August 15, 2018

Fiber Feeds Bacteria to Control Type 2 Diabetes Mellitus

Investigators pit high-fiber diet against standard diet for glycemic control
In a randomized clinical trial, participants with type 2 diabetes consuming a high-fiber diet showed better glycemic control, an expected finding. But the mechanism underlying fiber’s benefits, as suggested in the study, challenge conventional wisdom.

This paper is part of NMJ's 2018 Microbiome Special Issue. Download the full issue here.


Zhao L, Zhang F, Ding X, et al. Gut bacteria selectively promoted by dietary fibers alleviate type 2 diabetes. Science. 2018;359(6380):1151-1156.


To determine effects of a high-fiber diet on the gut microbiome and glucose regulation in individuals with established type 2 diabetes.


Randomized, open-label, parallel-group clinical trial


Forty-three patients with clinically diagnosed type 2 diabetes mellitus were randomized to either an intervention group (n=27) or the control group (n=16).

Study Medication and Dosage

The experimental group received a high-fiber diet of fresh vegetables, fruits, and nuts, supplemented with a gruel (which included whole grains, beans, peanuts, lotus seed, and yam), bitter melon, and prebiotics. The control group consumed an isocaloric diet based on Chinese Diabetes Society guidelines.

Both groups took the antidiabetic agent acarbose (an amylase inhibitor) and discontinued any previously used glycemic control medications. Acarbose transforms starch into fiber by reducing digestion and making it available as a fermentable carbohydrate to bacteria in the colon. The experimental group had a much higher intake of dietary fiber but daily energy and macronutrient intake was similar between groups.

To determine that interactions between the gut microbiota and the fiber were responsible for any observed changes in function, the gut microbiota from before and after the interventions were transplanted into germ-free mice. These mice ended up with gut biomes that more resembled the transplanted biome of the donor than they resembled each other.

Outcome Measures

Hemoglobin A1c (HbA1c) was the primary outcome measure. Additional outcomes included proportion of participants who achieved glycemic control, fecal short chain fatty acid (SCFA) levels, postprandial glucose, fasting blood glucose, lipid profiles, and other standard metabolic markers.

Key Findings

Increasing fermentable fiber by blocking carbohydrate digestion via the amylase inhibitor acarbose improved markers of type 2 diabetes in both groups, but the high-fiber group did significantly better. Hemoglobin A1c decreased in both groups but more so in the high-fiber group. Reduction in HbA1c was greater in the high-fiber group from 4 weeks onward. A greater portion of patients in the high-fiber group reached adequate glycemic control (HbA1c<7%) compared to the standard-diet group (89% vs 50%). The high-fiber group lost more weight and had better lipid profiles. Improvements came faster and were greater in patients who consumed a high-fiber diet in addition to the enzyme inhibitor.

Germ-free mice transplanted with post-intervention microbiota derived from either patient group did better, showing better metabolic health parameters than mice transplanted with pre-intervention microbiota. Mice transplanted with microbiota obtained post-intervention from the high-fiber group did the best, having the lowest fasting and postprandial blood glucose levels of all mice, mirroring the results in the human patient group.

Metagenomic sequencing was performed on 172 fecal samples collected at 4 time points (days 0, 28, 56, and 84), which led to a catalog of 4,893,833 nonredundant microbial genes. Both patient groups had a reduction in gene richness (the number of genes identified per sample) from day 0 to day 28, along with significant clinical improvements, with no further changes afterward.

These last data challenge the current notion that greater overall diversity implies better health. However, gene richness tended to be higher in the high-fiber group than in the normal diet group after day 28, and this trend was associated with better clinical outcomes in the intervention group.

We already know that high-fiber diets help control diabetes, but we have generally thought this benefit was because higher fiber would lower the glycemic index of carbohydrates.

The high-fiber diet favored the growth of bacteria that produce SCFAs, especially bacteria that produce butyric acid. This increase in acid production significantly lowered gut pH in the high-fiber group. Fifteen bacterial strains were significantly promoted by the high-fiber diet, and 47 strains were significantly reduced. This response was clearly strain-specific; for example, of the 6 strains of Faecalibacterium prausnitzii identified, only 1 strain was significantly promoted by the high-fiber diet. The 15 strains that were promoted were all significantly associated with increased SCFA production, which was inversely correlated with HbA1c.

Practice Implications

This paper further demonstrates that human intestinal microbiota affect blood sugar control. This study suggests that we can improve glycemic control by shifting bacterial populations in the intestinal microbiome. We already know that high-fiber diets help control diabetes, but we have generally thought this benefit was because higher fiber would lower the glycemic index of carbohydrates.1 We have repeated this idea for years even as we began to appreciate that there is little difference in glycemic index between whole grain and white flour products. We can now imagine that the difference in action was that the whole-grain versions provided more fermentable carbohydrates and shifted gut microbiota.

Acarbose is available in the United States and Canada by prescription only but rarely used. In China it is the most common prescription employed for treating early type 2 diabetes, the same indication used in the United States for which we might prescribe metformin.2 A 2014 Chinese study that compared acarbose against metformin showed that both agents decreased HbA1c levels to similar degrees and were equally effective in controlling the disease.3 Metformin also changes the gut microbiota,4 increasing levels of Akkermansia.5 Diabetes is now considered a disease of the gut microbiota.6,7 The botanical extract berberine, which we have often used clinically to replace metformin, also shifts the gut biome in a similar manner as metformin.8

We should note that pharmaceutical companies are experimenting with combining metformin and acarbose together in a single tablet.9

The present study suggests a benefit from inhibiting amylase enzyme action. In this study acarbose was used to block starch digestion. A number of fruits have a similar action. If we were willing to anthropomorphize fruits, we could see how they would prefer any of their consumers to get diarrhea. Such digestive upset increases the odds of the fruit seeds’ dispersal in the vicinity and provides an evolutionary advantage. Thus many fruits contain chemicals that act as amylase inhibitors, including baobab fruit,10 persimmon,11 mango,12 pomegranate,13 a range of anthocyanin-containing berries such as cherries,14 and, in general, polyphenols.15

This study provides a mechanistic explanation for why a diet high in vegetables, which provide fiber, and fruits, which may have anti-amylase action, is useful in treating type 2 diabetes: the gut biome shifts to increase SCFA production. Diabetics are often cautioned against eating fruit, but this advice may actually be counterproductive; in fact, sugar-sweetened fruit juices are significantly associated with the risk of developing type 2 diabetes, but whole fruits16 and 100% fruit juices17 are not. It may eventually prove useful to know the relative amylase inhibitory action and prebiotic content of various types of fruit. A number of classic antidiabetes botanicals are also amylase inhibitors, including Ocimum basilicum (basil)18 and mango.19

In 2013, Schor reviewed a study in this journal20 that suggested various berry jams lower glycemic impact of the bread they are eaten with. In the light of this present study, that earlier information makes better sense. The berry concentrates may have acted as anti-amylose agents, similar to acarbose, while providing fiber content. (Might we suggest toast, jam, and berberine as a possible breakfast for diabetics?)

This study’s results suggest that it may someday be possible to create a probiotic supplement that, taken in combination with a high-fiber diet, may have a significant effect on improving glycemic control. Our only evidence using direct microbial transplantation in humans is a study from 2012 that used fecal microbial transplant (FMT) from healthy, lean donors for men with metabolic syndrome. The men experienced temporary increases in peripheral insulin sensitivity, with a trend toward improved hepatic insulin resistance.21 These changes were related to fecal microbial diversity and increases in SCFAs.22

These data should also prompt us to rethink resistant starches and how they affect diabetes. In the past we thought their benefit was secondary to low glycemic index. Instead their benefit may be in reaching the colon and increasing SCFA production.

Categorized Under