BackgroundThe criterion standard gonadotropin-releasing hormone (GnRH) stimulation tests to diagnose central precocious puberty (CPP) are time-consuming, inconvenient, and expensive.ObjectivesTo determine predictive cut-off values codetermined by ultrasonographic parameters and basal gonadotropin levels in girls with premature sexual development and compare them results of criterion standard tests in a study of diagnostic accuracy.MethodsRetrospective review of hormonal investigations and ultrasonographic uterine and ovarian parameters in a consecutive sample of girls at a single center, tertiary care hospital in Bangkok, Thailand.ResultsWe separated data from 68 girls (age range 2–12 years) into 2 groups based on their response to a GnRH analogue agonist stimulation test. A “prepubertal response” group included girls with premature thelarche and thelarche variants (n = 18, 6.37 ± 1.77 years) and a “pubertal response” group, including girls with CPP (n = 50, 8.46 ± 1.46 years); excluding patients with pathological causes (n = 0). The basal level of luteinizing hormone (LH) had the largest area under receiver operating characteristic curves (AUC) of 0.84; 95% confidence interval [CI] 0.74–0.93) compared with basal levels of follicle stimulating hormone (AUC 0.77; 95% CI 0.64–0.90) or estradiol (0.70; 95% CI 0.56–0.85). An optimal cut-off of 0.25 IU/L LH was related to a pubertal response to GnRH analogue agonist stimulation tests with 75.0% sensitivity, 88.9% specificity, 94.7% positive predictive value (PPV), and 57.1% negative predictive value. Uterine and ovarian cut-off volumes of 3.5 cm3 and 1.5 cm3 were related to a pubertal response with 88.6% and 76.2% PPV, respectively. A uterine width cut-off of 1.7 cm combined with a basal LH cut-off of 0.25 IU/L increased specificity and PPV to 100%.ConclusionCombining uterine and ovarian ultrasonographic parameters with basal gonadotropin levels, especially uterine width and basal LH level, appears useful for diagnosis of CPP.