Dr. Axe Wire · Study decoded · gut

Food-Based Resistant Starch Shows No Metabolic Shift in Women

A pilot trial testing resistant starch from whole foods found no difference in blood pressure, metabolic markers or gut bacteria diversity after eight weeks.

By Dr. Josh Axe · November 14, 2026 · 1 source · Watch

Photo: Pexels

A food-based resistant starch trial published November 11, 2026 in Food & Function found no difference in blood pressure, metabolic syndrome markers or overall gut bacteria diversity after eight weeks, even though the high-intake group ate more than three times as much resistant starch as the control group.

30women, 8 weeks

7.9g vs 2.4gRS daily

The trial randomized 30 women aged 40 and older, each with at least two metabolic syndrome features, to a resistant starch-rich food pattern or a control diet for eight weeks. The high-intake group averaged 7.9 grams of resistant starch a day from whole foods, versus 2.4 grams a day in the control group. After adjusting for baseline differences and correcting for multiple comparisons, the researchers found no difference between groups in systolic blood pressure, where the adjusted gap between groups was small and the range of possible results included zero, in the primary metabolic syndrome measures, or in overall gut microbiome diversity and composition. Secondary measures, including urinary organic acids, a gut-inflammation marker called fecal calprotectin, and heart rate variability, also showed no difference after correction.

An exploratory look at individual bacterial genera found one, Veillonella, with higher levels in the high-resistant-starch group at week eight, but this did not remain significant once the researchers accounted for testing many genera at once. A possible link between a urinary marker called the lactate-to-pyruvate ratio and fasting glucose also weakened after removing one outlying data point.

Resistant starch reaches the colon largely intact, where gut bacteria ferment it into short-chain fatty acids, compounds linked in other research to blood sugar control and gut lining health. The authors describe the Veillonella and lactate-to-pyruvate findings as hypothesis-generating signals, not confirmed effects, since neither survived correction for testing many outcomes at once.

Resistant Starch Intake Compared

Food & Function, 2026, Lee et al.

This pilot trial of 30 women over eight weeks is small and short, so larger and longer trials are needed before drawing conclusions about whole-food resistant starch and metabolic health; the exploratory genus-level and urinary marker findings also need confirmation.

Dr. Axe’s Take

This trial found no shift in blood pressure, metabolic markers or gut bacteria diversity after eight weeks of eating more resistant starch from whole foods, and I take that result at face value rather than reading more into the Veillonella signal, which lost significance once the researchers corrected for testing many bacteria at once. Eight weeks is a short window for gut and metabolic changes to show up on standard tests. I'd still build meals around resistant-starch foods like cooked-and-cooled potatoes, green bananas, oats and legumes, since their fiber and fermentation benefits are supported by research well beyond this one small trial.

— Dr. Axe

Take the Metabolic Health Quiz →

Study summary

Who: 30 women aged 40 and older with at least two features of metabolic syndrome.
Design: Double-blind, randomized pilot trial; 1:1 assignment to high-resistant-starch or control food pattern for 8 weeks.
Dose: High-RS group averaged 7.9 g resistant starch per day from whole foods; control group averaged 2.4 g per day.
Primary result: No significant difference between groups in systolic blood pressure, with the adjusted gap between groups small and consistent with no effect, or in gut microbiome diversity and composition.
Secondary: Urinary organic acids, fecal calprotectin and heart rate variability showed no difference after correction. Veillonella abundance was higher in the High-RS group at week 8 (adjusted difference 3.24; 95% CI 1.48-5.00) but did not remain significant after correction for multiple comparisons (q=0.070).
Funding / conflicts: Funder listed in the journal record: Rural Development Administration. Conflicts not reported in the abstract.
Date: Nov 11, 2026