Fasting May Help Treat Type 1 Diabetes by Altering Gut Microbiome

Autoimmune type 1 diabetes (T1D) is about more than just insulin and blood sugar. A growing body of evidence links the gut microbiome to the risk, course, and associated inflammation of autoimmune diseases. Diet is one of the fastest ways to tweak the microbiota, so interest in therapeutic fasting is natural: it has already changed the composition of microbes and immune circuits in healthy people and in a number of autoimmune diseases. But how exactly the microbiome of people with T1D would respond to fasting has been unclear until now. A new study in Frontiers in Endocrinology closes part of this gap, showing that a week of medically supervised fasting dramatically and briefly restructures the microbiota in T1D, with the shift bringing it closer to the profile of healthy people - and surprisingly partially overlapping with what is seen in another autoimmune disease, multiple sclerosis (MS).

Background of the study

Type 1 diabetes mellitus (T1DM) is an autoimmune disease in which the immune system destroys β-cells of the pancreas; it is estimated that about 9 million people worldwide live with it. In addition to genetics, environmental factors significantly affect the risk and course of T1DM, and in recent years, the intestinal microbiome has become one of the key “suspects”: in people with T1DM, its composition and functions differ from healthy people, and changes in the microbiota have been described even before the onset of the disease; more often, increased intestinal permeability and a shift in metabolites that affect immunity (short-chain fatty acids, derivatives of vitamin A, tryptophan, etc.) are recorded. All this fits into the idea that “intestinal ecology” can twist the immune response and the course of autoimmunity.

Diet is the fastest lever for influencing the microbiota, so there is growing interest in therapeutic fasting and “post-mimetic” approaches. In models and healthy volunteers, prolonged food pauses restructure the microbial composition, and in animal experiments, repeated cycles of a “fasting-mimicking diet” reduced the pool of autoaggressive T cells and supported regulatory T cells; similar signals were also obtained in a model of multiple sclerosis. However, the question remained: how would the microbiome of people with T1D respond to fasting and whether the “microbial signatures” of fasting previously described in other groups would be repeated.

There is also a safety aspect. Historically, long-term dietary restrictions have been considered risky in T1D due to the risk of hypo/hyperglycemia and ketoacidosis. However, controlled safety data are accumulating: Ramadan fasting has been safely completed in selected patients, and no severe adverse events, including DKA, have been reported with medically supervised 7-day fasting. This opens the door for careful clinical protocols where the goal is not to “starve diabetes” but to study short, controlled interventions to understand mechanisms and potential adjuvant effects.

Against this backdrop, the pilot from Frontiers in Endocrinology formulates a clear hypothesis: if “deficiency of nutrient substrates” is a strong, disease-independent driver of microbiota reorganization, then a week-long fast should induce a signature of changes in T1D similar to those seen in healthy individuals and other autoimmune conditions. The next step is to test how reproducible these shifts are, how long they last, and whether they are at least associated with changes in clinical parameters (lipids, blood pressure), in order to decide whether to move on to larger and longer trials.

How the study is structured (who, what and when)

The pilot study included 19 adults with T1DM (95% women) and 10 healthy controls. All underwent a 7-day course of therapeutic fasting in an inpatient setting (not a hospital, but under observation): ~200 kcal/day due to vegetable broths, juices and oat broth; water and herbal teas - without restrictions. Feces were collected: on day 0 (before), day 7 (immediately after) and at day 150 (after ~5-6 months); the composition of the microbiota was assessed by 16S sequencing. Separately, the authors added a subsample from the NAMS study on MS: 10 patients with MS underwent two weeks of fasting with an interval of 6 months (between them - a daily interval window of 14 hours), the diet during the fasting phase was up to ~400 kcal/day.

What has changed in the microbiota - the main thing

The most notable finding: in patients with type 1 diabetes, the microbiota "jumped" after starvation - according to the beta-diversity, the composition on the 7th day had already converged to the profile of healthy people, while in controls the overall pattern for the same week hardly changed statistically (probably due to the small group). By the 150th day, the effect had subsided - a stable "new equilibrium" did not arise.

When broken down by genus, 21 taxa showed differential changes in people with T1D after fasting. Although the controls had lower significance, the direction of the shifts was the same. For example:

• Decrease: Agathobacter, Fusicatenibacter, Oscillospiraceae UCG-003;
• Growth: Escherichia/Shigella, Ruminococcus torques group, Ruminococcaceae UBA1819.

On a more subtle level (ASV, "almost species-specific"): only in DM1 did Bacteroides vulgatus and one of the Prevotella grow, while in controls Roseburia intestinalis and a number of other ASVs fell. In total, this confirms that fasting gives a short but powerful "click" on the microbiota, and the details depend on the initial status.

"Hunger Signature": Repeatable Shifts in T1DM, MS, and Healthy Individuals

Comparison with the MS group revealed a microbiome "starvation signature" independent of the disease. Seven genera changed in the same direction in all: Agathobacter, Bifidobacterium, Fusicatenibacter and Lachnospiraceae UCG-001 decreased, and Erysipelatoclostridium, Escherichia/Shigella, Eisenbergiella increased - and this is also shown by larger studies in non-autoimmune populations. In the second phase, MS showed high reproducibility: about half of the significant ASVs were repeated in both weeks of starvation. The picture coincides with the general biology of starvation: "plant fiber lovers" (many Lachnospiraceae) decline, and mucin- and glycosaminoglycan-destroyers ( R. gnavus, R. torques, Hungatella ) increase, switching to host resources; Eisenbergiella is associated with ketosis and may use β-hydroxybutyrate as a fuel.

Is this related to health indicators?

The authors compared the "bacterial" shifts with changes in clinical markers in T1DM and controls. After adjusting for multiple comparisons, they obtained 9 significant associations. For example, Oscillospiraceae UCG-002 correlated with the dynamics of LDL, and in controls - also HDL and diastolic pressure; the growth of Erysipelatoclostridium (controls) and Romboutsia (T1DM) coincided with a decrease in blood pressure; Lachnospira "went" along with a drop in urinary citrate in T1DM. These are correlations, not causality, but they rhyme with the literature on the effect of individual taxa on lipids and vascular tone.

How does this fit into the physiology of hunger?

The logic is simple: when there is a deficit of food substrates, microbes with broad metabolic capabilities and access to the host's resources - mucus (mucin), glycosaminoglycans, ketone bodies - win. Therefore, fasting naturally shifts the ecosystem from active fermenters of dietary fiber ( Agathobacter and its relatives are large producers of butyrate, they "love" fiber) to "generalists" and "slime eaters". Similar changes (including the growth of Akkermansia ) have already been described after 3-10-day fasts in other groups; the current work shows that the direction remains the same in type 1 diabetes.

What does this mean for people with T1D?

• This is about the microbiome, not "treating diabetes" with starvation. The changes are short-term and primarily concern the composition of bacteria; no stable long-term "restructuring" has been recorded over 5-6 months.
• Safety is key. A seven-day fast in type 1 diabetes is possible under monitoring conditions (no DKA was observed in the pilot studies), and there is data on the safety of Ramadan fasting in selected patients. But this is no reason to experiment at home - the risks of hypo/hyperglycemia and ketoacidosis are real.
• Where is the practical benefit? The researchers hint at two vectors: (1) to understand which taxa are associated with improvements in blood pressure and lipids; (2) to test whether the hunger signature can be mimicked by “soft” dietary measures (eating windows, diet composition) or probiotics/prebiotics, without a full week of fasting.

Restrictions

This is a pilot with small groups; the main statistics were "pulled" by DM1, the significance sagged in controls. Method - 16S (taxonomy, not functions); virus/mycobiome were not profiled. Correlations with clinical markers are associative; cause-and-effect relationships between specific bacteria and, say, LDL have yet to be verified. And, finally, the effect turned out to be transient - the "imprint" of starvation is erased within months.

What should science do next?

• Larger RCTs with clinical objectives (glycemic variability, blood pressure, lipids), multi-omics (metagenomics/metabolomics) and monitoring of the durability of effect.
• Comparison of regimens: fasting week vs. interval window (e.g. 14-16 hours), ketogenic phase, “post-mimetic” protocols.
• Microbiota targets: testing whether the 'hunger signature' can be recapitulated via diet/supplements without strict fasting in T1D.

Source: Graef FA et al. Fasting elicits gut microbiome signature changes that extend to type 1 diabetes patients. Frontiers in Endocrinology, August 13, 2025. DOI 10.3389/fendo.2025.1623800