Sex Differences in the Association Between Ultra-Processed Food Consumption and NAFLD: An Analysis of KNHANES 2013–2021 Data

Non-alcoholic fatty liver disease (NAFLD) is defined as excessive hepatic fat accumulation in the absence of significant alcohol intake. Importantly, NAFLD is the most common chronic liver disorder worldwide, prevalent in approximately more than 30% of adults [1] and especially more prevalent in individuals with obesity or type 2 diabetes mellitus than in the general population [2]. In recognition of its strong metabolic basis, the terminology for non-alcoholic fatty liver disease (NAFLD) has evolved. The term metabolic-associated fatty liver disease (MAFLD) was previously proposed to better capture the metabolic nature of the disease, and more recently, a multisociety Delphi consensus has endorsed the term metabolic dysfunction-associated steatotic liver disease (MASLD), providing a non-stigmatizing and inclusive nomenclature that highlights metabolic dysfunction as the key driver of disease pathogenesis [3]. NAFLD can progress from simple steatosis to non-alcoholic steatohepatitis and hepatic fibrosis and is associated with an elevated risk of cardiovascular disease and other complications. Alarmingly, young adults with NAFLD have a five-fold higher risk of mortality than those without the condition [4].

Ultra-processed foods (UPFs), classified as Group 4 in the NOVA food classification system [5], are industrial formulations produced through extensive mechanical and chemical processing, often incorporating additives to enhance shelf life, flavor, and texture. Over the past several decades, global UPF consumption has risen to strikingly high levels; in the United States, these products now provide approximately 53–58% of total daily energy intake among adults [6], with similar or higher proportions reported in other high-income countries. In some Western populations, they constitute more than half of all calories consumed [7]. This increasing dietary share has drawn attention to its potential health effects. High UPF consumption has been consistently linked to obesity, dyslipidemia, hypertension, cardiovascular disease, type 2 diabetes, and certain cancers [8]. Cohort studies have further demonstrated higher rates of metabolic syndrome and all-cause mortality in individuals with greater UPF intake than in those with lower UPF intake [9,10].

Genetic predisposition and lifestyle factors, particularly diet, are central to NAFLD pathogenesis. Emerging evidence suggests that higher UPF consumption is associated with an increased risk of NAFLD. A recent meta-analysis estimated a 22% higher risk of NAFLD among individuals in the highest UPF intake category than those among in the lowest [11]. However, substantial between-study heterogeneity was observed (I² = 78.5%), likely reflecting differences in study design, participant characteristics, food classification methods, and NAFLD diagnostic criteria. In addition, a hospital-based cross-sectional study from Israel found no significant association between UPF intake and NAFLD [12]. These inconsistencies highlight the need for studies that account for population-specific dietary patterns and metabolic susceptibilities.

Racial and ethnic disparities in the prevalence of NAFLD have also been reported [13,14]. However, most prior evidence has originated from Western populations, and data from Asian cohorts are scarce. Cultural and genetic differences may influence both UPF consumption patterns and metabolic responses to such diets, limiting the generalizability of Western findings to Asian populations. Moreover, NAFLD prevalence and risk factors differ according to sex [15], suggesting that sex-stratified analyses can reveal distinct associations between UPF intake and NAFLD.

This study aimed to examine the relationship between UPF consumption and NAFLD stratified by sex in a nationally representative sample of Korean adults using 2013–2021 Korea National Health and Nutrition Examination Survey (KNHANES) data. Toward this goal, a sex-stratified analyses of the association between NALFD and UPF consumption, determined from 24 h dietary recall data, were performed to assess potential effect modifications.

In this analysis of a large, nationally representative cohort of Korean adults spanning 2013 to 2021, greater consumption of UPFs was significantly associated with a higher prevalence of NAFLD. Compared with those in the lowest quartile, participants in the highest quartile of UPF intake had approximately 24% higher odds of NAFLD, even after adjusting for demographic, socioeconomic, and lifestyle factors. A clear dose–response relationship was observed, with NAFLD prevalence increasing from approximately 19% in the lowest UPF quartile group to 24% in the highest UPF quartile group. Importantly, this association differed by sex: while there was no significant association after multivariable adjustment in men, women in the highest UPF quartile group had approximately 50% greater odds of NAFLD compared to those in the lowest quartile group. These findings highlight sex-specific vulnerability, suggesting that the harmful effects of UPF consumption on liver health may be stronger in women, a critical pattern that would have been obscured without a stratified analysis. This investigation is one of the few large-scale studies in an Asian population. In addition, to our best knowledge, this study is the first to explore potential sex-specific differences in the association between UPF and NAFLD, providing evidence to inform tailored dietary guidelines and public health strategies.

Our results are consistent with previous evidence indicating that UPF consumption adversely affects metabolic health. Recent investigations have reported a positive association between high UPF intake and risk of NAFLD. In a prospective analysis of a Tianjin cohort in China (n = 16,000), individuals in the highest category of UPF consumption had a significantly higher risk of incident NAFLD [19]. Similarly, a recent systematic review and meta-analysis showed a significant association between UPF intake and NAFLD [20]. Moreover, a systematic review by Grinshpan et al. [21] summarized the limited and heterogeneous evidence, noting that only half of the available studies—mostly cross-sectional and conducted in Western populations—reported significant associations between UPF and NAFLD. Our findings extend the current evidence base by providing large-scale, nationally representative data from an Asian population, thereby addressing one of the key gaps identified in that review.

These associations are biologically plausible as UPFs can increase spontaneous energy intake owing to their high palatability and energy density, promoting weight gain and central adiposity, with consequent insulin resistance [22]. In addition, added sugars and refined fats may enhance hepatic de novo lipogenesis, thereby augmenting intrahepatic triglyceride accumulation [23] and contributing to NAFLD development and progression. Collectively, these mechanisms suggest that UPF-centered dietary patterns may increase the risk of hepatic fat accumulation, even among younger adults. In contrast, a cross-sectional study found no association between energy intake from UPF and fat mass [12,24], although the limited sample size of this study may have reduced the statistical power and increased the likelihood of type II errors.

To the best of our knowledge, no previous investigation has explicitly stratified the association between UPF and NAFLD according to sex, limiting direct comparisons. The current study found a positive association between UPF intake and NAFLD among females. Notably, this female-specific association contrasts with reports of related outcomes in which UPF effects appeared stronger in men. For example, Lee et al. found that the association between obesity and higher UPF consumption was stronger in men than in women [25]. Collectively, these findings suggest that the health effects of UPFs may vary by target organ and pathogenic pathways and are modulated by sex-related biological susceptibility.

Premenopausal women are relatively protected against visceral adiposity and insulin resistance through estrogenic effects; consequently, NAFLD is generally less prevalent even among women with obesity than among men [26]. Meanwhile, the risk of NAFLD among postmenopausal women is similar to or even exceeds that of men [27]. In addition, compared to men, women have higher peripheral and hepatic insulin sensitivity and show greater suppression of hepatic glucose production, conferring a metabolically favorable profile [28,29]. The hepatic de novo lipogenesis response to carbohydrate overfeeding is also lower in women than in men [30], consistent with a stronger intrinsic defense against hepatic fat accumulation. However, the energy surplus and poor nutrient composition of UPF diets can offset these protective effects. Women generally have smaller body size and lower basal metabolic rates than men [31]; therefore, a caloric excess may impose a proportionally greater storage burden. When the subcutaneous buffer capacity, supported in part by estrogen, is exceeded, fat partitioning can shift toward visceral and ectopic (hepatic) depots, potentially increasing the risk of NAFLD.

In contrast, men tend to have higher baseline propensities for visceral fat accumulation and NAFLD [26,32]; thus, the additional risk attributable to UPF may appear less pronounced on a relative scale. Collectively, these mechanisms may provide a biologically plausible explanation for the stronger association between UPF intake and NAFLD observed in women in our study.

This study leveraged a large, nationally representative sample of Korean adults from the KNHANES 2013–2021 (n = 24,587), with the full application of survey weights, stratification, and clustering, thereby maximizing statistical power and enhancing external validity. Beyond the overall association, sex-stratified analyses identified meaningful heterogeneity, revealing a stronger association between UPF and NAFLD in women that could have been obscured in the aggregate models. These findings provide a rationale for sex-tailored clinical and public health messaging. Finally, the evidence for the dose–response relationship was demonstrated on both categorical and continuous scales—through quartile-based logistic regression with tests for linear trend and complementary continuous modeling of %UPF in multivariable logistic regression—strengthening the robustness and biological plausibility of the findings.

This study has limitations that warrant consideration. First, its cross-sectional design precludes causal inferences. Second, the exposure is determined from a single 24 h dietary recall that may not have captured usual intake and is vulnerable to recall error. Moreover, applying NOVA categories to the Korean diet, where mixed dishes and composite recipes are common, introduces potential misclassification. Third, despite adjustments for key demographic, socioeconomic, lifestyle, and dietary factors, residual confounding factors remain. Unmeasured or imprecisely measured determinants such as genetic susceptibility, menopausal status, hormone therapy, detailed alcohol use patterns, sleep, and psychosocial stress can influence both UPF consumption and NAFLD risk. The magnitude of the association should be interpreted with caution given these limitations. Longitudinal studies with repeated dietary assessments, imaging-based endpoints, and finer control of sex-specific and genetic factors are needed.

Our findings indicate that reducing UPF intake may contribute to the prevention of NAFLD and promotion of liver health. Given the recent transition of nomenclature from non-alcoholic fatty liver disease (NAFLD) to metabolic dysfunction–associated steatotic liver disease (MASLD) [3], our results also align with the updated conceptual framework, highlighting UPF consumption as a modifiable metabolic risk factor under both terms. Clinicians should consider counseling that prioritizes eating a minimally processed diet and gradually reduces the frequency and portion size of high-contributing UPFs. Where feasible, dietitian referral, use of brief food diaries, and basic label-reading instructions can improve feasibility through patient-tailored strategies. Given the stronger association observed in women, sex-informed counseling such as focused education and follow-up for women in their midlife, a period marked by hormonal transition and fat redistribution, appears to be clinically reasonable.

Future research should use prospective cohorts with repeated dietary assessments to establish temporality and quantify dose–response relationships with greater precision. Interventional work spanning pragmatic nutrition counseling to structured UPF reduction trials should determine whether lowering UPF intake improves liver fat, ideally with imaging-based endpoints and related biomarkers. Studies should also explicitly examine effect modification by sex, menopausal status, and sex hormones and incorporate stratification by genetic susceptibility.

In this analysis of a nationally representative sample of Korean adults from the 2013–2021 KNHANES, higher UPF intake was associated with a higher prevalence of NAFLD. A consistent dose–response association was observed in both quartile-based logistic models and continuous analyses, with the association more pronounced in women after multivariable adjustment. Although the cross-sectional nature of the study limits its ability to establish causality, the convergence of categorical and continuous analyses, along with biological plausibility, supports the reduction in UPF consumption as a feasible approach to improve liver health. Prospective imaging-based cohort studies and intervention trials, along with analyses that consider sex and menopausal status, are needed to establish causality, quantify effect size, and inform sex-aware clinical guidance and policies.

The following abbreviations are used in this manuscript:

Conceptualization, B.S.K. and J.E.P.; methodology, B.S.K. and J.E.P.; software, N.G.K.; validation, B.S.K. and J.E.P.; formal analysis, N.G.K.; investigation, B.S.K. and J.E.P.; resources, B.S.K. and J.E.P.; data curation, B.S.K., N.G.K. and J.E.P.; writing—original draft preparation, B.S.K. and J.E.P.; writing—review and editing, B.S.K. and J.E.P.; visualization, B.S.K., N.G.K. and J.E.P.; supervision, B.S.K. and J.E.P.; project administration, B.S.K. and J.E.P. All authors have read and agreed to the published version of the manuscript.

Ethical review and approval were waived for this study due to the use of publicly available, de-identified data from the KNHANES, which contains no personally identifiable information.

Patient consent was waived due to the use of publicly available, de-identified data from the KNHANES.

The data used in this study are publicly available and can be freely downloaded from the KNHANES website (https://knhanes.kdca.go.kr/↗ (accessed on 1 August 2025)).

The authors declare no conflicts of interest.

This research received no external funding.