ultraprocessed foods: cause colon cancer in younger people??

More cancers are being diagnosed at earlier ages. This study found that ultraprocessed foods (UPFs) seem to predispose women to early-onset colorectal cancer (see colon cancer early onset in women jamaOnc2025 in dropbox, or doi:10.1001/jamaoncol.2025.4777)

 

Details:

--29,105 women included from the Nurses’ Health Study II study, an ongoing prospective cohort study of US female registered nurses, with data from 1991 to 2024

-- one advantage of this cohort is that there was a very extensive assessment of food intake (>130 items), from a validated food questionnaire administered every 4 years, as well as assessment of lifestyle factors, family, and medical history biennially; there was long-term follow up in the 90% range

    -- they divided the food intake questionnaire responses into 4 groups: unprocessed or minimally processed foods, processed culinary ingredients, processed foods, and ultraprocessed foods (UPFs) based on the extent of industrial processing

-- the women selected for this data-mining study had completed the baseline 1991 food-frequency questionnaire, had at least one endoscopy before age 50, had no history of cancer before the endoscopy, and had no history of colorectal polyps or inflammatory bowel disease

 

-- baseline characteristics at the time of first endoscopy:

    -- mean age 45, 92% white, BMI 25, height 165 cm

    -- family history of colorectal cancer 12%, history of diabetes 2%, physical activity 23 MET-h/week

    -- menopausal hormone use: 84% were premenopausal, 9% at the time used hormones, 5% never used them

    -- regular aspirin use 11%, regular other NSAID use varied from 27% in those with lowest quintile of ultraprocessed foods (UPF) and up to 38% in those in the highest quintile [aspirin/NSAIDs are important because many studies done over many decades of aspirin and other NSAIDs have documented a beneficial role in the prevention of CRC (colorectal cancer) and colonic adenomas (https://www.nature.com/articles/s41416-022-01882-8 ), with some studies finding regression of adenomas (https://www.gastrojournal.org/article/S0016-5085(02)00448-1/fulltext ].

    -- multivitamins in 53%

    -- alcohol: 6 g/d; ever smokers 35%, cigarette pack-yrs of ever-smokers: 12

    -- total number of endoscopies done: median of 1

    -- reason for endoscopy: screening in 53%, symptoms in 42%, but fewer endoscopies were done for screening in the highest UPF category vs in lowest UPF group (56.1% vs 50.3% in highest UPF group); and the opposite for endoscopies for symptoms (41.6% vs 47.4%)

    -- diet: total calorie intake 1800 kcal/d; fiber 19 g/d, folate 530 mcg/d, calcium 1100 mg/d, vitamin D 400 IU/d

    -- Alternative Healthy Eating Index-2010: mean of 58.0 in lowest UPF group and 50.8 in highest (this is a dietary scoring system that included healthy foods consisting of vegetables, fruits, whole grains, nuts, omega-3s, polyunsaturated fatty acids, and the unhealthy ones of sugar-sweetened drinks, red/processed meats, trans fats, and sodium; alcohol intake (which was scored differently), with total score up to 110. higher scores are associated with decreased risk of heart disease, cancer and all-cause mortality

 

--main outcomes and measures:

    -- incidence of early-onset colorectal cancers (EOCRC) precursors, including confirmed conventional adenomas (tubular, tubulovillous, or villous adenomas) and serrated lesions.

        -- the adenomas were also classified in terms of malignancy risk (high risk if >1cm or if villous or tubulovillous histology, high-grade dysplasia, or the presence of 3 or more adenomas); others were considered low-malignancy risk

    -- multivariable logistic regression models used to estimate the adjusted odds ratios (aORs) for known and putative risk factors

 

Results:

-- 29,105 women were followed over the 24 years of followup

-- 1189 cases were documented of early-onset conventional adenomas, and 1598 of serrated lesions.

-- ultraprocessed foods (UPFs) provided 34.8% of total daily calories and 28.2% of total dietary servings in this group, with median of 5.7 servings/day

-- there was a dose-response relationship between number of servings/day of UPF and conventional adenomas, p =0.02 for a nonlinear relationship. the solid vertical lines in the graph below (their Figure e5 in the supplement) reflect the 25th and 75th percentile of UPF intake, with increases in conventional adenomas starting at a low level of UPF consumption:

 -- the primary ultraprocessed foods consumed in the study:

    -- ultraprocessed breads and breakfast foods: 22.6% of the total UPFs

    -- sauces, spreads and condiments: 21.9%

    -- artificially sweetened beverages: 19.7%

    -- packaged sweet snacks and deserts: 14.3%

 

-- higher UPF intake was associated with higher risk of EOCRC precursors, with adjusted odds ratios (aORs) being adjusted for age, endoscopy period (2-year intervals), number of endoscopies, time since last endoscopy, reason for endoscopy, family history of colorectal cancer, baseline BMI, history of diabetes, menopausal hormone use, smoking history, alcohol intake, physical activity, use of vitamins, aspirin or NSAIDs, dietary factors including fiber, folate, calcium, vitamin D from foods and supplements, and Alternative Healthy Eating Index-2010 score:

    -- incidence of conventional adenomas: highest vs lowest UPF intake: 45% increased EOCRC risk, aOR 1.45 (1.19-1.77), p<0.001

    -- incidence of serrated lesions: no significant relation with UPF intake: aOR 1.04 (0.89-1.22), nonsignificant p=0.48 for trend

        -- findings were consistent after adjusting for BMI, type 2 diabetes, other dietary factors (fiber, folate, calcium and vitamin D) as noted above

        -- the following ultraprocessed foods were statistically associated with higher risk of EOCRC:

            -- sauces, spreads, and condiments: 23% increased risk, OR 1.23 (1.02-1.49), p <0.01 for trend

            -- sugar- or artificially-sweetened beverages: 25% increased risk, OR 1.25 (1.03-1.52), p=0.07  for trend

                -- of note, only artificially-sweetened beverages had a significant 21% increase, OR 1.21 (1.01-1.46), p for trend=0.01 (sugar-sweetened ones were not associated with increased risk)

            -- yogurt, cream cheese, and dairy-based desserts: 21% increased risk, OR 1.21 (1.00-1.46), p=0.10 for trend (though statistically significant)

            -- others (liquor, other artificial sweeteners): 25% increased risk, OR 1.25 (1.02-1.53), p=0.04 for trend

 

-- Findings remained consistent when restricting to EOCRC precursors diagnosed before age 45 years, with similar values as in those >45yo:

    --6% increase in those <45yo: OR 1.06 (1.02-1.10), p<0.01, as a continuous variable per 1 serving increase

-- associations with UPF intake did not differ significantly by histology, polyp size, malignancy risk, or anatomical location

Commentary:

-- The incidence of early-onset colorectal cancers (EOCRCs) diagnosed before age 50 has increased globally, especially in high-income countries

    -- the American Cancer Society (ACS) recommendation for colorectal cancer screening: start at age 45 https://www.cancer.org/cancer/types/colon-rectal-cancer/detection-diagnosis-staging/acs-recommendations.html

    -- the USPSTF also recommends screening all adults aged 45 to 75 for colorectal cancer, with a strong "A" rating for those aged 50-75 and a "B" rating for those aged 45-49 (the "B" rating is that there is a high certainty that the net benefit is moderate or there is moderate certainty that the net benefit is moderate to substantial)

-- EOCRC often occurs in younger people without family history or identifiable genetic predisposition to colorectal cancer, suggesting that there are environmental or lifestyle changes that predispose people with EOCRC

-- known colorectal cancer (CRC) risk factors include obesity, sedentary behavior, type 2 diabetes, and diets low in fiber but high in red and processed meat and added sugars

    -- the consumption of ultraprocessed foods (UPFs) has increased by about 10% of total daily calories in recent decades and now is about 60% of US adult energy intake (https://blogs.cdc.gov/nchs/2025/08/07/7825/#:~:text=A%20new%20NCHS%20report%20shows,and%20other%20serious%20health%20problems)!! And there has been a proliferation of the varieties (and addictiveness) of UPFs

    -- UPFs include industrially-produced ready-to-eat or ready-to-heat products that contain minimal whole foods

    -- overall UPFs are low in bioactive compounds such as fiber, vitamins and polyphenols (polyphenols are especially present in fruits, vegetables, teas and cocoas; they are antioxidants and reduce systemic inflammation), and high in unhealthy fats, refined starches, and food additives such as emulsifiers and artificial sweeteners. These changes disrupt the gut microbiome, increase intestinal inflammation and can lead to increased risk of colorectal cancers

-- this study was undertaken because of the lack of prospective data on the potential role of UPFs, though there has been increasing understanding of the potential mechanistic and epidemiological data

    -- the study included 2 potential mechanisms for increased colon cancer risk: through the sequence of conventional adenoma and serrated adenomas to carcinoma progression

-- this prospective cohort study found that of 29K women, higher consumption of UPFs was associated with increased EOCRC risk at age <50, specifically for women with colonic adenomas but not in those with serrated lesions. Of note, a large meta-analysis by the same authors accessed the Nurses' Health Study, Nurses' Health Study II (the one used in the current study), and also included the Health Professionals Follow-up Study (the latter being a study of men, from back in the days when nurses were women and doctors were men.....), with a mean age of 60yo, finding no difference in both adenomas and serrated lesions in the increased risk of colorectal cancer: https://academic.oup.com/jnci/artdiaicle/115/2/155/6881082. potential reasons:

    -- the microbiome disruptions, proinflammatory signaling and dietary genotoxins may perhaps be more common in younger individuals

    -- for example, the authors note that genomic analyses of colorectal tumors have colibactin-associated signatures (SBS88/ID18), reflecting possible exposure to polyketide synthetase-positive E coli, which are 2.4-4 times more common in EOCRC, especially in distal and rectal tumors. Early-life enrichment of colibactin-producing bacteria could leave enduring variant imprints that accelerate progression of the adenoma-carcinoma pathway (eg see https://www.nejm.org/doi/full/10.1056/NEJMra2200869  and especially https://pubmed.ncbi.nlm.nih.gov/40267983/)

        -- and the epigenetic alterations of the serrated pathway may require more time before resulting in later carcinomas

        -- they did find UPFs being associated with distal colon and rectal adenomas, and the finding that small and low-risk adenomas were found to be associated with UPF further supports the conclusion that detection of nascent polyps on endoscopy in a young cohort is a true precursor of colorectal cancers

-- other issues mechanistically supporting the relation between UPF and colorectal cancer:

    -- feeding studies in adults have found that in matched ad libitum diets, with participants having one diet with UPFs for 2 weeks then switching to the non-UPF diet found that the UPFs were associated with an increase in diet by 508 kcal and 0.9kg weight gain: doi.org/10.1016/j.cmet.2019.05.008

        -- and BMI >30 confers a 2-fold higher risk of colorectal cancer (though this study did not find a BMI association, perhaps because it is only part of the problem? perhaps the increased weight led to more glucose intolerance (not mentioned in the study but quite possible with the weight gain and that was the reason for more EOCRC precursors??), ?other weight-related potential comorbidities responsible for more EOCRC precursors?

    

-- this study adds to several others finding that non-sugar artificial sweeteners are worse than sugar-sweetened ones in terms of many clinical outcomes including diabetes, cardiovascular disease, hypertension and stroke: https://pmc.ncbi.nlm.nih.gov/articles/PMC4846275/

    -- a prior blog done in 2024 on the microbiome and cardiovascular disease reviewed the role of ultraprocessed foods, focusing on emulsifiers but noting that microbiome alterations may be involved in earlier onset colorectal and breast cancers: https://gmodestmedblogs.blogspot.com/2024/07/microbiome-in-cardiovascular-disease.html

    -- another blog reviewed the health effects of non-sugar sweeteners per the WHO, noting increased appetite, increased obesity, diabetes, all-cause mortality, cardiovascular disease, hypertension, end-stage renal disease: https://gmodestmedblogs.blogspot.com/2023/06/health-effects-of-non-sugar-sweeteners.html 

    -- and, another blog reviewed a study finding that soft drinks are associated with major depression disorder attributable to microbiome changes: https://gmodestmedblogs.blogspot.com/2025/12/soft-drinks-increase-depression-by.html

    -- which is not at all to say that sugar is good. just that there is the frequent misperception that non-sugar sweeteners are okay because they are not sugar and have no calories

-- another issue tangentially mentioned in this study is the role of breakfast, a meal notoriously deficient of healthy foods:

    -- we do know that children's cereals are getting nutritionally worse: https://gmodestmedblogs.blogspot.com/2025/07/childrens-cereals-are-getting.html

    -- and, for those who need to lose weight, it turns out that skipping breakfast completely does not lead to eating more at lunch as previously thought: the total energy intake is actually lower by skipping breakfast (https://gmodestmedblogs.blogspot.com/2019/03/breakfast-not-lead-to-weight-loss.html )

Limitations:

-- this rather rigorous study was based on questionnaires, with potential inaccuracies of data

-- some of the adjusted models above in the results are based on the initial baseline data (eg BMI, history of diabetes), which may have changed a lot over the course of the long study, and these changes might not reflect the subsequent reality, affecting the interpretation of the results.

-- this sample of professional largely white women does not reflect the general population, with their high levels of medical sophistication and health consciousness. for example, the UPF consumed by them (34.8% of total calories) is well below the current >60% in the population; the use of alcohol and cigarettes was pretty low; there was not much difference between the 2 groups in the Alternative Healthy Eating Index-2010

    -- the number of incident EOCRC case precursors did track with UPF consumed

    -- BUT, most significantly, this study is using a surrogate marker of pre-cancerous lesions (adenomas) and NOT ACTUAL CLINICAL EVENTS (ie, actual cancers), and we know:

        -- surrogate markers, even pretty well-accepted ones (eg A1c or LDL), do not necessarily stand in for actual clinical events: https://gmodestmedblogs.blogspot.com/2024/06/using-surrogate-markers-for-disease-are.html

        -- some very clear precancerous ones, such as HPV infection in women, do not always reflect true cancer risk:

            -- essentially all cervical cancer is associated with high risk HPV infections: https://pubmed.ncbi.nlm.nih.gov/31500479/

            -- but the vast majority or young women have transient HPV infections that self-resolve, leading for the recommendation not to check routinely for HPV in women younger than age 25

        -- so, do some/most/all colonic adenomas in younger people actually self-resolve, as with HPV??

        -- is the reason for no relationship between serrated lesions and early-onset colorectal cancers in this study explainable by the argument above of using surrogate markers? or do these lesions just not progress to cancer? could serrated lesions also self-resolve in younger people? after all, there are changes in the immune system with aging (https://medlineplus.gov/ency/article/004008.htm): are there enough changes that adenomas at age 40-50 are controlled by the immune system but by age 50 they are not????

-- also, there is some confusion in this study. almost all serrated lesions are called "serrated lesions", though in one case they are called "serrated adenomas". this is confusing. some serrated lesions are not adenomas, just hypertrophic lesions. and some are clearly associated with developing cancers. 

-- the number of actual EOCRC cases (zero reported in the paper, other than the comment that there were too few to have reliable data) does not reflect the incidence of EOCRC in the general population

    -- this is despite the fact that tubular adenomas (>80% of adenomas detected) were included in the precursor lesions for colorectal cancer, and we know that if only one or two tubular adenomas  <1cm are found, there is no increased risk of cancer at 5.5 years in a VA study: https://pubmed.ncbi.nlm.nih.gov/17698067/, with the result that the US Multi-Society Task Force on Colorectal Cancer recommended that further colonoscopy surveillance in these low risk people be 7-10 years (essentially the same as for those with completely normal colonoscopies): https://gmodestmedblogs.blogspot.com/2020/11/colon-cancer-screening-updated.html

-- though this study was based on the Food Frequency Questionnaires of the cohort, there is always the potential for participants to upgrade the quality of foods (as with downgrading the number of cigarettes smoked/alcohol consumed/perhaps UPFs...). and there was no granular data on the preparation of the foods, the brand consumed, or complexity of the foods consumed and their potential interactions, all limiting the accuracy and generalizability of the results (it is always a problem of reductionism in long-term dietary studies of focusing on a single food or supplement, given the complexity of food/supplement interactions in the actual meals eaten and the effects of other foods/supplements interactions that modify the effect of a single food/supplement. there is the example of b-carotene, where this presumably beneficial anti-oxidant has been found in a few studies to lead to an increase in lung cancers in smokers; there are about 50 carotenoids in the human diet and the reductionist approach of just adding b-carotene as a supplement undercuts the potential benefits of the array of carotenoids acting together as well as with other anti-oxidants or other beneficial components in actual diets)

-- we do not have data on the changes in UFPs over the expansive length of the study. how did the UPFs affect risk of EOCRC if they were started or stopped during the long followup? was there a threshold effect where it seemed okay to eat UFPs for a few years? did it matter?

-- this study utilized the current recommendation of screening people >45yo. there are clearly increases in those younger than that (as also found of EOCRC precursors in this study). we may well need to assess younger people. perhaps at some point in the near future?

    -- though colonoscopy does have adverse effects, these are found much more often in older individuals

so,

-- a rather disturbing article reflecting the rather large amount of ultraprocessed food consumption in the US (and globally), bringing up several issues:

    -- the food industry has been very successful in finding very tasty highly processed foods (lots of salt, sugar and lipids), leading to an increasing consumption trend, now to over 60% of daily calories in the US from UPFs: https://blogs.cdc.gov/nchs/2025/08/07/7825/#:~:text=A%20new%20NCHS%20report%20shows,and%20other%20serious%20health%20problems.

    -- this increase is associated with supermarket complicity (having the very profitable UPFs at eye level in the aisles)

    -- and the unfortunate decrease in home cooking (these high calorie UPFs are easy to get and eat, or just heat up for a few minutes), our lives are perhaps more stressful, parent(s) may well be working and not have the time to cook, and the future generations do not have the habit/knowledge of preparing healthy foods

-- and this study does bring up several other issues, eg:

    -- this study found similar results in those 40-45yo vs 45-50yo. should we be screening for colon cancer earlier?

    -- is FIT testing as sensitive and specific in the younger group? (though, of course, this study was pre-FIT testing)?

    -- it was notable that the group with the highest UPF consumption also had the most aspirin/NSAIDs consumption, which as noted above is actually protective against colorectal cancer. this skew brings up the possibility that UPFs by themselves are actually even more harmful. By the way, the 2016 USPSTF included low-dose aspirin in their 2016 colon cancer prevention guidelines, as well as direct colorectal cancer screening: https://gi.org/2016/04/11/low-dose-aspirin-for-prevention-colorectal-cancer-new-uspstf-recommendations/ , though subsequent recommendations eliminated the aspirin component given the increased risk of dangerous bleeding from the aspirin

    -- it is important to realize that there are many cancers besides colorectal ones that are found in younger people, including the earlier onset of cancers of the breast, pancreas, esophagus, gallbladder, stomach, endometrium, kidney, prostate, head and neck, and thyroid:  https://pmc.ncbi.nlm.nih.gov/articles/PMC12446525/ , and https://gmodestmedblogs.blogspot.com/2024/08/earlier-age-of-onset-in-17-cancers.html)

        -- i suspect that there are several very important risk factors that play a role in these early and later onset of cancers besides the ultraprocessed foods, including increased microplastics in our bodies (https://gmodestmedblogs.blogspot.com/2024/03/plastics-too-many-and-too-bad-for-our.html), as well as water/air/ground pollution, increasing social anomy/alienation, increasing occupational exposures.....these all can lead to a chronic inflammatory state, which predisposes people to developing cancer (and many of the other chronic diseases we are seeing, including diabetes, cardiovascular disease, chronic kidney disease, MAFLD, etc: https://www.nature.com/articles/s41591-019-0675-0

        -- in terms of air pollution, there is ample evidence that it is associated with chronic inflammatory states, and is related to lung cancer, breast cancer, GI malignancies, liver cancer, brain tumors, and ovarian cancer: https://bmjoncology.bmj.com/content/4/1/e000535 

-- this study reinforces the importance for us in primary care to counter the common popular misinformation that non-sugar, lower calorie artificial sweeteners are useful for patients with diabetes, overweight, etc. These are actually worse than actual sugar (which should also be reduced), per this and many other studies

-- and, of course, this all reinforces rather dramatically our need for a functional public health system. we need a public health system that strongly addresses and is empowered to act on these adverse social issues that are clearly associated with the huge morbidity/mortality increases we are all seeing all the time in our patients. unfortunately, this pretty evident public health imperative is being reversed through the current increasingly anti-regulation, anti-public health political climate. that is another huge climate change we as a country need to tackle...