Abstract
PURPOSE: The gonadotropin-releasing hormone (GnRH) stimulation test is regarded as the diagnostic gold standard for central precocious puberty (CPP); however, its invasiveness and limited accessibility restrict its routine use in clinical practice. Moreover, reported cutoff values for basal luteinizing hormone (LH) vary across studies, and simple screening models applicable to everyday clinical settings remain limited. This study aimed to evaluate routinely available clinical, biochemical, and imaging indicators and to develop a simplified screening model for early differentiation of CPP from premature thelarche (PT) in girls. PATIENTS AND METHODS: This retrospective study included 429 girls aged 4-8 years presenting with early breast development between January 2019 and January 2023. Basal sex hormone levels, anthropometric parameters, bone age, and pelvic ultrasound findings were collected. Univariable analysis, multivariable logistic regression, and receiver operating characteristic (ROC) curve analysis were performed to assess the screening performance of individual indicators and a combined model. RESULTS: Girls with CPP exhibited significantly higher age, BMI, bone age advancement, basal LH levels, and ovarian volumes than those with PT did (all p<0.05). Multivariable analysis identified age, BMI, and basal LH concentration as independent factors associated with CPP. Among individual indicators, basal LH concentration showed the strongest discriminatory ability (AUC=0.842), with a cutoff value of 0.15 IU/L. A simplified combined model integrating basal LH concentration, age, and BMI further improved screening performance (AUC=0.880). Pelvic ultrasound parameters did not emerge as independent predictors in multivariable analyses. CONCLUSION: A simplified screening model incorporating basal LH concentration, age, and BMI may serve as a practical initial screening tool to identify girls at higher risk of CPP and to reduce unnecessary GnRH stimulation testing. Prospective multicenter studies are required to validate the generalizability of this approach before broader clinical implementation.