Abstract
BACKGROUND: Antidepressive treatment outcomes can be assessed using sum scores from measurement scales, but symptom clusters may better capture the multidimensional structure of depression. Little is known about the comparative effectiveness of different treatment modalities on these clusters. We sought to evaluate the effects of rTMS and antidepressant medication on four symptom clusters and the extent to which these differ between treatments. In addition, we assessed whether baseline cluster scores predicted (non)response. METHODS: Data were obtained from two clinical trials (DETECT: rTMS vs. medication; PITA: tricyclic antidepressants). Primary outcomes were symptom cluster scores: 'General Depression', 'Anxiety', 'Somatic Symptoms' and 'Insomnia'. In the primary analysis based on DETECT, a MANCOVA comparing rTMS with medication was performed. For validation, a MANCOVA was performed replacing medication data from DETECT with data from PITA. Baseline symptom cluster scores were compared between responders and non-responders, as well as treatment groups. RESULTS: In both the primary and validation analyses, no difference was seen between rTMS and medication on the symptom clusters. Similar patterns of response were observed in all groups, with 'Insomnia' showing the greatest effect of treatment. Baseline cluster scores did not predict treatment response. CONCLUSIONS: We did not find a differential effect of rTMS or antidepressant medication on depressive symptom clusters. Both treatments demonstrated comparable response patterns for all clusters, and baseline cluster scores did not differ between responders and non-responders. Future studies with larger samples or more homogeneous treatments may elucidate the role of symptom clusters as a tool for more individualized treatment. TRIAL REGISTRATION: PITA NCT03548675 DETECT The Netherlands Trial Register NL7628.