Abstract
INTRODUCTION. Coolen’s equation has been used to calculate free cortisol (FC) in critical-illness but this does not include plasma albumin concentration (which has a high cortisol-binding capacity.) Vermeulen (1999) developed a formula to predict free-testosterone that included albumin plus sexhormone binding globulin (SHBG); we have modified this for FC calculation, replacing SHBG with cortisol binding globulin (CBG) and compared the two approaches. METHODS. Control subjects, and patients with septic shock, underwent a 1μg short synacthen test. Those with known adrenal failure: steroid, oestrogen or etomidate use: or pituitary disease were excluded. CBG was analysed by RIA (Biosource). Serum total cortisol was measured by an automated chemiluminesence assay (Bayer Centaur). Serum FC concentration was measured by steady-state gel filtration. Coolen’s equation was compared to that of Vermeulen, modified for FC, whereby FC = ([TC] - (N x [FC])) / (Kt{CBG - [TC]}). N = Ka Ca + 1 where Ka is the association constant for albumin and Ca is the albumin concentration. From Andersen (2002) we have taken the Kt for CBG as 76 x 106 / M and for albumin as 0.003 x 106 / M. Inter-group comparisons were performed by unpaired Student’s t test. Correlation analysis was determined by Pearson’s correlation coefficient. RESULTS. 9 patients (M6, F3, age 67.1±4.1yrs, APACHE 2 score 23.0±2.8) and 10 age-matched controls (M5, F5, age 63.3±1.9yrs) were studied. There was no difference in TC at zero, 30 or 60 mins between patients and controls (437.4±37.7 vs 492.3 ±63.4; 805±43.6 vs 829.3±139.7; 738.2±52.6 vs 814.6 ±176.2nmol/L.) Both CBG (26.6±3.8 vs 53.2±6.3μg/ml, P=0.001) and albumin (14.4±1.9 vs 42.3±0.6g/L; P<0.001) were significantly lower in patients than in controls. The FC was significantly higher in patients at zero minutes (207.5±44.9 vs 42.2±5.3nmol/L, P=0.003), 30 minutes (370.4±59.1 vs 136.5±10.7, P=0.002) and 60 minutes (335.5±48.6 vs 104.3±9.1, P<0.001) compared to controls. In the patient group, Coolen’s equation correlated to measured FC (r2 0.57, P=0.007) but there was a much closer correlation using the modified Vermeulen equation (r2 0.89, P<0.001). CONCLUSION. The modified Vermeulen formula (that includes albumin measurement) may provide a closer approximation of FC, in critically ill patients, than Coolen’s equation.