We observed that a group of detergent enzyme workers with known exposure to the subtilisin enzyme, Alcalase (Novo Industries, Bagsvaerde, Denmark), exhibited percutaneous sensitivity to Savinase (Novo Industries), a microbial protease, to which there was no previous occupational exposure. This was attributed to either cross-reactivity between these enzymes or to foreign enzyme contaminants contained in the Savinase antigen. The aims of this study were to determine the range of concentrations eliciting percutaneous responses to Alcalase and to another enzyme, Rapidase (an alpha-amylase) (Gist Brocades, Belgie, Netherlands); to compare the sensitivity of RAST and skin prick testing; and to characterize the relationship between wheal size and antigen concentration. Prick testing was conducted over six log10 antigen dilutions of Alcalase and Rapidase in 30 workers with previous exposure and skin reactivity to enzymes (group 1) and compared to nonexposed control groups, which included 60 atopic subjects (group 2) and 30 nonatopic subjects (group 3). The RAST was performed with Alcalase and Rapidase antigens. The percutaneous threshold concentrations in group 1 subjects varied widely from 10(3) to 10(-3) micrograms of protein per milliliter. Of 19 group 1 workers with skin test reactivity to Alcalase, 84% had positive RAST results; 83% of 24 workers who were reactive to Rapidase had positive RAST results. It was concluded that skin prick testing is preferred over in vitro methods for longitudinal monitoring of human sensitization to workplace allergens. In addition, the data predicted that based on a known Alcalase level of 0.07% in Savinase, 26% of Alcalase-sensitized subjects could react to Savinase. An excellent correlation (r > 0.97) was found between log concentration of antigen and wheal size parameters, with the log diameter and log area performing equally as well (r > 0.98). Analysis of variance revealed that more than 60% of intragroup variation represented human variability in wheal parameters at each concentration tested, whereas at least 95% of intergroup variation was due to regression. The excellent correlations of both wheal diameter and area with antigen concentrations were attributed to the very small changes observed between test concentrations.