Association of androgen deprivation therapy and depression in the treatment of prostate cancer: A systematic review and meta-analysis

Urologic Oncology: Seminars and Original Investigations ] (2017) ∎∎∎–∎∎∎

Original article

Association of androgen deprivation therapy and depression in the treatment of prostate cancer: A systematic review and meta-analysis Kevin T. Nead, M.D., M.Phil.a,*, Sumi Sinha, M.D.b, David D. Yang, B.A.b, Paul L. Nguyen, M.D.b a

Department of Radiation Oncology, Perelman School of Medicine, Perelman Center for Advanced Medicine, University of Pennsylvania, Philadelphia, PA b Department of Radiation Oncology, Dana-Farber Cancer Institute and Brigham and Women’s Hospital, Harvard Medical School, Boston, MA Received 6 April 2017; received in revised form 5 July 2017; accepted 17 July 2017

Abstract Background: There is increasing evidence that androgen deprivation therapy (ADT) may be associated with depression. Existing studies have shown conflicting results. Methods: PubMed, Web of Science, Embase, and PsycINFO were queried on April 5, 2017. Eligible studies were in English and reported depression among individuals with prostate cancer exposed to a course of ADT vs. a lesser-exposed group (e.g., any-ADT vs. no ADT and continuous ADT vs. intermittent ADT). We used the MOOSE statement guidelines and the Cochrane Review Group’s data extraction template. Study quality was evaluated by Newcastle-Ottawa Scale criteria. We conducted a random-effects meta-analysis to calculate summary statistic risk ratios (RRs) and 95% CIs. Heterogeneity was quantified using the I2 statistic and prespecified subgroup analysis. Small study effects were evaluated using Begg and Egger statistics. Results: A total of 1,128 studies were initially identified and evaluated. A meta-analysis of 18 studies among 168,756 individuals found that ADT use conferred a 41% increased risk of depression (RR ¼ 1.41; 95% CI: 1.18–1.70; P o 0.001). We found a consistent strong statistically significant association when limiting our analysis to studies in localized disease (RR ¼ 1.85; 95% CI: 1.20–2.85; P ¼ 0.005) and those using a clinical diagnosis of depression (RR ¼ 1.19; 95% CI: 1.08–1.32; P ¼ 0.001). We did not find an association for continuous ADT with depression risk compared to intermittent ADT (RR ¼ 1.00; 95% CI: 0.50–1.99; P ¼ 0.992). There was no statistically significant evidence of small study effects. Statistically significant heterogeneity in the full analysis (I2 ¼ 80%; 95% CI: 69–87; P o 0.001) resolved when examining studies using a clinical diagnosis of depression (I2 ¼ 16%; 95% CI: 0–60; P ¼ 0.310). Conclusion: The currently available evidence suggests that ADT in the treatment of prostate cancer is associated with an increased risk of depression. r 2017 Elsevier Inc. All rights reserved.

Keywords: Androgen deprivation therapy; Depression; Prostate cancer; Behavioral symptoms; Hormone therapy

1. Introduction Androgen deprivation therapy (ADT) has a demonstrated survival benefit for metastatic and locoregional prostate cancer [1]. An estimated 50% of men with prostate cancer ultimately use ADT [2], and this proportion may continue to grow with recent randomized evidence supporting a survival benefit in the salvage setting [3]. With over a million

Corresponding author. Tel.: þ1-215-662-2428; fax: þ1-215-349-5445. E-mail address: [email protected] (K.T. Nead). *

http://dx.doi.org/10.1016/j.urolonc.2017.07.016 1078-1439/r 2017 Elsevier Inc. All rights reserved.

new diagnoses of prostate cancer each year [4], the implications of adverse effects of ADT are substantial. In contrast to the demonstrated survival benefit, ADT has been associated with numerous adverse effects including cardiometabolic and neurocognitive dysfunction [5,6]. A recent large population-based study demonstrated a positive association between ADT and depression risk [7]. However, although some studies support this association [8–11], other studies do not [12–16]. Currently, the true association between ADT and depression is unclear. In this study, we undertake a systematic review and meta-analysis to examine the association of ADT in the treatment of prostate cancer and depression risk.

2

K.T. Nead et al. / Urologic Oncology: Seminars and Original Investigations ] (2017) ∎∎∎–∎∎∎

2. Methods

2.3. Statistical analysis

The MOOSE statement guidelines were used for this systematic review and meta-analysis [17].

We conducted a meta-analysis to calculate summary statistic RRs and 95% CIs for the risk of depression in patients with prostate cancer exposed to a course of ADT vs. a lesser-exposed comparison group (e.g., any-ADT vs. no ADT and continuous ADT vs. intermittent ADT). We included comparison groups exposed to a lesser amount of ADT given evidence for a dose effect of ADT on depression risk [7]. In our primary meta-analysis, we included all eligible studies reporting a binary depression outcome. Where eligible studies or subgroups were clearly overlapping, we included the study or subgroup with the largest number of events, or the largest sample size if the number of events was not available in subgroup. Additionally, if depression was evaluated at multiple time points, the effect estimate from the longest on ADT time point was used. Ratios of rates, odds ratios, and RRs were considered equivalent measures of risk [19,20]. Studies with zero events in a comparison arm were included per the Cochrane Handbook guidelines [22]. The proportion of heterogeneity due to study variation was quantified using the I2 statistic [21]. Random-effects meta-analytic models were selected a priori as a more conservative approach given expected variability in methods used to determine depression status and diversity of study populations. Heterogeneity was explored per the Cochrane Handbook for Systematic Reviews of Interventions [22]. We present summary data for prespecified subgroups including by study design (prospective, retrospective, and cross-sectional), whether confounding was accounted for in the analysis, by method used to determine depression (e.g., clinical diagnosis such as International Classification of Diseases-9 code vs. depression inventory), and whether studies were limited to localized disease. We did not undertake statistical comparisons within or across subgroups given the limitations of this approach in metaanalyses [22]. The presence of small study effects was evaluated by visualization of a funnel plot and calculating Begg and Egger statistics. Tests were considered significant if the two-sided P value was less than 0.05. All analyses were carried out using Stata version 12 (StataCorp, College Station, TX).

2.1. Eligibility criteria We included full-text articles in English that reported the outcome of depression (e.g., billing codes and depression inventory) among individuals with prostate cancer exposed to ADT vs. a lesser-exposed comparison group (e.g., any-ADT vs. no ADT and continuous ADT vs. intermittent ADT). Inclusion criteria for the quantitative metaanalysis were studies that reported an effect estimate (e.g., risk ratio [RR]) and measure of error (e.g., CI) or reported measurement of or adjustment for depression that could be used to calculate an effect estimate and measure of error. 2.2. Information sources We carried out electronic searches in PubMed (1966present), Web of Science (1945-present), Embase (1966present), and PsycINFO (1806-present). The search was undertaken on April 5, 2017. Each database was searched for prostate cancer and ADT and depression terms. Detailed search strings for each database can be found in Supplementary Fig. 1. We additionally queried the reference lists of included articles. Two investigators (S.S. and D.Y.) independently assessed the eligibility of each study by using the title and abstract for initial screening followed by review of the full text and data extraction with consensus reached by discussion with a third investigator (K.N.) as needed. We used a data extraction sheet developed based on the Cochrane Consumers and Communication Review Group’s data extraction template (http://cccrg.cochrane. org/author-resources). Searchers extracted the following items from each study: first author, type of article, study location, year of publication, dates of data collection or enrolment, study design (e.g., prospective or retrospective), local vs. metastatic disease, sample size, number of individuals on ADT, details of ADT use, how the outcome was delineated (e.g., International Classification of Diseases-9 codes and depression inventories), number of events in each group, type of effect statistic, measure of error, methodology to account for confounding, length of followup, whether comparison groups were derived from the same population, and whether prevalent depression was accounted for in the analysis. We assessed the internal validity of each study included in the qualitative review based on modified Newcastle-Ottawa Scale criteria [18]. We attempted to contact corresponding authors of all articles at least twice to provide missing descriptive details and summary data for inclusion in the quantitative metaanalysis.

3. Results Our study selection process is summarized in Fig. 1. In total, we reviewed 795 studies by title and abstract after duplicate removal with 52 studies undergoing full-text review. We excluded 26 studies after full-text review because they did not examine individuals with prostate cancer exposed to ADT vs. a lesser-exposed group (n ¼ 22) or did not examine depression as an outcome (n ¼ 4). Twenty-six studies meeting our review inclusion criteria

K.T. Nead et al. / Urologic Oncology: Seminars and Original Investigations ] (2017) ∎∎∎–∎∎∎

Fig. 1. Flow diagram of literature search and study selection for the systematic review and meta-analysis. ADT, androgen deprivation therapy.

were ultimately included and are summarized in Table 1. Studies were published between 2006 and 2017 and included 12 prospective studies (46%). Six studies (23%) relied on a clinical diagnosis of depression (e.g., billing codes or direct provider evaluation) and 12 (46%) were limited to localized disease. We assessed study quality using modified NewcastleOttawa Scale criteria as shown in Table 2. Six studies (23%) stated that they excluded prevalent depression at treatment initiation. Eleven studies (42%) used methods to account for confounding for the reported or calculated summary data used in this systematic review and metaanalysis. Our primary meta-analysis (Fig. 2) included 18 eligible studies with 77,017 (46%) ADT users among 168,756 total individuals. Using a random-effects model, we found that ADT use conferred a 41% increased risk of depression (RR ¼ 1.41; 95% CI: 1.18–1.70; P o 0.001). A funnel plot to evaluate small study effects (Supplementary Fig. 2) demonstrated possible asymmetry secondary to an over-representation of inventory-based studies with a positive direction of association. However, there was no evidence of statistically significant small study effects as evaluated by the Egger (P ¼ 0.07) or Begg (P ¼ 0.97) tests. Summary data in prespecified subgroups showed continued statistically significant results regardless of study design, whether studies accounted for confounding factors, the method used to determine binary depression status, and whether studies were limited to localized disease (Table 3).

3

We additionally stratified by whether studies compared any ADT vs. no ADT or continuous ADT vs. intermittent ADT. We found no association for continuous ADT with depression risk compared to intermittent ADT (Table 3). In sensitivity analyses, we excluded Donovan et al. [27] as the relevant comparison from this study examined radiation therapy with ADT vs. radical prostatectomy alone and differences in depression outcomes may be secondary to no ADT-related treatment difference. Exclusion of this study yielded a similar result (RR 1.46; 95% CI: 1.20–1.77; P o 0.001). We observed statistically significant heterogeneity in the primary analysis (I2 ¼ 80%; 95% CI: 69–87; P o 0.001). We explored heterogeneity in prespecified subgroups (Table 3). This demonstrated that when stratifying by the method used to determine depression status, the heterogeneity resolved in the subgroup of those diagnosed with depression in a clinical setting (I2 ¼ 16%; 95% CI: 0–60; P ¼ 0.310), but persisted in those using inventory-based metrics (I2 ¼ 85%; 95% CI: 74–91; P o 0.001). We further explored heterogeneity in the subgroup examining depression using inventory-based metrics by restricting our evaluation of heterogeneity to studies using the Hospital Anxiety and Depression Scale (HADS) inventory (n ¼ 6), which was the most commonly used inventory. However, even within this subgroup, heterogeneity persisted (RR ¼ 1.66; 95% CI: 1.01–2.73; P ¼ 0.04; I2 ¼ 78%; 95% CI: 52–90; P o 0.001). We additionally did not find statistically significant heterogeneity among studies comparing continuous vs. intermittent ADT. However, this may be secondary to a limited number of studies within this subgroup. Heterogeneity persisted in all other subgroups examined. Two studies were eligible for inclusion in the primary qualitative meta-analysis, but were excluded as they overlapped with a larger included study. Shahinian et al. [10], which overlapped with Dinh et al. [7], examined data in 50,613 men and found a statistically significantly increased risk of depression among ADT users following multivariable adjusted analysis (RR ¼ 1.08; 95% CI: 1.02–1.15). Chung et al. [26], which overlapped with Chen et al. [12], analyzed data in 1,714 individuals and found a statistically significantly increased risk of depression among ADT users in multivariable adjusted analysis (hazard ratio ¼ 1.93; 95% CI: 1.03–3.62). Eight studies were not eligible for inclusion in the quantitative meta-analysis. Sharpley et al.’s [11] study was excluded as it did not include a reported or calculable measure of error, but notably found a statistically significantly increased risk of depression among ADT users in multivariable adjusted analysis (β ¼ 0.723; P o 0.001). Five studies reported depression inventory–based score comparisons in ADT and no ADT groups [15,16,23–25], but did not report a binary depression outcome. HADS scores were similar in ADT and no ADT groups, respectively, in studies by Kerleau et al. [15] (HADS 5.5 vs.

4

Table 1 Summary of studies included in systematic review Location

Year

Dates of data collection/enrolment

Study design

Sample sizea

# on ADTa

Exposed definition

Control definition

Outcome delineation

Disease extent

Bill-Axelson [45] Chenc [12] Chipperfield [13] Chung [26] Couper [23] Dinh [7] Donovan [27] Hershman [14] Hervouet [46] Johansson [47] Karunasinghe [48] Kerleau [15] Langenhuijsen [42] Leed [49] Muresanu [43] Nelson [8] Saini [9] Salonen [40] Shahinian [10] Sharpley [11] Sharpley [25] Timilshina [50] Van Dam [24] van Tol-Geerdinkc [51] Watson [52] Wiechno [16]

Sweden Taiwan Australia Taiwan Australia United States UK United States Canada Sweden New Zealand France Europe United States Romania United States Italy Finland United States Australia Australia Canada Any Netherlands UK Poland

2011 2015 2013 2017 2009 2016 2016 2016 2013 2009 2016 2016 2013 2015 2016 2009 2013 2013 2006 2014 2014 2012 2016 2011 2016 2013

1997–2006 2000–2006b 2010–2011 2001–2010 2001–2005 1992–2006 1999–2009 1995–2008 2004–2006 1989–1996 2004–2014b 2011 1998–2001 2008–2012 2004–2014 Unknown 2007 1997–2003 1992–1997b Unknown Unknown 2004–2007 2011–2012 2002–2005 2011–2013 2005–2006

R R R R P R P P P P R CS P P P CS P P R CS CS P CS P R P

72,613 12,872 377 1,714 172 78,552 965 636 53 321 206 254 193 122 82 367 103 554 50,613 156 409 138 295 287 315 149

30,682 10,235b 203 868 51 33,882 478 311 23 45 75 34 96 61 42 174 49 280 15,748 93 117 66 82 198 117 88

Primary ADT AAc RT þ ADT ADT ADT RT/RP þ ADT RT þ 3–6 mo neoadjuvant ADT Continuous GnRHa þ AA RT þ GnRHa& 4 wk AA ADT or orchiectomy ADT or orchiectomy RT/RP þ ADT Continuous LHRHa þ AA GnRHa ≥ 6 mo Continuous ADT ADT RT/RP þ LHRHa& 1 mo AA Continuous GnRHa& 12.5 d AA ≥1 dose GnRHa or orchiectomy ADT ADT or orchiectomy ADT ≥ 1 year ADT RT þ ADT RT þ ADT or primary ADT RT þ LHRHa

Observation or curative treatment No ADT RT No ADT WW, RP, other No ADT RP Intermittent GnRHa þ AA RT No ADT No ADT No current ADT Intermittent LHRHa þ AA RP Intermittent ADT No ADT RP Intermittent GnRHa& 12.5d AA No ADT or orchiectomy No current ADT No current ADT No ADT No ADT RT No ADT RT

Billing codes Billing codes HADS ≥ 8 Billing codes BSI score Billing codes HADS ≥ 8 Billing codes Clinical diagnosis CESMD score DASS score HADS score EORTC QLQ-C30-version 2.0 CES-D score CTCAE events HADS ≥ 7 HADS ≥ 8 Clinical evaluation Billing codes (ICD-9) PHQ9 score Zung SDS GDS score ≥ 5 POMS score HADS ≥ 8 HADS ≥ 8 HADS ≥ 8

Local and metastatic Stage unknown Stage unknown Stage unknown Local Local Local Metastatic Local Local and metastatic Stage unknown Local Local and metastatic Local Local Local and metastatic Local Local and metastatic Local and metastatic Local and regional nodes Local and regional nodes Local Local Local Local and metastatic Local

AA ¼ antiandrogen; CES-D ¼ Center for Epidemiological Studies Depression Scale; CESMD ¼ Centre for Epidemiological Studies Measure of Depression was used to assess depression; CR ¼ crosssectional; CTCAE ¼ Common Terminology Criteria of Adverse Events; DASS ¼ Depression Anxiety Stress Scales; GDS ¼ Geriatric Depression Scale; GnRHa ¼ gonadotropin releasing hormone agonist; LHRHa ¼ luteinizing hormone-releasing hormone analogues; POMS ¼ Profile of Mood States; P ¼ prospective; RP ¼ radical prostatectomy; RT ¼ radiation therapy; R ¼ retrospective; SDS ¼ Self-Rating Depression Scale; SHRQLQ ¼ Summary of Health-Related Quality of Life Questionnaire. a Where relevant values correspond to time points or subgroups or both used in analysis. b Obtained via author correspondence. C Effect estimates taken from AA analysis, but some individuals received other forms of ADT in addition to AAs and in the reference group. d Effect statistics calculated through data obtained via author correspondence.

K.T. Nead et al. / Urologic Oncology: Seminars and Original Investigations ] (2017) ∎∎∎–∎∎∎

Study

K.T. Nead et al. / Urologic Oncology: Seminars and Original Investigations ] (2017) ∎∎∎–∎∎∎

5

Table 2 Internal validity of studies included in the systematic review using adapted Newcastle-Ottawa Scale Study

Defined source for Exposed and nonexposed groups from ascertainment of exposure same population

Stated, objective ascertainment of outcome

Outcome of interest not present at treatment initiation

Statistical methods to account for confounding

Bill-Axelson [45] Chen [12] Chipperfield [13] Chung [26] Couper [23] Dinh [7] Donovan [27] Hershman [14] Hervouet [46] Johansson [47] Karunasinghe [48] Kerleau [15] Langenhuijsen [42] Lee [49] Muresanu [43] Nelson [8] Saini [9] Salonen [40] Shahinian [10] Sharpley [11] Sharpley [25] Timilshina [50] Van Dam [24]

Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes

Yes Yes No Yes No Yes Yes Yes Yes Yes Yes Yes Yes Yes No Yes Yes No Yes Yes Yes Yes Yes

Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes

No Yes No Yes No Yes No No No No No NA (cross-sectional) No No No NA (cross-sectional) Yes No Yes NA (cross-sectional) NA (cross-sectional) Yes NA (cross-sectional)

van Tol-Geerdink [51] Watson [52] Wiechno [16]

Yes

Yes

Yes

No

None None Yes (multivariable regression) Yes (multivariable regression) None Yes (multivariable regression) Yes (randomization) Yes (multivariable regression) None None Yes (multivariable regression) None None None None Yes (multivariable regression) Yes (multivariable regression) None Yes (multivariable regression) None Yes (multivariable regression) None Yes (multivariate analysis of variance) None

Yes Yes

Yes Yes

Yes Yes

No No

None None

All quality metrics–based analyses/data are used in the current analysis. NA ¼ not applicable.

Fig. 2. Forest plot with summary risk ratio for the risk of depression among individuals on androgen deprivation therapy for the treatment of prostate cancer. RR, risk ratio.

K.T. Nead et al. / Urologic Oncology: Seminars and Original Investigations ] (2017) ∎∎∎–∎∎∎

6

Table 3 Risk of depression within subgroups

Study design Prospective Retrospective Cross-sectional Account for confoundingb Yes No Method to determine depression Clinical Inventory/scale Limited to localized disease Yes No ADT use comparison Any vs. none Continuous vs. intermittent

P value

I2 (95% CI)

P value

# of studies

RR (95% CI)

6 11 1

1.45 (1.13–1.85) 1.63 (1.10–2.41) 0.91 (0.84–1.00)

0.003 0.014 0.050

75% (44–89) 64% (32–81) naa

0.001 0.002 .

7 11

1.34 (1.02–1.75) 1.51 (1.17–1.96)

0.036 0.002

89% (81–94) 51% (2–75)

o0.001 0.026

7 11

1.19 (1.08–1.32) 2.16 (1.42–3.29)

0.001 o0.001

16% (0–60) 85% (74–91)

0.310 o0.001

8 10

1.85 (1.20–2.85) 1.36 (1.04–1.77)

0.005 0.024

66% (29–84) 81% (66–89)

0.004 o0.001

14 4

1.49 (1.22–1.81) 1.00 (0.50–1.99)

o0.001 0.992

83% (73–90) 39% (0–79)

o0.001 0.179

na ¼ not applicable. a Cannot calculate heterogeneity for subgroup with a single study. b As analyzed in the current study.

HADS 5.2; P ¼ 0.658) and Wiechno et al. [16] (HADS 5.93 vs. HADS 5.94; P ¼ 0.954). Couper et al. [23] compared Brief Symptom Inventory (BSI) depression subscale scores and found worse depression scores in the ADT group compared to definitive treatment and watchful waiting groups (BSI 0.355 vs. BSI 0.205; P o 0.05). Similarly, Sharpley et al. [25] compared Patient Health Questionnaire9 (PHQ9) scores and found worse depression scores in those currently on ADT compared to those not currently on ADT (PHQ9 14.53 vs. PHQ9 12.48; P o 0.05). Finally, Van Dam et al. [24] compared the short-form Profile of Mood States scores among those on ADT vs. not on ADT and found no difference between groups (P ¼ 0.31).

4. Discussion In this systematic review and meta-analysis of 18 studies including 168,756 individuals, we found a large, statistically significantly increased risk of depression in patients with prostate cancer exposed to ADT. This result was consistent when examining studies limited to individuals with localized disease. However, we did not find evidence that continuous ADT carried an increased risk of depression compared to intermittent ADT. This comprehensive analysis of published studies strongly supports the association of ADT in the treatment of prostate cancer with depression. Depression rates are high among cancer patients [28], and in this study, we demonstrate a clinically meaningful further increase in depression rates with the addition of ADT. This is of particular importance as depression is a significant predictor of mortality for patients with prostate cancer [29,30]. Although the causality of this association remains to be fully established, interventions to reduce depression have been shown to predict longer survival

among patients with cancer [31] and decrease the risk of cancer mortality among older individuals [32]. Although some approaches to mitigate depression in prostate cancer have been identified [33], further research is needed to determine effective interventions to prevent, identify, and treat depression, particularly among those undergoing ADT. One explanation for the observed association is that ADT is causally associated with depression. Low testosterone levels have been associated with increased rates of depression in healthy men [34,35]. Although an exact mechanism for this association is unclear, and the relationship between hormone levels and mood is likely complex, low testosterone levels may increase depression risk through alterations in central serotonin function and serotonin receptor density [34]. ADT may also indirectly affect depression risk through a global decrease in quality of life [36] given the demonstrated associations of ADT with decreased cognitive function, incident neurocognitive and cardiometabolic disease, insomnia, fatigue, gynecomastia, erectile dysfunction, and libido loss [5,6,37,38]. An alternative explanation for these findings is that there is no causal association between ADT and depression, and our results are secondary to bias. It is possible that individual studies including both local and metastatic disease were confounded by disease stage such that those on ADT had more metastatic disease and became more depressed for that reason. However, analysis of studies limited to localized disease showed a strong, positive association. Additionally, most studies were retrospective in nature, and not all were adjusted for confounding factors. Subgroup analysis of only prospective studies and only those adjusting for confounding factors continued to show a statistically significant association. Furthermore, a diverse array of methods was used to determine depression status including billings codes and various depression inventories.

K.T. Nead et al. / Urologic Oncology: Seminars and Original Investigations ] (2017) ∎∎∎–∎∎∎

As depression inventories do not represent a clinical diagnosis of depression, and include a variety of metrics and criteria, this could introduce an unmeasured bias. Visualization of the funnel plot demonstrated asymmetry secondary to inventory-based studies with a positive direction of association. This could be secondary to publication bias or other small study effects such as methodological quality or patient selection. However, even when limiting our analysis to only clinical diagnoses of depression, we continued to observe a statistically significant association of ADT and depression. Regardless, depression inventories deserve further study in this setting as they may have an important role as cost-effective tools to detect early depressive symptomology and guide patient care, particularly given the underdiagnosis of depression [39]. In subgroup analysis, we did not find evidence that continuous ADT carried an increased risk of depression compared to intermittent ADT. This is consistent with previous studies showing that intermittent ADT may not decrease adverse events [14]. Although one of the included studies did support an improvement in some quality of life metrics with intermittent ADT use, the risk of depression in that study was statistically significantly increased among intermittent ADT users compared to continuous ADT users [40]. Though low testosterone levels are associated with depression [33,34], testosterone recovery following ADT may be too slow to decrease depression risk before ADT reinitiation [41]. Therefore, based on existing studies [14,40,42,43], intermittent ADT may not be an effective strategy to alleviate depressive symptomology among ADT users. Our primary meta-analysis showed statistically significant between-study heterogeneity. Details of individual studies are summarized in Tables 1 and 2. We explored heterogeneity through prespecified subgroup analysis and found that heterogeneity resolved when limiting our analysis to those studies using a clinical diagnosis of depression (e.g., billing codes, interview-based, and Diagnostic and Statistical Manual of Mental Disorders-based diagnosis). Heterogeneity persisted in the subgroup using inventorybased metrics even when limiting the analysis to only those studies using the HADS inventory. These results are not surprising, as even within studies that used the HADS inventories, differences in cutoff scores to indicate depression existed. Furthermore, studies have shown disagreement among different depression inventories when compared to clinical diagnosis of depression [44]. Importantly, the analysis limited to the most relevant clinical outcome, a clinical diagnosis of depression, showed both a statistically significant association of ADT with depression and no statistically significant heterogeneity even when using the more conservative random-effects model. This study has several limitations that warrant consideration. First, most studies using a clinical diagnosis of depression were based on billing codes. This relies on physician recognition, which may lead to underdiagnosis [39]. Second,

7

we were unable to evaluate the association of different types of ADT with depression risk given the high variability of regimens used among included studies. Given that the various forms of ADT have differing effects on the hypothalamic-pituitary-gonadal axis, and the exact mechanisms for the association of ADT and depression remain unclear, this will be an important focus of future studies. Third, we were unable to specifically analyze the association of ADT and depression within specific risk groups, which may provide clinically useful information. Fourth, cultural and practicebased differences in reporting depression may affect depression rates, particularly in retrospective studies, and those relying on claims-based data. Finally, we cannot exclude the role of confounding in our analysis. However, as detailed above, subgroup analyses of prospective studies, including randomized trials, and those adjusting for confounding factors, continued to show a statistically significant association. In conclusion, we observed a statistically significantly increased risk of depression among patients with prostate cancer receiving ADT in a large meta-analysis that was robust to conservative methods and multiple subgroup and sensitivity analyses. Health care providers should be alert for and have a low threshold to intervene on depressive symptomology among patients with prostate cancer. The risk of depression secondary to ADT use should be discussed with patients and caretakers before initiating treatment. Conflicts of interest Paul Nguyen has consulted for Ferring, Medivation, Genome Dx, Augmenix, Bayer, and Dendreon and receives research funding from Astellas and Janssen. Author contributions Conceptualization and design, K.T.N., P.L.N.; Data collection, K.T.N., S.S., D.D.Y.; Drafting manuscript, K. T.N.; Reviewing and editing, K.T.N., S.S., D.D.Y., P.L.N.; Final approval: K.T.N., S.S., D.D.Y., P.L.N.; and Supervision, K.T.N., P.L.N. Appendix A. Supporting information Supplementary data associated with this article can be found in the online version at http://dx.doi.org/10.1016/j. urolonc.2017.07.016.

References [1] Bolla M, Collette L, Blank L, et al. Long-term results with immediate androgen suppression and external irradiation in patients with locally

8

[2]

[3]

[4]

[5]

[6]

[7]

[8]

[9]

[10]

[11]

[12]

[13]

[14]

[15]

[16]

[17]

[18]

[19]

[20] [21]

K.T. Nead et al. / Urologic Oncology: Seminars and Original Investigations ] (2017) ∎∎∎–∎∎∎ advanced prostate cancer (an EORTC study): a phase III randomised trial. Lancet 2002;360:103–6. Meng MV, Grossfeld GD, Sadetsky N, Mehta SS, Lubeck DP, Carroll PR. Contemporary patterns of androgen deprivation therapy use for newly diagnosed prostate cancer. Urology 2002;60:7–11:[discussion 11-12]. Shipley WU, Seiferheld W, Lukka HR, et al. Radiation with or without antiandrogen therapy in recurrent prostate cancer. N Engl J Med 2017;376:417–28. Ferlay J, Soerjomataram I, Dikshit R, et al. Cancer incidence and mortality worldwide: sources, methods and major patterns in GLOBOCAN 2012. Int J Cancer 2015;136:E359–86. Keating NL, O’Malley AJ, Smith MR. Diabetes and cardiovascular disease during androgen deprivation therapy for prostate cancer. J Clin Oncol 2006;24:4448–56. Nead KT, Gaskin G, Chester C, Swisher-McClure S, Leeper NJ, Shah NH. Association between androgen deprivation therapy and risk of dementia. J Am Med Assoc Oncol 2017;3:49–55. Dinh KT, Reznor G, Muralidhar V, et al. Association of androgen deprivation therapy with depression in localized prostate cancer. J Clin Oncol 2016;34:1905–12. Nelson C, Jacobson CM, Weinberger MI, et al. The role of spirituality in the relationship between religiosity and depression in prostate cancer patients. Ann Behav Med 2009;38:105–14. Saini A, Berruti A, Cracco C, et al. Psychological distress in men with prostate cancer receiving adjuvant androgen-deprivation therapy. Urol Oncol 2013;31:352–8. Shahinian VB, Kuo YF, Freeman JL, Goodwin JS. Risk of the “androgen deprivation syndrome” in men receiving androgen deprivation for prostate cancer. Arch Intern Med 2006;166: 465–71. Sharpley CF, Christie DR, Bitsika V. Do hormone treatments for prostate cancer cause anxiety and depression? Int J Clin Oncol 2014;19:523–30. Chen PM, Chen SC, Liu CJ, et al. The association between prostate cancer and mood disorders: a nationwide population-based study in Taiwan. Int Psychogeriatr. 2015;27:481–90. Chipperfield K, Fletcher J, Millar J, et al. Predictors of depression, anxiety and quality of life in patients with prostate cancer receiving androgen deprivation therapy. Psychooncology 2013;22:2169–76. Hershman DL, Unger JM, Wright JD, et al. Adverse health events following intermittent and continuous androgen deprivation in patients with metastatic prostate cancer. J Am Med Assoc Oncol 2016;2:453–61. Kerleau C, Guizard AV, Daubisse-Marliac L, et al. Long-term quality of life among localised prostate cancer survivors: QALIPRO population-based study. Eur J Cancer 2016;63:143–53. Wiechno PJ, Sadowska M, Kalinowski T, Michalski W, Demkow T. Does pharmacological castration as adjuvant therapy for prostate cancer after radiotherapy affect anxiety and depression levels, cognitive functions and quality of life? Psychooncology 2013;22: 346–51. Stroup DF, Berlin JA, Morton SC, et al. Meta-analysis of observational studies in epidemiology: a proposal for reporting Meta-analysis Of Observational Studies in Epidemiology (MOOSE) group. J Am Med Assoc 2000;283:2008–12. Stang A. Critical evaluation of the Newcastle-Ottawa scale for the assessment of the quality of nonrandomized studies in meta-analyses. Eur J Epidemiol 2010;25:603–5. Roerecke M, Rehm J. Cause-specific mortality risk in alcohol use disorder treatment patients: a systematic review and meta-analysis. Int J Epidemiol 2014;43:906–19. Rothman KJ, Greenland S, Lash TL. Modern epidemiology. Philadelphia, PA: Lippincott Williams & Wilkins, 2008;2008. Higgins JP, Thompson SG, Deeks JJ, Altman DG. Measuring inconsistency in meta-analyses. Br Med J 2003;327:557–60.

[22] Higgins J, Green S. Cochrane handbook for systematic reviews of interventions version 5.1.0. The Cochrane Collaboration, London, UK 2011. [23] Couper JW, Love AW, Dunai JV, et al. The psychological aftermath of prostate cancer treatment choices: a comparison of depression, anxiety and quality of life outcomes over the 12 months following diagnosis. Med J Aust 2009;190:S86–9. [24] Van Dam D, Wassersug RJ, Hamilton LD. Androgen deprivation therapy’s impact on the mood of prostate cancer patients as perceived by patients and the partners of patients. Psychooncology 2016;25: 848–56. [25] Sharpley CF, Bitsika V, Wootten AC, Christie DR. Differences in major depressive disorder and generalised anxiety disorder symptomatology between prostate cancer patients receiving hormone therapy and those who are not. Psychooncology 2014;23:1350–5. [26] Chung SD, Kao LT, Lin HC, Xirasagar S, Huang CC, Lee HC. Patients receiving androgen deprivation therapy for prostate cancer have an increased risk of depressive disorder. PLos One 2017;12: e0173266. [27] Donovan JL, Hamdy FC, Lane JA, et al. Patient-reported outcomes after monitoring, surgery, or radiotherapy for prostate cancer. N Engl J Med 2016;375:1425–37. [28] Pirl WF. Evidence report on the occurrence, assessment, and treatment of depression in cancer patients. J Natl Cancer Inst Monogr 2004:32–9. [29] Satin JR, Linden W, Phillips MJ. Depression as a predictor of disease progression and mortality in cancer patients: a meta-analysis. Cancer 2009;115:5349–61. [30] Jayadevappa R, Malkowicz SB, Chhatre S, Johnson JC, Gallo JJ. The burden of depression in prostate cancer. Psychooncology 2012;21: 1338–45. [31] Giese-Davis J, Collie K, Rancourt KM, Ner E, Kraemer HC, Spiegel D. Decrease in depression symptoms is associated with longer survival in patients with metastatic breast cancer: a secondary analysis. J Clin Oncol 2011;29:413–20. [32] Gallo JJ, Bogner HR, Morales KH, Post EP, Lin JY, Bruce ML. The effect of a primary care practice-based depression intervention on mortality in older adults: a randomized trial. Ann Intern Med 2007;146:689–98. [33] Chipperfield K, Brooker J, Fletcher J, Burney S. The impact of physical activity on psychosocial outcomes in men receiving androgen deprivation therapy for prostate cancer: a systematic review. Health Psychol 2014;33:1288–97. [34] Shores MM, Moceri VM, Sloan KL, Matsumoto AM, Kivlahan DR. Low testosterone levels predict incident depressive illness in older men: effects of age and medical morbidity. J Clin Psychiatry 2005;66:7–14. [35] Amore M, Innamorati M, Costi S, Sher L, Girardi P, Pompili M. Partial androgen deficiency, depression, and testosterone supplementation in aging men. Int J Endocrinol 2012;2012:280724. [36] Dacal K, Sereika SM, Greenspan SL. Quality of life in prostate cancer patients taking androgen deprivation therapy. J Am Geriatr Soc 2006;54:85–90. [37] Gonzalez BD, Jim HS, Booth-Jones M, et al. Course and predictors of cognitive function in patients with prostate cancer receiving androgen-deprivation therapy: a controlled comparison. J Clin Oncol 2015;33:2021–7. [38] Nguyen PL, Alibhai SM, Basaria S, et al. Adverse effects of androgen deprivation therapy and strategies to mitigate them. Eur Urol 2015;67:825–36. [39] Block SD. Assessing and managing depression in the terminally ill patient. ACP-ASIM End-of-Life Care Consensus Panel. American College of Physicians—American Society of Internal Medicine. Ann Intern Med 132; 209–18. [40] Salonen AJ, Taari K, Ala-Opas M, Viitanen J, Lundstedt S, Tammela TL. Advanced prostate cancer treated with intermittent or continuous androgen deprivation in the randomised FinnProstate

K.T. Nead et al. / Urologic Oncology: Seminars and Original Investigations ] (2017) ∎∎∎–∎∎∎

[41]

[42]

[43]

[44]

[45]

[46]

Study VII: quality of life and adverse effects. Eur Urol 2013;63: 111–20. Spry NA, Kristjanson L, Hooton B, et al. Adverse effects to quality of life arising from treatment can recover with intermittent androgen suppression in men with prostate cancer. Eur J Cancer 2006;42: 1083–92. Langenhuijsen JF, Badhauser D, Schaaf B, Kiemeney LALM, Witjes JA, Mulders PFA. Continuous vs. intermittent androgen deprivation therapy for metastatic prostate cancer. Urol Oncol 2013;31:549–56. Muresanu H. Benefits of intermittent/continuous androgen deprivation in patients with advanced prostate cancer. Clujul Med 2016;89: 419–22. Golden J, Conroy RM, O’Dwyer AM. Reliability and validity of the Hospital Anxiety and Depression Scale and the Beck Depression Inventory (Full and FastScreen scales) in detecting depression in persons with hepatitis C. J Affect Disord 2007;100:265–9. Bill-Axelson A, Garmo H, Nyberg U, et al. Psychiatric treatment in men with prostate cancer—results from a nation-wide, populationbased cohort study from PCBaSe Sweden. Eur J Cancer 2011;47: 2195–201. Hervouet S, Savard J, Ivers H, Savard MH. Depression and androgen deprivation therapy for prostate cancer: a prospective controlled study. Health Psychol 2013;32:675–84.

9

[47] Johansson E, Bill-Axelson A, Holmberg L, Onelov E, Johansson JE, Steineck G. Time, symptom burden, androgen deprivation, and selfassessed quality of life after radical prostatectomy or watchful waiting: the Randomized Scandinavian Prostate Cancer Group Study Number 4 (SPCG-4) clinical trial. Eur Urol 2009;55:422–30. [48] Karunasinghe N, Zhu Y, Han DY, et al. Quality of life effects of androgen deprivation therapy in a prostate cancer cohort in New Zealand: can we minimize effects using a stratification based on the aldo-keto reductase family 1, member C3 rs12529 gene polymorphism? BMC Urol 2016;16:48. [49] Lee M, Jim HS, Fishman M, et al. Depressive symptomatology in men receiving androgen deprivation therapy for prostate cancer: a controlled comparison. Psychooncology 2015;24:472–7. [50] Timilshina N, Breunis H, Alibhai S. Impact of androgen deprivation therapy on depressive symptoms in men with nonmetastatic prostate cancer. Cancer 2012;118:1940–5. [51] van Tol-Geerdink JJ, Leer JW, van Lin EN, Schimmel EC, Stalmeier PF. Depression related to (neo)adjuvant hormonal therapy for prostate cancer. Radiother Oncol 2011;98:203–6. [52] Watson E, Shinkins B, Frith E, et al. Symptoms, unmet needs, psychological well-being and health status in survivors of prostate cancer: implications for redesigning follow-up. BJU Int 2016;117: E10–9.