RESUMO
INTRODUCTION: The primary concern related to congenital adrenal hyperplasia (CAH) newborn screening (NBS) is the high rate of false-positive results (FPR) associated with prematurity; false-negative results (FNR) can also occur due to precocious sample collection. OBJECTIVE: To determine the neonatal 17-hydroxyprogesterone (N17OHP) normal range in newborns in Sao Paulo using different references according to age and birthweight (BW) and to establish the optimal NBS cut-off levels. METHODS: Neonatal 17-hydroxyprogesterone levels from 271 810 newborns (NBs) according to sample collection time (G1: 48-<72 h and G2: ≥72 h) and BW (≤1500 g, 1501-2000 g, 2001-2500 and >2500 g) were evaluated. N17OHP was measured by an fluoroimmunoassay, and serum 17OHP was measured by liquid chromatography-mass spectrometry. Affected and asymptomatic NBs with persistently increased 17OHP levels were submitted to CYP21A2-sequencing. RESULTS: Neonatal 17-hydroxyprogesterone levels in G1 were lower than G2 in all BW groups (P < 0·001). The FPR rate in G1/G2 was 0·2% using the 99·8th and 0·5% using the 99·5th percentile. The 99·8th percentile N17OHP value was the best cut-off for distinguishing between unaffected and affected NBs. Forty-four salt wasters, and five simple virilisers were diagnosed; N17OHP levels ranged from 93·3 to 2209·8 nmol/l, and no affected neonates with FNR were identified. The positive predictive value in G1 and G2 using the 99·8th percentile was 5·6% and 14·1%, respectively, and 2·3% and 7%, respectively, using the 99·5th percentile. Molecular tests identified two NBs with the nonclassical form among the 29 FPR. CONCLUSION: Neonatal 17-hydroxyprogesterone levels adjusted to sample collection age and birthweight reduced the FPR, and the use of N17OHP values based upon the 99·8th percentile improved the NBS efficacy.