Two weeks of antibiotics made mice leaner. Ten weeks made them fatter.

Most people think of an antibiotic prescription as a yes-or-no event: you either took the course or you didn't. A new mouse study suggests that framing misses the variable that may matter most for metabolism. Researchers at Sichuan University gave 100 mice the same six-drug antibiotic cocktail and changed exactly one thing between groups: how many weeks it ran. Two weeks left the animals leaner and better at burning calories as heat. Ten weeks of the identical drugs did the opposite.

Key Takeaways

• In high-fat-fed mice, a two-week antibiotic course lowered body fat percentage and fasting glucose and raised thermogenic gene expression, while a ten-week course of the same cocktail raised body fat and fasting glucose and suppressed those same genes.
• The mice ate the same number of calories in every high-fat group, so the divergence came from energy expenditure, not appetite.
• Gut bacteria largely bounced back after the short course but never recovered after the long one, which ended up dominated by Escherichia-Shigella.
• This is mouse work with no fecal transplant or germ-free arm, so the microbiome's causal role is inferred rather than proven.

Antibiotic course length flipped fat burning in opposite directions

A two-week antibiotic course and a ten-week course of the same drug cocktail produced opposite metabolic outcomes in obese mice, according to a study published in Frontiers in Microbiology in July 2026 by Ming Li's group at Sichuan University's West China School of Public Health. Every antibiotic-treated animal received the identical mixture — ampicillin, neomycin, bacitracin, vancomycin, imipenem/cilastatin and amphotericin B — by oral gavage. Only the exposure window differed.

The short-course animals started separating from the untreated high-fat controls at week 5, carrying less weight from that point on, along with a lower body fat percentage, lower fasting blood glucose and reduced triglycerides. The long-course animals moved the other way, weighing more than both the untreated high-fat group and the short-course group between weeks 5 and 10, with a higher body fat percentage, elevated fasting glucose and reduced HDL cholesterol.

The mechanism the researchers went looking for was thermogenesis (the process by which fat cells burn fuel to produce heat rather than storing it). In brown fat, the short course raised expression of Ucp1 (uncoupling protein 1, the protein that lets a fat cell dissipate energy as heat instead of banking it), along with Ppargc1a and Pparg, two genes that switch on the machinery. In white fat, the short course raised Ucp1 and Ppargc1a and increased mitochondrial DNA copy number, a signature of "browning" — white fat taking on the heat-generating character of brown fat. The long-course mice showed reduced expression of all of these compared with both other groups.

Function tracked the genes. After stimulation with a drug that maximally activates brown fat, the short-course mice showed a higher weight-adjusted metabolic rate than untreated high-fat controls, while the long-course group ran lower. A statistical analysis holding body weight constant found that body weight explained essentially none of this (p = 0.978) while group assignment explained a substantial share (p = 0.007).

Why this matters for anyone managing weight, blood sugar or repeat infections

The people this speaks to most directly are those who have taken long or repeated antibiotic courses — for recurrent urinary tract infections, acne, Lyme disease, sinusitis or dental prophylaxis — and who are also working on weight, insulin resistance or type 2 diabetes. The study's implication is that "antibiotics affect metabolism" is too blunt a statement. Direction of effect may depend on dose-duration, and the harmful direction is the prolonged one.

That fits an existing human literature. A birth cohort study of 1,016 children in the International Journal of Obesity found that infant antibiotic exposure was associated with 45% higher odds of childhood obesity, rising to nearly three-fold odds among children who received three or more courses — a dose-response pattern consistent with duration mattering. A review of early-life antibiotic exposure and obesity predisposition drawing on 17 epidemiological studies reached a similar conclusion and argued for judicious use specifically because the developing microbiota is vulnerable.

What is new here is the sign flip. Most prior work treats antibiotic exposure as uniformly obesogenic; this study says the damage is a function of how long the disruption is sustained.

One hundred mice, five groups, and a single changed variable

The design's strength is its restraint. The team used 100 male C57BL/6 mice, 10 weeks old, split into five groups of 20: standard chow with saline; standard chow with a two-week antibiotic course; high-fat diet with saline; high-fat diet with a two-week course; and high-fat diet with a ten-week course. The high-fat diet drew 60% of calories from fat. All antibiotic groups were dosed daily for the first two weeks; from weeks 3 to 10, only the long-course group continued, on an every-other-day schedule.

Crucially, caloric intake did not differ significantly among the three high-fat groups. The mice were not eating their way to different body compositions. Gut bacteria were profiled by 16S rRNA sequencing in 10 mice per group at week 2 and again at week 10.

The microbiome results are the cleanest part of the paper. At week 2, bacterial diversity had collapsed in every antibiotic group. By week 10, the short-course animals had recovered to the point of being statistically indistinguishable from untreated high-fat controls (p = 0.316), while the long-course group had the lowest diversity of any group and formed its own separate cluster, differing significantly from the chow controls, the antibiotic-only controls and the high-fat controls (p = 0.007, 0.007 and 0.008). Community-level tests confirmed the split (PERMANOVA p = 0.001; ANOSIM R = 0.7959). The long-course guts were markedly enriched for Escherichia-Shigella, a genus that includes opportunistic pathogens, along with Enterococcaceae and Erysipelatoclostridiaceae.

What this does and does not change about your next prescription

Nothing here argues for taking antibiotics to lose weight, and nothing here argues for stopping a prescribed course early. Cutting an antibiotic course short to protect your microbiome is a decision with real infection risk and belongs entirely to your prescriber. The honest takeaway is narrower: if you are facing an unusually long or open-ended antibiotic regimen, duration is a legitimate thing to raise with your doctor, alongside the question of whether a shorter evidence-based course would work.

It is also worth noticing what the short course did not do: it did not normalize these animals, which were still obese mice on a 60%-fat diet. The diet still drives the outcomes in this model.

No fecal transplant arm leaves the microbiome link inferred, not proven

The study's central weakness is that it never directly tested whether bacteria caused the metabolic changes. There was no fecal microbiota transplantation arm and no germ-free mice — the two experiments that would establish the microbiome as the mediator rather than a bystander. Antibiotics have documented direct effects on mitochondria and on the host independent of gut bacteria, and this design cannot rule those out. The authors are explicit that the responsible taxa and metabolites "remain to be identified."

Several other limits deserve naming. Only male mice were used. Metabolic rate was measured after pharmacological stimulation, so it reflects maximal thermogenic capacity rather than everyday resting expenditure. The sequencing was taxonomic only, without the functional metagenomics that would reveal which microbial pathways changed. And the group that would have been most informative — long-course antibiotics on normal chow — could not be run, because those mice had a survival rate below 60% by week 5.

Metagenomics and transplant experiments are the obvious next step

The follow-up this study demands is a transplant experiment: move the long-course microbiota into germ-free recipients and see whether the fat gain travels with the bacteria. Pairing that with functional metagenomics would identify which microbial products touch the thermogenic machinery. Until then, human relevance is a hypothesis rather than a finding.

Common Questions

Does taking antibiotics cause weight gain in humans? Observational human studies, particularly in infants and children, link antibiotic exposure to modestly higher obesity risk, with stronger associations after repeated courses. Those studies are correlational and cannot separate the antibiotics from the infections that prompted them.

Should I take probiotics during or after an antibiotic course? This study did not test probiotics and offers no evidence either way. Recovery of diversity in the short-course mice happened on its own, without any supplement.

How long does the gut microbiome take to recover after antibiotics? In this study, mice given a two-week course had largely recovered by week 10, while those on a ten-week course had not recovered at all. Human recovery timelines vary widely by drug and individual, and some species do not return.

Is Escherichia-Shigella dangerous? It is a genus that includes both harmless gut residents and opportunistic pathogens. Its expansion here is best read as a marker of a disrupted, low-diversity community rather than as an infection.

References

Tianle Feng, Yuanjie Cui, Xinwen Bi, et al. (2026). Duration-dependent gut microbiota disruption drives opposing effects on adipose thermogenesis in obese mice. Frontiers in Microbiology. https://doi.org/10.3389/fmicb.2026.1860759