Association Between Depression and Death in People With CKD: A Meta-analysis of Cohort Studies

Original Investigation Association Between Depression and Death in People With CKD: A Meta-analysis of Cohort Studies Suetonia C. Palmer, MBChB, PhD,1 Mariacristina Vecchio, MSc,2 Jonathan C. Craig, MBChB, PhD,3 Marcello Tonelli, MD,4 David W. Johnson, MBBS (Hons), PhD,5 Antonio Nicolucci, MD,2 Fabio Pellegrini, MSc,2,6 Valeria Saglimbene, MSc,2 Giancarlo Logroscino, PhD,7 S. Susan Hedayati, MD,8 and Giovanni F.M. Strippoli, MD, MM, MPH, PhD2,3,9,10 Background: Depression occurs relatively commonly in people with chronic kidney disease (CKD), but it is uncertain whether depression is a risk factor for premature death in this population. Interventions to reduce mortality in CKD consistently have been ineffective and new strategies are needed. Study Design: Systematic review and meta-analysis of cohort studies. Setting & Population: Adults with CKD. Selection Criteria for Studies: Cohort studies identified in Ovid MEDLINE through week 3 December 2012 without language restriction. Predictor: Depression status as determined by physician diagnosis, clinical coding, or self-reported scales. Selection Criteria for Studies: All-cause and cardiovascular mortality. Outcomes were summarized as relative risks (RRs) with 95% CIs using random-effects meta-analysis. Results: 22 studies (83,381 participants) comprising 12,063 cases of depression (mean prevalence, 27.4%; 95% CI, 20.0%-36.3%) with a follow-up of 3 months to 6.5 years were included. Methodological quality generally was good or fair. Depression consistently increased the risk of death from any cause (RR, 1.59; 95% CI, 1.35-1.87), but had less certain effects on cardiovascular mortality (RR, 1.88; 95% CI, 0.84-4.19). Associations for mortality were similar regardless of the diagnostic method used for depression, but were weaker in analyses controlled for preexisting cardiovascular disease (RR, 1.36; 95% CI, 1.23-1.50). Limitations: Meta-analyses adjusting for antidepressant medication use were not possible, and data for kidney transplant recipients and individuals with earlier stages of CKD not treated with dialysis were limited. Conclusions: Depression is associated with a substantially increased risk of death in people with CKD. Effective treatment for depression in people with CKD may reduce mortality. Am J Kidney Dis. 62(3):493-505. © 2013 by the National Kidney Foundation, Inc. INDEX WORDS: Depression; dialysis; mortality; outcomes.

hronic kidney disease (CKD) is a major public health problem that affects ⬃10% of the global population and increases the risk of death 2- to 4-fold, mostly due to excess cardiovascular-related mortality.1-5 Worldwide, more than 1.5 million people have advanced CKD requiring dialysis treatment and experience a mortality rate that is substantially higher than that in the general population.6 Approximately 10-20 of every 100 people treated with dialysis die each year, and the financial burden of care directly attributable to CKD is considerable.7 Despite a high global prevalence, clinician and patient awareness of CKD

C

remains very low,1,8 and the availability of randomized trials of treatment to inform practice and improve outcomes in CKD lags behind all other internal medical specialties.9,10 In addition to worse survival, people with CKD experience a high symptom burden, including fatigue, anorexia, pain, sleep disturbance, sexual dysfunction, and role impairment, that can lead to severely impaired quality of life.11,12 Approximately 20%-30% of people with severe CKD may fulfill the criteria for a depression diagnosis, although identifying depression may be more difficult in patients with CKD due to the

From the 1Department of Medicine, University of Otago Christchurch, New Zealand; 2Department of Clinical Pharmacology and Epidemiology, Consorzio Mario Negri Sud, S. Maria Imbaro, Italy; 3School of Public Health, University of Sydney, Sydney, Australia; 4Department of Medicine, University of Alberta, Edmonton, Canada; 5Department of Nephrology, University of Queensland at Princess Alexandra Hospital, Brisbane, Australia; 6“Casa Sollievo della Sofferenza” Hospital, IRCCS, San Giovanni Rotondo (FG); 7Department of Neurology and Psychiatry, University of Bari, Bari, Italy; 8Department of Medicine, Division of Nephrology, University of Texas Southwestern Medical Center, Dallas, TX; 9Diaverum Scientific Office, Lund, Sweden;

and 10Department of Emergency and Organ Transplantation, University of Bari, Bari, Italy. Received October 10, 2012. Accepted in revised form February 26, 2013. Originally published online April 29, 2013. Address correspondence to Giovanni F.M. Strippoli, MD, MM, MPH, PhD, Department of Clinical Pharmacology and Epidemiology, Consorzio Mario Negri Sud, Via Nazionale 8/a, 66030, S Maria Imbaro, Italy. E-mail: [email protected] © 2013 by the National Kidney Foundation, Inc. 0272-6386/$36.00 http://dx.doi.org/10.1053/j.ajkd.2013.02.369

Am J Kidney Dis. 2013;62(3):493-505

493

Palmer et al

coexistence of somatic symptoms.13,14 Although links between depression and increased mortality have been demonstrated in other populations with chronic disease, including cancer15-17 and cardiovascular disease,18-20 the effects of depression on mortality in people with CKD have not been definitively quantified, although observational studies suggest an association. Randomized trials of interventions for depression in CKD populations are few and are limited by sample size, duration, and the use of surrogate outcomes to evaluate efficacy.21-26 Applying evidence for treating depression from trials conducted in the general population may be inappropriate because people with CKD might experience different benefits and harms from interventions because of increased cardiovascular comorbidity and altered drug handling. Given the likely but uncertain prognostic implications of depression in CKD and the need to consider additional intervention trials in this area, we aimed to summarize the association between depression and premature mortality (total and cardiovascular) in people with CKD. We carefully considered the effects of diagnostic methods for depression and the severity of kidney disease.

METHODS Search Strategy and Study Selection Criteria We did a systematic review and meta-analysis based on a protocol designed by the authors and according to Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) guidelines.27 We searched MEDLINE (through week 3 in December 2012) using a search designed by a specialist information manager (Table S1, available as online supplementary material) for observational studies that examined depression in people with CKD (structural kidney or urinary abnormalities with or without a glomerular filtration rate ⬍60 mL/min/1.73 m2 as defined by the National Kidney Foundation–Kidney Disease Outcomes Quality Initiative [NKF-KDOQI] criteria28) and then recorded the clinical outcomes of total or cardiovascular mortality according to depression status at baseline. Two authors independently evaluated all citations by title and abstract and retrieved the full text in detail of any that appeared relevant. We included studies in which depression status was identified by self-rating scales, clinical record searching, and semistructured or clinical interview. We included only studies in which data for the risk of mortality were evaluated according to depression status at study baseline. Some of these studies also provided longitudinal data for depression status at follow-up. Studies also had to report risk estimates (relative risks [RRs], odds ratios [ORs], or hazard ratios [HRs]) with 95% confidence intervals (CIs) or provide sufficient data for these to be calculated. Descriptions of the key methods used to identify depressive symptoms or to diagnose depression are provided in Table S2. We excluded studies of children.

Data Extraction and Classification We extracted data about study participants (stage of kidney disease [non–dialysis-dependent CKD, receiving dialysis, or transplant recipient], age, sex, duration of dialysis, and baseline preva494

lence of depression) and study design (diagnostic methods for evaluating depression status, use of baseline or time-varying measurement of depression, duration of follow-up, and covariables controlled for in analyses). We assessed methodological quality using the Agency for Healthcare Research and Quality Methods Reference Guide for Effectiveness Reviews criteria to grade study quality as low, moderate, or high risk of bias.29 We applied quality items described in the STROBE (Strengthening the Reporting of Observational Studies in Epidemiology) statement,30 including specific items concerning measurement of depression status, adjustment for potential confounding factors, and clarity of reporting of outcome assessments and statistical analyses.

Outcome Measures We defined the main outcome of interest as all-cause mortality (including death with a functioning transplant in kidney transplant recipients). The secondary outcome was cardiovascular mortality. We extracted risk estimates and 95% CIs for outcomes with participants with CKD who did not have depression as the referent category. We selected results from the full statistical model that adjusted for the largest number of potential confounders. We also listed the confounders adjusted for in the study analyses in Table S3.

Statistical Analyses We did DerSimonian-Laird random-effects meta-analysis to summarize individual study data.31 We tested for heterogeneity between studies with the Cochran Q (heterogeneity ␹2) and I2 statistics.32 We performed additional prespecified stratified analyses to explore sources of heterogeneity between studies in the risk of all-cause mortality, accounting for each of the following studylevel covariates: stage of CKD (CKD stages 1-5, stage 5D, and transplantation), baseline prevalence of depression, duration of dialysis (when applicable), measurement of depression (baseline only or time varying), diagnostic method to identify depression (physician and interview or self-rated), age, sex, extent of adjustment for confounding variables, risk of bias, type of risk estimate (HR, RR, or OR), duration, and follow-up. For summary and subgroup analyses, we required a minimum of 3 or more independent studies to justify analyses. Because 4 cohorts reporting risks of mortality may have included participants from the same sources,33-36 we conducted sensitivity analyses by including studies from the same potential data source separately in summary analyses. Because Lopes et al35 reported outcome data based on 2 measurements of depression (the Center for Epidemiologic Studies Depression Scale [CES-D] and physician diagnosis, separately), we did meta-analyses for mortality using these data in separate analyses. For all analyses, a 2-tailed P ⬍ 0.05 indicated statistical significance. We did analyses using macro routines in SAS language37 (release 9.1, 2002-2003; SAS Institute Inc) and Comprehensive Meta-analysis (version 2, 2005; Biostat).

RESULTS We retrieved 1,803 unique citations and excluded 1,773 on screening or detailed review (Fig 1). We excluded 8 potentially valid studies because they did not provide risk estimates in a format that was extractable for analysis (Table S4). Table 1 summarizes the 22 included studies comprising 83,381 people with CKD with a median follow-up of 2 years (range, 3 months to 6.5 years). The investigators reported 12,063 cases of depression (mean prevalence, 27.4%; 95% Am J Kidney Dis. 2013;62(3):493-505

Depression and Mortality in CKD

1803

Unique cita
ons iden
fied in MEDLINE (1950 to December 2012) 1378 Cita
ons excluded on screening of
tles and abstracts using general criteria 256 305 272 412 133

Editorial, commentary or review Not appropriate exposure (depression) Not appropriate outcome Not chronic kidney disease Not observa
onal study

425 Poten
ally relevant cita
ons iden
fied for full-text review 395

Cita
ons excluded based on full-text screening using inclusion criteria 29 100 238 6 10 11 1

Editorial, commentary or review Not measuring depression as exposure Not mortality as outcome Not chronic kidney disease Not observa
onal study Pediatric Duplicate publica
on

30 Studies repor
ng associa
on between depression and mortality 8 Studies excluded as data for outcomes not in extractable format 22 Studies (83,381 par
cipants) included in the meta-analyses 22 All-cause mortality (83,381 par
cipants) 3 Cardiovascular mortality (1080 par
cipants)

Figure 1. Literature search for the metaanalysis. Flow chart shows number of citations retrieved by database searching and the number and type of studies included in this review. Abbreviations: BDI, Beck Depression Inventory; CES-D, Center for Epidemiologic Studies Depression Scale; CKD, chronic kidney disease; PRIME-MD, Primary Care Evaluation of Mental Disorders.

4 14 2 2

CKD stages 1-5 CKD stage 5D Kidney transplant Mixed

CI, 20.0%-36.3%) with a median study size of 249 (range, 37-47,899) participants. Four studies (n ⫽ 1,360 participants) were of populations with CKD stages 2-5 (mean estimated glomerular filtration rate, 21.7-43 mL/min/1.73 m2), 2 studies (n ⫽ 48,739) were of kidney transplant recipients, 2 studies (n ⫽ 157) were of mixed populations (dialysis and transplantation or earlier stages of CKD), and the other 14 studies (n ⫽ 33,125) were of dialysis patients. Depression status was extracted from clinical databases according to physician diagnosis or International Classification of Diseases or Diagnostic and Statistical Manual of Mental Disorders criteria in 7 studies (n ⫽ 55,207)36,38,39,43,45,47,53 and assessed using self-rating instrument scales in 13 studies (n ⫽ 5,942).33,34,41,44,46,48-52,54,55 One study reported outcomes according to both a physician-based diagnosis and self-rating scale separately,35 and in one study, the method for detecting depression was unclear.42 Four studies (n ⫽ 464) assessed depression using the goldstandard method of semistructured interviews by clinicians.38,39,45,47 Diagnostic cutoff scores to define depression using self-administered questionnaires were Am J Kidney Dis. 2013;62(3):493-505

Diagnos
c tool

Stage of kidney disease 4 1 7 1 3 6

Clinical record CES-D BDI PRIME-MD Structured interview Other

Dura
on follow-up 5 5 9 1 2

dura
on ≤12 mo dura
on 13-24 mo dura
on 25-60 mo dura
on >60 mo dura
on unclear

heterogeneous. The definitions and methods of adjudicating mortality events are provided in Table S5. Risk of bias in the included studies for risks of all-cause mortality (22 studies) and cardiovascular mortality (3 studies) are summarized in Fig 2. Risks of bias in individual studies are provided in Fig S1 and Fig S2 (available as online supplementary material). We judged 6 (27%), 13 (59%), and 3 (14%) of the 22 studies to be low, moderate, or high risk of bias overall, respectively. Common limitations in the studies with high or moderate risk of bias were unclear definitions and methods for adjudication of outcomes, unclear reporting of blinding of outcome assessors, lack of justification for final statistical models, and unclear reporting of attrition during follow-up. In included studies, the incidence of death from any cause ranged from 40-597 per 1,000 participants. Overall, there was a strong association between depression status and risk of all-cause mortality (RR, 1.59; 95% CI, 1.35-1.87; P ⬍ 0.001; Fig 3) with moderate heterogeneity (I2⫽76.3; P ⬍ 0.001). We used subgroup analyses to explore sources of the observed heterogeneity between studies, shown in Fig 4. We 495

496 Table 1. Characteristics of Studies Included in the Meta-analysis

Study & Location

Setting

No. With CKD

Cases

Duration of Analysis for Depressive Symptoms

Stage of CKD

Follow-up

Baseline Age (y)a

Male Sex (%)

Depression Measures

Outcomes Reported in Study

Outcome Adjudication

CKD Stage 1-5 2 wk

CKD4-5b

1 y (1997-1999)

72 ⫾ 10

52

BDI score ⱖ10; pts with BDI ⱖ10 underwent assessment by DIS

Mortality

Self-report

267

56 (21%)

2 wk

CKD2-5

1 y (2005-2006)

64.5 ⫾ 12.0

99

MINI for major depression

Death; hospitalization; initiation of dialysis

Medical record & self-report

Participants from AASK

628

42%

2 wk (administered baseline and annually)

CKDc

5 y (2002-2007)

60.1

62

BDI-II ⱖ11

Mortality; ESKD or Cr doubling

Clinic center personnel; adjudicated by outcome committee

Tsai et al41 (2012; TW)

All pts in an integrated CKD program from a tertiary hospital

568

160 (37%)

2 wk

CKD3-5

Mean, 25.2 mo (2007-2010)

57

62

Chinese BDI ⱖ11

Mortality; ESKD; hospitalization

Contact with pts supplemented by medical chart review & screening data from TSN and National Mortality File

Soucie & McClellan42 (1996; US)

Pts starting dialysis treatment identified through dialysis registry

2,774 (18%)

Unclear

Dialysis

90 d (1989-1992)

56.8 ⫾ 16.2

49

Clinical record

Mortality

Staff report

Lopes et al36 (2002; US & Europe)

Randomly selected pts from nationally representative dialysis facilities in DOPPS

HD

1996-1998

59-60

5157

Physician diagnosis by medical record; selfreported

Mortality; hospitalization

Unclear

Pts from prospective cohort admitted with heart failure

Hedayati et al39 (2010; US)

Consecutive pts attending single medical center

Fischer et al40 (2011; US)

37

CKD Stage 5D 15,245

5,256

931 (17%) by 12 mo (medical medical record record); 4 wk (self)

(Continued)

Palmer et al

Am J Kidney Dis. 2013;62(3):493-505

23 (55%) by BDI; 8 (19%) by DIS

Hedayati et al38 (2004; US)

Study & Location

Setting

No. With CKD

Duration of Analysis for Depressive Symptoms

Cases

Stage of CKD

Follow-up

Baseline Age (y)a

Male Sex (%)

Depression Measures

Outcomes Reported in Study

Outcome Adjudication

Lopes et al35 Randomly selected pts (2004; from nationally multinational) representative dialysis facilities in DOPPS II

9,382

NA

12 mo (medical record); 1 wk (CES-D)

HD

NA

NA

NA

Physician diagnosis by medical record CES-D ⱖ10

Mortality; hospitalization; withdrawal from dialysis

Unclear

Hedayati et al43 (2005; US)

Prevalent males on longterm HD identified from centralized VA data sets

1,588

233 (15%)

1y

HD

2 y to September 2002

61-62b

100

ICD-9 or CPT coding

Mortality; Hospitalization

Medical records

Boulware et al44 (2006; US)

Incident dialysis pts enrolled median 45 d after dialysis initiation

917

221 (24%)

Baseline, 6, 12, and 18 mo

HD or PD

Mean, 18 mo (enrolled 1995-1998)

Range, ⬍40-ⱖ80

53

MOS Short Form-36 (MHI-5) ⱕ52

Mortality

Medical records

Drayer et al45 (2006; US)

Prevalent HD pts at urban dialysis center

62

17 (27%)

2 wk

HD

Mean, 29 mo (2002-2005)

Range, 50-60

52

PRIME-MD mood module

Mortality; withdrawal from dialysis

Medical record & Social Security Death Index

Mahajan et al46 (2007; IN)

Prevalent pts on PD at tertiary center

52

34 (64%)

2 wk

PD

2 y (2003-2005)

60.2

64

BDI ⬎11

Mortality

Unclear

Diefenthaeler et Ambulatory HD pts at al33 (2008; single center BR)

40

22 (55%)

2 wk

HD

2 y (1996-1999)

54.5 ⫾ 15.1

73

BDI ⱖ14

Mortality

Unclear

Hedayati et al47 (2008; US)

Consecutive pts attending outpatient dialysis units

98

26 (27%)

2 wk

HD

1 y (2003-2005)

57.2 ⫾ 13.8

56

SCID

Death

Medical record

Kojima et al48 (2010; JP)

Prevalent pts on HD at 3 clinics

230

99 (43%)

2 wk

HD

Mean, 4.5 y (enrollment 2001-2002)

56.0 ⫾ 9.6

56

BDI ⱖ14

Mortality

Medical record

Riezebos et al49 (2010; NL)

Prevalence dialysis pts at single center

101

42 (42%)

1 wk

Dialysis

1 y (2006-2007)

63.6 ⫾ 13.6

52

HADS ⱖ7

Mortality; hospitalization; withdrawal from dialysis

Medical record

Balogun et al50 (2011; US)b

Pts undergoing long-term HD, aged ⱖ65 y, fluent in English

77

26 (34%)

1 wk

HD

Mean, 35.2 mo; (2005-2010)

⬎65

NA

15-item GDS ⱖ5

Mortality

Unclear

Chilcot et al51 (2011; UK)

Incident dialysis pts from 3 centers

160

41 (26%)

2 wk

Dialysis

Median, 511 d (2007-2009)

57.4 ⫾ 16.0

68

BDI ⱖ16

Mortality

Clinic nephrologist

Tsai et al52 (2012; TW)

Newly-diagnosed pts with ESKD

527 (23%)

During 2001-2007

HD, PD

Mean, 36 mo (2001-2007)

61.9 ⫾13.0

50

Use of Mortality antidepressant agents

Ending date of medical coverage

2,312

497

(Continued)

Depression and Mortality in CKD

Am J Kidney Dis. 2013;62(3):493-505

Table 1 (Cont’d). Characteristics of Studies Included in the Meta-analysis

498 Table 1 (Cont’d). Characteristics of Studies Included in the Meta-analysis

Study & Location

Setting

No. With CKD

Cases

Duration of Analysis for Depressive Symptoms

Stage of CKD

Follow-up

Baseline Age (y)a

Male Sex (%)

Depression Measures

Outcomes Reported in Study

Outcome Adjudication

Kidney Tx Recipients Dobbels et al53 (2008; US)

USRDS first kidney-only recipients with Medicare as primary payer

47,899

3,360 (7%)

3 y post-Tx

Tx

3 y (1995-2003)

Range, 0-65

60

Claims data for ICD-9 code 311 “depression not elsewhere classified”

Mortality; graft failure

Unclear

Novak et al54 (2010; HU)

Prevalent stable kidney Tx recipients at single university hospital

840

187 (22%)

1 wk

Tx

Median, 58 mo

49 ⫾ 13

59

CES-D ⱖ18

Mortality; graft loss

Medical record

Zimmermann et Pts on dialysis or Tx al34 (2006; recipients at university BR) center

47

34 (72%)

2 wk

HD, PD, Tx

Mean, 77.6 mo

45.3 ⫾ 12.2

54

BDI ⱖ10

Mortality

Medical record

110

22 (20%)

2 wk

CKD5/5D

Mean, 4.4 y (2001-2007)

Mean, 65.5

54

PHQ-9 for major depression

Mortality

State death data

CKD Stage 1-5, Including CKD Stage 5D and Tx Recipients

Young et al55 (2010; US)

Prevalent pts with severe CKD seen in primary care clinics

Palmer et al

Am J Kidney Dis. 2013;62(3):493-505

Abbreviations: AASK, African-American Study of Kidney Disease and Hypertension; BDI, Beck Depression Inventory; BR, Brazil; CES-D, Center for Epidemiologic Studies Depression Scale; CKD, chronic kidney disease (with stage given thereafter, if applicable), CKD5D, dialysis-dependent CKD stage 5; CPT, Current Procedural Terminology; Cr, creatinine; DIS, Diagnostic Interview Schedule; DOPPS, Dialysis Outcomes & Practice Patterns Study; DSM, Diagnostic and Statistical Manual of Mental Disorders; eGFR, estimated glomerular filtration rate; ESKD, end-stage kidney disease; GDS, Geriatric Depression Scale; HADS, Hospital Anxiety and Depression Scale; HD, hemodialysis; HU, Hungary; ICD-9, International Classification of Diseases, Ninth Revision; IN, India; JP, Japan; KLoSHA, Korean Longitudinal Study on Health and Aging; MINI, Mini International Neuropsychiatric Interview; MHI-5, mental health subscale; MOS, Medical Outcomes Study; NA, not available; NL, the Netherlands; PD, peritoneal dialysis; pts, patients; PHQ-9, Patient Health Questionnaire 9; PRIME-MD, Primary Care Evaluation of Mental Disorders; SCID, Structured Clinical Interview for DSM-IV; TSN, Taiwan Society of Nephrology; TW, Taiwan; Tx, transplant/transplantation; UK, United Kingdom; US, United States; USRDS, US Renal Data System; VA, US Department of Veterans Affairs. a Unless shown otherwise, baseline age given as mean or mean ⫾ standard deviation. b eGFR, ⬍30 mL/min/1.73 m2. c eGFR, 20-65 mL/min/1.73 m2.

Depression and Mortality in CKD Yes

No

Unclear or not reported

Sampling Scheme Described

2

20

Clear Descrip
on of Depression Assessment

1

21

Depression Assessors Blinded to Outcome Status

8

14

Outcome Assessors Blinded to Depression Status

19

3

Time from Diagnosis of Depression to Outcome was Reported

19

Results Controlled for Poten
al Confounding Factors

19

Jus
fica
on of Final Sta
s
cal Model was Reported

3 3 6

16

Clear Defini
on of Outcome

13

9

Authors Specified a Primary Outcome of the Study

18

Prospec
ve Collec
on of Data

18

<20% Loss to Follow-up

4 4

0%

20%

6

2

14

40% 60% Propor
on of Studies

Sampling Scheme Described

3

Clear Descrip
on of Depression Assessment

3

Depression Assessors Blinded to Outcome Status

80%

100%

3

Outcome Assessors Blinded to Depression Status

1

2

Time from Diagnosis of Depression to Outcome was Reported

3

Results Controlled for Poten
al Confounding Factors

3

Jus
fica
on of Final Sta
s
cal Model was Reported

2

1

Clear Defini
on of Outcome

2

1

Authors Specified a Primary Outcome of the Study

2

1

Prospec
ve Collec
on of Data

3

<20% Loss to Follow-up

1

1 0%

20%

40%

1 60%

80%

100%

Propor
on of Studies

Figure 2. Risk of bias assessment of observational studies examining the associations between depression status and risk of all-cause mortality (20 studies) and cardiovascular mortality (3 studies). For each quality item: “Yes” means that sufficient data were reported in the study to allow assessment of quality and fulfilled the criteria for the quality item. “No” means that sufficient data were reported in the study to assess quality but did not fulfill the criteria for the quality item. “Unclear” or “not reported” means that data used to assess the quality item were not reported in the study and that the study had a potentially high risk of bias.

found that results were consistent across stages of CKD (dialysis vs earlier stages of CKD), duration of dialysis, time point for depression assessment, method used to identify depression (specifically, interview based according to diagnostic criteria vs self-administered), age and sex, risk of bias, and duration of follow-up, although analyses were limited by few studies of people with earlier stages of CKD not treated with dialysis and recipients of a kidney transplant. In addition, estimates were not modified by the risk of bias in the ascertainment of all-cause mortality across studies. The extent of adjustment in multivariable analyses influenced results; associations with depression were weaker in studies that controlled for cardiovascular disease than those that did not (P ⬍ 0.001). We had insufficient data to analyze the effect of depression treatment or severity of depression on the association between depression and mortality. We identified 3 studies with moderate risk of bias in dialysis populations that reported the association between depression and cardiovascular mortality across a median follow-up of 18 (range, 12-29) months. Am J Kidney Dis. 2013;62(3):493-505

Cardiovascular mortality incidence ranged from 69161 per 1,000 participants. Findings in individual studies were mixed, and overall, the association between depression and cardiovascular mortality was relatively imprecise and not statistically significant (RR, 1.88; 95% CI, 0.84-4.19; P ⫽ 0.1; Fig 5), with low heterogeneity (I2⫽38.5%; P ⫽ 0.2). Results generally were consistent when we repeated analyses excluding data from potentially duplicative cohorts.33-36 We also repeated analyses for mortality using data from Lopes et al35 (2004) that used medical notes to identify physician diagnosis of depression in place of analyses using the CES-D from that study and found that the association between depression and all-cause mortality was unchanged (RR, 1.60; 95% CI, 1.32-1.95; P ⬍ 0.001). When we used a more graded duration of follow-up (ⱕ12, 13-24, and ⬎24 months) in subgroup analyses, we found no differences in risk estimates between subgroups. In studies in which outcomes for mortality were not extractable for inclusion in meta-analysis, findings generally were consistent with the overall 499

Palmer et al

Study

No. of par
cipants

Risk es
mate Random effects (95% CI)

Risk es
mate Random effects (95% CI)

Weight

Diefenthaeler et al, 2008

40

4.60 (0.51 to 41.1)

Riezebos et al, 2010

101

11.7 (1.63 to 83.8)

0.52% 0.64%

Mahajan et al, 2007

52

1.28 (0.27 to 6.05)

0.98%

Drayer et al, 2006

62

4.10 (1.22 to 13.7)

1.53%

Zimmermann et al, 2006

47

2.31 (0.74 to 7.24)

1.68%

Hedaya
et al, 2004

37

1.94 (0.62 to 6.06)

1.69%

Hedaya
et al, 2010

267

1.52 (0.53 to 4.35)

1.93%

Kojima et al, 2010

230

1.70 (0.64 to 4.50)

2.18%

Tsai 2012a et al., 2012

428

2.63 (1.00 to 6.95)

2.19%

Chilcot et al, 2011

160

2.70 (1.07 to 6.84)

2.34%

Young et al, 2010

110

2.95 (1.24 to 7.5)

2.59% 4.30%

Balogun et al, 2006

77

1.91 (1.05 to 3.47)

Boulware et al, 2006

917

2.22 (1.36 to 3.61)

5.33%

Hedaya
et al, 2008

98

1.21 (0.75 to 1.95)

5.45% 6.23%

Fischer et al, 2011

628

1.05 (0.70 to 1.58)

Novak et al, 2010

840

1.66 (1.12 to 2.47)

6.37%

Hedaya
et al, 2005

1588

0.98 (0.72 to 1.34)

7.46%

15,245

1.30 (1.00 to 1.70)

8.06%

2312

1.38 (1.16 to 1.65)

9.17%

Soucie et al, 1996 Tsai 2012 et al., 2012 Lopes et al, 2002 Dobbels et al, 2008 Lopes et al, 2004 Overall

5256

1.23 (1.08 to 1.40)

9.64%

47,899

2.24 (2.00 to 2.50)

9.80%

6987

1.42 (1.29 to 1.57)

9.91%

83,381

1.59 (1.35 to 1.87)

100.0%

Heterogeneity; Q sta
s
c=88.9, I2=76.3%, p value <0.001

0.5

Decreased mortality with depression

1

2

5

10

Increased mortality with depression

Figure 3. Risk ratios of total mortality for depressed people with chronic kidney disease compared with nondepressed people with chronic kidney disease. Summary estimates were obtained using a random-effects model. Data markers indicate the adjusted risk ratios (odds ratios, risk ratios, or hazard ratios) in depressed participants with chronic kidney disease compared with nondepressed people with chronic kidney disease. The size of the data markers indicates the weight of the study, which is the inverse variance of the effect estimate. The diamond marker indicates the summary risk estimate and its 95% confidence interval (CI).

conclusions in this review; increasing severity of depressive symptoms assessed by self-administered questionnaires was associated with lower survival (Table S4).We combined different types of risk estimates in the meta-analyses for all-cause mortality (HR, RR, and OR); in subgroup analyses, type of estimate did not statistically influence the observed relationship between depression and clinical outcome (P ⬎ 0.05 for subgroup difference).

DISCUSSION Depression is associated with an increased risk of premature death in people who have CKD. The effects of depression on cardiovascular mortality in CKD populations are less certain because few studies provide data. The association between depression and death is similar across age groups, in both men and women, and irrespective of the diagnostic tool used 500

for depression. It is important to note that diagnostic tools for depression vary widely, from chart diagnosis or self-administered questionnaires identifying depressive symptoms and the gold-standard approach using a semistructured interview by a trained clinician to identify depression according to diagnostic criteria. While we found no statistical difference between risk estimates based on the method used to identify depression symptoms (interview-based vs self-administered scale or medical record), the risk was numerically lower when depression was identified by interview, suggesting that it is possible there was less confounding due to somatic symptoms misclassified as depression in these analyses. Although the risk is lower when adjusted for cardiovascular disease, it remains statistically significant and clinically important, with a magnitude of ⬃30% increased risk of death in people with CKD and depression than for those with Am J Kidney Dis. 2013;62(3):493-505

Depression and Mortality in CKD

No. of studies

No. of par
cipants

Risk es
mate Random effects (95% CI)

CKD stages 1-5

4

1360

CKD stage 5D

14

Transplant

2

Study

Risk es
mate Random effects (95% CI)

Q sta
s
c

p value for heterogeneity

I2 value

1.34 (0.92-1.97)

3.6

0.30

17.4

33,125

1.40 (1.23-1.45)

24.1

0.03

46.1

48,739

Insufficient sample

-

-

-

p value between strata

Stage of chronic kidney disease

0.8

Baseline prevalence of depression 30% or above

8

8162

1.59 (1.13-2.25)

9.42

0.22

25.7

<30%

14

75,219

1.62 (1.32-2.00)

75.0

<0.001

02.7

0.9

Dura
on of dialysis Incident

4

18,634

1.52 (1.20-1.92)

5.58

0.13

46.2

Prevalent

11

14,919

1.38 (1.16-1.64)

19.6

0.03

49.1

Time-varying

3

49,444

1.86 (1.45-2.40)

12.4

0.002

83.9

Baseline

19

33,937

1.45 (1.27-1.67)

27.2

0.08

33.8

Interview or physician

7

55,207

1.51 (1.08-2.10)

63.2

<0.001

90.5

Self-ra
ng scale

13

5942

1.87 (1.40-2.48)

34.9

0.10

18.4

0.5

Measurement of depression 0.1

Method of measuring depression 0.3

Age (study mean) 60 years or older

9

5172

1.37 (1.09-1.73)

15.0

0.06

46.6

<60 years

10

22,406

1.51 (1.22-1.86)

12.6

0.18

28.5

0.6

Gender Propor
on men ≥60%

9

51,099

1.56 (1.21-2.02)

56.2

<0.001

85.8

Propor
on men<60%

12

25,218

1.60 (1.28-2.01)

19.4

0.05

43.3

0.9

Controlling for cardiovascular disease in mul
variate model Yes

13

34,326

1.36 (1.23-1.50)

24.2

0.02

50.3

No

9

49.055

2.21 (1.84-2.65)

2.26

0.97

<0.001

<0.001

Risk of bias Low

6

3495

1.44 (1.03-2.00)

12.6

0.03

60.3

Moderate

13

79,717

1.64 (1.34-1.99)

70.7

<0.001

83.0

High

3

169

1.93 (0.97-3.86)

0.87

0.65

0.0

0.6

Dura
on of follow-up 24 months or fewer

10

18,505

1.52 (1.14-2.04)

16.8

0.05

46.4

>24 months

10

52,633

1.80 (1.40-2.32)

32.8

<0.001

72.5

0.5

1

Decreased mortality with depression

2

0.4

5

Increased mortality with depression

Figure 4. Subgroup analyses for the effects of depression on all-cause mortality in people with chronic kidney disease (CKD). Abbreviation: CI, confidence interval.

CKD alone. Given that estimates suggest that ⬃10% of the global population has CKD and risks of mortality are 2-3 times that of the general population, the excess physical and financial burden of death due to depression in the presence of CKD may be considerable. Our finding that depression in the presence of CKD increases mortality after adjustment for sociodemographic and clinical variables is congruous with the excess mortality attributable to depression quantified in other populations with cancer,15,16 diabetes,56 myocardial infarction,20,57 and heart failure19 and in the general community.17,58 However, our study suggests that the excess mortality risk attributable to depression in those with CKD may be higher than that in these other populations, in which the risk typically is 20%-25%.15,16 Importantly, these earlier meta-analyses in settings of other chronic diseases, as well as our own study, have shown that some of the association between Am J Kidney Dis. 2013;62(3):493-505

depression and death can be attributed to confounding and bias, with adjusted estimates significantly smaller than unadjusted estimates of the association. In people with CKD specifically, depression is more likely when medical comorbidity or other clinical characteristics that are known to correlate with poor survival are present (eg, malnutrition,59,60 cardiovascular disease,36 inflammation,36 and elevated serum phosphorus level39) and which may not be completely accounted for by adjustment in analytical models. In addition, somatic symptoms that frequently accompany CKD, including tiredness, anorexia, sleep disturbance, and altered concentration, could lead to misclassification of depression when depression rating scales are used, resulting in additional bias in analyses. Accordingly, we explored the effects of methodological quality, depression classification, and multivariable adjustment in the original studies using subgroup analyses and found that the association between depression and death was robust to diagnostic method 501

Palmer et al

Study

No. of par
cipants

Risk es
mate Random effects (95% CI)

Risk es
mate Random effects (95% CI)

Weight

Riezebos et al, 2010

101

1.05 (0.25 to 4.46)

Drayer et al, 2006

62

1.13 (0.33 to 3.89)

28.1%

Boulware et al, 2006

917

3.27 (1.57 to 6.81)

49.3%

Overall

1080

1.88 (0.84 to 4.19)

100.0%

Heterogeneity; Q sta
s
c=3.25, I2=38.5%, p value =0.2

22.6%

0.5

Decreased cardiovascular mortality with depression

1

2

5

10

Increased cardiovascular mortality with depression

Figure 5. Risk ratios of cardiovascular mortality for depressed people with chronic kidney disease compared with nondepressed people with chronic kidney disease. Summary estimates were obtained using a random-effects model. Data markers indicate the adjusted risk ratios (odds ratios, risk ratios, or hazard ratios) in depressed participants with chronic kidney disease compared with nondepressed people with chronic kidney disease. The size of the data markers indicates the weight of the study, which is the inverse variance of the effect estimate. The diamond marker indicates the summary risk estimate and its 95% confidence interval (CI).

for depression ascertainment and risk of bias in contributing cohorts. While associations with mortality were weaker in individual studies that had controlled for cardiovascular disease, the relationship between mortality and depression in adjusted studies remained robust and statistically significant (P ⬍ 0.001). Depression may have a direct causal role in poor health outcomes. Previous studies have shown that depression is associated with impaired nutrition in dialysis patients and that antidepressant therapy in addition to supportive psychotherapy can improve nutritional status in depressed individuals.25,61,62 Depression is associated with elevated levels of C-reactive protein and inflammatory cytokines63 that may contribute in part to increased risk of subsequent cardiovascular events.64,65 Depression also is associated with lower levels of social support and increased social conflict,66 although improving depression and social support with cognitive behavioral and group therapy to date has not been shown to improve survival.67 Depressed individuals on dialysis therapy are significantly more likely to withdraw from treatment, which may increase the risk of premature death,68,69 and in earlier stages of CKD, depression is associated with a more rapid decline in kidney function.41 Nonadherence to medical treatment advice and recommendations is approximately 3-fold higher in depressed people70 and is linked to poorer outcomes, arguing for an additional possible mechanism between depression and the risk of increased health care utilization and death. Notably, adherence to antidepressant medication may be increased in settings in which participants are seen by mental health specialists,71 suggesting a broad approach to depression management, and research in CKD may be needed to improve both depression symptoms and related clinical outcomes in 502

this population. Although death by suicide is nearly doubled in people with kidney disease on dialysis therapy compared with community-dwelling adults,72 suicide is a rare event in this population and is unlikely to be a large contributor to excess mortality associated with depression in people with CKD. The association between depression and death in people with CKD suggests that treatment of depression has the potential to improve outcomes. However, current evidence for effective and safe strategies in the nephrology setting is limited to 6 small randomized controlled trials comprising 14-89 participants that have high medication discontinuation rates, often lack valid assessment of depression diagnosis, focus on surrogate outcomes, and are insufficiently powered to detect treatment-related hazards.21-26 Antidepressants improve depression symptoms in people with chronic diseases in general, but evidence for safety and tolerability and beneficial effects on quality of life and physical health outcomes in these populations are sparse.73 In the absence of high-quality evidence, broad adoption of interventions for depression known to be efficacious in the general population may not be appropriate in people with CKD because of increased cardiovascular comorbid conditions, heightened risks of drug interactions related to polypharmacy, and altered drug handling that might lead to unintended treatment-related harms.74 Nonpharmacologic interventions including exercise therapy additionally may be problematic due to lower effort tolerances and the high prevalence of cardiovascular disease in the CKD setting. Our finding that depression is associated with premature death suggests that additional randomized trials of treatment for depression in people with a broad range of CKD that evaluate end points includAm J Kidney Dis. 2013;62(3):493-505

Depression and Mortality in CKD

ing health care utilization and mortality, as well as treatment-related harms, should be high priority. The strengths of this review include the rigorous and standardized methodology defined a priori, the relevance of the findings to research prioritization, and the magnitude and robustness of the risk estimate for mortality observed. The limitations are largely due to the smaller quantity of information on which some findings are based, as well as the marked and partly unexplained differences in estimates provided by contributing studies. Data for the clinical effects of depression are sparse, particularly for patients with earlier stages of CKD and recipients of kidney transplants, for whom additional data would be informative. Based on available data, it is not possible to determine the influence of depression symptoms in clinical outcomes before and after the initiation of dialysis therapy. The variation in risk estimates between studies may relate to the methods used to identify depression. The investigators reported a wide range of tools to measure depression, including simple questions (reporting feeling sad or depressed), tools intended to screen for depression to identify people who may benefit from a more detailed assessment for depression, and formal diagnostic criteria identified using clinician interview. Although we did not observe statistical differences between ways in which depression was ascertained (physician interview or self-report scales) in subgroup analyses, studies also varied widely in self-report cutoff values used to classify a person as depressed and the duration of symptoms captured by diagnostic approaches, that both may have been sources of between-study heterogeneity. In addition, we could not analyze the relationship between persistence of depression symptoms and mortality outcomes. Existing meta-analyses in cardiovascular populations suggest that risks of death associated with depression are highest in the first 2 years after a diagnosis of depression and are attenuated over time.18,57 We assessed the duration of follow-up as an effect modifier and noted that risks of mortality associated with depression were persistent beyond 2 years of followup, although the statistical power to assess reliably the effects of follow-up length were small. We could not quantify the risk of suicidal mortality associated with CKD and concomitant depression due to insufficient data. We also were not able to separate out the study-level effects of antidepressant treatment on mortality because disaggregated data for treated patients were not provided consistently. Thus, depression in the presence of CKD is associated with substantially increased mortality. Given the relevance of depression on survival in excess of the already elevated mortality risks in CKD, together with the lack of specific evidence for safe and efficacious Am J Kidney Dis. 2013;62(3):493-505

treatment for depression with concomitant CKD, more and better studies of health care interventions for depression in CKD populations are urgently required.

ACKNOWLEDGEMENTS We acknowledge the specialist support from the Cochrane Renal Group. Ruth Mitchell, specialist information manager, designed and ran the electronic database searches. Narelle Willis, coordinating editor, provided administrative support. We also thank Michela Sciancalepore for administrative support. Support: Dr Palmer is supported by a Fellowship from the Consorzio Mario Negri Sud from an unrestricted grant from Amgen Dompé and is a 2012 L’Oreal UNESCO (United Nations Educational, Scientific and Cultural Organization) for Women in Science Australia and New Zealand Fellow. Dr Johnson is a current recipient of a Queensland Government Health Research Fellowship. This project received no external funding. Financial Disclosure: The authors declare that they have no other relevant financial interests.

SUPPLEMENTARY MATERIAL Table S1: Search strategy for Ovid MEDLINE. Table S2: Methods used to identify depressive symptoms or diagnose depression. Table S3: Variables controlled for and methods for adjustment in included studies. Table S4: Details of excluded studies. Table S5: Definitions of mortality reported in association with depression status at baseline. Figure S1: Bias assessment in individual studies for outcome of all-cause mortality. Figure S2: Bias assessment in individual studies for outcome of cardiovascular mortality. Note: The supplementary material accompanying this article (http://dx.doi.org/10.1053/j.ajkd.2013.02.369) is available at www.ajkd.org.

REFERENCES 1. Coresh J, Byrd-Holt D, Astor BC, et al. Chronic kidney disease awareness, prevalence, and trends among U.S. adults, 1999 to 2000. J Am Soc Nephrol. 2005;16(1):180-188. 2. Chen J, Wildman RP, Gu D, et al. Prevalence of decreased kidney function in Chinese adults aged 35 to 74 years. Kidney Int. 2005;68(6):2837-2845. 3. Chadban SJ, Briganti EM, Kerr PG, et al. Prevalence of kidney damage in Australian adults: the AusDiab Kidney Study. J Am Soc Nephrol. 2003;14(7)(suppl 2):S131-S138. 4. Singh NP, Ingle GK, Saini VK, et al. Prevalence of low glomerular filtration rate, proteinuria and associated risk factors in North India using Cockcroft-Gault and Modification of Diet in Renal Disease equations: an observational, cross-sectional study. BMC Nephrol. 2009;17(10):4. 5. Tonelli M, Wiebe N, Culleton B, et al. Chronic kidney disease and mortality risk: a systematic review. J Am Soc Nephrol. 2006;17(7):2034-2047. 6. Moeller S, Gioberge S, Brown G. ESRD patients in 2001: global overview of patients, treatment modalities and development trends. Nephrol Dial Transplant. 2002;17(12):2071-2076. 7. Smith DH, Gullion CM, Nichols G, Keith DS, Brown JB. Cost of medical care for chronic kidney disease and comorbidity among enrollees in a large HMO population. J Am Soc Nephrol. 2004;15(5):1300-1306. 503

Palmer et al 8. Plantinga LC, Boulware LE, Coresh J, et al. Patient awareness of chronic kidney disease: trends and predictors. Arch Intern Med. 2008;168(20):2268-2275. 9. Strippoli GF, Craig JC, Schena FP. The number, quality, and coverage of randomized controlled trials in nephrology. J Am Soc Nephrol. 2004;15(2):411-419. 10. Palmer SC, Sciancalepore M, Strippoli GF. Trial quality in nephrology: how are we measuring up? Am J Kidney Dis. 2011; 58(3):335-337. 11. Murtagh FE, Addington-Hall JM, Edmonds PM, et al. Symptoms in advanced renal disease: a cross-sectional survey of symptom prevalence in stage 5 chronic kidney disease managed without dialysis. J Palliat Med. 2007;10(6):1266-1276. 12. Evans RW, Manninen DL, Garrison LP Jr, et al. The quality of life of patients with end-stage renal disease. N Engl J Med. 1985;312(9):553-559. 13. Chan R, Steel Z, Brooks R, et al. Psychosocial risk and protective factors for depression in the dialysis population: a systematic review and meta-regression analysis. J Psychosom Res. 2011;71(5):300-310. 14. Murtagh FE, Addington-Hall J, Higginson IJ. The prevalence of symptoms in end-stage renal disease: a systematic review. Adv Chronic Kidney Dis. 2007;14(1):82-99. 15. Pinquart M, Duberstein PR. Depression and cancer mortality: a meta-analysis. Psychol Med. 2010;40(11):1797-1810. 16. Satin JR, Linden W, Phillips MJ. Depression as a predictor of disease progression and mortality in cancer patients: a metaanalysis. Cancer. 2009;115(22):5349-5361. 17. Wulsin LR, Vaillant GE, Wells VE. A systematic review of the mortality of depression. Psychosom Med. 1999;61(1):6-17. 18. Meijer A, Conradi HJ, Bos EH, Thombs BD, van Melle JP, de Jonge P. Prognostic association of depression following myocardial infarction with mortality and cardiovascular events: a metaanalysis of 25 years of research. Gen Hosp Psychiatry. 2011;33(3): 203-216. 19. Rutledge T, Reis VA, Linke SE, Greenberg BH, Mills PJ. Depression in heart failure a meta-analytic review of prevalence, intervention effects, and associations with clinical outcomes. J Am Coll Cardiol. 2006;48(8):1527-1537. 20. van Melle JP, de Jonge P, Spijkerman TA, et al. Prognostic association of depression following myocardial infarction with mortality and cardiovascular events: a meta-analysis. Psychosom Med. 2004;66(6):814-822. 21. Ancarani E, Biondi B, Bolletta A, et al. Major depression complicating hemodialysis in patients with chronic renal failure: a multicenter, double-blind, controlled clinical trial of S-adenosyl-Lmethionine versus placebo. Curr Ther Res Clin Exp. 1993;54(6): 680-686. 22. Baines LS, Joseph JT, Jindal RM. Prospective randomized study of individual and group psychotherapy versus controls in recipients of renal transplants. Kidney Int. 2004;65:1937-1942. 23. Blumenfield M, Levy NB, Spinowitz B, et al. Fluoxetine in depressed patients on dialysis. Int J Psychiatry Med. 1997;27(1): 71-80. 24. Duarte PS, Miyazaki MC, Blay SL, Sesso R. Cognitivebehavioral group therapy is an effective treatment for major depression in hemodialysis patients. Kidney Int. 2009;76:414-421. 25. Koo JR, Yoon JY, Joo MH, et al. Treatment of depression and effect of antidepression treatment on nutritional status in chronic hemodialysis patients. Am J Med Sci. 2005;329(1):1-5. 26. Lii YC, Tsay SL, Wang TJ. Group intervention to improve quality of life in haemodialysis patients. J Clin Nurs. 2007;16(11c): 268-275. 27. Moher D, Liberati A, Tetzlaff J, Altman DG. Preferred Reporting Items for Systematic Reviews and Meta-analyses: the PRISMA statement. Ann Intern Med. 2009;151(4):264-269, W264. 504

28. National Kidney Foundation. K/DOQI Clinical Practice Guidelines for Chronic Kidney Disease: evaluation, classification, and stratification. Am J Kidney Dis. 2002;39(2)(suppl 1):S1-S266. 29. Agency for Healthcare Research and Quality. Methods Reference Guide for Effectiveness and Comparative Effectiveness Reviews, version 1.0. 2007. Rockville, MD: US Department of Health and Human Services, Agency for Healthcare Research and Quality; 2007. http://effectivehealthcare.arhq.gov/repFiles/ 2007)10DraftMethodsGuide.pdf. Accessed March 26, 2012. 30. von Elm E, Altman DG, Egger M, Pocock SJ, Gotzsche PC, Vandenbroucke JP. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies. Lancet. 2007;370(9596):14531457. 31. DerSimonian R, Laird N. Meta-analysis in clinical trials. Control Clin Trials. 1986;7(3):177-188. 32. Higgins JP, Thompson SG, Deeks JJ, Altman DG. Measuring inconsistency in meta-analyses. BMJ. 2003;327(7414):557560. 33. Diefenthaeler EC, Wagner MB, Poli-de-Figueiredo CE, Zimmermann PR, Saitovitch D. Is depression a risk factor for mortality in chronic hemodialysis patients? Rev Bras Psiquiatr. 2008;30(2):99-103. 34. Zimmermann PR, Camey SA, Mari JJ. A cohort study to assess the impact of depression on patients with kidney disease. Int J Psychiatry Med. 2006;36(4):457-468. 35. Lopes AA, Albert JM, Young EW, et al. Screening for depression in hemodialysis patients: associations with diagnosis, treatment, and outcomes in the DOPPS. Kidney Int. 2004;66(5): 2047-2053. 36. Lopes AA, Bragg J, Young E, et al. Depression as a predictor of mortality and hospitalization among hemodialysis patients in the United States and Europe. Kidney Int. 2002;62(1): 199-207. 37. van Houwelingen HC, Arends LR, Stijnen T. Advanced methods in meta-analysis: multivariate approach and metaregression. Stat Med. 2002;21(4):589-624. 38. Hedayati SS, Jiang W, O’Connor CM, et al. The association between depression and chronic kidney disease and mortality among patients hospitalized with congestive heart failure. Am J Kidney Dis. 2004;44(2):207-215. 39. Hedayati SS, Minhajuddin AT, Afshar M, Toto RD, Trivedi MH, Rush AJ. Association between major depressive episodes in patients with chronic kidney disease and initiation of dialysis, hospitalization, or death. JAMA. 2010;303(19):1946-1953. 40. Fischer MJ, Kimmel PL, Greene T, et al. Elevated depressive affect is associated with adverse cardiovascular outcomes among African Americans with chronic kidney disease. Kidney Int. 2011;80:670-678. 41. Tsai Y-C, Chiu Y-W, Hung C-C, et al. Association of symptoms of depression with progression of CKD. Am J Kidney Dis. 2012;60(1):54-61. 42. Soucie JM, McClellan WM. Early death in dialysis patients: risk factors and impact on incidence and mortality rates. J Am Soc Nephrol. 1996;7(10):2169-2175. 43. Hedayati SS, Grambow SC, Szczech LA, Stechuchak KM, Allen AS, Bosworth HB. Physician-diagnosed depression as a correlate of hospitalizations in patients receiving long-term hemodialysis. Am J Kidney Dis. 2005;46(4):642-649. 44. Boulware LE, Liu Y, Fink NE, et al. Temporal relation among depression symptoms, cardiovascular disease events, and mortality in end-stage renal disease: contribution of reverse causality. Clin J Am Soc Nephrol. 2006;1(3):496-504. 45. Drayer RA, Piraino B, Reynolds CF III, et al. Characteristics of depression in hemodialysis patients: symptoms, quality of life and mortality risk. Gen Hosp Psychiatry. 2006;28(4):306-312. Am J Kidney Dis. 2013;62(3):493-505

Depression and Mortality in CKD 46. Mahajan S, Tiwari SC, Kalra V, et al. Analysis of depression and its effect on outcome among adult Indian peritoneal dialysis patients. Perit Dial Int. 2007;27(1):94-96. 47. Hedayati SS, Bosworth HB, Briley LP, et al. Death or hospitalization of patients on chronic hemodialysis is associated with a physician-based diagnosis of depression. Kidney Int. 2008; 74(7):930-936. 48. Kojima M, Hayano J, Suzuki S, et al. Depression, alexithymia and long-term mortality in chronic hemodialysis patients. Psychother Psychosom. 2010;79:303-311. 49. Riezebos RK, Nauta KJ, Honig A, Dekker FW, Siegert CE. The association of depressive symptoms with survival in a Dutch cohort of patients with end-stage renal disease. Nephrol Dial Transplant. 2010;25(1):231-236. 50. Balogun RA, Balogun SA, Kepple AL, et al. The 15-item Geriatric Depression Scale as a predictor of mortality in older adults undergoing hemodialysis. J Am Geriatr Soc. 2011;59(8): 1563-1565. 51. Chilcot J, Davenport A, Wellsted D, Firth J, Farrington K. An association between depressive symptoms and survival in incident dialysis patients. Nephrol Dial Transplant. 2011;26:1628-1634. 52. Tsai C-J, Loh E-W, Lin C-H, Yu T-M, Chan C-H, Lan T-H. Correlation of antidepressive agents and the mortality of end-stage renal disease. Nephrology. 2012;17(4):390-397. 53. Dobbels F, Skeans MA, Snyder JJ, Tuomari AV, Maclean JR, Kasiske BL. Depressive disorder in renal transplantation: an analysis of Medicare claims. Am J Kidney Dis. 2008;51(5):819828. 54. Novak M, Zsolt MM, Szeifert L, et al. Depressive symptoms and mortality in patients after kidney transplantation: a prospective prevalent cohort study. Psychosom Med. 2010;72(6): 527-534. 55. Young BA, Von KM, Heckbert SR, et al. Association of major depression and mortality in stage 5 diabetic chronic kidney disease. Gen Hosp Psychiatry. 2010;32(2):119-124. 56. Pan A, Lucas M, Sun Q, et al. Increased mortality risk in women with depression and diabetes mellitus. Arch Gen Psychiatry. 2011;68(1):42-50. 57. Barth J, Schumacher M, Herrmann-Lingen C. Depression as a risk factor for mortality in patients with coronary heart disease: a meta-analysis. Psychosom Med. 2004;66(6):802-813. 58. Cuijpers P, Smit F. Excess mortality in depression: a metaanalysis of community studies. J Affect Disord. 2002;72(3):227236. 59. Kalender B, Ozdemir AC, Koroglu G. Association of depression with markers of nutrition and inflammation in chronic kidney disease and end-stage renal disease. Nephron. 2006;102(3-4):c115121. 60. Koo JR, Yoon JW, Kim SG, et al. Association of depression with malnutrition in chronic hemodialysis patients. Am J Kidney Dis. 2003;41(5):1037-1042. 61. Friend R, Hatchett L, Wadhwa NK, Suh H. Serum albumin and depression in end-stage renal disease. Adv Perit Dial. 1997;13: 155-157.

Am J Kidney Dis. 2013;62(3):493-505

62. Koo JR, Yoon JW, Kim SG, et al. Association of depression with malnutrition in chronic hemodialysis patients. Am J Kidney Dis. 2003;41(5):1037-1042. 63. Hung KC, Wu CC, Chen HS, et al. Serum IL-6, albumin and co-morbidities are closely correlated with symptoms of depression in patients on maintenance haemodialysis. Nephrol Dial Transplant. 2011;26(2):658-664. 64. Vaccarino V, Johnson BD, Sheps DS, et al. Depression, inflammation, and incident cardiovascular disease in women with suspected coronary ischemia: the National Heart, Lung, and Blood Institute-sponsored WISE Study. J Am Coll Cardiol. 2007;50(21): 2044-2050. 65. Kaptoge S, Di Angelantonio E, Lowe G, et al. C-Reactive protein concentration and risk of coronary heart disease, stroke, and mortality: an individual participant meta-analysis. Lancet. 2010;375(9709):132-140. 66. Barefoot JC, Burg MM, Carney RM, et al. Aspects of social support associated with depression at hospitalization and follow-up assessment among cardiac patients. J Cardiopulm Rehabil. 2003;23(6):404-412. 67. Berkman LF, Blumenthal J, Burg M, et al. Effects of treating depression and low perceived social support on clinical events after myocardial infarction: the Enhancing Recovery in Coronary Heart Disease Patients (ENRICHD) Randomized Trial. JAMA. 2003;289(23):3106-3116. 68. Lacson E Jr, Li NC, Guerra-Dean S, Lazarus M, Hakim R, Finkelstein FO. Depressive symptoms associate with high mortality risk and dialysis withdrawal in incident hemodialysis patients. Nephrol Dial Transplant. 2012;27(7):2921-2928. 69. McDade-Montez EA, Christensen AJ, Cvengros JA, Lawton WJ. The role of depression symptoms in dialysis withdrawal. Health Psychol. 2006;25(2):198-204. 70. DiMatteo MR, Lepper HS, Croghan TW. Depression is a risk factor for noncompliance with medical treatment: metaanalysis of the effects of anxiety and depression on patient adherence. Arch Intern Med. 2000;160(14):2101-2107. 71. Robinson RL, Long SR, Chang S, et al. Higher costs and therapeutic factors associated with adherence to NCQA HEDIS antidepressant medication management measures: analysis of administrative claims. J Manag Care Pharm. 2006;12(1):43-54. 72. Kurella M, Kimmel PL, Young BS, Chertow GM. Suicide in the United States end-stage renal disease program. J Am Soc Nephrol. 2005;16(3):774-781. 73. Taylor D, Meader N, Bird V, Pilling S, Creed F, Goldberg D. Pharmacological interventions for people with depression and chronic physical health problems: systematic review and metaanalyses of safety and efficacy. Br J Psychiatry. 2011;198(3):179188. 74. Hedayati SS, Yalamanchili V, Finkelstein FO. A practical approach to the treatment of depression in patients with chronic kidney disease and end-stage renal disease. Kidney Int. 2012;81(3): 247-255.

505