Three methods of measuring the epidermolytic toxin Staphylococcus aureus-bioassay in newborn mice, radial immunodiffusion, and radioimmunoassay-were compared for reproducibility, specificity, and sensitivity. The bioassay is highly specific and remains the only functional assay. It is reproducible only if newborn mice of the same age are used. The time required for epidermolysis follows a dose-response relationship only if concentrations of toxin large enough to cause peeling in 90 min or less are used. This limits the sensitivity of the bioassay to about 5 mug per ml. Single radial immunodiffusion in agar is a specific and reproducible assay method, but its sensitivity is also about 5 mug per ml. A radioimmunoassay was established by the Farr technique using purified epidermolysin radiolabeled with 125iodine. This assay was highly reproducible and specific. The staphylococcal products, alpha-toxin and enterotoxins A and B, did not cross-react with anti-epidermolysin antibodies. The sensitivity of the radioimmunoassay is 20 ng per ml.

Radioiodinated staphylococcal epidermolytic toxin was found not to bind to erythrocytes, blood leukocytes, trypsin-dispersed keratinocytes, epidermis or whole skin. Moreover the toxin could not be found to bind to murine epithelia by indirect immunofluorescence methods. However, the toxin, measured by radioimmunoassay, could be eluted from the skin of mice undergoing epidermolysis following intraperitoneal injection of toxinogenic Staphylococcus aureus. Furthermore, epidemolysin was measured in the blister fluid of 3 of 5 children with bullous impetigo but not in blister fluid from control patients with other blistering eruptions. Thus epidermolysin has been demonstrated to be present in lesions of the staphylococcal epidermolytic toxin syndrome but its mechanism of action does not involve binding to cells.

Summary Measurement by radioimmunoassay of the epidermolytic toxin of Staphylococcus aureus and antibodies to epidermolysin was attempted in sera from patients with the staphylococcal scalded skin syndrome. Antibodies to epidermolysin were also measured in sera from normal adults and patients with drug‐related toxic epidermal necrolysis. Epidermolysin was not detectable by radioimmunoassay (in the range of 5‐100 ng/ml), radial immunodiffusion or bioassay in newborn mice. Antibodies to epidermolysin were detectable in sera from patients in all three groups studied, but there was a significant increase in the amount of antibody in convalescent, compared to acute phase sera, only in the patients with the staphylococcal scalded skin syndrome. We propose that neutralizing antibodies play a major role in termination or prevention of the staphylococcal scalded skin syndrome.