

Continue to: Total shoulder arthroplasty...

Total shoulder arthroplasty (TSA) is performed with increasing frequency in the United States and is considered to be cost-effective.^1-4 Following the procedure, patients generally achieve shoulder function and pain relief.^5-8 Despite the success of the procedure, the growing literature on TSA has also reported rates of complications between 3.6% and 25% of the treated patients.^9-16

In recent years, an increasing interest has focused on the rates and risk factors for unplanned hospital readmissions; these variables may not only reflect the quality of patient care but also result in considerable costs to the healthcare system. For instance, among Medicare patients, readmissions within 30 days of discharge occur in almost 20% of cases, costing $17.4 billion per year.¹⁷ Readmission rates increasingly factor into hospital performance metrics and reimbursement, including the Hospital Readmissions Reduction Program of the Patient Protection and Affordable Care Act that reduces Centers for Medicare and Medicaid Services payments to hospitals with high 30-day readmission rates.¹⁸

To date, only a few studies have evaluated readmission following TSA, with 30- to 90-day readmission rates ranging from 4.5% to 7.3%.^19-23 These studies comprised single institution series^20,22 and analyses of administrative databases.^19,21,23 Most studies have shown that readmission occurs more often for medical than surgical reasons, with surgical reasons most commonly including infection and dislocation.^19-23 However, only limited analyses have been conducted regarding risk factors for readmission.^21,23 To date and to our knowledge, no study has investigated reasons for readmission following TSA using nationwide data.

This study aims to determine the rates, risk factors, and reasons for hospital readmission following primary TSA in the United States using the American College of Surgeons National Surgical Quality Improvement Program (NSQIP) database.

METHODS

DATA SOURCE

The NSQIP database was utilized to address the study purpose. NSQIP is a nationwide prospective surgical registry established by the American College of Surgeons and reports data from academic and community hospitals across the United States.²⁴ Patients undertaking surgery at these centers are followed by the surgical clinical reviewers at the participating NSQIP sites prospectively for 30 days following the procedure to record complications including readmission. Preoperative and surgical data, such as demographics, medical comorbid diseases, and operative time, are also included. Previous studies have analyzed the complications of various orthopedic surgeries using the NSQIP data.^14,16,25-30

DATA COLLECTION

We retrospectively identified from NSQIP the patients who underwent primary TSA (anatomic or reverse) in 2013 to 2014. The timeframe 2013 to 2014 was used because NSQIP only began recording reasons for readmission in 2013. The inclusion criteria were as follows: Current Procedural Terminology (CPT) code for TSA (23472); preoperative diagnosis according to the International Classification of Diseases, Ninth Revision (ICD-9) codes 714.0, 715.11, 715.31, 715.91, 715.21, 715.89, 716.xx 718.xx, 719.xx, 726.x, 727.xx, and 733.41 (where x is a wild card digit); and no missing demographic, comorbidity, or outcome data. Anatomic and reverse TSA were analyzed together because they share the same CPT code, and the NSQIP database prevents searching by the ICD-9 procedure code.

The rate of unplanned readmission to the hospital within 30 postoperative days was characterized. The reasons for readmission in this 30-day period were only available in 2013 and were determined using the ICD-9 diagnosis codes. Patient demographics were recorded for use in identifying potential risk factors for readmission; the demographic data included sex, age, smoking status, body mass index (BMI), and comorbidities, including end-stage renal disease, dyspnea on exertion, congestive heart failure, diabetes mellitus, hypertension, and chronic obstructive pulmonary disease (COPD).

Continue to: Statistical analysis...

STATISTICAL ANALYSIS

Statistical analyses were performed using Stata version 13.1 (StataCorp). First, using bivariate and multivariate regression, demographic and comorbidity factors were tested for independent association with readmission to the hospital within 30 days of surgery. Second, among the readmitted patients, the reasons for readmission were tabulated. Of note, the reasons for readmission were only documented for the procedures performed in 2013. All tests were 2-tailed and conducted at an α level of 0.05.

RESTULTS

A total of 3627 TSA patients were identified. The mean age (± standard deviation) was 69.4 ± 9.5 years, 55.8% of patients were female, and mean BMI was 30.1 ± 7.0 years. Table 1 provides the additional demographic data. Of the 3627 included patients, 93 (2.56%) were readmitted within 30 days of surgery. The 95% confidence interval for the estimated rate of readmission reached 2.05% to 3.08%.

Table 1. Patient Population

Abbreviation: COPD, chronic obstructive pulmonary disease.

In the bivariate analyses (Table 2), the following factors were positively associated readmission: older age (60-69 years, relative risk [RR] = 1.6; 70-79 years, RR = 2.2; ≥80 years, RR = 3.3; P = .011), dependent functional status (RR = 2.9, P = .008), and anemia (RR = 2.2, P < .001).

Table 2. Bivariate Analysis of Risk Factors for Readmission

Abbreviations: CI, confidence interval; COPD, chronic obstructive pulmonary disease; RR, relative risk.

In the multivariate analyses (Table 3), the following factors were independent risk factors for readmission: older age (60-69 years, RR = 1.6; 70-79 years, RR = 2.3; ≥80 years, RR = 3.1; P =.027), male sex (RR = 1.6, P = .025), anemia (RR = 1.9, P = .005), and dependent functional status (RR = 2.8, P = .012). Interestingly, readmission showed no independent association with diabetes, dyspnea on exertion, BMI, COPD, hypertension, or current smoking status (P > .05 for each).

Table 3. Independent Risk Factors for Readmission on Multivariate Analysis

Abbreviations: CI, confidence interval; COPD, chronic obstructive pulmonary disease; RR, relative risk.

Continue to: Table 4...

The reasons for readmission were available for 84 of the 93 readmitted patients. The most common reasons for readmission included pneumonia (14 cases, 16.7%), dislocation (7 cases, 8.3%), pulmonary embolism (7 cases, 8.3%), and surgical site infection (6 cases, 7.1%) (Table 4).

Table 4. Reasons for Readmission

DISCUSSION

Our analysis of 3042 TSAs from the NSQIP database suggests that unplanned readmission to the hospital occurs following about 1 in 40 cases of TSA. The study also suggests that the most common reasons for readmission encompass pneumonia, dislocation, pulmonary embolism, and surgical site infection. Old age, male sex, anemia, and dependent functional status serve as risk factors for readmission, and patients with such factors should be counseled and monitored accordingly.

In recent years, an increasing emphasis has centered on reducing rates of hospital readmission, with programs such as the Hospital Readmissions Reduction Program of the Affordable Care Act cutting reimbursements for hospitals with high 30-day readmission rates.^17,18 To date, only a few studies have evaluated the reasons for readmission and readmission rates for TSA.^19-23 Initial reports consisted of single-institution TSA registry reviews. For example, Mahoney and colleagues²⁰ retrospectively evaluated shoulder arthroplasty procedures at their institution to document the readmission rates, finding a 5.9% readmission rate at 30 days. Readmission occurred more frequently in the first 30 days following discharge than in the 30- to 90-day period, with the most common reasons for readmission including medical complications, infection, and dislocation. Streubel and colleagues²² evaluated reoperation rates from their institution’s TSA registry, finding a 0.6% reoperation rate for primary TSA at 30 days and 1.5% for revision TSA. Instability and infection were the most common indications for reoperation. Our findings confirm these single-institution results and demonstrate their application to a nationwide sample of TSA, not just to high-volume academic centers. We similarly observed that dislocation, surgical site infection, and medical complications (mostly pneumonia and pulmonary embolism) were common causes of readmission, and that the 30-day readmission rate was about 1 in 40.

Several authors have since used statewide databases to analyze and determine risk factors for readmission following TSA. Lyman and colleagues¹⁹ used the New York State Database to show that higher hospital TSA surgical volume was associated with a lower rate of readmission when age and comorbidities were controlled for in a multivariate model. Old age was also associated with an increased readmission rate in their multivariate analysis, but comorbidities (as measured by the Charlson comorbidity index) presented a nonsignificant associative trend. These authors opted not to determine specific causes of readmission. Schairer and colleagues²¹ used State Inpatient Databases from 7 states, finding a 90-day readmission rate of 7.3%, 82% of which were due to medical complications and 18% of which were due to surgical complications (mostly infection and dislocation). Their multivariate regression revealed that male sex, reverse TSA, Medicaid insurance, patients discharged to inpatient rehabilitation or nursing facilities, medical comorbidities, and low-volume TSA hospitals were associated with readmission. Zhang and colleagues²³ used the same source to show that the 90-day readmission rate reached 14% for surgically treated proximal humerus fractures and higher for patients who underwent open reduction internal fixation, were female, were African American, were discharged to a nursing facility, possessed Medicaid insurance, or experienced medical comorbidities. Most recently, Basques and colleagues³¹ analyzed 1505 TSA cases from 2011 and 2012 in the NSQIP database, finding a 3.3% rate of readmission, with heart disease and hypertension as risk factors for readmission. Although the limitations of the NSQIP database prevented us from analyzing surgeon and hospital TSA volume or reverse vs anatomic TSA, our results confirm that the findings from statewide database studies apply to the United States nationwide NSQIP database. Old patient age, male sex, and medical comorbidities (anemia and dependent functional status) are independent risk factors for TSA readmission. We identified pneumonia, dislocation, pulmonary embolism, and surgical site infection as the most common reasons for readmission.

This study features several limitations that should be considered when interpreting the results. Anatomic and reverse TSA share a CPT code and were not separated using NSQIP data. A number of studies have reported that reverse TSA may place patients at higher risk for readmission;^20,21 however, confounding by other patient factors could play a role in this finding. The 30-day timeframe for readmission is another potential limitation; however, this timeframe is frequently used in other studies and is the relevant timeframe for the reduced reimbursement penalties from the Hospital Readmissions Reduction Program of the Affordable Care Act.¹⁸ Furthermore, the NSQIP database contains no information on surgeon or hospital TSA volume, which is a result of safeguards for patient and provider privacy. Additionally, readmission data were only available for 2011 to 2013, with causes of readmission only present in 2013. Although provided with such current information, we cannot analyze readmission trends over time, such as in response to the Affordable Care Act of 2010. Finally, although NSQIP surgical clinical reviewers strive to identify readmissions to other hospitals during their reviews of outpatient medical records, proportions of these readmissions are possibly missed. Therefore, our 30-day readmission rate may slightly underestimate the true rate.

Despite these limitations, the NSQIP database offers a unique opportunity to examine risk factors and reasons for readmission following TSA. The prior literature on readmission following TSA stemmed either from limited samples or administrative data, which feature known limitations.³² By utilizing a large, prospective, non-administrative, nationwide sample, our findings are probably both more reliable and generalizable to the country as a whole.

CONCLUSION

Unplanned readmission occurs following about 1 in 40 cases of TSA. The most common causes of readmission include pneumonia, dislocation, pulmonary embolism, and surgical site infection. Patients with old age, male sex, anemia, and dependent functional status are at a higher risk for readmission and should be counseled and monitored accordingly.

This paper will be judged for the Resident Writer’s Award.

ABSTRACT

An increasing interest focuses on the rates and risk factors for hospital readmission. However, little is known regarding the readmission following total shoulder arthroplasty (TSA). This study aims to determine the rates, risk factors, and reasons for hospital readmission following primary TSA. Patients undergoing TSA (anatomic or reverse) as part of the American College of Surgeons National Surgical Quality Improvement Program in 2011 to 2013 were identified. The rate of unplanned readmission to the hospital within 30 postoperative days was characterized. Using multivariate regression, demographic and comorbidity factors were tested for independent association with readmission. Finally, the reasons for readmission were characterized. A total of 3627 patients were identified. Among the admitted patients, 93 (2.56%) were readmitted within 30 days of surgery. The independent risk factors for readmission included old age (for age 60-69 years, relative risk [RR] = 1.6; for age 70-79 years, RR = 2.3; for age ≥80 years, RR = 23.1; P = .042), male sex (RR = 1.6, P = .025), anemia (RR = 1.9, P = .005), and dependent functional status (RR = 2.8, P = .012). The reasons for readmission were available for 84 of the 93 readmitted patients. The most common reasons for readmission comprised pneumonia (14 cases, 16.7%), dislocation (7 cases, 8.3%), pulmonary embolism (7 cases, 8.3%), and surgical site infection (6 cases, 7.1%). Unplanned readmission occurs following about 1 in 40 cases of TSA. The most common causes of readmission include pneumonia, dislocation, pulmonary embolism, and surgical site infection. Patients with old age, male sex, anemia, and dependent functional status are at higher risk for readmission and should be counseled and monitored accordingly.

Continue to: Total shoulder arthroplasty...

Total shoulder arthroplasty (TSA) is performed with increasing frequency in the United States and is considered to be cost-effective.^1-4 Following the procedure, patients generally achieve shoulder function and pain relief.^5-8 Despite the success of the procedure, the growing literature on TSA has also reported rates of complications between 3.6% and 25% of the treated patients.^9-16

In recent years, an increasing interest has focused on the rates and risk factors for unplanned hospital readmissions; these variables may not only reflect the quality of patient care but also result in considerable costs to the healthcare system. For instance, among Medicare patients, readmissions within 30 days of discharge occur in almost 20% of cases, costing $17.4 billion per year.¹⁷ Readmission rates increasingly factor into hospital performance metrics and reimbursement, including the Hospital Readmissions Reduction Program of the Patient Protection and Affordable Care Act that reduces Centers for Medicare and Medicaid Services payments to hospitals with high 30-day readmission rates.¹⁸

To date, only a few studies have evaluated readmission following TSA, with 30- to 90-day readmission rates ranging from 4.5% to 7.3%.^19-23 These studies comprised single institution series^20,22 and analyses of administrative databases.^19,21,23 Most studies have shown that readmission occurs more often for medical than surgical reasons, with surgical reasons most commonly including infection and dislocation.^19-23 However, only limited analyses have been conducted regarding risk factors for readmission.^21,23 To date and to our knowledge, no study has investigated reasons for readmission following TSA using nationwide data.

This study aims to determine the rates, risk factors, and reasons for hospital readmission following primary TSA in the United States using the American College of Surgeons National Surgical Quality Improvement Program (NSQIP) database.

METHODS

DATA SOURCE

The NSQIP database was utilized to address the study purpose. NSQIP is a nationwide prospective surgical registry established by the American College of Surgeons and reports data from academic and community hospitals across the United States.²⁴ Patients undertaking surgery at these centers are followed by the surgical clinical reviewers at the participating NSQIP sites prospectively for 30 days following the procedure to record complications including readmission. Preoperative and surgical data, such as demographics, medical comorbid diseases, and operative time, are also included. Previous studies have analyzed the complications of various orthopedic surgeries using the NSQIP data.^14,16,25-30

DATA COLLECTION

We retrospectively identified from NSQIP the patients who underwent primary TSA (anatomic or reverse) in 2013 to 2014. The timeframe 2013 to 2014 was used because NSQIP only began recording reasons for readmission in 2013. The inclusion criteria were as follows: Current Procedural Terminology (CPT) code for TSA (23472); preoperative diagnosis according to the International Classification of Diseases, Ninth Revision (ICD-9) codes 714.0, 715.11, 715.31, 715.91, 715.21, 715.89, 716.xx 718.xx, 719.xx, 726.x, 727.xx, and 733.41 (where x is a wild card digit); and no missing demographic, comorbidity, or outcome data. Anatomic and reverse TSA were analyzed together because they share the same CPT code, and the NSQIP database prevents searching by the ICD-9 procedure code.

The rate of unplanned readmission to the hospital within 30 postoperative days was characterized. The reasons for readmission in this 30-day period were only available in 2013 and were determined using the ICD-9 diagnosis codes. Patient demographics were recorded for use in identifying potential risk factors for readmission; the demographic data included sex, age, smoking status, body mass index (BMI), and comorbidities, including end-stage renal disease, dyspnea on exertion, congestive heart failure, diabetes mellitus, hypertension, and chronic obstructive pulmonary disease (COPD).

Continue to: Statistical analysis...

STATISTICAL ANALYSIS

Statistical analyses were performed using Stata version 13.1 (StataCorp). First, using bivariate and multivariate regression, demographic and comorbidity factors were tested for independent association with readmission to the hospital within 30 days of surgery. Second, among the readmitted patients, the reasons for readmission were tabulated. Of note, the reasons for readmission were only documented for the procedures performed in 2013. All tests were 2-tailed and conducted at an α level of 0.05.

RESTULTS

A total of 3627 TSA patients were identified. The mean age (± standard deviation) was 69.4 ± 9.5 years, 55.8% of patients were female, and mean BMI was 30.1 ± 7.0 years. Table 1 provides the additional demographic data. Of the 3627 included patients, 93 (2.56%) were readmitted within 30 days of surgery. The 95% confidence interval for the estimated rate of readmission reached 2.05% to 3.08%.

Table 1. Patient Population

Abbreviation: COPD, chronic obstructive pulmonary disease.

In the bivariate analyses (Table 2), the following factors were positively associated readmission: older age (60-69 years, relative risk [RR] = 1.6; 70-79 years, RR = 2.2; ≥80 years, RR = 3.3; P = .011), dependent functional status (RR = 2.9, P = .008), and anemia (RR = 2.2, P < .001).

Table 2. Bivariate Analysis of Risk Factors for Readmission

Abbreviations: CI, confidence interval; COPD, chronic obstructive pulmonary disease; RR, relative risk.

In the multivariate analyses (Table 3), the following factors were independent risk factors for readmission: older age (60-69 years, RR = 1.6; 70-79 years, RR = 2.3; ≥80 years, RR = 3.1; P =.027), male sex (RR = 1.6, P = .025), anemia (RR = 1.9, P = .005), and dependent functional status (RR = 2.8, P = .012). Interestingly, readmission showed no independent association with diabetes, dyspnea on exertion, BMI, COPD, hypertension, or current smoking status (P > .05 for each).

Table 3. Independent Risk Factors for Readmission on Multivariate Analysis

Abbreviations: CI, confidence interval; COPD, chronic obstructive pulmonary disease; RR, relative risk.

Continue to: Table 4...

The reasons for readmission were available for 84 of the 93 readmitted patients. The most common reasons for readmission included pneumonia (14 cases, 16.7%), dislocation (7 cases, 8.3%), pulmonary embolism (7 cases, 8.3%), and surgical site infection (6 cases, 7.1%) (Table 4).

Table 4. Reasons for Readmission