Plantains: the strongest evidence is about sick toddlers

The strongest evidence about plantain was collected in hospitals, in children. A meta-analysis of nine randomised trials in 3996 children aged 8 to 34 months with diarrhoea found that cooked green Musa paradisiaca added to standard treatment nearly doubled the rate of recovery (HR 1.96, 95% CI 1.62 to 2.37), with heterogeneity of 52 percent. Seven of those nine trials were open label, so the people scoring recovery knew who had eaten the banana. A randomised trial in 80 hospitalised children aged 1 to 28 months, whose diarrhoea had already lasted two weeks or more, shortened the illness by 18 hours against a yogurt diet of the same calories (p < 0.005), at a single centre in 2009. In every one of these trials the cooked green plantain was fed in hospital or under supervision and alongside standard treatment, not in place of it. The plantain in these trials was cooked green, the starchy state, so none of it speaks for ripe fried plantain, and none of it was tested in adults. For what is in the fruit, USDA measured six samples of overripe plantain and found vitamin C running from 8.07 to 28.4 mg per 100 g, a spread of more than three times from one fruit to the next. The three ripeness stages averaged within 1.5 mg of each other, so the variation sits between fruits rather than between stages.

The numbers, per 100 g

How much is actually in it

• USDA measured six samples of overripe plantain and found vitamin C ranging from 8.07 to 28.4 mg per 100 g, a spread of more than three times between one fruit and another. — USDA Foundation Foods samples, six per ripeness stage (n = 6 samples per stage) Evidence E source
  Overripe 19.08 mg per 100 g (range 8.07 to 28.4, median 18.9), ripe 20.08 mg (15.6 to 27.1), underripe 18.53 mg (15.3 to 23.4)

What a real portion looks like

• One whole raw plantain weighs 180 g in US dietary surveys, and the survey tables carry several separate plantain entries because frying, boiling and chip-making are counted as different foods. — US survey food records, FNDDS entries for plantain Evidence E source
  1 whole raw plantain = 180 g; separate records for raw, cooked with no added fat, cooked with oil, cooked with butter, chips, green with cracklings, ripe meat pie

What your body absorbs

• Measured against white bread, four Musa varieties gave glycaemic indices of 61 to 69, which the authors class as low, and a single fruit gave a low glycaemic load. — healthy adults tested with four banana varieties (Silk AAB, Mysore AAB, Gros Michel AAA, Pisang Awak ABB), white bread as the reference food Evidence C source
  GI 61 +/- 5, 61 +/- 6, 67 +/- 7 and 69 +/- 9 on the bread scale; starch 14% and sucrose up to 38% of available carbohydrate depending on variety; total dietary fibre 2.7% to 5.3%
• A drink made from cooked green banana pulp lowered plasma glucose by 15% over two hours in ten healthy women, while hunger, satiety and the gut hormones PYY and GIP did not move. — ten healthy women, glucose solution at 250 g/L alone, then one week later the same solution plus 75 g/L of cooked green banana pulp beverage, blood sampled every 15 minutes for 2 hours (n = 10) Evidence C source
  Plasma glucose -15% (p = 0.0232) versus the glucose solution alone; no change in insulin, PYY, GIP or hunger and satiety scores

What cooking and storage change

• By the USDA retention table for fruits, drying leaves 20% of the vitamin C in place while sauteing or stewing leaves 70%, so how a plantain is cooked decides most of what survives. — USDA Table of Nutrient Retention Factors Release 6, group Fruits and Fruit Juices, code 9 Evidence E source
  Vitamin C retention 20% dried, 50% canned, 70% sauteed or stewed; vitamin B6 95% baked when dried, 100% reheated

What it does to your health

• In a meta-analysis of nine randomised trials in 3996 children with diarrhoea, cooked green Musa paradisiaca added to standard treatment nearly doubled the rate of recovery (HR 1.96, 95% CI 1.62 to 2.37). — 3996 children aged 8 to 34 months with acute or persistent diarrhoea, nine randomised controlled trials (seven open label, two blinded), searched to January 2024 (n = 3996) Evidence A source
  HR 1.96 (95% CI 1.62 to 2.37) for resolution of diarrhoea, I2 = 52%; persistent diarrhoea HR 2.34 (1.78 to 3.08), acute diarrhoea HR 1.74 (1.45 to 2.09)
• In 80 hospitalised children with diarrhoea lasting two weeks or more, a cooked green plantain diet shortened the illness by 18 hours compared with a yogurt diet of the same calories. — 80 children of both sexes aged 1 to 28 months with 14 days or more of persistent diarrhoea, randomised to a cooked green plantain diet at 50 g/L or a yogurt diet, both isocaloric at 100 kcal/kg/d, for one week in hospital (n = 80) Evidence B source
  Diarrhoea 18 hours shorter (p < 0.005), less stool output and firmer stools (p < 0.002), greater daily weight gain (p < 0.001), lower cost (p < 0.005)
• In an open randomised trial in children with acute watery diarrhoea, 62.4% of those given cooked green banana recovered within 72 hours against 47.2% on standard care alone (RR 1.3, 95% CI 1.05 to 1.7, number needed to treat 7). — children aged 9 months to 5 years presenting within 48 hours of acute watery diarrhoea with no dehydration, cooked green banana supervised in hospital for 72 hours then continued at home Evidence B source
  Recovery at 72 hours 62.4% vs 47.2%, RR 1.3 (95% CI 1.05 to 1.7), NNT 7, fewer cases of dehydration and persistent diarrhoea

Safety, limits and interactions

• Among 200 people with pollen allergy, 30% reacted on skin testing to profilin, the pollen protein that cross-reacts with banana, and most of those who reacted to foods tolerated them once the food was cooked. — 200 consecutive patients with pollen allergy skin prick tested with purified date palm profilin, 60 profilin reactors of whom 34 had food allergy (n = 200) Evidence C source
  60 of 200 (30%) sensitised to profilin; 34 of 60 (57%) of them had food allergy, most often oral allergy syndrome only
• Among 1074 pregnant women in Cameroon, those whose most frequent meal was built on banana or plantain and tubers had 1.37 times the odds of total hypocalcaemia (adjusted OR 1.37, p = 0.012). — 1074 apparently healthy women in late pregnancy attending four maternities in the Nkongsamba Health District, Cameroon, November 2020 to September 2021 (n = 1074) Evidence C source
  Adjusted odds ratio 1.37 (p = 0.012) for total crude hypocalcaemia; total hypocalcaemia affected 61.6% of the sample while ionised hypocalcaemia affected 2.89%

What people believe that the data does not support

• A 2022 review cited for the idea that unripe plantain controls type 2 diabetes found heterogeneity of 98% across its 18 studies and reported that the combined effect sizes were not usable. — 18 studies of unripe plantain, bitter yam, okra and chitosan, seven of them on unripe plantain, interventions lasting 3 to 84 days (n = 18 studies) Evidence D source
  I2 = 98%, prediction intervals crossing zero (mean difference 4.4 mg/dL, 95% PI -6.65 to 15.50 and 3.4 mg/dL, 95% PI -23.65 to 30.50)

What is still unknown

• The same review states in its methods that no human studies were included, which is visible only in the full text and not in the abstract that gets quoted. — study selection criteria of the review, full text section 2.2 Evidence D source
  Zero human studies among the 18 included

The bottom line

Plantain is a staple starch, and the blood sugar numbers people quote for it come from the dessert side of the genus. Four Musa varieties tested at glycaemic indices of 61 to 69 against white bread, a reference that does not share a scale with the glucose-referenced tables most charts use, and those were dessert cultivars rather than cooking plantain. A drink made from cooked green banana pulp lowered plasma glucose by 15 percent over two hours in ten healthy women (p = 0.0232), one dose in one session with the cultivar richest in resistant starch. In the same ten women, insulin, PYY, GIP, hunger and satiety did not move. The 2022 review cited for the idea that unripe plantain controls type 2 diabetes reported heterogeneity of 98 percent across its 18 studies and said the combined effect sizes were not usable. Its methods section adds that no human studies were included at all, a sentence that is in the full text and not in the abstract people quote. Cooking decides much of the rest. The USDA retention table for fruits leaves 70 percent of vitamin C in place after sauteing or stewing and 20 percent after drying, a table built on food groups rather than on plantain itself. Release 6 has no carbohydrate row, so what boiling does to plantain starch is outside every table we hold. One whole raw plantain counts as 180 g in US dietary surveys, a survey convention rather than a measured fruit, and shop plantains run well under and well over that. Who should take care. Among 200 people with pollen allergy, 30 percent reacted on skin testing to profilin, the pollen protein that cross-reacts with banana, and most of the 34 of them who had food allergy tolerated the offending fruit once it was cooked. Banana stands in for plantain there, because no study in our sources tested plantain itself, and plantain is almost always eaten cooked. Among 1074 pregnant women in Cameroon, a most frequent meal built on banana or plantain and tubers went with 1.37 times the odds of total hypocalcaemia (adjusted OR 1.37, p = 0.012), odds rather than risk and measured at one moment in time. In that same sample, ionised calcium, the fraction that matters physiologically, was low in under 3 percent of the women. We found no trial and no cohort on plantain in kidney disease, and none on plantain with any medicine.

Sources

1. pltn-r1-10 · food composition dataset
   Plantains, overripe, raw: Vitamin C, total ascorbic acid 19.08 MG, range 8.07-28.4 (median 18.9), 6 samples; Plantains, ripe, raw 20.08 MG, range 15.6-27.1; Plantains, underripe, raw 18.53 MG, range 15.3-23.4
   USDA FoodData Central, Foundation Foods, release 2026-04-30 · checked 2026-09-24
2. pltn-r1-12 · food portion dataset
   Plantain, raw [fdc_id 2709560], Other starchy vegetables: 1 whole = 180 g
   USDA FNDDS 2021-2023, FDC ID 2709560 · checked 2026-09-24
3. pltn-r1-04 · clinical trial
   Glycemic indices of Silk, Mysore, Gros Michel and Pisang Awak were 61 ± 5, 61 ± 6, 67 ± 7, 69 ± 9 and can be categorized as low against white bread as the standard.
   Int J Food Sci Nutr, 2011 · checked 2026-09-24
4. pltn-r1-05 · controlled acute feeding study
   After 2 h of GBPd intake, no difference was observed in hunger/satiety sensation and plasma glycemic homeostasis indexes, except for a decrease in plasma glucose concentration (-15%, p = 0.0232) compared to stage-1.
   Plant Foods Hum Nutr, 2024 · checked 2026-09-24
5. pltn-r1-11 · retention factor dataset
   Fruits and Fruit Juices (group 9), vitamin C: code 158 FRUITS,DRIED 20% retained; code 154 FRUITS,CANNED 50% retained; code 253 FRUITS(DRIED),SAUTEED 70% retained; code 155 FRUITS,FRESH(NOT CITRUS),STEWED 70% retained
   USDA Table of Nutrient Retention Factors, Release 6, 2007 · checked 2026-09-24
6. pltn-r1-01 · systematic review and meta-analysis of RCTs
   GB-based food consumption significantly increased the hazard of resolution of diarrhea compared to standard treatment (HR 1.96, 95% CI [1.62; 2.37], p < 0.01; I2 = 52%).
   Eur J Pediatr, 2024 · checked 2026-09-24
7. pltn-r1-02 · randomised controlled trial
   The average duration of diarrhea in the plantain-based diet group was 18 hours shorter (p < 0.005) and it also had lower cost (p < 0.005).
   J Am Coll Nutr, 2009 · checked 2026-09-24
8. pltn-r1-03 · randomised controlled trial
   The proportion of children who recovered within 72 hours was significantly higher (62.4%) in the green banana group compared to the control group (47.2%) [RR 1.3 (95% CI 1.05-1.7), NNT=7].
   Indian Pediatr, 2020 · checked 2026-09-24
9. pltn-r1-06 · prevalence study
   Sixty patients (30%) showed skin reactivity to date palm profilin, Pho d 2...The large majority of profilin-allergic patients reported oral allergy syndrome as the only food-induced symptom and were able to tolerate the offending foods if they were cooked or otherwise processed.
   Clin Exp Allergy, 2008 · checked 2026-09-24
10. pltn-r1-07 · cross-sectional study
    while having banana/plantain and tubers as the content of their most consumed meal significantly increased the odds of total crude hypocalcaemia (AOR = 1.37, p-value = 0.012)
    PLoS One, 2022 · checked 2026-09-24
11. pltn-r1-08 · systematic review and meta-analysis
    Meta-analysis of the results found a large proportion of I2 values for all studies (98%), meaning heterogeneity. As a consequence, the combined effect sizes were not useful.
    Nutrients, 2022 · checked 2026-09-24
12. pltn-r1-09 · methods section
    Additionally, no human studies were included.
    Nutrients, 2022 · checked 2026-09-24