Dosa: a high glycaemic index breakfast, and the grain that changes it

The finding this page leans on is about which grain goes into the batter. A 2026 systematic review of six studies, searched to 14 January 2025, found that swapping rice for foxtail millet in dosa lowered the rise in blood glucose after the meal in adults with type 2 diabetes. That review pooled one randomised controlled study, three randomised crossovers and two non-randomised studies, and it reports the direction without a pooled effect size, so this page cannot tell you how much lower. The rice version has been measured directly. Rice dosa scored a glycaemic index of 76 in 19 healthy men aged about 24.7 years with a BMI of 21.7, nine Chinese and ten Indian, in a randomised crossover with 180 minutes of blood sampling. That is the high band, with rice idli at 85 and upma at 71 in the same study, while all five Western breakfast foods tested there came out low. There is no USDA composition record for dosa at all, and the one reference figure for a filled dosa is British: 97 kcal and 6.3 g of fat per 100 g in the 2021 CoFID entry for a homemade vegetable filling.

How much is actually in it

• Black gram sits among the decorticated pulses that supply most of the 18.22 mg of isoflavones an urban Indian adult eats per day. — sample urban Indian population assessed for legume consumption, with genistein and daidzein measured in the legumes they eat (n = not reported in abstract) Evidence C source
  Total genistein 0.60 to 8.65 mg per 100 g and total daidzein 1.10 to 30.37 mg per 100 g across the legumes analysed; average isoflavone intake 18.22 mg/day
• The only composition figure for a filled dosa in either reference table is British, at 97 kcal and 6.3 g of fat per 100 g, because USDA holds no composition record for dosa at all. — CoFID 2021 entry Dosa, filling, vegetable, homemade, code 15-513 Evidence E source
  Per 100 g: 97 kcal, 404 kJ, fat 6.3 g, protein 1.8 g, carbohydrate 10.2 g, starch 5.2 g, total sugars 4.1 g, NSP fibre 1.5 g, water 77.4 g, calcium 24 mg

What a real portion looks like

• A plain dosa weighs 35 g small, 80 g medium and 155 g large, while a filled one runs from 113 g to 523 g. — USDA FNDDS household measures for Dosa, plain (fdc_id 2708347) and Dosa, with filling (fdc_id 2709129) Evidence E source
  Plain 1 small 35 g, 1 medium 80 g, 1 large 155 g, unspecified quantity 80 g; with filling 1 small 113 g, 1 medium 267 g, 1 large 523 g

What cooking and storage change

• How fast the starch breaks down in a test tube predicted the blood glucose peak of these dishes closely, with correlations of 0.83 and 0.85. — six cereal-based South Indian foods tested for in vitro starch digestibility alongside the human arm of the same trial (n = 6 foods) Evidence D source
  Peak response versus in vitro rate of starch hydrolysis r 0.83, area under the curve r 0.85, P < 0.05
• In 30 people with type 2 diabetes, the cooking method changed the one-hour glucose response of semolina meals paired with pulse dhals, with steamed versions beating roasted ones. — 30 patients with non-insulin-dependent diabetes mellitus, five conventional Indian meals of 50 g carbohydrate, semolina cooked two ways or combined with black gram, green gram or Bengal gram dhal (n = 30) Evidence B source
  Only steam-cooked semolina and semolina with Bengal gram dhal gave a lower response than glucose at 1 hour, and only semolina with black gram dhal at 2 hours; no differences in triacylglycerol response
• A 2026 systematic review of six trials found that swapping rice for foxtail millet in dosa lowered the blood glucose rise after the meal in people with type 2 diabetes. — adults, one randomised controlled study, three randomised crossover studies and two non-randomised studies, searched to 14 January 2025 (n = 6 studies) Evidence B source
  Foxtail millet dosas reduced postprandial blood glucose; multigrain rotis reduced HbA1c and LDL cholesterol in the same review

What it does to your health

• Measured in healthy young men, rice dosa has a glycaemic index of 76, which puts it in the high band alongside rice idli at 85. — 19 healthy men, nine Chinese and ten Indian, aged 24.7 years with BMI 21.7, randomised crossover, 180 minutes of blood sampling (n = 19) Evidence B source
  Rice dosa GI 76 (sem 5); rice idli 85 (sem 4); upma 71 (sem 6); all five Western breakfast foods tested were low GI

Safety, limits and interactions

• Taking the antibiotic cefuroxime axetil after a fermented rice breakfast recovered 16.59% of the dose in urine against 28.44% after a fried wheat breakfast. — 8 healthy male volunteers, randomised crossover with one week washout, 500 mg cefuroxime axetil taken 10 minutes after either idly with chutney or poori with dal-fry (n = 8) Evidence B source
  8 hour urinary recovery 16.59% with idly and chutney versus 28.44% with poori and dal-fry, P < 0.005; AUC and Cmax also higher with the second breakfast (P < 0.01 and P < 0.02)
• Breaking black gram protein down with enzymes left only 2 of 6 sensitised people still reacting on a skin test, which shows the allergen survives ordinary handling but not hydrolysis. — people sensitised to kidney bean, black gram or peanut, tested by skin prick against raw extracts and against hydrolysates (n = 6 black gram sensitive individuals) Evidence C source
  2 of 6 black gram sensitive individuals still positive to the hydrolysate, against 1 of 10 for kidney bean and 1 of 7 for peanut

Who this works differently for

• Six traditional South Indian cereal foods, reported as a set rather than one by one, scored a glycaemic index of 67 to 90 in people with type 2 diabetes and only 44 to 69 in healthy volunteers. — adults with non-insulin-dependent diabetes mellitus and healthy volunteers, six traditional South Indian cereal foods including idli, each providing 50 g of carbohydrate (n = not reported in abstract) Evidence B source
  GI 67 to 90 in diabetes versus 44 to 69 in healthy subjects; responses at 30, 60 and 120 minutes were lower than glucose in both groups (P < 0.05)
• Among 216 asthma patients screened in India, rice and black gram, the two ingredients of dosa batter, were the two most common positive skin prick reactions. — 216 consecutive patients with allergic bronchial asthma, mean age 31.9 years, screened by skin prick test and serum IgE with blinded food challenges (n = 216) Evidence C source
  Rice positive in 24 patients (11%) and black gram in 22 (10%); blinded food challenges with rice or black gram confirmed sensitisation in 28 to 37% of cases

What is still unknown

• A 2025 systematic review put black gram among the protein sources carrying antinutritional factors and allergens, and said the effect of processing on them stays unquantified. — literature on faba bean, mung bean, lentil, black gram, cowpea, quinoa, hemp, leaf proteins, microalgae and duckweeds Evidence D source
  Antinutritional factors, phytoestrogens and oligosaccharides present in all sources reviewed; no conclusive data to quantitatively assess safety after processing
• USDA retention factors cannot say what cooking does to a dosa, because the reference table knows neither the dish nor a food group to place it in. — USDA Table of Nutrient Retention Factors Release 6, searched for dosa Evidence E source
  No method and no food group matched; the table covers 13 groups and none of them holds this dish

The bottom line

Start with the plate. US survey records put a plain dosa at 35 g small, 80 g medium and 155 g large, and a filled one at 113 g, 267 g and 523 g for the same three sizes. Cooking changes the curve. In 30 patients with non-insulin-dependent diabetes eating five conventional Indian meals of 50 g of carbohydrate each, only steam-cooked semolina and semolina with Bengal gram dhal gave a lower glucose response than glucose itself at one hour. At two hours only semolina with black gram dhal did. Those were semolina dishes rather than rice dosa, and the same trial found no differences in the triacylglycerol response. A laboratory test on six cereal-based South Indian foods found that the rate of starch breakdown in a test tube tracked the human response closely, with r 0.83 for the peak and r 0.85 for the area under the curve (P < 0.05). Who is eating matters as much as what. A set of six traditional South Indian cereal foods, reported together rather than one by one, scored a glycaemic index of 67 to 90 in adults with non-insulin-dependent diabetes and 44 to 69 in healthy volunteers. That abstract does not report how many people were tested, and its healthy range does not agree with the 76 measured for rice dosa in the other trial, so a single glycaemic index for dosa is a range rather than a value. Read the 76 above as a number measured in healthy young men. Two cautions here are real, one about medicine and one about allergy. In 8 healthy male volunteers, 500 mg of cefuroxime axetil taken 10 minutes after idly with chutney left 16.59 percent of the dose in urine over 8 hours, against 28.44 percent after poori with dal-fry (P < 0.005), with AUC and Cmax also lower. The breakfast in that trial was idly with chutney, described as a fermented rice breakfast, and dosa itself was never given. If you take cefuroxime axetil, when you take it against a fermented rice breakfast is a question for whoever prescribes it. Among 216 consecutive patients with allergic bronchial asthma in India, mean age 31.9 years, rice was skin-prick positive in 24 (11 percent) and black gram in 22 (10 percent). Those were the two most common positives in that screen, and blinded food challenges confirmed sensitisation in 28 to 37 percent of those cases. In a separate skin test study, breaking black gram protein down with enzymes left 2 of 6 sensitised people still reacting, which the authors read as the allergen surviving ordinary handling. What processing does to the rest is unquantified: a 2025 review placed black gram among protein sources carrying antinutritional factors, phytoestrogens and oligosaccharides, and stated there is no conclusive data to assess their safety after processing. The USDA retention table cannot help either, because its 13 food groups hold neither this dish nor a group to put it in.

Sources

1. dosa-r5-09 · nutrition survey
   Total genistein content of all legumes considered for analysis was found to be in the range of 0.60 to 8.65 mg per 100 g whereas total daidzein content was found to be in the range of 1.10 to 30.37 mg per 100 g in selected legumes...Average isoflavone consumption by selected population was found to be 18.22 mg/d through the consumption of legumes.
   Nutr Cancer, 2019 · checked 2026-09-24
2. dosa-r5-12 · dataset
   Dosa, filling, vegetable, homemade [CoFID 15-513]: Energy 97 kcal / 404 kJ; Fat 6.3 g; Protein 1.8 g; Carbohydrate 10.2 g; Starch 5.2 g; Total sugars 4.1 g; NSP 1.5 g; Water 77.4 g; Calcium 24 mg per 100 g.
   CoFID 2021, Public Health England, OGL v3.0, code 15-513 · checked 2026-09-24
3. dosa-r5-11 · dataset
   Dosa, plain [fdc_id 2708347], FNDDS code 55702100: 1 small 35 g; 1 medium 80 g; 1 large 155 g; 1 surface inch 1 g; Quantity not specified 80 g. Dosa, with filling [fdc_id 2709129], FNDDS code 58174100: 1 small 113 g; 1 medium 267 g; 1 large 523 g; 1 surface inch 3 g; Quantity not specified 267 g.
   USDA FNDDS, Food and Nutrient Database for Dietary Studies, fdc_id 2708347 and 2709129 · checked 2026-09-24
4. dosa-r5-03 · in vitro analysis
   Significant relationships were observed between peak responses and area under the curve for foods in patients with NIDDM and in vitro rate of starch hydrolysis (r 0.83, r 0.85, P < 0.05).
   Br J Nutr, 2000 · checked 2026-09-24
5. dosa-r5-04 · randomised clinical trial
   Compared with the blood glucose response for a 50 g glucose load, only meals based on steam-cooked semolina and semolina-Bengal gram dhal elicited a significantly lower response at 1 h postprandially, and only meals based on semolina-black gram dhal at 2 h postprandially.
   Br J Nutr, 1992 · checked 2026-09-24
6. dosa-r5-10 · systematic review
   In participants with type 2 diabetes (T2DM), multigrain rotis prepared from finger millet, foxtail millet, sorghum, jowar, oats and maize reduced glycated hemoglobin (HbA1c) and low density lipoprotein cholesterol (LDL-c) levels, foxtail millet dosas reduced postprandial blood glucose levels and food prepared from heat-treated barnyard grains reduced fasting blood glucose (FBG) and LDL-c levels.
   Diabetes Metab Syndr, 2026 · checked 2026-09-24
7. dosa-r5-01 · randomised crossover trial
   Asian breakfast foods were of medium (white bun filled with red bean paste=58 (sem 4); Chinese steamed white bun=58 (sem 3)) to high GI (rice idli=85 (sem 4); rice dosa=76 (sem 5); upma=71 (sem 6)), whereas Western breakfast foods were all of low GI
   Br J Nutr, 2018 · checked 2026-09-24
8. dosa-r5-05 · randomised crossover trial
   Diet-A included idly with chutney. Diet-B included poori and dal-fry...The 8 h urinary recovery of cefuroxime was 16.59 and 28.44 per cent (P < 0.005) with Diet-A and Diet-B respectively.
   Indian J Med Res, 2000 · checked 2026-09-24
9. dosa-r5-07 · case-control study
   Significant reduction in the biopotency of hydrolysates was also observed in SPT where only 1/10 kidney bean-sensitive individuals, 2/6 black gram-sensitive individuals and 1/7 peanut-sensitive individuals were found positive to their respective hydrolysates.
   Food Chem Toxicol, 2015 · checked 2026-09-24
10. dosa-r5-02 · randomised clinical trial
    The glycaemic index (GI) values ranged from 67 to 90 in NIDDM and from 44 to 69 in healthy subjects with no significant differences within the groups.
    Br J Nutr, 2000 · checked 2026-09-24
11. dosa-r5-06 · clinical diagnostic study
    Rice elicited marked positive skin prick test reactions (SPT) in 24 (11%) asthma patients followed by black gram 22 (10%), lentil 21 (9.7%) and citrus fruits 20 (9.2%).
    Asian Pac J Allergy Immunol, 2006 · checked 2026-09-24
12. dosa-r5-08 · systematic review of hazard data
    Further, the effects of food processing on these proteins remain poorly understood and no conclusive data are available to quantitatively assess their safety after processing.
    Compr Rev Food Sci Food Saf, 2025 · checked 2026-09-24
13. dosa-r5-13 · dataset
    USDA Table of Nutrient Retention Factors, Release 6 (2007), query dosa: no retention method matched and no FDC group is assigned to this food. Groups present in the table: Dairy and Egg Products; Poultry Products; Breakfast Cereals; Fruits and Fruit Juices; Pork Products; Vegetables and Vegetable Products; Nut and Seed Products; Beef Products; Beverages; Finfish and Shellfish Products; Legumes and Legume Products; Lamb, Veal, and Game Products; Cereal Grains and Pasta.
    USDA Table of Nutrient Retention Factors, Release 6, 2007 · checked 2026-09-24