SummaryThe mechanism by which Propionibacterium acnes induces inflammatory lesions in acne vulgaris is still unknown. To investigate the effect of erythromycin stearate in acne vulgaris, the composition of the skin surface lipids was analyzed before, during and after treatment in 17 patients given 500 mg daily for 1 week and 250 mg daily for 2 months. The clinical anti-inflammatory effect was also evaluated. Quantitative thin-layer chromatography showed that the fatty acid fraction of the skin surface lipids decreased significantly in all patients during treatment. The clinical anti-inflammatory effect was good in 14 of the 17 patients. ATR-infrared spectroscopy showed that the clinical effect precedes the detectable reduction in fatty acids in the skin surface lipids. Even if a toxic effect of the fatty acids is responsible for the inflammation in most patients, an allergy to the fatty acids in some cases cannot be excluded. Sensitization tests on guinea-pigs showed that short-chained and middle-chained fatty acids, especially, are highly allergenic. The effect of erythromycin stearate on the lipase activity was also studied in a model system (purified pancreatic lipase). The enzyme activity was not affected by exposure to the drug. Consequently, the effect of erythromycin stearate on P. acnes was evaluated in acne patients, the bacteria being identified by aid of gas-chromatography. All the strains isolated were highly sensitive to erythromycin stearate before and after treatment. Semi-quantitative screening of P. acnes isolates on Tween 80 and tributyrine agar dishes after erythromycin stearate therapy indicated that P. acnes strains lacked lipolytic activity. However, in none of the patients' pustules or cysts was P. acnes completely inhibited. Sub-culturing demonstrated no significant quantitative decrease in the numbers of P. acnes after 1 week of therapy but did so after 4 weeks. Erythromycin stearate, therefore, may decrease the viability and numbers of P. acnes.