Ultra-processed foods: Implications for gastrointestinal health

Table 1 NOVA classification of foods[1]

Category	Description	Examples
Unprocessed or minimally processed foods	Natural foods altered by removal of inedible parts, drying, freezing, or pasteurization without adding substances	Fresh fruits, vegetables, grains, eggs, milk, and meat
Processed culinary ingredients	Substances extracted from group 1 foods or nature, used in home cooking to prepare and season dishes	Oils, butter, sugar, salt, starch
Processed foods	Products are made by adding salt, sugar, oil, or other group 2 ingredients to group 1 foods, typically to increase shelf life or palatability	Canned vegetables, cheese, salted nuts, bread
UPFs	Industrial formulations with little or no whole foods, containing additives for flavor, texture, and shelf life. Designed for convenience and hyper-palatability	Soft drinks, packaged snacks, instant noodles, and ice cream

UPFs: Ultra-processed foods.

Table 2 Alternative classification systems for ultra-processed foods[1,4-7]

Classification system	Developed by/source	Basis of classification	Key features
NOVA	Monteiro et al[1], Brazil (FAO, 2018)	Extent and purpose of food processing	Categorizes foods into 4 groups; emphasizes the health risks of UPFs
UNC system	University of North Carolina	Processing categories based on ingredient lists and barcodes	Categorizes > 12000 United States foods from NHANES; focuses on industrial processing
IFPRI classification	International Food Policy Research Institute	Technological processes and ingredient function	Emphasizes processing techniques like extrusion, hydrolysis
EU FP7 (Food4Me Project)	European Commission	Combining processing level with nutritional and matrix properties	Assess the impact of the food matrix and nutrient profile
EPIC-Soft/GloboDiet	European Prospective Investigation into Cancer	Harmonized food description for dietary recall	Focuses on preparation and preservation methods

NHANES: National Health and Nutrition Examination Survey; UPFs: Ultra-processed food; FAO: Food and Agriculture Organization of the United Nations.

Table 3 Summary of available studies on the effect of ultra-processed food on the gut microbiome[43-46]

Ref.	Alpha diversity	Beta diversity	Bacterial composition changes about UPFs		Composition changes related to specific UPFs	Clinical outcome
			Increases	Decreases
Atzeni et al[44], 2022	No significant difference	No significant difference	Positive association between Alloprevotella spp. (P = 0.041) and Sutterella spp. (P = 0.116) vs tertile 2; Positive association between Alloprevotella spp. (P = 0.065), Negativibacillus spp. (P = 0.096), and Prevotella spp. (P = 0.116) vs tertile 3		No significant differences found between bacterial taxa and UPFs categories	UPFs consumption is positively associated with higher total energy intake; No significant differences in cholesterol, TG, or HbA1c between the tertiles. No association between bacterial taxa and cardiovascular risk factors
Cuevas-Sierra et al[43], 2021	Men consuming > 5 servings/day of UPFs showed lower richness compared to men consuming < 3 servings/day (observed P = 0.03, Shannon P = 0.01, Chao1 P = 0.04). No differences in women or the whole population	No significant difference	Gemmiger spp. (P < 0.001); Granulicatella spp. (P < 0.001); Parabacteroides spp. (P < 0.001); Shigella spp. (P < 0.001); Bifidobacterium spp. (P < 0.001); Anaerofilum spp. (P = 0.001); Cc_115 spp. (P = 0.007); Oxalobacter spp. (P = 0.008); Collinsella spp. (P = 0.008)	Lachnospira spp. (P = 0.003); Roseburia spp. (P = 0.003)	Women: Dairy and pizza positively correlated with Actinobacteria (P < 0.05), and pizza positively correlated with Bifidobacterium spp. (P < 0.05)	Consumption of UPFs is associated with ↑ serum triglycerides (P = 0.004) and ↓ HDL-c levels (P = 0.04). UPFs consumption is associated with ↑ depression, anxiety, and weight in women, while ↑ weight and BMI in men
Fernandes et al[45], 2023	No associations between food processing level and alpha diversity	NA	Clostridium butyricum; Odoribacter splanchnicus, Barnesiella intestinihominis, Alistipes onderdonkii, Alistipes indistinctus	Ruminococcus sp., (Ruminococcus) gnavus Bacteroides vulgatus Bacteroides plebeius	NA	Consumption of UPFs is associated with leptin resistance
García-Vega et al[46], 2020	Higher in females than males (Shannon, P = 0.046), higher in middle-aged than younger individuals (Shannon, P = 0.012). No significant association between diet quality (including UPFs intake) and alpha diversity	Differences according to participants’ city of origin (P = 0.001), sex (P = 0.001), socioeconomic level (P = 0.024) and BMI (P = 0.002). No significant association between diet quality (including UPFs intake) and beta diversity	Bifidobacterium adolescentis, Prevotella melaninogenica, Subdoligranulum variabile, Veillonella dispar, Ruminococcus spp., Bilophila spp., Oscillospira spp.	Prevotella copri Clostridium hathewayi, Ruminococcaceae unclassified spp. Gemella spp. Lachnospira spp. Oscillospira spp.	OTUs from Oscillospira spp., Unclassified Ruminococcaceae, Ruminococcus spp., Lachnospira spp. Positively associated with the intake of plant-derived food groups, rich in dietary fiber; Bifidobacterium adolescentis is associated with plant-derived food groups; Bile-tolerant Bilophila sp., Prevotella copri, and the opportunistic pathogen Prevotella melaninogenica were associated with increased intake of animal-derived foods	Diets enriched with plant-derived foods have more diverse gut microbiota and ↑ levels of SCFA-producing bacteria

HDL-c: High-density lipoprotein cholesterol; OTU: Operational taxonomic unit; SCFA: Short-chain fatty acid; TG: Triglyceride; UPFs: Ultra-processed food; BMI: Body mass index; HbA1c: Glycated hemoglobin.

Table 4 Summary of available studies on the use of ultra-processed food and risk of inflammatory bowel disease[13-17,51-69]

Ref.	Study design	Population (n)	Assessment method	Diagnostic methods	Main findings
Studies on UPFs and the risk of UC
Persson et al[51], 1992	Case-control	n = 755; Cases: 445, Controls: 310	FFQ	Medical records	Higher fast-food intake: ↑ risk of UC
[52], 1994	Case-control	n = 245; Cases: 146, Controls: 109	Dietary history questionnaire	Self-reported	Western diet: ↑ risk of UC
Rashvand et al[60], 2018	Case-control	n = 186; Cases: 81, Control: 105	FFQ	Medical report	Processed meat: ↑ risk of UC
Akbari et al[67], 2022	Case-control	n = 244; Cases: 85, Control: 158	FFQ	Diagnosis by gastroenterologists	Western dietary pattern: No increased risk of UC
Studies on UPFs and the risk of CD
Cohen et al[57], 2013	Cohort study	n = 6768	FFQ	Self-reported	Higher intake of sweetened beverages: ↑ risk of CD
Peters et al[65], 2022	Cohort study	n = 125445	FFQ	Self-reported	Western diet pattern: ↑ risk of CD
Narula et al[13], 2021	Cohort study	n = 116037	FFQ/NOVA classification	Self-reported	UPFs consumption: ↑ risk of CD
Studies on UPFs and risk of IBD (both UC and CD
Klein et al[53], 1998	Case-control	n = 87; Cases: 60, Control: 27	Dietary history questionnaire	Diagnosis by gastroenterologists	Higher total sugar intake: ↑ risk of both UC & CD
Russel et al[54], 1998	Case-control	n = 1304; Cases: 668, Control: 636	Dietary history questionnaire	Medical records	Chocolate intake: No increased risk of IBD
Sakamoto et al[55], 2005	Case-control	n = 319; Cases: 156, Control: 163	FFQ	Self-administered questionnaire	Higher sweets intake was associated with ↑ the risk of IBD; Risk for CD > UC
Maconi et al[56], 2010	Case-control	n = 243; Cases: 146, Control: 97	FFQ	Medical records	Processed meat and refined sugar intake: ↑ risk of IBD
Ng et al[58], 2015	Case-control	n = 1382; Cases: 775, Control: 607	Food habits questionnaire	Diagnosis by gastroenterologists	Western dietary pattern: No increased risk of IBD
Ananthakrishnan et al[59], 2015	Cohort study	n = 84803	FFQ	Self-reported and medical reports	Western dietary pattern: No increased risk of IBD
Racine et al[61], 2016	Case-control	n = 366351	FFQ	Hospital-based registries, pathology records	Sugar and soft drinks: No increased risk of IBD
Khalili et al[62], 2019	Cohort study	n = 83042	FFQ	Medical records	Sweetened beverage intake: ↑ risk of IBD
Preda et al[63], 2020	Case-control	n = 185	Dietary history questionnaire	Diagnosis by gastroenterologists	Intake of sweets/sweetened drinks, processed meat, fried foods, ice-cream, and mayonnaise: ↑ risk of IBD
Han et al[64], 2020	Case-control	n = 103789; Cases: 46456, Control: 57333	Dietary history questionnaire	Self-reported	Fast food, ice cream, processed meat, cookies, candy, sugar-sweetened beverages: ↑ risk of IBD
Vasseur et al[14], 2021	Cohort study	n = 105832	24-hour dietary/NOVA classification	Self-reported	UPFs consumption: No increased risk of IBD
Meyer et al[15], 2023	Cohort study	n = 413590	FFQ/NOVA classification	Self-reported	UPFs consumption: No increased risk of IBD
Dong et al[66], 2022	Cohort study	n = 413590	FFQ	Self-reported	Processed meat intake: No increased risk of IBD
Lo et al[16], 2022	Cohort study	n = 245112	FFQ/NOVA classification	Self-reported and medical report	UPFs consumption: ↑ risk of IBD

CD: Crohn’s disease; FFQ: Food frequency questionnaire; IBD: Inflammatory bowel disease; UC: Ulcerative colitis; UPFs: Ultra-processed foods.

Table 5 Summary of available studies on positive association of ultra-processed foods with non-alcoholic fatty liver disease/obesity/metabolic syndrome[77,79,80-102]

Ref.	Study design & population (n)	UPFs assessment method	Diagnosis of MASLD	Main findings
Zhang et al[77], 2024	Prospective cohort (United Kingdom Biobank, n = 143073)	24-hour diet recall, NOVA classification	Hospitalizations, mortality records (ICD codes)	26% ↑ risk of severe MASLD (HR: 1.26; 95%CI: 1.15-1.38)
Zhang et al[80], 2022, China	Prospective, n = 16168	NOVA (g/1000 kcal/day)	AUS	The highest UPFs quartile had 18% higher MASLD risk (HR 1.18); dose-response noted
Liu et al[81], 2023	Cross-sectional (NHANES 2011-2018, n = 5499)	24-hour dietary recall, NOVA classification	US-FLI > 30	83% ↑ odds of MASLD (OR: 1.83; 95%CI: 1.42-2.37)
Konieczna et al[79], 2022	Prospective cohort (PREDIMED-Plus subset), n = 5867	FFQ, NOVA classification	FLI	Greater UPFs consumption: Associated with ↑ ALT, AST, and hepatic fat accumulation, especially in high-risk metabolic groups
Rauber et al[82], 2018	Cross-sectional (United Kingdom, 2008-2014)	National Diet and Nutrition Survey data + NOVA classification	Nutrient profile analysis; indirect MASLD risk via dietary patterns	Diets high in UPFs had ↓ fiber, ↑ sugars, and ↑ fats-suggestive of ↑ MASLD risk
Hall et al[83], 2020, United States	RCT, n = 20	NOVA (controlled feeding trial)	MRS	No significant change in liver fat after 2-week UPFs or unprocessed diet
Fridén et al[84], 2022, Sweden	Cross-sectional, n = 286	NOVA (% of kcal)	MRI	Positive crude association with liver fat; not significant after adjustment
Ivancovsky-Wajcman et al[85], 2021, Israel	Cross-sectional, n = 789	NOVA (% of kcal)	AUS + FibroMax panel	No direct UPFs-MASLD link; higher UPFs linked to ↑ NASH and fibrosis in smokers and MASLD patients
Canhada et al[86], 2023, Brazil	Prospective, n = 8065	Semi-quantitative 114-item FFQ NOVA classification	-	Higher UPFs consumption was associated with a 19% ↑ risk of incident MetS. 150 g increase in UPFs consumption/day: Associated with a 4% ↑ risk of incident MetS
Pan et al[87], 2023, China	Prospective, n = 5147	24-hour dietary recall, Cumulative mean UPF intake, NOVA classification (g/day)	-	Higher UPFs consumption: Associated with 17% ↑ risk for MetS (HR 1.17, 95%CI: 1.01-1.35)
Martínez Steele et al[88], 2019, United States	Cross-sectional (NHANES), n = 6385	24-hour dietary recall NOVA classification	-	A 10% ↑ increase in UPFs consumption was associated with a 4% higher prevalence of MetS (PR 1.04, 95%CI: 1.02-1.07). Higher UPFs consumption: Associated with a ↑ prevalence of MetS (PR 1.28, 95%CI: 1.09-1.50)
Lavigne-Robichaud et al[89], 2018, Canada	Cross-sectional, n = 811	24-hour dietary recall NOVA classification	-	Higher UPFs consumption: Associated with ↑ prevalence of MetS (OR 1.90, 95%CI: 1.14-3.17; P for trend = 0.04)
Li et al[90], 2021, China	Prospective, n = 12451	24-hour dietary recall of 3 consecutive days at each survey, Cumulative mean UPF intake NOVA classification (g/day)	-	Higher UPFs consumption: Associated with ↑ risk of overweight/obesity and central obesity
Cordova et al[91], 2021, 9 European countries	EPIC study, prospective, n = 348748	Quantitative dietary questionnaires or semi-quantitative FFQ, or a combination of semi-quantitative FFQ and 7- and 14-day records, NOVA classification	-	Higher consumption of UPFs (per 1 SD increment) was positively associated with weight gain (0.12 kg/5 years, 95%CI: 0.09-0.15)
Rauber et al[92], 2021, United Kingdom	Prospective, n = 18218	24-hour dietary recall, NOVA classification	-	Higher UPFs consumption: Associated with ↑ risk for obesity (HR = 1.79, 95%CI: 106-3.03), and abdominal obesity (HR = 1.30, 95%CI: 113-1.48)
Sandoval-Insausti H et al[93], 2020, Spain	Prospective, n = 652	Face-to-face dietary history, recording all food consumed in a typical week in the preceding year, NOVA classification	-	Participants with a higher UPFs consumption were more likely to develop abdominal obesity (OR = 1.62, 95%CI: 104-2.54; P for linear trend = 0.037)
Beslay et al[94], 2020, France	NutriNet-Sante cohort, n = 110260	24-hour dietary recall, NOVA classification	-	Risk of overweight (HR for an absolute ↑of 10% of UPFs = 1.11, 95%CI: 1.08-1.14, P < 0.001), and for obesity (HR for an absolute increment of 10% of UPF = 1.09, 95%CI: 1.05-1.13, P < 0.001)
Canhada et al[95], 2020, Brazil	Prospective, n = 11827	Semi-quantitative 114-items FFQ NOVA classification	-	UPFs consumption: Associated with a ↑ risk of weight gain and waist gain, overweight/obesity incidence (RR = 1.20, 95%CI: 1.03-1.40), and obesity incidence (RR = 1.02, 95%CI: 0.85-1.21)
Mendonça et al[96], 2016, Spain	SUN project, prospective, n = 8451	Self-administered semi-quantitative 136-item FFQ, NOVA classification	-	↑ incidence of overweight and obesity with ↑ baseline quartiles of UPFs
Silva Meneguelli et al[97], 2022, Brazil	Cross-sectional, n = 325	24-hour dietary recall NOVA classification	-	Positive associations between UPF consumption and excessive body weight (PR = 1.004, 95%CI: 1.00-1.01), and abdominal obesity (PR = 1.004, 95%CI: 1.00-1.01)
Martinez-Perez et al[98], 2021, Spain	Cross-sectional, PREDIMED-Plus trial, n = 5636	Semi-quantitative 143-items FFQ NOVA, IARC, IFIC, and UNC classification	-	5% ↑ in UPFs consumption: Associated with 0.11 higher BMI (95%CI: 0.05-0.18)
Machado et al[99], 2020, Australia	Cross-sectional, NNPAS, n = 7411	24-hour dietary recall NOVA classification	-	UPFs consumption: Associated with higher BMI and WC and ↑ prevalence of obesity and abdominal obesity (P < 0.001 for all outcomes)
Nardocci et al[100], 2021, Canada	Cross-sectional, CCHS, n = 13608	24-hour dietary recall NOVA classification	-	10% ↑ in UPFs consumption: Associated with 6% ↑ odds of obesity (OR 1.06, 95%CI: 1.02-1.11)
Juul et al[101], 2018, United States	Cross-sectional, NHANES, n = 15977	24-hour dietary recall NOVA classification	-	Higher UPFs consumption is associated with a 161-unit increase in BMI (95%CI: 1.11-2.10), a 407 cm increase in WC (95%CI: 2.94-5.19), and greater odds of being overweight (OR 1.48, 95%CI: 1.25-1.76), obese (OR 1.53, 95%CI: 1.29-1.81), and having abdominal obesity (OR 1.62, 95%CI: 1.39-1.89)
Silva et al[102], 2018, Brazil	Brazilian Longitudinal Study of Adult Health (ELSA-Brazil), cross-sectional, n = 8977	Semi-quantitative FFQ, NOVA classification	-	Higher UPFs consumption: Associated with a higher BMI (b = 0.80, 95%CI: 0.53-1.07 kg/m2), WC (b = 1.71, 95%CI: 1.02-2.40 cm) and higher odds for being overweight (OR 1.31, 95%CI: 1.13-1.51), obese (OR 1.41, 95%CI: 1.18-1.69) and increased WC (OR 1.41, 95%CI: 1.20-1.66)

AUS: Abdominal ultrasound; FLI: Fatty Liver Index; US: Ultrasound; MRS: Magnetic resonance spectroscopy; MRI: Magnetic resonance imaging; NNPAS: Cross-sectional, National Nutrition and Physical Activity Survey; CCHS: Canadian Community Health Survey; AST: Aspartate aminotransferase; ALT: Alanine aminotransferase; BMI: Body Mass index; CI: Confidence interval; FFQ: Food frequency questionnaire; HOMA-IR: Homeostatic model assessment for insulin resistance; HR: Hazard ratio; HSI: Hepatic steatosis index; IARC: International Agency for Research on Cancer; IFIC: International food information council; IR: Insulin resistance; Mets: Metabolic syndrome; MASLD: Non-alcoholic fatty liver disease; NHANES: National Health and Nutrition Examination Survey; OR: Odds ratio; QUICKI: Quantitative insulin-sensitivity check index; RCT: Randomized controlled trial; RR: Relative risk; UNC: University of North Carolina; UPF: Ultra-processed food; WC: Waist circumference.

Table 6 Studies with a negative association between ultra-processed foods and non-alcoholic fatty liver disease/obesity/metabolic syndrome

Ref.	Design & population (n)	UPFs assessment method	Outcomes
Magalhães et al[103], 2022, Brazil	Prospective, n = 896	Semi-quantitative 83-item FFQ, NOVA classification	UPFs consumption: No association with MetS
Barbosa et al[104], 2023, Brazil	Cross-sectional, n = 895	24-hour dietary recall, NOVA classification, NOVA score	Higher UPFs consumption: Not associated with a ↑ prevalence of MetS
Nasreddine et al[105], 2018, Lebanon	Community-based survey, n = 302	Semi-quantitative 80-item FFQ NOVA classification	Ultra-processed dietary pattern: No association with MetS (OR 1.11, 95%CI: 0.26-4.65)
Asma et al[106], 2019, Malaysia	Cross-sectional, n = 200	24-hour dietary recall NOVA classification	UPFs consumption: Not associated with BMI, WC, and% body fat
Adams and White[74], 2015, United Kingdom	Cross-sectional, n = 2174	Food diary and UPF	UPFs consumption: No association with markers of body weight

BMI: Body mass index; CI: Confidence interval; FFQ: Food frequency questionnaire; Mets: Metabolic syndrome; OR: Odds ratio; UPF: Ultra-processed food; WC: Waist circumference.

Table 7 Summary of available studies on the use of ultra-processed food and risk of cancers

Ref.	Study design	Population characteristics, n	Outcome parameters	Main findings
Fiolet et al[109], 2018, France	Prospective cohort (NutriNet-Sante)	Adults > 18 years, n = 104980	Overall cancer, breast cancer, CRC, prostate cancer	A 10% ↑ increase in the proportion of UPFs in the diet is associated with; 13% ↑ risk of overall cancer, an 11% ↑ risk of breast cancer. No significant association between UPFs consumption and risk of CRC or prostate cancer
Wang et al[112], 2022, United States	Prospective cohort	Adults: 25-75 years, 206248	CRC	Males in the highest quintile of UPFs consumption had a 29% ↑ risk of CRC compared to males in the lowest quintile. No association was found between UPFs consumption and CRC risk in females
Zhong et al[113], 2023, United States	Prospective cohort	Adults: 55-74 years, n = 98265	Pancreatic cancer	Adults in the highest quartile of UPFs consumption had a 49% ↑ risk of pancreatic cancer compared to individuals in the lowest quartile
[114], 2023	Multicentric prospective cohort	n = 450111	Head and neck cancers, esophageal cancer, gastric cancer, colon cancer, rectal cancer, hepatocellular carcinoma, gallbladder cancer, pancreatic cancer, lung cancer, renal cell carcinoma, bladder cancer, glioma, thyroid cancer, multiple myeloma, non-Hodgkin lymphoma, leukaemia, melanoma, breast cancer (premenopausal and postmenopausal), cervical cancer, endometrial cancer, ovarian cancer, and prostate cancer	A substitution of 10% of processed foods with an equal amount of minimally processed foods was associated with ↓ a risk of: Overall cancer (HR 0.96, 95%CI: 0.95-0.97), head and neck cancers (HR 0.80, 95%CI: 0.75-0.85); Esophageal squamous cell carcinoma (HR 0.57, 95%CI: 0.51-0.64); Colon cancer (HR 0.88, 95%CI: 0.85-0.92); Rectal cancer (HR 0.90, 95%CI: 0.85-0.94); Hepatocellular carcinoma (HR 0.77, 95%CI: 0.68-0.87), and Postmenopausal breast cancer (HR 0.93, 95%CI: 0.90-0.97); Replacement of processed and ultra-processed foods and drinks with an equal amount of minimally processed foods might ↓ the risk of various cancer types
Chang et al[115], 2023, United Kingdom	Prospective cohort	Adults: 40-69 years	Overall, cancer and 34-site site-specific cancers	Every 10% point ↑ in UPF consumption was associated with: ↑ overall cancer (HR: 1.02, 95%CI: 1.01-1.04); ↑ ovarian cancer (HR: 1.19, 95%CI: 1.08-1.30)
El Kinany et al[118], 2022, Morocco	Case-control	Adults > 18 years; Cases (n = 1453); Controls (n = 1453)	CRC	Individuals in the highest tertile of UPFs consumption compared to the lowest tertile had: 40% ↑ OR of having overall CRC; 36% ↑ OR of having colon cancer; 44% ↑ OR of having rectal cancer
Jafari et al[119], 2023, Iran	Case-control	Adults: 40-75 years; Cases (n = 71); Controls (n = 142)	CRC	Individuals in the highest tertile of UPFs consumption had a 332% ↑ OR of CRC compared to individuals in the lowest tertile
Romaguera et al[117], 2021, Spain	Case-control	Adults: 20-85 years; CRC; Cases (n = 1852); Controls (n = 3447); Breast cancer; Cases (n = 1486); Controls (n = 1652); Prostate cancer; Cases (n = 953); Controls (n = 1283)	CRC; breast cancer; prostate cancer	Individuals in the highest tertile of UPFs consumption had a 30% ↑ OR of CRC compared to the lowest tertile; No significant association between UPFs consumption and prostate cancer or overall breast cancer
Romieu et al[121], 2022, Chile, Colombia, Costa Rica, Mexico	Case-control	Adults 20-45 years; Cases (n = 525); Controls (n = 525)	Premenopausal breast cancer	Participants in the highest tertile of UPFs consumption had a 93% ↑ OR of having overall premenopausal breast cancer compared to the lowest tertile
Solans et al[116], 2021, Spain	Case-control	Adults: 20-85 years; Cases (n = 230); Controls (n = 1634)	CLL	In incident cases only, a 10% ↑ in UPFs in the diet was associated with a 22% ↑ OR of being diagnosed with CLL
Trudeau et al[122], 2020, Canada	Case-control (PROtEuS)	Adults: 39-75 years; Cases (n = 1919); Controls (n = 1991)	Prostate cancer	No association was found between UPFs consumption and prostate cancer when comparing quartiles of UPF consumption
Esposito et al[123], 2023, Italy	Case-control	Adults, > 18 years; Cases (n = 44); Controls (n = 88)	CNS tumours	1% ↑ in UPFs in diet was associated with: 6% ↑ OR of overall CNS tumors; 9% ↑ OR of malignant CNS tumors
Jacobs et al[120], 2022, South-Africa	Case-control	Adults > 18 years; Cases (n = 396); Controls (n = 396)	Breast cancer	No statistically significant association between UPFs consumption and breast cancer when comparing tertiles of UPFs consumption

CLL: Chronic lymphocytic leukemia; CNS: Central nervous system; CRC: Colorectal cancer; HR: Hazard ratio; OR: Odds ratio; UPF: Ultra-processed foods; CI: Confidence interval.

Table 8 Available randomised studies on the dietary emulsifier restriction and their effect

Ref.	Type of study	Population characteristics, n	Outcome parameters	Key findings
Bhattacharyya et al[125], 2017	RCT	Cases (n = 5): UC in remission on 100mg carrageenan-containing capsule; Controls (n = 7): UC in remission on placebo capsules	Relapse in two groups at different time points (3, 6, 9, and 12 months); Relapse is defined as an increase of two (or more) points on the SCCAI	Relapses were higher with the carrageenan diet (P = 0.046); Increase in interleukin-6 (P = 0.02) and fecal calprotectin (P = 0.06) in carrageenan diet group
Chassaing et al[126], 2022	RCT	Cases (n = 7): Diet enriched with 15 g/day of CMC; Controls (n = 9): Emulsifier-free diets	Metabolic impact in healthy volunteers: Effect on human gut microbiota composition and gene expression	CMC reduced microbiota richness with a decrease in evenness and Shannon indices; CMC consumption affected the fecal metabolome with depletion of short-chain fatty acids and free amino acids
Fitzpatrick et al[127], 2025	RCT	HED (n = 12) vs LED (n = 12) was provided for 4 weeks in patients with CD		The emulsifier content did not influence disease activity in CD
Bancil et al[128], 2025	Multicentric RCT	LED (n = 75) vs LED plus emulsifier re-supplementation (Controls, n = 79) was provided for 8 weeks in active CD	Primary endpoint: Proportion of patients achieving CDAI response (≥ 70 reduction) at 8 weeks; Secondary endpoint: CDAI remission and fecal calprotectin	CDAI response was achieved in 39 (49.4%) on LED vs 23 (30.7%) in the control group (P = 0.019); Patients on LED are more than twice as likely to experience CDAI remission (adjusted RR: 2.1, 95%CI: 1.0-4.42) and > 50% reduction in FCP (adjusted RR: 2.9, 95%CI: 1.1-8.0)

CD: Crohn’s disease; CDAI: Crohn’s disease activity index; FCP: Fecal calprotectin; RCT: Randomized controlled trial; UC: Ulcerative colitis; SCCAI: Simple Clinical Colitis Activity Index; CMC: Carboxymethylcellulose; HED: High emulsifier diet; LED: Low emulsifier diet; RR: Relative risk.