Use of local data to enhance uptake of published recommendations: an example from the diagnostic evaluation of precocious puberty

Background It has been recommended that basal luteinising hormone (LH) levels be used as the initial test to identify cases of central precocious puberty (CPP) in children. However, in clinical practice, gonadotropin-releasing hormone (GnRH) stimulation tests are frequently still used. Objective To assess the diagnostic utility of a single LH to identify CPP in girls, as a means to safely reduce GnRH stimulation testing rates. Design Retrospective analysis of patients referred for GnRH stimulation between August 2007 and December 2010, with prospective 12-month follow-up of GnRH stimulation testing rates post implementation of management algorithm. Patients 57 girls (6.2+/-2.1 years) with early signs of puberty. Main outcome measure Ability of basal LH to predict clinical pubertal progression, 6 months following the GnRH stimulation test. Results Pubertal progression occurred in 18 patients. All patients with a basal LH level >= 0.3 IU/L had subsequent pubertal progression, while 39 of 41 patients with a basal LH <= 0.2 IU/L did not progress, resulting in 100% specificity (95% CI 92% to 100%) and 90.5% sensitivity (69.6% to 98.8%). Using the locally derived algorithm, GnRH stimulation testing was redirected to patients with pubertal progression that was discordant with basal LH data. Post intervention, there was a 75% reduction in GnRH stimulation testing without comprising the rate of diagnosis of CPP. Conclusions Our results confirm the diagnostic utility of basal LH levels in the diagnosis of CPP and demonstrate that dissemination and interpretation of local data may facilitate change in clinical practice, resulting in streamlined patient care and cost savings.