Trajectories of depression symptom improvement and associated predictor analysis: An analysis of duloxetine in double-blind placebo-controlled trials

- Trajectories of symptom improvement in patients with MDD were examined.
- Ten placebo-controlled trials of duloxetine were included; MDD symptoms were assessed with the HAM-D.
- We found different trajectories of improvement in remitters and nonremitters, starting at early time points.
- Early improvement on many items of the HAM-D predicted remission.
- Early monitoring of some individual symptoms of MDD may be useful in making treatment decisions.

BackgroundIn the treatment of major depressive disorder (MDD), it is not fully understood how individual symptoms improve over time (trajectory) in remitters. This study compared symptom improvement trajectories, as measured with the 17-item Hamilton Depression Rating Scale (HAM-D17), in remitters and nonremitters.MethodsThis analysis is based on 10 placebo-controlled, randomized, double-blind trials of duloxetine (40-60 mg/day) for treatment of MDD from baseline up to week 8. Remission was defined as a HAM-D17 total score ≤7 at week 8 (last observation carried forward). Trajectories of HAM-D17 items were assessed by mixed model repeated measures analysis for treatment and remitter-nonremitter comparisons. Grouping of the trajectories was performed by factor analysis. Predictor analysis using HAM-D17 items was conducted by logistic regression.ResultsThere were 1555 patients in the duloxetine group (489 [31.4%] remitters) and 1206 patients in the placebo group (290 [24.0%] remitters; P<.0001). For most items, the difference in trajectories between remitters and nonremitters appeared at early time points and increased over time. Treatment response trajectories were very similar for duloxetine and placebo remitters, while duloxetine nonremitters improved more than placebo nonremitters. For duloxetine remitters, we found 3 trajectory groups of HAM-D17 items. The predictor analysis showed that improvement in 6 individual items at week 1 or 2 was significantly associated with remission at week 8.LimitationsGeneralizability of these results may be limited by the relatively short observation period used to define remission.ConclusionsEarly monitoring of some symptoms of depression may prove useful in guiding treatment decisions.