Supplementary Material for: Identification of an Immune-Neuroendocrine Biomarker Panel for Detection of Depression: A Joint Effects Statistical Approach

<b><i>Background/Aims:</i></b> Less than half of depression patients are correctly diagnosed within the primary care setting. Previous proteomic studies have identified numerous immune and neuroendocrine changes in patients. However, few studies have considered the joint effects of biological molecules and their diagnostic potential. Our aim was to develop and validate a diagnostic serum biomarker panel identified through joint effects analysis of multiplex immunoassay profiling data from 1,007 clinical samples. <b><i>Methods:</i></b> In stage 1, we conducted a meta-analysis of two independent cohorts of 78 first-/recent-onset drug-naive/drug-free depression patients and 156 controls and applied the 10-fold cross-validation with least absolute shrinkage and selection operator regression to identify an optimal diagnostic prediction model (biomarker panel). In stage 2, we tested the discriminatory performance of this biomarker panel using the naturalistic Netherlands Study of Depression and Anxiety (NESDA) cohort of 468 depression patients and 305 controls. <b><i>Results:</i></b> An optimal panel of 33 immune-neuroendocrine biomarkers and gender was selected in the meta-analysis. Testing this biomarker-gender panel using the NESDA cohort resulted in a moderate to good performance to differentiate patients from controls (0.69 < AUC < 0.86), particularly the first-episode patients free of chronic non-psychiatric diseases or medications and following incorporation of sociodemographic covariates (0.76 < AUC < 0.92). <b><i>Conclusion:</i></b> Despite the need for additional validation studies, we demonstrated that a blood-based biomarker-sociodemographic panel can detect depression in naturalistic healthcare settings with good discriminatory power. Further refinements of blood biomarker panels aiding in the diagnosis of depression may provide a cost-effective means to increase accuracy of clinical diagnosis within the primary care setting.