Diagnostic Sensitivity and Specificity: 21-Deoxycortisol vs 17-Hydroxyprogesterone in Newborn Screening for 21-Hydroxylase Deficiency
In a second-tier LC-MS/MS newborn screening study encompassing 924 non-CAH infants and 17 confirmed 21-hydroxylase deficiency cases, 21-deoxycortisol (21DF) demonstrated perfect discriminatory power with an area under the ROC curve (AUC) of 1.0. At a cut-off of 7.9 nmol/L blood, 21DF achieved 100% sensitivity and 99% specificity. By comparison, 17-hydroxyprogesterone (17OHP) at its optimal cut-off (<20 nmol/L) yielded only 78% sensitivity and 98% specificity [1]. Two equivocal cases by 17OHP criteria were unequivocally identified by 21DF, confirming its superior diagnostic utility.
| Evidence Dimension | Diagnostic sensitivity and specificity for classical 21-hydroxylase deficiency |
|---|---|
| Target Compound Data | 21-deoxycortisol (corticosterone): AUC = 1.0; Sensitivity = 100%; Specificity = 99% (cut-off 7.9 nmol/L blood) |
| Comparator Or Baseline | 17-hydroxyprogesterone (17OHP): Sensitivity = 78%; Specificity = 98% (cut-off <20 nmol/L blood); AUC <1.0 |
| Quantified Difference | Sensitivity: +22 percentage points; Specificity: +1 percentage point; AUC: 1.0 vs lower (value not reported but inferred from sensitivity/specificity deficit) |
| Conditions | Second-tier LC-MS/MS using dried blood spots (one 3.2 mm punch) from newborns (median gestational age 37 weeks, range 22-43 weeks); non-parametric Kruskal-Wallis test and Dunn's post hoc test for steroid concentration comparisons |
Why This Matters
Procuring high-purity 21-deoxycortisol reference material eliminates false-negative CAH screening results and reduces false-positive referrals, directly impacting clinical decision-making and healthcare resource allocation.
- [1] Greaves RF, Kumar M, Mawad N, et al. Best Practice for Identification of Classical 21-Hydroxylase Deficiency Should Include 21 Deoxycortisol Analysis with Appropriate Isomeric Steroid Separation. Int J Neonatal Screen. 2023;9(4):58. doi:10.3390/ijns9040058. View Source
