Editorial NexoScience
What is the safer thing to do with a fresh dog bite: close it with stitches straight away, or leave it open and let it drain? Emergency medicine has leaned toward leaving it open for decades, on the reasoning that a mouth is a dirty place. A new pooled analysis of the published evidence suggests that reflex may not be buying the protection it promises.
The analysis, published June 4 in the journal World Journal of Emergency Surgery, gathers 26 studies of wounds inflicted by mammals, most of them dog bites. Of those, 17 studies covering 8,091 patients compared primary closure, stitching a wound shut at the first visit, against delayed closure or leaving it open altogether. The rest tackled a second question, whether preventive antibiotics stop wounds from becoming infected.
The old rule came from a fear of what lives in a mouth. Bites from mammals are among the most common injuries that walk into an emergency room anywhere in the world, and dogs alone bite about as many people every year as there are inhabitants of Egypt. Teeth drive bacteria deep into torn tissue, so surgeons either left the wound open or waited several days for the risk of infection to pass before suturing. The habit was never settled by good trials, which is why guidelines still contradict each other.
Against delayed stitching, closing the wound at once came out ahead. Wounds sutured on the first visit had roughly half the odds of becoming infected or healing badly compared with wounds stitched days later (odds ratio 0.49, confidence interval running from 0.27 to 0.90, sitting entirely below the line of no effect). That advantage held whether or not the patients also got antibiotics, and it was larger for wounds on the limbs and trunk.
Pooled across every comparison, the advantage washes out. Measured against delayed closure and no closure lumped together, primary closure produced an odds ratio of 0.79, with a confidence interval from 0.54 to 1.17 that comfortably includes the possibility of no difference at all. Set specifically against leaving the wound open for good, immediate stitches showed no clear benefit either, with two exceptions the authors flagged: wounds on limbs or trunk, and patients who reached a doctor within hours rather than the next day.
Appearance favored immediate closure, with one exception. Only three of the pooled studies scored how the healed wound looked. In the two where doctors had not used vacuum sealing drainage, a suction dressing that pulls fluid out of a wound while it waits to be closed, patients whose bites were stitched at once were markedly more satisfied with the result. In the third, which did use that technique, the delayed group ended up looking better. One included study also found raised, thickened scars several times more common when closure had been put off.
Preventive antibiotics did not earn their place across the board. Pooled across the studies that tested them, antibiotics prescribed to head off infection made no measurable difference. Two situations were the exception: bites from mammals other than dogs, and wounds on the face or head, where the drugs were tied to a substantial drop in infections. For a routine dog bite on an arm or a leg, properly washed out and debrided, the pooled data offer no support for reaching for a prescription.
The authors were direct about how far this can be pushed. Almost every significant result came out of subgroup analyses, which the researchers ran to explain the disagreement between studies rather than as the question the review set out to answer. About half of the pooled material is observational rather than randomized, so a surgeon’s decision to stitch or wait may itself reflect how clean, how deep or how recent the wound was, and the association cannot be read as proof that closing the wound caused the better outcome. The authors acknowledged that this heterogeneity is the main source of bias in the review, and that factors they could not test (wound length and depth, suture material, drainage, which antibiotic and for how long) were reported too rarely or in forms too different to control for. They also noted that the cosmetic finding rests on patients’ own satisfaction ratings in three studies, a subjective measure. And because the included studies excluded people with diabetes or weakened immune systems, along with the most severe injuries, none of this speaks to those patients.
What the field still does not have is a trial built to settle the question. The researchers called for high-quality randomized trials that control the confounding factors at baseline and measure more than infection alone, including healing time, scarring, length of hospital stay and cost, which almost no published study has reported. They also pointed to a gap nobody has tried to fill: not one study has compared bites from domestic animals with bites from wild ones, even though it is among the first things a patient reports.
The study, “Primary closure and prophylactic antibiotics for treatment of traumatic wounds caused by mammals, a systematic review and meta-analysis,” by Meng Cui, Yiqing Jia, Zhaoyang Chen, Jie Qu, Zonghong Zhu, Yan Xu, Shuyuan Liu, Ruifeng Chen and Yi Shan, was published June 4, 2025, in the World Journal of Emergency Surgery. DOI: 10.1186/s13017-025-00619-1.
