OR32-4 Diagnostic Performance of Macimorelin vs the Insulin Tolerance Test (ITT) in 140 Patients at Varying Growth Hormone (GH) Cutpoints: A Post Hoc Analysis

Abstract

Background: Macimorelin (MAC), an orally active ghrelin receptor agonist, is indicated for the diagnosis of adult growth hormone deficiency (AGHD) in the United States. The efficacy of MAC for AGHD diagnosis was previously demonstrated; comparing MAC at cutpoint values of 2.8 and 5.1 ng/mL with ITT at cutpoint values of 3.0 and 5.1 ng/mL.1,2Objective: This post hoc analysis evaluated the percent agreement and sensitivity and specificity of MAC vs ITT over a range of GH cutpoints. Methods: This analysis included data from a phase 3, open-label, randomized, 2-way crossover study of MAC vs ITT in subjects with high (Group A, n=38), intermediate (Group B, n=37), and low (Group C, n=40) likelihood for AGHD and healthy, matched controls (Group D, n=25).1 Percent agreement (negative, positive, and overall) and estimated sensitivity and specificity were determined using GH cutpoint values of 2.8, 4, 5.1, and 6.5 ng/mL for both MAC and ITT. Results: 74 subjects were classified as GH deficient, and 66 subjects were classified as GH sufficient. These subjects were classified based on the ITT using a cutpoint of 5.1 ng/mL.1 The highest negative, positive, and overall agreements between tests were observed when GH cutpoints chosen for MAC and ITT were identical to each other and were either 2.8 or 5.1 ng/mL. With a GH cutpoint value of 2.8 ng/mL for both tests, negative agreement was 94% (95% CI: 86%, 98%), positive agreement was 87% (95% CI: 76%, 94%), and overall agreement was 91% (95% CI: 85%, 95%). At a GH cutpoint value of 5.1 ng/mL for both tests, negative agreement was 92% (95% CI: 83%, 97%), positive agreement was 82% (95% CI: 72%, 90%), and overall agreement was 87% (95% CI: 80%, 92%). Assuming all Group A participants were cases and all Group D participants were controls, estimated specificities of MAC and ITT were identical (96%) at GH cutpoint values of 2.8, 4, or 5.1 ng/mL. Estimated sensitivity for ITT at GH cutpoint value of 5.1 ng/mL (97%) was higher than for MAC at cutpoint value of 2.8 ng/mL (87%); increasing the test cutpoint to 6.5 ng/mL increased sensitivity to 97% and 100% for MAC and ITT, respectively, but at the expense of specificity decreases to 92% (MAC) and 88% (ITT). Conclusions: Among the cutpoints examined, agreement between MAC and ITT was highest at either 2.8 or 5.1 ng/mL, with positive agreement declining modestly at the higher cutpoint. Sensitivity of MAC was maximal at 6.5 ng/mL but at the expense of a decline in specificity from 96% to 92%, which may be undesirable if the primary consideration is minimization of false-positive diagnosis of AGHD. A MAC cutpoint of 5.1 ng/mL provides maximal specificity (96%) and high sensitivity (92%) with good overall agreement to ITT at the same cutpoint (87%), making it clinically useful for the diagnosis of AGHD. Reference: 1. Garcia JM, et al. J Clin Endocrinol Metab. 2018;103(8):3083-3093. 2. Garcia JM, et al. Presented at ENEA; 17-20 October 2018; Wroclaw, Poland.