Back

An increase in C-reactive protein levels during antidepressant treatment predicts treatment non-response in major depressive disorder

Etzel, M.; Rosenblum, Y.; Dresler, M.; Steiger, A.; Mikoteit, T.; Zeising, M.

2025-08-29 psychiatry and clinical psychology
10.1101/2025.08.27.25331881 medRxiv
Show abstract

BackgroundOne-third of patients with major depressive disorder (MDD) exhibit low-grade inflammation as reflected by C-reactive protein (CRP) concentrations > 3 mg/L. We explored whether CRP changes from baseline to week one of antidepressant treatment ({Delta}CRP) can serve as a marker of treatment response. MethodsCRP plasma levels were measured at baseline and after the first week of treatment in 33 MDD patients and correlated with patients Hamilton depression scale (HAM-D), while adjusting for age, gender and body mass index. The early and final antidepressant responses were defined as a > 25% and > 50% HAM-D score reduction at week one and four of treatment compared to baseline, respectively. We compared baseline and week-one CRP levels with the paired t-test within responders and non-responders separately and {Delta}CRP between the groups with the ANCOVA. ResultsHigher {Delta}CRP correlated with lower final {Delta}HAM-D scores (r = -0.5, p = 0.006). Non-responders showed higher {Delta}CRP - but not baseline and week one CRP - than responders (p = 0.018, Cohens d = 1.1). A {Delta}CRP increase was observed in 13/16 (81%) non-responders and 7/17 (41%) responders (Fishers exact tests p = 0.03). A {Delta}CRP increase combined with an early non-response was observed in 13/16 (81%) non-responders and 1/17 (6%) responders (p < 0.0001). ConclusionsRather early {Delta}CRP at week one than baseline CRP might be indicative of treatment response at week four, especially if combined with early {Delta}HAM-D. In the future, {Delta}CRP could be introduced into psychiatric practice to guide treatment plans.

Matching journals

The top 4 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.