Prevalence of Suppressed Morning Serum Cortisol and Its Relationship With Cumulative Glucocorticoid Exposure in a Moderate-Severe Asthma Patient Cohort

Abstract

The extent to which inhaled glucocorticoid exposure (ICS) contributes to risk of adrenal insufficiency (AI) is not fully understood. The aim of this study was to establish the relative contribution of both oral (OCS) and inhaled (ICS) glucocorticoid exposure to risk of AI.82 patients with severe asthma treated with fluticasone propionate (FP) who participated in a 32-week prospective randomised trial INCA SUN, (NCT02307669) were studied. Cumulative ICS exposure was calculated using a unique digital device, which creates an acoustic recording of inhaler adherence and technique. Analysis of this data provides an exact measure of the ICS dose received by each patient. Morning serum samples collected during the final study visit (week 32) were analysed for serum cortisol concentration (cortisol) using Roche Elecsys Cortisol II immunoassay. Participants were then stratified into three groups based on cortisol concentration to predict risk of AI; cortisol < 100nmol/l (high risk), 100–315 nmol/l (indeterminate risk) and > 315 nmol/l (low risk) based on locally derived reference ranges. 21% participants were classified as low risk, 18% as high risk and the remaining 61% at indeterminate risk of AI. Median morning cortisol in the low risk group was ten-fold higher than those in the high risk group (380 vs 38.5 nmol/l, p=0.001). OCS exposure was a significant predictor of risk of AI (OR 1.1 [1.03–1.17] per mg/kg increase in prednisolone exposure, p=0.004)). Participants at high risk were more likely to be on maintenance OCS (33% vs 0%, p=0.015) and had a greater median cumulative OCS exposure over the study period (7.55 vs 0.66 mg/kg prednisolone, p=0.002). ICS exposure was also associated with risk of AI. Participants at high risk AI had a greater adherence to ICS therapy (78% vs 62%, p=0.049) and greater cumulative received ICS dose over the study duration than those at low risk AI (178.2 vs 127.9 mg, p=0.036). ICS exposure remained a significant predictor of AI even when OCS exposure is controlled for (OR 2.49 [1.06–5.82] per 1mg/kg increase in FP exposure). Both the asthma control test (ACT) & asthma quality of life questionnaire (AQLQ) scores correlate with morning cortisol concentration (ACT r=0.2, p=0.068, AQLQ r=0.26, p=0.019). Interestingly, participants with cortisol < 100nmol/l reported worse asthma control (ACT score 16 vs 20, p=0.07) and a lower AQLQ score (4.1 vs 5.8, p=0.02) than the low risk group despite objectively better lung function (FEV1 90.6 vs 77.6% predicted). Our data suggests that both cumulative oral and inhaled glucocorticoid exposure contribute independently to cortisol suppression and risk of AI. The discrepancy between objective (FEV1) and more subjective measures of asthma control (ACT score) in the high risk group suggests that undiagnosed AI, as well as other non-airway co-morbidities, may contribute to the symptom burden experienced by these patients.