Predictive comparison of specific depression symptoms for peripheral inflammation

Area of Science:

• Neuroscience
• Immunology
• Psychiatry

Background:

• Peripheral inflammation is increasingly recognized as a factor in depression.
• However, its relationship may be specific to certain depressive symptoms, not a uniform association with the overall depression score.
• Neurovegetative symptoms, like fatigue, are hypothesized to be more closely linked to inflammation.

Purpose of the Study:

• To investigate the relationship between peripheral inflammation markers and specific depressive symptoms versus the overall depression score.
• To compare the predictive power of depression sum scores against individual symptoms, particularly lethargy, in relation to inflammation.
• To explore a specific depression phenotype associated with inflammation using a data-driven approach.

Main Methods:

• Utilized a large dataset from the UK Biobank (n=311,474).
• Employed a residualized random forest framework for statistical comparison.
• Assessed peripheral inflammation using C-reactive protein, white blood cell counts, neutrophil-to-lymphocyte ratio, and systemic immune inflammation index.
• Measured depressive symptoms via a self-report questionnaire.

Main Results:

• Lethargy demonstrated significantly better prediction of peripheral inflammation levels compared to the overall depression sum score across various analytical settings.
• This indicates a stronger association between fatigue-related symptoms and peripheral inflammation.
• Despite statistical significance, the overall model performance was poor, limiting practical interpretability.

Conclusions:

• Findings suggest a specific depression phenotype where fatigue symptoms are closely linked to peripheral inflammation.
• Highlights the potential of symptom-level analyses for identifying depression biomarkers.
• Emphasizes the value of precision intervention strategies targeting specific symptom clusters in depression.