Reliability of aneuploidy estimates in human sperm: results of fluorescence in situ hybridization studies using two different scoring criteria.

Aneuploidy estimates for chromosomes 1, 12, X, and Y were obtained in human sperm from five donors using multicolor fluorescence in situ hybridization (FISH) analysis. Disomy frequencies were obtained by scoring a minimum of 10,000 sperm for each chromosome probe per donor. This analysis was replicated for two scoring criteria: one used one half of a signal domain as the minimum distance between two signals to be counted as two and thus disomic; the other set one signal domain as the minimum distance between two signals. A total of 120,870 sperm were assessed using one half of a domain as the criterion, and 113,478 were scored using one domain as the criterion. The percentage of disomy for chromosomes 1, 12, X, Y, and XY was 0.18, 0.16, 0.15, 0.19, and 0.25, respectively, using the one-half-domain criterion, and 0.08, 0.17, 0.07, 0.12, and 0.16, respectively, using the one-domain criterion. The percentage of disomy decreased significantly with use of one domain as the minimum distance for signal separation for all chromosomes except for chromosome number 12. These lower disomy frequencies correlated well with frequencies derived from human sperm karyotypes analyzed in our laboratory. This suggests that the fluorescent signals for chromosomes 1, X, and Y split into more than one domain in decondensed interphase sperm, and that the use of the one-half-domain criterion would lead to an overestimate of aneuploidy frequencies. The factors known to affect aneuploidy estimates derived from FISH studies are discussed, and recommendations for stringent scoring criteria are proposed.