The contribution of inhaled corticosteroid exposure to adrenal insufficiency in a severe asthma cohort monitored with an INCA device

Title: The contribution of inhaled corticosteroid exposure to adrenal insufficiency in a severe asthma cohort monitored with an INCA device
Author(s): V. Brennan J. Martin-Grace C. Mulvey G. Greene1 G. Collier T.McCartan L. Lombard J. Walsh S.Plunkett E. Mac Hale M. Sherlock RW. Costello
Institution: Beaumont
Poster: Click to view poster
Category: Asthma and COPD
Abstract: Inhaled corticosteroids (ICS) are reported to cause adrenal insufficiency (AI) 1. The prevalence is unknown and published studies are confounded by oral corticosteroids (OCS) and selection bias. This study aims to establish the relative contribution of OCS and ICS to adrenal suppression in a severe asthma cohort.
Morning cortisol samples were collected on completion of the INCA-SUN RCT and analysed using the Roche-Elecsys Cortisol-II assay. Results were classed suppressed (<100nmol/l), indeterminate (101-314nmol/l) or normal (>315nmol/l), based on published data predicting short Synacthen test outcomes using morning cortisol.
18.3% (15/82) of participants had a morning cortisol <100 nmol/l, suggesting AI. 12.2% of participants received maintenance OCS and 67% received at least one course of OCS during the study. The predicted prevalence of AI in patients without OCS exposure in the preceding week was 11.9%. Serum cortisol was strongly predicted by cumulative glucocorticoid exposure. ICS exposure adjusted for weight was significantly higher among adrenally suppressed patients (1.7 mg/kg compared to 1.16mg/kg , p=0.009. Logistic regression analysis demonstrated that ICS exposure is a significant predictor of adrenal insufficiency even when controlling for OCS exposure. OR = 2.49 [1.06, 5.82] per 1mg/kg FP increase for ICS, OR = 1.10 [1,03, 1.17] per 1mg/kg Prednisolone increase for OCS.
20% of patients with severe asthma may have undiagnosed AI. Our analysis shows a strong association between cumulative ICS exposure and the risk of AI, independent of OCS use.