Patient-centered care, defined by the Institute of Medicine as “health care that establishes a partnership among practitioners, patients, and their families to ensure that decisions respect patients’ wants, needs and preferences and that patients have the education and support they need to make decisions and participate in their own care,” has been recognized as an important factor in improving care transitions after discharge from the hospital.¹ Previous efforts to improve the discharge process for hospitalized patients and reduce avoidable readmissions have focused on improving systems surrounding the patient, such as by increasing the availability of outpatient follow-up or standardizing communication between the inpatient and outpatient care teams.^1,2 In fact, successful programs such as Project BOOST and the Care Transitions Interventions™ provide healthcare institutions with a “bundle” of evidence-based transitional care guidelines for discharge: they provide postdischarge transition coaches, assistance with medication self-management, timely follow-up tips, and improved patient records in order to improve postdischarge outcomes.^3,4 Successful interventions, however, may not provide more services, but also engage the patient in their own care.^5,6 The impact of engaging the patient in his or her own care by providing patient-friendly discharge instructions alone, however, is unknown.

A patient-centered discharge may use tools that were designed with patients, or may involve engaging patients in an interactive process of reviewing discharge instructions and empowering them to manage aspects of their own care after leaving the hospital. This endeavour may lead to more effective use of discharge instructions and reduce the need for additional or more intensive (and costly) interventions. For example, a patient-centered discharge tool could include an educational intervention that uses the “teach-back” method, in which patients are asked to restate in their own words what they thought they heard, or in which staff use additional media or a visual design tool meant to enhance comprehension of discharge instructions.^6,7 Visual aids and the use of larger fonts are particularly useful design elements for improving comprehension among non-English speakers and patients with low health literacy, who tend to have poorer recall of instructions.^8-10 What may constitute essential design elements to include in a discharge instruction tool, however, is not clear.

Moreover, whether the use of discharge tools with a specific focus on patient engagement may improve postdischarge outcomes is not known. Particularly, the ability of patient-centered discharge tools to improve outcomes beyond comprehension such as self-management, adherence to discharge instructions, a reduction in unplanned visits, and a reduction in mortality has not been studied systematically. The objective of this systematic review was to review the literature on discharge instruction tools with a focus on patient engagement and their impact among hospitalized patients.

The Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) Statement was followed as a guideline for reporting throughout this review.¹¹

Data Sources

A literature search was undertaken using the following databases from January 1994 or their inception date to May 2014: Medline, Embase, SIGLE, HTA, Bioethics, ASSIA, Psych Lit, CINAHL, Cochrane Library, EconLit, ERIC, and BioMed Central. We also searched relevant design-focused journals such as Design Issues, Journal of Design Research, Information Design Journal, Innovation, Design Studies, and International Journal of Design, as well as reference lists from studies obtained by electronic searching. The following key words and combination of key words were used with the assistance of a medical librarian: patient discharge, patient-centered discharge, patient-centered design, design thinking, user based design, patient education, discharge summary, education. Additional search terms were added when identified from relevant articles (Appendix).

Inclusion Criteria

We included all English-language studies with patients admitted to the hospital irrespective of age, sex, or medical condition, which included a control group or time period and which measured patient outcomes within 3 months of discharge. The 3-month period after discharge is often cited as a time when outcomes could reasonably be associated with an intervention at discharge.²

Exclusion Criteria

Studies that did not have clear implementation of a patient-centered tool, a control group, or those whose tool was used in the emergency department or as an outpatient were excluded. Studies that included postdischarge tools such as home visits or telephone calls were excluded unless independent effects of the predischarge interventions were measured. Studies with outcomes reported after 3 months were excluded unless outcomes before 3 months were also clearly noted.

All searches were entered into Endnote and duplicates were removed. A 2-stage inclusion process was used. Titles and abstracts of articles were first screened for meeting inclusion and exclusion criteria by 1 reviewer. A second reviewer independently checked a 10% random sample of all the abstracts that met the initial screening criteria. If the agreement to exclude studies was less than 95%, criteria were reviewed before checking the rest of the 90% sample. In the second stage, 2 independent reviewers examined paper copies of the full articles selected in the first stage. Disagreement between reviewers was resolved by discussion or a third reviewer if no agreement could be reached.

Data Analysis and Synthesis

The following information was extracted from the full reference: type of study, population studied, control group or time period, tool used, and outcomes measured. Based on the National Health Care Quality report’s priorities and goals on patient and/or family engagement during transitions of care, educational tools were further described based on method of teaching, involvement of the care team, involvement of the patient in the design or delivery of the tool, and/or the use of visual aids.¹² All primary outcomes were classified according to 3 categories: improved knowledge/comprehension, patient experience (patient satisfaction, self-management/efficacy such as functional status, both physical and mental), and health outcomes (unscheduled visits or readmissions, adherence with medications, diet, exercise, or follow-up, and mortality).

No quantitative pooling of results or meta-analysis was done given the variability and heterogeneity of studies reviewed. However, following guidelines for Effect Practice and Organisation of Care (EPOC) Risk of Bias criteria,¹³ studies that had a higher risk of bias such as uncontrolled before-after studies or studies with only 1 intervention or control site (historical controls, eg) were excluded from the final review because of the difficulties in attributing causation. Only primary outcomes were reported in order to minimize type II errors.

Our search revealed a total of 3699 studies after duplicates had been removed (Figure). A total of 714 references were included after initial review by title and abstract and 30 studies after full-text review. Agreement on a 10% random sample of all abstracts and full text was 79% (k=0.58) and 86% (k=0.72), respectively. Discussion was needed for fewer than 100 references, and agreement was subsequently reached for 100%.

There were 22 randomized controlled trials and 8 nonrandomized studies (5 nonrandomized controlled trials and 3 controlled before-after studies). Most of these studies were conducted in the United States (13/30 studies), followed by other European countries (5 studies), and the United Kingdom (4 studies). A large number of studies were conducted among patients with cardiovascular disease or risk factors (10 studies), followed by postsurgical patients such as coronary artery bypass graft surgery or orthopaedic surgery (5 studies). Five of 30 studies were conducted among individuals older than 65 years. Most studies excluded patients who did not speak English or the country’s official language; only 3 studies included patients with limited literacy, patients who spoke other languages, or caregivers if the patients could not communicate.

Most studies tested the impact of educational discharge interventions (28 of 30 studies) (Table 1). Quite often, it was a member of the research team who carried out the patient education. Only 3 studies involved multiple members of the care team in designing or reviewing the discharge tool with the patient. Almost half (12 studies) targeted multiple aspects of postdischarge care, including medications and side effects, signs and symptoms to consider, plans for follow-up, dietary restrictions, and/or exercise modifications. Many (19 studies) provided education using one-on-one teaching in association with a discharge tool, accompanied by a written handout (13 studies), audiotape (2 studies), or video (3 studies). While 13 studies had patients involved in creating what content was discussed and 14 studies had patients involved in the delivery of the tool, only 6 studies had patients involved in both design and delivery of the tool. Nine studies also used visual aids such as pictures, larger font, or use of a tool enhanced for patients with language barriers or limited health literacy.

Our systematic review found 30 studies that engaged patients during the design or the delivery of a discharge instruction tool and that tested the effect of the tool on postdischarge outcomes.^6-10,14–38 Our review suggests that there is sufficient evidence that patient-centered discharge tools improve comprehension. However, evidence is currently insufficient to determine if patient-centered tools improve adherence with discharge instructions. Moreover, though limited studies show promising results, more studies are needed to determine if patient engagement improves self-efficacy and healthcare utilization after discharge.

A major limitation of current studies is the variability in the level of patient engagement in tool design or delivery. Patients were involved in the design mostly through targeted development of a discharge management plan and the delivery by encouraging them to ask questions. Few studies involved patients in the design of the tool such that patients were responsible for coming up with content that was of interest to them. The few that did, often with the additional use of video media, demonstrated significant outcomes. Only a minority of studies used an interactive process to assess understanding such as “teach-back” or maximize patient comprehension such as visual aids. Even fewer studies engaged patients in both developing the discharge tool and providing discharge instructions.

Several previous studies have demonstrated that most complications after discharge are the result of ineffective communication, which can be exacerbated by lack of fluency in English or by limited health literacy.^2,39-43 As a result, poor understanding of discharge instructions by patients and their caregivers can create an important care gap.⁴⁴ Therefore, the use of patient-centered tools to engage patients at discharge in their own care is needed. How to engage patients consistently and effectively is perhaps less evident, as demonstrated in this review of the literature in which different levels of patient engagement were found. Many of the tools tested placed attention on patient education, sometimes in the context of bundled care along with home visits or follow-up, all of which can require extensive resources and time. Providing patients with information that the patients themselves state is of value may be the easiest refinement to a discharge educational tool, although this was surprisingly uncommon.^{6,9,10,17,23,33,37} Only 2 studies were found that engaged patients in the initial stage of design of the discharge tool, by incorporating information of interest to them.^23,32 For example, a study testing the impact of a computer-generated written education package on poststroke outcomes designed the information by asking patients to identify which topics they would like to receive information about (along with the amount of information and font size).²³ Secondly, although most of the discharge tools reviewed included the use of one-on-one teaching and the use of media such as patient handouts, these tools were often used in such a way that patients were passive recipients. In fact, studies that used additional video media that incorporated personalized content were the most likely to demonstrate positive outcomes.^17,34 The next level of patient engagement may therefore be to involve the patient as an interactive partner when delivering the tool in order to empower patients to self-care. For example, 1 study designed a structured education program by first assessing lifestyle risk factors related to hypertension that were modifiable along with preconceived notions through open-ended questions during a one-on-one interview.³⁷ Patients were subsequently educated on any knowledge deficits regarding the management of their lifestyle. Another level of patient engagement may be to use visual aids during discussions, as a well-known complement to verbal instructions.^45,46 For example, in a controlled study that randomized a ward of elderly patients with 4 or more prescriptions to predischarge counseling, the counseling session aimed to review reasons for their prescriptions along with corresponding side effects, doses, and dosage times with the help of a medicine reminder card. Other uses of visual aid tools identified in our review included the use of pictograms or illustrations or, at minimum, attention to font size.^{7,8,16,29,33,35} In the absence of a visual aid, asking the patient to repeat or demonstrate what was just communicated can be used to assess the amount of information retained.^18,33

An important result discovered in our review of the literature was also the lack of studies that tested the impact of discharge tools on usability of discharge information once at home. Conducting an evaluation of the benefits to patients after discharge can help objectify vague outcomes like health gains or qualify benefits in patient’s views. This might also explain why many studies with documented patient engagement at the time of discharge were able to demonstrate improvements in comprehension but not adherence to instructions. Although patients and caregivers may understand the information, this comprehension does not necessarily mean they will find the information useful or adhere to it once at home. For example, in 1 study, patients discharged with at least 1 medication were randomized to a structured discharge interview during which the treatment plan was reviewed verbally and questions clarified along with a visually enhanced treatment card.²⁶ Although knowledge of medications increased, no effect was found on adherence at 1 week postdischarge. However, use of the treatment card at home was not assessed. Similarly, another study tested the effect of an individualized video of exercises and failed to find a difference in patient adherence at 4 weeks.²⁸ The authors suggested that the lack of benefit may have been because patients were not using the video once at home. This is in contrast to 2 studies that involved patients in their own care by requiring them to request their medication as part of a self-medication tool predischarge.^16,30 Patients were engaged in the process such that increasing independence was given to patients based on their demonstration of understanding and adherence to their treatment while still in the hospital, a learning tool that can be applied once at home. Feeling knowledgeable and involved, as others have suggested, may be the intermediary outcomes that led to improved adherence.⁴⁷ It is also possible that adherence to discharge instructions may vary based on complexity of the information provided, such that instructions focusing solely on medication use may require less patient engagement than discharge instructions that include information on medications, diet, exercise modifications, and follow-up.⁴⁸

Our review has a few limitations. Previous systematic reviews have demonstrated that bundled discharge interventions that include patient-centered education have a positive effect on outcomes postdischarge.^2,5 However, we sought to describe and study the individual and distinct impact of patient engagement in the creation and delivery of discharge tools on outcomes postdischarge. We hoped that this may provide others with key information regarding elements of patient engagement that were particularly useful when designing a new discharge tool. The variability of the studies we identified, however, made it difficult to ascertain what level of patient engagement is required to observe improvements in health outcomes. It is also possible that a higher level of patient engagement may have been used but not described in the studies we reviewed. As only primary outcomes were included, we may have underestimated the effect of patient-centered discharge tools on outcomes that were reported as secondary outcomes. As we were interested in reviewing as many studies of patient-centered discharge tools as possible, we did not assess the quality of the studies and cannot comment on the role of bias in these studies. However, we excluded studies with study designs known to have the highest risk of bias. Lastly, we also cannot comment on whether patient-centered tools may have an effect on outcomes more than 3 months after a hospital discharge. However, several studies included in this review suggest a sustained effect beyond this time period.^8,25,32,37

Patient-centered discharge tools in which patients were engaged in the design or the delivery were found to improve comprehension of but not adherence with discharge instructions. The perceived lack of improved adherence may be due to a lack of studies that measured the usefulness and utilization of information for patients once at home. There was also substantial variability in the extent of patient involvement in designing the style and content of information provided to patients at discharge, as well as the extent of patient engagement when receiving discharge instructions. Future studies would benefit from detailing the level of patient engagement needed in designing and delivery of discharge tools. This information may lead to the discovery of barriers and facilitators to utilization of discharge information once at home and lead to a better understanding of the patient’s journey from hospital to home and onwards.

C.M.B. and this work were funded by a CIHR Canadian Patient Safety Institute Chair in Patient Safety and Continuity of Care. Funding was provided to cover fees to obtain articles from the Donald J. Matthews Complex Care Fund of the University Health Network in Toronto, Canada. The Toronto Central Local Health Integration Network provided funding for the design and implementation of a patient-oriented discharge summary. None of the funding or supportive agencies were involved in the design or conduct of the present study, analysis, or interpretation of the data, or approval of the manuscript.

Disclosures

The authors report no conflicts of interest.

Patient-centered care, defined by the Institute of Medicine as “health care that establishes a partnership among practitioners, patients, and their families to ensure that decisions respect patients’ wants, needs and preferences and that patients have the education and support they need to make decisions and participate in their own care,” has been recognized as an important factor in improving care transitions after discharge from the hospital.¹ Previous efforts to improve the discharge process for hospitalized patients and reduce avoidable readmissions have focused on improving systems surrounding the patient, such as by increasing the availability of outpatient follow-up or standardizing communication between the inpatient and outpatient care teams.^1,2 In fact, successful programs such as Project BOOST and the Care Transitions Interventions™ provide healthcare institutions with a “bundle” of evidence-based transitional care guidelines for discharge: they provide postdischarge transition coaches, assistance with medication self-management, timely follow-up tips, and improved patient records in order to improve postdischarge outcomes.^3,4 Successful interventions, however, may not provide more services, but also engage the patient in their own care.^5,6 The impact of engaging the patient in his or her own care by providing patient-friendly discharge instructions alone, however, is unknown.

A patient-centered discharge may use tools that were designed with patients, or may involve engaging patients in an interactive process of reviewing discharge instructions and empowering them to manage aspects of their own care after leaving the hospital. This endeavour may lead to more effective use of discharge instructions and reduce the need for additional or more intensive (and costly) interventions. For example, a patient-centered discharge tool could include an educational intervention that uses the “teach-back” method, in which patients are asked to restate in their own words what they thought they heard, or in which staff use additional media or a visual design tool meant to enhance comprehension of discharge instructions.^6,7 Visual aids and the use of larger fonts are particularly useful design elements for improving comprehension among non-English speakers and patients with low health literacy, who tend to have poorer recall of instructions.^8-10 What may constitute essential design elements to include in a discharge instruction tool, however, is not clear.

Moreover, whether the use of discharge tools with a specific focus on patient engagement may improve postdischarge outcomes is not known. Particularly, the ability of patient-centered discharge tools to improve outcomes beyond comprehension such as self-management, adherence to discharge instructions, a reduction in unplanned visits, and a reduction in mortality has not been studied systematically. The objective of this systematic review was to review the literature on discharge instruction tools with a focus on patient engagement and their impact among hospitalized patients.

The Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) Statement was followed as a guideline for reporting throughout this review.¹¹

Data Sources

A literature search was undertaken using the following databases from January 1994 or their inception date to May 2014: Medline, Embase, SIGLE, HTA, Bioethics, ASSIA, Psych Lit, CINAHL, Cochrane Library, EconLit, ERIC, and BioMed Central. We also searched relevant design-focused journals such as Design Issues, Journal of Design Research, Information Design Journal, Innovation, Design Studies, and International Journal of Design, as well as reference lists from studies obtained by electronic searching. The following key words and combination of key words were used with the assistance of a medical librarian: patient discharge, patient-centered discharge, patient-centered design, design thinking, user based design, patient education, discharge summary, education. Additional search terms were added when identified from relevant articles (Appendix).

Inclusion Criteria

We included all English-language studies with patients admitted to the hospital irrespective of age, sex, or medical condition, which included a control group or time period and which measured patient outcomes within 3 months of discharge. The 3-month period after discharge is often cited as a time when outcomes could reasonably be associated with an intervention at discharge.²

Exclusion Criteria

Studies that did not have clear implementation of a patient-centered tool, a control group, or those whose tool was used in the emergency department or as an outpatient were excluded. Studies that included postdischarge tools such as home visits or telephone calls were excluded unless independent effects of the predischarge interventions were measured. Studies with outcomes reported after 3 months were excluded unless outcomes before 3 months were also clearly noted.

All searches were entered into Endnote and duplicates were removed. A 2-stage inclusion process was used. Titles and abstracts of articles were first screened for meeting inclusion and exclusion criteria by 1 reviewer. A second reviewer independently checked a 10% random sample of all the abstracts that met the initial screening criteria. If the agreement to exclude studies was less than 95%, criteria were reviewed before checking the rest of the 90% sample. In the second stage, 2 independent reviewers examined paper copies of the full articles selected in the first stage. Disagreement between reviewers was resolved by discussion or a third reviewer if no agreement could be reached.

Data Analysis and Synthesis

The following information was extracted from the full reference: type of study, population studied, control group or time period, tool used, and outcomes measured. Based on the National Health Care Quality report’s priorities and goals on patient and/or family engagement during transitions of care, educational tools were further described based on method of teaching, involvement of the care team, involvement of the patient in the design or delivery of the tool, and/or the use of visual aids.¹² All primary outcomes were classified according to 3 categories: improved knowledge/comprehension, patient experience (patient satisfaction, self-management/efficacy such as functional status, both physical and mental), and health outcomes (unscheduled visits or readmissions, adherence with medications, diet, exercise, or follow-up, and mortality).

No quantitative pooling of results or meta-analysis was done given the variability and heterogeneity of studies reviewed. However, following guidelines for Effect Practice and Organisation of Care (EPOC) Risk of Bias criteria,¹³ studies that had a higher risk of bias such as uncontrolled before-after studies or studies with only 1 intervention or control site (historical controls, eg) were excluded from the final review because of the difficulties in attributing causation. Only primary outcomes were reported in order to minimize type II errors.

Our search revealed a total of 3699 studies after duplicates had been removed (Figure). A total of 714 references were included after initial review by title and abstract and 30 studies after full-text review. Agreement on a 10% random sample of all abstracts and full text was 79% (k=0.58) and 86% (k=0.72), respectively. Discussion was needed for fewer than 100 references, and agreement was subsequently reached for 100%.

There were 22 randomized controlled trials and 8 nonrandomized studies (5 nonrandomized controlled trials and 3 controlled before-after studies). Most of these studies were conducted in the United States (13/30 studies), followed by other European countries (5 studies), and the United Kingdom (4 studies). A large number of studies were conducted among patients with cardiovascular disease or risk factors (10 studies), followed by postsurgical patients such as coronary artery bypass graft surgery or orthopaedic surgery (5 studies). Five of 30 studies were conducted among individuals older than 65 years. Most studies excluded patients who did not speak English or the country’s official language; only 3 studies included patients with limited literacy, patients who spoke other languages, or caregivers if the patients could not communicate.

Most studies tested the impact of educational discharge interventions (28 of 30 studies) (Table 1). Quite often, it was a member of the research team who carried out the patient education. Only 3 studies involved multiple members of the care team in designing or reviewing the discharge tool with the patient. Almost half (12 studies) targeted multiple aspects of postdischarge care, including medications and side effects, signs and symptoms to consider, plans for follow-up, dietary restrictions, and/or exercise modifications. Many (19 studies) provided education using one-on-one teaching in association with a discharge tool, accompanied by a written handout (13 studies), audiotape (2 studies), or video (3 studies). While 13 studies had patients involved in creating what content was discussed and 14 studies had patients involved in the delivery of the tool, only 6 studies had patients involved in both design and delivery of the tool. Nine studies also used visual aids such as pictures, larger font, or use of a tool enhanced for patients with language barriers or limited health literacy.

Our systematic review found 30 studies that engaged patients during the design or the delivery of a discharge instruction tool and that tested the effect of the tool on postdischarge outcomes.^6-10,14–38 Our review suggests that there is sufficient evidence that patient-centered discharge tools improve comprehension. However, evidence is currently insufficient to determine if patient-centered tools improve adherence with discharge instructions. Moreover, though limited studies show promising results, more studies are needed to determine if patient engagement improves self-efficacy and healthcare utilization after discharge.

A major limitation of current studies is the variability in the level of patient engagement in tool design or delivery. Patients were involved in the design mostly through targeted development of a discharge management plan and the delivery by encouraging them to ask questions. Few studies involved patients in the design of the tool such that patients were responsible for coming up with content that was of interest to them. The few that did, often with the additional use of video media, demonstrated significant outcomes. Only a minority of studies used an interactive process to assess understanding such as “teach-back” or maximize patient comprehension such as visual aids. Even fewer studies engaged patients in both developing the discharge tool and providing discharge instructions.

Several previous studies have demonstrated that most complications after discharge are the result of ineffective communication, which can be exacerbated by lack of fluency in English or by limited health literacy.^2,39-43 As a result, poor understanding of discharge instructions by patients and their caregivers can create an important care gap.⁴⁴ Therefore, the use of patient-centered tools to engage patients at discharge in their own care is needed. How to engage patients consistently and effectively is perhaps less evident, as demonstrated in this review of the literature in which different levels of patient engagement were found. Many of the tools tested placed attention on patient education, sometimes in the context of bundled care along with home visits or follow-up, all of which can require extensive resources and time. Providing patients with information that the patients themselves state is of value may be the easiest refinement to a discharge educational tool, although this was surprisingly uncommon.^{6,9,10,17,23,33,37} Only 2 studies were found that engaged patients in the initial stage of design of the discharge tool, by incorporating information of interest to them.^23,32 For example, a study testing the impact of a computer-generated written education package on poststroke outcomes designed the information by asking patients to identify which topics they would like to receive information about (along with the amount of information and font size).²³ Secondly, although most of the discharge tools reviewed included the use of one-on-one teaching and the use of media such as patient handouts, these tools were often used in such a way that patients were passive recipients. In fact, studies that used additional video media that incorporated personalized content were the most likely to demonstrate positive outcomes.^17,34 The next level of patient engagement may therefore be to involve the patient as an interactive partner when delivering the tool in order to empower patients to self-care. For example, 1 study designed a structured education program by first assessing lifestyle risk factors related to hypertension that were modifiable along with preconceived notions through open-ended questions during a one-on-one interview.³⁷ Patients were subsequently educated on any knowledge deficits regarding the management of their lifestyle. Another level of patient engagement may be to use visual aids during discussions, as a well-known complement to verbal instructions.^45,46 For example, in a controlled study that randomized a ward of elderly patients with 4 or more prescriptions to predischarge counseling, the counseling session aimed to review reasons for their prescriptions along with corresponding side effects, doses, and dosage times with the help of a medicine reminder card. Other uses of visual aid tools identified in our review included the use of pictograms or illustrations or, at minimum, attention to font size.^{7,8,16,29,33,35} In the absence of a visual aid, asking the patient to repeat or demonstrate what was just communicated can be used to assess the amount of information retained.^18,33

An important result discovered in our review of the literature was also the lack of studies that tested the impact of discharge tools on usability of discharge information once at home. Conducting an evaluation of the benefits to patients after discharge can help objectify vague outcomes like health gains or qualify benefits in patient’s views. This might also explain why many studies with documented patient engagement at the time of discharge were able to demonstrate improvements in comprehension but not adherence to instructions. Although patients and caregivers may understand the information, this comprehension does not necessarily mean they will find the information useful or adhere to it once at home. For example, in 1 study, patients discharged with at least 1 medication were randomized to a structured discharge interview during which the treatment plan was reviewed verbally and questions clarified along with a visually enhanced treatment card.²⁶ Although knowledge of medications increased, no effect was found on adherence at 1 week postdischarge. However, use of the treatment card at home was not assessed. Similarly, another study tested the effect of an individualized video of exercises and failed to find a difference in patient adherence at 4 weeks.²⁸ The authors suggested that the lack of benefit may have been because patients were not using the video once at home. This is in contrast to 2 studies that involved patients in their own care by requiring them to request their medication as part of a self-medication tool predischarge.^16,30 Patients were engaged in the process such that increasing independence was given to patients based on their demonstration of understanding and adherence to their treatment while still in the hospital, a learning tool that can be applied once at home. Feeling knowledgeable and involved, as others have suggested, may be the intermediary outcomes that led to improved adherence.⁴⁷ It is also possible that adherence to discharge instructions may vary based on complexity of the information provided, such that instructions focusing solely on medication use may require less patient engagement than discharge instructions that include information on medications, diet, exercise modifications, and follow-up.⁴⁸

Our review has a few limitations. Previous systematic reviews have demonstrated that bundled discharge interventions that include patient-centered education have a positive effect on outcomes postdischarge.^2,5 However, we sought to describe and study the individual and distinct impact of patient engagement in the creation and delivery of discharge tools on outcomes postdischarge. We hoped that this may provide others with key information regarding elements of patient engagement that were particularly useful when designing a new discharge tool. The variability of the studies we identified, however, made it difficult to ascertain what level of patient engagement is required to observe improvements in health outcomes. It is also possible that a higher level of patient engagement may have been used but not described in the studies we reviewed. As only primary outcomes were included, we may have underestimated the effect of patient-centered discharge tools on outcomes that were reported as secondary outcomes. As we were interested in reviewing as many studies of patient-centered discharge tools as possible, we did not assess the quality of the studies and cannot comment on the role of bias in these studies. However, we excluded studies with study designs known to have the highest risk of bias. Lastly, we also cannot comment on whether patient-centered tools may have an effect on outcomes more than 3 months after a hospital discharge. However, several studies included in this review suggest a sustained effect beyond this time period.^8,25,32,37

Patient-centered discharge tools in which patients were engaged in the design or the delivery were found to improve comprehension of but not adherence with discharge instructions. The perceived lack of improved adherence may be due to a lack of studies that measured the usefulness and utilization of information for patients once at home. There was also substantial variability in the extent of patient involvement in designing the style and content of information provided to patients at discharge, as well as the extent of patient engagement when receiving discharge instructions. Future studies would benefit from detailing the level of patient engagement needed in designing and delivery of discharge tools. This information may lead to the discovery of barriers and facilitators to utilization of discharge information once at home and lead to a better understanding of the patient’s journey from hospital to home and onwards.

C.M.B. and this work were funded by a CIHR Canadian Patient Safety Institute Chair in Patient Safety and Continuity of Care. Funding was provided to cover fees to obtain articles from the Donald J. Matthews Complex Care Fund of the University Health Network in Toronto, Canada. The Toronto Central Local Health Integration Network provided funding for the design and implementation of a patient-oriented discharge summary. None of the funding or supportive agencies were involved in the design or conduct of the present study, analysis, or interpretation of the data, or approval of the manuscript.

Disclosures

The authors report no conflicts of interest.

Patient-centered care, defined by the Institute of Medicine as “health care that establishes a partnership among practitioners, patients, and their families to ensure that decisions respect patients’ wants, needs and preferences and that patients have the education and support they need to make decisions and participate in their own care,” has been recognized as an important factor in improving care transitions after discharge from the hospital.¹ Previous efforts to improve the discharge process for hospitalized patients and reduce avoidable readmissions have focused on improving systems surrounding the patient, such as by increasing the availability of outpatient follow-up or standardizing communication between the inpatient and outpatient care teams.^1,2 In fact, successful programs such as Project BOOST and the Care Transitions Interventions™ provide healthcare institutions with a “bundle” of evidence-based transitional care guidelines for discharge: they provide postdischarge transition coaches, assistance with medication self-management, timely follow-up tips, and improved patient records in order to improve postdischarge outcomes.^3,4 Successful interventions, however, may not provide more services, but also engage the patient in their own care.^5,6 The impact of engaging the patient in his or her own care by providing patient-friendly discharge instructions alone, however, is unknown.

A patient-centered discharge may use tools that were designed with patients, or may involve engaging patients in an interactive process of reviewing discharge instructions and empowering them to manage aspects of their own care after leaving the hospital. This endeavour may lead to more effective use of discharge instructions and reduce the need for additional or more intensive (and costly) interventions. For example, a patient-centered discharge tool could include an educational intervention that uses the “teach-back” method, in which patients are asked to restate in their own words what they thought they heard, or in which staff use additional media or a visual design tool meant to enhance comprehension of discharge instructions.^6,7 Visual aids and the use of larger fonts are particularly useful design elements for improving comprehension among non-English speakers and patients with low health literacy, who tend to have poorer recall of instructions.^8-10 What may constitute essential design elements to include in a discharge instruction tool, however, is not clear.

Moreover, whether the use of discharge tools with a specific focus on patient engagement may improve postdischarge outcomes is not known. Particularly, the ability of patient-centered discharge tools to improve outcomes beyond comprehension such as self-management, adherence to discharge instructions, a reduction in unplanned visits, and a reduction in mortality has not been studied systematically. The objective of this systematic review was to review the literature on discharge instruction tools with a focus on patient engagement and their impact among hospitalized patients.

The Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) Statement was followed as a guideline for reporting throughout this review.¹¹

Data Sources

A literature search was undertaken using the following databases from January 1994 or their inception date to May 2014: Medline, Embase, SIGLE, HTA, Bioethics, ASSIA, Psych Lit, CINAHL, Cochrane Library, EconLit, ERIC, and BioMed Central. We also searched relevant design-focused journals such as Design Issues, Journal of Design Research, Information Design Journal, Innovation, Design Studies, and International Journal of Design, as well as reference lists from studies obtained by electronic searching. The following key words and combination of key words were used with the assistance of a medical librarian: patient discharge, patient-centered discharge, patient-centered design, design thinking, user based design, patient education, discharge summary, education. Additional search terms were added when identified from relevant articles (Appendix).

Inclusion Criteria

We included all English-language studies with patients admitted to the hospital irrespective of age, sex, or medical condition, which included a control group or time period and which measured patient outcomes within 3 months of discharge. The 3-month period after discharge is often cited as a time when outcomes could reasonably be associated with an intervention at discharge.²

Exclusion Criteria

Studies that did not have clear implementation of a patient-centered tool, a control group, or those whose tool was used in the emergency department or as an outpatient were excluded. Studies that included postdischarge tools such as home visits or telephone calls were excluded unless independent effects of the predischarge interventions were measured. Studies with outcomes reported after 3 months were excluded unless outcomes before 3 months were also clearly noted.

All searches were entered into Endnote and duplicates were removed. A 2-stage inclusion process was used. Titles and abstracts of articles were first screened for meeting inclusion and exclusion criteria by 1 reviewer. A second reviewer independently checked a 10% random sample of all the abstracts that met the initial screening criteria. If the agreement to exclude studies was less than 95%, criteria were reviewed before checking the rest of the 90% sample. In the second stage, 2 independent reviewers examined paper copies of the full articles selected in the first stage. Disagreement between reviewers was resolved by discussion or a third reviewer if no agreement could be reached.

Data Analysis and Synthesis

The following information was extracted from the full reference: type of study, population studied, control group or time period, tool used, and outcomes measured. Based on the National Health Care Quality report’s priorities and goals on patient and/or family engagement during transitions of care, educational tools were further described based on method of teaching, involvement of the care team, involvement of the patient in the design or delivery of the tool, and/or the use of visual aids.¹² All primary outcomes were classified according to 3 categories: improved knowledge/comprehension, patient experience (patient satisfaction, self-management/efficacy such as functional status, both physical and mental), and health outcomes (unscheduled visits or readmissions, adherence with medications, diet, exercise, or follow-up, and mortality).

No quantitative pooling of results or meta-analysis was done given the variability and heterogeneity of studies reviewed. However, following guidelines for Effect Practice and Organisation of Care (EPOC) Risk of Bias criteria,¹³ studies that had a higher risk of bias such as uncontrolled before-after studies or studies with only 1 intervention or control site (historical controls, eg) were excluded from the final review because of the difficulties in attributing causation. Only primary outcomes were reported in order to minimize type II errors.

Our search revealed a total of 3699 studies after duplicates had been removed (Figure). A total of 714 references were included after initial review by title and abstract and 30 studies after full-text review. Agreement on a 10% random sample of all abstracts and full text was 79% (k=0.58) and 86% (k=0.72), respectively. Discussion was needed for fewer than 100 references, and agreement was subsequently reached for 100%.

There were 22 randomized controlled trials and 8 nonrandomized studies (5 nonrandomized controlled trials and 3 controlled before-after studies). Most of these studies were conducted in the United States (13/30 studies), followed by other European countries (5 studies), and the United Kingdom (4 studies). A large number of studies were conducted among patients with cardiovascular disease or risk factors (10 studies), followed by postsurgical patients such as coronary artery bypass graft surgery or orthopaedic surgery (5 studies). Five of 30 studies were conducted among individuals older than 65 years. Most studies excluded patients who did not speak English or the country’s official language; only 3 studies included patients with limited literacy, patients who spoke other languages, or caregivers if the patients could not communicate.

Most studies tested the impact of educational discharge interventions (28 of 30 studies) (Table 1). Quite often, it was a member of the research team who carried out the patient education. Only 3 studies involved multiple members of the care team in designing or reviewing the discharge tool with the patient. Almost half (12 studies) targeted multiple aspects of postdischarge care, including medications and side effects, signs and symptoms to consider, plans for follow-up, dietary restrictions, and/or exercise modifications. Many (19 studies) provided education using one-on-one teaching in association with a discharge tool, accompanied by a written handout (13 studies), audiotape (2 studies), or video (3 studies). While 13 studies had patients involved in creating what content was discussed and 14 studies had patients involved in the delivery of the tool, only 6 studies had patients involved in both design and delivery of the tool. Nine studies also used visual aids such as pictures, larger font, or use of a tool enhanced for patients with language barriers or limited health literacy.

Our systematic review found 30 studies that engaged patients during the design or the delivery of a discharge instruction tool and that tested the effect of the tool on postdischarge outcomes.^6-10,14–38 Our review suggests that there is sufficient evidence that patient-centered discharge tools improve comprehension. However, evidence is currently insufficient to determine if patient-centered tools improve adherence with discharge instructions. Moreover, though limited studies show promising results, more studies are needed to determine if patient engagement improves self-efficacy and healthcare utilization after discharge.

A major limitation of current studies is the variability in the level of patient engagement in tool design or delivery. Patients were involved in the design mostly through targeted development of a discharge management plan and the delivery by encouraging them to ask questions. Few studies involved patients in the design of the tool such that patients were responsible for coming up with content that was of interest to them. The few that did, often with the additional use of video media, demonstrated significant outcomes. Only a minority of studies used an interactive process to assess understanding such as “teach-back” or maximize patient comprehension such as visual aids. Even fewer studies engaged patients in both developing the discharge tool and providing discharge instructions.

Several previous studies have demonstrated that most complications after discharge are the result of ineffective communication, which can be exacerbated by lack of fluency in English or by limited health literacy.^2,39-43 As a result, poor understanding of discharge instructions by patients and their caregivers can create an important care gap.⁴⁴ Therefore, the use of patient-centered tools to engage patients at discharge in their own care is needed. How to engage patients consistently and effectively is perhaps less evident, as demonstrated in this review of the literature in which different levels of patient engagement were found. Many of the tools tested placed attention on patient education, sometimes in the context of bundled care along with home visits or follow-up, all of which can require extensive resources and time. Providing patients with information that the patients themselves state is of value may be the easiest refinement to a discharge educational tool, although this was surprisingly uncommon.^{6,9,10,17,23,33,37} Only 2 studies were found that engaged patients in the initial stage of design of the discharge tool, by incorporating information of interest to them.^23,32 For example, a study testing the impact of a computer-generated written education package on poststroke outcomes designed the information by asking patients to identify which topics they would like to receive information about (along with the amount of information and font size).²³ Secondly, although most of the discharge tools reviewed included the use of one-on-one teaching and the use of media such as patient handouts, these tools were often used in such a way that patients were passive recipients. In fact, studies that used additional video media that incorporated personalized content were the most likely to demonstrate positive outcomes.^17,34 The next level of patient engagement may therefore be to involve the patient as an interactive partner when delivering the tool in order to empower patients to self-care. For example, 1 study designed a structured education program by first assessing lifestyle risk factors related to hypertension that were modifiable along with preconceived notions through open-ended questions during a one-on-one interview.³⁷ Patients were subsequently educated on any knowledge deficits regarding the management of their lifestyle. Another level of patient engagement may be to use visual aids during discussions, as a well-known complement to verbal instructions.^45,46 For example, in a controlled study that randomized a ward of elderly patients with 4 or more prescriptions to predischarge counseling, the counseling session aimed to review reasons for their prescriptions along with corresponding side effects, doses, and dosage times with the help of a medicine reminder card. Other uses of visual aid tools identified in our review included the use of pictograms or illustrations or, at minimum, attention to font size.^{7,8,16,29,33,35} In the absence of a visual aid, asking the patient to repeat or demonstrate what was just communicated can be used to assess the amount of information retained.^18,33

An important result discovered in our review of the literature was also the lack of studies that tested the impact of discharge tools on usability of discharge information once at home. Conducting an evaluation of the benefits to patients after discharge can help objectify vague outcomes like health gains or qualify benefits in patient’s views. This might also explain why many studies with documented patient engagement at the time of discharge were able to demonstrate improvements in comprehension but not adherence to instructions. Although patients and caregivers may understand the information, this comprehension does not necessarily mean they will find the information useful or adhere to it once at home. For example, in 1 study, patients discharged with at least 1 medication were randomized to a structured discharge interview during which the treatment plan was reviewed verbally and questions clarified along with a visually enhanced treatment card.²⁶ Although knowledge of medications increased, no effect was found on adherence at 1 week postdischarge. However, use of the treatment card at home was not assessed. Similarly, another study tested the effect of an individualized video of exercises and failed to find a difference in patient adherence at 4 weeks.²⁸ The authors suggested that the lack of benefit may have been because patients were not using the video once at home. This is in contrast to 2 studies that involved patients in their own care by requiring them to request their medication as part of a self-medication tool predischarge.^16,30 Patients were engaged in the process such that increasing independence was given to patients based on their demonstration of understanding and adherence to their treatment while still in the hospital, a learning tool that can be applied once at home. Feeling knowledgeable and involved, as others have suggested, may be the intermediary outcomes that led to improved adherence.⁴⁷ It is also possible that adherence to discharge instructions may vary based on complexity of the information provided, such that instructions focusing solely on medication use may require less patient engagement than discharge instructions that include information on medications, diet, exercise modifications, and follow-up.⁴⁸

Our review has a few limitations. Previous systematic reviews have demonstrated that bundled discharge interventions that include patient-centered education have a positive effect on outcomes postdischarge.^2,5 However, we sought to describe and study the individual and distinct impact of patient engagement in the creation and delivery of discharge tools on outcomes postdischarge. We hoped that this may provide others with key information regarding elements of patient engagement that were particularly useful when designing a new discharge tool. The variability of the studies we identified, however, made it difficult to ascertain what level of patient engagement is required to observe improvements in health outcomes. It is also possible that a higher level of patient engagement may have been used but not described in the studies we reviewed. As only primary outcomes were included, we may have underestimated the effect of patient-centered discharge tools on outcomes that were reported as secondary outcomes. As we were interested in reviewing as many studies of patient-centered discharge tools as possible, we did not assess the quality of the studies and cannot comment on the role of bias in these studies. However, we excluded studies with study designs known to have the highest risk of bias. Lastly, we also cannot comment on whether patient-centered tools may have an effect on outcomes more than 3 months after a hospital discharge. However, several studies included in this review suggest a sustained effect beyond this time period.^8,25,32,37

Patient-centered discharge tools in which patients were engaged in the design or the delivery were found to improve comprehension of but not adherence with discharge instructions. The perceived lack of improved adherence may be due to a lack of studies that measured the usefulness and utilization of information for patients once at home. There was also substantial variability in the extent of patient involvement in designing the style and content of information provided to patients at discharge, as well as the extent of patient engagement when receiving discharge instructions. Future studies would benefit from detailing the level of patient engagement needed in designing and delivery of discharge tools. This information may lead to the discovery of barriers and facilitators to utilization of discharge information once at home and lead to a better understanding of the patient’s journey from hospital to home and onwards.

C.M.B. and this work were funded by a CIHR Canadian Patient Safety Institute Chair in Patient Safety and Continuity of Care. Funding was provided to cover fees to obtain articles from the Donald J. Matthews Complex Care Fund of the University Health Network in Toronto, Canada. The Toronto Central Local Health Integration Network provided funding for the design and implementation of a patient-oriented discharge summary. None of the funding or supportive agencies were involved in the design or conduct of the present study, analysis, or interpretation of the data, or approval of the manuscript.

Disclosures

The authors report no conflicts of interest.

In our current healthcare system, pressure to provide cost- and time-efficient care is immense. Inpatient care often focuses on assessing the patient’s presenting illness or injury and treating that condition in a manner that gets the patient on their feet and out of the hospital quickly. Because depression is not an indication for hospitalization so long as active suicidality is absent, inpatient physicians may view it as a problem best managed in the outpatient setting. Yet both psychosocial and physical factors associated with depression put patients at risk for rehospitalization.¹ Furthermore, hospitalization represents an unrecognized opportunity to optimize both mental and physical health outcomes.²

Indeed, poor physical and mental health often occur together. Depressed inpatients have poorer outcomes, increased length of stay, and greater vulnerability to hospital readmission.^3,4 Among elderly hospitalized patients, depression is particularly common, especially in those with poor physical health, alcoholism,⁵ hip fracture, and stroke.⁶ Yet little is known about how often depression goes unrecognized, undiagnosed, and, therefore, untreated.

The US Preventive Services Task Force (USPSTF) recommends screening for depression in the general adult population, including pregnant and postpartum women, and further suggests that screening should be implemented “with adequate systems in place to ensure accurate diagnosis, effective treatment, and appropriate follow-up.”² The USPSTF guidelines do not distinguish between inpatient and outpatient settings. However, the preponderance of evidence for screening comes from outpatient care settings, and little is known about screening among inpatient populations.⁷

This study had 2 objectives. First, we sought to examine the performance of depression screening tools in inpatient settings. If depression screening were to become routine in hospital settings, screening tools would need to be sensitive and specific as well as brief and suitable for self-administration by patients or for administration by nurses, resident physicians, or hospitalists. It is also important to consider administration by mental health professionals, who may be best trained to administer such tests. We, therefore, examined 3 types of studies: (1) studies that tested a self-administered screening instrument, (2) studies that tested screening by individuals without formal training, and (3) studies that compared screening tools administered by mental health professionals. Second, we sought to describe associations between depression and clinical or utilization outcomes among hospitalized patients.

We adhered to recommendations in the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) Statement,^8,9 including designing the analysis before performing the review. However, we did not post a protocol in an online registry, formally assess study quality, or perform a meta-analysis.

Data Sources and Searches

We searched PsycINFO and PubMed databases for articles published between 1990 and 2016 (as of July 31, 2016). In PubMed, 2 search term strings were used to capture studies of depression screening tools in inpatient settings. The first used the advanced search option to exclude studies related to primary care settings or children and adolescents, and the second used MeSH terms to ensure that a wide variety of studies were included. Specific search terms are included in the Appendix. A similar search was conducted in the PsycINFO database and these search terms are also included in the Appendix.

Study Selection

Articles were eligible if they were published in English in peer-reviewed journals, included at least 20 adults hospitalized for nonpsychiatric reasons, and described the use of at least 1 measure of depression. The studies must have either tested the validity of a depression screening tool or examined the association between depression screening and clinical or utilization outcomes. Two investigators reviewed each title, abstract, and full-text article to determine eligibility, then reached a consensus on which studies to include in this review.

Data Extraction

Two investigators reviewed each full-text article to extract information related to study design, population, and outcomes regarding screening tool analysis or clinical results. From articles that assessed the performance of depression screening tools, we extracted information related to the nature and application of the index test, the nature and application of the reference test, the prevalence of depression, and the sensitivity and specificity of the index test compared with the reference test. For articles that focused on the association between depression screening and clinical or utilization outcomes, the data on relevant clinical outcomes included symptom severity, quality of life, and daily functioning, whereas the data on utilization outcomes included length of stay, readmission, and the cost of care.

Altogether, the search identified 3226 records. After eliminating duplicates and abstracts not suitable for inclusion (Figure), 101 articles underwent full-text review and 32 were found to be eligible. Of these, 12 focused on the association between depression and clinical or utilization outcomes, while 20 assessed the performance of depression screening tools.

Depression Screening Tools

Table 1 describes the index and reference instruments as well as methods of administration, the prevalence of depression, and the sensitivity and specificity of the index instruments relative to the reference instruments. Across the 20 studies, the prevalence of depression ranged from 15% to 60%, with a median of 34%.^10–29 This finding may reflect different methods of screening or variation among diverse hospitalized populations. Many of the studies excluded patients with cognitive impairment or communication barriers.

The included studies tested a wide range of unique instruments, and compared them with diverse reference standards. Five studies examined instruments that were self-administered by patients^10–14; 9 studies assessed instruments administered by nurses, physicians, or research staff members without formal psychiatric training^15–23; and 6 studies evaluated instruments administered by mental health professionals.^24–29 Four studies compared different instruments that were administered in the same manner (eg, both self-administered by patients).^12–14,22 In the remaining studies, both instruments and methods of administration differed between the index and reference conditions.

Eight studies tested brief instruments with 5 or fewer items, most of which exhibited good sensitivity (range 38%–91%) and specificity (range 68%–86%) relative to longer instruments.^{12,14–19,22} In 2 of these studies, instruments were self-administered. In 1 case, a single self-administered item from the STOP-D instrument (“Over the past 2 weeks, how much have you been bothered by feeling sad, down, or uninterested in life?”) performed nearly as well as the 14-item Hospital Anxiety and Depression Scale.¹⁴ In the other 6 studies testing brief instruments, the instruments were administered by individuals without formal training.^15–19,22 In 1 such study, geriatricians asking 2 questions about depressed mood and anhedonia performed well compared with a formal psychiatric interview.¹⁷

Four studies tested variations of the Geriatric Depression Scale (GDS).^12,18,21,23 In 3 of these studies, abbreviated versions of the GDS exhibited relatively high sensitivity and specificity.^12,18,21 However, a study comparing the 15-item GDS (GDS-15) with the GDS-4 found that GDS-15 correctly classified 10% more patients with suspected depression.¹² Two studies examined variations of the Patient Health Questionnaire (PHQ). One study found that both the PHQ-2 and PHQ-9 obtained by staff nurses performed well relative to a comprehensive assessment by a trained advanced practice nurse.^13,19

When reported, positive predictive value, negative predictive value, and area under the receiver-operator curve were generally high.

Depression and Clinical or Utilization Outcomes

Of the 12 studies that reported either clinical or utilization outcomes for depression screening in an inpatient setting,^4,30–40 3 measured rates of rehospitalization.^4,31,39 The other 9 studies tested for associations between symptoms of depression and either health or treatment outcomes. Table 2 provides a more detailed description of the study designs and results.

Other studies found that depression was associated with reduced functional abilities such as mobility and self-care,^30,32–34 and increased hospital readmission³¹ as well as physical and mental health deficits.³⁷ Interestingly, although 1 study did not find that depression and hospital readmission were closely linked (frequency at 19%), it found that comorbid illness and previous hospitalizations predicted readmission.⁴

We also evaluated the associations between depression diagnosed in the inpatient studies and 2 types of outcomes. The first type includes clinical outcomes including symptom severity, quality of life, and daily functioning. Most studies we identified assessed clinical outcomes, and all detected an association between depression and worse clinical outcomes. The second type includes healthcare utilization, which can be measured with the patients’ length of hospital stay, readmission and cost of care. In 1 such study, Mitchell aet al.³¹ reported a 54% increase in readmission within 30 days of discharge among patients who screened positive for depression.³¹ Additionally, Cully et al.³⁰ found that depression may impinge on the recovery process of acute rehabilitation patients.

In our current healthcare system, pressure to provide cost- and time-efficient care is immense. Inpatient care often focuses on assessing the patient’s presenting illness or injury and treating that condition in a manner that gets the patient on their feet and out of the hospital quickly. Because depression is not an indication for hospitalization so long as active suicidality is absent, inpatient physicians may view it as a problem best managed in the outpatient setting. Yet both psychosocial and physical factors associated with depression put patients at risk for rehospitalization.¹ Furthermore, hospitalization represents an unrecognized opportunity to optimize both mental and physical health outcomes.²

Indeed, poor physical and mental health often occur together. Depressed inpatients have poorer outcomes, increased length of stay, and greater vulnerability to hospital readmission.^3,4 Among elderly hospitalized patients, depression is particularly common, especially in those with poor physical health, alcoholism,⁵ hip fracture, and stroke.⁶ Yet little is known about how often depression goes unrecognized, undiagnosed, and, therefore, untreated.

The US Preventive Services Task Force (USPSTF) recommends screening for depression in the general adult population, including pregnant and postpartum women, and further suggests that screening should be implemented “with adequate systems in place to ensure accurate diagnosis, effective treatment, and appropriate follow-up.”² The USPSTF guidelines do not distinguish between inpatient and outpatient settings. However, the preponderance of evidence for screening comes from outpatient care settings, and little is known about screening among inpatient populations.⁷

This study had 2 objectives. First, we sought to examine the performance of depression screening tools in inpatient settings. If depression screening were to become routine in hospital settings, screening tools would need to be sensitive and specific as well as brief and suitable for self-administration by patients or for administration by nurses, resident physicians, or hospitalists. It is also important to consider administration by mental health professionals, who may be best trained to administer such tests. We, therefore, examined 3 types of studies: (1) studies that tested a self-administered screening instrument, (2) studies that tested screening by individuals without formal training, and (3) studies that compared screening tools administered by mental health professionals. Second, we sought to describe associations between depression and clinical or utilization outcomes among hospitalized patients.

We adhered to recommendations in the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) Statement,^8,9 including designing the analysis before performing the review. However, we did not post a protocol in an online registry, formally assess study quality, or perform a meta-analysis.

Data Sources and Searches

We searched PsycINFO and PubMed databases for articles published between 1990 and 2016 (as of July 31, 2016). In PubMed, 2 search term strings were used to capture studies of depression screening tools in inpatient settings. The first used the advanced search option to exclude studies related to primary care settings or children and adolescents, and the second used MeSH terms to ensure that a wide variety of studies were included. Specific search terms are included in the Appendix. A similar search was conducted in the PsycINFO database and these search terms are also included in the Appendix.

Study Selection

Articles were eligible if they were published in English in peer-reviewed journals, included at least 20 adults hospitalized for nonpsychiatric reasons, and described the use of at least 1 measure of depression. The studies must have either tested the validity of a depression screening tool or examined the association between depression screening and clinical or utilization outcomes. Two investigators reviewed each title, abstract, and full-text article to determine eligibility, then reached a consensus on which studies to include in this review.

Data Extraction

Two investigators reviewed each full-text article to extract information related to study design, population, and outcomes regarding screening tool analysis or clinical results. From articles that assessed the performance of depression screening tools, we extracted information related to the nature and application of the index test, the nature and application of the reference test, the prevalence of depression, and the sensitivity and specificity of the index test compared with the reference test. For articles that focused on the association between depression screening and clinical or utilization outcomes, the data on relevant clinical outcomes included symptom severity, quality of life, and daily functioning, whereas the data on utilization outcomes included length of stay, readmission, and the cost of care.

Altogether, the search identified 3226 records. After eliminating duplicates and abstracts not suitable for inclusion (Figure), 101 articles underwent full-text review and 32 were found to be eligible. Of these, 12 focused on the association between depression and clinical or utilization outcomes, while 20 assessed the performance of depression screening tools.

Depression Screening Tools

Table 1 describes the index and reference instruments as well as methods of administration, the prevalence of depression, and the sensitivity and specificity of the index instruments relative to the reference instruments. Across the 20 studies, the prevalence of depression ranged from 15% to 60%, with a median of 34%.^10–29 This finding may reflect different methods of screening or variation among diverse hospitalized populations. Many of the studies excluded patients with cognitive impairment or communication barriers.

The included studies tested a wide range of unique instruments, and compared them with diverse reference standards. Five studies examined instruments that were self-administered by patients^10–14; 9 studies assessed instruments administered by nurses, physicians, or research staff members without formal psychiatric training^15–23; and 6 studies evaluated instruments administered by mental health professionals.^24–29 Four studies compared different instruments that were administered in the same manner (eg, both self-administered by patients).^12–14,22 In the remaining studies, both instruments and methods of administration differed between the index and reference conditions.

Eight studies tested brief instruments with 5 or fewer items, most of which exhibited good sensitivity (range 38%–91%) and specificity (range 68%–86%) relative to longer instruments.^{12,14–19,22} In 2 of these studies, instruments were self-administered. In 1 case, a single self-administered item from the STOP-D instrument (“Over the past 2 weeks, how much have you been bothered by feeling sad, down, or uninterested in life?”) performed nearly as well as the 14-item Hospital Anxiety and Depression Scale.¹⁴ In the other 6 studies testing brief instruments, the instruments were administered by individuals without formal training.^15–19,22 In 1 such study, geriatricians asking 2 questions about depressed mood and anhedonia performed well compared with a formal psychiatric interview.¹⁷

Four studies tested variations of the Geriatric Depression Scale (GDS).^12,18,21,23 In 3 of these studies, abbreviated versions of the GDS exhibited relatively high sensitivity and specificity.^12,18,21 However, a study comparing the 15-item GDS (GDS-15) with the GDS-4 found that GDS-15 correctly classified 10% more patients with suspected depression.¹² Two studies examined variations of the Patient Health Questionnaire (PHQ). One study found that both the PHQ-2 and PHQ-9 obtained by staff nurses performed well relative to a comprehensive assessment by a trained advanced practice nurse.^13,19

When reported, positive predictive value, negative predictive value, and area under the receiver-operator curve were generally high.

Depression and Clinical or Utilization Outcomes

Of the 12 studies that reported either clinical or utilization outcomes for depression screening in an inpatient setting,^4,30–40 3 measured rates of rehospitalization.^4,31,39 The other 9 studies tested for associations between symptoms of depression and either health or treatment outcomes. Table 2 provides a more detailed description of the study designs and results.

Other studies found that depression was associated with reduced functional abilities such as mobility and self-care,^30,32–34 and increased hospital readmission³¹ as well as physical and mental health deficits.³⁷ Interestingly, although 1 study did not find that depression and hospital readmission were closely linked (frequency at 19%), it found that comorbid illness and previous hospitalizations predicted readmission.⁴

We also evaluated the associations between depression diagnosed in the inpatient studies and 2 types of outcomes. The first type includes clinical outcomes including symptom severity, quality of life, and daily functioning. Most studies we identified assessed clinical outcomes, and all detected an association between depression and worse clinical outcomes. The second type includes healthcare utilization, which can be measured with the patients’ length of hospital stay, readmission and cost of care. In 1 such study, Mitchell aet al.³¹ reported a 54% increase in readmission within 30 days of discharge among patients who screened positive for depression.³¹ Additionally, Cully et al.³⁰ found that depression may impinge on the recovery process of acute rehabilitation patients.

In our current healthcare system, pressure to provide cost- and time-efficient care is immense. Inpatient care often focuses on assessing the patient’s presenting illness or injury and treating that condition in a manner that gets the patient on their feet and out of the hospital quickly. Because depression is not an indication for hospitalization so long as active suicidality is absent, inpatient physicians may view it as a problem best managed in the outpatient setting. Yet both psychosocial and physical factors associated with depression put patients at risk for rehospitalization.¹ Furthermore, hospitalization represents an unrecognized opportunity to optimize both mental and physical health outcomes.²

Indeed, poor physical and mental health often occur together. Depressed inpatients have poorer outcomes, increased length of stay, and greater vulnerability to hospital readmission.^3,4 Among elderly hospitalized patients, depression is particularly common, especially in those with poor physical health, alcoholism,⁵ hip fracture, and stroke.⁶ Yet little is known about how often depression goes unrecognized, undiagnosed, and, therefore, untreated.

The US Preventive Services Task Force (USPSTF) recommends screening for depression in the general adult population, including pregnant and postpartum women, and further suggests that screening should be implemented “with adequate systems in place to ensure accurate diagnosis, effective treatment, and appropriate follow-up.”² The USPSTF guidelines do not distinguish between inpatient and outpatient settings. However, the preponderance of evidence for screening comes from outpatient care settings, and little is known about screening among inpatient populations.⁷

This study had 2 objectives. First, we sought to examine the performance of depression screening tools in inpatient settings. If depression screening were to become routine in hospital settings, screening tools would need to be sensitive and specific as well as brief and suitable for self-administration by patients or for administration by nurses, resident physicians, or hospitalists. It is also important to consider administration by mental health professionals, who may be best trained to administer such tests. We, therefore, examined 3 types of studies: (1) studies that tested a self-administered screening instrument, (2) studies that tested screening by individuals without formal training, and (3) studies that compared screening tools administered by mental health professionals. Second, we sought to describe associations between depression and clinical or utilization outcomes among hospitalized patients.

We adhered to recommendations in the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) Statement,^8,9 including designing the analysis before performing the review. However, we did not post a protocol in an online registry, formally assess study quality, or perform a meta-analysis.

Data Sources and Searches

We searched PsycINFO and PubMed databases for articles published between 1990 and 2016 (as of July 31, 2016). In PubMed, 2 search term strings were used to capture studies of depression screening tools in inpatient settings. The first used the advanced search option to exclude studies related to primary care settings or children and adolescents, and the second used MeSH terms to ensure that a wide variety of studies were included. Specific search terms are included in the Appendix. A similar search was conducted in the PsycINFO database and these search terms are also included in the Appendix.

Study Selection

Articles were eligible if they were published in English in peer-reviewed journals, included at least 20 adults hospitalized for nonpsychiatric reasons, and described the use of at least 1 measure of depression. The studies must have either tested the validity of a depression screening tool or examined the association between depression screening and clinical or utilization outcomes. Two investigators reviewed each title, abstract, and full-text article to determine eligibility, then reached a consensus on which studies to include in this review.

Data Extraction

Two investigators reviewed each full-text article to extract information related to study design, population, and outcomes regarding screening tool analysis or clinical results. From articles that assessed the performance of depression screening tools, we extracted information related to the nature and application of the index test, the nature and application of the reference test, the prevalence of depression, and the sensitivity and specificity of the index test compared with the reference test. For articles that focused on the association between depression screening and clinical or utilization outcomes, the data on relevant clinical outcomes included symptom severity, quality of life, and daily functioning, whereas the data on utilization outcomes included length of stay, readmission, and the cost of care.

Altogether, the search identified 3226 records. After eliminating duplicates and abstracts not suitable for inclusion (Figure), 101 articles underwent full-text review and 32 were found to be eligible. Of these, 12 focused on the association between depression and clinical or utilization outcomes, while 20 assessed the performance of depression screening tools.

Depression Screening Tools

Table 1 describes the index and reference instruments as well as methods of administration, the prevalence of depression, and the sensitivity and specificity of the index instruments relative to the reference instruments. Across the 20 studies, the prevalence of depression ranged from 15% to 60%, with a median of 34%.^10–29 This finding may reflect different methods of screening or variation among diverse hospitalized populations. Many of the studies excluded patients with cognitive impairment or communication barriers.

The included studies tested a wide range of unique instruments, and compared them with diverse reference standards. Five studies examined instruments that were self-administered by patients^10–14; 9 studies assessed instruments administered by nurses, physicians, or research staff members without formal psychiatric training^15–23; and 6 studies evaluated instruments administered by mental health professionals.^24–29 Four studies compared different instruments that were administered in the same manner (eg, both self-administered by patients).^12–14,22 In the remaining studies, both instruments and methods of administration differed between the index and reference conditions.

Eight studies tested brief instruments with 5 or fewer items, most of which exhibited good sensitivity (range 38%–91%) and specificity (range 68%–86%) relative to longer instruments.^{12,14–19,22} In 2 of these studies, instruments were self-administered. In 1 case, a single self-administered item from the STOP-D instrument (“Over the past 2 weeks, how much have you been bothered by feeling sad, down, or uninterested in life?”) performed nearly as well as the 14-item Hospital Anxiety and Depression Scale.¹⁴ In the other 6 studies testing brief instruments, the instruments were administered by individuals without formal training.^15–19,22 In 1 such study, geriatricians asking 2 questions about depressed mood and anhedonia performed well compared with a formal psychiatric interview.¹⁷

Four studies tested variations of the Geriatric Depression Scale (GDS).^12,18,21,23 In 3 of these studies, abbreviated versions of the GDS exhibited relatively high sensitivity and specificity.^12,18,21 However, a study comparing the 15-item GDS (GDS-15) with the GDS-4 found that GDS-15 correctly classified 10% more patients with suspected depression.¹² Two studies examined variations of the Patient Health Questionnaire (PHQ). One study found that both the PHQ-2 and PHQ-9 obtained by staff nurses performed well relative to a comprehensive assessment by a trained advanced practice nurse.^13,19

When reported, positive predictive value, negative predictive value, and area under the receiver-operator curve were generally high.

Depression and Clinical or Utilization Outcomes

Of the 12 studies that reported either clinical or utilization outcomes for depression screening in an inpatient setting,^4,30–40 3 measured rates of rehospitalization.^4,31,39 The other 9 studies tested for associations between symptoms of depression and either health or treatment outcomes. Table 2 provides a more detailed description of the study designs and results.

Other studies found that depression was associated with reduced functional abilities such as mobility and self-care,^30,32–34 and increased hospital readmission³¹ as well as physical and mental health deficits.³⁷ Interestingly, although 1 study did not find that depression and hospital readmission were closely linked (frequency at 19%), it found that comorbid illness and previous hospitalizations predicted readmission.⁴

We also evaluated the associations between depression diagnosed in the inpatient studies and 2 types of outcomes. The first type includes clinical outcomes including symptom severity, quality of life, and daily functioning. Most studies we identified assessed clinical outcomes, and all detected an association between depression and worse clinical outcomes. The second type includes healthcare utilization, which can be measured with the patients’ length of hospital stay, readmission and cost of care. In 1 such study, Mitchell aet al.³¹ reported a 54% increase in readmission within 30 days of discharge among patients who screened positive for depression.³¹ Additionally, Cully et al.³⁰ found that depression may impinge on the recovery process of acute rehabilitation patients.

Acute kidney injury (AKI) is a major contributor to morbidity and mortality in hospitalized patients across the world.¹ Affecting up to 20% of all admissions (depending on which definition of AKI is used),² AKI is the most common reason for new-inpatient nephrology consultation. Recent data suggest that AKI incidence has risen rapidly, by up to 10% per year.^3,4

AKI is associated with a variety of serious short- and long-term complications. Approximately 33% to 60% of critically ill patients who develop dialysis-requiring AKI do not survive to hospital discharge, and mortality associated with dialysis-requiring AKI is greater than that associated with other serious conditions such as myocardial infarction or acute respiratory distress syndrome.⁵ Even relatively mild AKI in the acute inpatient setting appears to be an independent risk factor for mortality.⁶

For several decades, many physicians believed that AKI was a self-limited process followed by complete recovery of renal function to pre-AKI levels among survivors. (Numerous trainees have been taught some variant of the old adage: “If the patients survive, so will their kidneys.”) But studies linking AKI with the development of new-onset chronic kidney disease (CKD) or the accelerated progression of pre-existing CKD have changed this view.⁷ One important reason the long-term impact of AKI hasn’t been appreciated is that, traditionally, clinical studies of AKI examined inhospital outcomes such as short-term mortality and resource usage and did not consider what transpired months to years after discharge. More recently, epidemiologic studies linking inpatient events with outpatient outcomes have filled this knowledge gap.⁸ Contemporary animal models of AKI have shed light on potential mechanisms of maladaptive repair after AKI, characterized by fibrosis, vascular rarefaction, tubular loss, glomerulosclerosis, and chronic interstitial inflammation, all of which result in renal function decline. So over the last decade there has been a paradigm shift in how we think about AKI and CKD. Rather than distinct entities, AKI and CKD are now viewed as interconnected syndromes since AKI is a risk factor for CKD progression and CKD is a risk factor for new episodes of AKI.⁹

Two studies published in this issue of the Journal of Hospital Medicine augment our understanding of AKI and its clinical impact in hospitalized patients. Analyzing data from the National Inpatient Sample, Silver et al.¹⁰ found that hospitalizations that include AKI are substantially costlier and associated with longer lengths of stay than hospitalizations without AKI. The authors also highlight that the additional economic costs of AKI exceeded those of many other higher-profile yet less-common acute medical conditions, such as myocardial infarction and gastrointestinal bleeding. These results re-emphasize the important economic burden of AKI at a national level and expand on prior literature by confirming findings previously limited to single-center and regional studies. Better defining the impact AKI has on our healthcare system could help ensure that adequate resources are invested to combat AKI.

The second study, by Rutter et al.,¹¹ found that among hospitalized patients with normal baseline renal function, use of vancomycin in combination with piperacillin-tazobactam is associated with a higher incidence of AKI after antibiotic exposure than use of either agent as monotherapy. This association persisted even after adjusting for potential confounders such as underlying comorbidities, exposure to nephrotoxic agents, documented hypotension, and baseline renal impairment. This study adds to a growing body of literature that suggests synergistic nephrotoxicity between vancomycin and piperacillin-tazobactam. It underscores that any medical intervention—even treatments typically envisioned as non-hazardous and frequently life-saving—involve inherent risks and should prompt the medical community to promote proper antimicrobial stewardship. Whether such exposures to vancomycin or beta-lactam derivatives cause AKI via direct tubular damage, interstitial nephritis, or some other novel mechanism remains to be elucidated. Better delineation of the contemporary causes of AKI, including increased antibiotic exposure, is the first step toward identifying ways to reduce AKI incidence.

Both of these papers serve to highlight the clinical importance of AKI among hospitalized patients. Their findings re-emphasize the need for vigilance in detecting AKI and intervening early to achieve the best clinical outcomes.

Given recent understanding that survivors of AKI are at greater risk for more rapid loss of renal function long after hospital discharge, one goal the US Department of Health and Human Services put forth for Healthy People 2020 is to “increase the proportion of hospital patients who incurred AKI who have follow-up renal evaluation in 6 months post-discharge” (10% improvement targeted).¹² Transitions of care after hospitalizations complicated by AKI require special attention to ensure that patients’ needs are optimally monitored and managed during the critical post-discharge period. One recent study analyzing discharge documentation for hospitalizations including AKI found that fewer than half of the discharge summaries and patient instructions commented on the presence, cause, or course of AKI, indicating clear room for improvement.¹³ And currently, it appears that only a minority of patients with AKI—even AKI severe enough to require dialysis—are seen by a nephrologist within 90 days of discharge.¹⁴

Hospitalists play a crucial role in coordinating care as vulnerable patients transition from the inpatient to outpatient setting. We suggest that AKI should be properly documented in the discharge summary. In addition, patients should be informed that they experienced AKI so they can discuss with future caregivers potential strategies to avoid additional renal insults. Discharge referrals to nephrology should be arranged for high-risk patients, including those whose renal function remains decreased at discharge or those who had recurrent AKI episodes during prior hospitalizations. For patients with pre-hospitalization baseline CKD, nephrology should be consulted before indefinitely discontinuing medications like angiotensin-converting enzyme inhibitors or angiotensin receptor blockers. These medications are indispensable in retarding the progression of proteinuric CKD, even though they may predispose patients to AKI under certain circumstances (eg, in states of decreased renal perfusion). Adopting these simple steps may substantially improve the long-term outcomes of patients who experience AKI during hospitalization.

Acknowledgments

The authors are supported by NIH-NIDDK Grants T32DK007219 (BJL) and K24DK92291 (CYH).

Disclosure

Nothing to report.

Acute kidney injury (AKI) is a major contributor to morbidity and mortality in hospitalized patients across the world.¹ Affecting up to 20% of all admissions (depending on which definition of AKI is used),² AKI is the most common reason for new-inpatient nephrology consultation. Recent data suggest that AKI incidence has risen rapidly, by up to 10% per year.^3,4

AKI is associated with a variety of serious short- and long-term complications. Approximately 33% to 60% of critically ill patients who develop dialysis-requiring AKI do not survive to hospital discharge, and mortality associated with dialysis-requiring AKI is greater than that associated with other serious conditions such as myocardial infarction or acute respiratory distress syndrome.⁵ Even relatively mild AKI in the acute inpatient setting appears to be an independent risk factor for mortality.⁶

For several decades, many physicians believed that AKI was a self-limited process followed by complete recovery of renal function to pre-AKI levels among survivors. (Numerous trainees have been taught some variant of the old adage: “If the patients survive, so will their kidneys.”) But studies linking AKI with the development of new-onset chronic kidney disease (CKD) or the accelerated progression of pre-existing CKD have changed this view.⁷ One important reason the long-term impact of AKI hasn’t been appreciated is that, traditionally, clinical studies of AKI examined inhospital outcomes such as short-term mortality and resource usage and did not consider what transpired months to years after discharge. More recently, epidemiologic studies linking inpatient events with outpatient outcomes have filled this knowledge gap.⁸ Contemporary animal models of AKI have shed light on potential mechanisms of maladaptive repair after AKI, characterized by fibrosis, vascular rarefaction, tubular loss, glomerulosclerosis, and chronic interstitial inflammation, all of which result in renal function decline. So over the last decade there has been a paradigm shift in how we think about AKI and CKD. Rather than distinct entities, AKI and CKD are now viewed as interconnected syndromes since AKI is a risk factor for CKD progression and CKD is a risk factor for new episodes of AKI.⁹

Two studies published in this issue of the Journal of Hospital Medicine augment our understanding of AKI and its clinical impact in hospitalized patients. Analyzing data from the National Inpatient Sample, Silver et al.¹⁰ found that hospitalizations that include AKI are substantially costlier and associated with longer lengths of stay than hospitalizations without AKI. The authors also highlight that the additional economic costs of AKI exceeded those of many other higher-profile yet less-common acute medical conditions, such as myocardial infarction and gastrointestinal bleeding. These results re-emphasize the important economic burden of AKI at a national level and expand on prior literature by confirming findings previously limited to single-center and regional studies. Better defining the impact AKI has on our healthcare system could help ensure that adequate resources are invested to combat AKI.

The second study, by Rutter et al.,¹¹ found that among hospitalized patients with normal baseline renal function, use of vancomycin in combination with piperacillin-tazobactam is associated with a higher incidence of AKI after antibiotic exposure than use of either agent as monotherapy. This association persisted even after adjusting for potential confounders such as underlying comorbidities, exposure to nephrotoxic agents, documented hypotension, and baseline renal impairment. This study adds to a growing body of literature that suggests synergistic nephrotoxicity between vancomycin and piperacillin-tazobactam. It underscores that any medical intervention—even treatments typically envisioned as non-hazardous and frequently life-saving—involve inherent risks and should prompt the medical community to promote proper antimicrobial stewardship. Whether such exposures to vancomycin or beta-lactam derivatives cause AKI via direct tubular damage, interstitial nephritis, or some other novel mechanism remains to be elucidated. Better delineation of the contemporary causes of AKI, including increased antibiotic exposure, is the first step toward identifying ways to reduce AKI incidence.

Both of these papers serve to highlight the clinical importance of AKI among hospitalized patients. Their findings re-emphasize the need for vigilance in detecting AKI and intervening early to achieve the best clinical outcomes.

Given recent understanding that survivors of AKI are at greater risk for more rapid loss of renal function long after hospital discharge, one goal the US Department of Health and Human Services put forth for Healthy People 2020 is to “increase the proportion of hospital patients who incurred AKI who have follow-up renal evaluation in 6 months post-discharge” (10% improvement targeted).¹² Transitions of care after hospitalizations complicated by AKI require special attention to ensure that patients’ needs are optimally monitored and managed during the critical post-discharge period. One recent study analyzing discharge documentation for hospitalizations including AKI found that fewer than half of the discharge summaries and patient instructions commented on the presence, cause, or course of AKI, indicating clear room for improvement.¹³ And currently, it appears that only a minority of patients with AKI—even AKI severe enough to require dialysis—are seen by a nephrologist within 90 days of discharge.¹⁴

Hospitalists play a crucial role in coordinating care as vulnerable patients transition from the inpatient to outpatient setting. We suggest that AKI should be properly documented in the discharge summary. In addition, patients should be informed that they experienced AKI so they can discuss with future caregivers potential strategies to avoid additional renal insults. Discharge referrals to nephrology should be arranged for high-risk patients, including those whose renal function remains decreased at discharge or those who had recurrent AKI episodes during prior hospitalizations. For patients with pre-hospitalization baseline CKD, nephrology should be consulted before indefinitely discontinuing medications like angiotensin-converting enzyme inhibitors or angiotensin receptor blockers. These medications are indispensable in retarding the progression of proteinuric CKD, even though they may predispose patients to AKI under certain circumstances (eg, in states of decreased renal perfusion). Adopting these simple steps may substantially improve the long-term outcomes of patients who experience AKI during hospitalization.

Acknowledgments

The authors are supported by NIH-NIDDK Grants T32DK007219 (BJL) and K24DK92291 (CYH).

Disclosure

Nothing to report.

Acute kidney injury (AKI) is a major contributor to morbidity and mortality in hospitalized patients across the world.¹ Affecting up to 20% of all admissions (depending on which definition of AKI is used),² AKI is the most common reason for new-inpatient nephrology consultation. Recent data suggest that AKI incidence has risen rapidly, by up to 10% per year.^3,4

AKI is associated with a variety of serious short- and long-term complications. Approximately 33% to 60% of critically ill patients who develop dialysis-requiring AKI do not survive to hospital discharge, and mortality associated with dialysis-requiring AKI is greater than that associated with other serious conditions such as myocardial infarction or acute respiratory distress syndrome.⁵ Even relatively mild AKI in the acute inpatient setting appears to be an independent risk factor for mortality.⁶

For several decades, many physicians believed that AKI was a self-limited process followed by complete recovery of renal function to pre-AKI levels among survivors. (Numerous trainees have been taught some variant of the old adage: “If the patients survive, so will their kidneys.”) But studies linking AKI with the development of new-onset chronic kidney disease (CKD) or the accelerated progression of pre-existing CKD have changed this view.⁷ One important reason the long-term impact of AKI hasn’t been appreciated is that, traditionally, clinical studies of AKI examined inhospital outcomes such as short-term mortality and resource usage and did not consider what transpired months to years after discharge. More recently, epidemiologic studies linking inpatient events with outpatient outcomes have filled this knowledge gap.⁸ Contemporary animal models of AKI have shed light on potential mechanisms of maladaptive repair after AKI, characterized by fibrosis, vascular rarefaction, tubular loss, glomerulosclerosis, and chronic interstitial inflammation, all of which result in renal function decline. So over the last decade there has been a paradigm shift in how we think about AKI and CKD. Rather than distinct entities, AKI and CKD are now viewed as interconnected syndromes since AKI is a risk factor for CKD progression and CKD is a risk factor for new episodes of AKI.⁹

Two studies published in this issue of the Journal of Hospital Medicine augment our understanding of AKI and its clinical impact in hospitalized patients. Analyzing data from the National Inpatient Sample, Silver et al.¹⁰ found that hospitalizations that include AKI are substantially costlier and associated with longer lengths of stay than hospitalizations without AKI. The authors also highlight that the additional economic costs of AKI exceeded those of many other higher-profile yet less-common acute medical conditions, such as myocardial infarction and gastrointestinal bleeding. These results re-emphasize the important economic burden of AKI at a national level and expand on prior literature by confirming findings previously limited to single-center and regional studies. Better defining the impact AKI has on our healthcare system could help ensure that adequate resources are invested to combat AKI.

The second study, by Rutter et al.,¹¹ found that among hospitalized patients with normal baseline renal function, use of vancomycin in combination with piperacillin-tazobactam is associated with a higher incidence of AKI after antibiotic exposure than use of either agent as monotherapy. This association persisted even after adjusting for potential confounders such as underlying comorbidities, exposure to nephrotoxic agents, documented hypotension, and baseline renal impairment. This study adds to a growing body of literature that suggests synergistic nephrotoxicity between vancomycin and piperacillin-tazobactam. It underscores that any medical intervention—even treatments typically envisioned as non-hazardous and frequently life-saving—involve inherent risks and should prompt the medical community to promote proper antimicrobial stewardship. Whether such exposures to vancomycin or beta-lactam derivatives cause AKI via direct tubular damage, interstitial nephritis, or some other novel mechanism remains to be elucidated. Better delineation of the contemporary causes of AKI, including increased antibiotic exposure, is the first step toward identifying ways to reduce AKI incidence.

Both of these papers serve to highlight the clinical importance of AKI among hospitalized patients. Their findings re-emphasize the need for vigilance in detecting AKI and intervening early to achieve the best clinical outcomes.

Given recent understanding that survivors of AKI are at greater risk for more rapid loss of renal function long after hospital discharge, one goal the US Department of Health and Human Services put forth for Healthy People 2020 is to “increase the proportion of hospital patients who incurred AKI who have follow-up renal evaluation in 6 months post-discharge” (10% improvement targeted).¹² Transitions of care after hospitalizations complicated by AKI require special attention to ensure that patients’ needs are optimally monitored and managed during the critical post-discharge period. One recent study analyzing discharge documentation for hospitalizations including AKI found that fewer than half of the discharge summaries and patient instructions commented on the presence, cause, or course of AKI, indicating clear room for improvement.¹³ And currently, it appears that only a minority of patients with AKI—even AKI severe enough to require dialysis—are seen by a nephrologist within 90 days of discharge.¹⁴

Hospitalists play a crucial role in coordinating care as vulnerable patients transition from the inpatient to outpatient setting. We suggest that AKI should be properly documented in the discharge summary. In addition, patients should be informed that they experienced AKI so they can discuss with future caregivers potential strategies to avoid additional renal insults. Discharge referrals to nephrology should be arranged for high-risk patients, including those whose renal function remains decreased at discharge or those who had recurrent AKI episodes during prior hospitalizations. For patients with pre-hospitalization baseline CKD, nephrology should be consulted before indefinitely discontinuing medications like angiotensin-converting enzyme inhibitors or angiotensin receptor blockers. These medications are indispensable in retarding the progression of proteinuric CKD, even though they may predispose patients to AKI under certain circumstances (eg, in states of decreased renal perfusion). Adopting these simple steps may substantially improve the long-term outcomes of patients who experience AKI during hospitalization.

Acknowledgments

The authors are supported by NIH-NIDDK Grants T32DK007219 (BJL) and K24DK92291 (CYH).

Disclosure

Nothing to report.

HOUSTON – Surgical aortic valve replacement can be performed in intermediate-risk elderly patients with an operative mortality rate of 4.1%, which is better than expected, according to results from a large multicenter analysis. However, the rate of in-hospital stroke was 5.4% – twice what was expected.

“This is most likely secondary to neurologic assessment [that was conducted] for all patients postoperatively,” Vinod H. Thourani, MD, said at the annual meeting of the Society of Thoracic Surgeons.

Doug Brunk/Frontline Medical News

[email protected]

HOUSTON – Surgical aortic valve replacement can be performed in intermediate-risk elderly patients with an operative mortality rate of 4.1%, which is better than expected, according to results from a large multicenter analysis. However, the rate of in-hospital stroke was 5.4% – twice what was expected.

“This is most likely secondary to neurologic assessment [that was conducted] for all patients postoperatively,” Vinod H. Thourani, MD, said at the annual meeting of the Society of Thoracic Surgeons.

Doug Brunk/Frontline Medical News

[email protected]

HOUSTON – Surgical aortic valve replacement can be performed in intermediate-risk elderly patients with an operative mortality rate of 4.1%, which is better than expected, according to results from a large multicenter analysis. However, the rate of in-hospital stroke was 5.4% – twice what was expected.

“This is most likely secondary to neurologic assessment [that was conducted] for all patients postoperatively,” Vinod H. Thourani, MD, said at the annual meeting of the Society of Thoracic Surgeons.

Doug Brunk/Frontline Medical News

[email protected]

The Diagnosis

Carcinosarcoma is a rare malignant entity, representing less than 2% of all esophageal neoplasms. It usually shows a bulky appearance of an intraluminal polypoid lesion owing to predominant sarcomatous development with little stromal proliferation.

References

1. Hung J.J., Li A.F., Liu J.S., et al. Esophageal carcinosarcoma with basaloid squamous cell carcinoma and osteosarcoma. Ann Thorac Surg. 2008;85[3]:1102-4.

2. Madan A.K., Long A.E., Weldon C.B., et al. Esophageal carcinosarcoma. J Gastrointest Surg. 2001;5[4]:414-7.

The Diagnosis

Carcinosarcoma is a rare malignant entity, representing less than 2% of all esophageal neoplasms. It usually shows a bulky appearance of an intraluminal polypoid lesion owing to predominant sarcomatous development with little stromal proliferation.

References

1. Hung J.J., Li A.F., Liu J.S., et al. Esophageal carcinosarcoma with basaloid squamous cell carcinoma and osteosarcoma. Ann Thorac Surg. 2008;85[3]:1102-4.

2. Madan A.K., Long A.E., Weldon C.B., et al. Esophageal carcinosarcoma. J Gastrointest Surg. 2001;5[4]:414-7.

The Diagnosis

Carcinosarcoma is a rare malignant entity, representing less than 2% of all esophageal neoplasms. It usually shows a bulky appearance of an intraluminal polypoid lesion owing to predominant sarcomatous development with little stromal proliferation.

References

1. Hung J.J., Li A.F., Liu J.S., et al. Esophageal carcinosarcoma with basaloid squamous cell carcinoma and osteosarcoma. Ann Thorac Surg. 2008;85[3]:1102-4.

2. Madan A.K., Long A.E., Weldon C.B., et al. Esophageal carcinosarcoma. J Gastrointest Surg. 2001;5[4]:414-7.

Research with mice and rats in Germany may have found a new way to treat high blood pressure.

Using epifluorescence intravital video microscopy imaging, researchers examined mice to whom they had given angiotensin II hormones to induce arterial hypertension.

They determined that the mice with low levels of thrombin-driven factor XI (FXI) – either naturally or inhibited by antisense oligonucleotides – had healthier endothelium.

“Specificity of the effects of FXI depletion was confirmed by continuous in vivo supplementation with human FXI,” wrote Sabine Kossmann, PhD, of the Center for Thrombosis and Hemostasis, University Medical Center, Mainz (Germany), and her coauthors (Sci Transl Med. 2017 Feb 1. doi: 10.1126/scitranslmed.aah4923).

Targeting the “feedback loop” between the FXI and a receptor that helps thrombin propagate on platelets reduced both vascular inflammation and blood pressure.

“Our findings suggest that inhibiting the ... thrombin-FXI–amplifying loop may provide added cardiovascular benefits that are synergistic with those of established platelet inhibitors,” the authors wrote.

This work was supported by grants from the Stiftung Pathobiochemie und Molekulare Diagnostik and the Federal Ministry of Education and Research. The authors disclosed funding and grants from the German Research Society, the European Research Council, NIH, and other sources. One of the researchers is inventor of five patents related to the FXI inhibitor and equity holder in Aronora, and may have financial interest in the findings of the research.

[email protected]

Research with mice and rats in Germany may have found a new way to treat high blood pressure.

Using epifluorescence intravital video microscopy imaging, researchers examined mice to whom they had given angiotensin II hormones to induce arterial hypertension.

They determined that the mice with low levels of thrombin-driven factor XI (FXI) – either naturally or inhibited by antisense oligonucleotides – had healthier endothelium.

“Specificity of the effects of FXI depletion was confirmed by continuous in vivo supplementation with human FXI,” wrote Sabine Kossmann, PhD, of the Center for Thrombosis and Hemostasis, University Medical Center, Mainz (Germany), and her coauthors (Sci Transl Med. 2017 Feb 1. doi: 10.1126/scitranslmed.aah4923).

Targeting the “feedback loop” between the FXI and a receptor that helps thrombin propagate on platelets reduced both vascular inflammation and blood pressure.

“Our findings suggest that inhibiting the ... thrombin-FXI–amplifying loop may provide added cardiovascular benefits that are synergistic with those of established platelet inhibitors,” the authors wrote.

This work was supported by grants from the Stiftung Pathobiochemie und Molekulare Diagnostik and the Federal Ministry of Education and Research. The authors disclosed funding and grants from the German Research Society, the European Research Council, NIH, and other sources. One of the researchers is inventor of five patents related to the FXI inhibitor and equity holder in Aronora, and may have financial interest in the findings of the research.

[email protected]

Research with mice and rats in Germany may have found a new way to treat high blood pressure.

Using epifluorescence intravital video microscopy imaging, researchers examined mice to whom they had given angiotensin II hormones to induce arterial hypertension.

They determined that the mice with low levels of thrombin-driven factor XI (FXI) – either naturally or inhibited by antisense oligonucleotides – had healthier endothelium.

“Specificity of the effects of FXI depletion was confirmed by continuous in vivo supplementation with human FXI,” wrote Sabine Kossmann, PhD, of the Center for Thrombosis and Hemostasis, University Medical Center, Mainz (Germany), and her coauthors (Sci Transl Med. 2017 Feb 1. doi: 10.1126/scitranslmed.aah4923).

Targeting the “feedback loop” between the FXI and a receptor that helps thrombin propagate on platelets reduced both vascular inflammation and blood pressure.

“Our findings suggest that inhibiting the ... thrombin-FXI–amplifying loop may provide added cardiovascular benefits that are synergistic with those of established platelet inhibitors,” the authors wrote.

This work was supported by grants from the Stiftung Pathobiochemie und Molekulare Diagnostik and the Federal Ministry of Education and Research. The authors disclosed funding and grants from the German Research Society, the European Research Council, NIH, and other sources. One of the researchers is inventor of five patents related to the FXI inhibitor and equity holder in Aronora, and may have financial interest in the findings of the research.

[email protected]

Doubt is being cast on the efficacy of Diclegis – the only prescription drug approved in the United States for treating nausea and vomiting in pregnancy – after researchers exposed flaws in previously unpublished data that served as the basis for the drug’s approval.

But the larger point, according to the researcher who brought the unpublished study to light, is the danger of relying too heavily on hidden data.

“It’s not like there’s some special concern over the safety of Diclegis. It’s that there is this commonly prescribed medication that hasn’t been proven to be effective,” Navindra Persaud, MD, a family physician and researcher at St. Michael’s Hospital in Toronto, said in an interview.

©monkeybusinessimages/thinkstockphotos.com

[email protected]

On Twitter @whitneymcknight

Doubt is being cast on the efficacy of Diclegis – the only prescription drug approved in the United States for treating nausea and vomiting in pregnancy – after researchers exposed flaws in previously unpublished data that served as the basis for the drug’s approval.

But the larger point, according to the researcher who brought the unpublished study to light, is the danger of relying too heavily on hidden data.

“It’s not like there’s some special concern over the safety of Diclegis. It’s that there is this commonly prescribed medication that hasn’t been proven to be effective,” Navindra Persaud, MD, a family physician and researcher at St. Michael’s Hospital in Toronto, said in an interview.

©monkeybusinessimages/thinkstockphotos.com

[email protected]

On Twitter @whitneymcknight

Doubt is being cast on the efficacy of Diclegis – the only prescription drug approved in the United States for treating nausea and vomiting in pregnancy – after researchers exposed flaws in previously unpublished data that served as the basis for the drug’s approval.

But the larger point, according to the researcher who brought the unpublished study to light, is the danger of relying too heavily on hidden data.

“It’s not like there’s some special concern over the safety of Diclegis. It’s that there is this commonly prescribed medication that hasn’t been proven to be effective,” Navindra Persaud, MD, a family physician and researcher at St. Michael’s Hospital in Toronto, said in an interview.

©monkeybusinessimages/thinkstockphotos.com

[email protected]

On Twitter @whitneymcknight

CHICAGO – More than 53,000 people will develop pancreatic ductal adenocarcinoma in the United States this year, and upwards of 41,000 will die from the disease, many of them with tumors considered unresectable because they involve adjacent vessels. However, researchers at the University of California, Irvine, have found that careful removal of the tumor around involved veins and arteries, even in borderline cases, can improve outcomes for these patients.

Roy M. Fujitani, MD, updated previously published data on a single-center study he coauthored in 2015 of 270 patients who had undergone a Whipple operation, 183 for pancreatic adenocarcinoma (J Vasc Surg. 2015;61:475-80) at a symposium on vascular surgery sponsored by Northwestern University.

Courtesy Dr. Roy M. Fujitani

CHICAGO – More than 53,000 people will develop pancreatic ductal adenocarcinoma in the United States this year, and upwards of 41,000 will die from the disease, many of them with tumors considered unresectable because they involve adjacent vessels. However, researchers at the University of California, Irvine, have found that careful removal of the tumor around involved veins and arteries, even in borderline cases, can improve outcomes for these patients.

Roy M. Fujitani, MD, updated previously published data on a single-center study he coauthored in 2015 of 270 patients who had undergone a Whipple operation, 183 for pancreatic adenocarcinoma (J Vasc Surg. 2015;61:475-80) at a symposium on vascular surgery sponsored by Northwestern University.

Courtesy Dr. Roy M. Fujitani

CHICAGO – More than 53,000 people will develop pancreatic ductal adenocarcinoma in the United States this year, and upwards of 41,000 will die from the disease, many of them with tumors considered unresectable because they involve adjacent vessels. However, researchers at the University of California, Irvine, have found that careful removal of the tumor around involved veins and arteries, even in borderline cases, can improve outcomes for these patients.

Roy M. Fujitani, MD, updated previously published data on a single-center study he coauthored in 2015 of 270 patients who had undergone a Whipple operation, 183 for pancreatic adenocarcinoma (J Vasc Surg. 2015;61:475-80) at a symposium on vascular surgery sponsored by Northwestern University.

Courtesy Dr. Roy M. Fujitani

Study Overview

Objective. To determine if evolocumab, a PCSK9 inhibitor, affects the progression of coronary artery disease in patients treated with statins.

Design. Multicenter, international, double-blind, placebo-controlled, randomized clinical trial.

Setting and participants. 197 community and academic hospitals worldwide enrolled 978 participants who underwent serial intravascular ultrasounds (IVUS) to measure their burden of coronary atherosclerosis. A total of 2628 patients were screened. Patients were considered for inclusion if they were 18 years of age or older and had at least 1 coronary artery stenosis of at least 20% on a clinically indicated catheterization. Additionally, the target vessel had to meet IVUS imaging quality and visibility standards. Participants were required to have been on stable statin therapy for at least 4 weeks with an LDL level of > 80 mg/dL or between 60–80 mg/dL with either 1 major or 3 minor cardiovascular risk factors. Major risk factors were noncoronary atherosclerotic disease, myocardial infarction (MI) or hospitalization for unstable angina within the past 2 years, or type 2 diabetes. Minor risk factors included current tobacco use, hypertension, low HDL-C levels, family history of early coronary disease, hsCRP level of 2 mg/L or greater, and age older than 50 years for men and 55 years for women. Patients with uncontrolled hypertension, uncontrolled diabetes, heart failure, renal insufficiency, or liver disease were excluded.

Intervention. Patients were randomized to either treatment with monthly subcutaneous injections of 420 mg evolocumab or placebo injections for 76 weeks. Participants attended 7 follow-up visits during the study period and then underwent repeat IVUS imaging at the 78th week. Research staff, who were blinded to both treatment status and imaging sequence, collected and assessed target vessel measurements, including the vessel lumen and external elastic membrane dimensions. IVUS imaging has been used in numerous clinical studies and has been shown to be accurate and reliable [1].

Main outcome measures. The primary outcome was the target artery change in percent atheroma volume (PAV) from baseline to week 78. PAV was calculated from IVUS measurements. Nominal change in PAV was then determined by calculating the difference of the PAV at baseline and at week 78.

The secondary measure was the normalized total atheroma volume (TAV). TAV addresses variability in the length of vessel segments and the number of images collected during IVUS catheter pullback. The nominal change in TAV was then determined by the difference at baseline and at week 78.

Additional secondary efficacy endpoints included number of patients with regression of plaque and change in lipid parameters. Safety outcomes were investigated through evaluation of the incidence of adjudicated clinical events, including all-cause mortality, cardiovascular death, MI, unstable angina requiring hospitalization, coronary revascularization, stroke, transient ischemic attack, and heart failure requiring hospitalization. Post-hoc analysis compared baseline LDL-C level and change in PAV and regression of PAV. The association between LDL lowering and plaque progression was also assessed post hoc.

IVUS measurements were evaluated as least squares means. Comparison of treatment groups was conducted using analysis of covariance on rank transformed data that accounted for baseline value and geographic location. Investigators used a step-down statistical procedure to evaluate primary and secondary endpoints. The statistical model accounted for confounders such as baseline LDL-C, baseline PAV, intensity of statin therapy, geographic region, age, and sex.

Main results. 484 participants were randomized to the evolocumab group and 484 to the placebo group, and 423 participants in both groups completed both baseline and follow-up IVUS imaging. Treatment and control groups contained participants matched for age, gender, ethnicity, cardiovascular risk factors, and baseline medication use, including lipid-lowering agents, ACE inhibitors, ARBs, beta-blockers, and antiplatelet therapies. Both groups consisted of a majority of white (93.4% in placebo and 94.2% in treatment) males (72.3% in placebo and 72.1% in treatment). Approximately 80% of participants had hypertension (83.7% in placebo and 82.2% in treatment), about 35% had prior MIs (35.3% in placebo and 34.9% in treatment), and roughly a fifth of participants had diabetes (21.5% in placebo and 20.2% in treatment). At baseline 98.6% of participants were treated with statins, with 58.9% on high-intensity therapy and 39.4% on moderate-intensity. Mean LDL-C level at baseline was 92.5 (SD, 27.2) mg/dL.

After 76 weeks of treatment, mean LDL-C level in the placebo group was 93.0 mg/dL and 36.6 mg/dL in the treatment group, which corresponds to a 0.2 mg/dL increase in the placebo group and a 56.3 mg/dL reduction in the treatment group. The change in LDL-C level was statistically significant (P < 0.001).

Placebo group participants had no significant change in PAV (0.05%, P = 0.78), but the evolocumab group experienced a 0.95% decrease from baseline (P < 0.001). Similarly, the placebo group had no change in TAV from baseline (–0.9 mm³, P = 0.45), but the treatment group had a 5.8 mm³ reduction in TAV from baseline (P < 0.001). The treatment group had a greater proportion of patients who experienced PAV regression (64.3% vs. 47.3%, P < 0.001) and TAV regression (61.5% vs. 48.9%, P < 0.001).

Subgroup analysis did not demonstrate a significant association between change in PAV and specific study participant characteristics (eg, age, gender, ethnicity).

Post-hoc analysis using local regression (LOESS) curve revealed a linear relationship between achieved LDL-C level and change in PAV for LDL-C levels from 110 mg/dL to 20 mg/dL.

The treatment group did not exhibit a significant increase in adverse drug events, which included injection site reactions, myalgias, neurocognitive events, and incidence of diabetes. There was no significant difference in adverse cardiovascular outcomes between groups; however, there were numerically fewer nonfatal MIs and coronary revascularizations in the treatment group.

Conclusion. The use of evolocumab in statin-treated patients resulted in greater reduction of PAV than use of statins alone.

Commentary

Evolocumab is a monoclonal antibody that inhibits pro-protein convertase subtilisin-kexin type 9 (PCSK9), which is involved in LDL-C receptor recycling. By reducing removal of LDL-C receptors, evolocumab amplifies LDL-C clearance and has been shown to reduce LDL-C levels by approximately 61% from baseline with 12 weeks oftreatment [2]. Studies have shown that the lipid-lowering potential of evolocumab is superior to statins alone and to combination therapy with statins and ezetimibe [2]. Furthermore, PCSK9 inhibitors have been effective at LDL-lowering in patients who failed or could not tolerate standard of care therapy with statins and ezetimibe [3,4]. PCSK9 inhibitors hold great promise for reducing morbidity and mortality of cardiovascular disease; however, LDL-lowering is not equivalent to improved clinical outcomes.

The GLAGOV study moves toward demonstration of the clinical benefit of evolocumab. The study shows that combined therapy with statins and evolocumab, versus statins alone, not only achieves better stability of atherosclerotic plaque dimensions but actually results in regression of plaque size. In the study, plaque burden is extrapolated from vessel measurements obtained through IVUS, and nominal changes in PAV and TAV serve as markers for atherosclerosis, but these surrogates cannot be equated to a reduction in cardiovascular events. The GLAGOV trial does explore clinical outcomes such as MI, stroke, unstable angina, coronary revascularization, and death; however, the study is not powered to evaluate the statistical significance of these events. We await sufficiently powered phase 3 clinical trials to determine the clinical benefits of PCSK9 inhibitors on cardiovascular disease.

The GLAGOV trial has several strengths, including its design as an international, double-blind, placebo-controlled, randomized clinical trial. The intervention is simple and the outcomes are clearly defined. The statistical assessment yields significant results. Nonetheless, there are multiple limitations to the study. The lead author has received research support from Amgen, the maker of evolocumab. Amgen also participated in study design and maintenance of trial databases; however, data analysis was conducted by an independent statistician. Additionally, the majority of study participants were white males with very few minority patients despite inclusion of study sites around the globe. The homogeneity of the study cohort makes the data difficult to generalize to a larger population. Similarly, patients who lacked a clinical indication for coronary catheterization and those with uncontrolled diabetes, hypertension, and heart failure were excluded, which further limits application of this study to many patients with atherosclerosis. Another limitation is study attrition; only 87% of participants completed the 78-week IVUS and were included in the data analysis, and results may have differed if those lost to follow-up had completed the trial. Furthermore, study duration was limited to 76 weeks and the magnitude and durability of study outcomes after this time point remain unknown.

Applications for Clinical Practice

Reduction in PAV and TAV are surrogate endpoints and are not indicative of a clinical benefit. Nonetheless, the GLAGOV study demonstrates that evolocumab, when used in conjunction with statins, can promote regression of atherosclerosis greater than treatment with statins alone. More studies are needed to evaluate a clinical benefit of adding evolocumab to the regularly used arsenal of lipid-lowering therapies for the treatment of atherosclerosis. Furthermore, cost-effectiveness of evolocumab has not been shown. In 2015 the yearly wholesale price of evolcumab was $14,350. A cost-effectiveness analysis based on this price estimates that treatment of atherosclerotic coronary vascular disease with evolocumab has a cost of $414,000 per quality-adjusted life year [5]. Evolocumab is well tolerated, but additional studies for cardiovascular and mortality outcomes are needed before it can be considered part of the standard of treatment for coronary artery disease.

—Lauren Brooks, MD, University of Maryland School of Medicine, Baltimore, MD

Study Overview

Objective. To determine if evolocumab, a PCSK9 inhibitor, affects the progression of coronary artery disease in patients treated with statins.

Design. Multicenter, international, double-blind, placebo-controlled, randomized clinical trial.

Setting and participants. 197 community and academic hospitals worldwide enrolled 978 participants who underwent serial intravascular ultrasounds (IVUS) to measure their burden of coronary atherosclerosis. A total of 2628 patients were screened. Patients were considered for inclusion if they were 18 years of age or older and had at least 1 coronary artery stenosis of at least 20% on a clinically indicated catheterization. Additionally, the target vessel had to meet IVUS imaging quality and visibility standards. Participants were required to have been on stable statin therapy for at least 4 weeks with an LDL level of > 80 mg/dL or between 60–80 mg/dL with either 1 major or 3 minor cardiovascular risk factors. Major risk factors were noncoronary atherosclerotic disease, myocardial infarction (MI) or hospitalization for unstable angina within the past 2 years, or type 2 diabetes. Minor risk factors included current tobacco use, hypertension, low HDL-C levels, family history of early coronary disease, hsCRP level of 2 mg/L or greater, and age older than 50 years for men and 55 years for women. Patients with uncontrolled hypertension, uncontrolled diabetes, heart failure, renal insufficiency, or liver disease were excluded.

Intervention. Patients were randomized to either treatment with monthly subcutaneous injections of 420 mg evolocumab or placebo injections for 76 weeks. Participants attended 7 follow-up visits during the study period and then underwent repeat IVUS imaging at the 78th week. Research staff, who were blinded to both treatment status and imaging sequence, collected and assessed target vessel measurements, including the vessel lumen and external elastic membrane dimensions. IVUS imaging has been used in numerous clinical studies and has been shown to be accurate and reliable [1].

Main outcome measures. The primary outcome was the target artery change in percent atheroma volume (PAV) from baseline to week 78. PAV was calculated from IVUS measurements. Nominal change in PAV was then determined by calculating the difference of the PAV at baseline and at week 78.

The secondary measure was the normalized total atheroma volume (TAV). TAV addresses variability in the length of vessel segments and the number of images collected during IVUS catheter pullback. The nominal change in TAV was then determined by the difference at baseline and at week 78.

Additional secondary efficacy endpoints included number of patients with regression of plaque and change in lipid parameters. Safety outcomes were investigated through evaluation of the incidence of adjudicated clinical events, including all-cause mortality, cardiovascular death, MI, unstable angina requiring hospitalization, coronary revascularization, stroke, transient ischemic attack, and heart failure requiring hospitalization. Post-hoc analysis compared baseline LDL-C level and change in PAV and regression of PAV. The association between LDL lowering and plaque progression was also assessed post hoc.

IVUS measurements were evaluated as least squares means. Comparison of treatment groups was conducted using analysis of covariance on rank transformed data that accounted for baseline value and geographic location. Investigators used a step-down statistical procedure to evaluate primary and secondary endpoints. The statistical model accounted for confounders such as baseline LDL-C, baseline PAV, intensity of statin therapy, geographic region, age, and sex.

Main results. 484 participants were randomized to the evolocumab group and 484 to the placebo group, and 423 participants in both groups completed both baseline and follow-up IVUS imaging. Treatment and control groups contained participants matched for age, gender, ethnicity, cardiovascular risk factors, and baseline medication use, including lipid-lowering agents, ACE inhibitors, ARBs, beta-blockers, and antiplatelet therapies. Both groups consisted of a majority of white (93.4% in placebo and 94.2% in treatment) males (72.3% in placebo and 72.1% in treatment). Approximately 80% of participants had hypertension (83.7% in placebo and 82.2% in treatment), about 35% had prior MIs (35.3% in placebo and 34.9% in treatment), and roughly a fifth of participants had diabetes (21.5% in placebo and 20.2% in treatment). At baseline 98.6% of participants were treated with statins, with 58.9% on high-intensity therapy and 39.4% on moderate-intensity. Mean LDL-C level at baseline was 92.5 (SD, 27.2) mg/dL.

After 76 weeks of treatment, mean LDL-C level in the placebo group was 93.0 mg/dL and 36.6 mg/dL in the treatment group, which corresponds to a 0.2 mg/dL increase in the placebo group and a 56.3 mg/dL reduction in the treatment group. The change in LDL-C level was statistically significant (P < 0.001).

Placebo group participants had no significant change in PAV (0.05%, P = 0.78), but the evolocumab group experienced a 0.95% decrease from baseline (P < 0.001). Similarly, the placebo group had no change in TAV from baseline (–0.9 mm³, P = 0.45), but the treatment group had a 5.8 mm³ reduction in TAV from baseline (P < 0.001). The treatment group had a greater proportion of patients who experienced PAV regression (64.3% vs. 47.3%, P < 0.001) and TAV regression (61.5% vs. 48.9%, P < 0.001).

Subgroup analysis did not demonstrate a significant association between change in PAV and specific study participant characteristics (eg, age, gender, ethnicity).

Post-hoc analysis using local regression (LOESS) curve revealed a linear relationship between achieved LDL-C level and change in PAV for LDL-C levels from 110 mg/dL to 20 mg/dL.

The treatment group did not exhibit a significant increase in adverse drug events, which included injection site reactions, myalgias, neurocognitive events, and incidence of diabetes. There was no significant difference in adverse cardiovascular outcomes between groups; however, there were numerically fewer nonfatal MIs and coronary revascularizations in the treatment group.

Conclusion. The use of evolocumab in statin-treated patients resulted in greater reduction of PAV than use of statins alone.

Commentary

Evolocumab is a monoclonal antibody that inhibits pro-protein convertase subtilisin-kexin type 9 (PCSK9), which is involved in LDL-C receptor recycling. By reducing removal of LDL-C receptors, evolocumab amplifies LDL-C clearance and has been shown to reduce LDL-C levels by approximately 61% from baseline with 12 weeks oftreatment [2]. Studies have shown that the lipid-lowering potential of evolocumab is superior to statins alone and to combination therapy with statins and ezetimibe [2]. Furthermore, PCSK9 inhibitors have been effective at LDL-lowering in patients who failed or could not tolerate standard of care therapy with statins and ezetimibe [3,4]. PCSK9 inhibitors hold great promise for reducing morbidity and mortality of cardiovascular disease; however, LDL-lowering is not equivalent to improved clinical outcomes.

The GLAGOV study moves toward demonstration of the clinical benefit of evolocumab. The study shows that combined therapy with statins and evolocumab, versus statins alone, not only achieves better stability of atherosclerotic plaque dimensions but actually results in regression of plaque size. In the study, plaque burden is extrapolated from vessel measurements obtained through IVUS, and nominal changes in PAV and TAV serve as markers for atherosclerosis, but these surrogates cannot be equated to a reduction in cardiovascular events. The GLAGOV trial does explore clinical outcomes such as MI, stroke, unstable angina, coronary revascularization, and death; however, the study is not powered to evaluate the statistical significance of these events. We await sufficiently powered phase 3 clinical trials to determine the clinical benefits of PCSK9 inhibitors on cardiovascular disease.

The GLAGOV trial has several strengths, including its design as an international, double-blind, placebo-controlled, randomized clinical trial. The intervention is simple and the outcomes are clearly defined. The statistical assessment yields significant results. Nonetheless, there are multiple limitations to the study. The lead author has received research support from Amgen, the maker of evolocumab. Amgen also participated in study design and maintenance of trial databases; however, data analysis was conducted by an independent statistician. Additionally, the majority of study participants were white males with very few minority patients despite inclusion of study sites around the globe. The homogeneity of the study cohort makes the data difficult to generalize to a larger population. Similarly, patients who lacked a clinical indication for coronary catheterization and those with uncontrolled diabetes, hypertension, and heart failure were excluded, which further limits application of this study to many patients with atherosclerosis. Another limitation is study attrition; only 87% of participants completed the 78-week IVUS and were included in the data analysis, and results may have differed if those lost to follow-up had completed the trial. Furthermore, study duration was limited to 76 weeks and the magnitude and durability of study outcomes after this time point remain unknown.

Applications for Clinical Practice

Reduction in PAV and TAV are surrogate endpoints and are not indicative of a clinical benefit. Nonetheless, the GLAGOV study demonstrates that evolocumab, when used in conjunction with statins, can promote regression of atherosclerosis greater than treatment with statins alone. More studies are needed to evaluate a clinical benefit of adding evolocumab to the regularly used arsenal of lipid-lowering therapies for the treatment of atherosclerosis. Furthermore, cost-effectiveness of evolocumab has not been shown. In 2015 the yearly wholesale price of evolcumab was $14,350. A cost-effectiveness analysis based on this price estimates that treatment of atherosclerotic coronary vascular disease with evolocumab has a cost of $414,000 per quality-adjusted life year [5]. Evolocumab is well tolerated, but additional studies for cardiovascular and mortality outcomes are needed before it can be considered part of the standard of treatment for coronary artery disease.

—Lauren Brooks, MD, University of Maryland School of Medicine, Baltimore, MD

Study Overview

Objective. To determine if evolocumab, a PCSK9 inhibitor, affects the progression of coronary artery disease in patients treated with statins.

Design. Multicenter, international, double-blind, placebo-controlled, randomized clinical trial.

Setting and participants. 197 community and academic hospitals worldwide enrolled 978 participants who underwent serial intravascular ultrasounds (IVUS) to measure their burden of coronary atherosclerosis. A total of 2628 patients were screened. Patients were considered for inclusion if they were 18 years of age or older and had at least 1 coronary artery stenosis of at least 20% on a clinically indicated catheterization. Additionally, the target vessel had to meet IVUS imaging quality and visibility standards. Participants were required to have been on stable statin therapy for at least 4 weeks with an LDL level of > 80 mg/dL or between 60–80 mg/dL with either 1 major or 3 minor cardiovascular risk factors. Major risk factors were noncoronary atherosclerotic disease, myocardial infarction (MI) or hospitalization for unstable angina within the past 2 years, or type 2 diabetes. Minor risk factors included current tobacco use, hypertension, low HDL-C levels, family history of early coronary disease, hsCRP level of 2 mg/L or greater, and age older than 50 years for men and 55 years for women. Patients with uncontrolled hypertension, uncontrolled diabetes, heart failure, renal insufficiency, or liver disease were excluded.

Intervention. Patients were randomized to either treatment with monthly subcutaneous injections of 420 mg evolocumab or placebo injections for 76 weeks. Participants attended 7 follow-up visits during the study period and then underwent repeat IVUS imaging at the 78th week. Research staff, who were blinded to both treatment status and imaging sequence, collected and assessed target vessel measurements, including the vessel lumen and external elastic membrane dimensions. IVUS imaging has been used in numerous clinical studies and has been shown to be accurate and reliable [1].

Main outcome measures. The primary outcome was the target artery change in percent atheroma volume (PAV) from baseline to week 78. PAV was calculated from IVUS measurements. Nominal change in PAV was then determined by calculating the difference of the PAV at baseline and at week 78.

The secondary measure was the normalized total atheroma volume (TAV). TAV addresses variability in the length of vessel segments and the number of images collected during IVUS catheter pullback. The nominal change in TAV was then determined by the difference at baseline and at week 78.

Additional secondary efficacy endpoints included number of patients with regression of plaque and change in lipid parameters. Safety outcomes were investigated through evaluation of the incidence of adjudicated clinical events, including all-cause mortality, cardiovascular death, MI, unstable angina requiring hospitalization, coronary revascularization, stroke, transient ischemic attack, and heart failure requiring hospitalization. Post-hoc analysis compared baseline LDL-C level and change in PAV and regression of PAV. The association between LDL lowering and plaque progression was also assessed post hoc.

IVUS measurements were evaluated as least squares means. Comparison of treatment groups was conducted using analysis of covariance on rank transformed data that accounted for baseline value and geographic location. Investigators used a step-down statistical procedure to evaluate primary and secondary endpoints. The statistical model accounted for confounders such as baseline LDL-C, baseline PAV, intensity of statin therapy, geographic region, age, and sex.

Main results. 484 participants were randomized to the evolocumab group and 484 to the placebo group, and 423 participants in both groups completed both baseline and follow-up IVUS imaging. Treatment and control groups contained participants matched for age, gender, ethnicity, cardiovascular risk factors, and baseline medication use, including lipid-lowering agents, ACE inhibitors, ARBs, beta-blockers, and antiplatelet therapies. Both groups consisted of a majority of white (93.4% in placebo and 94.2% in treatment) males (72.3% in placebo and 72.1% in treatment). Approximately 80% of participants had hypertension (83.7% in placebo and 82.2% in treatment), about 35% had prior MIs (35.3% in placebo and 34.9% in treatment), and roughly a fifth of participants had diabetes (21.5% in placebo and 20.2% in treatment). At baseline 98.6% of participants were treated with statins, with 58.9% on high-intensity therapy and 39.4% on moderate-intensity. Mean LDL-C level at baseline was 92.5 (SD, 27.2) mg/dL.

After 76 weeks of treatment, mean LDL-C level in the placebo group was 93.0 mg/dL and 36.6 mg/dL in the treatment group, which corresponds to a 0.2 mg/dL increase in the placebo group and a 56.3 mg/dL reduction in the treatment group. The change in LDL-C level was statistically significant (P < 0.001).

Placebo group participants had no significant change in PAV (0.05%, P = 0.78), but the evolocumab group experienced a 0.95% decrease from baseline (P < 0.001). Similarly, the placebo group had no change in TAV from baseline (–0.9 mm³, P = 0.45), but the treatment group had a 5.8 mm³ reduction in TAV from baseline (P < 0.001). The treatment group had a greater proportion of patients who experienced PAV regression (64.3% vs. 47.3%, P < 0.001) and TAV regression (61.5% vs. 48.9%, P < 0.001).

Subgroup analysis did not demonstrate a significant association between change in PAV and specific study participant characteristics (eg, age, gender, ethnicity).

Post-hoc analysis using local regression (LOESS) curve revealed a linear relationship between achieved LDL-C level and change in PAV for LDL-C levels from 110 mg/dL to 20 mg/dL.

The treatment group did not exhibit a significant increase in adverse drug events, which included injection site reactions, myalgias, neurocognitive events, and incidence of diabetes. There was no significant difference in adverse cardiovascular outcomes between groups; however, there were numerically fewer nonfatal MIs and coronary revascularizations in the treatment group.

Conclusion. The use of evolocumab in statin-treated patients resulted in greater reduction of PAV than use of statins alone.

Commentary

Evolocumab is a monoclonal antibody that inhibits pro-protein convertase subtilisin-kexin type 9 (PCSK9), which is involved in LDL-C receptor recycling. By reducing removal of LDL-C receptors, evolocumab amplifies LDL-C clearance and has been shown to reduce LDL-C levels by approximately 61% from baseline with 12 weeks oftreatment [2]. Studies have shown that the lipid-lowering potential of evolocumab is superior to statins alone and to combination therapy with statins and ezetimibe [2]. Furthermore, PCSK9 inhibitors have been effective at LDL-lowering in patients who failed or could not tolerate standard of care therapy with statins and ezetimibe [3,4]. PCSK9 inhibitors hold great promise for reducing morbidity and mortality of cardiovascular disease; however, LDL-lowering is not equivalent to improved clinical outcomes.

The GLAGOV study moves toward demonstration of the clinical benefit of evolocumab. The study shows that combined therapy with statins and evolocumab, versus statins alone, not only achieves better stability of atherosclerotic plaque dimensions but actually results in regression of plaque size. In the study, plaque burden is extrapolated from vessel measurements obtained through IVUS, and nominal changes in PAV and TAV serve as markers for atherosclerosis, but these surrogates cannot be equated to a reduction in cardiovascular events. The GLAGOV trial does explore clinical outcomes such as MI, stroke, unstable angina, coronary revascularization, and death; however, the study is not powered to evaluate the statistical significance of these events. We await sufficiently powered phase 3 clinical trials to determine the clinical benefits of PCSK9 inhibitors on cardiovascular disease.

The GLAGOV trial has several strengths, including its design as an international, double-blind, placebo-controlled, randomized clinical trial. The intervention is simple and the outcomes are clearly defined. The statistical assessment yields significant results. Nonetheless, there are multiple limitations to the study. The lead author has received research support from Amgen, the maker of evolocumab. Amgen also participated in study design and maintenance of trial databases; however, data analysis was conducted by an independent statistician. Additionally, the majority of study participants were white males with very few minority patients despite inclusion of study sites around the globe. The homogeneity of the study cohort makes the data difficult to generalize to a larger population. Similarly, patients who lacked a clinical indication for coronary catheterization and those with uncontrolled diabetes, hypertension, and heart failure were excluded, which further limits application of this study to many patients with atherosclerosis. Another limitation is study attrition; only 87% of participants completed the 78-week IVUS and were included in the data analysis, and results may have differed if those lost to follow-up had completed the trial. Furthermore, study duration was limited to 76 weeks and the magnitude and durability of study outcomes after this time point remain unknown.

Applications for Clinical Practice

Reduction in PAV and TAV are surrogate endpoints and are not indicative of a clinical benefit. Nonetheless, the GLAGOV study demonstrates that evolocumab, when used in conjunction with statins, can promote regression of atherosclerosis greater than treatment with statins alone. More studies are needed to evaluate a clinical benefit of adding evolocumab to the regularly used arsenal of lipid-lowering therapies for the treatment of atherosclerosis. Furthermore, cost-effectiveness of evolocumab has not been shown. In 2015 the yearly wholesale price of evolcumab was $14,350. A cost-effectiveness analysis based on this price estimates that treatment of atherosclerotic coronary vascular disease with evolocumab has a cost of $414,000 per quality-adjusted life year [5]. Evolocumab is well tolerated, but additional studies for cardiovascular and mortality outcomes are needed before it can be considered part of the standard of treatment for coronary artery disease.

—Lauren Brooks, MD, University of Maryland School of Medicine, Baltimore, MD

SAN FRANCISCO – Post-surgery adjuvant treatment with gemcitabine and oxaliplatin (GEMOX) was feasible, but failed to significantly improve relapse-free survival when compared with surveillance among patients with localized biliary tract cancer in the randomized phase III PRODIGE 12-ACCORD 18 (UNICANCER GI) trial.

Relapse-free survival at a median of 44.3 months in 196 patients who were randomized within 3 months of R0 or R1 resection of a localized biliary tract cancer to receive either GEMOX for 12 cycles or surveillance was 30.4 months vs. 22.0 months in the groups, respectively. At 4 years, relapse-free survival was 39.3% and 33.2%, respectively, Julien Edeline, MD, of Oncology Medical Eugene Marquis Comprehensive Cancer Center, Rennes, France, reported at the symposium, sponsored by ASCO, ASTRO, the American Gastroenterological Association, and Society of Surgical Oncology.

No difference was seen between the groups with respect to the co-primary endpoint of 12- and 24-month global health-related quality of life scores (70.8 vs. 83.3, and 75.0 vs. 83.3, respectively), he said.

Study subjects had localized intra-hepatic, perihilar, or extra-hepatic cholangiocarcinoma, or gallbladder cancer and were enrolled from 33 centers between July 2009 and February 2014. They had ECOG performance status of 0-2, and adequate liver function. The treatment and surveillance arms were well balanced, with similar primary disease sites, Dr. Edeline said.

Those in the GEMOX arm received 12 cycles (6 months) of gemcitabine 1,000 mg/m² on day 1 and oxaliplatin 85 mg/m² on day 2. Those in the surveillance arm underwent ACE, CA19.9 testing, and CT scans every 3 months for 2 years then every 6 months for 3 years.

In the treatment and surveillance arms, respectively, R0 resection rates were 86.2% and 87.9%, and lymph node invasion was present in 37.2% and 36.4%.

The maximal grade of adverse events was 3 in 57.5% vs. 22.2% of patients in the groups, respectively, and grade 4 in 17.0% vs. 9.1%. One patient in each arm died during treatment.

The main grade 3 or greater adverse events were peripheral neuropathy in 50.0% vs. 1.1% and neutropenia in 22.3% vs. 0% for GEMOX vs. surveillance group patients.

“As you know, there is a high risk of relapse following surgery for localized biliary tract cancer. There is currently no proven adjuvant or neoadjuvant treatment,” Dr. Edeline said. “In the palliative setting, we know that the combination of gemcitabine and cisplatin improves overall survival. GEMOX is considered an active regimen based on data from phase II trials. At the time of the design of our study, GEMOX was considered the standard first line treatment for biliary tract cancer.”

Based on this background, the aim of the current phase III trial was to assess whether GEMOX would increase relapse-free survival while maintaining health-related quality of life in patients with localized disease.

“We showed that adjuvant GEMOX was feasible. Toxicities were as expected and manageable, and we didn’t see detrimental effects on quality of life. However, adjuvant GEMOX was not associated in the PRODIGE 12 trial with an improvement in relapse-free survival,” he said, noting that this was also true in subgroup analyses, which showed no benefit of GEMOX with respect to relapse-free survival in any predefined subgroups.

“Clearly, further research through international collaboration is required to improve outcomes in these patients,” he concluded.

Dr. Edeline reported having no disclosures.

[email protected]

SAN FRANCISCO – Post-surgery adjuvant treatment with gemcitabine and oxaliplatin (GEMOX) was feasible, but failed to significantly improve relapse-free survival when compared with surveillance among patients with localized biliary tract cancer in the randomized phase III PRODIGE 12-ACCORD 18 (UNICANCER GI) trial.

Relapse-free survival at a median of 44.3 months in 196 patients who were randomized within 3 months of R0 or R1 resection of a localized biliary tract cancer to receive either GEMOX for 12 cycles or surveillance was 30.4 months vs. 22.0 months in the groups, respectively. At 4 years, relapse-free survival was 39.3% and 33.2%, respectively, Julien Edeline, MD, of Oncology Medical Eugene Marquis Comprehensive Cancer Center, Rennes, France, reported at the symposium, sponsored by ASCO, ASTRO, the American Gastroenterological Association, and Society of Surgical Oncology.

No difference was seen between the groups with respect to the co-primary endpoint of 12- and 24-month global health-related quality of life scores (70.8 vs. 83.3, and 75.0 vs. 83.3, respectively), he said.

Study subjects had localized intra-hepatic, perihilar, or extra-hepatic cholangiocarcinoma, or gallbladder cancer and were enrolled from 33 centers between July 2009 and February 2014. They had ECOG performance status of 0-2, and adequate liver function. The treatment and surveillance arms were well balanced, with similar primary disease sites, Dr. Edeline said.

Those in the GEMOX arm received 12 cycles (6 months) of gemcitabine 1,000 mg/m² on day 1 and oxaliplatin 85 mg/m² on day 2. Those in the surveillance arm underwent ACE, CA19.9 testing, and CT scans every 3 months for 2 years then every 6 months for 3 years.

In the treatment and surveillance arms, respectively, R0 resection rates were 86.2% and 87.9%, and lymph node invasion was present in 37.2% and 36.4%.

The maximal grade of adverse events was 3 in 57.5% vs. 22.2% of patients in the groups, respectively, and grade 4 in 17.0% vs. 9.1%. One patient in each arm died during treatment.

The main grade 3 or greater adverse events were peripheral neuropathy in 50.0% vs. 1.1% and neutropenia in 22.3% vs. 0% for GEMOX vs. surveillance group patients.

“As you know, there is a high risk of relapse following surgery for localized biliary tract cancer. There is currently no proven adjuvant or neoadjuvant treatment,” Dr. Edeline said. “In the palliative setting, we know that the combination of gemcitabine and cisplatin improves overall survival. GEMOX is considered an active regimen based on data from phase II trials. At the time of the design of our study, GEMOX was considered the standard first line treatment for biliary tract cancer.”

Based on this background, the aim of the current phase III trial was to assess whether GEMOX would increase relapse-free survival while maintaining health-related quality of life in patients with localized disease.

“We showed that adjuvant GEMOX was feasible. Toxicities were as expected and manageable, and we didn’t see detrimental effects on quality of life. However, adjuvant GEMOX was not associated in the PRODIGE 12 trial with an improvement in relapse-free survival,” he said, noting that this was also true in subgroup analyses, which showed no benefit of GEMOX with respect to relapse-free survival in any predefined subgroups.

“Clearly, further research through international collaboration is required to improve outcomes in these patients,” he concluded.

Dr. Edeline reported having no disclosures.

[email protected]

SAN FRANCISCO – Post-surgery adjuvant treatment with gemcitabine and oxaliplatin (GEMOX) was feasible, but failed to significantly improve relapse-free survival when compared with surveillance among patients with localized biliary tract cancer in the randomized phase III PRODIGE 12-ACCORD 18 (UNICANCER GI) trial.

Relapse-free survival at a median of 44.3 months in 196 patients who were randomized within 3 months of R0 or R1 resection of a localized biliary tract cancer to receive either GEMOX for 12 cycles or surveillance was 30.4 months vs. 22.0 months in the groups, respectively. At 4 years, relapse-free survival was 39.3% and 33.2%, respectively, Julien Edeline, MD, of Oncology Medical Eugene Marquis Comprehensive Cancer Center, Rennes, France, reported at the symposium, sponsored by ASCO, ASTRO, the American Gastroenterological Association, and Society of Surgical Oncology.

No difference was seen between the groups with respect to the co-primary endpoint of 12- and 24-month global health-related quality of life scores (70.8 vs. 83.3, and 75.0 vs. 83.3, respectively), he said.

Study subjects had localized intra-hepatic, perihilar, or extra-hepatic cholangiocarcinoma, or gallbladder cancer and were enrolled from 33 centers between July 2009 and February 2014. They had ECOG performance status of 0-2, and adequate liver function. The treatment and surveillance arms were well balanced, with similar primary disease sites, Dr. Edeline said.

Those in the GEMOX arm received 12 cycles (6 months) of gemcitabine 1,000 mg/m² on day 1 and oxaliplatin 85 mg/m² on day 2. Those in the surveillance arm underwent ACE, CA19.9 testing, and CT scans every 3 months for 2 years then every 6 months for 3 years.

In the treatment and surveillance arms, respectively, R0 resection rates were 86.2% and 87.9%, and lymph node invasion was present in 37.2% and 36.4%.

The maximal grade of adverse events was 3 in 57.5% vs. 22.2% of patients in the groups, respectively, and grade 4 in 17.0% vs. 9.1%. One patient in each arm died during treatment.

The main grade 3 or greater adverse events were peripheral neuropathy in 50.0% vs. 1.1% and neutropenia in 22.3% vs. 0% for GEMOX vs. surveillance group patients.

“As you know, there is a high risk of relapse following surgery for localized biliary tract cancer. There is currently no proven adjuvant or neoadjuvant treatment,” Dr. Edeline said. “In the palliative setting, we know that the combination of gemcitabine and cisplatin improves overall survival. GEMOX is considered an active regimen based on data from phase II trials. At the time of the design of our study, GEMOX was considered the standard first line treatment for biliary tract cancer.”

Based on this background, the aim of the current phase III trial was to assess whether GEMOX would increase relapse-free survival while maintaining health-related quality of life in patients with localized disease.

“We showed that adjuvant GEMOX was feasible. Toxicities were as expected and manageable, and we didn’t see detrimental effects on quality of life. However, adjuvant GEMOX was not associated in the PRODIGE 12 trial with an improvement in relapse-free survival,” he said, noting that this was also true in subgroup analyses, which showed no benefit of GEMOX with respect to relapse-free survival in any predefined subgroups.

“Clearly, further research through international collaboration is required to improve outcomes in these patients,” he concluded.

Dr. Edeline reported having no disclosures.

[email protected]