Introduction

Cushing’s syndrome refers to a collection of symptoms that develop from a chronic overexposure to cortisol.¹ The most common causes in dogs are a pituitary tumor secreting adrenocorticotropic hormone (ACTH; pituitary-dependent hyperadrenocorticism [HAC]), an autonomous functioning adrenocortical tumor (adrenal-dependent HAC), or iatrogenic.¹ The suspicion of HAC is typically based on history and physical examination. Clinical manifestations of HAC include polydipsia, polyuria, polyphagia, abdominal distention, alopecia, excessive panting, and muscle weakness. These signs are the results of the gluconeogenic, lipolytic, protein catabolic, anti-inflammatory, and immunosuppressive effects of glucocorticoid hormones.² Because adrenal function tests lack specificity, such investigations should be performed only in dogs for whom there is an indication to pursue a diagnosis of HAC, in order to maximize the positive predictive value of these tests.^3–17 Recently, the American College of Veterinary Internal Medicine published a consensus on diagnosis of spontaneous canine HAC, in which the presence of compatible clinical signs (CSs) is emphasized as being the primary indication for pursuing a diagnosis of HAC.¹⁸

Previous studies have reported the prevalence of various CSs in canine HAC.^3,19,20 These studies help to define signs that clinicians should use to identify dogs to be tested for HAC. Nonetheless, the concern has been raised that, because of an increased awareness, dogs may be currently evaluated at earlier stages of disease development.^2,18 Therefore, the prevalence of individual CSs may be becoming lower than in studies conducted decades ago.

It has been suggested that large breeds of dogs with HAC appear to show subtler clinical manifestations than smaller breeds.^1,21 If this observation is true, then the positive predictive value of ACTH testing would be higher if the dog is from a large breed compared with a dog from a smaller breed showing the same number of compatible CSs. There are no studies comparing the clinical manifestations of HAC between large and smaller dogs.

The aim of this study was to investigate clinical and clinicopathological abnormalities in canine spontaneous HAC in a population of dogs more recently diagnosed in a primary care practice setting. It was hypothesized that (1) the dog’s weight is negatively correlated to the number of clinical features in spontaneous HAC, (2) the dog’s weight is negatively correlated to the severity of clinicopathological abnormalities in spontaneous HAC, and (3) prevalences of CSs in dogs with spontaneous HAC are lower than reported in earlier studies.

Materials and Methods

Results

Respondents and Study Population

Between December 2013 and November 2014, 124 responses were collected. Among dogs assessed for eligibility, 62 dogs fulfilled all the inclusion criteria and none of the exclusion criteria (Figure 1). The included cases were diagnosed a median of 5 (0–108) mo prior to questionnaire submission with 43/62 (69%) evaluated during the previous 12 mo. Respondents had a median experience as veterinarian of 9 (0–34) yr.

View Figure

• Download as Powerpoint
• Download Figure

Mean age of dogs included in the study was 10.3 (± 2.1) yr, and median body weight was 14.8 (3–39) kg. The weight of one dog was not properly reported. This dog was included in the analyses that did not involve weight. There were 29 spayed females, 19 castrated males, 10 intact males, and 4 intact females (Table 1). Breeds in the group of dogs ≤20 kg included mixed-breed (n = 4), Staffordshire bull terrier (n = 3), Scottish terrier (n = 3), English cocker spaniel (n = 3), dachshund (n = 3), bichon frise (n = 3), beagle (n = 2), Boston terrier (n = 2), Jack Russell terrier (n = 2), Pomeranian (n = 2), shih tzu (n = 2), Yorkshire terrier (n = 2), and one each of American Cocker spaniel, Australian shepherd, border collie, border terrier, Cairn terrier, Cavalier King Charles spaniel, Maltese, miniature poodle, and poodle. Breeds in the group of dogs >20 kg included mixed-breed (n = 7), boxer (n = 3), bearded collie (n = 2), border collie (n = 2), and one each of American cocker spaniel, basset hound, German shorthaired pointer, golden retriever, Labrador retriever, Lurcher, and Staffordshire bull terrier.

TABLE 1 Signalement of the Dogs Included in the Two Groups of the Study

View Fullscreen

Results of adrenal function tests included only an ACTH stimulation test consistent with spontaneous HAC in 29 cases (47%), only a LDDST consistent with spontaneous HAC in 20 cases (32%), an ACTH and a LDDST consistent with spontaneous HAC in 10 cases (16%), and an ACTH stimulation test not consistent but a LDDST consistent with HAC in three cases (5%).

Presenting Complaint and Prevalence of Individual Clinical Signs

A median of 18 (10–19) signs were assessed and 6.2 (± 2.9) were present for each animal. Owners’ presenting complaints included increased urination and/or water intake (n = 44, 71%), polyphagia (n = 22, 35%), abdominal enlargement (n = 20, 32%), dermatological signs (n = 8, 13%), and exercise intolerance (n = 7, 11%). The individual prevalences of various CSs are reported in Table 2.

TABLE 2 Prevalence of Various Clinical Signs in Dogs Included in This Study Calculated as (Number of Dogs in Whom the Clinical Sign was Observed) / (Number of Dogs in Whom the Clinical Sign was Assessed)

View Fullscreen

Influence of Weight on Clinical and Clinicopathological Findings

The variables CSR, cholesterol concentration, and platelet count were normally distributed, whereas ALKP, ALT, BUN, USG, and UPCR were not. The mean CSR for all dogs included in the study was 0.36 (± 0.17). The mean CSRs in dogs ≤20 kg and dogs >20 kg were, respectively, 0.37 (± 0.18) and 0.34 (± 0.16). Results showed no correlation between the weight and the CSR (P = .66). There was no statistically significant difference between dogs ≤20 kg (n = 40) and dogs >20 kg (n = 21) for the CSR (P = .47) or any of the individual CSs assessed (Table 3). Because a quantification of the water intake was only provided for three dogs ≤20 kg (90, 100, and 150 mL/kg/day), the severity of the polydipsia could not be compared between groups.

TABLE 3 Data Collected in Dogs ≤20 kg (n = 40) and Dogs >20 kg (n = 21) Regarding the Clinical Signs

View Fullscreen

Results did not show any significant correlation between the weight and ALKP activity (P = .94), ALT activity (P = .96), BUN (P = .84), USG (P = .39), UPCR (P = .2), and cholesterol concentration (P = .81). There was a significant negative correlation between the weight and the platelet count (P = .005, r² = 0.3). The platelet count was lower in dogs >20 kg compared with dogs ≤20 kg as shown in Table 4.

TABLE 4 Median (Range) and Mean (± Standard Deviation) of Clinicopathological Parameters of Dogs Included in the Study

View Fullscreen

There was no correlation between the CSR and BUN concentration (P = .11), USG (P = .77), UPCR (P = .34), cholesterol concentration (P = .57), and platelet count (P = .28). There were weak positive correlations between ALKP activity and the CSR (P = .05, r² = 0.001) and ALT activity and the CSR (P = .02, r² = 0.17).

Discussion

It has been suggested that large-breed dogs with HAC show subtler clinical manifestations than smaller breeds.^1,21 It has also been suggested that dogs may now be presented at earlier stages of disease development.^2,18 Therefore, we hypothesized that the prevalence of individual CSs may be lower than in studies conducted decades ago.

The results of this study did not support there being a difference between heavier and lighter dogs in clinical features of canine spontaneous HAC. The weight was not correlated to the CSR, and there was no statistical difference in the prevalence of individual CSs of HAC between dogs ≤20 kg and dogs >20 kg. It is possible that the lack of significant findings in this study is attributable to sample size and type II error. Post hoc power calculations were performed based on the mean CSR for small dogs of 0.37 ± 0.18 (powerandsamplesize.com). Based on this data and the sample size in this study, we would have been able to detect a significant difference, if one were present, and the mean CSR of the group of dogs >20 kg was ≤0.24 with a power of 80% and α of 5%. This equates to a mean of 7.4 CSs present in dogs ≤20 kg and 4.8 in dogs >20 kg. Although it is plausible that smaller differences than this are present, their presence would be of questionable clinical significance, and the sample size required to demonstrate them would be prohibitively large.

Weight has been used in other studies as a parameter to differentiate small and larger dogs, and it is reasonable to expect that this variable should be highly correlated to the size.^22–25 However, the body condition score was not taken into consideration in this study, as many clinicians were involved in the assessment of these dogs, and it was felt that this was less likely to provide a consistent assessment method when compared with weight. Consequently, dogs from the group >20 kg may not all be of large size and dogs in the group ≤20 kg may not all be of small size. Indeed, one American cocker spaniel and one Staffordshire bull terrier were included in the group of dogs >20 kg, but these breeds may be considered of small-to-medium size. Additionally, one Australian shepherd was included in the group of dogs weighing ≤20 kg, but this breed may be considered of medium-to-large size. However, apart from these three dogs, the groups of dogs ≤20 kg and >20 kg appear to reflect a population of dogs of small and large size, respectively.

The CSR was used to quantify the number of CSs. However, the difference between the clinical presentation of HAC in dogs >20 kg compared with dogs ≤20 kg may lie in the severity of the clinical features, not their number. Severity was not assessed in this study, as most assessments of this are subjective, and it would be difficult to assess severity using methods described here. If the hypothesis is correct that heavier dogs show subtler clinical features of HAC, then reaching a diagnosis would be likely to be less frequent than in lighter dogs because of a lower suspicion. This would create an inherent bias in any study resulting from the inclusion of a higher number of heavier dogs showing more clinical features. Prospective studies involving adrenal function testing in a larger proportion of potentially compatible cases would be required to address such a potential selection bias.

The only clinical parameter that was negatively correlated with weight was the platelet count. This is the only finding from our study supporting that clinicopathological abnormalities in dogs with HAC are less severe in heavier dogs. The platelet count method was not requested in the questionnaire. Automated platelet counts are not always reliable because of platelet clumping.²⁶ Additionally, the platelet count was reported only in a small number of dogs in this study. Based on the low r² value and visual inspection of the scatter plot, the correlation between weight and platelet count seems poor. For these reasons, the clinical relevance of the correlation between weight and platelet count identified in this study is questionable.

A recent study performed in primary care practice reported lower prevalence rates of common CSs compared with earlier reports.^3,19,20,27 However, this study was specifically evaluating survival of untreated dogs and may have been biased toward mildly affected cases. The prevalence of the individual CSs reported in our study are similar to the ones reported decades ago and to a recent study performed in a referral population.^3,19,20,28 However, the prevalence of polyphagia was slightly higher than previously reported (Table 3). Alopecia in this study was the only CS reported less frequently compared with earlier studies (42.6%; 95% confidence interval: 31–55.1); previous studies reported a prevalence of 63 and 74%.^3,19,20 These findings do not support the hypothesis that clinical manifestations in canine HAC are more subtle nowadays compared with several decades ago.^2,18 Some prevalence rates reported in this study (i.e., nocturia, recurrent pyoderma, recurrent urinary tract infection, suspected ligament rupture, pseudomyotonia, and anorexia) have not been previously reported. The prevalence of clinicopathological abnormalities could not be compared with the ones reported in previous studies because different assays with different methods of calculating reference intervals, sensitivities, and reliability would have been used in the various studies and the variations in how the data was reported.

ALKP and ALT activities were positively correlated to the CSR. The increase in ALKP activity in dogs with HAC is primarily caused by the steroid-induced isoenzyme.² The increase in ALT activity is thought to be secondary to damage caused by swollen hepatocytes, glycogen accumulation, and interference with hepatic blood flow.² The cause of this relationship between ALKP and ALT activity and the CSR is unclear, but both may be a consequence of exposure to higher cortisol concentration or to the longer duration of overexposure to endogenous cortisol.^2,29,30 The correlation between ALKP and ALT activities with the CSR is poor as shown by the low r² values. Therefore, ALT and ALKP activities should not be used to take clinical decisions following diagnosis of HAC. Additionally, the interpretation of these findings is limited by the fact that different assays/analyzers were likely used to measure liver enzymes activities in dogs included in this study. Although a variety of assays and laboratories were used to assess these parameters, there is no apparent reason to believe that this would have introduced any bias into the results relating to patient size or CSR; rather, it would have introduced more noise. As such, the persistence of a significant correlation may have been more pronounced had these assays been standardized. Interestingly, other serum biochemistry parameters in dogs have recently been compared internationally between reference laboratories and established in those instances that interlaboratory variation was acceptable, but reference ranges based on locally selected populations varied dramatically.³¹ Although it may have been possible to use the magnitude of change from the high end of the reference intervals instead of absolute values as an alternative assessment of ALT/ALKP elevation, it is not clear whether this would have been a superior method of analysis. ALKP activity is sometimes used as a screening test for HAC.³² Dogs with lower ALKP activity and, therefore, potentially a lower number of clinical manifestations may not have been tested for HAC, which may be a reason why the prevalences of CSs reported in this study are not lower than several decades ago.

One of the limitations of this study is that CSs were reported by many different primary care veterinarians, and some clinical features may have been missed or not recorded. The data set would perhaps have been more complete if we had focused on cases seen in referral hospitals, in which such data may be more consistently recorded. However, cases seen in referral hospitals may not reflect the overall population of dogs with HAC because of more frequent unusual presentations or less typical clinicopathological abnormalities. However, there have been no studies addressing this point, and most of the previously published studies into this subject have used cases seen at secondary or tertiary referral hospitals. Additionally, the presence or absence of a specific referral department (e.g., dermatology) could have resulted in overestimation or underestimation, respectively, of system specific signs. The authors believe that studying the clinical or clinicopathological characteristics only in cases seen in primary care practices better reflects the overall population of dogs with HAC. Nonetheless, the decision to investigate a primary care population may account for the lack of agreement with statements made by those working in referral centers regarding the prevalence of CSs and associations with patient size.

The diagnosis of HAC should be based on a combination of CSs, clinicopathological abnormalities, and adrenal function tests.¹⁸ As there is no gold standard for diagnosis of HAC, and notably because adrenal function tests lack specificity, some dogs incorrectly diagnosed with HAC might have been included in this study.^3–17 Notably, a minimal number of CSs was not necessary to include dogs in the study, as it would have resulted in a selection bias. However, the inclusion and exclusion criteria aimed to avoid the inclusion of incorrectly diagnosed dogs in the final data set, and the signalment (i.e., age, gender distribution) of the dogs included in this study, similar to that reported in other studies about HAC, suggests that we included a similar population.^3,20,33–38 Although most breeds diagnosed with HAC in this study have been reported in previous studies on canine spontaneous HAC, the relative higher prevalence of Staffordshire bull terriers and border collies is unusual.^3,20,33–38 Most of these earlier studies were performed on populations of dogs diagnosed in the United States. A recent epidemiological study on dogs diagnosed with HAC in primary care practice in the United Kingdom reported a similar proportion of Staffordshire bull terriers compared with the present study and some border collies (although in a smaller proportion).³⁸ This difference in the proportion of these specific breeds may be a result of a higher popularity in the United Kingdom, which is supported by the fact that all Staffordshire bull terriers and border collies included in our study were diagnosed in the United Kingdom.^39,40 Other possible causes include different genetic background in dogs in the United Kingdom, a possible different breed distribution of dogs with HAC in primary care practice compared with referral practices, or a change in breed predisposition over the last decades.

Only dogs with an ACTH stimulation test and/or LDDST consistent with HAC were included in this study. It has been acknowledged that the reference intervals and cut-off values for adrenal function tests should be re-established, and therefore, some dogs with HAC but adrenal function tests results not consistent with HAC might have been excluded from the analysis.¹⁸ There is no data supporting that these cases have different clinical features compared with cases with adrenal function tests consistent with HAC. In fact, there was no association between the post-ACTH cortisol concentration or the 8 hr LDDST cortisol concentration and a clinical score based on number and severity of CSs in a recent study on canine HAC.⁴¹ Additionally, the prevalences of individual CSs appear to be similar in dogs diagnosed with HAC despite normal adrenal function tests compared with dogs with traditionally diagnosed pituitary-dependent HAC in another study.⁴²

Other limitations include the absence of dogs from giant breeds and the small sample size. Additionally, some cases were excluded based on the presence of some CSs that were not historically associated with HAC. This may have biased the comparison of the clinical characteristics with earlier studies, as new clinical manifestations of HAC may not have been identified. However, it was thought that including such cases would have resulted in an unacceptable risk of including dogs with nonadrenal illnesses, either with comorbidities to HAC or in some instances with a disease causing false-positive testing for HAC. It is possible that some of these CSs may have been due to diseases that occur secondary to HAC such as vomiting during a bout of acute pancreatitis.^43,44 Because adrenal function tests are discouraged during such acute problems, the decision was made to exclude cases with such CSs at the time of apparent diagnosis, instead focusing on CSs that may or may not be present during a more chronic phase of disease manifestation, when testing is typically recommended.

Conclusion

In this study, the range of clinical manifestations and clinicopathological abnormalities was not significantly different between dogs ≤20 kg and dogs >20 kg. The prevalence of various clinical findings appears to be similar to those historically reported. This observation does not support the idea that cases of HAC are being detected at earlier stages of disease development, although the severity of clinical features was not specifically evaluated in this study. Dogs with more CSs tend to have higher ALT and ALKP activities.