Mediterranean Diet and Type 2 Diabetes: What the Research Shows, Foods, and a Sample Week

Quick answer: A Mediterranean-style way of eating is one of several patterns the American Diabetes Association (ADA) lists for people with diabetes, and the trials I could read suggest it may lower A1C by a few tenths of a percentage point compared with other diets. It is not a magic switch, and it does not remove the need to watch portions of the carbohydrate-containing foods it is full of: beans, whole grains, bread and fruit. The most useful way to use this guide is as a food map and a sample week, then check your own readings with your care team.

What this guide covers and what it does not

This is a plain-language look at the Mediterranean diet for adults with type 2 diabetes or prediabetes. I summarize what published studies report, say what I could and could not read in full, and give you a sample week with approximate carbohydrate totals that I calculated from the food data shown in the tables. I am Charles Kirkland, not a physician or a dietitian, so nothing here is a treatment plan, and I do not give advice on medicines. If you take glucose-lowering medicine, big changes in how you eat can change how much of it you need, so it is worth telling your care team before you change your plate, not after.

I have also not tested any of these meals on myself and I do not report any personal glucose results in this article. Where I describe study results, I give the design, the size and the length as best I could see them, and I mark anything I could only read in a summary. For the broader picture of how food fits into diabetes care, see the diet for type 2 diabetes guide and the diabetes food guide. If you want to see how this pattern stacks up against other approaches, the diabetes diets compared page puts them side by side.

What “Mediterranean diet” actually means

There is no single official Mediterranean diet. The phrase describes the traditional eating habits of people living around the Mediterranean Sea, and researchers have defined it in different ways from study to study. That matters, because when someone says “the Mediterranean diet worked,” the first question is which version of it the study used.

Harvard’s Nutrition Source describes the pattern as primarily plant-based, built around whole grains, olive oil, fruits, vegetables, legumes, nuts, herbs and spices. Olive oil is the main added fat, fish is the preferred animal protein, eaten at least twice a week, and poultry, eggs and dairy appear in smaller amounts. Red meat is eaten only a few times a month, water is the main drink, and wine is allowed in moderation with meals. Daily physical activity is part of the picture.

The ADA’s own page on eating patterns lists Mediterranean-style eating as one of seven patterns and says it can help people looking to reduce their risk of diabetes and reduce their A1C, along with lowering triglycerides and cardiovascular risk. The page describes olive oil as the main source of dietary fat, with plant-based foods, fish and seafood, and limited red meat, wine and concentrated sugars. It does not give this pattern a carbohydrate target, and it says the right plan depends on the individual and should be chosen with your healthcare team.

What the research shows

I will take these in order of how directly they speak to people who already have type 2 diabetes. I want to be upfront about one limit: the PubMed Central pages for the most recent meta-analysis were blocked when I tried to read them, so for several studies below I am reporting what the abstract or a summary says, not what the full paper says.

Meta-analyses of trials in type 2 diabetes

A 2015 meta-analysis in the European Journal of Clinical Nutrition pooled nine studies with 1,178 patients with type 2 diabetes. Compared with control diets, a Mediterranean-style diet was associated with a lower HbA1c (mean difference of -0.30 percentage points, 95% confidence interval -0.46 to -0.14) and lower fasting plasma glucose (-0.72 mmol/L, 95% CI -1.24 to -0.21), with a small difference in body weight (-0.29 kg, 95% CI -0.55 to -0.04). The abstract I read did not say what the control diets were, so I cannot tell you what the Mediterranean diet was being compared with in each study.

A newer meta-analysis in BMC Nutrition, published in April 2024 by Zheng and colleagues, pooled seven randomized controlled trials with 1,371 patients. The search ran up to December 2021. It reported a pooled HbA1c difference of -0.39 percentage points (95% CI -0.58 to -0.20) and lower systolic blood pressure (-4.17 mmHg, 95% CI -7.12 to -1.22) and diastolic blood pressure (-1.20 mmHg, 95% CI -2.21 to -0.19). Changes in cholesterol measures were not statistically significant. I am not quoting its body mass index or fasting glucose numbers, because the abstract text I saw had a body mass index interval that did not make sense (the upper bound was lower than the point estimate) and I could not check it against the full paper.

A 2014 network meta-analysis in the Journal of Human Nutrition and Dietetics by Carter and colleagues compared several diet strategies. It included eight studies, but only three had HbA1c data (487 people), and it concluded that none of the interventions was significantly better than the others at lowering glucose measures. It did find that the Mediterranean diet reduced HbA1c compared with usual care, though not compared with a Paleolithic diet. This is worth keeping in mind: the evidence says Mediterranean-style eating tends to beat usual care, and it does not say it beats every other careful diet.

Put together, these results point in the same direction, a few tenths of a point lower A1C on average, which could matter in a population but is not a promise for any one person. Differences between people, the way a diet is delivered, and how closely people follow it all change the outcome.

Trials that looked at medication need and diabetes risk

One of the most cited trials is by Esposito and colleagues in Annals of Internal Medicine (2009). It enrolled 215 overweight people with newly diagnosed type 2 diabetes and randomly assigned them to a Mediterranean-style diet with less than half of calories from carbohydrate, or to a low-fat diet with less than 30 percent of calories from fat. After four years, 44 percent of the Mediterranean-style group had needed to start diabetes drug treatment, compared with 70 percent of the low-fat group, and the Mediterranean-style group lost about 2 kg more. I read this through a trade-news summary, not the paper itself. The summary notes that the trial was not double-blind. It also tells us something important about the intervention: the Mediterranean-style arm was lower in carbohydrate than many people picture, so this trial may be testing a lower-carb version of the pattern as much as the traditional one.

The PREDIMED trial is the largest dietary trial of this pattern, and it has an unusual history. It enrolled 7,447 people at high cardiovascular risk and assigned them to one of three groups: a Mediterranean diet supplemented with extra-virgin olive oil, a Mediterranean diet supplemented with mixed nuts, or a control diet that advised reducing dietary fat. On June 13, 2018, the New England Journal of Medicine retracted the original 2013 paper because of randomization errors, and published a corrected, reanalyzed version the same day. According to Harvard’s summary of the episode, household members had been enrolled without randomization, whole clinics rather than individual patients had been assigned to groups at one of eleven sites, and randomization tables seemed to have been used inconsistently at another. The authors adjusted for those problems and reran the analysis after excluding 1,588 participants whose assignment was known or suspected to have departed from the protocol. The result did not change meaningfully: cardiovascular events were about 30 percent lower in the Mediterranean groups. That is a heart-health result in people at high risk, not an A1C result.

A PREDIMED sub-study from the Reus center looked at diabetes itself. In 418 participants without diabetes at the start who had at least three cardiovascular risk factors, the incidence of new diabetes over a median of 4.0 years was 17.9 percent in the low-fat group, 10.1 percent in the olive oil group and 11.0 percent in the mixed nuts group, which the authors summarized as an average 52 percent relative reduction with Mediterranean diet counseling (95% CI 27 to 86 percent). That result comes from an AAFP summary of the 2011 Diabetes Care paper. I could not confirm whether the 2018 retraction touched this specific analysis, and it was a single-center subgroup, so I would treat it as supportive, not decisive.

What the studies cannot tell us

There are real limits here. Diet trials cannot be blinded, so people know what they are eating. Many are run in settings where participants received free olive oil, nuts, counseling or food, which is more support than most people get at home. The comparison diets differ. Some studies are short. And “Mediterranean diet” can mean anything from a counseling program about the pyramid to a defined lower-carb plan. Harvard also points out that the pattern does not specify portion sizes, so there is a risk of excess calorie intake, and it notes that fat made up 39 to 42 percent of calories in PREDIMED, above the usual 20 to 35 percent guideline, adding that the type of fat matters more than the percentage. It advises adopting the whole pattern, not just adding olive oil or nuts to an existing diet.

The food map: what to eat how often

The Mediterranean Diet Foundation’s UNESCO page gives a simple structure, which I have reorganized into the table below. These are the pattern’s frequencies, not diabetes-specific rules.

Source for the frequencies: the Mediterranean Diet Foundation / UNESCO page. Harvard’s version of the wine allowance is about one to two glasses a day for men and one for women. The foundation page also gives a weekly potato allowance of about three servings, preferably fresh, which is worth knowing if you have read the potatoes and blood sugar guide.

Carbohydrate in the foods that matter most

Here is the part that is easy to miss. A Mediterranean-style plate is not a low-carbohydrate plate by default. Beans, lentils, whole grains, bread and fruit all carry meaningful carbohydrate, and a generous scoop of any of them can add up. The numbers below come from USDA reference data through GetFoodFacts (FoodData Central SR Legacy) and ReciPal. Per-serving values are rounded from the label view on those pages, and any half-cup figure is my own arithmetic.

Brown rice and bulgur show the point nicely. A cup of brown rice lists 52 g of carbohydrate and 3 g of fiber, while a cup of bulgur lists 33.8 g and 8.19 g. If you replace a half cup of brown rice with a half cup of bulgur, the carbohydrate drops by about 9.1 g. That is a simple swap, not a rule. For how rice itself behaves, see the white rice guide, and for bread see the bread guide. Pasta also appears in some Mediterranean menus; the pasta guide goes through portions.

Whole-wheat bread has almost the same carbohydrate as white bread per slice (12.91 g vs 12.65 g in the reference entries). The difference is fiber, 1.9 g compared with 0.6 g. If bread is part of your pattern, the fiber is worth having, but the carbohydrate count does not fall because the bread is darker.

The fish, fat and protein side

The sardine and salmon rows are USDA entries read on ReciPal, per 100 grams because that was the only measure those pages gave. The feta row is a single branded entry (King Cheese Inc.), so check the package you buy, since sodium and fat vary between brands. Canned fish packed in oil, brined fish and olives can all add sodium, and I did not tabulate sodium for most of these foods.

Olive oil and calories: the part nobody mentions

The pattern leans on olive oil, and rightly so, since the research suggests the type of fat matters. But a tablespoon is 119 calories. Two tablespoons on a salad and roasting vegetables add 238 calories, and three tablespoons come to 357. None of that is carbohydrate, so it does not directly move a glucose reading, but it counts toward your overall energy intake. Harvard notes that the pattern does not specify portions and that losing weight on it generally depends on pairing it with calorie awareness. If weight is part of your goal with your care team, measure the oil with a spoon for a week to see what you actually use.

Nuts follow the same logic. One ounce of almonds is about 164 calories and 6.11 g of carbohydrate; a “handful” that is really two ounces is about 328 calories. Nuts are a good fit for the pattern, but they are also easy to over-eat.

A worked example: one Mediterranean lunch bowl

Here is a bowl built from the foods above: half a cup of chickpeas, half a cup of bulgur, one serving of tomato, one tablespoon of olive oil, and one ounce of feta, plus any leafy greens or cucumber you want, which I did not tabulate.

So a bowl that looks like a salad is a meal with roughly 46 g of carbohydrate, almost all of it from the chickpeas and bulgur. Now compare it with portion creep. A full cup of chickpeas plus a full cup of brown rice on the same plate lists 97 g of carbohydrate (45 plus 52), more than double the bowl above. Both plates are “Mediterranean.” The difference is the scoop. That is why this guide keeps pushing you to measure the starchy parts at least until you know what your usual portions are.

A sample week with approximate carbohydrate

The table below is a sample, not a prescription. I added up carbohydrate only from the foods tabulated in this guide (eggs, whole-wheat bread, tomatoes, lentils, chickpeas, brown rice, bulgur, almonds, oranges and feta). I did not count non-starchy vegetables beyond tomato, yogurt, olive oil (zero), fish (zero), herbs, or any wine, and I did not count sauces or soups beyond the listed lentils. The totals are therefore the minimum of what your day would have, not a full count, and they are not a target. Your own targets belong to you and your care team.

The days range from about 33.6 g to about 83 g in tabulated carbohydrate, mostly because of how many portions of beans, grains, bread and fruit appear. Day 5 has no bread or rice at all, and Day 6 has the sandwich and the rice. If you are working with a carbohydrate budget per meal, a dietitian can scale this up or down in minutes. If you keep track of your readings after meals, the post-meal glucose tracker is a place to write down what you ate and what you saw.

Swaps for common American foods

Grocery list and a budget approach

You do not need specialty products. Many of the most useful items are shelf-stable, canned or frozen, and I am not quoting any prices because I did not verify prices for this article.

Canned and dried: chickpeas, lentils, white beans, canned sardines or salmon, canned or boxed tomatoes, whole grains such as bulgur and brown rice, and plain nuts. Frozen: spinach, mixed vegetables, peas and fish fillets. Fresh: tomatoes, cucumbers, onions, greens, lemons, oranges, and whatever fish is on offer. Pantry: olive oil, vinegar, herbs, garlic and spices. If budget is the constraint, the pattern leans on beans, lentils and canned fish, which are among the lower-cost sources of protein in most stores. Store brands of canned beans and frozen vegetables are fine, and labels matter: pick canned beans without added salt or rinse them, and read sodium on canned fish and cheese. The nutrition label carb guide walks through reading those labels, and the grocery store guides look at specific stores.

Wine and alcohol

Wine shows up in the pattern as a moderate, with-meals option: Harvard gives about one to two glasses a day for men and one for women, and the Mediterranean Diet Foundation page gives a similar general reference. That is a description of the traditional pattern, not advice to begin drinking, and Harvard does not add further warnings on its page. For people with diabetes, alcohol raises its own questions, including its effect on blood sugar hours later and its interaction with some medicines, which my alcohol and blood sugar guide covers in detail. If you do not drink, you do not need to start in order to follow this pattern. If you do drink, ask your clinician what is safe for you.

How this fits different situations

If you have prediabetes

The best-known evidence for this pattern in people without diabetes is from PREDIMED, where the Mediterranean groups developed diabetes less often than the low-fat group, though the sub-study I found was a single-center subgroup in high-risk people. The ADA lists the pattern as a way to reduce the risk of diabetes. For prediabetes, the practical takeaway is that the pattern is a reasonable framework, and portions of starchy foods and calorie-dense extras still matter. See the diet for prediabetes guide for the wider context.

If you have type 2 diabetes and do not take medicine

The meta-analyses suggest a modest A1C benefit on average, and the Esposito trial reported that fewer people on the Mediterranean-style diet needed drug treatment over four years. But that trial’s diet was lower in carbohydrate, and it was newly diagnosed, overweight participants. A traditional version loaded with large portions of grains and legumes may behave differently. Ask your clinician how often to check your A1C and whether a dietitian visit makes sense while you try a new way of eating.

If you use a CGM or check your glucose

A continuous glucose monitor or a meter can help you see how your own meals behave, since two people can eat the same lentil soup and see different curves. If you use one, try one change at a time (for example, swapping the rice for bulgur) and keep the rest of the meal the same for a few days before drawing a conclusion, and keep in mind that sleep, stress, activity and medicine all affect readings. The guide to reading CGM data explains what the graph does and does not show. One meal is a data point, not a verdict.

If you take glucose-lowering medicine

This is the group for whom a major change in eating can matter most. I do not give medication advice. What I can say is that an abrupt move toward fewer refined carbohydrates, or more fasting or alcohol, can change how medicine works, and your care team is the right place to ask about timing and monitoring before you shift.

Common mistakes

Treating every Mediterranean food as free. Chickpeas, lentils, bread, couscous, rice, pasta and fruit all carry carbohydrate, and a generous bowl can reach 80 or more grams, as the portion-creep example showed.

Pouring olive oil without measuring. It is not carbohydrate, but at 119 calories a tablespoon it adds up quickly.

Adding olive oil and nuts to an otherwise unchanged diet. Harvard specifically advises adopting the whole pattern, not just adding those foods.

Assuming “whole grain” means “low carbohydrate.” Whole-wheat bread lists 12.91 g per slice, almost the same as white bread.

Counting wine as part of the plan. It is part of the traditional pattern, not part of a diabetes plan, and it is optional.

Reading PREDIMED as a diabetes treatment trial. It was a cardiovascular prevention trial in high-risk adults, with diabetes incidence as a secondary result in a sub-study.

Using canned or restaurant versions without checking sodium. Canned fish, cheeses, olives and store-bought dressings can be high in sodium; I did not tabulate it for most foods.

Myth versus fact

Questions to bring to your dietitian or care team

How many grams of carbohydrate should each meal have for me, and how should I count beans and whole grains? Would a Mediterranean-style pattern fit my medicines, kidney health and heart history? How often should I check my glucose or A1C when I change how I eat? Do I need to limit sodium from canned fish, cheese and olives? If I want to lose weight, how should I measure olive oil and nuts? If I drink, what is a safe amount for me, if any? Can you help me turn the sample week above into one that fits my readings?

Frequently asked questions

Is the Mediterranean diet good for type 2 diabetes?

The ADA lists Mediterranean-style eating among several patterns that can help with diabetes, and meta-analyses of trials in people with type 2 diabetes report modestly lower HbA1c on average (about -0.30 to -0.39 percentage points). It is a reasonable pattern to discuss with your care team, though it is not the only one.

Is the Mediterranean diet low in carbohydrate?

Not by default. It includes legumes, whole grains, bread and fruit. A cup of boiled chickpeas lists 45 g of carbohydrate and a cup of brown rice lists 52 g, so portions matter.

What did the PREDIMED retraction mean?

In June 2018 the New England Journal of Medicine retracted and republished the trial because of randomization problems at some sites. The authors reanalyzed the data, including a version excluding 1,588 participants, and reported that results did not change meaningfully, with cardiovascular events about 30 percent lower in the Mediterranean groups.

Does it lower A1C?

Meta-analyses of randomized trials report average reductions of about 0.30 to 0.39 percentage points compared with control diets. Individual results vary, and a network meta-analysis found no significant difference between several diets for glucose measures.

How much olive oil should I use?

The traditional pattern uses olive oil as the main added fat, about one tablespoon per person for cooking and dressings per the Mediterranean Diet Foundation page. A tablespoon is about 119 calories, so measuring helps if weight is a goal.

Can I eat bread, rice and beans on this diet?

Yes, they are part of the pattern, but they carry carbohydrate. A slice of whole-wheat bread lists about 12.9 g, and a half cup of cooked lentils about 20 g. Portion size is the key.

Do I have to drink wine?

No. Wine appears in the traditional pattern as an optional, moderate part, and alcohol raises separate questions for people with diabetes. Ask your clinician before drinking, and see the alcohol guide.

Is it expensive?

It does not have to be. Beans, lentils, canned fish, frozen vegetables, oats and store-brand olive oil are staples. I did not verify prices for this article, so compare in your own store.

Can I use a CGM to see how these foods affect me?

Yes. A CGM or meter can show your own response to a meal such as lentil soup versus rice. Change one thing at a time and remember that sleep, stress, activity and medicine affect readings too.

Should I change my diet if I take diabetes medicine?

Talk to your care team first. Large changes in eating can change how medicine works, and I do not give medication advice.

Sources

• American Diabetes Association: eating for diabetes management (meal patterns)
• Harvard Nutrition Source: Mediterranean diet
• Harvard Nutrition Source: PREDIMED study retraction and republication (June 22, 2018)
• PubMed: Mediterranean-style diet meta-analysis in type 2 diabetes, European Journal of Clinical Nutrition 2015
• DOAJ record: Zheng et al., BMC Nutrition 2024 meta-analysis of Mediterranean diet trials in type 2 diabetes
• AGRIS record: Carter et al., Journal of Human Nutrition and Dietetics 2014 network meta-analysis
• NutraIngredients summary of Esposito et al., Annals of Internal Medicine 2009
• AAFP summary of the PREDIMED-Reus diabetes incidence result (Diabetes Care 2011)
• Mediterranean Diet Foundation / UNESCO: food groups and frequencies
• GetFoodFacts (USDA SR Legacy): chickpeas, bulgur, brown rice, tomato, navel orange, feta (branded)
• ReciPal: lentils, olive oil, sardines, salmon

About the author: Charles Kirkland is the founder of Metabolic Clarity Labs and author of The Journey Back. He is not a physician. This is general education, not a medically reviewed treatment plan.

Sources checked October 31, 2026. Discuss personal decisions with a qualified clinician. Do not start, stop, or change prescribed treatment based on this article.