Second line tests in the differential diagnosis of neoplastic and non-neoplastic hypercortisolism: a systematic review and meta-analysis

Introduction: The clinical and hormonal overlap between neoplastic (CS) and non-neoplastic (NNH/pCS) hypercortisolism is a well-known challenge. The gold standard for the diagnosis of NNH/pCS is the resolution/lack of progression of the hypercortisolism over an adequate follow-up period. Various dynamic tests have been proposed to allow an early discrimination between these conditions, but to date there is no agreement on which of them should be used. Aim: To provide an overview of the available tests and to obtain a quantitative synthesis of their diagnostic performance in discriminating NNH/pCS from CS. Methods: An extensive search in Ovid (including Cochrane, Medline and Embase databases) was performed for articles published between 1990 and 2022. The included articles applied one or more second-line tests to differentiate NNH/pCS from CS patients. For the NNH/pCS group, we admitted the inclusion of patients presenting clinical features and/or biochemical findings suggestive of hypercortisolism despite apparent lack of a pCS-related condition. Results: The electronic search identified 339 articles. After references analysis and study selection, we identified 9 studies on combined dexamethasone-corticotropin releasing hormone (Dex-CRH) test, 4 on Desmopressin test and 3 on CRH test; no study on Dex-Desmopressin met the inclusion criteria. Dex-CRH test provided the highest sensitivity (97%, 95CI% [88%; 99%]). CRH tests showed excellent specificity (99%, 95%CI [0%; 100%]), but its sensitivity was lacking. Although metaregression analysis based on diagnostic odds ratio failed to provide a gold standard, CRH test (64.77, 95%CI [0.15; 27174.73]) seemed to lack in performance compared to the others (Dex-CRH 138.83, 95%CI [49.38; 390.32] and Desmopressin 110.44, 95%CI [32.13; 379.63]). Discussion: Both Dex-CRH and Desmopressin tests are valid tools in discriminating NNH/pCS from CS patients, with the first probably achieving slightly better performance and the latter being more suitable for clinical practice. We do not recommend CRH test alone in this setting, although its combination with other tests could be further researched. Further studies are needed on this topic, possibly focusing on mild Cushing’s Disease and clinically well-characterized NNH/pCS patients.