Serum Cytokines as Biomarkers for Comorbid Anxiety in Postpartum Depression: A Machine Learning Approach

Area of Science:

• Immunology
• Psychiatry
• Computational Biology

Background:

• Postpartum depression (PPD) is a significant mental health concern.
• The role of inflammatory markers in PPD and comorbid anxiety requires further investigation.

Purpose of the Study:

• To investigate serum cytokine levels, specifically interleukin-6 (IL-6), in patients with PPD.
• To explore the potential of IL-6 as a biomarker for PPD and comorbid anxiety.
• To apply machine learning techniques to predict PPD and anxiety based on cytokine profiles.

Main Methods:

• Serum samples were collected from 53 PPD patients and 35 healthy controls.
• Cytokine levels (IL-2, IL-6, IL-10, TNF-alpha) were measured using flow cytometry.
• Machine learning models (Logistic Regression, Decision Trees, Random Forest, SVMs) were employed for prediction.

Main Results:

• PPD patients showed significantly higher serum IL-6 levels compared to controls.
• A positive correlation was observed between anxiety scores and IL-6 levels (r = 0.483, P < .001).
• Random Forest and SVM models achieved high accuracy in predicting PPD and anxiety, with IL-6 as a crucial feature.

Conclusions:

• Serum cytokine activation, particularly IL-6, is present in PPD.
• IL-6 shows potential as an auxiliary biomarker for diagnosing PPD and associated anxiety.
• Machine learning enhances understanding of cytokine-PPD relationships and IL-6's correlation with symptom severity.