Unbound or free cortisol constitutes a small fraction of total plasma cortisol, but is believed to represent the biologically active portion of this circulating glucocorticoid. We tested the hypothesis that the percentage free cortisol was altered in plasma from dogs with hyperadrenocorticism, which could account for a greater target tissue response to this circulating hormone. The percentage free cortisol in plasma samples from human beings, healthy dogs, and dogs with hyperadrenocorticism was estimated, using centrifugal ultrafiltration-dialysis. Total cortisol concentrations were determined by use of radioimmunoassay. Total cortisol concentrations appeared greater in plasma from human beings than in plasma from either group of dogs. However, the percentage free cortisol was lower in plasma from human beings, resulting in a calculated concentration of free cortisol that was quite similar between plasma from human beings and healthy dogs. Total plasma cortisol concentrations were greater (P < 0.01) in samples from dogs with hyperadrenocorticism (190 +/- 113 nmol/L; mean +/- SD) than in healthy dogs (102 +/-85 nmol/L), but the percentage free cortisol was not different between these 2 groups (dogs with hyperadrenocorticism, 16 +/- 9%; healthy dogs, 13 +/- 6%). However, plasma free cortisol concentrations (product of total and the percentage of free cortisol) were greater (P < 0.01) in samples from dogs with hyperadrenocorticism (36 +/- 41 nmol/L) than in those from healthy dogs (16 +/- 9 nmol/L). Significant (P < 0.001) positive linear relationships were found between total cortisol concentrations and percentage free cortisol in plasma samples from healthy dogs and dogs with hyperadrenocorticism. Furthermore, the slope of these lines was not different between the 2 groups, providing no evidence for alterations in cortisol binding associated with hyperadrenocorticism. The higher total cortisol concentrations in dogs affected with this disease do, however, result in greater concentrations of free cortisol in circulation, contributing to the development of clinical signs observed in this disease.

Veterinary diagnostic endocrinology laboratories frequently receive hemolyzed plasma, serum, or blood samples for hormone analyses. However, except for the previously reported harm done by hemolysis to canine insulin, effects of hemolysis on quantification of other clinically important hormones are unknown. Therefore, these studies were designed to evaluate effects of hemolysis on radioimmunoassay of thyroxine, 3,5,3'-triiodothyronine, progesterone, testosterone, estradiol, cortisol, and insulin in equine, bovine, and canine plasma. In the first experiment, hormones were measured in plasma obtained from hemolyzed blood that had been stored for 18 hours. Blood samples were drawn from pregnant cows, male and diestrous female dogs, and male and pregnant female horses. Each sample was divided into 2 equal portions. One portion was ejected 4 times with a syringe through a 20-gauge (dogs, horses) or 22-gauge (cows) hypodermic needle to induce variable degrees of hemolysis. Two subsamples of the blood were taken before the first and after the first, second, and fourth ejections. One subsample of each pair was stored at 2 to 4 C and the other was stored at 20 to 22 C for 18 to 22 hours before plasma was recovered and stored at -20 C. The second portion of blood from each animal was centrifuged after collection; plasma was recovered and treated similarly as was blood. Concentrations of thyroxine in equine plasma, of 3,5,3'-triiodothyronine, estradiol, and testosterone in equine and canine plasma, and of cortisol in equine plasma were not affected by hemolysis. Storage of bovine blood at either temperature and equine blood at 20 to 22 C caused progesterone concentrations to decrease (P < 0.05); the effect was not enhanced or diminished by hemolysis. Insulin concentration in equine blood decreased (P < 0.05) at both temperatures; this effect was exacerbated by hemolysis. In the second experiment, blood samples from horses and dogs were hemolyzed and plasma was immediately recovered and stored for 18 to 22 hours at 2 to 4 C or 20 to 22 C. Storage of hemolyzed equine plasma did not affect concentrations of progesterone, insulin, or thyroxine at either temperature. Whereas progesterone concentration was not affected in hemolyzed canine plasma, hemolysis decreased (P < 0.05) insulin concentration when plasma was stored at 20 to 22 C. These results emphasize the importance of examining effects of sample collection and handling procedures on hormone stability and the danger of extrapolating results of such studies from one species to another and from one hormone to another.

Hyperlipemic serum and plasma samples often are received by clinical laboratories for endocrinologic analysis by radioimmunoassay. We designed a study to determine what effect, if any, hyperlipemia has on estimation of lipid-soluble hormone concentrations determined by solid-phase radioimmunoassays. Progesterone, testosterone, thyroxine, and cortisol concentrations were determined in canine plasma and serum with various degrees of lipemia. Samples of serum, heparinized plasma, and EDTA-treated plasma were obtained from blood collected from 4 female and 4 male Beagles by use of evacuated tubes. To induce hyperlipemia in vitro, IV fat emulsion was diluted in deionized water to produce 0 (water only), 33, 67, or 100% mixtures. Twenty microliters of each mixture then was added to the subsamples of serum and plasma from each dog. Hormone concentrations were determined, using validated radioimmunoassays. Triglyceride concentrations were determined by enzymatic assay. Addition of IV fat emulsion in vitro was an accurate and reproducible means of altering triglyceride concentrations in the samples. Triglyceride concentrations as high as 700 mg/dl had no effect on radioimmunoassays for progesterone, testosterone, and thyroxine in serum, heparinized plasma, or EDTA-treated plasma. Addition of 100% (but not 33 or 67%) fat emulsion reduced the mean cortisol concentration in heparinized plasma by 12% (P < 0.05). This severe hyperlipemia did not affect quantification of cortisol in serum or EDTA-treated plasma.