Benefits
Stool output and recovery time with cooked green banana in infant diets
In a double-blind trial of 62 hospitalized infant boys in Bangladesh with prolonged diarrhea, a rice diet with cooked green banana for 7 days led to recovery by day 4 in 78% versus 23% on rice alone, with less stool output and fluid need. A pooled review of nine randomized trials in young children, most of them open-label, found faster recovery when green banana foods were added to standard care.
Stool form with green banana biomass in children
In a randomized study of 80 children and adolescents with hard, infrequent stools, green banana biomass alone reduced hard stools, straining and painful bowel movements from baseline, but did not raise the share with more than three bowel movements a week. Combined with laxatives, it allowed lower laxative doses in many children. There was no placebo group.
Post-meal blood sugar response by ripeness and banana starch
In 10 adults with type 2 diabetes, 120 g of under-ripe banana raised blood sugar less than over-ripe banana or white bread (glycemic index about 43 versus 74 for over-ripe), with similar insulin responses. In 28 healthy adults, a drink of 40 g native banana starch gave lower glucose and insulin responses than digestible corn starch, though it held far less digestible carbohydrate.
Body weight and later food intake with native banana starch
In a blinded crossover of 30 adults with obesity and type 2 diabetes, 24 g a day of native banana starch for 4 weeks lowered body weight by about 1.6 kg versus 0.3 kg on a soy milk control; insulin measures did not differ from control. In 28 healthy adults, one 40 g dose led to lower intake at a free meal 3 hours later, without changing appetite ratings.
HbA1c and body composition with green banana biomass over 24 weeks
In a randomized trial in 113 Brazilian adults with prediabetes or type 2 diabetes on usual treatment, 40 g a day of green banana biomass (about 4.5 g resistant starch) plus diet advice for 24 weeks was followed by lower HbA1c, fasting glucose, weight and fat mass from baseline. The diet-only group also lowered HbA1c, and a between-group test was not reported in the abstract.
Carbohydrate fuel during long cycling compared with a sports drink
In two randomized crossover trials funded by a banana company, trained cyclists who ate ripe bananas during a 75 km time trial performed and recovered about the same as with a 6% carbohydrate drink at the same carbohydrate dose. Bananas left riders feeling more full and bloated. Against water only, one trial found a 5% faster time with bananas and another found no difference.
A food source of potassium and vitamin B6
By USDA figures, a medium banana (118 g) supplies about 420 mg of potassium and 0.43 mg of vitamin B6, plus about 27 g of carbohydrate. The adult potassium intake target is 2,600 mg a day for women and 3,400 mg for men. In 9 men after exercise in the heat, one or two bananas raised blood potassium only slightly and slowly.
Mechanism of action
Ripening turns starch into sugar
In under-ripe bananas, starch makes up most of the carbohydrate, and as the fruit ripens it changes into free sugars. Native green banana starch is a type 2 resistant starch, made of raw granules that human amylase digests slowly, which fits the lower blood sugar rise seen with under-ripe fruit than with over-ripe fruit.
Resistant starch carries fewer digestible calories to the small intestine
A 40 g dose of the native banana starch used in trials held about 26.8 g of resistant starch and only 9.3 g of digestible starch, compared with 37.8 g digestible starch in corn starch. Resistant starch passes to the colon for fermentation, which is the proposed basis for the lower glucose response and lower later food intake seen in short studies.
Potassium from bananas enters the blood slowly
In 9 men after an hour of cycling in the heat, eating 150 or 300 g of banana left plasma potassium about 0.1 to 0.2 mmol/L higher at 60 minutes than eating nothing (4.5 and 4.6 versus 4.4 mmol/L), within the normal range. The author concluded these changes come too slowly to relieve an acute exercise cramp.
Clinical trials
Randomized crossover of bananas versus a 6% carbohydrate sports drink, each at 0.2 g/kg carbohydrate every 15 minutes during a 75 km time trial; funded by Dole Foods, with Dole scientists on the team (Nieman et al. 2012, PLoS One)
14 trained cyclists.
Time trial performance (2.41 vs 2.36 hours, P = 0.258) and blood glucose did not differ between bananas and the drink, and inflammation, oxidative stress and immune measures changed similarly. Bananas led to higher IL-10, IL-8 and plasma antioxidant capacity and higher dopamine. Riders felt more full and bloated with bananas. No placebo; the comparison was another carbohydrate.
Randomized, counterbalanced crossover of four 75 km time trials (Cavendish banana, mini-yellow banana, 6% sugar drink, water only) with 2-week washouts; funded by Dole Foods (Nieman et al. 2018, PLoS One)
20 cyclists, overnight fasted.
All three carbohydrate trials raised post-exercise glucose and lowered some inflammation markers compared with water, and bananas and the sugar drink reduced overall metabolic disturbance to a similar degree. Lab tests of monocytes cultured in plasma showed lower COX-2 gene expression after bananas but not the sugar drink, versus water. In the full paper, performance times did not differ among the four trials, including water only.
Randomized, blinded within-subject crossover: 24 g/day native banana starch or 24 g/day soy milk powder in water for 4 weeks each, with no washout; funded by the Mexican Social Security Institute (Ble-Castillo et al. 2010, Int J Environ Res Public Health)
30 adults with obesity and type 2 diabetes in Mexico (28 completed, 24 of them women).
Body weight fell 1.57 kg on banana starch versus 0.3 kg on soy milk (P = 0.0021), with BMI also lower. Fasting insulin and HOMA-IR fell from baseline on banana starch but did not differ significantly from the control. Three people in each group had flatulence, more frequent or soft stools in the first week. Soy milk is not an inert control.
Randomized, single-blind crossover: a drink with 40 g native banana starch (from green Dwarf Cavendish) or 40 g digestible corn starch, 1-week washout (Ble-Castillo et al. 2017, Nutrients)
28 healthy young adults.
Energy intake at a free test meal 3 hours later was lower after banana starch, and glucose and insulin responses over 180 minutes were lower. Appetite ratings did not favor banana starch (some satiety and fullness ratings were higher with corn starch), and gut hormones did not differ overall, apart from higher PYY at 60 minutes and GLP-1 at 120 minutes after banana starch. The banana drink held less digestible starch and fewer calories (104 vs 151 kcal), so the arms were not carbohydrate-matched.
Test-meal comparison on separate days of 120 g under-ripe banana, 120 g over-ripe banana and 40 g white bread (Hermansen et al. 1992, Diabet Med)
10 adults with type 2 diabetes in Denmark.
The blood glucose response area was lowest with under-ripe banana (62), higher with over-ripe banana (106) and highest with white bread (181), and the glycemic index was 43 for under-ripe versus 74 for over-ripe banana. Insulin responses were similar across the three meals. Single test meals only.
Randomized controlled trial of nutritional support plus 40 g/day green banana biomass (about 4.5 g resistant starch) versus diet alone for 24 weeks, on top of usual treatment (Costa et al. 2019, Br J Nutr)
113 middle-aged adults in Brazil with prediabetes or type 2 diabetes (62 biomass, 51 diet only).
Within the biomass group, HbA1c, fasting glucose, diastolic blood pressure, weight, BMI, waist and fat mass fell from baseline. The diet-only group also lowered HbA1c and waist. The abstract reports within-group changes and does not give a between-group test for the main outcomes. Not blinded; no placebo food.
Double-blind randomized trial of a rice-based diet with 250 g/L cooked green banana, with 4 g/kg pectin, or rice diet alone for 7 days, in hospital (Rabbani et al. 2001, Gastroenterology)
62 boys aged 5 to 12 months in Bangladesh with persistent diarrhea (22 banana, 19 pectin, 21 control).
By day 3, 55% on green banana, 59% on pectin and 15% on rice alone had recovered; by day 4, 78%, 82% and 23% (P < 0.001 versus control). Green banana and pectin also reduced stool amount, oral and intravenous fluid needs, vomiting and illness duration. Sick infants under medical care, not a general gut health population.
Randomized five-arm study: green banana biomass alone, biomass plus PEG 3350, biomass plus sodium picosulfate, PEG 3350 alone or sodium picosulfate alone (Cassettari et al. 2019, J Pediatr (Rio J))
80 children and adolescents in Brazil with functional constipation (Rome IV).
With biomass alone, the share of children with hard stools (Bristol 1 or 2), straining, painful bowel movements and abdominal pain fell, but fecal incontinence and blood in stool did not, and the share with more than three bowel movements a week did not rise. Lower laxative doses were needed by 87% (with picosulfate) and 63% (with PEG). No adverse effects reported. No placebo arm.