What Should You Eat While Taking Antibiotics?
The two moves worth making at the dinner table while you're on antibiotics are to cut the added sugar and to lean on fiber-rich plants and fermented foods. A new Nature study makes the sugar part specific: in hospitalized blood-cancer patients on antibiotics, more sugar in the two days before a dose was tied to a further drop in gut bacterial diversity on top of what the drug itself did. The food that feeds your beneficial bacteria does the opposite. These are small changes, but the balance of upside to effort makes them one of the few useful things you can do at home while a course runs.
If you're on antibiotics this week, the useful rule at the dinner table is simple. Cut the added sugar and keep eating plants. The reason that advice turned specific is a Nature paper published September 30, 2026. In an observational study of 173 adults hospitalized for blood-cancer treatment, each extra 100 grams of sugar in the 48 hours before an antibiotic dose was tied to roughly a further 21% drop in gut bacterial diversity on top of the antibiotic's own damage. The other half of the move, fiber and fermented foods, is where the recovery evidence has quietly been building.
What the new study actually measured
The paper came out of groups at NYU Langone, City of Hope, and Memorial Sloan Kettering. They tracked 9,419 meals and daily stool samples across adults going through blood-cancer treatment, most of them around stem-cell or bone-marrow transplant. Then they matched each antibiotic dose to what the patient had eaten in the previous 48 hours. More sugar before a dose predicted less gut diversity after it. The size of the effect worked out to about a 21% further diversity drop for every extra 100 grams of sugar, a large exposure, more than most Americans eat in a typical day.
That's an observational signal in a sick, hospitalized group, not a randomized trial in healthy outpatients. The same team ran a mouse arm to check the mechanism. Giving mice the antibiotic biapenem alongside sucrose let the pathobiont Enterococcus bloom about 16-fold by day 3 and 33-fold by day 6, compared with antibiotic alone. One of those Enterococcus species, E. faecium, is a leading cause of bloodstream infection in exactly this patient group. The human association and the mouse mechanism point the same way, with the caveat that whether the sugar signal holds in healthier outpatients taking a short course of amoxicillin has not been tested.
Why sugar lands harder when the drugs are working
Antibiotics do two things that matter here. They thin out the butyrate-producing anaerobes that feed your colon lining. And they leave simple sugars pooling in the colon that would normally have been absorbed higher up. Those leftover sugars become fuel for the species you don't want after a course: Enterococcus, Enterobacteriaceae, sometimes C. difficile. Add a soda or a slice of cake and you're pouring more fuel on.
Meanwhile, the beneficial bacteria that usually crowd out those opportunists are being starved of their preferred food, dietary fiber. When they run out of fiber, some switch to eating the gut's own mucus layer. That thins the barrier between you and your microbes, which means more permeability, more inflammation, and more open space for pathogens to attach.
What to eat instead
Fiber is the dependable one. Vegetables, legumes, whole grains, oats. In animal and small human studies, fiber-containing diets after antibiotic exposure track with faster return of beneficial species. The absence of fiber does the opposite: it prolongs the dysbiosis and keeps the mucus barrier thin.
Fermented foods are a stronger lever than they used to look. A 17-week Stanford trial in healthy adults showed that stepping fermented food intake up to about six servings a day, yogurt, kefir, kimchi, sauerkraut, cottage cheese, raised microbial diversity in a way an isolated fiber bump didn't, and dropped several inflammatory markers along with it. The sample was small and the participants were not on antibiotics, so the effect in a course of amoxicillin is extrapolated, not proven.
Probiotic pills are a more specific and more modest tool, and the right strain matters. Meta-analyses show that Saccharomyces boulardii and Lactobacillus rhamnosus GG, started within 48 hours of the first antibiotic dose, cut antibiotic-associated diarrhea by roughly a third to a half. For C. difficile diarrhea specifically, the most recent Cochrane review puts the relative reduction at about half, with a small absolute reduction of around 1.6%, and benefit concentrated where baseline risk is high. In that same Cochrane subgroup, only S. boulardii showed a significant effect on C. difficile; L. rhamnosus GG did not. One real caveat: probiotics reduce symptoms. They don't reliably rebuild overall gut diversity. Food carries most of that weight.
A safety note that gets lost in headlines about the Nature cohort: the same transplant and blood-cancer patients the study enrolled are the group in whom probiotics are explicitly not recommended. Current guidance from the American Society for Transplantation and Cellular Therapy and from the European Conference on Infections in Leukaemia advises against probiotic supplements in neutropenic, significantly immunocompromised, or central-line patients because of the risk of bloodstream infections, including Saccharomyces fungemia. Probiotic advice below is for immunocompetent outpatients only.
| Food or supplement | What the evidence shows | Practical move during a course |
|---|---|---|
| Added sugar and sugary drinks | More sugar in the 48 hours before a dose linked to further loss of gut diversity in hospitalized patients; feeds opportunists in mice | Cut sodas, sweetened coffee, candy, and desserts, especially in the hours around each dose |
| Fiber-rich plants (vegetables, legumes, whole grains, oats) | Feeds butyrate producers; animal and small human studies suggest faster post-antibiotic recovery | Make most meals plant-forward and vary your fiber sources |
| Fermented foods (yogurt, kefir, kimchi, sauerkraut) | 17-week trial in healthy adults raised diversity and lowered inflammatory markers at about six servings per day | Work up to multiple servings a day if you tolerate it |
| Probiotics (immunocompetent outpatients only) | S. boulardii and L. rhamnosus GG cut antibiotic-associated diarrhea by roughly a third to a half; for C. difficile, only S. boulardii has a proven effect | If you're otherwise healthy and on a broad-spectrum drug, start S. boulardii within 48 hours; avoid probiotics if you are neutropenic, immunocompromised, or have a central line |
| Ultra-processed, low-fiber foods | Little substrate for beneficial microbes; often high in added sugar | Default away from them for the course and about two weeks after |
Who benefits most, and what's still unknown
The sugar signal was measured where the gut was under maximum pressure: chemo, transplant, broad-spectrum intravenous antibiotics. If sugar makes things worse there, it probably matters for someone taking a week of amoxicillin for a sinus infection, but nobody has tested that directly yet. The gap worth watching is a randomized trial of a low-sugar, fiber-forward diet started with antibiotic initiation in outpatients, with endpoints that matter: diarrhea, C. difficile infection, time to microbiome recovery.
Even without that trial, the trade is lopsided. Cutting soda and dessert for two weeks costs almost nothing. Vegetables, beans, oats, and yogurt during a course cost little. If you are an immunocompetent outpatient on a broad-spectrum drug, starting S. boulardii inside the first 48 hours has the strongest supplement evidence available for preventing antibiotic-associated diarrhea. If you are hospitalized, neutropenic, or have a central line, skip the probiotic and lean harder on the dietary side instead; the fiber and fermented-food move still applies and current guidance for these patients favors prebiotic fiber over live-organism supplements.
Keep taking the antibiotic. The drug is doing what it's meant to do, and stopping early breeds resistance. The dietary moves are there to give the gut better substrate to rebuild on while the course finishes its job. That is a smaller claim than curing antibiotic side effects, but it's the specific, measurable one the new Nature data support.