Horm Metab Res 2025; 57(03): 170-176
DOI: 10.1055/a-2509-3766
Original Article: Endocrine Care

Reassessing the Role of Morning Cortisol in Adrenal Insufficiency Diagnosis: Insights from a Multicentric Cohort

1   Endocrinology, Centro Hospitalar de Lisboa Ocidental EPE, Lisboa, Portugal (Ringgold ID: RIN58410)
,
Catarina Regala
2   Endocrinology, Instituto Português de Oncologia de Lisboa Francisco Gentil EPE, Lisboa, Portugal (Ringgold ID: RIN37838)
,
Clotilde Limbert
1   Endocrinology, Centro Hospitalar de Lisboa Ocidental EPE, Lisboa, Portugal (Ringgold ID: RIN58410)
,
Tiago Nunes Silva
2   Endocrinology, Instituto Português de Oncologia de Lisboa Francisco Gentil EPE, Lisboa, Portugal (Ringgold ID: RIN37838)
,
João Sequeira Duarte
1   Endocrinology, Centro Hospitalar de Lisboa Ocidental EPE, Lisboa, Portugal (Ringgold ID: RIN58410)
,
Valeriano Leite
2   Endocrinology, Instituto Português de Oncologia de Lisboa Francisco Gentil EPE, Lisboa, Portugal (Ringgold ID: RIN37838)
,
Susana Prazeres
3   Laboratory of Endocrinology, Pathology Department, Instituto Português de Oncologia de Lisboa Francisco Gentil EPE, Lisboa, Portugal (Ringgold ID: RIN37838)
› Author Affiliations
Zoom Image

Abstract

The diagnosis of adrenal insufficiency (AI) is challenging due to nonspecific symptoms. Measuring 8 AM serum cortisol levels is a common screening test, but its accuracy in predicting AI remains uncertain. This study aimed to evaluate the predictive diagnostic value of basal morning cortisol levels in suspected AI cases and compare them with levels in healthy individuals. We have conducted a retrospective multicentric cohort study. The study included an epidemiological cohort (patients with AI suspicion) and a control cohort (healthy volunteers). In the epidemiological cohort an ACTH-stimulation test was performed to confirm (AI group) or exclude (non-AI group) the disease. We have included 273 individuals (168 suspected AI cases and 105 controls). Basal cortisol levels were higher in the control group compared to the AI-suspected group (9.9±3.2 μg/dl vs. 7.7±3.3 μg/dl, p<0.001), though both were below the 15 μg/dl recommended threshold for excluding AI. Within the epidemiological cohort, even after a propensity score matching, taking into consideration sex, age and AI symptoms complaints, no significant difference in basal cortisol levels was found between patients with and without AI [6.10 (4.43; 8.45) vs. 7.14 (4.68; 12.15), p=0.128]. Logistic regression and ROC curve analyses showed a low predictive value for basal cortisol, with a positive predictive value of 18.9% and low specificity. AI patients more frequently experienced hypotension and nausea, compared with non-AI patients. Morning cortisol levels alone are unreliable for diagnosing AI. Revising cortisol cutoffs and incorporating symptom-based criteria may improve diagnostic accuracy.

Supplementary Material



Publication History

Received: 05 October 2024

Accepted after revision: 28 December 2024

Article published online:
23 January 2025

© 2025. Thieme. All rights reserved.

Georg Thieme Verlag KG
Oswald-Hesse-Straße 50, 70469 Stuttgart, Germany