Caveats for the Use of Humerus Length in the Prediction of Fetal Down Syndrome

Abstract

Objective: To assess the reliability and reproducibility of fetal humerus length in the diagnosis of trisomy 21. Methods: Cohort study inclusive of 22 trisomy 21 fetuses, who underwent ultrasonographic examination between 14 and 22 weeks’ gestation, and 457 euploid controls. Regression analysis was performed for humerus length as function of biparietal diameter. Based on the generated regression equation in euploid fetuses, expected values of humerus length for a given biparietal diameter were calculated. The ratios of observed to expected (O/E) humerus length values were compared between euploid and trisomy 21 fetuses using Student’s t test. Receiver operating characteristic (ROC) curve analysis was used to detect optimal thresholds of O/E humerus length for diagnosis of trisomy 21. In addition, a MEDLINE search was conducted for articles published on humerus length as predictor of trisomy 21. Results: No differences were present between the regression lines of trisomy 21 and euploid fetuses (mean ± standard deviation O/E humerus length in euploid and aneuploid fetuses: 1.00 ± 0.10 vs. 0.97 ± 0.11, p = 0.21). The optimal threshold O/E humerus length <0.88 identified by ROC curve analysis had a sensitivity of 18% and a false-positive rate of 9% for the diagnosis of trisomy 21. From a review of the evidence provided by the 17 published series on humerus length as predictor of Down syndrome, the following caveats emerge: (1) with a median false-positive rate of 5% (range 1–12%), the median sensitivity of humerus length was only 28% (range 15–64%); (2) differences were present among centers in the regression lines of euploid fetuses and in the optimal diagnostic thresholds of humerus length, suggesting inter-center variability, and (3) most populations studied were at high genetic risk for trisomy 21, hence the diagnostic ability of humerus length in low risk populations has not been tested. Conclusions: The ability of humerus length to predict trisomy 21 is inconsistent. Only institutions with locally generated regression equations and documented predictive ability of this marker should utilize humerus length as a screening test for trisomy 21, alone or incorporated into diagnostic algorithms with serum or other sonographic markers of trisomy 21. The diagnostic ability of humerus length in low risk populations is currently unknown.