S100B serum levels predict treatment response in patients with melancholic depression

Background: There is an ongoing search for biomarkers in psychiatry, for example, as diagnostic tools or predictors of treatment response. The neurotrophic factor S100 calcium binding protein B (S100B) has been discussed as a possible predictor of antidepressant response in patients with major depression, but also as a possible biomarker of an acute depressive state. The aim of the present study was to study the association of serum S100B levels with antidepressant treatment response and depression severity in melancholically depressed inpatients. Methods: After a wash-out period of 1 week, 40 inpatients with melancholic depression were treated with either venlafaxine or imipramine. S100B levels and Hamilton Depression Rating Scale (HAM-D) scores were assessed at baseline, after 7 weeks of treatment, and after 6 months. Results: Patients with high S100B levels at baseline showed a markedly better treatment response defined as relative reduction in HAM-D scores than those with low baseline S100B levels after 7 weeks (P=.002) and 6 months (P=.003). In linear regression models, S100B was a significant predictor for treatment response at both time points. It is of interest to note that nonresponders were detected with a predictive value of 85% and a false negative rate of 7.5%. S100B levels were not associated with depression severity and did not change with clinical improvement. Conclusions: Low S100B levels predict nonresponse to venlafaxine and imipramine with high precision. Future studies have to show which treatments are effective in patients with low levels of S100B so that this biomarker will help to reduce patients' burden of nonresponding to frequently used antidepressants