Peripheral Biomarkers of Anorexia Nervosa: A Meta-Analysis

Abstract

The pathogenesis of anorexia nervosa (AN) has been hypothesized to involve several biological systems. However, reliable biomarkers for AN have yet to be established. This study was aimed to identify statistically significant and clinically meaningful peripheral biomarkers associated with AN. A systematic literature search was conducted to identify studies published in English from inception until 30 June 2022. We conducted two-level random-effects meta-analyses to examine the difference between AN and comparison groups across 52 distinct biomarkers and found that acylated ghrelin, adrenocorticotropic hormone (ACTH), carboxy-terminal collagen crosslinks (CTX), cholesterol, cortisol, des-acyl ghrelin, ghrelin, growth hormone (GH), obestatin, and soluble leptin receptor levels were significantly higher in cases of AN compared with those in non-AN controls. Conversely, C-reactive protein (CRP), CD3 positive, CD8, creatinine, estradiol, follicle-stimulating hormone (FSH), free thyroxine, free triiodothyronine, glucose, insulin, insulin-like growth factor 1 (IGF-1), leptin, luteinizing hormone, lymphocyte, and prolactin levels were significantly lower in AN compared with those in non-AN controls. Our findings indicate that peripheral biomarkers may be linked to the pathophysiology of AN, such as processes of adaptation to starvation. Scientific investigation into peripheral biomarkers may ultimately yield breakthroughs in personalized clinical care for AN.

Keywords: anorexia nervosa; eating disorder; peripheral biomarkers; starvation.

Figure 1

Flowchart of the inclusion procedure in a PRISMA diagram. AN = anorexia nervosa, BN = bulimia nervosa, BED = binge-eating disorder, ED = eating disorder.

Figure 2

Forest plot of all biomarkers’ summary effect sizes. The forest plot was organized by the direction of effect for biomarkers higher in the AN group, those higher in the comparison group, and those that were not statistically significant. N = number of studies; Tau² = Tau-squared for examining between-study heterogeneity; I² = heterogeneity index for examining between-study heterogeneity; PEESE p-value = p-value for the precision-effect estimate for the SE publication bias test (<0.05 indicates possible publication bias); SMD = standardized mean difference; 95% CI = 95% confidence interval.