Skip navigation

Dog bites Man

A classic headline, and an event that is common enough. But what happens next? Obviously treatment for the immediate effects of the bite are needed, but what of the possibility of infection? Is there a role for prophylactic antibiotics, and are they effective if used?


A meta-analysis of randomised controlled trials of the use of prophylactic antibiotics found eight RCTs [1], six in the USA and two in the UK. All the studies used oral antibiotics, three using penicillin and three pencillinase-resistant penicillins. The overall occurrence of infection rates ranged from 3.2% to 45.8% (the latter in Middlesborough, which had much the highest rate of infection; in this study 93% of patients had dog bites, the remainder being bitten by other humans, cats, a ferret and a rabbit). The cumulative incidence of infection was 16%.

Limited effect of prophylactic antibiotics

There were 323 patients who received placebo, of whom 52 became infected. There were 306 patients who received prophylactic antibiotics, of whom 30 became infected.

The relative risk of infection with antibiotics was 0.56 (95% confidence intervals 0.4 - 0.8), so antibiotics significantly reduced the risk of infection. It was interesting that while the meta-analysis demonstrated this clearly, only one of the eight trials individually showed results for antibiotics to be statistically superior to placebo.


The NNT for antibiotic prophylaxis to prevent infection after dog bites was 16 (95% CI 9 - 92). If 100 patients with dog bite wounds seen in casualty are given oral antibiotics, on average 84 patients would escape infection regardless of therapy, nine will become infected despite medication, and seven will avoid infection because of the medication.

Is it worth treating 100 patients with antibiotics to prevent infection in seven? Cost benefit analysis is missing, so the answer is left to the reader. In the meantime, try to avoid being bitten in Middlesborough!


  1. P Cummings. Antibiotics to prevent infection in patients with dog bite wounds: a meta-analysis of randomized trials. Annals of Emergency Medicine 1994 23: 535-40.

previous or next story in this issue