Eighty-five persons at risk for Huntington disease (HD) have enrolled in a predictive-testing pilot program. Informativeness of the test has been determined for 41 of these candidates by using linked DNA probes. Nine (21.9%) of these persons have been excluded from the test as a result of the unavailability of DNA from crucial family relatives. Homozygosity for all of the three DNA markers (D4S10, D4S62, and D4S95) was not found in any affected parent. Only one (2%) of the 41 test candidates has had an uninformative result. Results have been given to 20 persons, of whom 12 (60%) received a decreased risk and eight (40%) received an increased risk of having inherited the mutant gene for HD. The combined use of three DNA markers significantly increases the informativeness of family structures such that some change in the estimation of genetic risk is now possible for approximately 75% of all persons who request predictive testing.