In the present study, the prevalence and risk factors of MAFLD were explored, and the significant difference in the prevalence of MAFLD among different groups of gender, age, BMI and female menopausal status were revealed. To our knowledge, this is the first study focusing on the prevalence and associated metabolic characteristics of MAFLD since the new definition of MAFLD was set up [4, 5].
The prevalence of MAFLD was 26.1% (23.8% after age and sex adjustment), and it was significantly higher in males than in females (35.4% vs. 14.1%, and 32.3% vs. 13.4% after age and sex adjustment), which was consistent with previous studies of NAFLD in Asian populations . revealing that male gender is still predominate in MAFLD. The total prevalence reached peak at the age of 55–59 (Fig. 1). In males, the prevalence rises steeply between the age of 18–39, then rose slowly after 40 years and reached peak at the age of 50–54, revealing that MAFLD is more prevalent in men in their middle ages and they were at an increasing risk of having MALFD while in a younger age which need to be paid special attention. We observed that older men had a lower prevalence of MAFLD than middle-aged men, the possible reasons might be as follows: Some people died of other disease in their older age as fatty liver can significantly increase overall mortality , hence these people cannot be counted in; Compared with old men who usually retire, middle-aged men who are at peak of their careers have much more social engagement like dinner and party to take part in, which might increase their risk of having metabolic disorders. Whereas in females, the prevalence of MAFLD is relatively low before the age of 50, however, it rose sharply between the age of 50–69 and reached peak at the age of 65–69. Notably, the sharply increasing trend in females was consistent with the presence of perimenopausal period, indicating that change of estrogen levels was an important factor of the rapid increase in the prevalence of MAFLD. Besides, after stratification by menopausal status, the prevalence of MAFLD in females rose steady from premenopausal period (6.1%) to perimenopausal period (16.8%), then to postmenopausal period (30.2%) (Fig. 2). This phenomenon consists with former studies , indicating that estrogen might have protective effects against MAFLD in females . Conversely, the decrease of estrogen in perimenopausal and postmenopausal women can lead to fat redistribution and thus causing metabolic disorders including dyslipidemia and glucose intolerance , which can also be demonstrated in our study that the prevalence of dyslipidemia (22.9% for premenopausal period, 51.6% for perimenopausal period and 68.8% for postmenopausal period) was parallel with the trend of the MAFLD prevalence stratified by menopausal status. This finding indicated that the perimenopausal and postmenopausal status might be substantial risk factors for MAFLD in women, suggesting estrogen deficiency in perimenopausal and postmenopausal women may also play an important role in the development of MAFLD.
Previous studies have found that presence of NAFLD is closely related to MS components such as obesity, insulin resistance, hypertension and dyslipidemia, and is considered to be the liver manifestation of MS . Our study has also found that after stratified by BMI, the prevalence of MAFLD increased sharply with the increase of BMI, reaching 59.8% in obese people (Fig. 3). In binary logistic regression (Table 4), BMI and waist circumference were also significantly associated with MAFLD, indicating obesity is closely associated with MAFLD and obesity management should be emphasized as weight loss was proved to reduce steatosis .
In people with MAFLD, the proportions of the abnormal metabolic features were all significantly higher than those in people without MAFLD (Table 2), confirming that MAFLD is closely associated with MS components including abdominal obesity, hypertension, dyslipidemia, and dysglycemia. Among them, in addition to waist circumference, the most significant difference was found in elevated triglyceride, and triglyceride were also shown to be significantly associated with MAFLD in logistic regression (Table 4), with the highest OR value of 1.776, suggesting that elevated triglyceride may be an important risk factor of MAFLD. Moreover, the difference in the proportion of elevated fasting glucose was also very significant, and fasting glucose was also significantly associated with MAFLD in logistic regression with the OR value of 1.403, consisting with the previous study that there’s correlation between fatty liver and dyslipidemia and dysglycemia . Studies have also shown that NAFLD is not only closely related to the cardiovascular and renal diseases associated with MS, but also precedes the presentation of metabolic derangements , considering the high proportions of abnormal metabolic features (Table 2) and the significant high prevalence of MS (53.2% vs. 10.1%) in people with MAFLD (Fig. 4) in our study, MAFLD may be a significant risk factor of MS. Notably, we noticed that the prevalence of MS in females with MAFLD was significantly higher than that in males with MAFLD (61.4% vs. 50.6%, P < 0.001) (Fig. 4), indicating that in patients with MAFLD, females may be more susceptible to MS than males. Furthermore, we also found that the prevalence of MAFLD in people without any MS risk factors was 3.3%, while those met all 5 MS criteria had a prevalence of 82.5% (Table 3), and the prevalence of MAFLD in people with MS was significantly higher than those without MS (65.1% vs. 15.5%, P < 0.001), which also demonstrated that MS and its components are important risk factors of MAFLD.
It was also shown in our study that the prevalence of dyslipidemia and hyperuricemia was significantly higher in people with MAFLD than those without MAFLD (80.0% vs. 41.7% for dyslipidemia and 45.0% vs. 16.8% for hyperuricemia) (Fig. 4). Dyslipidemia is a well-known risk factor for NAFLD , and this can also be reflected in the sharp rise of the prevalence of MAFLD in perimenopausal and postmenopausal women in our study, which might be related to the dyslipidemia due to estrogen deficiency (Fig. 2). In binary logistic regression, uric acid was showed to be significantly associated with MAFLD. Previous cross-sectional and prospective studies have found that elevated serum uric acid could independently predict the increased risk of NAFLD, even serum uric acid levels within normal range were closely related to the presence of NAFLD independently [23–25]. Hence, combining the findings in our study with previous studies, serum uric acid might be considered as an independent risk factor of MAFLD.
The present study revealed that people with MAFLD are more likely to have elevated liver enzymes especially elevated ALT compared with those without MAFLD (42.5% vs. 11.1% for elevated ALT, P < 0.001) (Table 2), which stands for the higher proportion of abnormal liver function in people with MAFLD, consisting with the previous study that the most common cause of unexplained elevated ALT is NAFLD . Moreover, Table 4 showed that ALT was significantly associated with MAFLD and studies have showed that elevated ALT is associated with the development of NAFLD into steatohepatitis and even liver fibrosis , indicating that elevated ALT has an important clinical significance for MAFLD. Platelets are considered to be elevated during inflammation, and studies have found a linear correlation between platelet count and the severity of liver fibrosis in people with NAFLD , in our study, we also found that platelet count was significantly associated with MAFLD (Table 4), indicating that platelet count and ALT levels may be used as a reference indicator of MAFLD development and the resulting liver fibrosis.
Although the occurrence of NAFLD is closely related to obesity, non-obese people may also suffer from NAFLD, especially in the Asia-Pacific region . In our study, the proportions of abnormal metabolic features in non-obese people with MAFLD were all significantly higher than those in non-obese people without MAFLD (Table 2), suggesting that metabolic disorders also play an important role in the occurrence of MAFLD in non-obese people. Notably, the proportion of elevated total cholesterol was significantly higher in non-obese people with MAFLD than that in obese people with MAFLD, suggesting total cholesterol may play an important role in the presence of non-obese MAFLD which need to further exploration.
Our study also has certain limitations. First, this is a cross-sectional study, therefore, the natural course of MAFLD and causal relationships can’t be determined. There was a large sample size with a wide range of clinical data included in our study, making it possible to adjust for underlying confounding factors. Second, the diagnosis of MAFLD was based on ultrasonography, which might be partially insensitive to mild hepatic steatosis. However, ultrasonography has been widely used in the epidemiological investigations of fatty liver since it is safe, non-invasive, widely available, and has an acceptable sensitivity and specificity in the detection of hepatic steatosis , and it was also recommended as the first-line imaging method by the Association for the Study of the Liver (APASL) in the clinical guideline for MAFLD . Third, there might be certain selection bias because the population who participated in health examination included in our study tend to be more concerned about their health. Furthermore, some information was not available from the current data of health examinations, like medication history of participants.