Enhancing Accuracy of Newborn Screening for 21OHD: Strategic Use of 21-Deoxycortisol in a Large-Scale Tokyo Cohort.

Abstract

Newborn screening (NBS) for 21-hydroxylase deficiency (21OHD) has historically shown high false-positive (FP) rates, especially in low-birth-weight (LBW) infants. In 2022, we proposed a second-tier liquid chromatography-tandem mass spectrometry (LC-MS/MS)-based algorithm to improve screening specificity; however, its real-world performance remains unassessed prospectively. This work aimed to prospectively evaluate our LC-MS/MS-based screening algorithm for 21OHD and develop a refined version addressing newly identified clinical limitations. We prospectively analyzed 326 006 newborns screened in Tokyo from 2021 to 2025. Based on the findings, a revised algorithm was retrospectively validated in 946 246 newborns screened between 2015 and 2025. The original algorithm showed 100% sensitivity but lower-than-expected positive predictive values (PPVs) (9.0% for initial positives; 44.8% for final referrals). This was mainly due to FPs in LBW infants, who made up 75.7% of initial positives despite being unaffected. Steroid profiling showed elevated 21-deoxycortisol (21DOC) in all true positives but undetectable levels in 99.2% of LBW FPs. A refined algorithm prioritized a low 21DOC cutoff as the primary exclusion criterion, with the 11-deoxycortisol to 17α-hydroxyprogesterone (11DOC/17OHP) ratio as a secondary safeguard. This revision reduced LBW-related recalls by more than 99% and raised PPVs to 87.2% for initial positives and 89.1% for final referrals. This large-scale prospective study showed that a refined LC-MS/MS algorithm using 21DOC and 11DOC/17OHP greatly improved 21OHD screening specificity, eliminating LBW-related FPs and enhancing NBS efficiency.