NAFLD, predominantly caused by metabolic syndrome, is closely associated with obesity, insulin resistance and hyperlipidemia (1). Over the past four decades, the incidence of NAFLD has steadily increased (2, 3), currently affecting 25% of the global adult population (3). The disease prevalence is about 30% in Asia (4), the US, and South America, 24% in Europe, and 13% in Africa (2, 5–7), making it the most widespread chronic liver disease worldwide (2). An estimated 20% of non-alcoholic steatohepatitis (NASH) patients progress to cirrhosis, and the risk of hepatocellular carcinoma in NASH patients surged 7.7·times between 2002 and 2016, from 2.1 to 16.2% (8). In the US, it is expected that NASH medical expenses per patient will jump from $3,636 to $6,968 between 2020 and 2039 (9). Similarly, in Japan, the annual healthcare cost for NASH ranged from 322,206 to 340,399 yen per patient between 2011 and 2017 (10), imposing a substantial economic burden. Therefore, attention must be focused on early disease detection in primary healthcare settings.

Obesity, over nutrition, a high-calorie diet, and a sedentary lifestyle contribute to the accumulation of liver fat (11, 12), and are crucial risk factors for NAFLD (13). The regulatory mechanism of NAFLD is connected with metabolism, heredity, intestinal microorganisms, and other factors (11, 14, 15). At present, the management of NAFLD is centered around reducing insulin resistance and limiting oxidative stress (13). Treatment strategies are founded on lifestyle management, such as modifying diet and increasing physical activity, with the intent of controlling weight and managing risk factors pertinent to metabolic syndrome (16, 17). Consequently, adhering to a balanced diet and maintaining a healthy lifestyle have become pivotal in treating and delaying the progression of this disease (18, 19).

Fruits and vegetables are plant-based foods, rich in dietary fiber, which can help maintain the balance of intestinal flora, reduce inflammation, and decrease fat accumulation in the liver. Moreover, fruits and vegetables are abundant in antioxidants that neutralize free radicals and diminish oxidative stress damage to the liver. The antioxidants (20) and anti-inflammatory compounds (21) in fruits and vegetables enhance insulin sensitivity, accelerate beta-oxidation, and inhibit new fat production (22). As a result, it has been hypothesized that an intake of fruits and vegetables correlates with a lower prevalence of NAFLD (23). However, the relationship between fruit and vegetable intake and NAFLD risk remains a subject of debate. Several observational studies suggest that higher dietary vegetable intake is associated with a lower NAFLD risk (24, 25). Yet, some research indicates that there is no such relationship (23, 26). A similar controversy exists in regard to the correlation between fruit intake and the risk of NAFLD (26–31). Even though the role of vegetable and fruit intake in NAFLD has drawn considerable public attention, no meta-analyses have demonstrated a correlation between vegetable and fruit consumption and NAFLD risk. Therefore, we undertook this meta-analysis to summarize the results of observational studies regarding the association between vegetable and fruit consumption and the risk of NAFLD.

This systematic review and meta-analysis statement followed Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA), and the protocol was registered with PROSPERO (ID: CRD42023460430).

This represents the inaugural meta-analysis study investigating the connection between the intake of vegetables and fruits and the risk of NAFLD. A total of 11 studies involving 493,682 participants were included in the analysis, and the results suggest a negative correlation between the consumption of vegetables and fruits and the risk of NAFLD.

Dietary fiber, which is abundant in fruits and vegetables, is a type of short-chain fatty acids (SCFAs) produced by the fermentation of intestinal microorganisms in the gastrointestinal tract. These SCFAs, like propionic acid and butyric acid, have numerous benefits, including maintaining the integrity of the intestinal barrier, reducing the inflammatory reaction in the liver, regulating appetite, and maintaining glucose balance at the systemic level. This is all helpful in maintaining the energy balance of the liver, improving insulin sensitivity, and regulating liver lipid metabolism (34, 35). Furthermore, dietary fiber also enhances satiety, thus promoting calorie restriction (36). Additionally, fruits and vegetables are rich in antioxidants, including vitamin C, vitamin E, beta-carotene, polyphenols, which neutralize free radicals and mitigate the detrimental effects of oxidative stress on the liver (37–41). It should also be noted that NAFLD is frequently accompanied by an inflammatory response. Therefore, inhibiting inflammation is crucial to alleviating NAFLD symptoms. Fruits and vegetables are rich in polyphenols and flavonoids, which possess anti-inflammatory properties and can reduce the severity of inflammation in the liver (41, 42).

It is observed that in some of the included studies, the relationship between intake levels of vegetables (23, 26, 28, 29, 32) and fruits (23, 28, 30–33) and the prevalence of NAFLD were not consistent with the conclusion in this meta-analysis. In particular, the proportion of related studies on fruit intake with inconsistent conclusions with this meta-analysis is relatively high. This discrepancy may be attributed to variations in disease diagnosis methods, study populations, adjustment factors, and dietary assessment methods.

The diagnosis of NAFLD differed across the included studies, and differences in the definition of fatty liver and the degree of medical diagnosis may have led to discrepancies in conclusions. Chan's study (27) measured intrahepatic triglyceride content (IHTG) using 1H-MRS within 8 weeks after the baseline visit of included participants, and the result showed that higher levels of vegetable and fruit intake were associated with a lower prevalence of NAFLD. Liu et al. (28) used B-ultrasound as the diagnostic basis, Giraldi et al. (32) used bright liver pattern and ultrasound features of liver steatosis as the diagnostic basis, and Noureddin et al. (29) identified eligible NAFLD patients through Medicare claims. Their results showed that vegetable and fruit intake levels were not associated with the prevalence of NAFLD. It is worth discussing that in the study with a large sample size conducted in South Korea by Kim and Shin (25), the incidence risk of NAFLD in the female population was related to vegetable and fruit intake, while in the male population, only fruit intake was found to be related. In the overall population, the incidence risk of NAFLD was associated with the intake of both vegetables and fruits. However, this conclusion may have significant bias because the diagnosis of NAFLD is based on the Fatty Liver Index (FLI), using a cutoff value of 60. However, in the study by Kim et al., 46% of the participants were of normal weight or underweight, and the accuracy of FLI in diagnosing NAFLD in lean NAFLD patients was low. This is because patients with low body mass index and NAFLD would not be able to reach the FLI limit for diagnosing NAFLD in the absence of increased GGT or triglycerides (included in the FLI calculation) (43). Furthermore, the accuracy of the critical value of 60 for diagnosing NAFLD is low in Asian populations, including in South Korea, because after the study by Kim et al., the ideal critical value for South Koreans was described to be equal to 29 (44–46).

Furthermore, results may vary between populations. Asians may have a stronger preference for leafy vegetables such as spinach and cabbage, while Europeans and Americans tend to consume vegetables like corn, squash, potatoes, onions, and broccoli, which contain higher levels of starch. This dietary difference could explain the diverse findings regarding the correlation between vegetable and fruit intake and NAFLD among participants in Asia, Europe, and America. Notably, there is limited literature from Europe and the Americas, necessitating cautious interpretation of these results and future confirmation through additional relevant studies. Additionally, we observed different results between males and females (23, 25, 31, 33), which may be attributed to notable differences in their dietary patterns (33). Studies have shown that females tend to increase their intake of fruits and vegetables more than males (47, 48), and there are sex-specific disparities in fatty acid oxidation and regulation of liver de novo lipogenesis (DNL), with males inhibiting DNL less rapidly than females, leading to a shift in cellular metabolism from fatty acid oxidation to esterification (49). Since gender-specific studies are relatively scarce, further clinical data are required to validate these conclusions.

Results may vary when adjusting factors in a study. After adjusting for social and economic status and other factors in Li et al.'s study (33), green leafy vegetables (GLV) were negatively correlated with NAFLD. However, further adjustment for BMI eliminated this negative correlation. We found that adjusting for BMI did not significantly alter the results in multiple studies, including Noureddin et al.'s study (29). Although GLV intake is negatively correlated with NAFLD in normal/overweight individuals, obesity-related metabolic complications such as hyperlipidemia and insulin resistance may significantly increase liver lipids, resulting in decreased insulin sensitivity. These complications cannot be regulated by lipid metabolism and GLV intake (50). Additionally, some reports show that obese individuals significantly underestimate their dietary intake in self-recording or interview evaluation (51), which may explain why some studies did not observe the relationship between GLV intake and NAFLD after adjusting for BMI. Liu et al. (28) investigated the relationship between dietary patterns and NAFLD in Chinese adolescents, finding no association between fruit and vegetable intake and the occurrence of NAFLD. Teenagers usually have excellent physiological functions and efficiently absorb nutrients from food. Meanwhile, teenagers are more prone to unhealthy eating patterns that could affect their BMI and contribute to differences in conclusions. Therefore, additional studies are required to confirm the reliability of this conclusion.

In studies examining the intake levels of total vegetables and total fruits (25) and their association with the prevalence of NAFLD, results show a negative correlation. However, research by Liu et al. (28) and Giraldi et al. (32) indicates no relationship between vegetable and fruit intake levels and the prevalence of NAFLD. Guo et al.'s study (26) found significant results for fruit intake but not for vegetable intake. These studies included vegetables and fruits within dietary patterns, which are typically composed of independent or interactive foods and complex nutrient combinations that affect human metabolism. Therefore, it is challenging to exclude the synergistic effects of nutritional foods on NAFLD. Furthermore, not all types of vegetables and fruits are associated with a reduced risk of chronic diseases, as they contain different components and bioactive phytochemicals (52). While vegetables are generally considered low-carbohydrate foods and those with high dietary fiber levels may reduce the risk of NAFLD (33), excessive intake of starchy vegetables might increase blood glucose and insulin resistance, which is detrimental to NAFLD patients (53). Similarly, fruits contain natural fructose, and high fructose intake increases hepatic de novo lipogenesis (DNL), reduces fatty acid β-oxidation (FAO), and leads to fatty acid deposition (54). Excessive fructose intake can also promote the development of NAFLD. Additionally, the type of fructose—natural fructose from fruits vs. industrial fructose—might lead to different research conclusions. In Tajima et al.'s study (23), fruit intake was negatively correlated with the fatty liver index in the elderly, whose dietary fructose mainly came from fruits (55). Despite inquiries about fruit intake in studies, younger individuals might mistakenly count industrial fructose and fruit juice consumption as fruit intake. This could mean that, among the younger population, the harmful effects of industrial fructose and soft drinks on NAFLD might outweigh the protective effects of fruits on NAFLD (55). These factors may explain the inconsistencies between the conclusions regarding fruit intake in the included studies and the results of this meta-analysis.

In addition, there are some factors that cannot be avoided in the included studies. For example, in Giraldi et al. (32), there may be large differences in samples and high variability, resulting in wide confidence intervals, and outcomes may be affected. In Li et al.'s study (33), fruit intake was recorded for both NAFLD patients and the control group. However, the results were not significant, possibly due to differences in the study population, adjustment factors, dietary assessment methods, or recall bias in participants' reporting of fruit intake, leading to the omission of ORs. This does not meet the inclusion criteria for studies on fruit intake and NAFLD prevalence in our research and consequently contributes to a degree of selection bias in this study. Furthermore, the assessment questionnaire itself has self-reporting bias and subjectivity, and different assessment methods have different and limited contents, such as the lack of eating methods of vegetables and fruits, the choice of types of vegetables and fruits, the combination of food and the time of eating, which may also lead to different outcomes.

Our study has various strengths. Firstly, it is the first meta-analysis to investigate the association between vegetable and fruit intake and incidence of NAFLD, utilizing large sample sizes. Secondly, in most of the studies included in the meta-analysis, the incidence of major NAFLD risk factors was controlled. Finally, we conducted subgroup analysis and stratification analysis of confounding factors to explore the sources of heterogeneity in the association between vegetable and fruit intake and NAFLD events.

Our study has a few limitations. Firstly, the included studies in our analysis encompassed case-control studies. As these studies employed food assessment questionnaires to estimate dietary intake, we cannot entirely rule out the possibility of measurement errors resulting from under- or over-reporting of food group intake due to participants' subjective judgments or memory biases. This could introduce recall and selection bias. Secondly, even though most studies adjusted for potential risk factors of NAFLD, residual confounding is always a concern in all observational studies. Finally, ORs were pooled from the highest and lowest intake levels, but intake levels were not always consistent across studies. Because of limited data, we were unable to include all studies in the dose-response analysis.

In conclusion, this systematic review and meta-analysis provide evidence that higher fruit and vegetable intake is linked to a lower risk of NAFLD. However, given that the relationship between vegetable intake and NAFLD incidence varies across different populations (age, sex, and ethnicity), types of vegetables, and fruits, more high-quality prospective studies are desired to further elucidate this connection. Additionally, there are studies suggesting that excessive fruit intake may actually promote the development of NAFLD.

The original contributions presented in the study are included in the article/Supplementary material, further inquiries can be directed to the corresponding authors.

RW: Data curation, Formal analysis, Investigation, Methodology, Software, Visualization, Writing – original draft. RY: Data curation, Formal analysis, Methodology, Software, Writing – review & editing. JJ: Investigation, Project administration, Software, Writing – review & editing. FL: Data curation, Investigation, Software, Writing – review & editing. HZ: Investigation, Software, Visualization, Writing – review & editing. ZC: Conceptualization, Methodology, Resources, Supervision, Writing – review & editing. HW: Project administration, Supervision, Validation, Writing – review & editing. SY: Conceptualization, Methodology, Project administration, Supervision, Validation, Writing – review & editing. JL: Conceptualization, Funding acquisition, Methodology, Project administration, Resources, Supervision, Validation, Writing – review & editing.