Skip navigation
Link to Back issues listing | Back Issue Listing with content Index | Subject Index

Treatments for impetigo

Systematic review
Topical antibiotic vs placebo
Topical antibiotic vs oral antibiotic
Mupirocin vs fusidic acid

Impetigo is a common bacterial skin infection that affects about 1 in 35 of under-4s and 1 in 60 of under-15s each year. Treatment options vary. They include topical and oral antibiotics, antifungals, perhaps with steroids, and some treatments that are less conventional. The situation is complicated by diagnosis, extent of affected skin, and outcome. A new systematic review [1] shows us that there is limited clinical information, but does a great job of making sense of it.

Systematic review

The search strategy was sensitive and extensive, examining six databases, including the Cochrane Library, as well as evidence from other sources. The authors also asked pharmaceutical companies for any additional published or unpublished trials.

For inclusion studies had to be randomised double blind trials for bullous or non-bullous impetigo, irrespective of the extent of the disease. The outcome was cure or improvement of the condition within seven to 14 days. Patient selection by skin swab or other bacteriological testing was an exclusion criterion.


There were 16 studies included in the review, all published. Most of the studies were of high reporting quality, minimising possibility of bias. There were three main groups of comparisons: topical antibiotics versus placebo, topical antibiotic versus oral erythromycin, and a comparison of two different topical preparations. These 10 trials had 664 patients treated with topical antibiotic, 119 with placebo and 96 with oral antibiotic.

Topical antibiotic vs placebo

There were three trials (Figure 1, Table 1). Topical antibiotic produced 81% cures or improvements at 7-14 days, significantly more than with topical placebo (61%). The number needed to treat to produce an additional patient cured or improved at 7-14 days was 5 (95%CI 3.2 to 11).

Figure 1: Topical antibiotics vs placebo for impetigo

Table 1: Results of comparisons with topical antibiotics for impetigo

Improved/total (%)
Treatment 1
Treatment 2
Number of trials
Treatment 1
Treatment 2
Relative benefit
Topical antibiotic Placebo
1.2 (1.1 to 1.6)
5.0 (3.2 to 11)
Topical antibiotic Oral antibiotic
1.2 (1.01 to 1.3)
9 (4.4 to 194)
Topical mupirocin Topical fusidic acid
1.0 (0.98 to 1.09)
not calculated

Topical antibiotic vs oral antibiotic

Three trials compared topical antibiotic with oral erythromycin (Table 1). Topical antibiotic produced 87% cures or improvements at 7-14 days, significantly more than with oral erythromycin (75%). The number needed to treat to produce an additional patient cured or improved at 7-14 days was 9 (95%CI 4.4 to 194). This may not be a robust conclusion for two reasons. First the numbers are small, and different calculations can show a non-significant difference. Moreover, the authors omitted one small study that showed better results with another oral antibiotic.

Mupirocin vs fusidic acid

Four trials compared topical mupiricin with topical fusidic acid. Both had cure or improvement rates of over 90% (Table 1) and there was no difference between them.


Topical antibiotics are better than placebo at producing a cure or improvement at 7-14 days in patients with impetigo, and are probably at least as good as oral antibiotics. Rates were higher for topical antibiotics if all the arms of the trials are gathered together (Figure 2). Overall 91% of patients treated with topical antibiotics were improved or cured, 75% with oral antibiotics, and 61% with topical placebo.

Figure 2: Treatments arms compared

So what should we think of the information to hand? The authors make a sensible comment, that while trials were generally well reported, with adequate scores for reporting quality, the design of most of the included trials was less than adequate, with considerable clinical heterogeneity in outcomes used and description of the extent of impetigo at the start of treatment. Moreover, the 16 included trials came from 359 studies examined, almost all of which were excluded because of fundamental flaws like lack of randomisation or blinding.

Studies were also small, so though we have 16 studies, we have only a handful of patients on particular treatments. This means we cannot say whether topical antibiotics are better than oral antibiotics, or which of either is best. It makes us think about our lack of knowledge, and is a pointer for the type of study needed to be done within the NHS to provide clear clinical guidelines.

And it makes us think about the outcomes and NNTs. With topical placebo the improvement or cure rate was 60%. This suggests that the outcome used is not sensitive. To determine any real differences between treatments we need a much better outcome with a higher hurdle for efficacy. If 60% of people get better anyway, that leaves room for only 40% to be cured with treatment. If they were all so cured, the best NNT would be 100/40, or 2.5.

What this little review does is to do what all systematic reviews should do, not just dig around in the medical archaeology, but give us a vision of what we need to do to get better. Here it is simple: more, better, and bigger trials.


  1. A George, G Rubin. A systematic review and meta-analysis of treatments for impetigo. British Journal of General Practice 2003 53: 480-487.

previous or next story