Screening for 21-hydroxylase–deficient nonclassic adrenal hyperplasia among hyperandrogenic women: a prospective study

Objective: To prospectively establish the specificity, sensitivity, and positive predictive value (PPV) of a basal 17-hydroxyprogesterone (17-HP) level for the screening of 21-hydroxylase–deficient nonclassic adrenal hyperplasia (NCAH) among hyperandrogenic women. Design: Prospective observational trial. Setting: Tertiary care academic medical centers. Patient(s): Eight healthy controls, 20 patients with NCAH, and 284 consecutively seen patients with hyperandrogenism. Intervention(s): All controls and patients with NCAH, and select patients with hyperandrogenism, underwent acute ACTH (1-24) stimulation. Main Outcome Measure(s): Specificity was determined by measuring 17-HP every other day during one menstrual cycle in 8 healthy women with normal ovulation (107 samples). Sensitivity was determined by measuring 17-HP between 7 and 9 a.m. and 3 and 5 p.m. on the same day in 20 patients with genetically confirmed NCAH. The PPV was determined by prospectively measuring 17-HP in 284 consecutively seen hyperandrogenic women at their initial evaluation. The diagnosis of NCAH was established by a stimulated 17-HP level of >10 ng/mL. Result(s): Among controls, 17-HP levels of 4 ng/mL had sensitivities of 100%, 90%, and 90%, respectively, for the detection of the disorder when obtained in the morning, and sensitivities of 95%, 90%, and 85%, respectively, when obtained in the afternoon. Among the 284 consecutively seen hyperandrogenic women, the PPVs of the first and second 17-HP levels were 7.3% and 19% for a cutoff level of >2 ng/mL, 13% and 43% for a cutoff level of >3 ng/mL, and 33% and 40% for a cutoff level of >4 ng/mL, respectively. Conclusion(s): A basal 17-HP level is a useful screening tool for NCAH. A cutoff level of 4 ng/mL has maximum specificity and PPV, with little loss in sensitivity if testing is performed in the morning and during the follicular phase. However, a lower cutoff level (e.g., 2 or 3 ng/mL) is preferable if testing is performed at odd hours of the day, as is common in many practices, and maximum sensitivity is desired.