Plasma Free Cortisol Vs. Total Cortisol in Healthy Individuals and in States of High and Low Cortisol Binding Globulin, Including Oral Contraceptive Use, Cirrhosis and Critical Illness: Implications for Diagnosing Adrenal Insufficiency

Presentation Number: SAT 402
Date of Presentation: April 1st, 2017

Laura E. Dichtel*1, Melanie Schorr1, Claudia Loures de Assis2, Jessica K. Sims2, Kathleen E. Corey1, Richard E. Reitz3, Michael J. McPhaul3 and Karen K. Miller1
1Massachusetts General Hospital/Harvard Medical School, Boston, MA, 2Massachusetts General Hospital, Boston, MA, 3Quest Diagnostics Nichols Institute, San Juan Capistrano, CA


Accurate diagnosis of adrenal insufficiency (AI) is critical, as there are risks to both over and under diagnosis. Data using LC-MS/MS assays for total cortisol (TC) and free cortisol (FC) in states of high or low cortisol binding globulin (CBG) levels, including oral estrogen use, cirrhosis and critical illness, are needed. The goals of this study were to determine: 1) the relationship between TC and FC levels in healthy individuals and groups with abnormal CBG levels, and 2) the FC level by LC-MS/MS that best predicts a TC of 18 μg/dl (the standard AI diagnostic cut-off on ACTH-stimulation testing) in healthy individuals.

338 subjects in 4 groups were studied: healthy controls (HC)(n=243), oral contraceptive users (OCP)(n=31), patients with cirrhosis (n=38) and intensive care unit patients (ICU)(n=26). FC and TC were measured by LC-MS/MS and albumin by spectrophotometry (Quest Diagnostics Nichols Institute, San Juan Capistrano, CA), and CBG by ELISA (Biovendor, Ashville, NC).

Mean age was 48±17 y and mean BMI 28±7 kg/m2. Mean albumin (g/dL) and CBG (μg/mL) were higher in OCP and HC than in cirrhosis and ICU [(albumin: OCP 4.6±0.3, HC 4.7±0.3 versus cirrhosis 3.5±0.9 and ICU 3.1±0.7, p<0.001) and (CBG: OCP 57±18 and HC 28±7 versus ICU 22±4 and cirrhosis 21±7, p<0.0001)]. CBG correlated weakly with albumin in HC (R=0.32, p<0.0001) and cirrhosis (R=0.46, p<0.02) but not in OCP or ICU. Mean random TC (μg/dL) was 18.9±8.5 in OCP, 14.7±6.6 in ICU, 11.5±7.3 in HC and 9.9±5.7 in cirrhosis. Mean FC (μg/dL) and % FC were highest in ICU subjects (FC: ICU 1.29±1.08, cirrhosis 0.59±0.47, HC 0.58±0.67 and OCP 0.33±0.15, p<0.01 vs. ICU; % FC: ICU 7.7±4.8%, cirrhosis 5.4±2.7%, HC 4.2±2.2% and OCP 1.8±0.8%, p<0.05 vs. ICU). TC strongly correlated with FC in the ICU, HC and cirrhosis groups and more weakly in the OCP group (ICU R=0.91, HC R=0.90, cirrhosis R=0.86, OCP R=0.70, all p<0.0001). Consistent with Hamrahian et al, ICU subjects with albumin <2.5 gm/dL had higher FC despite similar TC. In receiver operator curve (ROC) analysis in the HC group, an FC cutoff of 1.0 μg/dL predicted a TC of ≥18 μg/dL with 91% sensitivity and 98% specificity (AUC 0.98, p<0.0001). No HC subject with an FC ≥ 1.6 μg/dL had a TC < 18 μg/dL (67% sensitivity; 100% specificity). An FC cutoff of ≥1.0 μg/dL predicted a TC of ≥15.1 mcg/dL in ICU and ≥13.4 mcg/dL in cirrhosis with 100% sensitivity but only 92% and 83% specificity, respectively. In a separate group of 8 HC who underwent ACTH-stimulation testing, all had stimulated TC ≥ 18 μg/dL and FC levels ≥1.0 μg/dL, except one, who had a TC of 16.7 μg/dL with an FC of 1.4 μg/dl.

In conclusion, an FC of ≥ 1.0 μg/dL in this LC-MS/MS assay predicts a TC level of ≥ 18 μg/dL with maximum sensitivity and specificity. Clinicians should use caution in interpreting TC levels in patient groups with altered CBG levels, particularly in women on OCPs, given the variable resulting increase in CBG.


Disclosure: RER: Employee, Quest Diagnostics, Employee, Quest Diagnostics. MJM: Employee, Quest Diagnostics, Employee, Quest Diagnostics. Nothing to Disclose: LED, MS, CL, JKS, KEC, KKM