Nonrandom distribution of genotypes among red cell indices.

Venous blood samples were obtained from 25,302 healthy adults in Kentucky, USA. The red cell indices measured on these samples were evaluated by multiple stepwise regression analysis to derive an algorithm capable of discriminating the 138 individuals within this population who had genotypes AA, AC, AS or AA beta-thalassemia. The simple discriminant MCV2 x MCH with a cut-off set at 1530 detected 137 out of 138 of the heterozygotes with a false positive rate in this population of 4.4%. Other discriminants tested produced fewer false positives but also missed a sufficient number of heterozygotes to be unacceptable for genetic counselling purposes.