Impact of old and current immunoassays on the 1 mg overnight dexamethasone suppression test: comparison with LC-MS/MS.

OBJECTIVE

The influence of immunoassay performance in hypercortisolism and dexamethasone suppression test (DST) settings was scarcely investigated. We evaluated the effectiveness of 2 immunoassays in detecting hypercortisolism compared to recommended liquid chromatography-tandem mass spectrometry (LC-MS/MS), and compared immunoassay analytical performance in basal and post-DST conditions.

METHODS

We measured cortisol in post-DST sera of patients with suspected hypercortisolism or adrenal incidentalomas by Elecsys gen I (n = 260), and by Access (n = 217). All samples were also measured by a validated LC-MS/MS method. We estimated hypercortisolism rate according to the established 50 nmol/L cutoff, and generated immunoassay-specific cutoffs providing >95% sensitivity and >80% specificity. Finally, we compared cortisol measurements in basal and post-DST samples.

RESULTS

Using the 50 nmol/L cutoff, both immunoassays detected lower rates of hypercortisolism compared with LC-MS/MS, particularly in patients with adrenal adenomas (P < .050). Elecsys gen I and Access determined 6.9% and 6.4% possible false negatives, respectively. Elecsys gen I also caused 3.8% possible false positives. Optimal cutoff was 41 nmol/L for Elecsys gen I (sensitivity: 97.7%; specificity: 80.8%), and 33 nmol/L for Access (sensitivity: 97.5%; specificity: 78.3%). In basal and post-DST samples, Elecsys gen I overestimated by 32.5% and 6.1%, whereas Access underestimated by -4.7% and -5.9% compared to LC-MS/MS cortisol measurements, respectively. Sex differences in method deviations were noted.

CONCLUSIONS

Both immunoassays demonstrated remarkable underdetection of hypercortisolism, suggesting the application of a method-specific cutoff. Immunoassay performance may not be uniform in basal and post-DST conditions and should be purposely examined. Accurate LC-MS/MS methods should be preferred in hypercortisolism settings.