Background: Short synacthen tests (SSTs) are routinely performed for the investigation of adrenal insufficiency (AI), but are costly and time consuming. The frequency of SST testing could be reduced by prioritising testing for specific indications (based on likelihood of failure of the test), and also by identification of an assay-specific 9am cortisol level that obviates the need to proceed to SST.
Methods: Retrospective review of indications and results of all SSTs performed over one year at our institution. Receiver-operating characteristic (ROC) curve analysis was performed based on the baseline serum cortisol level to obtain a predictive cut-off value. Cortisol was assayed by Centaur immunoassay (Siemens).
Results: 450 SSTs were performed in one year. Failure rates varied according to indication; only 2% of patients failed SSTs if indicated for fatigue/malaise; substantially more patients failed for other indications such as following adrenal surgery (54%), suspected HPA axis suppression (57%), or reassessment of known HPA axis suppression (67%).
ROC curve analysis identified that a 9am cortisol of <56 nmol/L predicted failure with 100% sensitivity and >374 nmol/L provided 100% specificity to pass the test; for 99% sensitivity and specificity the values were 114 and 362 nmol/L respectively.
Conclusion: Pass rates for SSTs varied considerably depending on indication. The low failure rate for those with fatigue suggests that the test should only be performed if additional clinical parameters support the diagnosis of AI. A baseline serum cortisol level of <56 nmol/L predicted SST failure in all individuals; levels >374 nmol/L predicted a 100% pass rate. If SSTs were avoided in all individuals with a 9am cortisol outside these cutoffs, 154 tests would have been avoided, resulting in significant cost benefit.