Tacos: what the tortilla and filling studies show

Nobody has studied the taco as a dish, and we have no USDA record for one, so this page has no nutrient table. What researchers have measured are its parts: the corn tortilla, the filling and the salt. Read everything below as evidence about those parts. Salt is the part that shows up in national data. In the United States, burritos and tacos together supply 4.3 percent of all the sodium people eat, the fifth largest source after pizza, breads, cured meats and soups. The corn tortilla has been tested more closely. In a trial in 10 Mexican women, about 35 percent of the zinc in a tortilla meal was absorbed, and flour fortified with zinc raised the total they took in without changing that share. The alkaline cooking that turns maize into tortilla dough, called nixtamalization, cut two fumonisins in visibly mouldy maize by 95 percent. Whether that means your body takes in less is still open. In the same 40-person study, toxin markers in urine came out no different from those of a group eating control maize, so how much less the body actually takes in is still unclear.

How much is actually in it

• In NHANES 2017-2018 dietary recalls from 7,081 Americans, burritos and tacos supplied 4.3% of all dietary sodium, the fifth largest food source after pizza, breads, cured meats and soups. — US population, NHANES 2017-2018, 24 h recalls (n = 7081) Evidence C source
  Burritos and tacos 4.3% of population sodium intake
• At a US taco chain, restaurants with calorie labels on menus sold 24.7 fewer calories per transaction than matched unlabelled restaurants over 3 to 24 months, with the largest drop in the number of tacos bought. The labelled restaurants were mostly in California, and outside California no association was found. — Transactions from 2,329 Taco Bell restaurants, 474 with menu labelling, 2007-2014 (n = 2329) Evidence C source
  -24.7 kcal per transaction (95% CI 23.6 to 25.7)

What a real portion looks like

• In a survey of 915 Mexican adults aged 40 and over, tortillas made up 77% of all maize eaten, with a median of 252 g of maize tortillas a day. — Representative sample of 915 adults aged 40+ in south and eastern Mexico, 2018-2019 (n = 915) Evidence C source
  Median maize tortilla intake 252 g/d, 77% of total maize
• In a blind randomised tasting with 147 consumers, taco fillings with part of the beef swapped for mushrooms were liked as much as 100% beef filling, while cutting salt by 25% lowered overall liking. — 147 consumers tasting six taco blend recipes blind (n = 147) Evidence B source
  No difference in liking among full-salt blends; 25% less sodium liked less overall

What your body absorbs

• In a randomised crossover trial in 10 Mexican women, about 35% of the zinc in a corn tortilla meal was absorbed, whether the tortilla was unfortified or fortified with 20 mg/kg zinc oxide or zinc sulfate. — Healthy Mexican women aged 21-51, 10 per phase, single stable-isotope test meals (n = 10) Evidence B source
  Fractional zinc absorption 0.35, 0.36 and 0.37 (unfortified, zinc oxide, zinc sulfate), no difference
• In the same 10-woman trial, tortillas made from flour fortified with 20 mg/kg zinc delivered more absorbed zinc than unfortified tortillas, because the absorbed fraction stayed the same while zinc content rose. — Healthy Mexican women aged 21-51, staff and students of INCMN, no zinc supplements for 6 months (n = 10) Evidence B source
  Total absorbed zinc significantly higher with either fortified tortilla than with plain tortilla
• Adding 30 or 60 mg/kg of reduced iron to zinc-fortified corn tortilla flour did not change zinc absorption in 10 women (fractional absorption 0.32 to 0.33 in all three arms). — Healthy Mexican women aged 21-51, phase 2 of the trial (n = 10) Evidence B source
  Fractional zinc absorption 0.32, 0.33 and 0.32 with 0, 30 and 60 mg/kg added iron
• In 5 healthy women eating about 140 g of tortillas, calcium absorption was 50% from low-phytate maize tortillas versus 35% from tortillas made of ordinary maize. — 5 healthy adult women, two test meals one month apart (n = 5) Evidence B source
  Fractional calcium absorption 0.50 vs 0.35, P = 0.003

What cooking and storage change

• Nixtamalization, the alkaline cooking step used to make tortilla dough, cut fumonisin B1 and B2 in visibly mouldy maize by 95% and zearalenone by 89% in a 40-person intervention study. — 40 healthy adults in Kliptown, Soweto, South Africa (n = 40) Evidence B source
  FB1 -95%, FB2 -95%, ZEN -89% in maize; FB3 and DON to undetectable

What it does to your health

• In adults with overweight or obesity, tortillas with 30% of the refined flour replaced by soy flour lowered blood glucose 90 to 120 minutes after eating (p = 0.0009). — 10 adults with overweight or obesity who tested the tortilla, out of 17 in the study (n = 10) Evidence B source
  Lower plasma glucose at 90-120 min for tortillas (p = 0.0009), size not reported in the abstract
• In 35 adults with a mean age of 72, a breakfast tortilla filled with 135 g of black or red kidney beans did not differ significantly in fullness from one with 80 g of beef, with no difference in lunch or 24-hour energy intake. — 35 older adults, mean age 72.4, mean BMI 25.1 (n = 35) Evidence B source
  No significant difference in appetite AUCs, ad libitum lunch or 24 h energy intake between fillings

Safety, limits and interactions

• Among 915 adults aged 40 and over in south and east Mexico, each maize tortilla eaten per day was linked to 2.0% higher blood aflatoxin adducts, and 2.8% for tortillas made from homemade masa. — Representative sample of 915 adults aged 40+ in south and eastern Mexico, 2018-2019 (n = 915) Evidence C source
  +2.0% AFB1-lysine per tortilla (95% CI 0.6 to 3.4); +2.8% for homemade masa (95% CI 1.1 to 4.6)
• In a 2023 Finnish school outbreak with 721 people ill, flour tortillas were linked to illness (adjusted odds ratio 3.3), and samples held over 24,000 mg/kg of the preservative calcium propionate against a legal maximum of 2,000 mg/kg. — Pupils and staff of 17 schools in a Finnish municipality, August 2023 (n = 721) Evidence C source
  AOR 3.3 (95% CI 1.4 to 7.4) for tortillas; propionate above 24,000 mg/kg vs 2,000 mg/kg limit

Who this works differently for

• In NHANES 2011-2016 data on 3,112 US children aged 6 to 11, burritos and tacos were the main high glycaemic index food for Mexican American children, while pizza led for most other groups. — 3,112 US children aged 6-11, NHANES 2011-2016 (n = 3112) Evidence C source
  Top high-GI food type differed by group; burritos/tacos for Mexican American children

What is still unknown

• Eating nixtamalized mouldy maize left urinary mycotoxin markers no different from eating control maize in 40 adults, and deoxynivalenol stayed detectable in over 90% of urine samples, so how much nixtamalization lowers what the body absorbs is still unclear. — 40 healthy adults in Kliptown, Soweto, South Africa (n = 40) Evidence B source
  No significant difference in urinary FB1, ZEN, DON between groups (p > 0.05); detection above 90%

The bottom line

If you want to use less beef in your tacos, the studies here speak to taste and fullness, not to health. In a blind tasting with 147 people, fillings with part of the beef swapped for mushrooms were liked as much as all beef. Cutting the salt by a quarter was noticed, and people liked those versions less. In 35 adults with an average age of 72, swapping 80 g of beef in a breakfast tortilla for 135 g of black or red kidney beans made no significant difference to fullness or to what they ate later that day. The study did not test whether the two are equal. A tortilla with 30 percent of its refined flour replaced by soy flour lowered blood sugar 90 to 120 minutes after eating, and the abstract does not say by how much. That result comes from 10 adults with overweight or obesity, which is a promising start and too small to build a habit on. The tortilla itself carries two safety findings, both observational and both about how it is made. Among 915 Mexican adults aged 40 and over, who ate a median of 252 g of maize tortillas a day, each tortilla a day went with 2.0 percent more aflatoxin in the blood, measured as AFB1-lysine adducts. Aflatoxin is a mould toxin. For tortillas made from homemade masa the figure was 2.8 percent. That is a link in people eating tortillas at every meal, and it does not show that an occasional taco harms anyone. In Finland in 2023, 721 pupils and staff fell ill after school lunches, and the flour tortillas held more than 24,000 mg/kg of the preservative calcium propionate against a legal maximum of 2,000. That was a production failure. In US data on children aged 6 to 11, burritos and tacos were the main high glycaemic food for Mexican American children, while pizza led for most other groups. Some things we looked for and did not find. There is no study on tacos in pregnancy, in kidney disease or alongside medicines, so we cannot tell you who should be careful beyond the findings above. We also found no popular taco belief that anyone has tested, and no research on how cooking changes the finished dish.

Sources

1. tco-b5-10 · cross-sectional
   More than 50% of US population-level dietary sodium intake was contributed by: pizza (5.3%); breads, rolls and buns (4.7%); cold cuts and cured meats (4.6%); soups (4.4%); burritos and tacos (4.3%)
   Nutrients, 2023 · checked 2026-09-23
2. tco-b5-15 · cohort
   customers purchased 24.7 (95% CI, 23.6-25.7) fewer calories per transaction from restaurants in the menu labeling group in the 3- to 24-month follow-up period vs the comparison group ... with a larger decrease in the number of tacos vs other items purchased
   JAMA Netw Open, 2023 · checked 2026-09-23
3. tco-b5-08 · cross-sectional
   Maize tortillas represented 77% of total maize consumption, with a median consumption of 252 g/d
   Am J Clin Nutr, 2025 · checked 2026-09-23
4. tco-b5-11 · randomised controlled trial
   The taco blend recipes with full sodium were also liked more overall than those with 25% less sodium. But there was no significant difference in overall liking among the three full-salt recipes, nor among the three reduced-salt recipes, indicating that across the consumer population we tested, acceptance of the mushroom-containing recipes was on par with that of the 100% beef recipe.
   Appetite, 2016 · checked 2026-09-23
5. tco-b5-01 · randomised crossover trial
   The mean (+/- SEM) fractional zinc absorption from unfortified tortilla, tortilla fortified with zinc oxide, and tortilla fortified with zinc sulfate did not differ among treatments: 0.35 +/- 0.07, 0.36 +/- 0.05, and 0.37 +/- 0.07, respectively.
   Food Nutr Bull, 2012 · checked 2026-09-23
6. tco-b5-02 · randomised crossover trial
   The subjects absorbed significantly less zinc with CT than with fortified tortillas, and absorption from both fortified tortillas (CT + ZnO and CT + ZnSO4) was similar.
   Food Nutr Bull, 2012 · checked 2026-09-23
7. tco-b5-03 · randomised crossover trial
   The three treatment groups with 0, 30, and 60 mg/kg of added iron had similar fractional zinc absorption (0.32 +/- 0.04, 0.33 +/- 0.02, and 0.32 +/- 0.05, respectively)
   Food Nutr Bull, 2012 · checked 2026-09-23
8. tco-b5-04 · clinical trial (crossover)
   Mean fractional absorption of calcium from tortillas prepared from the low-phytate maize (0.50 +/- 0.03) was significantly (P = 0.003) greater than that from tortillas prepared from the control maize (0.35 +/- 0.07).
   Am J Clin Nutr, 2005 · checked 2026-09-23
9. tco-b5-05 · randomised controlled trial
   Nixtamalization achieved a reduction in fumonisin B3 and deoxynivalenol (DON) to unquantifiable or undetectable levels in maize, while reducing fumonisin B1 (FB1), fumonisin B2, and zearalenone (ZEN) by 95%, 95%, and 89%, respectively.
   Toxins (Basel), 2025 · checked 2026-09-23
10. tco-b5-12 · randomised controlled trial
    substituting 30% of refined flour with soy flour significantly reduced the area under the concentration-time curve for postprandial glucose for breads (p = 0.03) and arepas (p = 0.04), and reduced plasma glucose concentrations at 90-120 min for tortillas (p = 0.0009)
    Nutrients, 2026 · checked 2026-09-23
11. tco-b5-13 · randomised controlled trial
    Appetite sensation AUCs, ad libitum pizza intake, and 24-h energy intake did not significantly differ between the test meals.
    J Nutr, 2025 · checked 2026-09-23
12. tco-b5-07 · cross-sectional
    For every maize tortilla consumed, the concentration of circulating AFB1-lys was 2.0% (95% CI: 0.6%, 3.4%) higher. The corresponding estimate for homemade masa tortilla was 2.8% (95% CI: 1.1%, 4.6%).
    Am J Clin Nutr, 2025 · checked 2026-09-23
13. tco-b5-09 · retrospective cohort
    consumption of flour tortillas and vegetable filling served at lunch were statistically associated with illness (adjusted odds ratio (aOR) = 3.3; 95% confidence interval (CI): 1.4-7.4 for the tortillas ... high concentrations of calcium propionate (E 282) were measured (> 24,000 mg/kg), exceeding the regulatory maximum limit of 2,000 mg/kg.
    Euro Surveill, 2026 · checked 2026-09-23
14. tco-b5-14 · cross-sectional
    Types of high GI foods differed meaningfully across racial/ethnic groups (ie, Mexican American: burritos/tacos; other Hispanic, White, and Black: pizza; Asian American: rice).
    Prev Chronic Dis, 2020 · checked 2026-09-23
15. tco-b5-06 · randomised controlled trial
    urinary levels of FB1, ZEN, and its metabolites α- and β-zearalenol (α- and β-ZEL) did not show significant differences between groups (p > 0.05). DON and tenuazonic acid levels were not affected by the intervention (p > 0.05), with urinary detection frequencies remaining above 90%.
    Toxins (Basel), 2025 · checked 2026-09-23