Ascertainment bias in Turner syndrome: new insights from girls who were diagnosed incidentally in prenatal life.

Abstract

To evaluate differences in phenotype and other clinical features between patients who have Turner syndrome diagnosed incidentally (on the basis of a prenatal karyotype performed for reasons unrelated to suspicion of Turner syndrome, eg, advanced maternal age) or traditionally (on the basis of either a prenatal karyotype performed for abnormal ultrasound findings or a postnatal karyotype performed for clinical findings suggesting Turner syndrome). Analysis was performed on baseline data from 88 girls, aged 9 months to 4 years, who were randomized into a multicenter growth hormone intervention trial. Baseline information included a detailed medical history (especially of cardiac and renal anomalies), examination for presence of phenotypic features characteristic of Turner syndrome, length or height (depending on age) expressed as a standard deviation score, and weight standard deviation score. Patients were classified to either the "incidental" (N = 16) or the "traditional" (N = 72) diagnosis group as described above. The incidental group had significantly fewer total phenotypic features of Turner syndrome than the traditional group (3.6 +/- 2.9 vs 6.7 +/- 2.9). When subgrouped by karyotype, the proportion of patients who manifested phenotypic features was greatest in patients with a 45,X nonmosaic karyotype, lowest in the patients with 45,X/46,XX mosaicism, and intermediate in those with "other" karyotypes. Fewer congenital cardiac defects were observed in the incidental group (31%) compared with the traditional group (64%), but there was no difference between the groups in the prevalence of renal defects. Karyotype distribution differed significantly between the traditional and incidental groups: 74% versus 19% had the nonmosaic 45,X karyotype in the traditional group and 7% versus 56% had the mosaic 45,X/46,XX karyotype. Patients whose Turner syndrome was diagnosed incidentally had significantly fewer phenotypic features and cardiac defects, as well as a greater proportion of mosaic karyotypes, compared with patients whose Turner syndrome was diagnosed clinically. These results support the theory that significant ascertainment bias exists in our understanding Turner syndrome, with important implications for prenatal counseling.