ECE2013 Poster Presentations Adrenal cortex (64 abstracts)
1Department of Endocrinology, Bordeaux, France; 2USMR, Bordeaux, France; 3Department of Nuclear Medicine, Bordeaux, France.
The performance of late-night salivary cortisol (LNSC) and optimal number of samples to accurately diagnose post-operative recurrence of Cushings disease (CD) at an early-stage are unknown. We performed a retrospective analysis in a single tertiary-care center to compare the accuracy of multiple salivary sampling strategies to diagnose early-stage recurrences of CD. 36 patients in remission of CD after surgery were followed up for 69.2±10.6 mo with multiple successive measurements of LNSC as part of long-term follow-up evaluation. Following an extensive biochemical evaluation, patients were classified as being in remission or in early-stage recurrence. The diagnostic accuracy of three diagnostic strategies combining 2, 3 or 4 LNSC was estimated and compared using areas under the ROC curves (AUC), sensitivity, specificity and predictive values. 44 sequences of assays (recurrence prevalence 52.3%) were available in 36 patients. The intra-sequence variability of LNSC was higher during recurrence (medians of S.D.s: 2.1 vs 0.5 nmol/l; P<0.0001, Wilcoxon test). AUCs ranged from 0.93 to 0.96 depending on the strategy. For 90% sensitivity, the best specificity (90.9%) was achieved when taking into account the maximum LNSC value in the four-first dosages strategy. For 25% recurrence prevalence, the proportion of false-negatives would be 3.7% and of false-positive 23.4%.
Conclusion: Due to a major within-patient variability of LNSC from one day to another, a strategy using 4 samples collected on successive days may detect early-stage recurrence of CD with a high accuracy, avoiding unnecessary hospitalization for complex testing in more than 75% of patients.