Associations between ultra-processed food (UPF) consumption and weight change and obesity risk among consumers of plant-based diets: a systematic review

Abstract

Interest in plant-based diets for health promotion continues to grow, emphasizing whole and minimally processed foods. Consistent with broader U.S. trends, plant-based consumers derive many calories from UPFs, whose impact on weight change, overweight, and obesity remains unclear. This review aimed to evaluate the association between UPF consumption and weight change, overweight, and obesity among consumers of plant-based diets. A systematic literature search was performed in Web of Science, PubMed, and Scopus databases, following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) protocol. This review analysed twelve studies published between 2020 and 2025 involving 206,727 participants aged 18 and older across six countries. There was a mixed association between UPF consumption and weight change, overweight, and obesity among plant-based consumers, with some suggestion that higher diet quality UPFs were inversely or not associated with body weight. Among the 12 studies included, 4 specifically examined the research question, while the rest separately reported baseline UPF intake and weight-related outcomes. The quality of evidence was weak. While the four studies addressing the research questionwere rated fair to good quality individually, it was not possible to draw definitive conclusions as the studies were observational and had a high risk of bias. Across the studies, there was also variability in consumption patterns and potential misclassification of plant-based consumers. Further research evaluating the quality of plant-based UPFs and UPF subcategories is warranted, as diet quality appears to be an influencing factor of the association.

Introduction

The rising prevalence of overweight and obesity has become a major trigger of chronic diseases globally (1,2). The literature consistently suggests that the consumption of ultra-processed foods (UPFs) is adversely associated with overweight and obesity (3–7), and chronic disease risk (8). UPFs are defined by NOVA as industrially manufactured foods that are formulated with multiple additives derived from little to no whole foods (9). Growth in UPF consumption has paralleled the growth in obesity prevalence, with intake significantly increasing from 53.5% to 57% of energy intake between 2001 and 2018 in the US (10). While many have posited that the poor nutritional quality of most UPFs underlies the adverse association between UPFs and adverse weight outcomes, emerging research suggests that nutritional composition may not fully explain adverse associations (11). Other potential biological pathways include gut microbiome alterations, gut-brain satiety signalling disruptions, and hormonal imbalances (12,13).

Nevertheless, emerging mixed evidence about the association between some UPFs and adiposity and diet quality is currently a point of contention(14). Despite widespread use of NOVA by researchers, critics argue that it is overly simplistic and fails to adequately distinguish between processed foods based on their actual nutritional value(15). Some researchers contend that NOVA’s broad categorization of UPFs may mislead consumers by grouping potentially health-promoting or nutritionally balanced products, such as plant-based meat and dairy alternatives, alongside genuinely unhealthy options(15).

Alongside the rise in UPF consumption and ongoing debates around NOVA classification systems, interest in plant-based diets has grown as a strategy for weight control, chronic disease prevention (16), and reduction in greenhouse gas emissions from animal-based foods (17). The term ‘plant-based diet’ describes a wide spectrum of dietary patterns which emphasize plant products and limit or exclude animal-derived products (18). As such, plant-based diets include a diverse range of dietary patterns that prioritize plant-derived foods with varying inclusion of animal-derived products (19), including plant-forward patterns like the Mediterranean and Lacto-vegetarian patterns (Table 1) (19–23).

The growing interest in plant-based dietary patterns has led to shifts in the food landscape, including growth in new plant-based UPF products (24) to address taste preferences, convenience, and shelf life (25,26). Plant-based dietary patterns are intended to consist mainly of whole and minimally processed foods, such as fruits, vegetables, nuts, and whole grains (27). However, these foods are often replaced by new plant-based products, including meat analogs, plant-based burgers, and alternative dairy products, most of which are classified as UPFs. Their production typically involves chemical modification and frequent use of cosmetic additives, among other ultra-processing methods (28,29). The emergence of these products, particularly those with improved nutrient profiles, may interest consumers trying to reduce their consumption of animal-based foods to promote health (30). However, plant-based UPFs have an uncertain relationship with health outcomes as compared to fruits, vegetables, legumes, grains, and non-tropical oils, which are rich in antioxidants, fiber, phytochemicals, and low in saturated fats, and generally cardioprotective (30,31).

The growth in plant-based UPFs is relevant to plant-based consumers, who, like the general population, consume a substantial proportion of calories from UPFs, which are nutritionally heterogeneous (24,32). Emerging research has found that some subgroups of UPF, such as whole grain breads and breakfast cereals, are not associated or inversely associated with disease risk (33). Conversely, other subcategories of UPF, like sugar-sweetened beverages, have been associated with increased overweight among vegetarians (34). Given these complexities, evaluating the overall association between UPF consumption and weight-related outcomes in plant-based consumers is essential.

Therefore, this systematic review (SR) aims to evaluate the current evidence on the associations between UPF consumption and its impact on weight change or obesity risk among consumers of plant-based diets.

Materials and Methods

Protocol

This SR adhered to the PRISMA (Preferred Reporting for Systematic Reviews and Meta-Analyses) guidelines (35). The Web of Science, PubMed, and Scopus were systematically searched from inception to February 2025, with an updated search conducted in March 2025 to capture new studies published between February and March 2025. The electronic search was carried out using the following keywords and combinations. (ultraprocessed food* OR ultra-processed food* OR ultra processed food* OR Processed* OR NOVA*) AND (plant-based diet OR vegetarian OR vegan). See the Supplementary section for the search strategy. For abstract screening, the following outcomes related to overweight, and obesity were considered: “obesity,” “BMI,” “body fat,” “adipose tissue,” “fat mass,” “fat-free mass,” “muscle mass,” “body composition,” and “waist circumference.”

Inclusion and exclusion criteria

The SR included observational studies and randomized controlled trials with the following criteria: (1) ultra-processed food exposure; (2) reporting body weight change, overweight, or obesity; (3) including adult (≥ 18 years old) plant-based consumers; (4) being published in English; (6) having some or all of the sample consume a plant-based diet, such as a vegan, vegetarian, or a Mediterranean diet. See Table 1 for a complete list of plant-based diets considered in this review.

Exclusion criteria were as follows: (1) gray literature, including book chapters, letters, and comments; (2) animal, in vitro, and cell culture studies; (3) review articles. Articles that met the inclusion criteria underwent full-text screening and were excluded based on the following: (a) review studies, (b) irrelevant outcomes reported, (c) full text published not in English. Details of the population, intervention, comparator, and outcome (PICO) are provided in Table S1.

Study selection

Two investigators (T.S.G. and Q.Y.) independently conducted abstract screening. T.S.G., Q.Y., and Y.M.H. conducted full-text reviews. Discrepancies for inclusion were resolved through consensus or discussion with the senior investigator (M.K.V.). A standardized process was employed based on the inclusion and exclusion criteria, considering the individual studies’ setting and population and evaluating exposures and outcomes.

Data extraction

Two investigators (T.S.G. and Q.Y.) used a standardized method to extract data from eligible articles: (1) first author’s name; (2) year of publication; (3) study location; (4) study design; (5) number of participants; (6) study sample size; (7) duration of follow up/intervention (8) dietary intake assessment (9) ultra-processed food exposure assessment methods; (10) UPF exposure estimation; (11) outcome measures related to overweight or obesity.

Quality assessment

Two investigators (T.S.G. and Y.M.H.) evaluated the quality of the included studies, and any discrepancies were resolved through discussion with the senior investigator (M.K.V.). The Newcastle–Ottawa Scale (NOS) (36) was used to assess cohort studies, and an adapted version of NOS (37) to evaluate cross-sectional studies separately (Supplementary Table S2 and Table S3). The NOS checklist consists of selection, comparability, and outcome. We used the threshold for converting NOS scoring to the Agency for Healthcare Research and Quality standards, yielding overall quality assessments of good, fair, or poor. The quality of the randomized controlled trial was assessed using the Cochrane Risk of Bias tool (38), as shown in Supplementary Tables S4 and S5.

Results

Study characteristics

The literature search yielded 2651 studies (1311 from PubMed, 1021 from Web of Science, and 319 from Scopus). Duplicates (596) were excluded, and 2055 articles underwent abstract screening, with 64 articles undergoing full-text screening. Out of the 64 fully reviewed articles, 52 were excluded, resulting in 12 articles included in this study (24,39–49), as shown in Figure 1. These 12 articles were published between 2020 and 2025 with 206727 participants, aged 18 years and above. To avoid double-counting participants in the UK Biobank, which was utilized in three studies (40,46,47), participants from Chang et al. (46) were counted as their research included the relevant population and spanned a broad timeframe from 2006 to 2010. A summary of the included studies is presented in Table 2. The studies were conducted across various countries, including France (41), The United Kingdom, (40,46,47) Brazil, (44,45), the United States (24,42,48), the United States and Canada (39), and Spain (43,49). Four studies were prospective cohort studies (39,40,47,49), six were cross-sectional studies (24,41,43–46), and two studies were randomized controlled trials (42,48). For dietary intake assessment, seven articles used 24-hour dietary records (40,41,44–48), four articles used food frequency questionnaires (39,43,45,48), and one article weighed food and beverages from each meal (42). UPF consumption was presented as a proportion of energy intake from UPF (% UPF _energy), (24,39,41,42,44,47), as a percentage by food weight contribution in the total daily food intake (% UPF _g/day), (40,43,46,48,49) or as daily UPF consumption frequency (times/day)(45). The classification of ultra-processed foods in all studies was based on the NOVA classification system.

Overall, there was a limited body of research exploring the association between UPF and weight change, overweight, or obesity among plant-based consumers and the quality of evidence was deemed low which limits the overall strength of conclusion. To best assess the quality of the evidence evaluating these associations, quality evaluation scales were applied in the context of the primary research question. Among the four studies that evaluated this research question, two (41,48) were rated fair quality and two were rated good quality (45,47). All the eight studies that reported baseline UPF and weight-related outcomes were rated poor (24,39,40,42–44,46,49), given that their primary outcomes did not align with our a priori outcome of interest.

Evidence from studies directly assessing the association between UPF consumption and obesity and weight change among plant-based consumers.

The four studies (41,45,47,48) that specifically evaluated this research question are shown in Table 2. Among these studies, one found no association between UPF consumption and obesity among plant-based consumers (47); another found that for vegans, complete avoidance of animal-based UPF was associated with weight reduction and low-fat plant-based foods were not associated with weight gain (48); obesity was a predictor of UPF consumption among vegetarians (41); and one reported that UPF consumption was associated with being overweight among vegetarians (45). The evidence provided a mixed association between UPF and overweight or obesity, and confidence in the quality of the studies was limited, as the evaluation of each study suggested they were fair to good quality and primarily observational, making it challenging to draw definitive conclusions. Additionally, the evidence derived from the four studies had high risk of bias, inconsistent results, small samples size, and other methodological limitations summarized in Tables 4–6.

In a study among vegetarian volunteers of the UK Biobank prospective cohort, Navratilova et al. (47) found that consumers of plant-based meat alternatives (PBMA) vs. non-PBMA consumers had no difference in obesity risk (RR = 0.88, 95% CI: 0.55–1.42, p= 0.60) after adjusting for confounding factors such as age, sex, BMI, ethnicity group, Townsend quintiles group, physical activity, smoking status, and proportion of total energy intake from UPFs. Additionally, the total UPF intake was 46% among PBMA consumers and 40% among non PMBA Consumers. Notably, controlling for total % kcal from UPF (which was marginally higher in PBMA vs. Non-PBMA) showed no difference in risk of obesity. Similarly, a secondary analysis of RCT noted that consumption of plant-based processed foods, regardless of NOVA category, was not associated with weight gain (48). Additionally, reducing ultra-processed animal-based foods led to weight loss: every 120 g/day decrease was linked to 1 kg weight loss over 16 weeks (p = 0.001). After adjusting for energy intake, every ~156–158 g/day reduction corresponded to 1 kg loss (p = 0.02) (48). The study highlighted that low-fat plant-based foods were not associated with weight gain regardless of the degree of processing (48). Among the two cross-sectional studies (41,45) that addressed the research question, Dos Santos et al. (45) found that higher UPF intake (Q5 vs. Q1) was associated with an increased prevalence of overweight (28.1% vs. 22.7%) after adjusting for sex and age among Brazilian vegetarian adults. Although vegetarians had a lower average UPF intake and overweight prevalence than the general population, UPF exposure still contributed to excessive weight gain or maintenance (45). Similarly, Gehring et al.(41) found that obesity was a predictor of UPF consumption after controlling for the type of vegetarian diet in the NutriNet-Santé cohort. UPF consumption was higher among participants with obesity vs. normal weight (β = 0.098, p = 0.001). The higher intake of salty snacks and biscuits in the subgroup of vegetarians and vegans resulted in higher UPF consumption, suggesting that diet quality may be associated with these findings. Notably, this study assessed the association in reverse (examining obesity as a predictor of UPF intake rather than the opposite) (41).

Studies providing baseline, unadjusted or indirect evidence of association between UPF and obesity and weight change among plant-based consumers.

Across the eight studies (24,39,40,42–44,46,49), that did not directly address this research question, the evidence was generally of poor quality and a weak or no consistent signal observed (See Table 3). The available data consisted only of unadjusted baseline results, rather than a comprehensive analysis targeting the outcome of interest. While these preliminary findings suggest a weak potential signal of association worthy of further investigation, the lack of direct evaluation and adjustment for confounding variables limits the strength of evidence these studies can provide for answering our research question. While both the UK Biobank (40) and the Adventist Health Study 2 (AHS-2) (39) cohort studies found unadjusted positive associations between UPF consumption and weight outcomes, a prospective study of patients with metabolic-dysfunction-associated steatotic liver disease(49) found no intergroup differences in changes in BMI between baseline and 6 months by tertiles of reduction of UPF intake (T1: −1.4 kg/m²; T2/T3: −1.1 kg/m², p= 0.697) among those with high adherence to the Mediterranean diet. Similarly, cross-sectional studies (24,43,44,46) reported null associations between UPF and obesity among plant-based consumers. Interestingly, vegans have a lower rate of overweight and obesity, consuming fewer calories from UPF than their vegetarian counterparts (46). Finally, a randomized crossover trial involving 20 individuals at an inpatient facility examined ad libitum consumption of two diets low in UPFs (26% and 32% kcal) over 2 weeks, comparing a plant-based low-fat diet with an animal-based low-carbohydrate diet (42). The plant-based low-fat diet resulted in a total weight loss of 1.09 ± 0.32 kg (p = 0.003), which was not significantly different (p =0.15) from the weight loss observed with the low-carbohydrate diet (1.77 ± 0.32 kg, p =0.0001) (42). There were no differences in weight loss outcomes between low-carbohydrate and low-fat diet arms when UPF intake was low, between 26-32% of kcal (42).

Discussion

This systematic review found limited and mixed evidence of the association between UPF consumption and weight change, overweight, or obesity among plant-based consumers, with studies providing weak quality evidence and inconsistent findings. While few studies have specifically examined this question, existing research suggests that only certain categories (i.e., refined grains, sugar-sweetened beverages, sweets, and desserts) were adversely associated with obesity (41), whereas others (i.e, PBMAs) appeared protective (47), providing preliminary support for diet quality being a mediator in the relationship between plant-based UPF and weight-related outcomes among plant-based consumers. This preliminary evidence suggests that plant-based UPFs with higher diet quality may be inversely or not associated with body weight. However, due to the scarcity of evidence, high risk of bias, limitations of dietary assessment tools, misclassification of plant-based consumers (24,44), and variability in UPF consumption patterns compared to minimally processed foods, no definitive conclusions can be drawn. Further research is needed to clarify these relationships and assess the role of diet quality in mediating these associations.

Among the 4 studies (41,45,47,48) that evaluated our research question, there is some suggestion that there is heterogeneity in the association between UPF intake and obesity risk among plant-based consumers. This heterogeneity appears to stem from several key factors. First, the sources of UPFs may be the major factor as UPF drives from animal foods were associated with weight gain vs plant-based UPFs (48). Secondly, methodological limitations and differences among the studies, including FFQ design, outcome measurement and inadequate adjustment for potential confounding factors may explain the discrepant findings. For example, a study of Brazilian vegetarian adults noted a positive association between UPFs (assessed via qualitative FFQ) and prevalence overweight (calculated from self-reported height and weight) (45). Thirdly, overall diet quality may mediate the association as suggested by findings that PBMA consumption vs non-PBMA consumption among vegetarians showed no difference in obesity risk in adjusted analyses (47), and UPFs intake was associated with unhealthy (but not healthy) plant-based diet scores across all dietary groups (41).

Notably, the prospective study by Navratilova et al. (47) examined the wider health impact of consuming PBMAs among vegetarians and found no difference in obesity risk between vegetarians who consumed plant-based meat alternatives (PBMAs) vs. those who did not despite higher total energy intake (8593 kJ/day vs 8090, p<0.001), which may be attributed to the higher overall diet quality of PBMAs consumers, physical activity and health-conscious consumers. Previous research suggests that plant-based consumers often prioritize health, and thus, they may be more likely to choose healthier UPF subcategories (20,44). This assertion is further supported by data from the National Health and Nutrition Examination Survey (NHANES), which showed that plant-based consumers had stable UPF intake, obesity prevalence, and increasing diet quality (52.1 to 55.8 p <0.001) between 1999 to 2020 (24). Similarly, Leitão et al.,(44) found that UPFs were important contributors to protein adequacy, and no associations between overweight and obesity were observed in a healthy sample of vegans.

The mixed associations between some subgroups of UPFs and body weight outcomes are consistent with emerging evidence showing similar associations between UPF subgroups and morbidity and mortality. Previous research has shown that ultra-processed breakfast cereals and bread were not associated with the risk of multimorbidity, cancer, cardiometabolic disease (50), cardiovascular outcomes (51), or incident type 2 diabetes mellitus (33,52). Interestingly, Li et al. (53) found that higher consumption of UPFs was associated with shorter leukocyte telomere length (LTL). Nevertheless, in subgroup analyses, consuming 1 serving per day of certain UPF subclasses (e.g., breakfast cereals and vegetarian alternatives) was associated with a protective, significantly longer LTL in both males and females. While the mechanisms require further research, the authors speculate that plant-derived UPFs have antioxidant-rich substances and trace elements that could affect health. Additionally, ultraprocessed whole grains have been inversely associated with T2D risk (54). The reason for heterogeneity in the association with subcategories of plant-based UPFs may be due to differences in the contents of fiber, minerals, vitamins, and phytochemicals, and could potentially be driven by residual confounding associated with greater health consciousness of individuals consuming these foods (54).

Another potential explanation for the mixed association is that plant-based consumers tend to eat fewer calories and more low-energy-dense foods, which can help promote satiety and weight management (55). Moreover, UPF intake was generally lower among plant-based consumers than the national average of 67%(56), which could attenuate associations with obesity and potentially suggest threshold effects. A recent study by Pan et al. (57) observed that the impact of higher UPF consumption on the risk of overweight or obesity was relatively weak among the overall adult population in China, where the UPF consumption was low (5.82 to 15.53 g/day) compared to the average in the US and Europe.

Lastly, understanding these associations requires not only examining specific plant-based UPFs but also considering the underlying mechanisms through which UPFs contribute to obesity and weight gain. Several biobehavioural pathways linking UPFs to weight gain have been proposed (e.g., appetite dysregulation, gut microbiome alterations) (58). However, existing studies in plant-based populations have not directly assessed these mechanisms. Future research should incorporate objective measures, such as satiety hormones, eating behaviours, and metabolic responses, to clarify whether observed weight outcomes are driven by UPF properties or mediated by these biobehavioural pathways.

Mechanistic path of UPF

The included studies did not evaluate the mechanistic pathways underlying the association between UPF consumption and weight change, overweight, or obesity. However, recent studies have highlighted potential mechanisms, including poor nutrient density, high energy density, low satiety potential, hyperpalatability – which predisposes individuals to overconsume and may adversely impact the gut microbiota (59). It is established that energy dense foods high in fat, salt and sugar combined with nutrient-deficient dietary patterns are associated with the prevalence and development of obesity and related conditions through inflammatory and oxidative stress pathways (59–61). Detailed studies on these mechanistic pathways are needed to inform public health advice and create national dietary guidance on UPFs (61).

Study strengths and limitations

Limitations of this systematic review include methodological weaknesses in included studies, such as reliance on qualitative food frequency questionnaires (45) or single 24-hour recalls (24) and misclassification plant-based consumers (24,44) which limits the robustness of findings. Additionally, most studies did not specifically examine the research question or control for confounders (24,39,40,42–44,46) and did not account for the mediating effect of diet quality. More research across diverse samples, using multiple 24-hour recalls or longitudinal tracking and considering diet quality is needed to better understand the consistency and strength of the association between total and UPF subgroup consumption and weight-related outcomes among plant-based consumers. Lastly, comparing these associations across countries where UPF consumption amounts and types vary can provide a broader understanding of its impact on health outcomes, including among populations whose UPF consumption represents a higher percentage of total caloric intake (e.g., at levels meeting or exceeding the national average). Understanding these patterns could provide more targeted insights into the health implications of high UPF consumption and inform public health recommendations tailored to these populations.

Nevertheless, several strengths are worth mentioning. This systematic review adhered to best practice criteria for searching, selecting, and evaluating studies for inclusion and assessing the quality of evidence. The included studies were diverse, from multiple countries, encompassing a wide age range, and featuring various study designs with differing levels of UPF consumption. Dietary survey data were drawn from large samples, such as the UK Biobank from the United Kingdom (40,46,47) and the AHS-2 cohort recruited from Seventh-day Adventist churches in the United States and Canada (39).

Conclusion

The growing interest in plant-based diets has transformed the food landscape, leading to the emergence of many plant-based products classified as ultra-processed. In the studies reviewed, there was a mixed association between UPF consumption and weight change, overweight, or obesity among plant-based consumers. While evidence was mixed and inconclusive, there were some suggestions that higher diet quality UPFs were inversely or not associated with body weight, consistent with previous studies where some higher quality ultra-processed subgroups were favorably associated with health outcomes, while others, like sugar-sweetened beverages, were strongly associated with adverse health outcomes (62). The quality of evidence was weak, with only four studies addressing this research question rated between fair (41,48) to good quality (45,47). Nevertheless, due to the scarcity of evidence, the observational nature of the studies, high risk of bias, self-administered 24-h recall questionnaire (47) or single 24-hour recalls (24), misclassification plant-based consumers (24,44), and variability in UPF consumption patterns compared to minimally processed foods, no definitive conclusions can be drawn. As the food landscape continues to change and dietary guidelines shift their emphasis to plant-based dietary patterns, more well-designed studies are needed to comprehensively evaluate whether UPF consumption—particularly potentially healthier UPF subgroups— is associated with adverse weight outcomes.

KEY TEACHING POINTS.

1. Few studies evaluated the association between UPF consumption and weight-related outcomes among plant-based consumers and the quality of evidence was weak.

2. The studies that evaluated the research question found a mixed association between ultra-processed food (UPF) consumption and overweight and obesity among plant-based consumers.

3. Diet quality may partly to mediate the association between UPF consumption and weight-related outcomes in the population.

Appendix

Table 1.

Types of plant-based diets and definitions.

Types of plant-based diet	Definition
Vegan	Exclude all animal foods and dairy products. Does not require the consumption of whole foods or restrict fat or refined sugar (20,63)
Lacto-ovo-vegetarian	Excludes meat but includes dairy products, eggs, and honey (20,63)
Lacto-vegetarian	Exclude eggs, meat, seafood, and poultry, and include milk products (20,63)
Ovo-vegetarian	Exclude meat, seafood, poultry, and dairy products but include milk (20,63)
Semi-vegetarian	Varies widely. Involve consuming red meat, poultry, or fish no more than once per week, no red meat at all, or limiting red meat to once per week and poultry to five times per week (21).
Mediterranean diet	Like a whole plant-based diet, this approach permits small amounts of chicken, dairy products, eggs, and red meat once or twice a month. It also encourages the consumption of fish and olive oil, with no restrictions on fat intake (20)

Figure 1.

PRISMA Flow Diagram

Table 2:

Evidence from studies directly assessing the association between UPF consumption and obesity or weight change among plant-based consumers.

Author	Country	Study design	N participant	Study population	Dietary intake assessment	Ultra-processed Food Classification	UPF Exposure estimation	Outcome-related Weight change/Obesity	NOS or CRB quality assessment
Kahleova et al.,(48)	United Staes	Randomized Control trial	244 participants	Overweight adults	Dietary records	The classification of ultra-processed foods was based on the NOVA classification system	UPF intake (g/day)	The randomized controlled trial involved the vegan group (avoid animal products and minimize oils) and a control group (No dietary changes). Reducing animal UPF led to weight loss: every 120 g/day decrease was linked to 1 kg weight loss over 16 weeks (p = 0.001). After adjusting for energy intake, every ~156–158 g/day reduction corresponded to 1 kg loss (p = 0.02)	Fair quality
Navratilova et al.,(47)	United Kingdom	Prospective Cohort	3342 self-reported Vegetarians	UK Biobank participants	self-administered 24-h recall questionnaire	The classification of ultra-processed foods was based on the NOVA classification system	% of energy from UPFs. This study directed evaluated the research question	Participants were categorized into two groups: vegetarian PBMA consumers and vegetarian PBMA nonconsumers. UPF intake was higher among PBMA vegetarian consumers vs. non-PBMA vegetarian consumers (46% vs 40%, p=0.001), and without adjustment, the distribution of body weight differed between the 2 groups (BMI 24.9 vs 24.2, p = 0.001). PBMA intake was not associated with obesity risk (RR = 0.88, 95% CI: 0.55–1.42, p = 0.60) after adjusting for confounding factors.	Good quality
Gehring et al. (41),	France	Cross-sectional	21212 participants men =26.9% and women 73.1%	The Nutri Net-Sante cohort age > 18y, included meat-eaters = 19,812, pes-co-vegetarians = 646, vegetarians =500, and vegans =254. Subsample without meat eaters =1400.	24-h dietary records (more than 3000 items): Completion of 3 non-consecutive 24-h dietary records for 1 weekend day and 2 weekdays. Validated by comparison with 24-hour urine and blood biomarkers	The classification of ultra-processed foods was based on the NOVA classification system. NOVA group 4 which includes plant-based meat and dairy substitutes. All food and beverages were reviewed by 3 dieticians and 5 researchers, specialists in nutritional epidemiology in 4 NOVA groups.	% of energy from UPFs.	UPF consumption was significantly higher for vegetarians (37%) and vegans (39.5%) than meat-eaters (33%), p<0.001. Across quartiles of UPFs consumption among the vegetarian subgroup (n=1400), vegetarians (β = 0.024, p=0.002) and vegans (β = 0.042, p< 0.001) had a significantly higher consumption of UPFs using pesco-vegetarians as reference. There were significant demographic differences across quartiles of UPF consumption among vegetarian subgroups (pesco-vegetarians, vegetarians, and vegans). In a significant multivariable-adjusted association obesity was a predictor of UPF consumption after controlling for duration, age, and sociodemographic characteristics. Obesity vs normal weight was associated with 0.098 greater UPF intake across quartile of UPF intake.		Fair quality
Dos Santos et al., (45)	Brazil	Cross-sectional	925 vegetarians, Female =80.8% and Male =19.2%	The participants are aged 18 years and above from the Brazilian Study on Health, Diet, and Nutrition of Vegetarians	Frequency questionnaire (FFQ). Examined 29 UPF items	The classification of UPF was based on the NOVA classification system	Daily UPF consumption(times/day)	This study examined the association between UPF consumption and overweight among vegetarians. The daily UPF consumption among participants ranged from 0 to 13.1 times per day, with a median of 1.8 times. Age-sex adjusted analysis revealed a significant exposure-response relationship between UPF consumption and overweight, with a prevalence ratio (PR) of 1.09 (95% CI: 1.01-1.17) after adjusting for age (PR 1.03, 95% CI: 1.02-1.04). When comparing consumption quintiles, vegetarians in Q1 (0.00-0.81 times/day) had a 22.7% overweight prevalence, while those in Q5 (3.47-13.12 times/day) showed a 28.1% prevalence.	Good quality

KEY: NOVA refers to the food classification system as defined by Martinez-Steele et al, (28,29) PF: ultra-processed food (includes foods and beverages), BMI: Body Mass Index [weight (kilograms)/height (meters)²], FFQ: Food frequency Questionnaire, WK: Week, %: Percentage, DASH: Dietary Approaches to Stop Hypertension, NAFLD: Non-alcoholic fatty liver disease patients, Kg: Kilogram, NOS: Newcastle Ottawa Scale, CRB: Cochrane Risk of bias. NHANES: National Health and Nutrition Survey

Table 3:

Studies providing baseline, unadjusted or indirect evidence of association between UPF and obesity or weight change among plant-based consumers.

Author	Country	Study design	N participant	Study population	Dietary intake assessment	Ultra-processed Food Classification	UPF Exposure estimation	Outcome-related Weight change/Obesity	NOS or CRB quality assessment
Hall et al., (42)	United States	Randomized controlled trial	20 Volunteers Male = 11 Female = 9	Participants were admitted as inpatients to the Metabolic Clinical Research Unit at the NIH Clinical Center where they resided in individual rooms. Age 18–50 years; weight stable (<±5% over past 6 months); body mass index ≥20 kg/m2; body weight ≥53 kg; no signs of arrhythmia and without diabetes.	Weighing remaining food and beverages from each meal. The nutrient and energy intake were calculated using nutrition software.	The study compared a Low-fat (plant-based) diet formulated with low UPF (26%) and a low-carbohydrate (anima-based) diet with low UPF (32%) per the NOVA classification system.	% energy from UPFs.	The study compared a Low-fat (plant-based) diet formulated with low UPF (26%) and a low-carbohydrate (animal-based) diet with low UPF (32%). Two-week weight loss outcomes were comparable between the 2 diet groups. The plant-based low-fat diet resulted in a total weight loss of 1.09 ± 0.32 kg (p = 0.003), which was not significantly different (p =0.15) from the weight loss observed with the low-carbohydrate diet (1.77 ± 0.32 kg, p =0.0001) There were no differences in weight loss outcomes between low-carbohydrate and low-fat diet arms when UPF intake was between 26-32% of kcal.	Poor quality
Tu et al., (40)	United Kingdom	Prospective cohort	121 300 participants Female 56.4%	UK Biobank participants. Age 40 to 69 years with adherence to dietary patterns (Mediterranean-stye, DASH, or plant-based diet)	24-hour dietary recall questionnaires of 238 common food and beverage items. As a means of two or more 24-hour dietary recalls.	The classification of ultra-processed foods was based on the NOVA classification system. NOVA group 4 represents food and beverages.	UPF intake (% by food weight).	The median UPF intake (% by food weight) of total food mass was 11.5. Across quintiles of Mediterranean diet score at baseline, participants with higher vs. lower (Q5 vs. Q1) adherence to the Mediterranean diet consumed lower UPFs (9.8% vs. 11.5% by weight, p<0.001) and had lower BMI (25.8 vs. 27.5, p < 0.001). Similarly, across quintiles of DASH diet score (Q5 vs. Q1), participants with higher vs. lower adherence consumed lower UPFs (8.4 vs. 11.5 kcal, p<0.001) and had lower BMI (25.1 vs. 26.7, p < 0.001). Across quintiles of UPF intake, participants in Q5 vs. Q1 consumed higher UPF (25.1 vs. 11.5, p < 0.001) and had higher BMI (27.8 vs. 26.7, p < 0.001)	Poor quality
Orlich et al., (39)	North America	Prospective Cohort	77437 men and women participants	AHS-2 cohort recruited from Seventh-day Adventist churches in the United States and Canada. Non-vegetarians and vegetarians (comprising vegans, lacto-ovo vegetarians, pesco vegetarians, and semi vegetarians)	FFQ: dietary intake was assessed at baseline by a previously validated self-administered quantitative FFQ of >200 food items	The classification of ultra-processed foods was based on the NOVA classification system	% of energy from UPFs.	UPF contributed 27.4% median of total dietary intake. Participants across UPF intake quintiles (Q1 vs. Q5), had a higher BMI (25.24 vs. 28.91, p < 0.001). All analyses unadjusted for confounding factors	Poor quality
Garcia et al.,(49)	Spain	Prospective Cohort	70 participants	FLIPAN participants following the Mediterranean diet, aged 40 and 60	validated 143-item semi-quantitative FFQ	The classification of ultra-processed foods was based on the NOVA classification system	%UPF intake (g/day)	Participants were divided into three groups (T1-T3) based on changes in UPF consumption over 6 months. Mediterranean diet adherence was inversely correlated with UPF intake. There were no significant differences in BMI by tertiles of reduction of UPF intake across all groups (T1: −1.4 kg/m2; T2/T3: −1.1 kg/m2, p= 0.697)	Por Quality
Monserrat-Mesquida et al., (43)	Spain	Cross-sectional	100 Non-alcoholic fatty liver disease patients. Men = 56% and women =43.5%	NAFLD patients from the Balearic Islands, with mean age of 51.8 years	Semiquantitative FFQ: Dietary intakes were obtained using a validated 148-item FFQ	The classification of ultra-processed foods was based on the NOVA classification system	% UPF Intake (g/day)	Participants were classified into tertiles based on their adherence to the Mediterranean diet. Individuals with low vs. high adherence to the Med diet had higher % UPF consumption. (33.3 vs. 12.7, p<0.001). However, no association was observed between UPF consumption and BMI, with mean BMI values of 32.7, 32.5, and 33.4 for low, moderate, and high adherence groups, respectively	Poor quality
Leitao et al., (44)	Brazil	Cross-sectional	774 participants Male= 17.7% Female = 82.3%	The participants were 24-35 years old from the Vegan Eating Habits Evaluation Survey with self-reported adherence to a vegan diet.	24-hour dietary recalls.	The classification of ultra-processed foods was based on the NOVA classification system	% of energy from UPFs.	UPF intake was 13.2% of food intake in the vegan population. In both men and women there was no association between UPF and mean BMI	Poor quality
Sullivan et al., (24)	United States	Cross-sectional	51,698 participants	The participants aged 20 years and above from NHANES data who consumed plant-based food.	Completion of at least one 24-hour dietary recall	The classification of ultra-processed foods was based on the NOVA classification system	% of energy from UPFs	Plant-based consumers were defined as individuals obtaining ≥50% of their total protein from plant sources.There was no association between UPF consumption and weight-related outcomes in plant-based consumers. Over 2 decades, UPF intake remained stable among plant-based consumers, consistently comprising over half of their total energy intake (50.7–57% kcal; p =0.34) while UPF consumption among non-PB consumers significantly increased from 51 to 55.4 (p <0.001). The proportion of individuals with normal and overweight increased, while obesity rates remained stable among PB consumers.	Poor quality
Chang et al., (46)	United Kingdom	Cross-sectional	45,057 flexitarians, 4932 pescatarians, 4119 vegetarians and 159 vegans	UK Biobank participants’ mean age 58.2	24-h recalls	The classification of ultra-processed foods was based on the NOVA classification system	% of energy from UPFs	The study found that UPF consumption varied across dietary groups, but no clear linear trend linked higher UPF intake to higher BMI classifications. Vegans, who consumed 50.3% of their energy from UPF, had the lowest rates of obesity (6.2%) and overweight (31.4%) compared to vegetarians who consumed 51.7% of UPF and were 31.8% overweight and 12.3% obese (p<0.001)	Poor quality

KEY: NOVA refers to the food classification system as defined by Martinez-Steele et al, (28,29) UPF: ultra-processed food (includes foods and beverages), BMI: Body Mass Index [weight (kilograms)/height (meters)²], FFQ: Food frequency Questionnaire, WK: Week, %: Percentage, DASH: Dietary Approaches to Stop Hypertension, NAFLD: Non-alcoholic fatty liver disease patients, Kg: Kilogram, NOS: Newcastle Ottawa Scale, CRB: Cochrane Risk of bias. NHANES: National Health and Nutrition Survey

Table 4:

Newcastle-Ottawa Scale (NOS) for Cohort

	Selection				Comparability	Outcome assessment
	Representativeness of the Exposed Cohort (Max.1)	Selection of the Non-Exposed Cohort (Max.1)	Ascertainment of Exposure (Max.1)	Demonstration that Outcome of Interest was Not Present at Start of Study (Max.1)	Comparability of Cohorts Based on Design or Analysis (Max.2)	Assessment of Outcome (Max.1)	Was Follow-Up Long Enough for Outcomes to Occur (Max.1)	Adequacy of Follow-Up of Cohorts (Max.1)	Score (Max.9)	Quality Score
Orlich et al., (39)	1	1	0	0	0	1	1	1	5	Poor
Tu et al., (40)	1	1	0	0	0	1	1	1	5	Poor
*Navratilova et al., (47)	1	1	1	1	2	1	1	1	9	Good
Garcia et al., (49)	1	1	1	0	0	0	1	1	5	Poor

Key: Scoring Interpretation

Good Quality: 3 or 4 stars in the Selection domain AND 1 or 2 stars in the Comparability domain AND 2 or 3 stars in the Outcome domain

Fair Quality: 2 stars in the Selection domain AND 1 or 2 stars in the Comparability domain AND 1 or 2 stars in the Outcome domain

Poor Quality: 0 or 1 star in Selection domain OR 0 stars in Comparability domain OR 0 or 1 star in Outcome domain

^*The star symbol represents studies that directly address the research question

Table 5:

Newcastle-Ottawa Scale (NOS) adapted for Cross-sectional StudiesKey: Scoring Interpretation

	Selection				Comparability	Outcome assessment
	Representativeness of the Sample (Max.1)	Sample Size (Max.1)	Non-Respondents (Max.1)	Ascertainment of Exposure (Risk Factor) (Max.1)	Control for Confounding Factors (Max.2)	Assessment of Outcome (Max.2)	Statistical Test (Max.1)	Score (Max.9)	Quality score
*Gehring et al., (41)	1	1	0	0	1	2	1	6	Fair
Monserrat-Mesquida et al., (43)	0	0	0	0	0	2	1	3	Poor
Leitao et al., (44)	0	0	1	0	0	0	1	2	Poor
Sullivan et al., (24)	1	1	1	1	0	2	1	7	Poor
Chang et al., (46)	1	1	1	1	0	2	1	6	poor
*Dos Santos et al., (45)	1	1	1	0	1	2	1	7	Good

Key: Scoring Interpretation

Good Quality: 3 or 4 stars in the Selection domain AND 1 or 2 stars in the Comparability domain AND 2 or 3 stars in the Outcome domain

Fair Quality: 2 stars in the Selection domain AND 1 or 2 stars in the Comparability domain AND 1 or 2 stars in the Outcome domain

Poor Quality: 0 or 1 star in Selection domain OR 0 stars in Comparability domain OR 0 or 1 star in Outcome domain.

^*The star symbol represents studies that directly address the research question

Table 6:

Cochrane Risk of Bias for Randomized Controlled studies

	Selection bias		Performance bias	Detection bias	Attrition bias	Reporting bias
	Random sequence generation	Allocation concealment	Blinding of Participants and Researchers	Blinding of outcome assessment	Incomplete outcome Data	Selective Reporting	Other bias	Quality Score
Hall et al, (42)	Low risk	High risk	Unclear risk	Low risk	Low risk	Low risk	Unclear risk	Poor
*Kahleova et al,(48)	Low risk	Low risk	High risk	Low Risk	Low Risk	Low Risk	Low Risk	Fair

Key: Thresholds for Converting the Cochrane Risk of Bias Tool to AHRQ Standards (Good, Fair, and Poor)

Good quality: All criteria met (i.e. low for each domain) Using the Cochrane ROB tool, it is possible for a criterion to be met even when the element was technically not part of the method. For instance, a judgment that knowledge of the allocated interventions was adequately prevented can be made even if the study was not blinded, if EPC team members judge that the outcome and the outcome measurement are not likely to be influenced by lack of blinding.

Fair quality: One criterion not met (i.e. high risk of bias for one domain) or two criteria unclear, and the assessment that this was unlikely to have biased the outcome, and there is no known important limitation that could invalidate the results.

Poor quality: Two or more criteria listed as high or unclear risk of bias

^*The star symbol represents studies that directly address the research question

Data Availability Statement

The original contributions presented in the study are included in the article and Supplementary Materials. Further inquiries can be directed to the corresponding author.