Antidepressants for treating depression in adults with end‐stage kidney disease treated with dialysis

Abstract

Background

Depression affects approximately one‐quarter of people treated with dialysis and is considered an important research uncertainty by patients and health professionals. Treatment for depression in dialysis patients may have different benefits and harms compared to the general population due to different clearances of antidepressant medication and the severity of somatic symptoms associated with end‐stage kidney disease (ESKD). Guidelines suggest treatment of depression in dialysis patients with pharmacological therapy, preferably a selective serotonin reuptake inhibitor. This is an update of a review first published in 2005.

Objectives

To evaluate the benefit and harms of antidepressants for treating depression in adults with ESKD treated with dialysis.

Search methods

We searched Cochrane Kidney and Transplant's Specialised Register to 20 January 2016 through contact with the Information Specialist using search terms relevant to this review.

Selection criteria

Randomised controlled trials (RCTs) comparing antidepressant treatment with placebo or no treatment, or compared to another antidepressant medication or psychological intervention in adults with ESKD (estimated glomerular filtration rate < 15 mL/min/1.73 m²).

Data collection and analysis

Data were abstracted by two authors independently onto a standard form and subsequently entered into Review Manager. Risk ratios (RR) for dichotomous data and mean differences (MD) for continuous data were calculated with 95% confidence intervals (95% CI).

Main results

Four studies in 170 participants compared antidepressant therapy (fluoxetine, sertraline, citalopram or escitalopram) versus placebo or psychological training for 8 to 12 weeks. In generally very low or ungradeable evidence, compared to placebo, antidepressant therapy had no evidence of benefit on quality of life, had uncertain effects on increasing the risk of hypotension (3 studies, 144 participants: RR 1.72, 95% CI 0.75 to 3.92), headache (2 studies 56 participants: RR 2.91, 95% CI 0.73 to 11.57), and sexual dysfunction (2 studies, 101 participants: RR 3.83, 95% CI 0.63 to 23.34), and increased nausea (3 studies, 114 participants: RR 2.67, 95% CI 1.26 to 5.68). There were few or no data for hospitalisation, suicide or all‐cause mortality resulting in inconclusive evidence. Antidepressant therapy may reduce depression scores during treatment compared to placebo (1 study, 43 participants: MD ‐7.50, 95% CI ‐11.94 to ‐3.06). Antidepressant therapy was not statistically different from group psychological therapy for effects on depression scores or withdrawal from treatment and a range of other outcomes were not measured.

Authors' conclusions

Despite the high prevalence of depression in dialysis patients and the relative priority that patients place on effective treatments, evidence for antidepressant medication in the dialysis setting is sparse and data are generally inconclusive. The relative benefits and harms of antidepressant therapy in dialysis patients are poorly known and large randomised studies of antidepressants versus placebo are required.

Plain language summary

Antidepressants for treating depression in adults with end‐stage kidney disease treated with dialysis

Background

People treated with dialysis frequently experience depression and anxiety. Depression in this situation is linked to poor quality of life and increased complications, such as needing to be admitted to hospital, or stopping dialysis treatment. Patients, their families, and health care workers agree that caring for depression symptoms appropriately and finding effective treatments is really important. Antidepressant drugs may not be removed from the body as quickly for people with kidney disease and so may cause more side effects. Despite depression being very common and treatment having potentially different side‐effects compared with people without kidney disease, a previous version of this review in 2005 found only a single research study. It is unknown whether antidepressant treatment works and is safe for people with kidney failure.

A summary of whether antidepressant therapy works and is safe in people with kidney failure would be relevant to patients and their families, health care workers, and policy makers to generate patient‐centred treatment policies.

This review looks at whether we know whether drug treatment works to improve symptoms of depression in adults treated with dialysis without causing common and severe side effects.

Study characteristics

We included all studies which have looked at drug treatment against placebo (sugar pill) or other kinds of mental health support. People included in the studies had an equal chance of receiving either treatment.

Key results

Unfortunately, even though depression is very common and finding good treatments for depression are highly valued by patients on dialysis, there are only a few small studies to tell us about whether drug treatments are both safe and reduce symptoms. Based on this information, we still don't know whether depression treatment works well for people treated with dialysis and is safe (doesn't cause excess and serious side effects).

Quality of the evidence

The question of whether drugs can reduce symptoms of depression and improve quality of life for people on dialysis is still important. We need a big study that involves dialysis patients and assesses a commonly‐used antidepressant drug with a placebo and measures the treatment effects based on what patients and their families value most.

Summary of findings

Summary of findings for the main comparison. Antidepressant therapy versus placebo for depression in dialysis patients.

Antidepressant therapy versus placebo for depression in dialysis patients
Patient or population: dialysis patients with depression Settings: dialysis patients Intervention: selective serotonin reuptake inhibitor (SSRI) antidepressant (8 to 12 weeks) Comparison: placebo
Outcomes	Illustrative comparative risks* (95% CI)		Relative effect (95% CI)	No of participants (studies)	Quality of the evidence (GRADE)	Comments
	Assumed risk per 1000 patients treated for 8 to 12 weeks	Corresponding risk per 1000 patients treated for 8 to 12 weeks
	Placebo	SSRI
Dizziness or hypotension	135	97 more (34 fewer to 394 more)	1.72 (0.75 to 3.92)	114 (3)	⊕1,2,3 very low	Downgraded as follows: 1Study limitations (studies generally at unclear or high risk of bias for many domains) 2Severe imprecision (2 grades): risk estimate includes null effect and estimate consistent with both appreciable benefit and harm 3Directness: data derived from small number of studies in specific settings which may not be generalisable
Nausea	118	197 more (30 more to 473 more)	2.67 (1.26 to 5.68)	114 (3)	⊕⊕⊕1,2 moderate	Downgraded as follows: 1Study limitations (studies generally at unclear or high risk of bias for many domains) 2Directness: data derived from small number of studies in specific settings which may not be generalisable
Sexual dysfunction	19	54 more (7 fewer to 423 more)	3.83 (0.63 to 23.34)	101 (2)	⊕1,2,3 very low	Downgraded as follows: 1Study limitations (studies generally at unclear or high risk of bias for many domains) 2Severe imprecision (2 grades): risk estimate includes null effect and estimate consistent with both appreciable benefit and harm 3Directness: data derived from small number of studies in specific settings which may not be generalisable
All‐cause mortality	20	No difference (18 fewer to 280 more)	1.00 (0.07 to 15.12)	50 (1)	⊕1,2,3 very low	Downgraded as follows: 1Study limitations (studies generally at unclear or high risk of bias for many domains) 2Severe imprecision (2 grades): risk estimate includes null effect and estimate consistent with both appreciable benefit and harm 3Directness: data derived from small number of studies in specific settings which may not be generalisable
*The basis for the assumed risk (e.g. the median control group risk across studies) calculated from data in the meta‐analyses. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: Confidence interval; RR: Relative risk
GRADE Working Group grades of evidence High quality: Further research is very unlikely to change our confidence in the estimate of effect. Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. Very low quality: We are very uncertain about the estimate.

Background

Description of the condition

Adults who have end‐stage kidney disease (ESKD) ‐ defined as estimated glomerular filtration rate (eGFR) < 15 mL/min/1.73 m² ‐ commonly experience a heavy burden of symptoms that includes fatigue/tiredness, depression and anxiety (Murtagh 2007). In a recent systematic review of observational studies (Palmer 2013a), approximately 25% of adults with chronic kidney disease (CKD) fulfilled diagnostic criteria for major depression based on a clinical interview, a prevalence that may exceed that of other settings including primary care, cancer, heart failure and diabetes (Anderson 2001; Mitchell 2011; O'Connor 2009). Depression associated with kidney disease is linked to poorer health outcomes including increased mortality (Palmer 2013b), worse quality of life (Lopes 2002; Tsay 2002), lower adherence to recommended dietary and fluid restrictions (Sensky 1996), increased utilisation of health care (Abbas Tavallaii 2009), and hospitalisation (Hedayati 2005). Health‐related quality of life may correlate more strongly with depression than with measures of dialysis adequacy (Martin 2000; Steele 1996). Depressed patients treated with peritoneal dialysis have higher complication rates from peritonitis (Juergenson 1996), while dialysis patients are more likely to commit suicide than the general population (Abram 1971; Haenel 1980). Among older dialysis patients, rates of hospitalisation for depression as a primary diagnosis are significantly greater than among patients with ischaemic heart disease, cerebrovascular disease and peptic ulcer disease (Kimmel 1998).

In addition to prevalence and severity of depressed mood in dialysis patients, adult patients on or nearing dialysis, their caregivers and health care professionals collectively cite depression among their top 10 current research uncertainties in kidney disease (Manns 2014). Similarly, patients consider the psychosocial aspects of living with CKD and reducing symptoms of kidney disease as priorities to be considered within current research agendas (Tong 2008), suggesting the optimal management depression is an important and unanswered question in the care of patients with ESKD.

Description of the intervention

Potential treatments for depression in adults with ESKD include psychosocial and psychological interventions (e.g. cognitive behavioural therapy (CBT), psychodynamic psychotherapy, interpersonal psychotherapy, non‐directive or supportive therapy and counselling). In addition, antidepressant medication may be used to treat depression including selective serotonin reuptake inhibitors (SSRI), serotonin‐norepinephrine (noradrenaline reuptake inhibitors (SNRI), tricyclic antidepressants (TCA), and monoamine oxidase inhibitors (MAOI). Clinical guidelines suggest treatment with antidepressant medication (with SSRIs as first‐line) in adults with CKD who meet diagnostic criteria for depression, based on observational studies and limited evidence of harm (ERBP 2012). Drug clearance of many antidepressants or their active metabolites is substantially reduced by kidney failure (selegiline, amitripylinoxide, venlafaxine, desvenlafaxine, milnacipran, bupropion, reboxetine) and may be altered by dialysis treatment (ERBP 2012). Accordingly the balance of harm and benefit of antidepressant therapy in adults with kidney disease may be different from that of the general population.

How the intervention might work

Associations between depression and treatment adherence, hospitalisation, mortality and quality of life in the setting of kidney disease suggest adequate treatment of depressed mood has a potential role to play in improved outcomes in this clinical setting. Existing studies have shown that depression is linked to impaired nutrition in adults with advanced kidney disease and that antidepressant therapy together with psychotherapy might improve nutritional parameters in depressed patients (Friend 1997; Koo 2003; Koo 2005). In addition, depression is associated with inflammation in dialysis patients which has been suggested as a putative contributor to cardiovascular disease in the setting of other non‐communicable diseases (Emerging Risk Factors Collaboration 2010; Vaccarino 2007). Depression is also linked to reduced social support, elevated social conflict and a greater likelihood of withdrawing from dialysis treatment (Lacson 2012). Reduced adherence with medical treatment may be caused by depression and is in turn linked to poorer outcomes, suggesting an additional causal mechanism between depression and health care utilisation that might be ameliorated by effective treatment for depression (DiMatteo 2000).

Why it is important to do this review

The high prevalence of depressive symptoms in adults treated with dialysis, associations with poor health outcomes, the markedly altered pharmacology of drug interventions for depression caused by kidney failure, together with the prioritisation of depression by patients as a key contemporary uncertainty in the care of people with CKD mandates the need for robust evidence synthesis of available treatments. In an earlier version of this review, current to February 2005, a single short‐term study comparing fluoxetine with placebo was identified. This review aims to update the current evidence for antidepressant agents to treat depression in adults with ESKD treated with dialysis and to evaluate our confidence in the evidence using the Grading of Recommendations Assessment, Development and Evaluation (GRADE) approach (Guyatt 2011).

Objectives

To evaluate the benefit and harms of antidepressants for treating depression in adults with ESKD treated with dialysis.

Methods

Criteria for considering studies for this review

Types of studies

We included RCTs and quasi‐RCTs (in which allocation to treatment was obtained by alternation, use of alternate medical records, date of birth, or other predictable methods) measuring the effect of dietary interventions in adults with ESKD treated with dialysis. We also included cross‐over studies but used only data from the first randomisation period in the review. In the event that outcome data were absent, we excluded the study from meta‐analysis but included the study in the narrative systematic review.

Types of participants

Inclusion criteria

Patients aged 18 years and over with ESKD (eGFR < 15 mL/min/1.73 m²) treated with dialysis (either haemodialysis or peritoneal dialysis) were included. Depressive disorder was considered present when investigators assessed participants using any diagnostic tool including interview or depression scale.

Exclusion criteria

We excluded studies including patients with bipolar affective disorder.

Types of interventions

Inclusion criteria

We included studies comparing antidepressants to placebo, no intervention or psychological interventions. The following drugs were considered.

• TCA and related antidepressant drugs (amitriptyline hydrochloride, amoxapine, clomipramine hydrochloride, dosulepin/dothiepin hydrochloride, doxepin, imipramine hydrochloride, nortriptyline, lofepramine, maprotiline, trazodone hydrochloride)

• SSRI (citalopram, escitalopram, fluoxetine, fluvoxamine maleate, paroxetine, sertraline)

• MAOI (moclobemide, phenelzine, isocarboxazid, tranylcypromine)

• Others (mirtazapine, nefazodone, reboxetine, venlafaxine).

Studies where there was equivalent supportive or other psychotherapy in the two arms were included (i.e. antidepressant plus psychotherapy compared with placebo plus psychotherapy). Studies with electroconvulsive therapy (ECT) in combination with antidepressants versus antidepressants alone were not included. The dosages of the drugs given and the duration of treatment were recorded.

Exclusion criteria

Studies of euphoriants (e.g. amphetamines) and adjuvants (tryptophan, lithium, carbamazepine) were excluded.

Types of outcome measures

Primary outcomes

• Health‐related quality of life

• Adverse events.

Secondary outcomes

Patient centred outcomes

• Hospitalisation

• All‐cause mortality

• Suicides or suicide attempts

• Withdrawal from study intervention

• Withdrawal from dialysis.

Surrogate outcomes

• Depression score.

Search methods for identification of studies

Electronic searches

We searched the Cochrane Kidney and Transplant's Specialised Register to 20 January 2016 through contact with the Information Specialist using search terms relevant to this review. The Specialised Register contains studies identified from the following sources.

1. Monthly searches of the Cochrane Central Register of Controlled Trials CENTRAL

2. Weekly searches of MEDLINE OVID SP

3. Handsearching of renal‐related journals and the proceedings of major renal conferences

4. Searching of the current year of EMBASE OVID SP

5. Weekly current awareness alerts for selected renal journals

6. Searches of the International Clinical Trials Register (ICTRP) Search Portal and ClinicalTrials.gov.

Studies contained in the Specialised Register are identified through search strategies for CENTRAL, MEDLINE, and EMBASE based on the scope of Cochrane Kidney and Transplant. Details of these strategies as well as a list of handsearched journals, conference proceedings and current awareness alerts are available in the Specialised Register section of information about the Cochrane Kidney and Transplant.

See Appendix 1 for search terms used in strategies for this review.

Searching other resources

1. Reference lists of clinical practice guidelines, review articles and relevant studies

2. Letters seeking information about unpublished or incomplete studies to investigators known to be involved in previous studies

3. For the original review the American College of Physicians database and PsycINFO were also searched.

Data collection and analysis

Selection of studies

For this update review (2016), study titles and abstracts were reviewed by two authors. Full text articles of studies considered relevant were obtained and reviewed for eligibility by both authors.

Data extraction and management

For this update, data extraction and assessment of risk of bias was performed by two authors using a previously prepared standard data extraction form which were piloted before use. We assessed and extracted characteristics regarding the study design, participants and methods, intervention and outcome details, summary statistics and associated commentaries. We resolved any disagreements through consultation with review author.

We extracted the following information.

1. Characteristics of participants: number of participants randomised, age, sex, multi‐morbidity (diabetes, prior myocardial infarction, prior stroke, hypertension]) depression score, tool used to assess depression, medication, type of dialysis, time on dialysis treatment

2. Characteristics of interventions: description of intervention(s), route, daily dose, duration of treatment

3. Study design: setting, year of publication, crossover or parallel study, primary outcome, months of follow up, risks of bias

4. Outcome measures: primary and secondary outcome measures, summary statistics of continuous data (mean, standard deviation (SD)) and dichotomous data (number who experienced endpoint and number at risk).

Studies reported in non‐English language journals were translated before assessment. Where more than one publication of one study existed, reports were grouped together and the publication with the most complete data was used in the analyses. Where relevant outcomes were only published in earlier versions, these data were used. Any discrepancy between published versions was highlighted.

Assessment of risk of bias in included studies

The following items were assessed independently by two authors using the risk of bias assessment tool (Higgins 2011) (seeAppendix 2).

• Was there adequate sequence generation (selection bias)?

• Was allocation adequately concealed (selection bias)?

• Was knowledge of the allocated interventions adequately prevented during the study?

  • Participants and personnel (performance bias)

  • Outcome assessors (detection bias)

• Were incomplete outcome data adequately addressed (attrition bias)?

• Are reports of the study free of suggestion of selective outcome reporting (reporting bias)?

• Was the study apparently free of other problems that could put it at a risk of bias?

We resolved disagreements regarding the risk of bias adjudications by consultation with a third review author.

Measures of treatment effect

Dichotomous data

For dichotomous outcomes (hospitalisation, mortality, suicide or suicide attempts, withdrawal from study treatment, withdrawal from dialysis, adverse events), results were expressed as risk ratio (RR) with 95% confidence intervals (CI).

Continuous data

Where continuous scales of measurement were used to assess the effects of treatment (health‐related quality of life, depression score), the mean difference (MD) was used, or the standardised mean difference (SMD) if different scales had been used.

Unit of analysis issues

Studies with multiple treatment groups

In studies comparing the efficacy of more than two interventions we considered the following:

1. If the different antidepressant interventions were of the same class (e.g. SSRI), we summarised the different experimental conditions into a single group that was compared with the control group for dichotomous outcomes (we summed the sample sizes and the number of people with events across the treatment groups). For continuous data, we entered the means and standard deviations of a single intervention group (usually the highest dose) for comparison with the control group. Where appropriate, we considered sensitivity analyses, testing the impact of including the alternative intervention group in analyses.

2. If the different interventions were of different classes (e.g. SSRI or TCA versus placebo), we included each treatment group in separate meta‐analyses, ensuring we did not include outcome data for the control group participants more than once in a single meta‐analysis

Cross‐over studies

We included cross‐over studies in meta‐analyses only if it was possible to extract data for the treatment and control groups from the first treatment period.

Dealing with missing data

Any further information required from the original author was requested by written correspondence and any relevant information obtained was to be included in the review. Evaluation of important numerical data such as screened, randomised patients as well as intention‐to‐treat, as‐treated and per‐protocol population were carefully performed. Attrition rates, for example drop‐outs, losses to follow‐up and withdrawals were investigated. Issues of missing data and imputation methods (for example, last‐observation‐carried‐forward) were critically appraised (Higgins 2011).

Assessment of heterogeneity

We first assessed for statistical heterogeneity visually by inspecting forest plots of standardised mean effect sizes and of RR. Furthermore, we applied a Chi² test to assess heterogeneity. The test has low power in general but especially when the sample size of the included studies is low or there are only a few included studies. Therefore, we used a P value of 0.10 to determine statistical significance. In addition, we used the I² statistic. The I² statistic describes the percentage of variability in effect estimates that is due to heterogeneity rather than sampling error. We used conventions of interpretation that were defined by Higgins (Higgins 2003). In the case of substantial levels (50% to 90%) and considerable levels (75% to 100%) of heterogeneity, we examined data by subgroup and sensitivity analyses (see Subgroup analysis and investigation of heterogeneity; Sensitivity analysis) for different aspects of clinical and methodological heterogeneity.

Assessment of reporting biases

In order to minimise publication bias, we made every attempt to include unpublished studies (e.g. by searching online trial registries). In order to assess for publication bias, we implemented funnel plots (effect versus standard error of the effect size) when a sufficient number of studies was available (according to recommendations of the Higgins 2011 section 10 Addressing reporting biases). For the analysis and the interpretation of the funnel plots, other reasons for asymmetry besides publication bias were considered (e.g. differences in methodological quality; true heterogeneity in intervention effects).

Data synthesis

Data were summarised using the random‐effects model and the fixed‐effect model was also used to ensure robustness of the model chosen and susceptibility to outliers.

Subgroup analysis and investigation of heterogeneity

Although subgroup analyses have to be treated with caution, as they are hypothesis‐forming rather than hypothesis‐testing, we considered a priori defined analyses in order to explore whether methodological and clinical differences between the studies may have systematically influenced the differences that were observed in the treatment outcomes. However, insufficient data were available to conduct subgroup analyses for the primary outcomes.

Sensitivity analysis

We considered sensitivity analyses to explore the influence of the following factors on effect size, however insufficient data were available.

• Repeating the analysis excluding unpublished studies

• Repeating the analysis taking account of risk of bias, as specified

• Repeating the analysis excluding any very long or large studies to establish how much they dominate the results

• Repeating the analysis excluding studies using the following filters: diagnostic criteria, language of publication, source of funding (industry versus other), and country.

'Summary of findings' tables

We have presented the main results of the review in a 'Summary of findings' table. This table presents key information concerning the quality of the evidence, the magnitude of the effects of the interventions examined, and the sum of the available data for the main outcomes (Schunemann 2011a). The 'Summary of findings' table also include an overall grading of the evidence related to each of the main outcomes using the GRADE (Grades of Recommendation, Assessment, Development and Evaluation) approach (GRADE 2008). The GRADE approach defines the quality of a body of evidence as the extent to which one can be confident that an estimate of effect or association is close to the true quantity of specific interest. The quality of a body of evidence involves consideration of within‐trial risk of bias (methodological quality), directness of evidence, heterogeneity, precision of effect estimates and risk of publication bias (Schunemann 2011b). We have presented the following outcomes.

• Dizziness or hypotension

• Nausea

• Sexual dysfunction

• All‐cause mortality

Results

Description of studies

Results of the search

2005 review

A total of 807 titles were identified in our initial search (MEDLINE (125), PSYCHINFO (570), EMBASE (234), ACP (3), The Cochrane Library (The Cochrane Database of Systematic Reviews (34), DARE (2), CENTRAL (23)), CINAHL (7), UK National Research Register (1), International Pharmaceutical Abstracts (8)). After screening, 29 full papers were retrieved for assessment for inclusion. Twenty seven studies (28 papers) were excluded as they were not RCTs (10 studies), the interventions were not relevant to this review (14), or the populations were not ones being assessed by this review (3). One study (Blumenfield 1997) with a total of 12 randomised patients was eligible for inclusion. This study compared fluoxetine with placebo in a subset of depressed dialysis patients. The selected patients were derived from a dialysis population who were either referred for psychiatric consultation, or were patients during one dialysis shift who were assessed for major depressive disorder by a psychiatrist. The study did not give standard deviations for the results but we were able to calculate it from the other data provided (mean and exact P values) in the paper. The study duration was eight weeks.

2016 review update

A search was conducted in January 2016 to update the review (Figure 1). This new search identified 144 new reports. Of these, 15 reports were of six new excluded studies (CAST Study 2013; Ciarambino 2012; dos Rios Santos 2013; SMILE Study 2010; Solak 2012; Turk 2010), 120 reports were of two existing excluded studies (ADEMEX Study 2002; HEMO Study 1997) and three reports were of three new included studies (Hosseini 2012; Taraz 2013; Yazici 2012). One study was translated to English before assessment and data extraction (Yazici 2012).

1.

Study selection flow diagram

Prior to publication four reports of two new studies (Gharekhani 2014; Solak 2010) and two ongoing studies (NCT02358343; NCT02407821) were identified. These will be assessed in a future update of this review.

Included studies

See: Characteristics of included studies

Overall, a total of four studies were included in the present review (involving 170 participants) (Blumenfield 1997; Hosseini 2012; Taraz 2013; Yazici 2012).

Design

Three of the four studies were reported as double‐blind and placebo‐controlled (Blumenfield 1997; Taraz 2013; Yazici 2012). One study reported psychological training as the control intervention (Hosseini 2012). Studies were all of short duration. Participants were followed up for eight weeks in two studies (Blumenfield 1997; Yazici 2012) and 12 weeks in two studies (Hosseini 2012; Taraz 2013). The studies included 14 (Blumenfield 1997), 44 (Hosseini 2012), 50 (Taraz 2013) and 62 (Yazici 2012) participants.

Settings

Participants were all treated with haemodialysis. One study was conducted in the USA (Blumenfield 1997), two were conducted in Iran (Hosseini 2012; Taraz 2013) and one was conducted in Turkey (Yazici 2012).

Participants

The diagnosis of major depression as a criterion for participant inclusion varied among the studies. Blumenfield 1997 diagnosed major depression if participants fulfilled 16 of the first 17 items of the Hamilton Depression Scale. Hosseini 2012 included participants who had a Hospital Anxiety and Depression Score of 8 or above. Taraz 2013 identified major depression using a Beck Depression Inventory (BDI‐II) score of 16 or above. Yazici 2012 included participants meeting DSM‐IV (Diagnostic and Statistical Manual of Mental Disorders, 4th Edition) criteria during an interview. The ages of participants were as follows.

• Blumenfield 1997: unclear

• Hosseini 2012: mean ages were 49.1 ± 14.5 years and 52.3 ± 15.6 years for the treatment and control groups, respectively

• Taraz 2013: median ages were 60 and 65 years (IQR 22 to 24.5)

• Yazici 2012: mean ages were 49.3 ± 10.3 and 52.8 ± 11.8 years.

The proportion of men ranged between 43% (Hosseini 2012) and 58% (Taraz 2013).

Intervention and comparators

The antidepressant medications prescribed in the studies together with the suggested treatment doses in ESKD (Hedayati 2012) are shown in Table 2. Prescribed doses in two included studies were comparatively low given the suggested dose range (Blumenfield 1997; Yazici 2012), one study used an antidepressant not recommended for use in people with an eGFR < 20 mL/min (citalopram) (Hosseini 2012) and in one study the dose of escitalopram (caution with use in the dialysis setting) was unclear (Taraz 2013).

1. Comparison of drug doses prescribed in studies against recommended dosing for people with end‐stage kidney disease (ESKD).

Study	Antidepressant	Dose prescribed in study	Dose suggested in ESKD	Dose level in study compared to recommendation1
Blumenfield 1997	Fluoxetine	20 mg/d	20 to 80 mg/d; no dose adjustment recommended but long half‐life; use with caution	Low
Hosseini 2012	Citalopram	20 mg/d	Initial dose 10 mg/d; active metabolite; not recommended for eGFR < 20 mL/min	High
Taraz 2013	Escitalopram	Uncertain	Use with caution in severe renal impairment	High
Yazici 2012	Sertraline	50 to 100 mg/d	50 to 200 mg/d; no dose adjustment recommended but active metabolite is renally excreted.	Low‐moderate

¹The reference doses of antidepressant medication for treatment of depression in patients with ESKD are derived from Hedayati 2012.

In three studies, an antidepressant was compared with matching placebo involving 126 participants (Blumenfield 1997; Taraz 2013; Yazici 2012) and in one study an antidepressant was compared with psychological training in 44 participants (Hosseini 2012). All studies evaluated an SSRI antidepressant including fluoxetine 20 mg daily (Blumenfield 1997), citalopram 20 mg/d (Hosseini 2012), escitalopram uncertain dose daily (Yazici 2012) and sertraline 50 mg daily for 2 weeks then 100 mg daily (Taraz 2013). In Hosseini 2012 the comparator treatment included group participation in six sessions of one hour teamwork training every other day including explaining the anatomy of the kidney, pathophysiology and causes of kidney failure, treatment modalities with advantages and disadvantages, the mechanism of haemodialysis, the required care in haemodialysis, the stages of adaptation, and techniques of problem‐solving, stress management and muscle relaxation. The participants who received citalopram did not attend these sessions.

Outcomes

All studies were included in the meta‐analysis. None of the studies reported data for health‐related quality of life, while all studies reported some adverse events related to treatment although this was not systematic. No study reported hospitalisation, suicide or attempted suicide, withdrawal from dialysis, or adherence to the recommended dialysis treatment as study endpoints. One study reported deaths during follow up (Taraz 2013), all reported withdrawal from study treatment and all four reported effects of treatment on depression scores (either change in score during follow up and/or end of treatment values) (Blumenfield 1997; Hosseini 2012; Taraz 2013; Yazici 2012). Data for change in depression scores for an antidepressant versus placebo was only possible for one study (Taraz 2013).

Excluded studies

See: Characteristics of excluded studies.

Thirty two studies did not meet our inclusion criteria and were excluded because of one of the following reasons: not assessing antidepressant medication, not including participants treated with dialysis, not including participants meeting criteria for depression, and were not RCTs.

Risk of bias in included studies

See Characteristics of included studies and Figure 2 and Figure 3. The risk of bias for many domains was generally unclear or high.

2.

Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.

3.

Risk of bias summary: review authors' judgements about each risk of bias item for each included study.

Allocation

Random sequence generation

Only one of the four studies reported the methods used to generate the random sequence for allocation (Taraz 2013).

Allocation concealment

All studies failed to report details of allocation concealment and were classified as unclear risk.

Blinding

Two studies reported that both participants and investigators were blinded to treatment allocation (Blumenfield 1997; Taraz 2013), one did not report blinding (Yazici 2012) and one used a comparator intervention (group‐based counselling) that made it unlikely that treatment allocation was blinded (Hosseini 2012). None of the studies specifically reported blinded outcome assessment for any outcome.

Incomplete outcome data

Three studies reported loss to follow‐up < 10% of participants, and were classified as low risk (Blumenfield 1997; Hosseini 2012; Yazici 2012) while Taraz 2013 was classified as high risk.

Selective reporting

Three studies provided detailed descriptions of adverse events and depression scores and were adjudicated as low risk (Blumenfield 1997; Taraz 2013; Yazici 2012). Hosseini 2012 indicated that adverse events were not different between groups but provided no additional specific details and was classified high risk.

Other potential sources of bias

None of the studies appeared to have other sources of bias.

Effects of interventions

See: Table 1

Antidepressant versus placebo

Primary outcomes

Quality of life

This outcome was not reported in the available studies.

Adverse events attributable to treatment or control intervention

The number of participants experiencing adverse events in each study is reported in Table 3, Table 4, and Table 5. In meta‐analysis, antidepressant therapy had uncertain effects on increasing the risk of on dizziness or hypotension (Analysis 1.1.1 (3 studies, 114 participants): RR 1.72, 95% CI 0.75 to 3.92; I² = 0%), increased nausea (Analysis 1.1.2 (3 studies, 114 participants): RR 2.67, 95% CI 1.26 to 5.68; I² = 0%), and had uncertain risks of sexual dysfunction (Analysis 1.1.3 (2 studies, 101 participants): RR 3.83, 95% CI 0.63 to 23.3; I² = 0%) and headache (Analysis 1.1.4 (2 studies, 56 participants): RR 2.91, 95% CI 0.73 to 11.57; I² = 0%).

2. Number of patients experiencing adverse events (Blumenfield 1997).

System	Event	Fluoxetine (6)	Placebo (7)
Autonomic	Dry mouth	0	1
Cardiovascular	Hypotension	4	1
Gastrointestinal	Abdominal pain	1	2
	Constipation	0	1
	Diarrhoea	1	1
	Gastroenteritis	0	2
	Nausea	5	2
	Vomiting	3	3
Musculoskeletal	Myalgia	1	1
Neurological	Dizziness	1	0
	Headache	3	0
	Insomnia	2	1
	Sensation disturbance	1	0
	Tremors	1	0
Psychiatric	Abnormal thought	1	0
	Anxiety	0	1
	Nervousness	1	1
Respiratory	Bronchitis	1	0
	Cough	0	2
	Dyspnoea	1	0
	Pharyngitis	1	0
	Rhinitis	1	0
	Upper respiratory tract infection	1	0
Skin	Furunculosis	1	0
	Pruritis	1	0
	Skin ulcer	0	1
Other	Dehydration	0	1
	Oedema	0	1
	Flu syndrome	1	0
	Tooth infection	1	0

3. Number of participants experiencing adverse events (Taraz 2013).

System	Event	Sertraline (21)	Placebo (22)
Autonomic	Sexual dysfunction	2	1
Cardiovascular	Dizziness	5	3
Gastrointestinal	Dyspepsia	6	4
	Anorexia	2	4
	Nausea	7	3
Other	Headache	4	2
	Hair loss	1	1

4. Number of patients experiencing adverse events (Yazici 2012).

System	Event	Escitalopram (28)	Placebo (30)
Autonomic	Ejaculation disorder or impotence	4	0
Cardiovascular	Feeling of dizziness	4	4
Gastrointestinal	Nausea	5	2
	Diarrhoea	2	1
Psychiatric	Insomnia	4	2
	Somnolence	2	2
Other	Flu‐like symptoms	1	2

1.1. Analysis.

Comparison 1 Antidepressant versus placebo, Outcome 1 Adverse events.

Secondary outcomes

All‐cause mortality

Taraz 2013 reported antidepressant therapy had uncertain effects on all‐cause mortality with few events (Analysis 1.2 (50 participants): RR 1.00, 95% CI 0.07 to 15.12).

1.2. Analysis.

Comparison 1 Antidepressant versus placebo, Outcome 2 All‐cause mortality.

Withdrawal from antidepressant medication

There was no difference in withdrawal from study treatment between antidepressant therapy and placebo, although the confidence interval was wide (Analysis 1.3 (3 studies, 118 participants): RR 1.39, 95% CI 0.44 to 4.47; I² = 0%).

1.3. Analysis.

Comparison 1 Antidepressant versus placebo, Outcome 3 Withdrawal from treatment (antidepressant).

Depression score

Only Taraz 2013 reported this outcome was reported in an extractable form. Sertraline therapy reduced the Beck Depression Inventory score significantly compared to placebo (Analysis 1.4 (43 participants): MD ‐7.50, 95% CI ‐11.94 to ‐3.06).

1.4. Analysis.

Comparison 1 Antidepressant versus placebo, Outcome 4 End of treatment depression score.

In narrative results, Blumenfield 1997 reported that depression scores were not significantly different between treatment groups at the end of the study (including the Beck Depression Inventory, the Brief Symptom Inventory, the Hamilton Depression Inventory, the Montgomery Asberg Depression Scale, and the self‐evaluating depression score assessed using an electronic visual analogue instrument). Yazici 2012 reported a significantly lower Hamilton Rating Scale for Depression score with antidepressant treatment at end of follow up (10.5, minimum 4, maximum 35 with escitalopram versus 28, minimum 7, maximum 35) (P = 0.001 for difference).

Outcomes not reported

Hospitalisation, suicide or attempted suicide, and withdrawal from dialysis therapy were not reported by any of the included studies.

Antidepressant versus psychological counselling

Primary outcomes

Quality of life

This outcome was not reported in one available study (Hosseini 2012).

Adverse events attributable to treatment or control intervention

Hosseini 2012 reported that antidepressant treatment "was well tolerated by all of the patients and no severe adverse effects were reported".

Secondary outcomes

Withdrawal from antidepressant medication

Hosseini 2012 reported no difference in withdrawal from study treatment between antidepressant therapy and psychological intervention, although the confidence interval was wide (Analysis 2.1 (44 participants): RR 3.00, 95% CI 0.34 to 26.6).

2.1. Analysis.

Comparison 2 Antidepressant versus psychological support, Outcome 1 Withdrawal from treatment (antidepressant).

Depression score

Hosseini 2012 reported the depression score (Hospital Anxiety and Depression Scale) was similar at the end of follow up for the antidepressant treatment and psychological support groups (Analysis 2.2 (40 participants): MD ‐1.07, 95% CI ‐3.85 to 1.71).

2.2. Analysis.

Comparison 2 Antidepressant versus psychological support, Outcome 2 End of treatment depression score.

Outcomes not reported

All‐cause mortality, hospitalisation, suicide or attempted suicide, and withdrawal from dialysis therapy were not reported by any of the included studies.

Subgroup analysis

The paucity of studies precluded planned subgroup analyses.

Assessment of heterogeneity

For all outcomes in which two or more studies could be included in meta‐analysis, there was no evidence of statistical heterogeneity in any analyses.

Assessment of publication bias

Insufficient data precluded assessment of funnel plots to evaluate for evidence of small study effects.

Discussion

Summary of main results

In this systematic review and meta‐analysis, evidence for the effectiveness and safety of antidepressant medication in adults with ESKD treated with dialysis is very sparse. Three studies have compared an SSRI with placebo involving 126 participants and a single study has compared SSRI therapy with psychological support in 44 participants. All studies were short, limited to 12 weeks or fewer and meta‐analyses for all outcomes included 118 participants or fewer. We generally had very low confidence in the estimated treatment effects or effects were not estimable due to limitations in study reporting, substantial imprecision in treatment effects and small numbers of studies, limiting the applicability of the findings to the general dialysis population.

Compared to placebo, short‐term SSRI therapy may reduce clinical depression scores in ungradeable evidence, at the expense of incurring nausea. Effects on all‐cause mortality, suicide, other adverse events such as hypotension or dizziness, headache and sexual dysfunction, withdrawal from dialysis, and hospitalisation were not estimable due to insufficient data. Compared to group psychological training, SSRI therapy had no statistical effect on depression scores, and withdrawal from treatment was difficult to ascertain due to few events in the available study.

Overall, despite the potential prevalence of major depression in adults treated with dialysis, estimated to affect up to one‐quarter of patients, and the priority patients and caregivers place on management of depression as an important clinical uncertainty, available study evidence for the safety and efficacy of antidepressant therapy is currently lacking.

Overall completeness and applicability of evidence

While our review was based on a highly sensitive electronic search strategy and included studies from the Cochrane Kidney and Transplant's Specialised Register of studies, which includes results from handsearching and journal alerts, the paucity of existing studies evaluating pharmacological treatment of depression in the setting of dialysis patients means there are considerable evidence gaps. First, studies evaluated depression using heterogeneous methods including depression scales and interview techniques. Due to the heavy symptom burden of ESKD, depression scales in the general population may lack validity in the setting of kidney failure and falsely detect the presence of depression due to somatic symptoms associated with kidney disease, such as fatigue and anorexia. Few studies exist to validate depression tools in the context of ESKD. Our recent meta‐analysis (although not specifically designed to assess diagnostic test accuracy for depression scales) suggested that depression scales are likely to over‐estimate depression prevalence in this population (Palmer 2013a).

Studies were short in duration (12 weeks or fewer). The relative efficacy and safety of longer term antidepressant treatment is unknown. Outcome data were sparse and confidence intervals were frequently very wide, including for important adverse events including hypotension. The development of a core set of patient‐prioritised outcomes for depression in adults with CKD, which are required to be reported in clinical studies would assist with building evidence for these drugs in dialysis patients as studies cumulate. Additionally, as clearance of antidepressant drugs or their metabolites is commonly impaired in ESKD, safety data for antidepressant medication are required. Notably, in this review only SSRI were studied. The effectiveness and safety of other agents, including TCA remains unknown. Information was limited to small studies conducted in the USA, Turkey and Iran and results may not be generalisable to other national settings.

Quality of the evidence

We graded our confidence in the evidence using the Grading of Recommendations Assessment, Development and Evaluation (GRADE) approach (Guyatt 2011), which considers study limitations, imprecision, indirectness, inconsistency and publication bias. Overall, most studies had high or unclear risks of bias for most domains of study reporting assessed. Estimated effects on efficacy and safety outcomes were frequently imprecise with confidence intervals that were consistent with both considerable benefit or harm. The generalisability (directness) of the evidence was limited by a gross paucity of available studies and publication bias (the effects of small studies on treatment effects) could not be assessed. While treatment estimates showed no evidence of statistical heterogeneity, the small number of studies limited the power of statistical testing and important inconsistency between studies could not excluded. Overall, based on important limitations, we have generally very low confidence in the evidence for the benefits and harms of antidepressant therapy in dialysis patients and any estimations of effect are very uncertain.

Potential biases in the review process

While this review was conducted according to a prespecified protocol and is reported using Cochrane methods, the review has limitations which should be considered when interpreting the results. First, relatively few data were available resulting in inconclusive evidence for many outcomes including mortality, suicide, hospitalisation, depression scores and adverse events. Second, while the included studies appeared similar in their treatment approaches to depression, there was insufficient power in analyses to detect any heterogeneity between differing studies. Third, adverse event data were collated ad hoc in the included studies and important adverse events may not have been captured. Finally, outcome data for patient‐important outcomes were not available in most studies, which reduced our confidence in the reliability of these treatment effects. This was particularly the case for reporting of quality of life domains.

Agreements and disagreements with other studies or reviews

In 2012, a meta‐analysis of antidepressant therapy for adults with stage 3 to 5 CKD (eGFR < 60 mL/min/1.73 m²) included both randomised and non‐randomised studies (ERBP 2012). In that meta‐analysis, three RCTs were identified including one ongoing study in adults not treated with dialysis (CAST Study 2013) and two studies in dialysis patients (Blumenfield 1997; Yazici 2012). In that review, insufficient outcome data were available for Yazici 2012 which was an abstract only publication at the time of publication, and, similar to this review, Blumenfield 1997 was adjudicated to show similar effects of fluoxetine on depression scores as placebo, with numerically higher numbers of adverse events with fluoxetine including hypotension. Based on evidence from nine additional non‐randomised studies, the review authors concluded that there was evidence suggesting benefit for antidepressant therapy with commonly‐experienced but mild side‐effects. The review concluded that evidence for treatment for depression is currently insufficient to guide treatment and that a well‐designed RCT is greatly needed. Guidelines also suggest early re‐assessment of therapy to assess effectiveness and to avoid prolonged use (ERBP 2012). Based on this updated Cochrane review, our findings are consistent with the need to evaluate antidepressant treatment in dialysis patients within the setting of new RCTs.

We note also that a relative lack of evidence for efficacy of antidepressant treatment in the setting of CKD is consistent with a broader concern about the efficacy of antidepressant treatment in the general population (Turner 2008). Given that the effects of antidepressants are modest even for people with normal kidney function, it is quite plausible that even small increases in treatment‐related harm in dialysis patients might offset this small benefit, even if the efficacy of treatment is similar for dialysis patients compared to the general population.

Authors' conclusions

Implications for practice.

Our main finding was that there is insufficient evidence to support prescribing of antidepressant medications in adults treated with dialysis outside of participation in RCTs. The potential for increased harm with antidepressant medication (due to impaired clearance of drugs and their metabolites) is also insufficiently evaluated by current evidence and application of research information derived from the general population may not be appropriate. Data from a few studies in this review suggests nausea may be caused by antidepressant therapy. Identification of effective treatments for depression is highly valued by patients mandating the need for further placebo‐controlled RCTs evaluating both drug effectiveness and safety based on study endpoints prioritised by patients and clinicians.

Implications for research.

There is a great need for undertaking further RCTs that assess the effectiveness of antidepressant treatment in patients treated with dialysis. New studies are needed to improve our confidence about the effectiveness and safety of drug treatments for depression. Studies evaluating drug therapy should be sufficiently powered to detect adverse events and patient‐centred outcomes. Patient and clinician priorities for clinical endpoints should be considered in the design of future RCTs. Standardisation of study design and outcomes would facilitate prospective and collaborative meta‐analysis of treatment effects. Ideally, new studies would include participants meeting diagnostic criteria for depression based on interview methods and evaluate longer term treatment and in patients with a range of depression severity including severe forms of depression. Greater understanding of the barriers to treatment implementation and the sustainability of interventions and follow up would be helpful.

What's new

Date	Event	Description
28 February 2016	New search has been performed	Changed authorship and updated review
28 February 2016	New citation required and conclusions have changed	Three new studies added, no change to conclusion. Adverse event data reported

History

Protocol first published: Issue 4, 2003
 Review first published: Issue 2, 2005

Date	Event	Description
14 October 2008	Amended	Converted to new review format.

Appendices

Appendix 1. Electronic search strategies

Database	Search terms
CENTRAL	MeSH descriptor: [Renal Dialysis] explode all trees MeSH descriptor: [Hemofiltration] explode all trees MeSH descriptor: [Kidney Failure, Chronic] explode all trees dialysis:ti,ab,kw in Trials (Word variations have been searched) hemodialysis or haemodialysis:ti,ab,kw in Trials (Word variations have been searched) hemofiltration or haemofiltration:ti,ab,kw in Trials (Word variations have been searched) hemodiafiltration or haemodiafiltration:ti,ab,kw in Trials (Word variations have been searched) CAPD or CCPD or APD:ti,ab,kw in Trials (Word variations have been searched) "end‐stage kidney" or "end‐stage renal" or "endstage kidney" or "endstage renal":ti,ab,kw in Trials (Word variations have been searched) eskd or eskf or esrd or esrf:ti,ab,kw in Trials (Word variations have been searched) #1 or #2 or #3 or #4 or #5 or #6 or #7 or #8 or #9 or #10 MeSH descriptor: [Depression] explode all trees MeSH descriptor: [Depressive Disorder] explode all trees MeSH descriptor: [Adjustment Disorders] explode all trees MeSH descriptor: [Adaptation, Psychological] explode all trees depression or depressed or depressive or anxiety or anxious:ti,ab,kw in Trials (Word variations have been searched) #12 or #13 or #14 or #15 or #16 #11 and #17
MEDLINE	exp Renal Dialysis/ exp Hemofiltration/ Kidney Failure, Chronic/ dialysis.tw. (hemodialysis or haemodialysis).tw. (hemofiltration or haemofiltration).tw. (hemodiafiltration or haemodiafiltration).tw. (CAPD or CCPD or APD).tw. (end‐stage kidney or end‐stage renal or endstage kidney or endstage renal).tw. (ESKD or ESKF or ESRD or ESRF).tw. or/1‐10 Depression/ exp Depressive Disorder/ Adjustment Disorders/ exp Adaptation, Psychological/ (depression or depressed or anxiety or anxious).tw. exp Antidepressive Agents/ or/12‐17 and/11,18
EMBASE	exp Renal Replacement Therapy/ (hemodialysis or haemodialysis).tw. (hemofiltration or haemofiltration).tw. (hemodiafiltration or haemodiafiltration).tw. dialysis.tw. (CAPD or CCPD or APD).tw. Chronic Kidney Disease/ Kidney Failure/ Chronic Kidney Failure/ (end‐stage renal or end‐stage kidney or endstage renal or endstage kidney).tw. (ESRF or ESKF or ESRD or ESKD).tw. or/1‐11 exp depression/ and/12‐13

Appendix 2. Risk of bias assessment tool

Potential source of bias	Assessment criteria
Random sequence generation Selection bias (biased allocation to interventions) due to inadequate generation of a randomised sequence	Low risk of bias: Random number table; computer random number generator; coin tossing; shuffling cards or envelopes; throwing dice; drawing of lots; minimisation (minimisation may be implemented without a random element, and this is considered to be equivalent to being random).
	High risk of bias: Sequence generated by odd or even date of birth; date (or day) of admission; sequence generated by hospital or clinic record number; allocation by judgement of the clinician; by preference of the participant; based on the results of a laboratory test or a series of tests; by availability of the intervention.
	Unclear: Insufficient information about the sequence generation process to permit judgement.
Allocation concealment Selection bias (biased allocation to interventions) due to inadequate concealment of allocations prior to assignment	Low risk of bias: Randomisation method described that would not allow investigator/participant to know or influence intervention group before eligible participant entered in the study (e.g. central allocation, including telephone, web‐based, and pharmacy‐controlled, randomisation; sequentially numbered drug containers of identical appearance; sequentially numbered, opaque, sealed envelopes).
	High risk of bias: Using an open random allocation schedule (e.g. a list of random numbers); assignment envelopes were used without appropriate safeguards (e.g. if envelopes were unsealed or non‐opaque or not sequentially numbered); alternation or rotation; date of birth; case record number; any other explicitly unconcealed procedure.
	Unclear: Randomisation stated but no information on method used is available.
Blinding of participants and personnel Performance bias due to knowledge of the allocated interventions by participants and personnel during the study	Low risk of bias: No blinding or incomplete blinding, but the review authors judge that the outcome is not likely to be influenced by lack of blinding; blinding of participants and key study personnel ensured, and unlikely that the blinding could have been broken.
	High risk of bias: No blinding or incomplete blinding, and the outcome is likely to be influenced by lack of blinding; blinding of key study participants and personnel attempted, but likely that the blinding could have been broken, and the outcome is likely to be influenced by lack of blinding.
	Unclear: Insufficient information to permit judgement
Blinding of outcome assessment Detection bias due to knowledge of the allocated interventions by outcome assessors.	Low risk of bias: No blinding of outcome assessment, but the review authors judge that the outcome measurement is not likely to be influenced by lack of blinding; blinding of outcome assessment ensured, and unlikely that the blinding could have been broken.
	High risk of bias: No blinding of outcome assessment, and the outcome measurement is likely to be influenced by lack of blinding; blinding of outcome assessment, but likely that the blinding could have been broken, and the outcome measurement is likely to be influenced by lack of blinding.
	Unclear: Insufficient information to permit judgement
Incomplete outcome data Attrition bias due to amount, nature or handling of incomplete outcome data.	Low risk of bias: No missing outcome data; reasons for missing outcome data unlikely to be related to true outcome (for survival data, censoring unlikely to be introducing bias); missing outcome data balanced in numbers across intervention groups, with similar reasons for missing data across groups; for dichotomous outcome data, the proportion of missing outcomes compared with observed event risk not enough to have a clinically relevant impact on the intervention effect estimate; for continuous outcome data, plausible effect size (difference in means or standardised difference in means) among missing outcomes not enough to have a clinically relevant impact on observed effect size; missing data have been imputed using appropriate methods.
	High risk of bias: Reason for missing outcome data likely to be related to true outcome, with either imbalance in numbers or reasons for missing data across intervention groups; for dichotomous outcome data, the proportion of missing outcomes compared with observed event risk enough to induce clinically relevant bias in intervention effect estimate; for continuous outcome data, plausible effect size (difference in means or standardized difference in means) among missing outcomes enough to induce clinically relevant bias in observed effect size; ‘as‐treated’ analysis done with substantial departure of the intervention received from that assigned at randomisation; potentially inappropriate application of simple imputation.
	Unclear: Insufficient information to permit judgement
Selective reporting Reporting bias due to selective outcome reporting	Low risk of bias: The study protocol is available and all of the study’s pre‐specified (primary and secondary) outcomes that are of interest in the review have been reported in the pre‐specified way; the study protocol is not available but it is clear that the published reports include all expected outcomes, including those that were pre‐specified (convincing text of this nature may be uncommon).
	High risk of bias: Not all of the study’s pre‐specified primary outcomes have been reported; one or more primary outcomes is reported using measurements, analysis methods or subsets of the data (e.g. sub‐scales) that were not pre‐specified; one or more reported primary outcomes were not pre‐specified (unless clear justification for their reporting is provided, such as an unexpected adverse effect); one or more outcomes of interest in the review are reported incompletely so that they cannot be entered in a meta‐analysis; the study report fails to include results for a key outcome that would be expected to have been reported for such a study.
	Unclear: Insufficient information to permit judgement
Other bias Bias due to problems not covered elsewhere in the table	Low risk of bias: The study appears to be free of other sources of bias.
	High risk of bias: Had a potential source of bias related to the specific study design used; stopped early due to some data‐dependent process (including a formal‐stopping rule); had extreme baseline imbalance; has been claimed to have been fraudulent; had some other problem.
	Unclear: Insufficient information to assess whether an important risk of bias exists; insufficient rationale or evidence that an identified problem will introduce bias.

Data and analyses

Comparison 1. Antidepressant versus placebo.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Adverse events	3		Risk Ratio (IV, Random, 95% CI)	Subtotals only
1.1 Dizziness/hypotension	3	114	Risk Ratio (IV, Random, 95% CI)	1.72 [0.75, 3.92]
1.2 Nausea	3	114	Risk Ratio (IV, Random, 95% CI)	2.67 [1.26, 5.68]
1.3 Sexual dysfunction	2	101	Risk Ratio (IV, Random, 95% CI)	3.83 [0.63, 23.34]
1.4 Headache	2	56	Risk Ratio (IV, Random, 95% CI)	2.91 [0.73, 11.57]
2 All‐cause mortality	1		Risk Ratio (IV, Random, 95% CI)	Totals not selected
3 Withdrawal from treatment (antidepressant)	3	118	Risk Ratio (IV, Random, 95% CI)	1.39 [0.44, 4.47]
4 End of treatment depression score	1		Mean Difference (IV, Random, 95% CI)	Totals not selected

Comparison 2. Antidepressant versus psychological support.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Withdrawal from treatment (antidepressant)	1		Risk Ratio (IV, Random, 95% CI)	Totals not selected
2 End of treatment depression score	1		Mean Difference (IV, Random, 95% CI)	Totals not selected

Characteristics of studies

Characteristics of included studies [ordered by study ID]

Blumenfield 1997.

Methods	Study design: parallel RCT Time frame: not reported Follow‐up period: 8 weeks
Participants	Country: USA Setting: Westchester Artificial Kidney Center in Valhalla, New York and New York Hospital Center of Queens, New York Patients 18 to 70 years; normal liver function tests; Major Depressive Disorder according to Hamilton Depression Scale; informed written consent Number: treatment group (7); control group (7) Mean age ± SD (years); not reported Sex (M/F): not reported Mean time treated with dialysis: not reported Depression criteria: Major depressive disorder with a total score of at least 16 on the first 17 items on the Hamilton Depression Scale Exclusion criteria: chronic illness other than CKD or DM causing CKD; axis I psychiatric disorder other than Major Depressive Disorder; serious suicidal risk; on psychotropic medication other than lorazepam; pregnant subjects or women of child bearing age not using contraception; involvement in any other drug study for 4 weeks prior to starting this study
Interventions	Treatment group Fluoxetine 20 mg/d Control group Placebo capsules (identical in appearance to fluoxetine capsules)
Outcomes	Withdrawal from medication Adverse events *Mean change in Beck Depression Inventory score, Montgomery Asberg Depression Scale, Brief Symptom Inventory (Depression) Scale, Hamilton Depression Scale and the Visual Analogue Scale. *Plasma concentrations of fluoxetine and norfluoxetine *Not included in meta‐analyses
Notes	Funding: placebo capsules supplied by Eli Lilly
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Method of randomisation was not mentioned in the original paper. However the author informed through personal communication that patients were allotted to their respective groups with randomisation done by pharmacy. Method unknown
Allocation concealment (selection bias)	Unclear risk	The random assignment was done by the hospital pharmacy
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Psychiatrists (i.e. both investigators and outcome assessors) and patients were blind, Placebo was similar in appearance to fluoxetine capsules; double blind
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not described
Incomplete outcome data (attrition bias) All outcomes	Low risk	Lost to follow‐up: 1/14 (7.1%)
Selective reporting (reporting bias)	Low risk	No study protocol was available. The study investigators described adverse events associated with treatment and treatment effects on depression scores. Other patient‐prioritised outcomes were not reported such as quality of life
Other bias	Low risk	None observed

Hosseini 2012.

Methods	Study design: parallel RCT Time frame: not reported Follow up period: 12 weeks
Participants	Country: Iran Setting: Hemodialysis Center of Imam Khomeini Hospital, Sari, Iran Patients with depressive and anxiety symptoms; depression measured using Hospital Anxiety and Depression Scale (score ≥ 8); long‐term HD Number (randomised/analysed): treatment group (22/19); control group (22/21) Mean age ± SD (years): treatment group (52.3 ± 15.6); control group (49.1 ± 14.5) Sex (M): treatment group (42%); control group (45%) Mean time treated with dialysis: not reported Exclusion criteria: did not signed the consent form for participation; previous history of psychiatric disorders; suffered from stressors other than ESKD in the past 6 months; new anxiety episode developed during the study
Interventions	Treatment group Citalopram group 20 mg/d for 3 months Control group Training group attend 6 sessions of 1‐hour psychological training
Outcomes	Adverse events Anxiety and depression scores measured using the Hospital Anxiety and Depression Scale
Notes	Funding: Manzandaran University of Medical Sciences
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Not described
Allocation concealment (selection bias)	Unclear risk	Not described
Blinding of participants and personnel (performance bias) All outcomes	High risk	Unlikely to be blinded as active comparison including psychological training
Blinding of outcome assessment (detection bias) All outcomes	High risk	Not described for any outcome and unlikely to be blinded
Incomplete outcome data (attrition bias) All outcomes	Low risk	Lost to follow‐up: 4/44 (9.1%)
Selective reporting (reporting bias)	High risk	No study protocol was available. The study investigators described adverse events associated with treatment, although actual numbers of events and a detailed description were not provided. Treatment effects on depression scores were detailed. Other patient‐prioritised outcomes were not reported such as quality of life
Other bias	Low risk	None observed

Taraz 2013.

Methods	Study design: parallel group RCT Time frame: not reported Follow up period: 12 weeks
Participants	Country: Iran Setting: Tehran University of Medical Sciences Adults 18 and 80 years; at least 3 months on HD using AV fistula; diagnosis of depression based on the Beck Depression Inventory (2nd edition) Number (randomised/analysed): treatment group (25/21); control group (25/22) Median age, IQR (years); treatment group (60, 22); control group (65, 24.5) Sex (M): treatment group (57%); control group (59%) Time treated with dialysis (median, IQR): treatment group (42, 59 months); control group (34, 76 months) Exclusion criteria: inflammatory cause of ESKD; autoimmune disease; active infection; malignancy; severe mental illness; cognitive dysfunction; haemorrhage/clotting disorder; hypersensitivity to sertraline; treatment with antibiotics; non‐steroid anti‐inflammatory drugs; steroids, immunosuppression or antidepressant medications within 1 month of study
Interventions	Treatment group Sertraline 50 mg for first 2 weeks and then 100 mg for next 10 weeks Control group Identical placebo
Outcomes	Adverse events Mortality Withdrawal from antidepressant treatment Depression score
Notes	Funding: Tehran University of Medical Sciences
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Computer generated randomisation list
Allocation concealment (selection bias)	Unclear risk	Not described
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Double‐blinded. Placebo tablets were visibly identical to sertraline tablets
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not described for any outcome
Incomplete outcome data (attrition bias) All outcomes	High risk	7/50 participants were not included in final analysis (14.0%)
Selective reporting (reporting bias)	Low risk	No study protocol was available. The study investigators described adverse events associated with treatment and treatment effects on depression scores. Other patient‐prioritised outcomes were not reported such as quality of life. The study reported mortality during follow up
Other bias	Low risk	None observed

Yazici 2012.

Methods	Study design: parallel RCT Time frame: unclear Follow up period: 8 weeks
Participants	Country: Turkey Setting: Calismaya Mersin Renal Nephrology Centre Patients aged 18 to 65 years with ESKD; depression (patients with Zung Depression Scale > 50% were interviewed and those diagnosed with major depressive disorder based DSM‐IV Number: treatment group (30); control group (28) Mean age ± SD (years): treatment group (49.3 ± 11.8); control group (52.8 ± 10.3) Sex (M/F): 23/35 Median, min to max time treated with dialysis (months): treatment group (36.5, 8 to 94); control group (32.5, 2 to 85) Exclusion criteria: any other comorbid axis I or II disorders; patients taking any psychotropic medications
Interventions	Treatment group Escitalopram daily for 8 weeks Control group Placebo
Outcomes	Adverse events Withdrawal from antidepressant treatment *Depression score
Notes	Funding: Calismaya Mersin Renal Nephrology Centre
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Not described
Allocation concealment (selection bias)	Unclear risk	Not described
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not described
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not described for any outcome
Incomplete outcome data (attrition bias) All outcomes	Low risk	Lost to follow‐up: 4/62 (6.5%)
Selective reporting (reporting bias)	Low risk	No study protocol was available. The study investigators described adverse events associated with treatment and treatment effects on depression scores. Other patient‐prioritised outcomes were not reported such as quality of life
Other bias	Low risk	None observed

AV ‐ arteriovenous; CKD ‐ chronic kidney disease; DM ‐ diabetes mellitus; ESKD ‐ end‐stage kidney disease; HD ‐ haemodialysis; IQR ‐ interquartile range; M/F ‐ male/female; RCT ‐ randomised controlled trial; SD ‐ standard deviation

Characteristics of excluded studies [ordered by study ID]

Study	Reason for exclusion
ADEMEX Study 2002	RCT: intervention CAPD vs CAPD + CrCl target; not assessing treatment of depression
Ancarani 1993	The active drug (S‐adenosyl‐L‐methionine) is not one of the groups of drugs considered for inclusion in this review
Briggs 2004	RCT; psychosocial intervention
CAST Study 2013	Not dialysis population
Chang 2004a	RCT; intervention not relevant to the review
Cho 2004	RCT; psychosocial intervention
Ciarambino 2012	Not CKD population
De Felice 1996	Not RCT (survey)
Deniston 1990	Not RCT; outcomes not relevant to this review
dos Rios Santos 2013	Not evaluating antidepressants
HEMO Study 1997	All dialysis patients; quality of life a secondary outcome
Kennedy 1989	Not RCT
Koo 2005	Not RCT
Kurella 2005	Not RCT
Levy 1996	Not RCT
Lieh 2004	Not RCT; mixed population
Miller 2005	Mixed population; psychosocial interventions
Moug 2004	Pilot RCT; psychosocial intervention
Pearlman 1988	Not RCT (case report)
Seabolt 2001	Not RCT (case report)
Sharp 2005	RCT; assessing interventions for fluid restriction compliance
SMILE Study 2010	Not RCT; not physical measures
Solak 2012	Not evaluating antidepressant
Tsay 2004a	Not evaluating antidepressant
Tsay 2004b	Not evaluating antidepressant
Tsay 2004c	Not evaluating antidepressant
Tsay 2005	RCT; psychosocial intervention
Turk 2010	Not evaluating antidepressant
van Vilsteren 2005	RCT; psychosocial intervention
Wuerth 2001	Not RCT
Zetin 1980	The active drug (zinc) used in this study is not one of the group of drugs considered eligible for inclusion
Zetin 1982	Study population consisted of all dialysis patients, not just depressed dialysis patients

CAPD ‐ continuous ambulatory peritoneal dialysis; CKD ‐ chronic kidney disease; CrCl ‐ creatinine clearance; RCT ‐ randomised controlled trial

Differences between protocol and review

'Summary of findings' table has been incorporated.

Contributions of authors

2016 review update

• SP and PN undertook searches

• PN, MR, VS, and KSR identified studies and helped with data extraction

• SP and PN drafted the review

• All authors provided intellectual review of the drafts

• SP and PN revised the review in response to comments

• SP and PN generated the figures

• GS provided input in case of any disagreement during screening process

• All authors read and approved the review before submission

2005 review

• Allison MacLeod, Conal Daly and Paul Roderick conceived the review with input from the Renal Association (UK), Standards and Audit sub‐committee.

• AMM helped with resolving differences in study inclusion and data extraction.

• KSR and Sheila Wallace undertook searches.

• KSR and CD screened search results, assessed retrieved articles, assessed quality of papers, extracted data and wrote the review.

• PR gave advice on outcome measures, study inclusion criteria and helped with resolving differences in study inclusion and data extraction.

• Janet Butler gave advice on the background for the review, types of intervention and outcome measures and wrote the review.

Declarations of interest

• Suetonia C Palmer: none known

• Patrizia Natale: none known

• Marinella Ruospo: none known

• Valeria Saglimbene: none known

• Kannaiyan S Rabindranath: none known

• Jonathan C Craig: none known

• Giovanni FM Strippoli: none known

New search for studies and content updated (conclusions changed)