Anti-Inflammatory Diet: Foods to Avoid

Adding anti-inflammatory foods is only half of the equation. The other half is removing the foods that aggressively drive inflammation higher — and on a dose-response basis, removing pro-inflammatory foods often produces a larger measurable change in serum high-sensitivity C-reactive protein (hs-CRP) and interleukin-6 (IL-6) than adding any single anti-inflammatory food. The 2019 Hall metabolic-ward inpatient trial demonstrated that an ultra-processed diet, matched calorie-for-calorie with a minimally processed diet, caused participants to spontaneously eat 500 extra calories per day and gain weight, while the same people on the minimally processed diet lost weight. Six food categories are the usual targets for reduction in an anti-inflammatory pattern: ultra-processed packaged foods, industrial seed oils, refined sugar and high-fructose corn syrup, processed and charred red meat, artificial trans fats, and excess alcohol. This page maps each category to its inflammatory mechanism and lists the swaps people following the pattern commonly make.


Table of Contents

  1. Ultra-Processed Foods (NOVA Category 4)
  2. Industrial Seed Oils and the Linoleic Acid Debate
  3. Refined Sugar and High-Fructose Corn Syrup
  4. Processed and Charred Red Meat
  5. Artificial Trans Fats
  6. Excess Alcohol
  7. Refined Grains and High Glycemic Loads
  8. Practical Swaps for Each Category
  9. Key Research Papers
  10. Connections
  11. Featured Videos

Ultra-Processed Foods (NOVA Category 4)

The NOVA classification, developed by Carlos Monteiro and colleagues at the University of São Paulo, sorts foods by degree of industrial processing rather than by nutrient composition. Category 4 — ultra-processed — refers to formulations of substances extracted from foods (oils, fats, sugars, starches, proteins), modified with industrial processes (hydrogenation, hydrolysis, extrusion, pre-frying), and combined with cosmetic and palatability additives (flavors, colors, emulsifiers, sweeteners, thickeners). Examples include soft drinks, packaged snacks, mass-produced bread and confectionery, instant noodles, reconstituted meats (hot dogs, chicken nuggets), and most ready-to-heat frozen meals.

The Hall et al. 2019 inpatient metabolic-ward trial, published in Cell Metabolism, is the most rigorous human test of the ultra-processed hypothesis. Twenty adults lived on the U.S. National Institutes of Health (NIH) metabolic ward for 4 weeks, randomized to alternating 2-week blocks of ultra-processed and minimally processed diets, matched calorie-for-calorie, macro-for-macro, and for fiber, sugar, and sodium. On the ultra-processed diet, participants spontaneously consumed about 508 extra calories per day and gained 0.9 kg in 2 weeks. On the minimally processed diet, the same participants lost 0.9 kg in 2 weeks. The drivers appear to be (1) faster eating rate (ultra-processed foods are softer and easier to chew), (2) lower satiety hormone response (lower peptide YY (PYY)) per calorie, and (3) higher palatability triggering hedonic overconsumption.

The inflammatory and chronic-disease epidemiology of ultra-processed food is now extensive. The NutriNet-Santé cohort (Srour et al. BMJ 2019; 105,159 French adults followed for a median of 5.2 years) found that each 10-point rise in the percentage of ultra-processed foods in the diet was associated with 12% higher cardiovascular disease risk (hazard ratio 1.12), 13% higher coronary heart disease risk and 11% higher cerebrovascular disease risk. A 2024 BMJ umbrella review of 45 pooled analyses (Lane et al.) graded the evidence: convincing for links between higher ultra-processed intake and cardiovascular-disease death (risk ratio 1.50), type 2 diabetes, anxiety and common mental disorders; highly suggestive for all-cause death (risk ratio 1.21), heart-disease death, depression, poor sleep, wheezing and obesity. No cancer link reached the convincing or highly suggestive grades.

A common rule of thumb: a packaged food with more than five ingredients, with ingredients not recognisable as food, or with industrial additives (carrageenan, polysorbate, soy lecithin, mono- and diglycerides, high-fructose corn syrup, hydrolyzed protein, "natural flavors") is likely NOVA category 4. People following the pattern swap these for minimally processed equivalents from the same culinary niche: whole-grain bread with a 4-ingredient label instead of mass-market bread; popcorn instead of chips; plain yogurt with fruit instead of flavored yogurt; nuts instead of granola bars.

Back to Table of Contents

Industrial Seed Oils and the Linoleic Acid Debate

The omega-6 / omega-3 ratio problem (covered in detail on the omega-3 page) is driven primarily by industrial seed oils — soybean, corn, sunflower, safflower, cottonseed, grapeseed — that supply approximately 8–12% of total calories in the modern American diet. U.S. food-supply data show that soybean oil consumption rose more than a thousand-fold between 1909 and 1999, and linoleic acid rose from 2.79% to 7.21% of energy over the same period (Blasbalg 2011). These oils are 50–70% linoleic acid (the parent omega-6 fatty acid), pressed and refined with hexane, deodorized with high-temperature steam, and oxidatively damaged in the process.

The mechanistic concerns are: (1) linoleic acid competes with omega-3s for the same desaturase and elongase enzymes, suppressing endogenous eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA) synthesis from alpha-linolenic acid (ALA); (2) industrial processing produces oxidized linoleic acid metabolites (OXLAMs — 9-HODE, 13-HODE, 4-HNE) which are pro-atherogenic and pro-inflammatory; (3) repeated heating in restaurant deep-fryers produces additional aldehydes, polymers, and trans isomers.

The clinical evidence is mixed because most large epidemiology studies have shown a modest benefit of replacing saturated fat with polyunsaturated fat (the original American Heart Association (AHA) recommendation). However, the reanalysis of historic randomized trials by Ramsden and colleagues (Sydney Diet Heart Study, Minnesota Coronary Experiment) suggested that the low-density lipoprotein-lowering effect of seed oils was not accompanied by reduced mortality — and may have increased it in some subgroups, possibly due to oxidation. In the recovered Minnesota Coronary Experiment data (1968–73), the diet that replaced saturated fat with linoleic-acid-rich corn oil lowered serum cholesterol by 13.8%, against 1.0% on the control diet, yet showed no mortality benefit; each 30 mg/dL fall in serum cholesterol went with a 22% higher risk of death (hazard ratio 1.22). The same paper's meta-analysis of five randomized trials (10,808 people) found no benefit for death from coronary heart disease (risk ratio 1.13) or from all causes (risk ratio 1.07). People following an anti-inflammatory pattern commonly replace seed oils with extra-virgin olive oil and small amounts of cold-pressed avocado oil for cooking and dressings, with butter or ghee used sparingly for high-heat cooking.

The replacement of restaurant deep-fried foods (French fries, fried chicken, fried fish, doughnuts, tempura) with home-cooked equivalents in olive oil or oven-baking is the single largest practical lever for seed oil reduction. Restaurant fryers are the densest source of OXLAMs in the modern food supply.

Back to Table of Contents

Refined Sugar and High-Fructose Corn Syrup

Refined sucrose is 50% glucose and 50% fructose. High-fructose corn syrup (HFCS-55, the common soft-drink formulation) is 55% fructose and 45% glucose. From an inflammatory standpoint, the fructose half is the more concerning. Unlike glucose, fructose is metabolized almost exclusively in the liver, where high doses (above ~50 g/day) saturate the normal glycolytic pathway and shunt excess substrate into hepatic de novo lipogenesis — producing triglycerides, very-low-density lipoprotein (VLDL), and a non-alcoholic fatty liver disease (NAFLD) phenotype within weeks.

The Stanhope and colleagues 2009 trial (J Clin Invest) gave overweight and obese adults either fructose-sweetened or glucose-sweetened beverages providing 25% of energy requirements for 10 weeks. Both groups gained similar weight, but only the fructose group showed more visceral (deep belly) fat, more fat-making in the liver (de novo lipogenesis), higher fasting apoB, low-density lipoprotein (LDL), small dense LDL and oxidized LDL, and lower insulin sensitivity. A 2012 cohort study of 42,883 U.S. men (de Koning et al., Circulation) found that sugar-sweetened beverage intake was associated with higher plasma triglycerides, C-reactive protein, interleukin-6 and tumor necrosis factor receptors 1 and 2; artificially sweetened beverages showed no such association.

Two limits come up most often: (1) free sugars under 5% of total energy intake (~25 g/day for most adults — the World Health Organization (WHO) conditional recommendation), and (2) the anti-inflammatory-diet target of sugar-sweetened beverages at zero or near-zero, replaced by water, sparkling water, unsweetened tea, or coffee. Whole-fruit fructose is not the problem — the dose per serving is low (~10 g in a whole apple), the fiber slows absorption, and the polyphenols and micronutrients offset any small metabolic cost. The problem is concentrated fructose: sodas, fruit juices, sweetened coffee drinks, candy, baked goods, and the long list of processed foods containing HFCS as the third or fourth ingredient.

Back to Table of Contents

Processed and Charred Red Meat

The 2015 International Agency for Research on Cancer (IARC) Monograph 114 classified processed meat (bacon, sausage, hot dogs, deli ham, salami, pepperoni, jerky — defined as meat transformed through salting, curing, fermentation, smoking, or other processes to enhance flavor or preservation) as Group 1: carcinogenic to humans, on evidence for colorectal cancer. Each 50 g daily serving of processed meat (one hot dog, two slices of bacon) was estimated to raise colorectal cancer risk by ~18%. Red meat itself was classified Group 2A (probably carcinogenic), with weaker but still concerning evidence for colorectal, pancreatic, and prostate cancer.

Three mechanisms appear to converge: (1) nitrites and nitrates added as preservatives form N-nitroso compounds in the gastric environment, several of which are direct mutagens; (2) heme iron from red meat catalyzes lipid peroxidation in the colon and contributes to oxidative DNA damage; (3) high-temperature cooking (grilling, pan-searing, broiling at >200°C) produces heterocyclic amines (HCAs) from amino acid and creatine pyrolysis, and polycyclic aromatic hydrocarbons (PAHs) from fat dripping onto flame. Both HCA and PAH activate the aryl hydrocarbon receptor (AhR), increase oxidative stress, and drive low-grade systemic inflammation.

Harm-reduction steps described for people who continue to eat red meat: (1) fresh whole-muscle cuts in place of processed meat; (2) lower-temperature cooking — braising, stewing, sous-vide, slow-cooking — rather than grilling and pan-searing; (3) marinating with rosemary, oregano, lemon, vinegar, or wine before high-temperature cooking, which reduces HCA formation 50–90% (covered on the spices page); (4) portions of ~100–150 g (palm-of-hand) no more than 2–3 times per week; (5) red meat eaten with abundant cruciferous vegetables, which upregulate phase II detoxification of HCAs.

Back to Table of Contents

Artificial Trans Fats

Artificial (industrial) trans fats are produced by partial hydrogenation of vegetable oils — an industrial process that converts liquid oils into semi-solid fats (margarine, shortening, baked-good fats) by chemically rearranging some cis-double-bonds to trans configuration. The Mozaffarian 2006 NEJM review estimated that artificial trans fats at 2% of energy intake increased coronary heart disease risk by ~23%, with mechanisms including raised low-density lipoprotein, lowered high-density lipoprotein (HDL), increased lipoprotein(a) (Lp(a)), increased systemic inflammation (high-sensitivity C-reactive protein, interleukin-6, tumor necrosis factor-alpha), and impaired endothelial function.

In 2015 the U.S. Food and Drug Administration (FDA) determined that partially hydrogenated oils are no longer "Generally Recognized as Safe." For most uses, manufacturers could not add them to foods after June 18, 2018, and the final compliance date for products already in distribution was January 1, 2021, so industrial trans fats have largely disappeared from packaged foods sold in the United States. Small amounts remain in (1) ruminant trans fats naturally present in dairy and beef fat (these are biochemically different and not clearly harmful at typical intakes), and (2) imported products from countries without trans fat regulations.

On labels, any product listing "partially hydrogenated" oil in the ingredient list contains trans fats, even if the "Trans Fat" line on the Nutrition Facts panel says 0 g (U.S. Food and Drug Administration allows rounding to zero if the amount is under 0.5 g per serving, and serving sizes can be gamed). Imported pastries, cookies, and crackers from regions with weaker regulation are where such oils can still turn up.

Back to Table of Contents

Excess Alcohol

Alcohol's relationship to inflammation is dose-dependent and tissue-dependent. Light drinking (1 drink per day for women, 1–2 for men) was historically associated with lower high-sensitivity C-reactive protein and lower cardiovascular event rates in observational studies — the J-shaped curve. The 2018 Lancet Global Burden of Disease alcohol analysis (Griswold et al.) reframed this by including cancer, liver disease, tuberculosis, and injury endpoints: the safest level of alcohol consumption for overall health is zero, and risk increases monotonically from there.

Mechanistically, alcohol at higher doses drives inflammation by (1) compromising the gut epithelial barrier and increasing lipopolysaccharide (LPS) translocation (the "leaky gut" mechanism behind alcoholic liver disease); (2) directly inducing hepatic Kupffer cell activation; (3) generating acetaldehyde, a Group 1 carcinogen; (4) suppressing immune surveillance against early cancers, particularly in the upper aerodigestive tract.

How anti-inflammatory diet writers usually treat it: for a person who already drinks 1 small glass of wine with meals and has no personal or family history of breast cancer or alcohol use disorder, continuing is commonly described as tolerable, while the Global Burden of Disease analysis above found the lowest overall risk at zero. Above 1 drink per day for women or 2 for men, reduction is reported to produce measurable high-sensitivity C-reactive protein improvement within weeks. Above 3 drinks per day, the risks described in that analysis climb further, regardless of inflammation.

Back to Table of Contents

Refined Grains and High Glycemic Loads

Refined white flour, white rice, and the products made from them (white bread, white pasta, pastries, breakfast cereals, crackers, pretzels) are stripped of bran and germ during milling. The result is rapid glucose absorption, large postprandial glucose excursions, compensatory insulin spikes, reactive hypoglycemia, and downstream effects on the inflammatory pathways already discussed: glycation pressure, hepatic de novo lipogenesis from carbohydrate overflow, and feedback on hunger and satiety hormones.

In the Prospective Urban and Rural Epidemiology (PURE) study of 137,130 people in 21 countries (Swaminathan et al. BMJ 2021), eating 350 g or more of refined grains a day (about 7 servings) was associated with 27% higher total mortality (hazard ratio 1.27) and 33% more major cardiovascular events (hazard ratio 1.33), compared with less than 50 g a day; higher refined-grain intake also went with higher systolic blood pressure. In PURE, intakes of whole grains and of white rice showed no significant association with these outcomes. A 2016 BMJ meta-analysis of prospective studies (Aune et al.) found that each 90 g/day of whole grains (three servings) was associated with 22% lower cardiovascular disease risk, 17% lower all-cause mortality and 15% lower total cancer mortality.

The usual swaps: stone-ground whole-wheat or sprouted-grain bread in place of white bread; whole-grain or legume-based pasta in place of white pasta; brown rice, farro, bulgur, or quinoa in place of white rice; steel-cut or rolled oats in place of breakfast cereals. The Mediterranean staples covered on the foundations page — bulgur, farro, barley — are the historical templates.

Back to Table of Contents

Practical Swaps for Each Category

The cumulative effect of these swaps, sustained over 3–6 months, typically produces measurable reductions in high-sensitivity C-reactive protein (often 30–50%), fasting insulin, triglycerides, and waist circumference, without explicit calorie counting. The pattern works because each swap simultaneously removes a pro-inflammatory input and adds an anti-inflammatory one from the same culinary niche — the patient never feels deprived of food category, only redirected within it.

Back to Table of Contents

Key Research Papers

  1. Hall KD et al., Ultra-Processed Diets Cause Excess Calorie Intake and Weight Gain: An Inpatient Randomized Controlled Trial (Cell Metab 2019) — PubMed PMID: 31105044
  2. Srour B et al., Ultra-processed food intake and risk of cardiovascular disease: prospective cohort study (NutriNet-Santé, BMJ 2019) — PubMed PMID: 31142457
  3. Bouvard V, Loomis D, Guyton KZ et al.; International Agency for Research on Cancer Monograph Working Group, Carcinogenicity of consumption of red and processed meat (IARC Monograph 114; Lancet Oncol 2015) — PubMed PMID: 26514947
  4. Mozaffarian D, Katan MB, Ascherio A et al., Trans Fatty Acids and Cardiovascular Disease (NEJM 2006) — PubMed PMID: 16611951
  5. Stanhope KL et al., Consuming fructose-sweetened, not glucose-sweetened, beverages increases visceral adiposity and lipids (J Clin Invest 2009) — PubMed PMID: 19381015
  6. Ramsden CE et al., Re-evaluation of the traditional diet-heart hypothesis: analysis of recovered data from Minnesota Coronary Experiment (BMJ 2016) — PubMed PMID: 27071971
  7. Griswold MG et al., Alcohol use and burden for 195 countries and territories, 1990-2016 (Lancet GBD, 2018) — PubMed PMID: 30146330
  8. Aune D et al., Whole grain consumption and risk of cardiovascular disease, cancer, and all cause mortality: systematic review and dose-response meta-analysis (BMJ 2016) — PubMed PMID: 27301975
  9. Lustig RH, Schmidt LA, Brindis CD, Public health: The toxic truth about sugar (Nature 2012) — PubMed PMID: 22297952
  10. Lane MM et al., Ultra-processed food exposure and adverse health outcomes: umbrella review (BMJ 2024) — PubMed PMID: 38418082
  11. Monteiro CA et al., The UN Decade of Nutrition, the NOVA food classification and the trouble with ultra-processing (Public Health Nutr 2018) — PubMed PMID: 28322183
  12. Swaminathan S, Dehghan M, Raj JM et al., Associations of cereal grains intake with cardiovascular disease and mortality across 21 countries in Prospective Urban and Rural Epidemiology study: prospective cohort study (BMJ 2021) — PubMed PMID: 33536317
  13. de Koning L, Malik VS, Kellogg MD, Rimm EB, Willett WC, Hu FB, Sweetened beverage consumption, incident coronary heart disease, and biomarkers of risk in men (Circulation 2012) — PubMed PMID: 22412070
  14. Blasbalg TL, Hibbeln JR, Ramsden CE, Majchrzak SF, Rawlings RR, Changes in consumption of omega-3 and omega-6 fatty acids in the United States during the 20th century (Am J Clin Nutr 2011) — PubMed PMID: 21367944
  15. U.S. Food and Drug Administration, Trans Fat (partially hydrogenated oils: June 18, 2018 and January 1, 2021 dates) — fda.gov

PubMed Topic Searches

  1. PubMed: Ultra-processed cardiovascular
  2. PubMed: Seed oils inflammation
  3. PubMed: HFCS hepatic
  4. PubMed: Red meat colorectal
  5. PubMed: Trans fat mortality
  6. PubMed: Alcohol GBD

Back to Table of Contents

Connections

Back to Table of Contents