Hair levels of steroid, endocannabinoid, and the ratio biomarkers predict viral suppression among people living with HIV/AIDS in China.

ABSTRACT

BACKGROUND: Predicting viral suppression early is crucial to improving treatment outcomes among people living with HIV/AIDS (PLWH) in clinics. Viral suppression is affected by stress, making stress indicators a potential predictive factor. Most of previous studies used the self-report questionnaire as stress indicators, but there were great drawbacks due to its subjective. In contrast, end products of neuroendocrine systems such as hypothalamic-pituitaryadrenal (HPA) and hypothalamic-pituitary-gonadal (HPG) axes and endogenous cannabinoid system (ECS) that involved in regulating stress as objective stress indicators are urgently needed to predict viral suppression. Therefore, this study aimed to investigate whether neuroendocrine indictors can strongly predict viral suppression among PLWH in China. METHODS: This cross-sectional study recruited 1198 PLWH on antiretroviral therapy (ART) in Guangxi, China. The concentrations of steroids (i.e., cortisol, cortisone, dehydroepiandrosterone, testosterone and progesterone) and endocannabinoids (i.e., N-arachidonoyl-ethanolamine and 1-arachidonyl glycerol) in hair were quantitated using the LC-APCI+-MS/MS method. To screen biomarkers that were used to predict viral suppression, association between hair biomarkers and viral suppression was examined by Mann-Whitney U test and partial correlation analyses. Receiver operating characteristic (ROC) curves and binary logistic regression based on the optimal classification threshold determined with ROC curves were used to estimate the prediction effects of the screened biomarkers on viral suppression (HIV-1 RNA < 200 copies/mL). RESULTS: Hair levels of dehydroepiandrosterone (DHEA), and N-arachidonoyl-ethanolamine (AEA), and the cortisol to DHEA ratio exhibited significant intergroup differences (ps < 0.05) and were correlated with HIV viral load (ps < 0.05). Hair DHEA concentrations strongly predicted viral suppression, showing good classification performance (area under the ROC curve = 0.651, p < 0.01) and strong predictive utility (adjusted odd ratio = 2.324, 95 % confidence interval = 1.211-4.899, p < 0.05) with an optimal threshold of 10.5 pg/mg. A hair AEA concentration of 2.4 pg/mg was the optimal threshold for predicting viral suppression based on good classification performance (area under the ROC curve = 0.598, p < 0.05) and predictive power (adjusted odd ratio = 2.124, 95 % confidence interval = 1.045-4.244, p < 0.05). In hair levels of cortisol to DHEA, viral suppression was observed to be highly predictive, with a threshold of 10.5 pg/mg being optimal for classification (area under the ROC curve = 0.624, p < 0.05) and prediction (adjusted odd ratio = 0.421, 95 % confidence interval = 0.201-0.785, p < 0.05). CONCLUSION: Hair levels of DHEA, and AEA and the cortisol to DHEA ratio were screened and verified to have significant predictive power with optimal thresholds for predicting viral suppression in a large-scale cohort. The data may provide new insights into predictors of successful virological outcomes and inform public health intervention and clinical practice to assist PLWH in achieving and sustaining viral suppression.