Serum progesterone predicts abnormal gestations in clomiphene citrate conception cycles as well as in spontaneous conception cycles.

Objective: To determine if a "discriminatory" P concentration could be established that predicted abnormal early pregnancies in clomiphene citrate (CC)-conceived cycles.

Design: Progesterone concentrations of gestations between 28 and 49 days from last menstrual period in both spontaneously conceived and CC-stimulated cycles were analyzed using a relative-operating characteristic (ROC) curve.

Interventions: Serum P concentrations were measured in 222 pregnant patients from the first 49 days of gestation. One hundred sixteen patients conceived in a spontaneous cycle and 106 patients conceived in CC-treated cycles. Two by two contingency tables were used to calculate the true-positive (sensitivity) and false-positive rates at 20 specific P at 20 specific P concentrations. A ROC curve was then generated by plotting the sensitivity of the test against the percent of normal patients incorrectly classified (false positive) at each P level. The best discriminatory value was estimated in each curve at a point of high sensitivity associated with a minimal false-positive value. The areas under the curve and SE were calculated for each group and compared by the critical ratio z-test.

Results: The best discriminatory P concentration was 10 ng/mL (32 nmol/L) for spontaneously conceived pregnancies and 30 ng/mL (95 nmol/L) for CC-treated pregnancies. The area under each ROC curve was significantly predictive. Comparison of the two curves indicated that the ability of P measurements to predict gestational complications was independent of follicular stimulation.

Conclusions: Follicular stimulation with CC increases the discriminatory P value that predicts gestational normalcy but does not alter the clinical utility of the test.