We conducted a cross-sectional study of patients with bipolar disorder who were hospitalized at the Anhui Provincial Mental Health Center. Anonymous data regarding general demographic characteristics and test outcomes were gathered. The criteria for inclusion were as follows:1) The diagnostic criteria were in accordance with the International Classification of Diseases, 10th Edition (ICD-10), and confirmed by two psychiatrists holding professional titles above the attending level; 2) Age range of 18 to 60 years; 3) No use of liver-protective medications; 4) Absence of hepatitis or any other conditions potentially contributing to fatty liver disease; 5) No recent (within the past three months) use of medications known to induce MAFLD, including psychiatric drugs; 6) No documented history of alcohol-related disorders that fulfill diagnostic standards. Exclusion criteria included: 1) Presence of a diagnosis or history related to substance misuse or dependency; 2) Pregnant individuals; 3) Presence of neurodegenerative conditions, such as congenital developmental delays or Alzheimer’s disease;4) Comorbid organic brain diseases or severe somatic diseases; 5) Malnutrition, hepatic bean nucleus degeneration, and other conditions that may lead to fatty liver. Adhering to the specified criteria, 1,072 individuals were recruited for the study during the recruitment period from July 2020 to December 2022. The relevant screening flowchart is shown in Figure 1 .

The BMI was determined through the division of body weight, measured in kilograms with participants in light attire. Participants were categorized into two groups based on the criteria established by the Chinese Working Group on Obesity (WGOC) (19): the lean group (BMI < 24 kg/m²) and the overweight or obese group (BMI ≥ 24 kg/m²).

In this study, we systematically evaluated the following biochemical indicators: blood glucose (GLU), uric acid (UA), triglycerides (TG), high-density lipoprotein cholesterol (HDL-C), apolipoprotein B (ApoB), apolipoprotein A1 (ApoA1), gamma-glutamyltransferase (GGT), levels of aspartate aminotransferase (AST), alkaline phosphatase (ALP), alanine aminotransferase (ALT) and total cholesterol (CHOL) were assessed. These blood tests are part of standard clinical treatment and nursing procedures after patients are admitted to the hospital. This study protocol collects and analyzes these routinely collected clinical data that meet the study criteria. On the first day of hospitalization, qualified medical staff collected venous blood samples from the antecubital vein of each participant after a minimum 12-hour fast. To ensure sample stability and prevent clotting, blood was drawn into vacuum-sterile plastic or polypropylene tubes containing sodium or potassium ethylenediaminetetraacetic acid (EDTA-K2/K3) at a concentration sufficient to chelate calcium ions. The collected samples were promptly transferred to an independent laboratory for flow cytometry (FCM) analysis under strict quality control conditions. The following definitions were used in the study: Hyperuricemia: Serum uric acid > 420 μmol/L in men and > 360 μmol/L in women (20, 21).Hypertension: Systolic blood pressure ≥ 130 and diastolic blood pressure ≥ 85. Diabetes: Clinically diagnosed as diabetes by an endocrinologist. Marital status: Married, other (Widowed/Divorced) and Single.

Ultrasound imaging, as the main method for diagnosing fatty liver disease, is part of the routine assessment upon admission. It is usually scheduled shortly after blood samples are collected upon admission to ensure the relative synchrony of metabolic status assessment. The ultrasound imaging indicated the presence of fatty liver. After excluding other potential causes such as excessive alcohol consumption, malnutrition and hepatolenticular degeneration. The diagnosis of MAFLD was guided by the most recent international consensus on MAFLD diagnostic criteria (22). Specifically, in addition to having fatty liver, one of the following three conditions must be met: obesity or overweight, non-obese patients with metabolic abnormalities, or patients with type 2 diabetes mellitus. In our analysis, metabolic abnormalities were identified based on the following indicators: elevated blood pressure (with systolic blood pressure greater than or equal to 130 mmHg and diastolic blood pressure greater than or equal to 85 mmHg), reduced high-density lipoprotein cholesterol (HDL-C) (below 1.0 mmol/L for men and below 1.3 mmol/L for women), increased triglyceride levels (at or above 1.70 mmol/L) and established impaired fasting glucose (fasting glucose levels ranging from 5.6 to 6.9 mmol/L).

All data analyses were conducted using R version 4.4.1. Initially, general demographic data and test results were analyzed descriptively. Following this, QQ plots were utilized to examine the distribution of continuous variables for normality. For continuous variables approximating a normal distribution (Age, BMI, Education, Disease course, Age of onset, FBG, ALT, AST, ALP, GGT, HDL, ApoA1, SB (systolic blood pressure), DB (diastolic blood pressure.), the mean ± SD was reported; for those not conforming to normality (TC, Triglycerides, ApoB, Uric acid), the median (IQR) was used. Categorical data were summarized in terms of counts and proportions (%). In statistical analyses, data following a Gaussian pattern were evaluated with the t-test; for those deviating from this pattern, the Mann-Whitney U test (z) was utilized. Meanwhile, categorical data were assessed via the chi-square test (χ²). Multicollinearity was assessed using the variance inflation factor (VIF); variables with a VIF exceeding 10 were excluded from the analysis. Stepwise binary logistic regression of retained predictors (VIF <10) identified MAFLD risk/protective factors, with a significance level of p < 0.05 (two-sided) for variable retention. ROC curves were plotted using GraphPad Prism version 9.5 to evaluate the discriminatory ability of identified risk or protective factors.

This research involved 1,072 participants who had been diagnosed with bipolar disorder, of whom 562 (52.4%) had lean bipolar disorder and 510 (47.6%) were overweight or obese. The prevalence of MAFLD comorbidity among all patients was 41.8%, with rates of 76.3% among overweight or obese patients and 10.5% among lean patients. Significant differences were observed in laboratory results between overweight or obese patients and lean patients (p < 0.05), particularly in relation to hyperuricemia and hypertension. Overweight or obese patients exhibited worse performance on multiple biochemical measures, including higher BMI, blood glucose levels, liver function indicators, total cholesterol, triglyceride levels, and uric acid levels. In terms of sociodemographic characteristics, overweight or obese patients were older at disease onset, had longer disease duration, and were more likely to be married. Regarding gender distribution, the proportion of overweight or obese patients was slightly higher among males, while the proportion of lean patients was slightly higher among females (52%). In terms of bipolar disorder subtypes, the proportion of manic episodes was higher among overweight or obese patients (68%), whereas the proportion of depressive episodes was higher among lean patients (38%). (The detailed results are presented in Table 1 ).

The results indicated that patients with bipolar disorder comorbid with MAFLD were, on average, older, had a longer duration of illness, and exhibited an older age at disease onset compared to those without MAFLD. In terms of metabolic indicators, patients with bipolar disorder and comorbid MAFLD showed significantly higher levels of glucose, ALT, ALP, TC, ApoB, triglycerides, and GGT, while exhibiting slightly lower levels of ApoA1 and significantly lower levels of HDL. Notably, the prevalence of diabetes mellitus was significantly higher in patients with comorbid MAFLD (17% vs. 1.0%), and hypertension was more common (15% vs. 4%). Although hyperuricemia did not show a significant difference between the two groups (P = 0.10), its prevalence was higher in patients with MAFLD (32% vs. 22%). (For a comprehensive overview, refer to Table 2 ).

In our analysis, we designated the occurrence of MAFLD as the outcome variable and examined several factors that might influence this condition in individuals with lean bipolar disorder. These factors included blood glucose levels, ALP, ALT, GGT, total cholesterol, HDL, triglycerides, ApoA1, ApoB, sex, age, BMI, the length of illness, initial age of disease manifestation, marital condition, presence of diabetes, and hypertension. A variance inflation factor (VIF) > 10 was used to detect multicollinearity before performing stepwise binary logistic regression analyses to identify and exclude variables causing covariate issues. The conclusive analysis highlighted several key risk factors associated with MAFLD comorbidity among lean patients suffering from bipolar disorder. Fasting blood glucose (per 1 mmol/L: OR = 1.04, 95% CI 1.02 - 1.06) and female gender (OR = 1.08, 95% CI 1.03 - 1.13) were independent risk factors for MAFLD (both p < 0.01); the core strong risk factors were triglycerides (per 1 mmol/L: OR = 1.24, 95% CI 1.19 - 1.28) and diabetes (OR = 1.35, 95% CI 1.17 - 1.57); HDL showed a strong protective effect (per 1 mmol/L: OR = 0.82, 95% CI 0.76 - 0.88); although GGT was statistically significant (OR = 1.001, 95% CI 1.000 - 1.002, p = 0.048), the effect size was close to the null value (OR = 1), and its clinical significance was unclear (see Table 3 ).

Furthermore, the receiver operating characteristic (ROC) curves were utilized to evaluate the ability of the identified risk factors to differentiate between individuals with emaciated bipolar disorder who have metabolic-associated fatty liver disease (MAFLD) and those who do not. The findings are depicted in Figure 2 . The AUC values were 0.9008 for TG, 0.7708 for FBG, 0.7463 for HDL, 0.6602 for GGT, 0.5798 for Diabetes and 0.5147 for Sex. Detailed results for each variable, including the area under the curve (AUC), 95% confidence interval (CI), and corresponding p-value, are summarized in Table 4 . Figure 3 displays a predictive nomogram for MAFLD, highlighting how different factors contribute to the risk of developing MAFLD. The diagram indicates that factors such as blood glucose levels, triglyceride levels, GGT, the presence of diabetes, and being female are associated with an increased risk of MAFLD, whereas HDL appears to have a protective effect. It’s important to note that a wider range of scores for a variable in the nomogram signifies a greater influence on MAFLD risk. The graph reveals considerable score variation for blood glucose and triglycerides, emphasizing their critical role in influencing MAFLD risk.

As far as we are aware, this research marks the initial comprehensive evaluation of MAFLD co-occurring with lean BD within a major public medical facility in China. This research not only examines the prevalence of MAFLD, but also explores potential risk and protective factors associated with the co-morbidity of BD and MAFLD. There is currently a gap in this area of research, and our findings are expected to provide new insights into the association between BD and MAFLD, offering valuable guidance for clinical practice. The results of this study revealed that MAFLD is prevalent among patients with bipolar disorder (prevalence: 41.8%) and has a notable presence in lean bipolar disorder patients (prevalence: 10.5%). In the population of lean bipolar disorder patients with co-morbid MAFLD, several distinctive features were observed: an older average age, a later age of onset, and a longer duration of illness. Additionally, these patients exhibited significantly higher levels of blood glucose, total cholesterol (TC), triglycerides, apolipoprotein B (ApoB), and liver function markers, while ApoA1 levels were slightly lower and high-density lipoproteins (HDL) were significantly lower. Further analyses indicated that fasting blood glucose levels, triglyceride levels, gamma-glutamyl transpeptidase (GGT) levels, female sex, and diabetic status are significant risk factors for MAFLD, whereas high-density lipoprotein (HDL) levels serve as protective factors against MAFLD.

Recent research examining the frequency of MAFLD in urban Chinese populations reported an overall prevalence of 26.1%, with a prevalence of 4.1% among lean individuals (23). Analyzing data from the National Health and Nutrition Examination Survey (NHANES) spanning 1999 to 2018, the research found that among the 21,034 participants aged 20 and above who were in excellent health, the weighted prevalence of MAFLD stood at 29.9% (24). Notably, even in specific disease groups, such as patients with schizophrenia, NAFLD was identified in up to 32.9% of cases, with 25% prevalence observed in non-obese individuals (25). Studies focusing on the BD population have also shown a relatively high incidence of NAFLD, affecting approximately 28.4% of individuals (26). In our study, the prevalence of MAFLD was 41.8% among patients with bipolar disorder and 10.5% among lean patients, highlighting the significant burden of co-morbid MAFLD in this patient group. It is important to note that patients with BD often exhibit irregular eating habits, lack of exercise, disturbed sleep patterns, frequent hospitalizations, and the use of antipsychotic medications due to the nature of their condition. As per our information, this research marks the initial comprehensive evaluation of MAFLD co-occurring with lean BD within a major public medical facility in China (27). Thus, the bipolar disorder population may be more susceptible to MAFLD, with even lean individuals showing a higher prevalence than the general population. This underscores the necessity for MAFLD screening among individuals with bipolar disorder, particularly in the lean subgroup. Early identification and intervention can improve prognosis and reduce associated risks, such as liver and cardiovascular complications linked to MAFLD.