THU030 Evaluation Of Combined CRH And HDDST Test Performance

Abstract

Abstract Disclosure: H. Elenius: None. R. McGlotten: None. L.K. Nieman: None. Introduction: Determining the etiology of ACTH-dependent Cushing’s Syndrome (CS) is frequently a diagnostic challenge. Non-invasive tests to distinguish Cushing’s Disease (CD) from Ectopic ACTH Syndrome (EAS) are the CRH stimulation test and the 8 mg overnight high dose dexamethasone suppression test (HDDST). The CRH test is more accurate (sensitivity (Se) 83-93%, specificity (Sp) 85-100%) than the HDDST (Se 56-95%, Sp 67-100%), depending on the type of CRH and criteria used. We retrospectively reviewed the results of each test at our institution between 2005-2019 to evaluate the tests' performance at established cut offs, identify optimal response criteria and to see whether combining the two tests provided additional diagnostic accuracy. Methods: CD or EAS was confirmed by either surgical pathology or biochemical cure after tumor resection or imputed from IPSS results (n=7). We identified 100 patients with CD and 34 with EAS. Two EAS patients with cyclic CS each underwent tests with or without active hypercortisolism. Test criteria to identify CD were ≥50% suppression of cortisol for the HDDST and a ≥35% increase in mean ACTH at 15-30 minutes or a ≥20% increase in mean cortisol at 30-45 minutes for the ovine CRH test. Results: The HDDST had a Se of 81%, Sp of 81% and diagnostic accuracy (DA) of 81%, while the CRH test had Se 95%, Sp 75% and DA 90%. Each test improved with modified response criteria. Using a HDDST suppression criterion of ≥63% gave Se 78%, Sp 89% and DA 81%. Using CRH test response criteria of ≥39% increase in ACTH or ≥30% increase in cortisol gave Se 94%, Sp 92% and DA 93%. Performance improved further by combining the two tests, using the optimized criteria (≥63% for HDDST and ≥39% or ≥30% for the CRH test). Test results were concordant in 80%, giving a Se of 96%, Sp 97% and DA 96%. Among EAS patients, this approach only misidentified one subject with cyclic CS who was tested while hypocortisolemic. Removing this test improved combined test performance to Se 96%, Sp 100% and DA 97%. Among the 26 patients with discordant results (CD = 22, EAS = 4), the HDDST result was likelier to be correct in EAS vs CD (50% vs 14%, PPV 60%), while the CRH test was more often correct in CD vs EAS (86% vs 50%, PPV 90%). Conclusions: Based on this series that included a large proportion of EAS patients, the CRH test performed better than the HDDST, in line with earlier data. In cyclic CS, tests performed better in hypercortisolemic patients. The optimal response criteria were ≥63% cortisol suppression for HDDST and ≥39% ACTH or ≥30% cortisol increase for the CRH test. In the 81% of hypercortisolemic cases with concordant test results, these combined optimized cut offs provided the best diagnostic performance, improving it to Se 96%, Sp 100%, DA 97%. When results were discordant, the CRH test worked better to identify patients with CD. Presentation: Thursday, June 15, 2023