Dietary Inflammatory Index Food List: What Is Scored
The Dietary Inflammatory Index scores 45 nutrients, spices and flavonoid groups, not foods. Each direction, the foods carrying it, and what is missing.
The Dietary Inflammatory Index (DII) does not contain a list of foods. It scores 45 dietary parameters — mostly nutrients, plus a few herbs, spices, tea and flavonoid groups — and gives each one a direction, anti-inflammatory or pro-inflammatory, drawn from published associations between that parameter and inflammatory markers measured in blood. The everyday foods below are the ones that carry those parameters, which is why so many foods people expect to find are simply absent.
How the index decides which direction a parameter goes
The index was assembled by reviewing published research linking dietary components to six inflammatory biomarkers: C-reactive protein, interleukin-1 beta, interleukin-4, interleukin-6, interleukin-10 and tumor necrosis factor alpha. Two of those, interleukin-4 and interleukin-10, are themselves anti-inflammatory, so raising them counted toward the anti-inflammatory side.
Each parameter came out of that review with an overall effect score, negative for anti-inflammatory and positive for pro-inflammatory. To score a person, their intake of each parameter is compared against a reference distribution built from dietary data collected in several countries, converted to a percentile, multiplied by the parameter's effect score, and summed. The result is one number for that person's diet: below zero on the anti-inflammatory side, above zero on the pro-inflammatory side.
Two consequences matter here. The input is a whole diet — habitual intake from a questionnaire or several days of food records, not a plate. And a whole food appears only where the authors created a parameter for it, which happened for a handful of herbs, spices and tea and nothing else. What the Dietary Inflammatory Index is covers the method rather than the output.
Parameters scored in the anti-inflammatory direction
| Parameter | Direction | Everyday foods that carry it | Effect on a diet-level score | Caveat on measurement |
|---|---|---|---|---|
| Fiber | Anti-inflammatory | Beans, lentils, oats, barley, vegetables | Pulls down, on one of the larger weights | Databases disagree on fiber values |
| Omega-3 fatty acids | Anti-inflammatory | Salmon, sardines, mackerel, walnuts, flaxseed | Pulls down | Needs the type of fish, often missed |
| Omega-6 and polyunsaturated fat | Anti-inflammatory | Corn, soybean and sunflower oil, nuts, seeds | Pulls down, contradicting most seed-oil advice | Follows the biomarker literature, not food lists |
| Monounsaturated fat | Anti-inflammatory | Olive oil, avocado, almonds, canola oil | Pulls down, on a small weight | Hard to split from total fat, scored the other way |
| Magnesium, zinc, selenium | Anti-inflammatory | Pumpkin seeds, spinach, black beans, oysters, Brazil nuts | Pull down | Selenium varies with the soil |
| Vitamins A, C and E, beta-carotene | Anti-inflammatory | Sweet potato, carrots, greens, citrus, peppers, almonds | Pull down | Degrade with storage, light and cooking |
| Vitamin D | Anti-inflammatory | Oily fish, egg yolk, fortified milk and cereal | Pulls down | Sun exposure is invisible to a food record |
| Thiamin, riboflavin, niacin, B6, folate | Anti-inflammatory | Whole grains, legumes, greens, meat, enriched flour | Pull down modestly | US enrichment shifts the score of the same bread |
| Flavan-3-ols, flavones, flavonols, flavanones, anthocyanidins, isoflavones | Anti-inflammatory | Tea, cocoa, onions, apples, citrus, berries, soy | Pull down | Need a flavonoid database few studies have |
| Green and black tea | Anti-inflammatory | Brewed tea of either kind | Pulls down | Brew strength is rarely recorded |
| Garlic, onion, ginger, turmeric, saffron, black pepper, thyme and oregano, rosemary, eugenol | Anti-inflammatory | The herbs and spices themselves; eugenol from cloves | Pull down; turmeric carries one of the strongest weights | Few questionnaires ask how much saffron someone ate |
| Caffeine | Anti-inflammatory | Coffee, tea | Pulls down slightly | Scored as caffeine, not as coffee |
| Alcohol | Anti-inflammatory | Beer, wine, spirits | Scores anti-inflammatory, which surprises most readers | From biomarker findings at moderate intakes; the index models one narrow literature, not overall health |
Parameters scored in the pro-inflammatory direction
| Parameter | Direction | Everyday foods that carry it | Effect on a diet-level score | Caveat on measurement |
|---|---|---|---|---|
| Total energy | Pro-inflammatory | Any large portion | Pushes up | An energy-adjusted version scores per 1,000 calories |
| Saturated fat | Pro-inflammatory | Butter, fatty meat, cheese, cream, coconut oil | Pushes up, the largest positive weight among fats | Fat intake is poorly reported |
| Trans fat | Pro-inflammatory | Partially hydrogenated oils, largely gone in the US | Pushes up | Finds almost nothing in recent US data |
| Total fat | Pro-inflammatory | Any fat-heavy meal | Pushes up | Sits awkwardly beside fats scored the other way |
| Cholesterol | Pro-inflammatory | Eggs, shellfish, organ meats | Pushes up | Dietary and blood cholesterol differ |
| Carbohydrate | Pro-inflammatory | Bread, rice, pasta, potatoes, sugar, soda | Pushes up | Table sugar and intact whole grain count alike |
| Protein | Pro-inflammatory | Meat, dairy, eggs, legumes | Pushes up weakly | Plant protein counts the same as animal |
| Iron | Pro-inflammatory | Red meat, fortified cereal, legumes | Pushes up weakly | Heme and non-heme are not separated |
| Vitamin B12 | Pro-inflammatory | Meat, fish, dairy, fortified foods | Pushes up | The only B vitamin on this side |
Foods people expect to find, and why they are not there
| Food | Why it is absent | Closest parameter that does cover it |
|---|---|---|
| Red meat | No meat parameter; the index scores nutrients | Saturated fat, cholesterol, iron, B12 |
| Processed meat | Same; curing agents are not modeled | Saturated fat, total energy |
| Added sugar | No added-sugar parameter, though the US panel lists one | Carbohydrate, total energy |
| Salt and sodium | Not among the 45 parameters at all | Nothing covers it |
| Ultra-processed foods | A manufacturing category, not a nutrient | Whatever nutrients the product delivers |
| Gluten and wheat | No parameter for either | Carbohydrate, fiber |
| Dairy | No parameter for milk, cheese or yogurt | Saturated fat, protein, B12 |
| Seed oils | No parameter for any single oil | Polyunsaturated and omega-6 fat, both anti-inflammatory |
| Broccoli and most vegetables | Only onion and garlic got parameters | Fiber, vitamin C, folate, carotenoids, flavonols |
| Coffee | No coffee parameter | Caffeine, part of the flavonoid load |
A food missing from the index still moves a score, but only through the nutrients it delivers. The complete list of inflammatory foods by category covers the foods themselves, which is usually the real question.
What a Nutrition Facts panel does and does not map onto
You cannot read a DII score off a package, and the reason is worth knowing precisely. The US Nutrition Facts panel is mandatory and fairly detailed: calories, total fat, saturated fat, trans fat, cholesterol, sodium, total carbohydrate, dietary fiber, total sugars, added sugars, protein, and four micronutrients.
Line that up against the 45 parameters and the overlap is lopsided. Calories, total fat, saturated fat, trans fat, cholesterol, carbohydrate, protein and iron are all parameters, and every one of them scores in the pro-inflammatory direction. Only two panel lines score the other way: dietary fiber, which carries one of the heavier anti-inflammatory weights, and vitamin D. The rest of the anti-inflammatory side — magnesium, zinc, selenium, most vitamins, every flavonoid subclass, tea, herbs and spices — is not on a label at all.
Two lines mislead in the other direction. Sodium is among the most prominent numbers on a US panel and is not a parameter in any form, so a very salty product loses nothing here. Added sugars, a line the US requires and most other countries' labels do not carry, has no parameter either; sugar enters only through carbohydrate and total energy. And a panel is per serving, at sizes set by regulation rather than by what you ate, while the index scores habitual daily intake.
A diet-level score is not a verdict on one food or one meal
The published index scores parameters, not foods. Any table giving a DII number for blueberries or for pizza is somebody's derivation from nutrient data rather than part of the instrument, so treat such values as illustrative. It cannot score a meal either: a score describes habitual intake over a period, because that is what the reference distribution was built from.
Most importantly, this is an epidemiological tool. It exists so researchers can rank groups of people by the estimated inflammatory potential of their diets and look for associations across a population. There is no validated cut-off saying an individual is or is not inflamed, and a score is not a diagnostic result. Systems built for this job also disagree with each other; why food inflammation scores disagree sets several side by side.
Why the amount matters, and what actually moves a score
Total energy is itself a parameter, on the pro-inflammatory side. Eat more of everything and you take in more of the anti-inflammatory parameters as well, while pushing the score up through energy at the same time. That is why an energy-adjusted version exists: intake is expressed per 1,000 calories and compared against a reference adjusted the same way, so the question becomes what a diet is made of rather than how much of it there is. Both versions are called a DII score in published work, and they are not the same measure.
How far any single parameter moves a score depends on two things. First, the effect weight it came out of the literature review with: fiber, saturated fat and turmeric weigh considerably more than protein or iron. Second, the percentile step, which bounds what any parameter can contribute at either end. In practice, going from almost none of something to an ordinary amount moves the score noticeably, while going from a lot to a great deal barely registers.
The index therefore rewards closing a gap rather than doubling down on something already covered — a property of how the score is constructed rather than anything visible in the number, and the clearest reason a small change in the right place can outweigh a large one somewhere else.
Where the index is least precise
Most published studies use fewer than all 45 parameters, because ordinary questionnaires cannot capture saffron, eugenol or individual flavonoid subclasses, and scores built from different parameter counts are not strictly comparable. The comparison is relative too: intake is scored against a multi-country reference, so the same breakfast lands at a different percentile depending on the reference used. And each direction reflects the research that existed when the review was done — a reasonable way to build an instrument, a poor way to settle an argument about one food. The inflammatory versus anti-inflammatory foods chart shows where everyday sorting and research-based sorting part ways.
What research using the index has and has not found
The index has been applied mostly in observational research. Cohorts in different countries have generally pointed the same way: people whose diets fall on the pro-inflammatory side show, on average, somewhat higher rates of several chronic conditions. The associations are usually modest, and they come from data in which diet does not travel alone — smoking, physical activity, income and education all correlate with diet and with the outcomes being counted. Statistical adjustment shrinks that problem without removing it.
Two limitations matter more for a skeptical reading than the results themselves. Many cohorts measure diet once, at enrollment, sometimes years before the outcomes they later count, so the idea that a person went on eating roughly that way is an assumption rather than a measurement. And studies checking the score directly against measured inflammatory markers in the same people have come out less consistent than the literature on outcomes — worth pausing on, because that link is the assumption the instrument rests on. Taken together, this is a serviceable tool for ranking groups and a poor basis for a claim about one person.
Using the list as a direction rather than a ranking
Read practically, the parameter list points where most dietary advice already points: more fiber, more plants, more herbs and spices, more of the fats in fish and nuts, less saturated fat, smaller totals. It tells you almost nothing about your own body, and closing that gap takes observation over time — what you ate, how you felt afterwards, recorded consistently enough that repetition separates itself from noise. That is the narrow job NutriAI does. You photograph a meal, the app estimates the foods and portions and gives a letter grade for estimated inflammatory potential from one consistent framework, and you log symptoms afterwards. The estimate carries real error bars and is weakest on oils, dressings, sauces and sugar in mixed dishes. The grade is not a clinical measurement, and any pattern the app surfaces is a correlation inside your own log rather than a diagnosis.
When this stops being a scoring question
Food scores and food logs suit ordinary, non-alarming symptoms. Some signs belong with a doctor instead: unintended weight loss, blood in the stool, persistent fever, joint swelling with redness and heat, severe or waking pain, or symptoms that worsen steadily rather than fluctuate. If you have a diagnosed inflammatory condition or take medication for one, discuss dietary changes with your doctor or a registered dietitian first.
What this is based on
- Published methodology describing how the Dietary Inflammatory Index was constructed, including the literature review of dietary components against inflammatory biomarkers and the parameter scoring procedure
- Peer-reviewed research on associations between dietary components and circulating inflammatory markers such as C-reactive protein and interleukin-6
- Epidemiological studies applying the index to cohort data, including their accounts of how many parameters their dietary assessment tools could supply
- Food composition reference data for the nutrient content of the foods named in the tables
- General clinical guidance on symptoms that warrant medical assessment rather than self-tracking
Frequently asked questions
- Is there an official Dietary Inflammatory Index food list?
- No. The official object is a list of 45 dietary parameters: nutrients, a few herbs and spices, tea, and flavonoid subclasses. Foods enter only through the parameters they carry, so any food list you see is somebody's translation of the parameter list into groceries rather than part of the published instrument.
- What is the DII score of a specific food, such as blueberries?
- The published index does not assign scores to individual foods. Per-food numbers circulating online are third-party derivations calculated from nutrient data, and different people making that calculation arrive at different answers. Treat any such value as illustrative rather than official.
- Why do seed oils score in the anti-inflammatory direction?
- Because the omega-6 and polyunsaturated fat parameters both received anti-inflammatory scores from the biomarker literature the index was built on. This is a real disagreement between the index and most popular food lists, and it is worth knowing before treating a DII score as confirmation of either view.
- Is salt included in the Dietary Inflammatory Index?
- No. Sodium is not one of the parameters in any form. A diet high in salt can still score well, which is one of the clearest reasons not to read a DII score as a general judgement on how healthy a diet is.
- Can I calculate my own DII score?
- Doing it properly needs complete nutrient intake data plus the reference distribution the index compares against, which is why it is normally done inside a study rather than by an individual. Approximation tools exist, but they are not the validated instrument and none of them are diagnostic.