Prophylactic antibiotics for the prevention of cellulitis (erysipelas) of the leg. A commentary


  • Conflicts of interest
    None declared.


Erysipelas, also called superficial cellulitis in the American literature, is an acute superficial dermal–hypodermal infection that usually affects the leg and is commonly caused by streptococci.1,2 It may be severe, especially in the elderly with comorbidities, but, unlike necrotizing infections, is generally not life-threatening. Oral or intravenous antibiotics, the therapeutic keystone, e.g. mostly penicillin or macrolides/streptogramins, are prescribed for 10–20 days to out- or inpatients. However, the 2010 Cochrane review concluded that the antibiotic administration mode and duration warranted more thorough evaluation, especially in the community.3 Management should also include the control of erysipelas risk factors, mainly cutaneous barrier disruption and leg oedema.4 In fact, recurrence is the main specific complication of erysipelas; it occurs in 10–50% of the patients and apparently shares similar risk factors with the incident cases.5,6 What should we do to avoid such recurrent erysipelas, which may lead to major lymphoedema (with functional and aesthetic handicaps) and, subsequently, higher risk of recurrence and possibly more severe episodes?

This issue of BJD reports a randomized clinical trial comparing 6 months of penicillin V (250 mg b.i.d.) vs. placebo for patients who had been treated for cellulitis of the leg (first and recurrent episodes).7 The hazard ratio indicated that the antibiotic reduced the recurrence risk by 47%, with repeat episodes occurring in 12 of 60 (20%) penicillin V-group patients vs. 21 of 63 (33%) placebo recipients (absolute difference: 13·3%, 95% confidence interval −2·4 to 28·1). For this disease with little available data, the authors should be commended for their efforts in conducting this trial and clearly reporting its findings. However, their recruitment goals were not met and the trend favouring prophylactic antibiotics is difficult to interpret and extrapolate.

Notably, hospital recruitment of patients with erysipelas seems extremely difficult in western countries, as a majority of patients are seen by office-based general practitioners and most guidelines now recommend home-administered oral or intravenous treatments for convenience and cost considerations, at least for routine cases. Indeed, we are presently encountering the same difficulties in a French erysipelas trial ( Randomized noninferiority study comparing short course of oral amoxicillin to one of the usual treatments of lower leg erysipelas. NCT01059123). Therefore, the authors’ design and logistics recommendations for future studies should be followed.

In addition, although prophylactic antibiotics may be valuable, the effect-size reported was slight, for three main reasons. Firstly, as stipulated by the inclusion criteria, 80% of the population had not experienced a previous erysipelas episode, thereby making the probability of recurrence lower than for a recurrent erysipelas population and explaining the trial’s approximately 70% nonrecurrence rate under placebo. It would probably be better to identify patients at high risk of recurrence after a first episode rather than expose more than two-thirds of the naive population to unnecessary antibiotics for 6 months. The presence of severe oedema or lymphoedema could be a criterion, as incident erysipelas may be associated with subclinical lymphoedema.8 The results of the ongoing PATCH I trial on patients with at least two recurrences should be informative. Secondly, several prophylactic antibiotic regimens are available worldwide, e.g. oral phenoxymethylpenicillin, oral amoxicillin and intramuscular benzathine penicillin, and their efficacies, costs and compliance rates might not be similar. Thirdly, optimal risk-factor control is essential to increase the benefit of any antibiotic strategy for recurrent erysipelas, e.g. effective elastic compression stockings and/or antifungals (whether topical or oral, according to the extent and localization of dermatomycosis).9