Can an antibiotic you did not need raise your risk of dying?
Yes. In this study of 2,851 hospital patients with bloodstream infections, those who received broad antibiotics that kill gut bacteria, without any medical reason to need them, had about 41 percent higher odds of dying within 90 days (odds ratio 1.41). Their gut also lost a large share of its protective bacteria.
Most of the bacteria living in your intestines are anaerobes. That word simply means they live without oxygen. They make up the largest share of the gut microbiome, and they help digest food, train the immune system, and crowd out bacteria that cause disease. Some antibiotics kill them along with the target infection. Doctors often reach for those drugs as a precaution, even when the infection they are treating does not call for anaerobic coverage.
What the Data Show
Among the 2,851 patients studied, 2,106 of them, or 74 percent, received antibiotics with anti-anaerobic activity. Only 745, or 26 percent, did not. After the researchers used statistical weighting to make the two groups comparable on measured factors like age and illness severity, the patients who got these antibiotics had about 41 percent higher odds of dying within 90 days (odds ratio 1.41). The range around that estimate makes it very likely the true increase falls somewhere between 3 percent and 92 percent higher (95% CI 1.03 to 1.92). There was about a 3 in 100 chance of seeing a result like this by coincidence alone (p = 0.03).
The gut findings pointed the same direction. Metagenomic sequencing, which reads the DNA of every microbe in a stool sample, showed that anti-anaerobic antibiotics were linked to a drop of about 16.6 points in the share of the gut occupied by anaerobes (fixed effect estimate -16.59). That decline was very likely somewhere between 2.5 and 30.7 points (95% CI -30.67 to -2.52), with roughly a 2 in 100 chance of being a fluke (p = 0.02). The longer patients stayed on these drugs, the higher their death risk climbed and the more gut anaerobes they lost.
Dr. Kumar’s Take
Seventy-four percent is a remarkable figure. Three out of four patients here got a drug aimed at bacteria they had no reason to be treating. In the hospital, broad coverage feels like the safe choice. This study says that instinct has a price, and the price shows up as both a thinner microbiome and more deaths at 90 days.
I want to be careful about how far I push this. Patients were randomized in BALANCE to a duration of therapy, not to which antibiotic they received. That choice was made by their doctors, which means sicker patients may have been steered toward the broader drug in ways no statistical adjustment fully erases. I read this as a strong signal that deserves its own trial, not as proof that the antibiotic itself killed anyone.
What makes it more than a statistical curiosity is the dose-response pattern. More days of exposure meant more damage to the gut and more risk of death. That is the kind of pattern you would expect if the drug were truly causing harm, and it is much harder to explain away as a quirk of who got prescribed what.
How the study was done
This was a planned sub-study of BALANCE, a multisite randomized controlled trial testing how long antibiotics should be given for bloodstream infection. The researchers looked only at patients who had no clinical reason to need anaerobic coverage and who survived at least seven days past their index blood culture. The exposure they tracked was receipt of any anti-anaerobic antibiotic from three days before that culture through seven days after it. They then compared 90-day death rates and sequenced stool samples to measure what happened to the gut community.
Limits worth knowing
The comparison here is observational. Doctors chose the antibiotics, and the reasons behind those choices are not fully captured in a dataset. Inverse probability of treatment weighting can balance the factors that were measured, but nothing balances the factors that were not. Restricting the analysis to patients who lived seven days also removes the earliest and often sickest deaths from view. The authors themselves stop at association and call for trials that deliberately test narrower antibiotic strategies.
Practical Takeaways
- If you or a family member is started on antibiotics in the hospital, it is fair to ask the team whether the drug chosen covers anaerobic bacteria and whether that coverage is actually needed for this infection.
- Ask whether the antibiotic can be narrowed once blood culture results come back, since this study found that every additional day of anti-anaerobic exposure was linked to more gut damage and more risk.
- Never stop or change a prescribed antibiotic on your own, because an undertreated bloodstream infection is far more dangerous than the risk described here.
- If you have been through a long hospital course of broad antibiotics, mention persistent digestive symptoms to your doctor rather than assuming they will settle on their own.
Related Studies and Research
- Does a probiotic actually protect your gut bacteria from antibiotics?
- Inside the colonic bioreactor: how S. boulardii preserves gut anatomy during antibiotics
- Can preventing strep help kids with PANDAS? A closer look at prophylactic antibiotics
- Wim Hof method for depression: randomized clinical trial
FAQs
Why did nearly three out of four patients get an antibiotic they did not need?
Broad-spectrum antibiotics are often started before anyone knows which bacteria are causing an infection. Covering more possibilities feels like insurance, and in a patient with a bloodstream infection the stakes are high enough that most clinicians err wide. The problem is that the wide choice frequently never gets narrowed once the culture results arrive. This study defined its group specifically as patients with no clinical indication for anaerobic coverage, which means the extra coverage was doing no work against their actual infection.
Does this mean broad-spectrum antibiotics are dangerous?
Not in general. These drugs save lives every day in infections where the responsible bacteria are unknown or where anaerobes are truly involved. The finding here is narrower and more specific: when there is no reason to target anaerobes, using a drug that kills them anyway was linked to worse survival. The useful takeaway is about matching the drug to the infection, not about avoiding broad antibiotics as a class.
Does a longer course of antibiotics cause more harm?
In this analysis, yes. Increased duration of anti-anaerobic antibiotics was associated with both greater mortality risk and additional loss of gut anaerobes, so the effect scaled with exposure rather than switching on after a single dose. This matters because duration is one of the easiest things to change in practice. It is also the exact question the parent BALANCE trial was built to study, which is part of why these patients were being tracked so closely in the first place.
Bottom Line
In 2,851 patients with bloodstream infections who had no reason to need anaerobic coverage, 74 percent got it anyway, and those patients had about 41 percent higher odds of dying within 90 days along with a measurable loss of protective gut bacteria. The damage grew with every additional day of exposure. This is association rather than proof, but it is a well-designed look at a habit that affects millions of hospital patients, and it argues that the broadest antibiotic is not automatically the safest one.
