Inflammation marker shows U-shaped link to medication use in acute depression

Patients admitted to hospital during an acute phase of major depressive disorder show a U-shaped relationship between blood levels of high-sensitivity C-reactive protein (hs-CRP) and whether they are currently taking psychiatric medication, according to a retrospective cross-sectional study published in BMC Psychiatry.

The single-center study included 492 acute-phase inpatients treated between 2020 and 2025. Of these, 355 were classified as medicated and 137 as unmedicated, based on their use of antipsychotics, antidepressants, or sedative-hypnotics at admission. The cohort comprised 114 men (23.2%) and 378 women (76.8%), with a mean age of 53.15 years and a median duration of illness of six years.

The researchers collected sociodemographic data, illness duration, and a range of biochemical parameters, including hs-CRP, blood counts, glucose-lipid metabolism measures, and thyroid function. The unmedicated group had significantly lower hs-CRP levels than the medicated group, with a median of 0.83 mg/L compared with 1.13 mg/L (p = 0.04, Cohen’s d = 0.30).

In multivariable logistic regression, each 1.0 mg/L increase in hs-CRP was associated with an 18% increase in the odds of current medication use (odds ratio 1.18, 95% confidence interval 1.00–1.38, p = 0.048). But the linear summary masked a more complex pattern that emerged when the team applied generalized additive model-based smooth curve fitting and piecewise regression to detect nonlinearity and threshold effects.

Those analyses revealed a U-shaped association, with a statistically significant nonlinearity (P for nonlinearity = 0.04) and a turning point at 0.79 mg/L. Below that threshold, each 0.1 mg/L increment in hs-CRP was associated with 17% lower odds of being on medication (OR 0.83, 95% CI 0.72–0.95, p = 0.007). At or above 0.79 mg/L, the direction reversed: each 1.0 mg/L increase corresponded to 172% higher odds of current pharmacotherapy (OR 2.72, 95% CI 1.11–6.65, p = 0.03).

In other words, among patients with relatively low hs-CRP, higher values were linked to a reduced likelihood of being on medication, while above the turning point higher values were linked to an increased likelihood. The authors describe the 0.79 mg/L value as an exploratory turning point, and the study’s conclusion emphasizes that the association is non-linear rather than monotonic.

Subgroup analyses stratified by sex, age, duration of MDD, diabetes, and hypertension yielded consistent directional findings across all strata, with all interaction p-values greater than 0.05, suggesting the pattern did not differ significantly across these patient characteristics.

Several features of the study design deserve emphasis when interpreting the findings. The data are cross-sectional and retrospective, drawn from a single center, so the results show an association between hs-CRP and medication status at a single point in time rather than a causal effect in either direction. It is not possible to determine from this design whether inflammation influences prescribing decisions, whether medication use affects hs-CRP levels, or whether both are shaped by other factors such as illness severity. The authors note that the study was designed to investigate the correlation and potential non-linear relationship between serum hs-CRP and medication status, not to test whether any intervention changes outcomes.

The work was published as an open-access research article on 23 October 2026, and the version shared ahead of the final record is citable and carries a permanent DOI, subject to further edits.

Subject of Research: The non-linear association between serum hs-CRP levels and psychiatric medication status in patients with acute-phase major depressive disorder.

Article Title: U-shaped association between serum hs-CRP and medication status in acute-phase major depressive disorder

Article References: Zhu, Q., Yao, B., Zhang, W., & Zheng, Y. (2026). U-shaped association between serum hs-CRP and medication status in acute-phase major depressive disorder. BMC Psychiatry. https://doi.org/10.1186/s12888-026-08787-3

Image Credits: AI Generated

DOI: 10.1186/s12888-026-08787-3

Keywords: hs-CRP, major depressive disorder, inflammation, medication status, cross-sectional study, logistic regression, threshold effect, psychiatry, BMC Psychiatry, C-reactive protein, U-shaped, association