Fracture risk by cortisol excess status in patients with adrenal incidentalomas: a population-based cohort study.

Abstract

Adrenal incidentalomas (AIs) may secrete excess cortisol, representing an elevated endogenous exposure to glucocorticoids, which could decrease bone mineral density and increase fracture risk. However, measurement of cortisol excess is not routinely done in patients with AI; thus, those with hormonally active AI at increased risk for fracture are under-identified. We sought to examine the association between excess cortisol levels and the incidence of fragility fracture in people with AI. This retrospective cohort study, conducted within two Kaiser Permanente regions (Southern California and Georgia), comprised women and men aged ≥50 yr with identified AI in the study period January 1, 2015-August 31, 2022. Patients' cortisol excess status was categorized by the type of test conducted (if any) and the test result. Fractures and relevant covariates were ascertained via International Classification of Diseases (ICD)-9/10 codes. Hazard ratios (HR) were estimated using Cox proportional hazard models with mortality as a competing risk. Among the cohort of 14 886 patients with AI, 273 (1.8%) had autonomous cortisol secretion (ACS) confirmed by dexamethasone suppression test (DST) results >1.8 μg/dL (>50 nmol/L), and another 201 (1.4%), tested with urine free or random cortisol tests, had results suggestive of excess cortisol production. Most of the cohort (n = 9353, 62.8%) were untested around AI diagnosis or during follow-up. Compared to patients with normal DST results (and adjusted for age, sex, race/ethnicity, and several other clinical characteristics), the estimated HR of fracture risk for patients with ACS (HR 1.42, CI 0.86-2.32), evidence of cortisol excess (1.41, 0.85-2.32), and untested patients (1.28, 0.88-1.87) were suggestive of elevated risk. However, none of the elevated hazard rates were statistically significant at the 95% significance level. The apparent elevated risk in the untested patients suggests that many untested patients may have hormonally active AI that puts them at risk for fracture from secondary osteoporosis.