Symptom-specific Temporal Phenotyping of Anti-depressant Response to Citalopram in STAR*D Trial

AbstractImportance: Depression treatments do not have equal potency for all depression symptoms. To inform research design and improve clinical care, it is critical to determine if these symptom-specific effects occur at different rates.Objective: To determine the symptom-specific time-to-improvement of depression symptoms during citalopram treatment.Design: The study uses STAR*D data, which prescribed citalopram as the first stage of an open-label, multi-stage clinical trial of stepwise treatment strategies for depression. This first stage typically lasted 12 weeks and follow-up assessments were, on average, every-other-week.Setting: Multi-centerParticipants: Over 4,000 adults aged 18-75 years diagnosed with Major Depressive Disorder without psychotic features participated in the STAR*D trial. Participants were prospectively enrolled patients in both primary and specialty care venues.Interventions: All participants started 20 mg/day of citalopram. If this dose was well tolerated, it could be increased to a maximum of 60 mg/day. Stage 1 typically lasted 12 weeks.Main Outcome(s) and Measure(s): This exploratory secondary data analysis focuses on change in the total score and each of the individual items, of the Quick Inventory of Depressive Symptomatology-SR.Results: 3,248 depressed adults were available for the present analyses (Mean age = 41 year (SD=13 years), 62% female)). The number of observations differed for each model. The median time for 1 standard deviation improvement in depression symptoms was 28 days. Median time for 1 point improvement (all possible thresholds are reported in text) varied widely for individual symptoms. The quickest improving symptom had a median-time-to-improvement of 15 days (increased weight) and the longest took 41 days (mid-nocturnal insomnia). The rank-order of which symptoms improved the quickest was highly consistent across models with different improvement thresholds (Spearman’s ρ = .66-.91, all p <.005) and when stable improvement across multiple follow-ups was required (Spearman’s ρ=.97, p=.003). Conclusions and Relevance: Results indicate that different depression symptoms respond to citalopram at different speeds. In addition to implications for designing clinical trials, this information is essential for developing realistic treatment expectations and to facilitate shared decision-making between clinicians and their patients about how to select treatments.Trial Registration: STAR*D is registered on ClinicalTrials.gov under ID NCT00021528 (https://clinicaltrials.gov/study/NCT00021528).