Toward Optimized and Cost-Efficient Protocols for Inferior Petrosal Sinus Sampling in the Diagnosis of Cushing Disease

Abstract

BACKGROUND AND OBJECTIVE: Conventional protocols for inferior petrosal sinus sampling (IPSS) during the workup of hypercortisolemia require adrenocorticotrophic hormone (ACTH) measurement at multiple time points with corticotropic-releasing hormone (CRH) used as a stimulator. Modernized evidence-based protocols must also reflect the increased utilization of desmopressin (DDAVP) for ACTH stimulation as the manufacturing shortage of traditionally used CRH continues. We model the diagnostic accuracy and potential economic savings of DDAVP utilization and reduced time point sampling. METHODS: A single-institution, retrospective review of patients undergoing IPSS between 2007 and 2021 was performed. A computational search for the minimal set of time points which preserves conventional diagnostic accuracy was performed by testing all 1 through 5-point combinations generated using a binomial expansion. Economic savings were modeled using a publicly available hospital chargemaster. RESULTS: A total of 50 patients qualified for inclusion, 47 of whom were diagnosed with Cushing disease and 3 with ectopic Cushing syndrome. Single-point diagnostic accuracy for DDAVP-stimulated tests was 71.4%–92.9%, and seven 2-point combinations (5-25 procedural minutes) were found, which preserved conventional diagnostic accuracy. Single-point accuracy for CRH-stimulated tests was 67.9%–89.3%, and two 2-point combinations (t = −15, t = +2 and t = −15, t = +10) preserved accuracy. For every time point removed, the cost of ACTH laboratory tests was reduced by $507 from $3042 for 6-point IPSS). The shortest and most economical stimulator and time point combination that preserved conventional accuracy was DDAVP sampling at t = 0 and t = +5 or t = +2 and t = +5 minutes, which cost $2028 total compared with the most expensive 6-point IPSS option with CRH ($6392). CONCLUSION: DDAVP is a cost-efficient and effective alternative to CRH during IPSS. Our results encourage prospective evaluation of potentially fewer sampling time points in the interest of time and cost efficiency balanced with preserved accuracy.