Objective To assess the implications of a change in prenatal diagnosis policy from full karyotype analysis to rapid trisomy testing for women referred primarily for increased risk of Down's Syndrome. Design Retrospective collection and review of data.Setting The four London Regional Genetics Centres.Population Pregnant women (32,674) in the London area having invasive prenatal diagnosis during a six-year three-month period. Methods Abnormal karyotypes and total number of samples referred for raised maternal age, raised risk of Down's Syndrome following serum screening or maternal anxiety were collected. Abnormal karyotypes detected by molecular trisomy detection were removed, leaving cases with residual abnormal karyotypes. These were assessed for their clinical significance. Pregnancy outcomes were ascertained by reviewing patient notes or by contacting obstetricians or general practioners. Main outcome measures Proportion of prenatal samples with abnormal karyotypes that would not have been detected by rapid trisomy testing, and the outcome of those pregnancies with abnormal karyotypes. Results Results from 32,674 samples were identified, of which 24,891 (76.2%) were from women referred primarily for Down's Syndrome testing. There were 118/24,891 (0.47%) abnormal sex chromosome karyotypes. Of the samples with autosomal abnormalities that would not be detected by rapid trisomy testing, 153/24,891 (0.61%) were in pregnancies referred primarily for Down's Syndrome testing. Of these, 98 (0.39%) had a good prognosis (46/98 liveborn, 3/98 terminations, 1/98 intrauterine death, 1/98 miscarriage, 47/98 not ascertained); 37 (0.15%) had an uncertain prognosis (20/37 liveborn, 5/37 terminations; 12/37 not ascertained) and 18 (0.07%) had a poor prognosis (1/18 liveborn, 2/18 miscarriage, 11/18 terminations, 4/18 not ascertained). Conclusions For pregnant women with a raised risk of Down's Syndrome, a change of policy from full karyotype analysis to rapid trisomy testing would result in the failure to detect chromosome abnormalities likely to have serious clinical significance in approximately 0.06% (1 in 1659) cases. However, it should be noted that this figure may be higher (up to 0.12%; 1 in 833) if there were fetal abnormalities in some of the pregnancies in the uncertain prognosis group for which outcome information was not available.