Anxiety and Response to Treatment of Depression in People Undergoing Maintenance Hemodialysis

Introduction

Generalized anxiety disorder (GAD) is a prevalent anxiety disorder (2.7% point prevalence in the United States) noted for its hallmark symptom, excessive worry. GAD is more common in individuals undergoing maintenance hemodialysis (HD) and has been associated with worse health outcomes.¹ However, the importance of co-occurring anxiety in HD patients with major depressive disorder remains understudied.²

The ASCEND (A Trial of Sertraline vs. Cognitive Behavioral Therapy for End-stage Renal Disease Patients with Depression) trial^3,4 compared the efficacy of cognitive behavioral therapy (CBT) with sertraline drug therapy for the treatment of depression (major depressive disorder or dysthymia) in patients undergoing maintenance HD. The study found that depressive symptoms and other patient-reported outcomes improved with both treatments. The outcome scores at 12 weeks were modestly better with sertraline but with higher incidence of mild and moderate severity adverse events. The original study also reported substantive reduction in anxiety as measured by GAD-7⁵ with both treatments, with no significant differences between the two groups (effect size estimate −1.2, 95% confidence interval [CI], −3.1 to 0.8). Prespecified analyses to determine whether the baseline severity of anxiety symptoms modified the response to treatment of depression among participants enrolled in the ASCEND study are presented.

Methods

The ASCEND trial randomized 120 participants with depression undergoing maintenance HD to CBT or sertraline drug therapy.^3,4 This analysis uses data from 116 participants who completed the GAD-7 questionnaire at baseline. Participants were grouped into three anxiety categories as determined by GAD-7 score: no or low anxiety (<10), moderate anxiety (10–14), and severe anxiety (15–21). Depressive symptoms, measured by quick inventory of depressive symptomatology–clinician rating (QIDS-C), were ascertained by a blinded assessor at baseline and 6 and 12 weeks.

To compare 12-week effectiveness for treatment of depression by baseline anxiety, a longitudinal model of QIDS-C score was fit with interactions between time and anxiety category and interactions between time, anxiety category, and treatment group. Baseline differences between treatment groups were constrained to be zero. An unstructured covariance was allowed to differ with time. A second longitudinal model was fit with GAD-7 score replacing anxiety category to test for trend. Models were adjusted for clinical site. A significance level of 0.05 was used for testing. Analyses were performed with R version 4.0.2 (the R Foundation for statistical computing).

Results

Treatment groups were of similar size within anxiety category (no/mild anxiety: CBT/sertraline n=19/21; moderate anxiety: CBT/sertraline n=18/15; severe anxiety: CBT/sertraline n=22/21). In total, 76 (66%) participants had moderate or severe anxiety at baseline (GAD-7 ≥10; CBT n=40; sertraline n=36). On average, depression levels at baseline were higher in participants with greater baseline anxiety (mean [SD] no/mild anxiety QIDS-C=8.7 [4], moderate anxiety QIDS-C=11.3 [3.7], severe anxiety QIDS-C=14.7 [5.3]).

Most (77%) of those with severe depression (QIDS-C score >15) at baseline were also severely anxious, as compared with only 19% of those with mild depression. Depression scores remained elevated at 12 weeks among those with severe anxiety at baseline as compared with those with no/mild or moderate anxiety (mean [95% CI]; no/mild anxiety QIDS-C=5.67 [95% CI, 4.17 to 7.18], moderate anxiety QIDS-C=5.61 [95% CI, 4 to 7.21], severe anxiety QIDS-C=8.72 [95% CI, 7.32 to 10.13]).

Looking at the intervention arms individually (Figure 1), the treatment advantage for sertraline over CBT at 12 weeks was the greatest among those with severe anxiety (difference in QIDS-C score=−3.32, 95% CI, −6.12 to −0.52). There were no significant differences at 12 weeks between CBT and sertraline groups for those with moderate anxiety (difference=−1.71, 95% CI, −4.88 to 1.46) and no/mild anxiety (difference=−1.23, 95% CI, −4.17 to 1.71). A test for trend was significant, indicating increased effect for sertraline compared with CBT with higher baseline GAD-7 score (slope=−0.32, P = 0.039).

Figure 1.

Longitudinal QIDS-C scores stratified by baseline anxiety and randomization arm. CBT, cognitive behavioral therapy; GAD, generalized anxiety disorder; QIDS-C, quick inventory of depressive symptomatology–clinician rating.

Discussion

The examination of comorbid anxiety on response to depression treatment in the ASCEND study indicated higher baseline anxiety scores, when compared with lower anxiety scores, were associated with less response to treatment of depression with both CBT and sertraline. For the subsample with severe anxiety, there was a larger improvement in depressive symptoms with sertraline than with CBT. However, owing to the high rate of co-occurring severe depression and anxiety, and the limitations of our sample size, this study cannot determine the extent to which baseline depression severity and/or anxiety comorbidity drives the observed reduced treatment efficacy.

These data are significant in that they argue for the consideration of comorbid anxiety, in addition to the severity of depression, when treating major depressive disorder in patients treated with maintenance dialysis. Furthermore, this may explain some of the treatment advantage observed for sertraline because sertraline is a first-line treatment for anxiety and has been shown to improve GAD.⁶ While CBT is also a first-line treatment for anxiety, effective intervention for depression may require the incorporation of anxiety-specific skills. Future depression intervention trials should consider co-occurring anxiety because it is associated with reduced treatment response in individuals treated with maintenance dialysis.

Disclosures

D. Cukor reports employment with Rogosin Institute, NIH research funding, and royalties from Elsevier–Academic Press. P. Heagerty reports serving as a member of a data safety monitoring board for Verily and serving as a member of Board of Directors for Cancer Research and Biostatistics (CRAB). S.S. Hedayati reports honoraria from American College of Physicians for participation in Nephrology MKSAP and from American Society of Nephrology Postgraduate Education Program and serving in advisory or leadership roles for American College of Physicians MKSAP Nephrology Committee and American Heart Association study sections. R. Mehrotra reports consultancy for Light Line Medical; serving as Chair of the Nephrology Longitudinal Assessment Approval Committee of the American Board of Internal Medicine, President-Elect of the International Society for Peritoneal Dialysis, and the Editor-in-Chief of CJASN; serving on the Editorial Boards for Journal of Renal Nutrition and Peritoneal Dialysis International; and reports other interests or relationships as Chair of the Board of Trustees for Northwest Kidney Centers. All remaining authors have nothing to disclose.

Funding

R. Mehrotra: Patient-Centered Outcomes Research Institute (ID0EKKBG948).

Author Contributions

Conceptualization: Daniel Cukor, Patrick Heagerty, S. Susan Hedayati, Rajnish Mehrotra, Tessa Rue, Mark Unruh.

Funding acquisition: Rajnish Mehrotra.

Formal analysis: Patrick Heagerty, Tessa Rue.

Writing – original draft: Daniel Cukor, Tessa Rue.

Writing – review & editing: Daniel Cukor, Patrick Heagerty, S. Susan Hedayati, Rajnish Mehrotra, Tessa Rue, Mark Unruh.