Results

Fatima Cody Stanford; Chris Taylor; Sarah Booth; Deanna M. Hoelscher; Cheryl A.M. Anderson; Andrea Deierlein; Teresa Fung; Christopher Gardner; Edward Giovannucci; Hollie Raynor; Sameera Talegawkar; Deirdre Tobias; Laural Kelly English; Molly Higgins; Emily Callahan; Nicole Reigh; Ramkripa Raghavan; Gisela Butera; Nancy Terry; Julie Obbagy

NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

Stanford FC, Taylor C, Booth S, et al. Dietary Patterns with Ultra-Processed Foods and Growth, Body Composition, and Risk of Obesity: A Systematic Review [Internet]. Alexandria (VA): USDA Nutrition Evidence Systematic Review; 2024 Nov.

Cover of Dietary Patterns with Ultra-Processed Foods and Growth, Body Composition, and Risk of Obesity: A Systematic Review

Dietary Patterns with Ultra-Processed Foods and Growth, Body Composition, and Risk of Obesity: A Systematic Review [Internet].

Show details

Contents

< Prev Next >

Results

Literature search and screening results

The articles included in this systematic review were identified from two literature searches (Appendix 2). The primary literature search was conducted to identify articles examining dietary patterns with UPF and growth, body composition, and risk of obesity. That literature search yielded 3920 search results after the removal of duplicates (see Figure 2). Dual-screening resulted in the exclusion of 3335 titles, 363 abstracts, and 191 full-texts articles from that search. The results of a second literature search that was conducted for another systematic review^† on dietary patterns and growth, body composition, and risk of obesity was manually searched. NESR analysts identified 17 articles from that literature search that examined dietary patterns with varying amounts of UPF. Reasons for full-text exclusion are in Appendix 3. The body of evidence included 48 articles in infants and young children up to age 24 months, n=5¹^–⁵; children and adolescents, n=25⁶^–³⁰; adults and older adults, n=16³¹^–⁴⁶; and individuals during pregnancy and/or postpartum, n=2⁴⁷^,⁴⁸

Image of flowchart of literature search and screening described fully in the text.

Figure 2

Literature search and screening flow chart.

Infants and young children up to age 24 months

Five articles met inclusion criteria that examined the relationship between dietary patterns with varying amounts of UPF consumed by infants and young children up to age 24 months and growth, body composition, and risk of obesity (also see Table 2). All 5 articles reported data from prospective cohort study designs.¹^–⁵

Description of the evidence

Population

Sample size of studies ranged from 449 to 14,989 participants, conducted in 4 countries (United Kingdom (n=2), Norway (n=1), Scotland (n=1), and United States (n=1)). Two articles from the Southampton Women’s Survey cohort were included but examined dietary patterns (exposure) and outcomes at different ages. Other cohorts represented in the included articles were: Norwegian Mother, Father and Child Cohort Study; Nurture Observational Study; and Growing Up in Scotland (GUS) Study.

Studies examined dietary patterns consumed at ages ranging from 4 months up to 2 years, with most examining 6 to 12 months. Three of the articles reported some information on the racial or ethnic background of participants as follows: 1) 4% non-white; 2) 65.2% non-Hispanic Black; 3) 94% White. One study, conducted in the United States, enrolled majority (~59%) low-income participants. Socioeconomic position (SEP) was reported in 4 of the 5 articles, with maternal education being the most used proxy.

Intervention/exposure and comparator

All studies used food-frequency questionnaires to assess habitual consumption of individual foods and/or food groups, administered by researchers or self-administered by caregivers. All but one study collected diet at more than one time point, which had a single assessment at age 19-24 months. Approaches used to examine dietary patterns were factor/cluster analyses (n=3) or index/score analyses (n=2).

Outcomes

Weight, height, and/or body composition of participants was measured in-person using standard protocols, instruments and trained staff in all but one article, which collected information via parent-report.² Cut points and measures used to categorize or classify participants’ outcome status were: BMI percentile, BMI z-score, or other ratios of weight-to-height (e.g., BMI reference curves vs. WHO vs. Cole’s method^*). The following outcomes were reported across the body of evidence:

BMI, BMI z-score, or weight-for-length z-score¹^,⁴^,⁵
Adiposity (e.g. fat mass),³^,⁴
Risk of overweight and/or obesity,¹^,²
Weight/weight gain,¹^–⁵
Length/length gain³

Synthesis of the evidence

None of the studies and/or dietary patterns were designed specifically to examine varying amounts of UPF within dietary patterns and their association with the outcomes of interest. The mean intakes of UPF in the study populations were not reported.

Across all studies, the results varied for both the direction and magnitude of effect estimates from analyses comparing participants consuming dietary patterns that reflected higher compared to lower amounts of foods that are typically classified as UPF.¹^–⁵ Two articles reported no associations between dietary patterns varying in UPF and risk of overweight or obesity.¹^,² Three studies reported no association with dietary patterns varying in UPF and BMI z-score, weight-for-length z-score, and/or BMI.¹^,²^,³^,⁴ One study observed that a dietary pattern with fewer UPF was associated with lower weight-for-length z-score, but other dietary patterns were not associated with weight-for-length z-score when compared to the dietary pattern with the most UPF.⁵ One study observed a positive association between dietary patterns higher in UPF and weight gain between 6-12 months of age, but not weight through follow-up at age 12 months.³

Conclusion statement and grade

The 2025 Dietary Guidelines Advisory Committee did not develop a conclusion statement to answer the question because of substantial concerns with consistency and directness in the available evidence (Table 5). Studies also had numerous concerns due to risk of bias (Table 7). This body of evidence includes both large and small studies, including studies with smaller sample sizes and null findings, which makes publication bias less likely.

Table 5

Conclusion statement, grades for dietary patterns with varying amounts of ultra-processed food consumed by infants and young children, up to age 24 months, and growth, body composition, and risk of obesity.

Table 6

Studies examining the relationship between dietary patterns with varying amounts of ultra-processed food consumed by infants and young children up to age 24 months and growth, body composition and composition and risk of obesity.

Table 7

Risk of bias for observational studies examining dietary patterns with varying amounts of ultra-processed food consumed by infants and young children up to age 24 months and growth, body composition and risk of obesity.

Children and adolescents

Twenty-five articles met inclusion criteria and examined the relationship between dietary patterns with varying amounts of UPF consumed by children and adolescents and growth, body composition, and risk of obesity (also see Table 4). All articles analyzed prospective cohort studies.⁶^–³⁰

Description of the evidence

Population

Studies examined dietary patterns varying in UPF consumed by participants ranging in age between 2 up to 19 years. Race and/or ethnic background of participants was reported as mostly or predominantly non-White and/or Hispanic in 3 articles and predominantly White and/or non-Hispanic in 5 articles. Information on socioeconomic position (SEP) of participants widely varied across studies.

Data from 18 different countries were represented across included articles (Australia; Belgium; Brazil; Cyprus; Estonia; France; Germany; Hungary; Ireland; Italy; Netherlands; Norway; Portugal; Spain; Sweden; United Kingdom; United States; Uruguay). Multiple articles from a single cohort study were included from the following studies: Avon Longitudinal Study of Parents and Children (ALSPAC),⁹^,²⁷ the EDEN study²⁴^,²⁵, Generation XXI¹⁴^,¹⁵^,²⁰^,²⁸^,²⁹; Identification and prevention of Dietary- and lifestyle-induced health EFfects In Children and infantS (IDEFICS) and/or the Kiel Obesity Prevention Study (KOPS)²³^,³⁰ and the Pelotas Birth Cohort study.¹¹^,¹²^,¹⁷ Across these cohort studies, the different articles examined dietary patterns varying in UPF differently, such as per 100 grams or per 100 kcal increase, by different outcomes or ages, and/or by different exposure classifications for dietary patterns varying in UPF (e.g., derived scores by Nova classification groups).

Intervention/exposure and comparator

Dietary intake assessment methods included food frequency questionnaire (n=13), 24-hour recalls (n=5), and/or diet history/diaries or other methods (n=7). Dietary patterns varying in UPF were examined by deriving a priori score-based indices (e.g., alignment or adherence based on a pre-determined index/score),⁷^,⁹^–¹³^,¹⁷^,¹⁸^,²¹^,²⁸^,²⁹ and data-driven methods of pattern development such as factor/cluster analysis⁶^,⁸^,¹⁶^,¹⁹^,²²^–²⁷^,³⁰ or latent class analysis¹⁵^,²⁰ Ten articles examined dietary patterns with different amounts of UPF, where UPF were defined by the Nova system. Exposure to UPF within a dietary pattern was based on contribution to total energy or weight and examined continuously (such as per 100 grams amount consumed) and/or categorically (such as by tertiles or quartiles).⁷^,⁹^–¹³^,¹⁷^,¹⁸^,²⁸^,²⁹ Four articles examined consumption of a dietary pattern with UPF compared to different dietary pattern without or low in UPF.¹⁴^,¹⁵^,¹⁹^,²⁰

All included studies described the types of UPF contributing to the dietary pattern. However, the groupings of foods that were classified as UPF widely varied across studies. Common sources or types of UPF contributing to the dietary patterns included:

-: sugar-sweetened beverages
-: processed meats and meat products
-: sugar-sweetened foods
-: packaged salty/savory and/or sweet snack foods
-: ready-to-eat meals and dishes

Outcomes

All of the included studies reported using standardized procedures to measure (in-person) body size and/or composition of participants. The following outcomes were reported in the included studies:

Risk of overweight and/or obesity¹⁹
BMI, BMI z-score, weight-to-height ratio, and/or weight-for-length z-score⁶^–¹⁰^,¹²^–¹⁶^,¹⁸^,¹⁹^,²³^,²⁴^,²⁸^–³⁰
Body composition (such as z-scores for fat mass, fat mass index, fat free mass, skinfold thickness)¹⁰^–¹⁵^,²⁴^,²⁵^,²⁷^,²⁹^,³⁰
Waist circumference (and waist circumference z-score or waist-hip-ratio)⁷^,¹⁰^,¹²^,¹⁴^,¹⁵^,²⁰^,²⁹^,³⁰
Weight and/or height/length⁹^,¹²^,¹⁹

Synthesis of the evidence

This evidence synthesis focuses on the studies with fewer limitations, which included studies that: (1) were designed to directly examine dietary patterns that vary in UPF; (2) used 24-hour recall or dietary records to collect dietary intake of UPF; and (3) used a food processing classification system to categorize items as UPF; Across evidence, the direction of reported effect estimates were similar, regardless of statistical significance, and suggested that dietary patterns with higher relative to lower intakes of foods (either explicitly or likely) classified as UPF were associated with greater adiposity and/or risk of overweight.⁶^,⁹^,¹²^,¹⁸^,²⁸ Specifically, dietary patterns higher in UPF were associated with significantly greater waist circumference, fat mass index, body fat percentage,⁹^–¹¹ BMI or BMIz-score,⁹^,¹²^,¹⁸^,²⁸ weight and/or lower length/height for age z-score,⁹^,¹² and risk of overweight.²⁶ Seven of these studies based UPF categorization on the Nova classification group 4. The amount or contribution of UPF was analyzed differently in each article and included continuously per 100g,¹¹ per SD,²⁶ per 10%,¹⁰ or per 100 kcal²⁸ increments; by % weight,⁹ or dichotomously as either 1300 kcal vs. 300 kcal¹⁸ or 6 or more compared to 5 or less Nova 4 sub-groups.¹² The mean percent of total energy from UPF within dietary patterns consumed ranged from 27% to 42% among the studies reporting that information. One study reported a range of 1 to 27 servings/day from UPF across participants.⁷ One study reported mean energy in kcal/day from UPF as 698 kcal/d.¹⁸ One study reported the range of UPF items per day from 9 to 15.¹³

The magnitude of effect estimates across the body of evidence varied, but tended to be small (e.g., <10% relative risk). In addition, many of the articles reported both statistically significant results as well as comparisons or findings that did not reach statistical significance. For example, Vedovato and colleagues²⁸ found that a dietary pattern higher in UPF (per 100 kcal) at age 4 years, but not at age 7 years, was significantly associated with higher BMI z-score at age 10 years. In contrast, Heerman et al. 2023¹⁸ observed that a dietary pattern with 1300 vs. 300 kcal from UPF in children at age 3 years or 4 years, but not at age 5 years, was significantly associated with higher BMI z-score at age ~ 6 to 8 years.

Two articles each examined consumption of dietary patterns that were comprised primarily of foods/items likely classified as UPF.⁶^,²⁷ Smith and colleagues²⁷ found that the ‘Packed Lunch’ dietary pattern at age 9 years was associated with higher fat mass gains in girls and less lean mass gain in boys at age 11 years (among valid-reporters). Results examining under-reporters were not statistically significant. Arruda and colleagues⁶ found that the ‘Western’, but not the ‘Snacks’ dietary pattern, at age 10 to 14 years was significantly associated with higher BMI z-score at age 13-17 years.
Two articles from the same cohort examined the ‘Energy-dense foods (EDF)” dietary pattern comprised primarily of foods/items likely classified as UPF (e.g., soft drinks, salty pastry, sweets, processed meats) compared to the ‘Healthier’ dietary pattern (low in EDF components and comprised of non-processed foods such as fruits and vegetables).¹⁴^,¹⁵ Consuming the ‘EDF’ v. ‘Healthier’ dietary pattern at age 4 years was significantly associated with statistically higher risk of overweight or obesity, greater fat mass index, BMI z-score, and weight-to-height-ratio in girls; higher risk of overweight/obesity in boys at age 7 years¹⁴ and higher BMI, fat mass index, weight-to-height-ratio and risk of obesity in girls at age 10 years.¹⁵ However, no significant associations were found for fat mass percentage in girls or boys at either follow-up, and for most outcomes in boys (BMI z-score, fat mass index, weight-to-height-ratio) at both follow-up points.

Two articles reported that dietary patterns with minimal or no UPF were associated with lower BMI z-score, waist circumference,⁷^,²⁹ and fat mass z-score.²⁹ Vilela and colleagues²⁹ found that consumption of less/minimal UPF (Nova 2+1 per 100g) was significantly associated with lower fat mass z-score, BMI z-score, waist circumference z-score at age 10 years. No significant associations were reported when examining dietary patterns based on Nova 3 classification of foods in relation to outcomes. No significant associations were found between consumption of more UPF (Nova 4 per 100g) and fat mass z-score, BMI z-score, waist circumference z-score at age 10 years. Also, none of the dietary patterns examined in this article were associated with fat-free mass z-score. Bawaked and colleagues⁷ found that consuming less vs. more UPF (medium or low UPF compared to high UPF based on Nova) at age 4 years associated with lower waist circumference at age 7 years, but this was not statistically significant.

One article reported that a dietary pattern higher in UPF was associated with lower BMI and body fat percentage, but complete data were not reported.¹³ Cunha and colleagues¹³ found that consuming more vs. less UPF (Q4 vs. Q1 Nova 4) at mean age 15.7 years was significantly associated with lower BMI and lower body fat percentage after 3-year follow-up. Notably, values and/or data supporting those results were not reported. In addition, total energy intake and BMI at baseline were lower in those with the highest consumption levels of UPF.

Findings that were not statistically significant (or mostly non-significant) between dietary patterns with varying amounts of UPF and outcomes generally supported a similar direction as significant results described above.

González and colleagues¹⁷ found that consuming greater amounts of UPF (Nova 4 per 100g) at age 24.4 months was not significantly associated with incidence of obesity at age 47.8 months.
Saldanha-Gomes and colleagues²⁵ reported that consumption of a dietary pattern comprised of many UPF at age 2 years’ was associated with higher risk of early adiposity rebound at age 5.5 years, but this result did not reach statistical significance. In another article by Saldanha-Gomes and colleagues,²⁴ consumption of that dietary pattern at age 2 years was not associated with BMI or % body fat at age 5 years.
Wolters and colleagues³⁰ found that consumption of a dietary pattern at age 5-7 years (labelled “Fast-food”) was not associated with BMI, fat mass index, or weight-to-height ratio at age 9-11 years. Change toward greater consumption of that pattern from baseline to follow-up was significantly associated with greater fat mass index at follow-up but not with BMI or weight-to-height ratio. Additionally, consumption of a different pattern at age 5-7 years (labelled, “Snack”) was not associated with fat mass index and weight-to-height ratio, but weakly associated with higher BMI at follow-up 2 years later. Change toward greater consumption of that pattern from baseline to follow-up was not associated with BMI, fat mass index, or weight-to-height ratio at follow-up.
Biazzi and colleagues⁸ found that consuming “DP II” comprised of foods mostly classified as UPF, at age 7-10 years was not significantly associated with change in BMI z-score at age 12-15 years.
Gasser and colleagues¹⁶ found that consuming the ‘Unhealthy’ dietary pattern, comprised of “snacks, sugary drinks, and other UPF”, at age 4-5 years was not significantly associated with BMI z-score or weight-to-height ratio at age 14-15 years. In addition, this dietary pattern at age 2-3 years was not significantly associated with BMI z-score at age 10-11 years, but it was (weakly) associated in some waves with higher weight-to-height ratio at age 10-11 years.
Hennessy and colleagues¹⁹ found no significant difference in body weight, weight z-score, height, height z-score, BMI, BMI z-score, prevalence of overweight, prevalence of obesity, prevalence of overweight/obesity, or prevalence of underweight at age 5 years between consumption of the ‘Low Nutrient-Density’ dietary pattern, which was comprised of confectionary, processed meat, and convenience foods compared to the ‘Traditional’ dietary pattern of whole meal breads, fresh meat, and fruit at age 2 years.
Marks and colleagues²¹ found that consuming a dietary pattern including chips and other snacks, candies, and other UPF, at age 11-13 years was not significantly associated with prevalence of overweight/obesity 1 year later.
Oellingrath and colleagues²² found that consuming a dietary pattern characterized by high-energy processed fast foods, refined grains, cakes and sweets at age 12-13 years was not significantly associated with prevalence of overweight 3 years later in those with healthy weight at baseline, nor in those with overweight at baseline.
Pala²³ found that consumption of a dietary pattern comprised of street foods, savory pastries, and chocolate bars (‘Snacking’) or a dietary pattern of chocolate spreads, biscuits and sweets/candy, fried meat, and soft drinks (‘’Sweet’), at age 2-10 years were not associated with risk of overweight/obesity or change in BMI approximately 2 years later.
Marinho²⁰ found no significant association between the EDF pattern described previously (primarily items likely to be UPF such as soft drinks, salty pastry, sweets, processed meats) compared to a ‘Healthier’ dietary pattern (low in EDF components and comprised of non-processed foods such as fruits and vegetables) and waist-to-weight ratio as a mediator of intelligence quotient.

Conclusion statement and grade

The 2025 Dietary Guidelines Advisory Committee developed a conclusion statement to answer the question based on their review of evidence examining dietary patterns with varying amounts of UPF consumed during childhood and adolescence and measures of growth, body composition, and risk of overweight/obesity (Table 8).

Table 8

Conclusion statement, grades for dietary patterns with varying amounts of ultra-processed food consumed by children and adolescents and growth, body composition, and risk of obesity.

Assessment of Evidence

This body of evidence includes studies with both larger and smaller sample sizes as well as null findings, which makes publication bias less likely. As outlined and described below, the body of evidence was assessed for the following elements used when grading the strength of evidence.

Consistency

The direction of findings were similar regardless of statistical significance suggesting that dietary patterns higher in foods classified as UPF were associated with a higher risk of overweight and/or greater BMI, waist circumference, and/or fat mass. However, the size of effects ranged widely and confidence intervals often included the null. Only one article reported results in the opposite direction, but those findings were likely explained by methodological inconsistencies.

Precision

Studies ranged in analytic sample sizes from n=243 to n=9,427. Several articles had either smaller sample sizes, minimal variability between the exposure and comparator to detect an effect, and/or limited number of events.

Risk of bias

Various potential sources of bias were identified (Table 5). Most of the key confounders were accounted for across studies, with exception of race and/or ethnicity of participants. Studies that classified UPF based on a single baseline diet assessment are more prone to misclassification of their usual intake over time, particularly among children. Assessment of foods and beverages as UPF from food-frequency questionnaires has not been validated in all studies and may also contribute to potential exposure misclassification. All of the studies objectively measured the outcomes, but several articles did not account for missing data in the analysis.

Directness

Studies using Nova classification of UPF were more direct in addressing the relationship of interest compared to the studies that used factor/cluster/latent class methods for deriving dietary patterns. Studies that used a posteriori methods were not directly intending to examine variation in UPF, but rather examine consumption of dietary patterns comprised of many foods often consumed in highly processed versions and/or likely to be classified as UPF with few exceptions. No studies directly compared dietary patterns with the same types of foods in a non-ultra-processed compared to ultra-processed version.

Generalizability

Only one study was conducted in the United States. Socioeconomic position of participants varied across studies and was most commonly based on parental education. Most of the dietary patterns compared and components among the dietary patterns are commonly consumed in the United States, such as cookies, cakes, candy/confectionary, luncheon (cold) meats, nuggets, fish sticks, and sugar-sweetened beverages, with relatively few elements that may be less generalizable such as curry-sausage or muesli. Outcomes were generalizable to those experienced in the United States population.

Table 9

Studies examining the relationship between dietary patterns with varying amounts of ultra-processed food consumed in children and adolescents and growth, body composition and composition and risk of obesity.

Table 10

Risk of bias for observational studies examining dietary patterns with varying amounts of ultra-processed food consumed by children and adolescents and growth, body composition and risk of obesity.

Adults and older adults

Sixteen articles met inclusion criteria that examined the relationship between dietary patterns with varying amounts of UPF consumed by adults and older adults and body composition and risk of obesity (also see Table 8). One article came from an RCT³² and 15 came from prospective cohort studies.³¹^,³³^–⁴⁶

Description of the evidence

Population

Studies were conducted in the following countries: Australia; Brazil (n=4); China; France; Iran; Korea; Spain (n=4); and the United Kingdom (n=2). One study was conducted across multiple countries (Denmark, France, Germany, Greece, Italy, Netherlands, Norway, Spain, Sweden, and the United Kingdom). Two included articles were based on data from the same cohort study but examined dietary patterns differently (e.g., per 100g vs. per 100 kcal increase). Studies ranged in sample size from 661 participants to 348,748 participants.

In 9 articles, the mean BMI of the study population was ≥ 25 kg/m2 (i.e., participants with overweight or obesity) or were at high-risk for diet-related chronic disease. Socioeconomic position of participants was reported most commonly from educational attainment level of participants, which was relatively high (e.g., >50% of participants had ~ 12 years or more) in 9 articles but widely varied in other articles. Race and/or ethnic information about participants was reported in 8 of 17 articles as follows:

“16.2% Black; Brown, 28.1%; 52.2% White; 2.6% Asian, 1% Indigenous”³³
“14% Black, 27% Brown, 55% White, 3% Asian, 1% Indigenous"³⁴
“6% “Black”, 31% “Brown”, 62% “White”, 1% “Yellow/Oriental”⁴⁴
“White, Spanish”³⁷
“White British”³⁹
100% Caucasian³²
“Brazilian”⁴⁵
“Korean”⁴⁶

Studies examined dietary patterns consumed by participants ranging in age between 18 up to 75 years. Mean follow-up duration varied widely across studies, with a range from 6 months to approximately 15 years.

Intervention/exposure and comparator

Dietary intake assessment methods included food-frequency questionnaire,³²^,³³^–³⁷^,⁴⁰^–⁴²^,⁴⁴^,⁴⁶ and 24-hour recalls.³¹^,³⁸^,³⁹^,⁴³^,⁴⁵ Dietary pattern methods included reduced rank regression,³⁹ factor/cluster analysis³² and a priori score derivation in the remaining studies that used the Nova^* classification to define UPF intake. In the one article from an RCT,³² dietary pattern changes were examined by factor analysis 6 months after participants followed the investigator-assigned “Atlantic” diet, which was intended to reflect traditional patterns in northwestern Spain and Portugal (composed of home-cooked, local, fresh, and minimally processed foods).

Intake of UPF was analyzed as a continuous exposure variable (e.g., per 10% or 15% different in energy from UPF,³¹^,³³^,³⁶^,³⁷^,⁴¹^,⁴³ per standard deviation,³⁵ or by weight;³⁴^,³⁸^,⁴⁴^,⁴⁶) and/or categorically (e.g., 700 vs. 234 g/d, g/day, servings/day, and/or quantiles).³¹^,³³^,³⁵^–³⁸^,⁴⁰^–⁴⁶ All included studies described the types of UPF contributing to the dietary pattern, but the number and types of UPF items and/or sub-groups within dietary patterns varied (e.g., 14 total items; 6 main sources; 3 sub-groups with 23 total items).

Outcomes

Included articles reported standardized procedures to measure body size and/or composition measurements in-person or used self-reported measures of weight and height. The following outcomes were reported across the body of evidence:

Risk of overweight and/or obesity⁴⁰^,⁴⁶
BMI³¹^,³⁶^,⁴³^–⁴⁶
Body composition (such as fat mass, fat mass index, % body fat)³⁷^,³⁹^,⁴³^,⁴⁴
Waist circumference³³^,³⁴^,³⁶^,³⁸^,³⁹^,⁴³
Weight³³^,³⁵^,³⁶

Synthesis of the evidence

This evidence synthesis focuses on the studies with fewer limitations, which included 1) studies that were designed to directly examine dietary patterns that vary in UPF, 2) studies that used 24-hour recall to collect dietary intake of UPF, and 3) studies that used a food processing classification system to categorize items as UPF. Across evidence, the direction of reported effect estimates were similar, regardless of statistical significance, and suggested that dietary patterns with higher compared to lower amounts of foods classified as UPF were significantly associated with higher risks of developing overweight and/or obesity, larger waist circumference, greater adiposity (e.g., fat mass, fat mass index, % body fat), and/or higher BMI.³¹^–³⁸^,⁴⁰^,⁴¹^,⁴³^–⁴⁶ In 2 of these 14 articles, dietary patterns with few or no foods classified as UPF were significantly associated with lower risks of developing obesity and/or less BMI gain.⁴⁵^,⁴⁶

The mean amount of UPF within dietary patterns ranged from ~5% up to 74% of total energy consumed in the populations for studies reporting this information. Common sources or types of UPF contributing to the dietary patterns included:

sources of added sugar, primarily from refined grains and/or sugar-sweetened foods (e.g., mass-produced/packaged cookies, breads, buns, rolls, cakes, pastry)³¹^,³³^–⁴⁶
processed meats and meat products (e.g., ham, sausage, hot dog, nuggets)³¹^–³⁸^,⁴⁰^–⁴²^,⁴⁴^–⁴⁶
sugar-sweetened beverages (e.g., soft drinks, sports drinks, fruit drinks)³¹^–³⁸^,⁴⁰^–⁴²^,⁴⁴^–⁴⁶
packaged (salty or savory) snack foods (e.g., chips, crackers),³¹^–³⁴^,³⁶^–³⁸^,⁴¹^–⁴⁵
pre-prepared, “fast”, ready-to-eat meals and dishes (e.g., pizza, instant noodles)³²^–³⁹^,⁴¹^–⁴⁶

Two of 16 articles reported a consistent direction of findings, but did not observe statistically significant associations between dietary patterns with varying amounts of UPF and risk of obesity and/or BMI.³⁹^,⁴² One of these articles used Nova 4 to classify UPF intake, and the other used reduced rank regression to examine a dietary pattern of foods typically classified as UPF (e.g., confectionary; buns, cakes, and pastries; sugar-sweetened soft drinks, processed meats). Therefore, the lack of statistical significance may be explained by these methodological inconsistencies.

Conclusion statement and grade

The 2025 Dietary Guidelines Advisory Committee developed a conclusion statement to answer the question based on their review of evidence that examined dietary patterns with foods classified as UPF consumed by adults and older adults in relation to outcomes including risk of obesity, overweight, and/or adiposity (fat mass, waist circumference, BMI) (Table 11).

Table 11

Conclusion statement, grades for dietary patterns with varying amounts of ultra-processed food consumed by adults and older adults and body composition, and risk of obesity.

Assessment of Evidence

As described below, the body of evidence was assessed for the following elements used when grading the strength of evidence. This body of evidence includes both large and small studies, including studies with smaller sample sizes and null findings, which makes publication bias less likely.

Consistency

Direction and magnitude of effects were consistent in ~ 88% (16) of the studies: dietary patterns with higher compared to lower UPF were significantly associated with higher risk of overweight, greater BMI, waist circumference, and/or fat mass. Two studies also analyzed dietary patterns that were definitively low in UPF and observed they were related to a lower risk of obesity/overweight, and/or less gain in BMI or weight. For the 2 studies that did not report statistically significant findings, the direction of the relationship was similar to the other studies, supporting a positive association between dietary patterns with varying amounts of UPF and these health outcomes.

Precision

The studies were generally well-powered to address the research question and demonstrated effects from a wide range of sample sizes (n=661 to 348,748).

Risk of bias (Table 7 and Table 8)

Across domains, various risks of bias were identified. Most of the key confounders were accounted for across studies with few exceptions: race and/or ethnicity of participants was not reported or accounted for in 8 articles and anthropometry at baseline in 3 articles. Nine of the studies assessed diet only once (at baseline), and therefore do not fully account for potential change in dietary patterns that may have occurred over time. Assessment of foods and beverages as UPF based on dietary intakes collected via food-frequency questionnaires contributes to potential for exposure misclassification. Most studies objectively measured outcomes, but 6 used self-reported weight/height.

Directness

Studies were directly addressing the relationship of interest by examining dietary patterns with varying amounts of UPF, which were based on the Nova classification system in most studies. A few studies examined dietary patterns comprised of foods that are typically classified as “UPF” or likely consumed in ultra-processed versions but were not explicitly designated or cited as such with a specific classification system.

Generalizability

No studies were conducted in the U.S. Nine articles included a majority or all participants with overweight/obesity and/or high-risk for diet-related chronic disease. Nine studies did not report information on the racial or ethnic background of participants. Education was the most common information provided on SEP of participants and varied across studies. Most of the dietary patterns compared are generalizable to the U.S. population. However, a few sources of UPF (from select studies) may not be as generalizable (e.g., ‘acarajé’ or ‘instant pork mince dumpling’). The outcomes examined are generalizable to the U.S. population.

Table 12

Studies examining the relationship between dietary patterns with varying amounts of ultra-processed food consumed in adults and older adults and growth, body composition and composition and risk of obesity.

Table 13

Risk of bias for interventions examining dietary patterns with varying amounts of ultra-processed food consumed by adults and older adults and growth, body composition and risk of obesity.

Table 14

Risk of bias for observational studies examining dietary patterns with varying amounts of ultra-processed food consumed by adults and older adults and growth, body composition and risk of obesity.

Pregnancy and postpartum

Two articles met the inclusion criteria examining the relationship between dietary patterns with varying amounts of UPF consumed during pregnancy and/or postpartum and gestational weight gain and postpartum weight change. Both articles analyzed prospective cohort studies.⁴⁷^,⁴⁸

Description of the evidence

Population

One study was conducted in the United States⁴⁷ and another one in Brazil⁴⁸ and included 367 and 584 participants, respectively. Both articles included >50% of the participants with overweight and obesity and Dias and colleagues, included only participants with gestational diabetes mellitus.⁴⁸ Cummings and colleagues, noted that participants were 17% “Black”, 5.3% “Asian”, 6.3% “Other or multiracial”, and 8.6% Hispanic and income-poverty ratio of 3.84.⁴⁷ Dias and colleagues, reported that 49% of the participants were non-White and 21% of the participants had family income below minimum wage.⁴⁸

Intervention/exposure and comparator

Cummings and colleagues⁴⁷ collected dietary data using 24-hour dietary recall at multiple time-points during pregnancy and postpartum period and examined scores on the “Instant” dietary pattern comprised of ready-to-heat pre-prepared pies, pasta, and pizza dishes; mass-produced packaged breads; reconstituted meats; sweet or savory packaged snacks; confectionery desserts; sweetened drinks. Dias and colleagues⁴⁸ used a brief food-frequency questionnaire to collect dietary data during pregnancy and 6-mo postpartum periods and derived the “Risk” dietary pattern from factor/cluster analysis comprised of Fried foods; Cookies and sweets; Sweetened beverages; Processed meat; Red meat with visible fat.

Outcomes

Cummings and colleagues assessed weight during pregnancy and categorized gestational weight gain as inadequate, adequate or excessive, based on 2009 Institute of Medicine guidelines^*. Postpartum weight change was defined as the change in weight between baseline (<12 weeks of pregnancy) and one year postpartum, both measured by the investigators. Percent of gestational weight gain retained was calculated by multiplying 100 by the difference in weight between the last prenatal medical visit (investigator measured) and one year postpartum (investigator measured), divided by gestational weight gain. Dias et al. reported the variation in BMI between 2-mo and 12-mo postpartum. These outcome measures were calculated based on the self-reported weight by the participants.

Synthesis of the evidence

Pregnancy

One study assessed the association between dietary patterns with varying amounts of UPF and gestational weight gain. The authors reported that a dietary pattern with greater amounts of foods classified as UPF (per SD) during pregnancy was significantly associated with higher risk of excessive gestational weight gain, but not significantly associated with the risk of inadequate gestational weight gain.

Postpartum

Two studies reported that there was no association between consumption of dietary patterns with UPF during postpartum and either % gestational weight gain retained and postpartum weight change⁴⁷ or with BMI at 12 months postpartum⁴⁸

Conclusion statement and grade

Insufficient evidence is available to adequately assess the body of evidence for consistency, directness, risk of bias (see Table 17), precision, and generalizability. This body of evidence is less likely to have publication bias as it includes only 2 studies with mixed and/or null findings from relatively small sample sizes.

Table 15

Conclusion statement, grades dietary patterns with varying amounts of ultra-processed food consumed during pregnancy and gestational weight gain.

Table 16

Conclusion statement, grades dietary patterns with varying amounts of ultra-processed food consumed during postpartum and postpartum weight change.

Table 17

Studies examining the relationship between dietary patterns with varying amounts of ultra-processed food consumed during pregnancy and postpartum and postpartum and gestational weight gain and postpartum weight change.ⁱ.

Table 18

Risk of bias for observational studies examining dietary patterns with varying amounts of ultra-processed food consumed during pregnancy and/or postpartum and gestational weight and postpartum weight change.

Footnotes

†: Hoelscher DM, Anderson, Booth, et al. Dietary Patterns and Growth, Body Composition, and Risk of Obesity: A Systematic Review. Date TBD. U.S. Department of Agriculture, Food and Nutrition Service, Center for Nutrition Policy and Promotion, Nutrition Evidence Systematic Review. Available at: https://nesr.usda.gov/
*: Cole, T.J., Bellizzi, M.C., Flegal, K.M., Dietz, W.H. Establishing a standard definition for child overweight and obesity worldwide: International survey. British Medical Journal 2000;320(7244):1240–1243. [PMC free article: PMC27365] [PubMed: 10797032]
*: Additional funding reported in Saldanha-Gomes et al. 2017 and 2022: French Ministry of Health, French Ministry of Research, INSERM Bone and Joint Disease National Research, Human Nutrition National Research Programs, Paris-Sud University, Nestle, French National Institute for Population Health Surveillance (InVS), French National Institute for Health Education (INPES), the European Union FP7 programmes, Diabetes National Research Program, French Agency for Environmental Health Safety, Mutuelle G´en´erale de l’Education Nationale, French national agency for food security, French-speaking association for the study of diabetes and metabolism
*: Details about the Nova food classification system can be found in this publication: Monteiro CA, Cannon G, Levy RB, et al. Ultra-processed foods: what they are and how to identify them. Public Health Nutr. 2019;22(5):936–941. doi:10.1017/S1368980018003762 [PMC free article: PMC10260459] [PubMed: 30744710] [CrossRef]
*: Cordova, 2021 additional funding: National cohorts supported by individual funders: Ligue Contre le Cancer, Institut Gustave-Roussy, Mutuelle Générale de l’Education Nationale, Institut National de la Santé et de la Recherche Médicale (INSERM) (France); German Cancer Aid, German Cancer Research Center (DKFZ), Federal Ministry of Education and Research (BMBF) (Germany); Associazione Italiana per la Ricerca sul Cancro-AIRC-Italy and National Research Council (Italy); Dutch Ministry of Public Health, Welfare and Sports (VWS), Netherlands Cancer Registry (NKR), LK Research Funds, Dutch Prevention Funds, Dutch ZON (Zorg Onderzoek Nederland), World Cancer Research Fund (WCRF), Statistics Netherlands (The Netherlands); Health Research Fund (FIS-ISCIII), the Regional Governments of Andalucía, Asturias, Basque Country, Murcia, Navarra, and the Catalan Institute of Oncology (Barceloan), Spain); Cancer Research UK and Medical Research Council (EPIC-Norfolk; EPIC-Oxford)
*: Possible ratings of low, some concerns, or high determined using the “Cochrane Risk-of-bias 2.0” (RoB 2.0) (August 2019 version)” (Sterne JAC, Savović J, Page MJ, Elbers RG, Blencowe NS, Boutron I, Cates CJ, Cheng H-Y, Corbett MS, Eldridge SM, Hernán MA, Hopewell S, Hróbjartsson A, Junqueira DR, Jüni P, Kirkham JJ, Lasserson T, Li T, McAleenan A, Reeves BC, Shepperd S, Shrier I, Stewart LA, Tilling K, White IR, Whiting PF, Higgins JPT. RoB 2: a revised tool for assessing risk of bias in randomised trials. BMJ 2019; 366: l4898 [PubMed: 31462531].
*: Institute of Medicine. Weight gain during pregnancy: reexamining the guidelines. Washington, DC: National Academies Press; 2009 [PubMed: 20669500]
*: Abbreviations: FFQ, Food frequency questionnaire; GDM, Gestational diabetes mellitus; GWG, Gestational weight gain; PPWR, Postpartum weight retention; PPWC, Postpartum weight change; TEI, total energy intake; UPF, ultra-processed food
*: Possible ratings of low, moderate, serious, critical, or no information determined using the “Risk of Bias in Non-randomized Studies of Exposures (ROBINS-E)” tool (ROBINS-E Development Group (Higgins J, Morgan R, Rooney A, Taylor K, Thayer K, Silva R, Lemeris C, Akl A, Arroyave W, Bateson T, Berkman N, Demers P, Forastiere F, Glenn B, Hróbjartsson A, Kirrane E, LaKind J, Luben T, Lunn R, McAleenan A, McGuinness L, Meerpohl J, Mehta S, Nachman R, Obbagy J, O’Connor A, Radke E, Savović J, Schubauer-Berigan M, Schwingl P, Schunemann H, Shea B, Steenland K, Stewart T, Straif K, Tilling K, Verbeek V, Vermeulen R, Viswanathan M, Zahm S, Sterne J). Risk Of Bias In Non-randomized Studies - of Exposure (ROBINS-E). Launch version, 1 June 2022. Available from: https://www.riskofbias.info/welcome/robins-e-tool.) *Low risk of bias except for concerns about uncontrolled confounding.

The contents of this document may be used and reprinted without permission. Endorsements by NESR, NGAD, CNPP, FNS, or USDA of derivative products developed from this work may not be stated or implied.

Bookshelf ID: NBK611033

Contents

< Prev Next >

PubReader
Print View
Cite this Page
Stanford FC, Taylor C, Booth S, et al. Dietary Patterns with Ultra-Processed Foods and Growth, Body Composition, and Risk of Obesity: A Systematic Review [Internet]. Alexandria (VA): USDA Nutrition Evidence Systematic Review; 2024 Nov. Results.
PDF version of this title (1.3M)

Results - Dietary Patterns with Ultra-Processed Foods and Growth, Body Compositi...
Results - Dietary Patterns with Ultra-Processed Foods and Growth, Body Composition, and Risk of Obesity: A Systematic Review

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

Dietary Patterns with Ultra-Processed Foods and Growth, Body Composition, and Risk of Obesity: A Systematic Review [Internet].

Results

Literature search and screening results

Infants and young children up to age 24 months

Description of the evidence

Population

Intervention/exposure and comparator

Outcomes

Synthesis of the evidence

Conclusion statement and grade

Children and adolescents

Description of the evidence

Population

Intervention/exposure and comparator

Outcomes

Synthesis of the evidence

Conclusion statement and grade

Assessment of Evidence

Consistency

Precision

Risk of bias

Directness

Generalizability

Adults and older adults

Description of the evidence

Population

Intervention/exposure and comparator

Outcomes

Synthesis of the evidence

Conclusion statement and grade

Assessment of Evidence

Consistency

Precision

Risk of bias (Table 7 and Table 8)

Directness

Generalizability

Pregnancy and postpartum

Description of the evidence

Population

Intervention/exposure and comparator

Outcomes

Synthesis of the evidence

Pregnancy

Postpartum

Conclusion statement and grade

Footnotes

Views

In this Page

Other titles in this collection

Recent Activity