An evaluation of the accuracy and precision of relative potency (RP) estimated by parallel-line skin test (PLST) was carried out in 171 subjects by 26 operators in 9 countries and 12 states. Each clinic tested 5 or more subjects with both histamine base (HB) 1.8 and 0.1 mg/ml using a common PLST protocol. Best-fit lines of sum of erythema (E) on log3 dose were calculated in each subject from intradermal titrations of each HB using 4 serial 3-fold dilutions in singlicate producing graded E bracketing 50 mm. RP was calculated within-subject from the horizontal distance between acceptable parallel regression lines (slope equal to or greater than 10 and not different at p=0.01, r equal to or greater than 0.85). The geometric mean RP=14.5 (nominal RP=18). The pooled within-lab standard deviation (SD) were combined as specified in Amer. Soc. for Testing & Materials Std. E-177 as a measure of between-lab precision (reproducibility). The 95% reproducibility limits (RL) of RP when 4 subjects are tested is 41-160% of the predicted RP. The D50 (3-fold dilution for E=50 mm) was calculated for each HB as an estimate of test sensitivity. Mean D50 of HB 1.8 is 4.46 (RL 5.14- 7.39 for 4 tests). Mean D50 of HB 0.1 is 3.98 (RL 2.86-4.76 for 4 tests). Each subject was also puncture (P) tested with each HB using a bifurcated needle. Mean PE 1.8 = 50 mm (RL 27-29 mm for 4 tests). Mean PE 0.1=10 mm (RL 1-59 mm for 4 tests). These data are applicable to establishing the proficiency of operators using this method as well as defining the accuracy, sensitivity, and precision of this method within and between clinics.
