Diagnosis of hyperadrenocorticism (HAC) in the

Transcription

1 J Vet Intern Med 2012;26: Concentrations of Noncortisol Adrenal Steroids in Response to ACTH in Dogs with Adrenal-Dependent Hyperadrenocorticism, Pituitary-Dependent Hyperadrenocorticism, and Nonadrenal Illness W.E. Monroe, D.L. Panciera, and K.L. Zimmerman Background: Increases of adrenal hormone concentrations other than cortisol have been reported in dogs with hyperadrenocorticism (HAC). Hypothesis/Objectives: Measuring noncortisol adrenal hormone concentrations will help identify HAC in dogs. The objective was to determine plasma cortisol, androstenedione, estradiol, progesterone, testosterone, and 17-hydroxyprogesterone concentrations during ACTH stimulation testing of dogs with clinical signs of HAC to ascertain their utility in diagnosis of the disease. Animals: Ninety dogs with clinical findings consistent with HAC had ACTH stimulation tests performed. Results from 29 dogs were excluded from analysis because diagnoses were inconclusive for a variety of reasons. Results from 32 dogs with HAC and 29 dogs with disease other than HAC were analyzed. Methods: Prospective observational study. Concentrations of adrenocortical hormones were determined before and 1 hour after injecting 5 lg/kg ACTH IM. Diagnoses were determined by response to therapy, histopathology or both. Results: Post-ACTH cortisol (P <.001), progesterone (P =.001), and 17-hydroxyprogesterone (P <.001) concentrations were associated with a diagnosis of HAC. Sensitivity and specificity, respectively, for diagnosing HAC for post-acth cortisol were 84 and 59%, progesterone 88 and 55%, and 17-hydroxyprogesterone 91 and 59%, and for post- ACTH cortisol, progesterone and 17-hydroxyprogesterone combined were 88 and 55%. Of 5 dogs with HAC and normal post-acth cortisol concentrations, 5 had increased progesterone and 4 had increased 17-hydroxyprogesterone. Conclusions and Clinical Importance: Serum progesterone and 17-hydroxyprogesterone concentrations were useful to diagnose HAC in this study, but were not more sensitive or specific than cortisol concentration. Key words: 17-hydroxyprogesterone; Progesterone. Diagnosis of hyperadrenocorticism (HAC) in the dog is based on appropriate clinical findings, results of adrenal function tests, and response to treatment. The most common tests of adrenal function, ACTH stimulation, and low-dose dexamethasone suppression tests (LDDST) are not positive in all cases of HAC. Results of ACTH stimulation tests are diagnostic of HAC in 85 95% of dogs with pituitarydependent HAC (PDH) 1,2 and in only 33 60% of dogs with functional adrenal tumors (ADH). 1,3 6 The LDDST is positive in 85 96% of dogs determined to have HAC. 1,2,7,8 Cortisol is the major glucocorticoid hormone produced by most mammalian species. 9 The biosynthetic From the Department of Small Animal Clinical Sciences (Monroe, Panciera) and the Department of Biomedical Sciences and Pathobiology (Zimmerman), Virginia-Maryland Regional College of Veterinary Medicine, Virginia Tech, Blacksburg, VA. Clinical work was done at the Veterinary Teaching Hospital of the Virginia- Maryland Regional College of Veterinary Medicine, Virginia Tech, Blacksburg, VA. Endocrine assays were performed by the Clinical Endocrinology Service, College of Veterinary Medicine, University of Tennessee, Knoxville, TN. Preliminary data from this study were presented as an abstract at the Society for Comparative Endocrinology Biannual Meeting, June 2, 2003, Ashville, NC. Corresponding author: W.E. Monroe, Department of Small Animal Clinical Sciences, Virginia-Maryland Regional College of Veterinary Medicine, Virginia Tech, Blacksburg, VA 25061; monroewe@vt.edu. Submitted September 16, 2011; Revised April 22, 2012; Accepted May 9, Copyright 2012 by the American College of Veterinary Internal Medicine /j x Abbreviations: ADH adrenal-dependent hyperadrenocorticism LDDST low-dose dexamethasone suppression test HAC hyperadrenocorticism PDH pituitary-dependent hyperadrenocorticism ROC receiver operator characteristic curves UTCVM University of Tennessee College of Veterinary Medicine VTH Veterinary Teaching Hospital VMRCVM Virginia-Maryland Regional College of Veterinary Medicine pathway for the adrenocortical hormones, however, includes many intermediate and other hormones including pregnenolone, 17-hydroxypregnenolone, dehydroepiandrosterone, androstenedione, estradiol, progesterone, testosterone, and 17-hydroxyprogesterone 10 that may also have physiologic or pathophysiologic effects (Fig 1). Dogs with PDH and ADH often have increased plasma concentrations of adrenal steroids other than cortisol during dynamic adrenal function testing In some cases, increases in other adrenal steroids have occurred in the absence of increased plasma cortisol concentrations. 11,13,15 17 The most likely explanation for this observation is that adrenal neoplasia or hyperplasia results in defective activity of steroidogenic enzymes. For example, deficiency of the enzymes 21-hydroxylase or 11-hydroxylase, both of which are necessary for production of cortisol from 17-hydroxyprogesterone, would decrease plasma cortisol and increase plasma 17-hydroxyprogesterone concentrations (Fig 1). Indeed,

2 946 Monroe, Panciera, and Zimmerman Fig 1. Adrenocortical hormones and metabolic pathways. Enzymes are noted in bold print. DHEA = dehydroepiandrosterone, 3bHSD = 3b-hydroxysteroid dehydrogenase, P450arom = P450 aromatase (modified from Guyton 9 and Auchus 10 ). adrenal tumors in humans have been shown to have decreased 21-hydroxylase activity, resulting in increased 17-hydroxyprogesterone concentration within the tumor. 19 Increases in adrenal steroids other than cortisol therefore may account for cases of HAC that have negative results for adrenal function tests based on measurement of cortisol concentration. Dogs with nonadrenal diseases and clinical signs similar to those of patients with HAC may have cortisol concentrations during adrenal function testing consistent with a diagnosis of HAC. 1,2,8,20,21 This may result in misdiagnosis of HAC, leading to inappropriate treatment. Ideally, dogs with nonadrenal disease should not be tested for HAC until their other diseases have resolved. However, the clinical signs of HAC often overlap with those of other diseases. It therefore is important to evaluate the plasma concentrations of adrenal steroids other than cortisol during adrenal function testing of dogs with nonadrenal diseases to compare those results to the results of dogs with HAC. The purpose of this study was to determine the utility of measurement of serum concentrations of adrenal steroids, including cortisol, androstenedione, estradiol, progesterone, testosterone, and 17-hydroxyprogesterone during ACTH stimulation testing for diagnosis of HAC in dogs presenting with clinical findings consistent with HAC. The hypothesis of this study was that dogs with HAC have increased serum concentrations of noncortisol adrenal steroid hormones and that the concentrations of these hormones are increased in some cases when cortisol concentrations are normal or decreased. Materials and Methods All dogs presented serially to the Veterinary Teaching Hospital (VTH) of the Virginia-Maryland Regional College of Veterinary Medicine (VMRCVM), Virginia Tech, between June 2001 and September 2003 for which HAC was determined to be a primary differential diagnosis by 1 of 5 attending board-certified small animal internists or a 3rd year resident, did not have an alternative diagnosis made by routine laboratory testing, and that underwent ACTH stimulation testing as the initial test for HAC were evaluated. Dogs were excluded from the study (4 of 94 tested) after final case review by the authors if they did not exhibit 1 of the following clinical signs or routine laboratory abnormalities: history of polyuria, polydipsia, polyphagia, muscle weakness, hepatomegaly, increased abdominal girth, decreased appendicular muscle mass, hair coat changes such as alopecia or coarse texture, excessive panting, comedones, calcinosis cutis, and serum alkaline phosphatase (ALP) activity >2 times the upper limit of the reference interval. Dogs in which an adrenal mass was noted incidentally on ultrasonographic examination also were included. All dogs had routine laboratory assessment including serum biochemistry profiles, CBC, and urinalysis. An ACTH response test was performed after a 12-hour fast in all dogs. Blood samples were obtained immediately before and 1 hour after IM administration of 5 lg/kg synthetic ACTH. a22 Serum was harvested after the sample was allowed to clot for 30 minutes. Serum was stored at 20 C until transported for analysis by overnight courier. All assays were performed at the University of Tennessee College of Veterinary Medicine (UT- CVM) Clinical Endocrinology Service. Concentrations of cortisol, androstenedione, estradiol, progesterone, testosterone, and 17-hydroxyprogesterone were determined by previously validated radioimmunoassay. 23,24 Sex-specific reference intervals were established previously at the UTCVM Clinical Endocrinology Service Laboratory (Table 1). Unless contraindicated by a coexisting clinical problem, all study dogs had abdominal ultrasonography performed to aid in differentiating PDH from ADH, and to aid in the diagnosis of other diseases that could have mimicked HAC. In addition, some dogs with clinical and laboratory signs, along with results of ACTH stimulation testing consistent with the diagnosis of HAC, had a high dose dexamethasone suppression test or endogenous plasma ACTH concentration performed to differentiate ADH from PDH.

3 Noncortisol Adrenal Steroids 947 Table 1. Reference intervals; 2.5th to 97.5th percentile, established by sex from healthy dogs *24 Intact Male n = 20 Neutered Male n = 37 Intact Female n = 20 Spayed Female n = 36 Hormone Pre- Post- Pre- Post- Pre- Post- Pre- Post- Cortisol ng/ml Andro ng/ml Estradiol pg/ml Progest ng/ml OHprog ng/ml Testost ng/ml *The Clinical Endocrinology Service, College of Veterinary Medicine, University of Tennessee, results from a small number of healthy dogs were added since publication of reference 24 and the range represents the percentile; personal communication with the former service director, J. W. Oliver (deceased). N, the number of dogs in each group; 17OHprog, 17-hydroxyprogesterone; Testost, testosterone; Progest, progesterone; Andro, androstenedione. Follow-up on all cases was obtained by telephone interviews with the attending veterinarian or client or by examination at the VTH for at least 6 months. The final diagnosis of HAC was based on the presence of previously stated clinical and routine laboratory signs and the absence of other diseases based on routine laboratory results or other clinical findings, a positive response to treatment for HAC, histopathologic findings consistent with adrenocortical neoplasia or hyperplasia or pituitary neoplasia when tissues were available, and exclusion of other diseases that could account for the clinical findings. The final diagnosis of not HAC was based on a combination of clinical or routine laboratory signs, response to treatment, biopsy or necropsy results that confirmed a diagnosis other than HAC, and failure of clinical signs of HAC to progress during at least 6 months of follow-up. Because of the low number of dogs with ADH (3) or both forms of disease (1), all dogs with HAC were analyzed together as a single group. In addition, data were analyzed when dogs with adrenal tumors were excluded. Differences in hormone concentrations between groups (HAC versus dogs with other diseases) were evaluated using quantitative hormone concentration data normalized by division by central tendency 25,26 using the following formula: [result (midvalue of sex specific reference interval)]/[(high low value of sex specific reference interval)/2]. The data were assessed for normality using the Anderson-Darling test using a statistical software package. b Because the data were not normally distributed, Mood s median test (P <.05) was used to identify hormones having significantly different group median results. Data are presented as median and range unless otherwise stated. The clinical diagnostic utility of hormone concentrations identified as having significant differences between the 2 groups were assessed by calculating the sensitivity, specificity, and related confidence intervals using both qualitative results and the previously standardized quantitative data. The qualitative data were prepared by assigning a value of 0 if the specific hormone concentration was within the sex specific reference interval, 1 if below, and +1 if above. Receiver operator characteristic curves (ROC) were created 27 using both the quantitative and qualitative results and then used for calculating optimal sensitivity and specificity. Because the qualitative and quantitative sensitivity and specificity were similar (overlapping confidence intervals) and because clinicians more often use qualitative results (comparison to reference interval values) when making clinical decisions, only the qualitative sensitivity and specificity results are reported. Results ACTH stimulation tests were performed on 90 dogs presented consecutively with clinical, laboratory, and ultrasonographic abnormalities consistent with HAC. Twenty-nine dogs were excluded from data analysis because the diagnosis could not be definitively determined. Reasons for exclusion included dogs not treated for suspected HAC and adrenal and pituitary histopathology was either not obtained or was inconclusive, and insufficient information such as necropsy evidence to confirm that clinical signs were caused by a condition other than HAC. A final diagnosis of HAC was made in 32 dogs: 28 with pituitary dependent hyperadrenocorticism (PDH), 3 with adrenal dependent hyperadrenocorticism (ADH), and 1 with both an adrenal adenoma and multinodular adrenocortical hyperplasia associated with a pituitary carcinoma. All but 1 dog diagnosed with HAC had 2 clinical signs of the disorder. One dog diagnosed with HAC had only polyuria, but also had increased ALP activity. None of the dogs with HAC were identified for ACTH stimulation testing based solely on incidentally finding an adrenal mass on ultrasound examination. Twenty-five dogs with PDH had a good response with decreased cortisol concentrations and resolution of clinical signs in response to mitotane treatment at conventional loading dosages (25 50 mg/kg/day) followed by maintenance dosages ( mg/kg/week). Three dogs with PDH had the diagnosis confirmed histologically at necropsy. One of those had a partial response to l-deprenyl, and one had a good response to mitotane for several months until it was withdrawn several months before death and necropsy because of complicating disease. Diagnosis was confirmed by histopathology at necropsy for the dog with ADH and PDH. Three dogs with ADH were diagnosed by adrenal histopathology, one after reduction of the size of the adrenal mass and clinical signs with mitotane treatment. Disease other than HAC was diagnosed in 29 dogs from which hormonal results were utilized for data analysis. Diagnoses were based on clinical and routine laboratory signs, response to treatment, and biopsy or necropsy results. Although not analyzed statistically, this group seemed to have fewer clinical signs of HAC than the dogs that ultimately were diagnosed

4 948 Monroe, Panciera, and Zimmerman with HAC. Four dogs had only 1 clinical sign as well as increased ALP activity. Three dogs had incidental adrenal masses found on ultrasonography; one of these had increased ALP activity. One dog had only markedly increased ALP activity, and 1 dog had bilateral adrenomegaly noted on ultrasound examination, along with increased ALP activity. All of the other dogs in this group had at least 2 clinical signs of HAC, most also with increased ALP activity. Follow-up information was obtained for all but 3 dogs for at least 6 months and as long as 5 years to assure the dogs did not develop clinical evidence of HAC. The 3 dogs that were not followed died or were euthanized within days to several weeks of testing, and had necropsy or sufficient clinical and laboratory evidence to make a diagnosis of another disease that caused the clinical findings. Many different diseases were diagnosed. Ten dogs had urinary tract infection or some form of chronic renal disease, along with at least 1 other disease in most cases. Nine dogs had hepatic disease including 2 with metastatic neoplasia and 2 with concurrent bilateral adrenomegaly noted with abdominal ultrasonography. One dog with hepatic disease had been treated with l-deprenyl without response. The remaining 10 dogs had a variety of disorders including diabetes mellitus, central diabetes insipidus, pheochromocytoma, sudden acquired retinal degeneration syndrome, hypertension, and obesity. Thirteen of the dogs had more than 1 disease, including 3 with incidental adrenal masses noted on ultrasonography. Dogs with HAC had a median age of 11 years (range: 5 15 years). Their median body weight was 14.4 kg (range: kg). The dogs included 12 neutered males, 15 spayed females, 3 intact females, and 2 intact males. Breeds represented in the group with HAC included 9 mixed breeds, 4 Labrador Retrievers, 4 Shetland Sheepdogs, 3 Miniature Schnauzers, 3 Boston Terriers, and 1 each of American Eskimo, Brittany Spaniel, Chihuahua, Dachshund, Golden Retriever, Husky, Mastiff, Pekingese, and Fox Terrier. The median age of the dogs with illnesses other than HAC was 11 years (range: 2 14 years). This group of dogs had a median body weight of 24.1 kg (range: kg). Ten dogs were neutered males and 19 were spayed females. There were 10 mixed breed dogs, 3 Brittany Spaniels, 2 dogs each of the Beagle, Boxer, Labrador Retriever and Golden Retriever breeds, and 1 dog each of Cocker Spaniel, Doberman Pinscher, German Shepherd, Lhasa Apso, Shetland Sheepdog, Spitz, Standard Schnauzer, and Yorkshire Terrier breeds. Significant differences between those dogs with and without HAC were found for post-acth cortisol (P <.001), post-acth progesterone (P =.001), and post-acth 17-hydroxy- progesterone (P <.001). Although the difference between the affected and nonaffected groups was significant for pre-acth cortisol (P =.029), the specificity was very low (14%), and further analysis is not presented. Quantitative results for post-acth cortisol, progesterone, 17-hydroxyprogesterone, estradiol, and androstenedione for dogs with HAC compared to those with nonadrenal illness are presented as scatterplots (Fig 2). Sensitivity and specificity (95% confidence interval) for all dogs with HAC for the post-acth cortisol concentration were 84% ( ) and 59% ( ), respectively. Those same values for the post-acth progesterone concentration were 88% ( ) and 55% ( ), respectively. For the post-acth 17-hydroxyprogesterone, the values for sensitivity and specificity were 91% ( ) and 59% ( ), respectively. The sensitivity and specificity for making the diagnosis of HAC when post-acth cortisol, progesterone, and 17-hydroxyprogesterone all were increased were 88% ( ) and 55% ( ), respectively. Sensitivity and specificity (95% confidence interval) for the dogs with PDH for the post-acth cortisol concentration were 89% ( ) and 66% ( ), respectively. Those same values for the post-acth progesterone concentration were 86% ( ) and 55% ( ), respectively. For post-acth 17-hydroxyprogesterone the values for sensitivity and specificity were 89% ( ) and 59% ( ), respectively. The sensitivity and specificity for making the diagnosis of HAC (PDH) when post-acth cortisol, progesterone, and 17-hydroxyprogesterone all were increased were 86% ( ) and 55% ( ), respectively. Of the 29 dogs with diseases other than HAC, 19 (66%) had post-acth cortisol concentrations that were within the reference interval. Of these 19 dogs, 8 (42%) had high post-acth progesterone (median: 1.9; range: ng/ml), and 6 (32%) had an increased post-acth 17-hydroxyprogesterone concentration (median: 1.92; range: ng/ml). Of the 10 (34%) dogs with diseases other than HAC with increased post-acth cortisol concentrations, 5 had increased post-acth progesterone concentrations and 6 increased post-acth 17-hydroxyprogesterone concentrations. Five of the 32 dogs with HAC, 3 with PDH and 2 ADH, had post-acth cortisol concentrations within the reference interval. All 5 dogs had increased post-acth progesterone concentrations (median: 2.15; range: ng/ml) and 4 had increased post-acth 17-hydroxyprogesterone concentrations (median: 9.91; range: ng/ ml). Only 1 of the 5 dogs with increased post- ACTH progesterone concentrations, and 2 of the 4 dogs with increased post-acth 17-hydroxyprogesterone concentrations had results above the maximum value found in dogs with nonadrenal illness. From the ROC curves constructed using quantitative results, cut-off values for post-acth progesterone and 17-hydroxyprogesterone concentrations that would provide a diagnosis of HAC with a specificity of 100% were identified. Using those cut-off values, the sensitivity of progesterone and 17-hydroxyprogesterone for making the diagnosis of HAC would be 3%. Those cutoff values were approximately 3 times the upper limit

5 Noncortisol Adrenal Steroids 949 Fig 2. Scatterplots of post-acth steroid hormone concentrations of dogs with hyperadrenocorticism (HAC) versus nonadrenal illness (NAI) for androstenedione (median, range); HAC (46.4 ng/ml, ), NAI (26.3 ng/ml, ); cortisol; HAC (247 ng/ml, ), NAI (145.6 ng/ml, ); estradiol; HAC (58.5 pg/ml, ), NAI (60.4 pg/ml, ), progesterone; HAC (2.62 ng/ml, ), NAI (1.43 ng/ml, ), 17-hydroxyprogesterone; HAC (5.8 ng/ml, ), NAI (1.58, ); no dog or diagnostic group of dogs (PDHversus ADH) consistently had the outlier or extreme values. of the reference intervals for progesterone, and about 9 times the upper limits for 17-hydroxyprogesterone. None of the 5 dogs with HAC that had post-acth cortisol concentrations within the reference interval had progesterone concentrations that exceeded the cut-off, and only 1 of these 5 dogs had a 17-hydroxyprogesterone concentration that exceeded the cut-off. Of the 4 dogs with adrenal tumors, 2 had normal post-acth cortisol concentrations. All of these dogs had increased post-acth progesterone (median: 2.82; range: ng/ml) and post-acth 17-hydroxyprogesterone (median: 7.25; range: ng/ml) concentrations. Only 1 had a post-acth progesterone concentration, and none had post-acth 17-hydroxyprogesterone concentrations above the maximum concentration found in dogs with nonadrenal illness (Table 2). Discussion Of the adrenocortical steroid hormones measured before and after ACTH in this study, post-acth cortisol, progesterone, and 17-hydroxyprogesterone appear to be useful to identify the presence of HAC. Other investigators also have found increased 17-hydroxyprogesterone or progesterone concentrations or both in dogs with HAC, both with and without concurrent increases in cortisol concentration ,28 The sensitivity of the post-acth cortisol concentration was similar to that reported previously, 1,2 but the specificity in our study of post-acth cortisol concentration for making the diagnosis of HAC was lower than previously reported, 2 possibly as a result of a lower upper limit of the reference range as has been established by other laboratories. 1,2 Because of the relatively low specificity of increased post-acth progesterone and 17-hydroxyprogesterone concentrations found in the present study as well as in other studies, it seems most appropriate that noncortisol steroid hormones be measured only in cases where post-acth cortisol is within the reference range, but HAC remains the primary differential diagnosis. Approximately 16% of dogs with a final diagnosis of HAC in the present study would fall into this category. Post- ACTH serum progesterone concentration was increased in all 5 of the dogs, and 17-hydroxyprogesterone was increased in 4 of them. This finding seems to confirm the findings of Ristic et al, 16 where 17-hydroxyprogesterone was found to be a sensitive test for confirming HAC in dogs with normal cortisol response

6 950 Monroe, Panciera, and Zimmerman Table 2. Dogs with HAC (hyperadrenocorticism, all causes), ADH (adrenal tumors, includes 1 dog with an adrenal tumor and PDH) or NAI (nonadrenal illness) with post-acth cortisol (Cort) increased (Hi) or within or below the reference interval (N) that had increased post-acth concentrations of progesterone, 17OH-progesterone, androstenedione, or estradiol. Hormone values are expressed as (median, range) for 3 or more dogs, (range) for 2 dogs, and (result) for 1 dog. Progest ng/ml 17OH-progest ng/ml Androstene ng/ml Estradiol pg/ml HAC Cort Hi (27 dogs) 23 (2.66, ) 25 (5.86, ) 15 (64.2, ) 8 (87.9, ) HAC Cort N (5 dogs) 5 (2.15, ) 4 (9.91, ) 4 (76.1, ) 2 ( ) NAI Cort Hi (10 dogs) 5 (2.7, ) 6 (4.1, ) 6 (37, ) 5 (74.6, ) NAI Cort N (19 dogs) 8 (1.9, ( ) 6 (1.92, ) 5 (43.7, ) 4 (78.9, ) ADH Cort Hi (2 dogs) 2 ( ) 2 ( ) 1 (74.6) 0 ADH Cort N (2 dogs) 2 ( ) 2 ( ) 2 ( ) 1 (101.5) to ACTH administration. However, in our study the sensitivity and specificity of progesterone or 17-hydroxyprogesterone for confirming the diagnosis of HAC were no better than cortisol alone. Perhaps these and some of the other noncortisol steroids that were measured would have proven more useful had we been able to include a larger number of cases with adrenal tumors. 11,18 Poor specificity was noted for all hormones measured, including cortisol. This problem is common to all tests of adrenal function, emphasizing the need for careful selection of patients to test. Ideally, dogs should not have adrenal function testing until nonadrenal illness is resolved. Because the present study was comprised of cases that were suspected to have HAC, the results likely reflect the specificity in most clinical situations. It is difficult to compare the sensitivity and specificity of the tests evaluated in the current report with others because of different methods of setting reference intervals or cut-off values for interpretation. Despite this, other studies show a similarly low specificity of 17-hydroxyprogesterone in dogs with clinical signs of HAC. 17,28,c The specificity of progesterone has not been evaluated by others, to the authors knowledge. Of the 5 dogs with HAC that had post-acth cortisol concentrations within the reference interval, only 1 of the 5 with increased post-acth progesterone concentration, and 2 of the 4 dogs with increased post- ACTH 17-hydroxyprogesterone concentrations had results above the maximum value for dogs with nonadrenal illness. When ROC curves were constructed and cut-off values were set to provide for 100% specificity, only 1 of these 5 dogs had a 17-hydroxyprogesterone concentration and none of the 5 had a progesterone concentration that exceeded these cutoffs. This implies that there are not clinically useful cut-off values for progesterone and 17-hydroxyprogesterone concentrations that will reliably allow differentiation of HAC from nonadrenal illness. It remains unclear why excess secretion of so many adrenocortical hormones occurs commonly in dogs with HAC. Abnormal secretion of adrenal androgens and progestins has been documented to occur in humans with adrenocortical tumors, possibly because of deficiencies of specific enzymes. For example, deficiency of 21-hydroxylase, responsible for catalyzing the conversion of 17-hydroxyprogesterone to steroids eventually leading to cortisol, results in an increase in 17-hydroxyprogesterone. 19 Although our data could not confirm it because of a low number of dogs with ADH, alterations of steroidogenic enzymes may be more common in those dogs with adrenal tumors than in those with PDH. 11,13,15,18 However, most dogs in the present study had PDH. A relative deficiency of 21-hydroxylase, 11-hydroxylase, or both induced by adrenocortical hyperplasia might explain the noncortisol steroid hormone pattern of increases seen in those dogs with PDH that do not have an increased concentration of cortisol (Fig 1). It is not clear if dogs that have documented increases in progesterone or other steroid hormones without an increase in cortisol have a different disease than dogs that secrete excess cortisol. It seems likely that adrenocortical hyperplasia induced by excessive ACTH secretion results in abnormal steroidogenesis in some dogs and not in others. This is possibly the result of some dogs with PDH secreting a relatively greater amount of a proopiomelanocortin fragment that may have more influence than ACTH on adrenal sex steroid production as has been suggested in human patients with hyperandrogenism. 29 Additionally, because normal and neoplastic canine adrenal cortical tissues have been shown to contain cellular receptors other than ACTH receptors, such as luteinizing hormone (LH) receptors or gastric inhibitory peptide receptors, 30 ectopic expression or eutopic overexpression of such receptors in adrenocortical tissue may lead to overproduction of noncortisol steroid hormones in HAC. Excess LH in neutered dogs may preferentially stimulate areas of the adrenal cortex that produce sex steroids to become hyperplastic. Prolonged stimulation of cortical cells could lead to transformation of hyperplastic tissue into adenomas or adenocarcinomas that produce sex steroids autonomously. As shown in Table 1, the sex steroid hormones in the panels performed in this study have values in healthy dogs that vary by sex, with higher values in animals that are sexually intact. It is presumed that the source of this additional hormone is the gonads.

7 Noncortisol Adrenal Steroids 951 What effect this may have on the clinical utility of measuring these hormones for the diagnosis of HAC remains uncertain. The power of the study was not strong enough to answer this question because we had so few sexually intact dogs. Atypical or occult HAC has been described in dogs with historical, clinical, and routine laboratory findings of HAC, yet results of adrenal function testing (ie, ACTH stimulation and LDDST) are negative. 31,32 Atypical HAC has been reported in dogs with PDH and ADH. 11,13 18 Five of the dogs in the current study with HAC would be considered atypical; however, we did not routinely perform LDDST, which may have documented the presence of HAC. Because many of the dogs in this study with nonadrenal illness had increases in cortisol and noncortisol adrenal steroids, it would appear that any increase of ACTH (HAC or nonadrenal illness) that causes increases of adrenal steroid production and release is likely to cause increased cortisol and noncortisol steroid concentrations. Dogs with atypical HAC of pituitary origin may have excessive 24-hour cortisol secretion that is causing the clinical signs rather than or perhaps in conjunction with increases in noncortisol adrenal steroids, as has been suggested by others. 31 A recent study demonstrated that progesterone, 17-hydroxyprogesterone, estradiol, and androstenedione did not induce the expression of the gene for corticosteroid-induced alkaline phosphatase in canine hepatocytes cultured ex vivo. d Therefore, rather than a truly different syndrome, atypical HAC of pituitary origin in many cases may be the result of dogs with HAC that happen to be 1 of the 15% of PDH cases that have a normal cortisol at the time of the test. However, nearly all dogs with HAC are likely to have at least 1 adrenal function test (ie, ACTH stimulation or LDDST) that is positive, 1,6,33 and very few dogs with adrenal tumors have been reported to have a normal or negative LDDST. 11,18 Therefore, although we did not perform LDDST in the dogs of the present report, it is suggested that dogs suspected to have HAC that have normal cortisol concentrations after ACTH, routinely have LDDST performed before an ACTH stimulation test in which noncortisol adrenal steroids are measured. Because no gold standard for diagnosis of HAC exists, the authors chose to confirm the diagnosis based on clinical response to treatment and histopathologic findings. Although many dogs were eliminated from analysis because this information was not available, it resulted in a conservative selection of cases. Cases were enrolled consecutively as tested in our hospital, and the clinical findings that led to adrenal function testing were reviewed in each case to ensure that the dog had clinical findings consistent with HAC rather than diseases such as alopecia X. Determining that the dogs with other diseases did not in fact have HAC was problematic, and could have caused inaccuracies or introduced bias into our results because we did not require that they failed treatment for HAC. The criteria for ruling out HAC included the presence of clinical and routine laboratory signs, response to treatment, biopsy or necropsy results consistent with another diagnosis that explained the clinical presentation, and failing to progressively show additional signs of HAC within several months, and in many cases years of followup. We cannot be certain that some of these dogs did not have a more subtle, slowly progressive form of HAC such as has been reported in a dog with suspected food-dependent HAC, 34 but the method chosen was the only practical way to include all dogs commonly tested for the presence of HAC by clinicians. The results of this study support the measurement of progesterone and 17-hyroxyprogesterone for the diagnosis of HAC in dogs that have normal cortisol concentrations after ACTH administration. However, the frequent finding of increases in these hormones in dogs with nonadrenal illness makes their use unsuitable for routine diagnosis of HAC. In the current study, measuring these 2 adrenal steroids was not more sensitive or specific than merely determining cortisol concentration in a standard ACTH simulation test. The most appropriate use of the ACTH stimulation test in which a panel of noncortisol adrenal steroids is measured would be for the dog with suspected HAC for which results of a routine ACTH stimulation and LDDST have not confirmed the diagnosis. It should be kept in mind, however, that the specificity of an increased concentration of progesterone or 17- hydroxyprogesterone was not better than cortisol for making the diagnosis of HAC, and based on construction of ROC curves, there were not cut-off values above the limits of the reference intervals for these hormones that were clinically useful. Therefore, measuring these adrenocortical hormones may be more useful for ruling out HAC than for making a definitive diagnosis of atypical HAC. Footnotes a Cortrosyn, Amphastar Pharmaceuticals, Inc, Rancho Cucamonga, CA b Minitab 15 Statistical Software, Minitab Inc, State College, PA c Behrend EN, Kemppainen RJ, Kennis RA, et al. Evaluation of measurement of 17-hydroxy-progesterone and estradiol for diagnosis of typical and occult hyperadrenocorticism. J Vet Intern Med 2011; 25:681 (abstract) d Behrend EN, Kemppainen RJ, Kennis RA, et al. Assessment by quantitative PCR of ability of sex hormones to induce expression of classic glucocorticoid-induced genes in canine hepatocytes. J Vet Intern Med 2011; 25:681 (abstract) Acknowledgment The authors thank Dr Jack W. Oliver (posthumously) for his insight and interest in canine adrenal disease and for assay performance.

8 952 Monroe, Panciera, and Zimmerman References 1. Feldman E. Comparison of ACTH response and dexamethasone suppression as screening tests in canine hyperadrenocorticism. J Am Vet Med Assoc 1983;182: Van Liew CH, Greco DS. Comparison of results of adrenocorticotropic hormone stimulation and low-dose dexamethasone suppression tests with necropsy findings in dogs: 81 cases ( ). J Am Vet Med Assoc 1997;211: Feldman EC. Effect of functional adrenocortical tumors on plasma cortisol and corticotropin concentrations in dogs. J Am Vet Med Assoc 1981;178: Meijer JC, Lubberink A, Rijnberk A, et al. Adrenocortical function tests in dogs with hyperfunctioning adrenocortical tumours. J Endocrinol 1979;80: Peterson M, Gilbertson S, Drucker W. Plasma cortisol response to exogenous ACTH in 22 dogs with hyperadrenocorticism caused by adrenocortical neoplasia. J Am Vet Med Assoc 1982;180: Reusch CE, Feldman EC. Canine hyperadrenocorticism due to adrenocortical neoplasia: Pretreatment evaluation of 41 dogs. J Vet Intern Med 1991;5: Peterson ME. Hyperadrenocorticism. Vet Clin North Am Small Anim Pract 1984;14: Rijnberk A, van Wees A, Mol JA. Assessment of two tests for the diagnosis of canine hyperadrenocorticism. Vet Rec 1988;122: Guyton AC, Hall JE. Adrenocortical hormones. In: Textbook of Medical Physiology, 11th ed. Philadelphia, PA: Elsevier Saunders; 2006: Auchus RJ, Miller WL. The principles, pathways, and enzymes of human steroidogenesis. In: degroot LJ, Jameson JL, eds. Endocrinology. Philadelphia, PA: Elsevier Saunders; 2006: Norman EJ, Thompson H, Mooney CT. Dynamic adrenal function testing in eight dogs with hyperadrenocorticism associated with adrenocortical neoplasia. Vet Rec 1999;144: Frank L, Schmeitzel L, Oliver J. Steroidogenic response of adrenal tissues after administration of ACTH to dogs with hypercortisolemia. J Am Vet Med Assoc 2001;218: Hill KE, Scott-Moncrieff JC, Koshko MA, et al. Secretion of sex hormones in dogs with adrenal dysfunction. J Am Vet Med Assoc 2005;226: Behrend EN, Kemppainen RJ, Boozer AL, et al. Serum 17-alpha-hydroxyprogesterone and corticosterone concentrations in dogs with nonadrenal neoplasia and dogs with suspected hyperadrenocorticism. J Am Vet Med Assoc 2005;227: Benitah N, Feldman EC, Kass PH, et al. Evaluation of serum 17-hydroxyprogesterone concentration after administration of ACTH in dogs with hyperadrenocorticism. J Am Vet Med Assoc 2005;227: Ristic JM, Ramsey IK, Heath EM, et al. The use of 17- hydroxyprogesterone in the diagnosis of canine hyperadrenocorticism. J Vet Intern Med 2002;16: Chapman PS, Mooney CT, Ede J, et al. Evaluation of the basal and post-adrenocorticotrophic hormone serum concentrations of 17-hydroxyprogesterone for the diagnosis of hyperadrenocorticism in dogs. Vet Rec 2003;153: Syme HM, Scott-Moncrieff JC, Treadwell NG, et al. Hyperadrenocorticism associated with excessive sex hormone production by an adrenocortical tumor in two dogs. J Am Vet Med Assoc 2001;219: Racz K, Pinet F, Marton T, et al. Expression of steroidogenic enzyme messenger ribonucleic acid and corticosteroid production in aldosterone-producing and non-functioning adrenal adenomas. J Clin Endocrinol Metab 1993;77: Kaplan AJ, Peterson ME, Kemppainen RJ. Effects of disease on the results of diagnostic tests for use in detecting hyperadrenocorticism in dogs. J Am Vet Med Assoc 1995;207: Chastain C, Franklin R, Ganjam V, et al. Evaluation of the hypothalmic pituitary-adrenal axis in clinically stressed dogs. J Am Anim Hosp Assoc 1986;22: Behrend EN, Kemppainen RJ, Bruyette DS, et al. Intramuscular administration of a low dose of ACTH for ACTH stimulation testing in dogs. J Am Vet Med Assoc 2006;229: Schmeitzel L, Lothrop C. Hormonal abnormalities in Pomeranians with normal coat and in Pomeranians with growth hormone-responsive dermatosis. J Am Vet Med Assoc 1990;197: Frank LA, Rohrbach BW, Bailey EM, et al. Steroid hormone concentration profiles in healthy intact and neutered dogs before and after cosyntropin administration. Domest Anim Endocrinol 2003;24: Colantuoni C, Henry G, Zeger S, et al. Local mean normalization of microarray element signal intensities across an array surface: Quality control and correction of spatially systematic artifacts. Biotechniques 2002;32: Zimmerman KL, Panciera DL, Panciera RJ, et al. Hyperphosphatasemia and concurrent adrenal gland dysfunction in apparently healthy Scottish Terriers. J Am Vet Med Assoc 2010;237: Witten IH, Frank E. Data Mining: Practical Machine Learning Tools and Techniques, 2nd ed. Boston: Morgan Kaufman; Sieber-Ruckstuhl NS, Boretti FS, Wenger M, et al. Evaluation of cortisol precursors for the diagnosis of pituitary-dependent hypercortisolism in dogs. Vet Rec 2008;162: McKenna TJ, Cunningham SK. The pathogenesis of adrenal and extra adrenal hyperandrogenism. J Steroid Biochem Mol Biol 1993;45: Galac S, Kars VJ, Klarenbeek S, et al. Expression of receptors for luteinizing hormone, gastric-inhibitory polypeptide, and vasopressin in normal adrenal glands and cortisol-secreting adrenocortical tumors in dogs. Domest Anim Endocrinol 2010;39: Scott-Moncrieff JC. Atypical and subclinical hyperadrenocorticism. In: Bonagura JD, Twedt DC, eds. Kirk s Current Veterinary Therapy, 14th ed. St. Louis, MO: Saunders Elsevier; 2009: Behrend EN, Kennis R. Atypical Cushing s syndrome in dogs: Arguments for and against. Vet Clin North Am Small Anim Pract 2010;40: Feldman E, Nelson R, Feldman M. Use of low- and highdose dexamethasone tests for distinguishing pituitary-dependent from adrenal tumor hyperadrenocorticism in dogs. J Am Vet Med Assoc 1996;209: Galac S, Kars VJ, Voorhout G, et al. ACTH-independent hyperadrenocorticism due to food-dependent hypercortisolemia in a dog: A case report. Vet J 2008;177: