Krista Varady has been studying intermittent fasting for two decades at the University of Illinois Chicago, and in a 2024 commentary, she and her co-authors set out to systematically debunk four common safety myths about the practice. Eating disorders. Muscle loss. Sex hormone disruption. Poor nutrient intake. Each one, the paper argues, collapses under the weight of the clinical evidence. What the paper does not address — because the clinical trials it draws from largely did not measure it — is what happens to the trillions of bacteria living inside the people doing the fasting.
That absence is the story.
Intermittent fasting is now one of the most-studied dietary interventions of the last twenty years. And yet the microbiome question, arguably the most consequential one for long-term health, is running roughly a decade behind the metabolic and cardiovascular literature. The gut is where the research is only now catching up.
The metabolic case arrived first, and it arrived loudly
The reason microbiome data lagged is partly a matter of what the field was looking for. Fasting research took off in the 2010s as a metabolic story: blood glucose, insulin sensitivity, LDL cholesterol, waist circumference, body weight. Those endpoints were cheap to measure, familiar to clinicians, and mapped cleanly onto the obesity epidemic that funders wanted addressed.
By 2024, a systematic review and network meta-analysis of randomised clinical trials on intermittent fasting had been published in the BMJ. Alternate-day fasting, 5:2, time-restricted eating with an 8-hour or 10-hour window — each variant had its own trials, its own weight-loss curves, its own patient population.
The verdict on weight loss was almost boring in its consistency. Intermittent fasting produced weight loss roughly equivalent to standard calorie restriction, without requiring people to count anything. For a population that had spent thirty years being told to log every gram of fat, this was a marketable finding.
The heart data was messier. A 2024 preliminary analysis of more than 20,000 US adults suggested that eating windows of under 8 hours per day were associated with a higher risk of cardiovascular death compared with 12-to-16-hour windows. Other reviews found the opposite: reductions in LDL cholesterol, triglycerides, and fat mass. The signal was mixed enough that the American Heart Association urged caution without condemning the practice.
What almost none of these headline studies measured, at least until recently, was the stool sample.
Why the microbiome research was slower
Gut microbiome science is expensive. Sequencing the bacterial DNA in a single sample requires significant investment per participant. Longitudinal studies — the kind that would track how a person’s microbial community shifts over 8, 12, or 24 weeks of daily fasting — require multiple samples per subject, careful storage, and bioinformatics pipelines that most nutrition labs did not have in-house.
The result is a literature that skewed toward small, short-duration studies, often in specific populations. The most cited early evidence came from Ramadan cohorts: Muslim adults who fast from dawn to sunset for roughly 30 days, providing a natural experiment for extended daily fasting windows.
Studies on Ramadan fasters have documented measurable reductions in gut microbial diversity across the fasting month. Some genera — particularly those associated with anti-inflammatory short-chain fatty acid production — declined. A subset rebounded after the fast ended. Others did not return to baseline within the follow-up window.
That distinction matters more than it sounds. Microbial diversity is not just a health metric. It is a stability metric. A less diverse gut community is a more fragile one — more reactive to a single restaurant meal, more prone to inflammatory flare-ups, less able to metabolise the fibre that eventually returns to the diet.
The fuel problem
The mechanism connecting fasting duration to microbiome damage is not mysterious. The bacteria that keep the gut lining alive feed on fermentable fibre. When food stops arriving, so does the fibre. The bacteria that depend on it — the fibre fermenters that produce butyrate, propionate, and acetate — start to shrink in number relative to bacteria that can survive on mucin, the protective layer coating the intestinal wall.
Those short-chain fatty acids are the primary fuel source for colonocytes, the epithelial cells that form the single-cell-thick barrier between the contents of your gut and the rest of your body. That barrier renews itself every few days. Renewal requires fuel. Fuel requires bacteria. Bacteria require fibre. Fibre requires eating.
Extend a daily fast to 18 hours, then 20, then push it into alternate-day territory, and the fuel supply thins. The mucin-degrading species — Akkermansia muciniphila is among them — expand their share of the community. Some evidence suggests this is beneficial in moderation. Other evidence suggests that when mucin-degraders dominate for extended periods, they thin the mucus layer itself, exposing the epithelium to bacterial products it was never meant to touch.
That exposure has a name: metabolic endotoxemia. It is the low-grade inflammation that the fasting literature was supposed to be reducing.
Population responses vs individual damage
The tension inside the research is that population-level averages keep looking fine. Trials of 12 or 16 weeks of time-restricted eating tend to show either neutral or modestly positive effects on inflammatory markers. Body weight drops. Insulin sensitivity improves. Inflammation markers trend downward in most cohorts.
But population averages hide the tails. The individual variation in microbiome response to fasting is enormous, and it correlates with what the participant was eating during the feeding window. A person eating whole grains, legumes, and vegetables inside an 8-hour window is running a fundamentally different experiment on their gut than a person eating ultra-processed food inside the same 8-hour window. The fasting duration is identical. The microbial outcome is not.
Almost no fasting trial has been designed to separate these two variables cleanly. Most studies let participants eat what they want during the feeding window and measure what happens on average. That design decision, made for good reasons of ecological validity, has left a specific question unanswered: how much of the observed benefit — or harm — is the fasting itself, and how much is the food composition?
The behavioural layer nobody wants to touch
The other reason microbiome data lagged is that fasting research has always had to fight for methodological respect. For most of the 2000s, extended fasting was coded as a fad, adjacent to juice cleanses and detox teas. Serious nutrition scientists were reluctant to attach their names to it. The shift from fad to scientifically supported eating pattern happened only in the last decade, driven largely by animal studies at institutions like the Salk Institute and human trials at UIC, Johns Hopkins, and King’s College London.
Once the metabolic evidence firmed up, the funding followed. But microbiome measurement was not standard practice in those early trials, and by the time it became standard, the field had already produced hundreds of studies with no stool data at all. Reconstructing the microbiome story is now a retrospective exercise for the interventions that were designed and completed before anyone was thinking about it.
A 2026 study reported by Science Daily found that intermittent fasting produces measurable changes in the brain, and while the headline focus was neurological, the paper joins a growing cluster of work treating fasting as a whole-system intervention rather than a metabolic lever. The gut-brain axis makes it impossible to ignore the microbiome much longer.
What the practice looks like when it goes wrong
The clinical picture of fasting-induced microbiome damage is not dramatic. There is no acute crisis. What people describe, months into an aggressive fasting protocol, is chronic bloating on refeeding, food sensitivities that were not there before, brain fog that does not lift with sleep, and — the paradox — weight loss that stalls and then reverses despite eating less than they used to.
Some of that is caloric adaptation. Some of it is behavioural drift. But a growing body of case-series evidence points to what happens when a diverse microbial community collapses into a narrow one that struggles to metabolise reintroduced food. The gut becomes reactive. Small amounts of fibre produce disproportionate bloating. Meals feel heavier than their calorie count suggests.
None of this is captured in a 12-week trial that measures body weight and LDL cholesterol.
The comparison to other over-studied supplements
The pattern of a health intervention accumulating hundreds of trials while a fundamental biological question goes unmeasured is not unique to fasting. Silicon Canals has explored a similar structural gap in the supplement world, where creatine has been tested across more than a thousand trials without the field settling several basic questions about how it interacts with vegetarian and vegan physiology. The interventions get studied; the sub-questions that would change clinical advice for specific populations wait their turn.
Fasting is arguably a worse case, because the microbiome sub-question is not niche. It affects everyone who fasts. And unlike creatine, where the muscle physiology endpoints are well-characterised, the gut endpoints of fasting are still being defined.
What the Varady paper does and does not settle
The UIC commentary is careful. It explicitly limits its claims to the four myths it addresses. It does not claim that fasting is safe for every population. It excludes people with a history of eating disorders. It notes that adolescents require special caution. It defers on cardiovascular disease patients, echoing the concerns raised in the time-restricted eating cardiovascular signal that appeared in 2024.
What the paper does not do — and what Varady’s team has been transparent about in interviews — is claim that the microbiome question has been resolved. The clinical trials the commentary draws on were not designed to answer it. The answer, when it arrives, will come from a different generation of studies now underway in labs that treat the stool sample as a primary endpoint rather than an afterthought.
What the next decade of research will have to answer
The questions that fasting science has not yet resolved are practical ones. How long does it take a damaged microbiome to recover after a person stops fasting? Which species come back and which do not? Does the composition of the feeding-window diet predict microbial outcomes more strongly than the fasting duration itself? At what point does an intermittent practice become a chronic caloric restriction that behaves differently in the gut?
None of these questions are unanswerable. They are just expensive, and they require longitudinal designs that funders have not historically prioritised for behavioural and dietary interventions. That is beginning to change. The labs now treating the stool sample as a primary endpoint rather than an afterthought are the ones that will decide whether the microbiome turns out to be a footnote to the fasting story or its central chapter.
Watch nextRUNTIME 8:05 · VEGOUT
Worth watching · VegOut on YouTube
Fasting Done Wrong Destroys Your Gut This video goes deeper on the same story Watch on YouTube
