Psychiatry · Medical Literature
Association between anxious depression and clinical outcomes in subjects with treatment-resistant depression: a comparative effectiveness research trial for antidepressant incomplete and non-responders with treatment-resistant depression (ASCERTAIN-TRD).
Abstract
Anxious depression is a negative prognostic factor linked to worse outcomes in major depressive disorder (MDD). However, its role in treatment-resistant depression (TRD) remains unclear. This secondary analysis of the ASCERTAIN-TRD trial investigates the role of anxious depression as a predictor or moderator of treatment response. A Hamilton Depression Rating Scale Anxiety and Somatization Subscale score of at least 7 identified subjects with the anxious depressive subtype. Treatment response was evaluated based on changes in Montgomery-Asberg Depression Rating Scale scores using mixed-effects models with repeated measures and included additional terms accounting for the presence or absence of anxious depression. Two hundred fifty-three patients, who had at least one post-baseline Montgomery-Asberg Depression Rating Scale and Hamilton Depression Rating Scale score were included in the analysis. Of them, 182 subjects presented anxious and 71 nonanxious depression. Treatment outcomes were similar across the three treatment groups regardless of anxious MDD status ( P -value > 0.5 for all comparisons) suggesting that anxious depression was not a significant predictor or moderator of treatment outcome. These findings indicate that anxious depression alone may be prognostic rather than predictive. These findings may be a useful guidance for clinicians, suggesting that standard TRD treatments remain applicable in the anxious MDD subtype.
Clinical relevance
In patients with treatment-resistant depression (TRD) enrolled in the ASCERTAIN-TRD trial, does the anxious depression subtype predict or moderate differential treatment response across three active interventions (aripiprazole augmentation, rTMS augmentation, or switch to venlafaxine XR/duloxetine)? __SNAPSHOT__ {"studyDesign":"Prespecified secondary analysis of the ASCERTAIN-TRD randomized controlled trial using mixed-effects models with repeated measures (MMRM)","setting":"Multi-site academic outpatient setting; principal investigative center at Massachusetts General Hospital/Harvard Medical School with multiple U.S. co-investigators","sampleSize":"n=253 (anxious depression: n=182; nonanxious depression: n=71); parent trial enrolled n=278","population":"Adults with TRD, defined as inadequate response to ≥2 antidepressant trials of adequate dose and duration (per the MGH Antidepressant Treatment Response Questionnaire); participants required to have ≥1 post-baseline MADRS and HDRS score. Anxious depression subtype defined by Hamilton Depression Rating Scale Anxiety and Somatization Subscale (HDRS-ASS) score ≥7 at baseline.","intervention":"Three active arms from the parent ASCERTAIN-TRD trial: (1) aripiprazole augmentation of current antidepressant, (2) repetitive transcranial magnetic stimulation (rTMS) augmentation, or (3) switch to venlafaxine XR (or duloxetine if venlafaxine was contraindicated)","comparator":"Within-analysis comparator: nonanxious depression subgroup (HDRS-ASS <7); between-treatment comparisons conducted across all three arms stratified by anxious/nonanxious status","primaryOutcome":"Change in Montgomery-Åsberg Depression Rating Scale (MADRS) total score from baseline, assessed via MMRM; anxious depression status evaluated as a predictor (main effect) and moderator (interaction with treatment arm) of MADRS change","followUp":"8 weeks (consistent with the parent ASCERTAIN-TRD trial protocol)","fundingSource":"Not reported in the provided abstract; the parent ASCERTAIN-TRD trial was funded by the Patient-Centered Outcomes Research Institute (PCORI), a non-profit comparative effectiveness research organization"}
Key findings
Of the 253 evaluable patients, 182 (71.9%; calculated) met criteria for anxious depression (HDRS-ASS ≥7) and 71 (28.1%; calculated) were classified as nonanxious. Using MMRM analyses that included interaction terms for anxious depression status by treatment group, treatment outcomes as measured by MADRS change were similar across all three treatment arms regardless of anxious MDD status (P > 0.5 for all comparisons). The abstract reports that anxious depression was not a statistically significant predictor (main effect on outcome) or moderator (differential treatment response by anxious subtype) of treatment response in this TRD population. No absolute MADRS change scores by anxious subtype, response rates, remission rates, confidence intervals, effect sizes, or NNT/NNH values are reported in the available abstract. Adverse event data specific to this secondary analysis are not reported.
Practice implications
This secondary analysis addresses a clinically meaningful and previously unresolved question: whether the anxious depression subtype—consistently identified as a negative prognostic marker in non-TRD major depressive disorder—also confers differential treatment response or moderates treatment selection in established TRD. The finding that anxious depression status did not significantly predict or moderate MADRS outcomes across three distinct treatment modalities (pharmacological augmentation, neuromodulation, and antidepressant switching) suggests that clinicians managing TRD need not routinely deprioritize rTMS or SNRI switching strategies specifically on the basis of an anxious presentation. Notably, this is consistent with at least one prior rTMS study reporting that patients with and without anxious depression demonstrated similar improvement rates, and with some antidepressant trials that failed to demonstrate anxious depression as a reliable predictor of poorer outcomes. However, several important caveats apply: (1) only the abstract is available; full model estimates, confidence intervals, and secondary outcomes cannot be evaluated; (2) the study is a secondary analysis of a trial that was underpowered relative to its original design target (n=253 vs. a planned n=639 in the parent trial), raising the possibility of a Type II error—i.e., a true but small moderating effect of anxious depression may exist but could not be detected with the available sample size; (3) patients were not blinded to treatment assignment in the parent trial; (4) the nonanxious group was substantially smaller (n=71), limiting the statistical power of subgroup comparisons; (5) adolescents, elderly patients, and those with psychotic features were excluded from ASCERTAIN-TRD, limiting generalizability to those populations; and (6) the HDRS-ASS ≥7 threshold is one of several competing operational definitions of anxious depression, with evidence that common categorical definitions show poor agreement and may imperfectly capture the construct. The absence of placebo or treatment-as-usual arms means that the prognostic effect of anxious depression on absolute outcomes (as opposed to differential effects across active arms) cannot be fully evaluated in this dataset. __LIMITATIONS__ Secondary analysis of an underpowered parent trial (enrolled n=278 vs. planned n=639); the nonanxious subgroup (n=71) is particularly small, limiting detection of meaningful moderating effects.,Open-label parent trial: participants were not blinded to treatment assignment, introducing performance and expectancy bias that may differentially affect self- and clinician-rated outcomes.,Only the abstract is publicly available; full statistical outputs, confidence intervals, effect sizes, secondary outcome data, and sensitivity analyses cannot be independently reviewed.,Single categorical definition of anxious depression (HDRS-ASS ≥7) used; competing definitions show poor agreement and different predictive validity, limiting generalizability of the null finding.,rTMS arm was curtailed early due to COVID-19 pandemic restrictions, creating unequal group sizes and potential selection bias in who received that treatment.,Exclusion of adolescents, elderly patients, and those with active psychotic features limits generalizability to these clinically relevant TRD subpopulations.,No placebo or treatment-as-usual arm prevents determination of whether anxious depression predicts absolute (as opposed to differential) treatment outcomes.,Funding source for this specific secondary analysis is not reported in the abstract; the parent trial was PCORI-funded (non-profit), which reduces but does not eliminate the risk of selective reporting.,The analysis does not report on clinically important secondary outcomes such as remission rates, response rates, anxiety symptom burden, suicidality, or functional outcomes stratified by anxious subtype. __BOTTOM_LINE_CATEGORY__ promising_but_preliminary __BOTTOM_LINE__ In this secondary analysis of the ASCERTAIN-TRD trial, anxious depression (defined by HDRS-ASS ≥7) was not a statistically significant predictor or moderator of MADRS-based treatment outcomes across aripiprazole augmentation, rTMS augmentation, or SNRI switching in TRD. Clinicians should not use the presence of the anxious depression subtype alone to preferentially select or exclude specific TRD treatment modalities, though this null finding must be interpreted cautiously given the relatively small sample size, open-label design, and availability of only the abstract. Confirmation in larger, adequately powered TRD cohorts with full moderator analyses and validated multidimensional anxiety measures is needed before definitive practice guidance can be issued.