Free and Total Plasma Cortisol Measured by Immunoassay and Mass ...

ORIGINAL

ARTICLE

E n d o c r i n e

C a r e

Free and Total Plasma Cortisol Measured by Immunoassay and Mass Spectrometry Following ACTH1–24 Stimulation in the Assessment of Pituitary Patients Morton G. Burt, Brenda L. Mangelsdorf, Anne Rogers, Jui T. Ho, John G. Lewis, Warrick J. Inder, and Matthew P. Doogue Southern Adelaide Diabetes and Endocrine Services (M.G.B., B.L.M., J.T.H., M.P.D.) and Flinders University (M.G.B., A.R., J.T.H., M.P.D.), Adelaide, Australia 5041; Steroid and Immunobiochemistry Laboratory (J.G.L.), Canterbury Health Laboratories, Christchurch, New Zealand 8140; and Department of Endocrinology and Diabetes (W.J.I.), St Vincent’s Hospital, Melbourne, Australia 3065

Context: Measurement of plasma cortisol by immunoassay after ACTH1–24 stimulation is used to assess the hypothalamic-pituitary-adrenal (HPA) axis. Liquid chromatography-tandem mass spectrometry (LCMS) has greater analytical specificity than immunoassay and equilibrium dialysis allows measurement of free plasma cortisol. Objective: We investigated the use of measuring total and free plasma cortisol by LCMS and total cortisol by immunoassay during an ACTH1–24 stimulation test to define HPA status in pituitary patients. Design and Setting: This was a case control study conducted in a clinical research facility. Participants: We studied 60 controls and 21 patients with pituitary disease in whom HPA sufficiency (n ⫽ 8) or deficiency (n ⫽ 13) had been previously defined. Intervention: Participants underwent 1 ␮g ACTH1–24 intravenous and 250 ␮g ACTH1–24 intramuscular ACTH1–24 stimulation tests. Main Outcome Measures: Concordance of ACTH1–24-stimulated total and free plasma cortisol with previous HPA assessment. Results: Total cortisol was 12% lower when measured by immunoassay than by LCMS. Female sex and older age were positively correlated with ACTH1–24-stimulated total and free cortisol, respectively. Measurements of total cortisol by immunoassay and LCMS and free cortisol 30 minutes after 1 ␮g and 30 and 60 minutes after 250 ␮g ACTH1–24 were concordant with previous HPA axis assessment in most pituitary patients. However, free cortisol had greater separation from the diagnostic cutoff than total cortisol. Conclusions: Categorization of HPA status by immunoassay and LCMS after ACTH1–24 stimulation was concordant with previous assessment in most pituitary patients. Free cortisol may have greater clinical use in patients near the diagnostic threshold. (J Clin Endocrinol Metab 98: 1883–1890, 2013)

lucocorticoids regulate metabolic, developmental, and immune function and play a pivotal role in preserving basal and stress-related homeostasis. Deficiency of cortisol results in substantial morbidity and may be fatal

G

if untreated (1). Cortisol deficiency can be due to adrenal (primary) or hypothalamic-pituitary (secondary) dysfunction. Although the diagnosis of primary adrenal failure is generally straightforward, evaluation of the hypothalam-

ISSN Print 0021-972X ISSN Online 1945-7197 Printed in U.S.A. Copyright © 2013 by The Endocrine Society Received October 9, 2012. Accepted March 6, 2013. First Published Online March 28, 2013

Abbreviations: BMI, body mass index; CBG, corticosteroid-binding globulin; CI, confidence interval; HPA, hypothalamic-pituitary-adrenal; IHT, insulin hypoglycemia test; LCMS, liquid chromatography-tandem mass spectrometry.

doi: 10.1210/jc.2012-3576

J Clin Endocrinol Metab, May 2013, 98(5):1883–1890

jcem.endojournals.org

1883

1884

Burt et al

Free Cortisol in Pituitary Patients

ic-pituitary-adrenal (HPA) axis in patients at risk of secondary adrenal dysfunction can be a diagnostic challenge. Assessment for secondary adrenal dysfunction relies on interpretation of basal cortisol concentration and change in cortisol concentration in response to physiological or pharmacological stimuli. The insulin hypoglycemia test (IHT) is considered the gold-standard test of the HPA axis in patients with pituitary disease (2, 3). However, the IHT is resource intensive, is unpleasant for patients, and carries a small risk of seizure or loss of consciousness. This has contributed to the common use of ACTH1–24 tests to assess HPA function in patients with pituitary disease (4, 5). The peak cortisol after ACTH1–24 stimulation provides important prognostic information regarding HPA function in pituitary patients (6). However, it is primarily a test of adrenal (rather than pituitary) function. Furthermore, the sensitivity of ACTH1–24 stimulation testing to define secondary adrenal failure is only 60% when compared with the IHT or metyrapone tests (7). Current assessment of HPA function is based on the measurement of total cortisol by immunoassay, often using diagnostic thresholds derived from the literature. However, commercially available cortisol immunoassays vary in their precision and may cross-react with other steroid hormones; hence, diagnostic thresholds should be assay-specific (8). Furthermore, only about 5% of circulating cortisol is free and in equilibrium with the cortisol receptor, with approximately 80% bound to corticosteroid-binding globulin (CBG) and the remainder to albumin (9). Interindividual variability in CBG can result in a difference in the total (measured) cortisol concentration when there is no difference in the free (physiologic) cortisol concentration (10). Additionally, the capacity of CBG to bind cortisol is saturable at physiological concentrations. Thus, at higher concentrations, a moderate increase in total cortisol may be associated with a large increase in free cortisol (11). Liquid chromatography-tandem mass spectrometry (LCMS) provides greater analytical specificity than immunoassays (12). The costs of LCMS assays are decreasing and they are becoming increasingly available to measure steroid hormones. Furthermore, LCMS has sufficient sensitivity in combination with equilibrium dialysis to measure plasma free cortisol accurately. However, the role of LCMS during ACTH1–24 stimulation testing in pituitary patients is not known. Other patient variables are often not considered when interpreting the results of ACTH1–24 stimulation tests. However, they may be important. For example, sex was reported to influence ACTH1–24-stimulated total cortisol secretion in some (13) but not all (14) studies. Furthermore, age, body mass index (BMI), and sex influence ad-


renal sensitivity to endogenous ACTH pulses (15). The effect of these variables on ACTH1–24-stimulated total and free cortisol secretion is worthy of investigation. The primary aim of this study was to investigate whether the measurement of total and free plasma cortisol by LCMS during the low- and high-dose ACTH1–24 stimulation tests better defines HPA status in pituitary patients than total cortisol measured by immunoassay. We also established a reference range for basal and ACTH1–24stimulated plasma total and free cortisol measured by LCMS and assessed the relationship between age, BMI, and sex and ACTH1–24-stimulated total and free cortisol concentrations.

Materials and Methods This open cross-sectional study was undertaken within the Endocrine Research Unit, Repatriation General Hospital, Adelaide, Australia and the Department of Endocrinology, St Vincent’s Hospital, Melbourne, Australia. The study was approved by the Southern Adelaide Clinical Human Research Ethics Committee, Adelaide and the Human Research Ethics Committee, St Vincent’s Hospital, Melbourne. All subjects provided written informed consent in accordance with the Declaration of Helsinki.

Subjects We prospectively recruited 60 healthy volunteers aged 20 to 80 years from the general community who had no history of pituitary or adrenal disease and were not taking oral glucocorticoids or estrogens and 21 patients with established pituitary disease. Twelve pituitary patients had had nonfunctioning tumors, 4 acromegaly, 1 Cushing’s disease, 1 prolactinoma, 2 congenital hypopituitarism, and 1 hypophysitis. Thirteen pituitary patients were taking glucocorticoid replacement (8 subjects hydrocortisone, 19 ⫾ 6 mg/d; 4 subjects cortisone acetate, 22 ⫾ 3 mg/d; 1 subject prednisolone 5 mg/d), 11 thyroxine, 5 testosterone, 1 oral estrogen, 1 transdermal estrogen, 3 GH, and 4 desmopressin. Fifteen pituitary patients had undergone pituitary surgery and 2 had received pituitary radiotherapy. In all pituitary patients HPA sufficiency (n ⫽ 8) or deficiency (n ⫽ 13) had been defined by IHT (glucose ⬍2.2 mmol/L, cortisol cutoff 550 nmol/L) or morning cortisol (deficient ⱕ100 nmol/L, sufficient ⱖ500 nmol/L).

Study design Subjects attended the Endocrine Research Unit, Repatriation General Hospital, or the Department of Endocrinology, St Vincent’s Hospital at 9 AM after an overnight fast on 2 occasions about 7 days apart. For subjects taking glucocorticoids, the afternoon glucocorticoid dose was withheld the day before ACTH1–24 testing, to ensure testing occurred more than 24 hours after last glucocorticoid administration. On the first visit, height, weight, waist and hip circumference, pulse, and blood pressure were recorded. A cannula was inserted into an antecubital fossa vein for blood sample collections. At each visit, subjects were administered either 1 ␮g ACTH1–24 (Synacthen; Novartis Pharmaceuticals, Sydney, Australia) IV or 250 ␮g ACTH1–24 IM, in

doi: 10.1210/jc.2012-3576

random order. For the 1 ␮g ACTH1–24 stimulation test, 250 ␮g ACTH1–24 was added to 99 mL 0.9% sodium chloride, producing a solution with an ACTH1–24 concentration of 2.5 ␮g/mL. Following agitation for at least 2 minutes, 0.4 mL of the diluted ACTH1–24 solution was drawn up using a 50 U insulin syringe and injected as a bolus directly into the IV cannula, which was then flushed with normal saline. Blood samples were collected at baseline, and 30 and 60 minutes after ACTH1–24 administration.

jcem.endojournals.org

Table 1. Subject Characteristics of Controls and Pituitary Patients

Number, n Sex (female/ male) Age, ya Weight, kga Height, ma BMI, kg/m2a Waist, cma Waist/hip ratioa CBG, nmol/La

Laboratory analysis Plasma total cortisol was measured by immunoassay (Elecsys 2010; Roche Diagnostics, Mannheim, Germany; coefficient of variation ⬍ 3%) and by in-house LCMS assay as previously described (16). In brief, following the addition of an internal standard d4-cortisol and extraction with tert-butylmethyl-ether, plasma total cortisol was measured by LCMS (API 3200 Q Trap; Applied Biosystems, Scoresby, Australia) with selective ion monitoring of mass-to-charge ratio 363-121 for cortisol and 367-121 for d4-cortisol. The coefficient of variation was 3.4% to 6.2% at cortisol concentrations between 70 and 1000 nmol/L. Plasma free cortisol was isolated at 37°C using a commercial equilibrium dialysis kit (Pierce, Rockford, Illinois) and measured by LCMS. Plasma free cortisol was calculated using the following equation: free cortisol ⫽ total cortisol dialysate cortisol/retentate cortisol. The Elecsys Cortisol CalSet calibrators were used for both the immunoassay and the LCMS assays. Plasma CBG was measured using a 2-site noncompetitive monoclonal antibody– based ELISA assay, as previously described (17).

Statistical analysis Most data were analyzed using SPSS 20.0 for Windows (SPSS Inc, Chicago, Illinois) with 95% confidence intervals (CI) calculated using Analyze-It for Microsoft Excel (version 2.26; Microsoft Corp, Leeds, United Kingdom) and sensitivities and specificities calculated at http://www.vassarstats.net/clin1.html. Results are reported as mean ⫾ SD with P ⬍ .05 considered statistically significant. ␹2 tests were used to assess categorical variables and unpaired and paired t tests were used to assess independent and dependent continuous variables, respectively. The relationships between age and BMI with CBG and basal and maximal ACTH1–24-stimulated (30-minute concentration during 1 ␮g test, 60-minute concentration during 250 ␮g test) total and free cortisol concentrations were assessed by simple linear regression analyses. As the relationships for total cortisol by immunoassay and LCMS were similar, only the LCMS results are reported. If more than one variable (sex, age, BMI) were significantly related to CBG or cortisol in univariate analyses, these were then entered into a multivariate analysis. The sensitivity and specificity of each cortisol assay to predict ACTH deficiency and concordance with previous HPA assessment at each time point were calculated based on the lower limits of the 95% CI derived from normal subjects, with the patients’ previous assessment of the HPA axis considered the gold standard.

Results Subject characteristics (Table 1) Subjects with pituitary disease had a significantly higher mean waist circumference and waist/hip ratio than

1885

a

Controls

Pituitary Patients

P Value

60 34/26

21 14/7

.42

60.5 ⫾ 11.5 80.6 ⫾ 19.1 1.70 ⫾ 0.09 27.7 ⫾ 5.6 92.9 ⫾ 14.0 0.88 ⫾ 0.09 611 ⫾ 147

57.5 ⫾ 15.4 87.9 ⫾ 18.7 1.70 ⫾ 0.06 30.4 ⫾ 6.6 102.8 ⫾ 14.2 0.92 ⫾ 0.08 566 ⫾ 109

.34 .14 .98 .08 .006 .046 .21

Mean ⫾ SD.

controls. In women, mean waist circumference was 13 cm greater in pituitary patients than in normal subjects (P ⫽ .001). In men, mean waist circumference was 7 cm greater in pituitary patients, but the difference was not statistically significant (P ⫽ .26). The mean BMI of the pituitary patients was 2.7 kg/m2 higher than controls; however, this difference was not statistically significant. There were no other significant differences in other demographic variables between the 2 groups. In particular, mean CBG concentration was not significantly different in the controls and pituitary patients. Relationship between sex, age, BMI, CBG, and cortisol in controls Mean CBG concentration was significantly higher in women than in men (673 ⫾ 135 vs 530 ⫾ 123 nmol/L, P ⬍ .0001). Basal total (464 ⫾ 143 vs 419 ⫾ 93 nmol/L, P ⫽ .17) and free (14.6 ⫾ 7.4 vs 16.3 ⫾ 6.1 nmol/L, P ⫽ .34) cortisol concentrations were not significantly different in women and men. Thirty minutes after 1 ␮g ACTH1–24 mean total (814 ⫾ 111 vs 699 ⫾ 108 nmol/L, P ⬍ .0001), but not free (52.3 ⫾ 10.9 vs 53.1 ⫾ 16.3 nmol/L, P ⫽ .83), cortisol concentration was significantly higher in women than in men. Sixty minutes after 250 ␮g ACTH1–24 mean total cortisol concentration was significantly higher (940 ⫾ 122 vs 840 ⫾ 116, P ⫽ .002) and mean free cortisol concentration was significantly lower (72.9 ⫾ 12.5 vs 81.3 ⫾ 19.4 nmol/L, P ⬍ .05) in women than in men. The concentrations of CBG (r ⫽ 0.05, P ⫽ .74), basal total cortisol (r ⫽ 0.19, P ⫽ .15), and basal free cortisol (r ⫽ 0.21, P ⫽ .11) were not significantly correlated with age. Similarly, total cortisol concentrations 30 minutes after 1 ␮g ACTH1–24 (r ⫽ 0.15, P ⫽ .26) and 60 minutes after 250 ␮g ACTH1–24 (r ⫽ 0.23, P ⫽ .07) were not significantly correlated with age. In contrast, free cortisol concentrations 30 minutes after 1 ␮g ACTH1–24 (r ⫽ 0.38, P ⫽ .003) and 60 minutes after 250 ␮g ACTH1–24 (r ⫽ 0.49, P ⬍ .0001; Figure 1) were both positively corre-

1886

Burt et al

140

Free Cortisol in Pituitary Patients


CBG concentration was significantly and negatively correlated with BMI (r ⫽ ⫺0.28, P ⫽ .03). However, in a multiple regression analysis CBG concentration was significantly independently related to sex (P ⬍ .0001), but not to BMI (P ⫽ .052). Basal and 1 ␮g and 250 ␮g ACTH1–24stimulated total and free cortisol concentration were not significantly correlated with BMI (data not shown).

r=0.49, p