In this elderly men-based cross-sectional study, we observed an increase in the risk of NAFLD among those who had higher circulating FSH levels. This association was independent of age, demographic factors, lifestyle factors, history of hypertension and diabetes, BMI, WC, nutritional risk, and other sex hormones.
Prior data investigating FSH in metabolic diseases, including NAFLD, have primarily focused on postmenopausal women.20,10−13 Indeed, FSH is more often measured and concerned in females, and postmenopausal women have higher metabolic risks than pre-menopausal women. Our study represents an important advance in our understanding of FSH and NAFLD, highlighting the association of FSH levels in an aging male population with the risk of NAFLD.
Few other studies have looked at FSH levels in relation to NAFLD in males or postmenopausal women. In line with our findings, a study of community-dwelling men aged 20–69 years observed a progressive increase in FSH with aging.8 However, no association between FSH and NAFLD was found, which may be explained by the limited adjustment for only age, LDL-c, and other sex hormones in the logistic regression analyses. In contrast with our data, a previous study conducted in a postmenopausal female population revealed that FSH was anti-steatotic, with low FSH conferring a higher risk of ultrasonography-confirmed NAFLD.13 It remains unknown why FSH plays sex-dependent roles predisposing individuals to NAFLD, although androgens may also affect men and women differently.6,21
The precise mechanism behind the association of FSH and NAFLD remains unknown. FSH is now regarded to play direct roles in a variety of nongonadal tissues by interacting with the FSH receptors in those tissues.22–24 For example, FSH promotes fat accumulation and fat redistribution (from subcutaneous fat depots to visceral fat depots) in aging, and blocking FSH with polyclonal antibodies induces thermogenic adipose tissue and reduces body fat.18,25 The positive association between FSH and obesity was confirmed in human studies in which circulating FSH levels were observed to correlate with increases in BMI of both males and females.25 Notably, we failed to find this association in our elderly male population, possibly because non-obese NAFLD (BMI < 28 kg/m2) was more common (81.1%) in this population and BMI is a poor predictor of adiposity in some elders with height loss. Our findings from epidemiological investigations seem suggest that FSH also drive ectopic (such as liver) fat deposition besides its regulation in adipose tissue metabolism, although no evidence of direct association between FSH and hepatis steatosis was reported previously. Considering the aging-associated changes in circulating FSH level and prevalence of NAFLD, elevated FSH may be one of the possible mechanisms to explain the more NAFLD subjects found in elderly. Further large prospective studies and basic research are required to confirm our hypothesis in the future.
Dietary protein deficiency has been associated with excessive TG storage and NAFLD in population studies, and older adults are at risk for protein malnutrition.26 Therefore, in addition to common pathogenic risks of NAFLD, we also included nutritional risk as a variable in our analysis of the association between FSH and NAFLD. Using the scores of the MNA-SF, we showed that participants in the present study had a high prevalence of malnutrition/nutritional risk (30.5%). However, to our surprise, elderly subjects with NAFLD had a lower nutritional risk than those who were non-NAFLD (Table 1). This lack of association between protein insufficiency and NAFLD may be due to adequate protein intake in the studied population who lived in Shanghai, a large city with a thriving economy. Nevertheless, a possible negative correlation of circulating FSH and cores of MNA-SF was observed in this study (after adjusting for age, r = −0.092, P = 0.059), and cores of MNS-SF were included in the multivariate logistic regression analysis.
Limitations of this study include that the subjects are from a population receiving regular medical examination as opposed to a community or a general population, so our results may not reflect the general population. Liver ultrasonography but not a liver biopsy was used to diagnose NAFLD in our subjects. This was chosen since healthy subjects would likely not accept to perform liver biopsies routinely. This was a cross-sectional single-center study with a relatively small sample size, which is a significant flaw, and further prospective studies with larger sample sizes are required to confirm our present findings.