The Relation Between Cortisol and Anthropometric Measurements Throughout Lifespan: A Systematic Review and Meta-Analysis

Introduction: Recently, cross-sectional studies report associations between long-term glucocorticoid levels in scalp hair (HairGC) and obesity. However, there is a wide variation in studied outcomes and associations, possibly caused by differences in population characteristics, e.g. age, sex, dispersion of adiposity, and used laboratory methods. The aim of this systematic review and meta-analysis was to investigate the relation between HairGC and anthropometrics and to explore possible moderators of this association. Methods: We searched the Medline, Embase, Cochrane, Web of Science, Scopus, Cinahl, PsycInfo, and Google Scholar databases for articles that relate HairGC to measures of adiposity (date 11-16-2020). Primary outcomes were correlations between hair cortisol (HairF) and cortisone (HairE), and anthropometrics: BMI, waist circumference (WC) and waist-hip-ratio (WHR). Authors were contacted to provide missing outcome information. Pooled correlation coefficients were calculated using random effects models. Assessment of heterogeneity was performed using the I(2) statistic. Exploratory moderator analyses were performed with subgroup analyses and meta-regression. This systematic review was performed in accordance to the PRISMA guidelines. Results: Our systematic search identified 150 cohorts, comprising a total of 37,107 unique individuals, of which 15,033 sampled from population-based cohorts. For BMI, the pooled correlation for HairF was 0.121 (95% CI 0.083–0.158, n=26,941; I(2) 94.2%, p<0.001) and for HairE 0.108 (95% CI 0.047–0.167, n=7,250; I(2) 52%, p<0.01). For WC, the pooled correlation for HairF was 0.111 (95% CI 0.058–0.164, n=10,290; I(2) 63%, p<0.01) and for HairE 0.200 (95% CI 0.137–0.264, n=2,198; I(2) 0%, p=0.42). For WHR, the pooled correlation for HairF was 0.102 (95% CI 0.040–0.163, n=6,865; I(2) 27%, p=0.14) and for HairE 0.261 (95% CI 0.195–0.330, n=1,314; I(2) 0%, p=0.40). A higher percentage of male participants was related to stronger correlations with WC (p<0.001), but not with BMI and WHR. Mean age, mean BMI, and mean HairGC levels of the cohorts did not significantly moderate the pooled correlations, neither did the used laboratory techniques (immunoassays vs mass spectrometry-based assays). Conclusion: This unique, large meta-analysis demonstrates that long-term endogenous glucocorticoids as assessed by HairGC show small but consistent correlations to measures of obesity, despite a large heterogeneity between the included cohorts. The strongest associations were found between HairE and WC and between HairE and WHR. This suggests that glucocorticoid levels in the high-normal range, especially cortisone, may contribute to or reflect the state of specifically central adiposity, even within the general population.