Reduction of error rates in the microlymphocytotoxicity test.

The overall error rate of the microlymphocytotoxicity test has been reduced from 1.08% in 1971 to approximately 0.35%, based on comparisons of 882 replicate test pairs (different bleedings) and on family studies, with 0.11% attributable to “reading” errors. This reduction was attributed to automation for lymphocyte adding, better incubation temperature control, better technical preparation of cells and stronger, more monospecific sera. Three types of errors contribute to the overall error rate: random, sera-related (affected by strength and specificity), and lymphocyte-related (possible B-cell reactivity and donor/lymphocyte physiological condition). Strongly skewed error distributions, which seem to be lymphocyte-related, occurred in less than 4% of the test pairs.