Non-alcoholic fatty liver disease (NAFLD) is a spectrum of liver disorders characterized by excessive hepatic fat accumulation in the absence of excessive alcohol consumption or any other secondary causes diagnosed by liver biopsy (1). Considering the global obesity and metabolic syndrome epidemics, NAFLD is becoming the leading cause of chronic liver damage and cannot be understated due to its rising prevalence, accounts for 25% worldwide (2) and 15%-40% in Asia (3), epidemiologically. The disease continuum of NAFLD spans from simple steatosis or fatty liver, progressing to non-alcoholic steatohepatitis (NASH) accompanied by varying degrees of fibrosis, and can eventually culminate in cirrhosis (4). Identifying individuals at risk early in the disease trajectory is crucial for timely intervention and prevention of its progression.

Although it’s widely known that NAFLD is commonly associated with obesity, it is reported that 40% of the NAFLD patients worldwide (5) and 17.5% of those in Chinese population are non-obese (6). As they often overlook the need for clinical evaluation for NAFLD and are likely to be overlooked in risk assessments for the condition, this population may pose greater challenges than obese individuals. Existing studies have proposed limited indices, including AIP (Atherogenic index of plasma) and RLP-C (Remnant lipoprotein cholesterol), as potential predictors for risk assessment in non-obese population (7, 8). First proposed by Wang et al. in a cross-sectional study involving 9,602 Chinese participants, the Zhejiang University (ZJU) Index proved to be a proficient tool for screening NAFLD (AUC=0.822) (9, 10). Subsequent researches have confirmed the association between the ZJU index and the incidence of NAFLD in both Chinese and international populations, demonstrating superior predictive accuracy compared to the aforementioned indices (11, 12). However, its efficacy in non-obese populations remains invalidated.

Therefore, this study aims to assess the relationship between the ZJU index and the incidence of NAFLD in non-obese populations and explore whether this relationship exhibits sex differences. Additionally, we aim to compare the effectiveness of the ZJU index with some preexisting indicators (AIP, RLP-C) in predicting the onset of NAFLD.

In this longitudinal study, we observed that the ZJU index is a significant risk factor for new-onset NAFLD in non-obese populations, showing superior predictive efficacy compared to existing indices AIP and RLP-C. Notably, there was a significant interaction between sex and new NAFLD diagnoses, with the correlation being more substantial in the female subgroup. In clinical practice, the ZJU index presents a valuable tool for assessing NAFLD risk in non-obese populations, aiding in the early identification of individuals at high risk for NAFLD and facilitating early interventions to improve their prognosis.

The ZJU index, initially proposed by Wang et al. in a cross-sectional study involving 9,602 Chinese participants who underwent hepatic ultrasound examination to diagnose NAFLD, is recognized as a proficient tool for screening NAFLD in the Chinese population (AUC=0.822) (9). Several prior studies have validated that the ZJU index effectively predicted NAFLD in both domestic and international populations. In a cross-sectional study involving 19,804 Chinese adults, the ZJU index, fatty liver index, hepatic steatosis index, lipid accumulation product, and visceral adiposity index of each participant were calculated. The ZJU index showcased superior diagnostic accuracy in identifying NAFLD with an AUC of 0.925 (95% CI 0.919-0.931), outperforming four other indexes. This superior diagnostic performance was consistent in age and sex subgroups, showing the ZJU index as a reliable and efficient tool for NAFLD detection (11). In another cohort involving 28,729 Chinese adults without fatty liver disease at baseline, the ZJU index was found to be significantly associated with an elevated risk of developing fatty liver disease over a median follow-up of 3.01 years. As the ZJU quartiles increased, the HRs increased correspondingly, with 4.87 (4.24-5.59) for women and 6.23 (5.56-6.98) for men (12). Moreover, validation in North American settings was evident in a study involving 107 obese women, which revealed an AUC of 0.742 for the ZJU index, emphasizing its applicability in predicting NAFLD among this population (10). However, few studies have examined its correlation with NAFLD in non-obese individuals. Our findings on the ZJU index align with these past studies, underscoring its reliable predictive value for new-onset NAFLD in non-obese group. Notably, while previous studies predominantly focused on obese or overweight participants, our research breaks new ground by delving into the relationship between the ZJU index and NAFLD in non-obese individuals.

Furthermore, when we conducted subgroup analyses on various factors, we noted a significant interaction in the sex subgroup with respect to the ZJU index. There was a larger effect size of the association between the ZJU index and new-onset NAFLD risk in women than men. The result may be attributed to the distinct metabolic profiles between males and females. Women generally exhibit a gynoid gluteo-femoral subcutaneous fat distribution, while men tend to have an android visceral fat distribution (14). On the other hand, estrogen, prevalent in females, provides protection against NAFLD (15). In this process, estrogen-related receptor alpha (ERRα) is a critical modulator in the downstream of estrogen/ERα signaling, particularly influencing hepatic very low-density lipoprotein (VLDL) assembly and secretion, with its dysfunction leading to hepatosteatosis and potential progression to NAFLD, especially in men and postmenopausal women (16). In additional analyses, we employed the age of 50 years as a surrogate marker for the onset of menopause. he results demonstrated a significant interaction between sex and the association between the ZJU index for new-onset NAFLD in the population aged over 50. Conversely, this interaction was not significant in the population under 50 years. In the research by Li et al. (12), the findings revealed a higher HR for men, at 6.23 (5.56-6.98), compared to women, who had an HR of 4.87 (4.24-5.59). This conclusion stands in contrast to the results of our study. However, in further subgroup analyses, both the ≤50 and >50 age groups exhibited higher HRs for women compared to men. Notably, in the >50 age group, the gender difference in HRs was more pronounced, aligning with the conclusions of our study, which suggested that the ZJU index may be a more effective predictor of NAFLD in postmenopausal women.

Previous research indicated that the development of fatty liver is associated with disruptions in lipid metabolism, anomalies in glucose processing, and impaired functionality of adipose tissues (17, 18). The components of ZJU index, including ALT, AST, TG, and FPG, serve as representative markers for hepatic metabolic irregularities (19, 20). Besides, in a study of 3,329 Chinese adults, the ZJU index revealed a profound link with insulin resistance (IR). With ascending ZJU quartiles, significant increases were observed in parameters like BP, waist circumference, and cholesterol levels, while a decline was noted in HDL-C levels. The ZJU index stood out not only for signifying an amplified risk of IR in its highest quartile but also for its superior predictive accuracy, boasting ROC curve values of 0.833 in males and 0.788 in females (21). Insulin resistance acts as a risk factor for NAFLD (22). It paves the way for NAFLD both directly, by amplifying de novo lipogenesis, and indirectly, by boosting the free fatty acid flux to the liver due to diminished lipolysis inhibition (23). Consequently, the ZJU index emerges as an optimized predictive model that encapsulates various forecasting factors.

NAFLD is progressively identified in non-obese individuals, and these cases can sometimes have more adverse outcomes than their obese counterparts, including a quicker progression to cirrhosis (24, 25). The progression of fibrosis occurs more rapidly in lean individuals with NAFLD compared to those who have a higher BMI (24). Genetic variations in Patatin-like phospholipase domain-containing protein 3 (PNPLA3), transmembrane 6 superfamily 2 (TM6SF2) or other genes associated with steatosis could possibly explain the association between lean NAFLD and the increased risk of future development of severe liver disease (26). Moreover, early detection and intervention can be more challenging in these non-obese patients than in those who are obese. Miao et al. (8) using the same source data as ours investigated the relationship between Remnant lipoprotein cholesterol (RLP-C) and NAFLD, revealing that higher RLP-C levels were associated with an increased risk of NAFLD, with HRs for 1.6 and 1.3 in high and middle RLP-C groups. Li et al. (7) involved 16,173 participants and use the atherogenic index of plasma (AIP) to predict NAFLD. Over a span of 5 years, 14.4% of the non-obese individuals developed NAFLD. With increasing AIP values, the risk of new-onset NAFLD rose significantly, especially evident in those within the lower BMI range. In this study, when comparing the ROC curves of three indicators, the AUC values highlighted the ZJU index’s superior performance over RLP-C and AIP in predicting NAFLD. This underscores the potential of the ZJU index as a more effective tool for detecting NAFLD risk among the non-obese population.

Publicly available datasets were analyzed in this study. This data can be found here: the Dryad data repository at http://datadryad.org/ under the doi: 10.5061/dryad.1n6c4.14.

The studies involving humans were approved by the ethics committee of Beijing Anzhen Hospital. The studies were conducted in accordance with the local legislation and institutional requirements. Written informed consent for participation was not required from the participants or the participants’ legal guardians/next of kin in accordance with the national legislation and institutional requirements.

KZ: Methodology, Writing – original draft. YY: Methodology, Writing – original draft. HG: Formal analysis, Writing – original draft. LM: Validation, Writing – original draft. RL: Data curation, Writing – review & editing. TZ: Data curation, Writing – review & editing. YW: Visualization, Writing – review & editing. ZZ: Visualization, Writing – review & editing. WC: Project administration, Writing – review & editing.