J Vet Intern Med 2007;21:1168–1173

Abnormalities of Serum Electrolyte Concentrations in Dogs with Hypoadrenocorticism Jennifer A. Adler, Kenneth J. Drobatz, and Rebecka S. Hess Background: The sensitivity and specificity of the sodium to potassium ratio (Na:K ratio) as a cutoff for recommendation of an adrenocorticotropic hormone (ACTH) stimulation test in dogs suspected of having hypoadrenocorticism (HA) is unknown. Additionally, abnormalities in plasma ionized calcium (iCa2+) and ionized magnesium (iMg2+) concentrations and venous pH of dogs with HA are incompletely documented. Objectives: To define the sensitivity and specificity of the Na:K ratio as a diagnostic aid for HA in dogs and to examine for associations between venous pH and the Na:K ratio, iCa2+ concentration, or iMg2+ concentration in dogs with HA. Animals: Seventy-six dogs with HA and 200 dogs randomly selected from the general hospital population. Methods: Retrospective study. Dogs were included in the study if results of an ACTH stimulation test confirmed a diagnosis of HA, the dog had a serum sodium concentration below the reference range or a serum potassium concentration above the reference range, and the dog was treated with mineralocorticoids. Receiver operating curve analysis was used to determine optimal cutoffs of sensitivity and specificity for the Na:K ratio in diagnosing HA. Results: Use of Na:K ratios of 27 or 28 classified 95% of dogs correctly as diseased or not diseased. The sensitivity of a Na:K ratio of 28 was 93% (CI, 85–98%) and that of 27 was 89% (CI, 80–95%). The specificity of a Na:K ratio of 28 was 96% (CI, 92– 98%) and that of 27 was 97% (CI, 93–99%). The sensitivity and specificity of a Na:K ratio of 24 were 79% (95% CI, 67–86%) and 100% (98% CI, 97%–100%), respectively. Conclusions and Clinical Importance: Na:K ratios of 27 or 28 identify the highest percentage of dogs with suspected mineralocorticoid and glucocorticoid deficiency correctly. In dogs with a Na:K ratio of 24 or less, the likelihood of confirming a diagnosis of HA with an ACTH stimulation test is high. Key words: Acidosis; Addison’s disease; Glucocorticoid; Mineralocorticoid; Na:K ratio.

aturally occurring hypoadrenocorticism (HA) in dogs is a well-described endocrinopathy characterized by lack of cortisol and aldosterone secretion from the adrenal glands.1–15 Aldosterone promotes distal renal tubular reabsorption of sodium and chloride and facilitates secretion of potassium and hydrogen ions.16 Therefore, lack of aldosterone leads to hyponatremia and hyperkalemia.16 Anecdotally, a sodium to potassium ratio (Na:K ratio) of 27 has been suggested as a cutoff ratio or screening tool for detecting HA, and an adrenocorticotropic hormone (ACTH) stimulation test has been recommended in dogs with a Na:K ratio of 27 or less,1–3,7–9,17–19 but this has not been specifically evaluated. One of the goals of this study was to assess the sensitivity and specificity of the Na:K ratio as a diagnostic aid for HA in dogs and to determine which Na:K ratio is most useful as a cutoff ratio below which an ACTH stimulation test should be recommended. Sodium and potassium concentrations are readily and inexpensively measured. Knowledge of the sensitivity and specificity of the Na:K ratio as a diagnostic aid for HA in dogs would provide clinicians with a rapid and easy screening test that could be of therapeutic importance while awaiting results of an ACTH stimulation test or other test results. Another goal of this study was to investigate a possible association between venous pH and the Na:K ratio,

N

From the Department of Clinical Studies, Matthew J. Ryan Veterinary Hospital, School of Veterinary Medicine, University of Pennsylvania, Philadelphia. Reprint requests: Rebecka S. Hess, 3850 Delancey St, Philadelphia, PA 19104; e-mail: [email protected]. Submitted November 10, 2006; Revised February 20, 2007, May 4, 2007; Accepted June 21, 2007. Copyright E 2007 by the American College of Veterinary Internal Medicine 0891-6640/07/2106-0001/$3.00/0

serum creatinine concentration, plasma ionized calcium (iCa2+) concentration, or ionized magnesium (iMg2+) concentration. Although elevations in total calcium concentration have been previously reported in dogs with HA,1–3,5,9,11 plasma iCa2+ and iMg2+ concentrations, and venous pH have not been previously reported in a large group of dogs with HA. Establishing whether venous pH and the Na:K ratio, serum creatinine concentration, iCa2+ concentration, or iMg2+ concentration are associated with one another could improve understanding of the pathophysiology of electrolyte disturbances in dogs with HA.

Materials and Methods Criteria for Selecting Cases A computer search was performed of all dogs admitted to the Matthew J. Ryan Veterinary Hospital of the University of Pennsylvania (MJR-VHUP) between January 1997 and December 2005 with a coded medical diagnosis of HA. Medical records were reviewed in detail by one of the authors (JAA). Dogs were included in the study if results of an ACTH stimulation test confirmed a diagnosis of HA and the dog had a serum sodium concentration below the reference range or a serum potassium concentration above the reference range on initial blood gas analysis or biochemical profile performed at MJR-VHUP or the referral biochemical profile performed just before referral. Only dogs treated with mineralocorticoids by the time of discharge from the hospital or at the time of follow-up were included in the analysis. Results of an ACTH stimulation test were considered consistent with a diagnosis of HA if serum cortisol concentrations before and after ACTH stimulation were #1.0 mg/dL.17 Serum cortisol concentration was measured before, and 1 or 2 hours after, IV or IM administration of synthetic ACTH, respectively.a–c An ACTH stimulation test was performed at MJR-VHUP for 75 dogs, and for 1 dog results of an ACTH stimulation testd performed immediately before referral were used for the purpose of a diagnosis. Dogs were excluded from the study if a diagnosis of HA had been previously established, HA had been induced iatrogenically, the record was miscoded and the dog did not really have HA, cortisol

Hypoadrenocorticism in Dogs concentrations before or after ACTH stimulation were .1.0 mg/dL, the complete medical record was not available for review, serum sodium and potassium concentrations were within reference range on initial blood gas analysis and biochemical profile performed at MJR-VHUP and on referral biochemical profile, or if the dog was not treated with mineralocorticoid supplementation during hospitalization or at the time of the last follow up.

Procedures Signalment, clinicopathologic findings, endocrine test results, abdominal ultrasonographic findings, and outcome were recorded. Findings are reported from the time of initial examination at MJRVHUP. Testing was performed at MJR-VHUP, unless otherwise noted.

Clinicopathologic Findings. Na and K concentration from the biochemical profilee and data reported by a venous blood gas analyzer,f including venous pH, lactate concentration, plasma iCa2+ concentration, plasma iMg2+ concentration, and base deficit, were recorded in most dogs included in the study. For calculating the Na:K ratio, the lowest Na:K ratio for each dog was recorded from the initial biochemical profile or venous blood gas analysis performed at MJR-VHUP. If these values were clinically normal, referral serum sodium and potassium concentrations were used for the calculation. For determining associations between venous pH and plasma iCa2+ concentration, plasma iMg2+ concentration, or the Na:K ratio, all values were obtained from the initial venous blood gas analysis performed at MJR-VHUP. For determining associations between plasma iCa2+ concentration, plasma iMg2+ concentration, and serum creatinine concentration, the serum creatinine concentration was obtained from initial biochemical profile performed at MJR-VHUP. Endocrine Testing. The use of automated chemiluminescent enzyme immunoassayg for measuring cortisol in dogs has been previously validated.20 One of three protocols was used for ACTH stimulation, depending on the product being used at the time each dog was admitted to MJR-VHUP. An ACTH gela was administered at a dose of 2.2 U/kg IM, and cortisolg was measured before and 2 hours after ACTH injection, as previously reported.17 In a different protocol, synthetic ACTHb was administered at a dose of 0.25 mg/ dog IV, and cortisolg was measured before and 1 hour after ACTH stimulation.17 Most recently, synthetic ACTHc was administered at a dose 125 mg for dogs with a body weight less than 15 kg or 250 mg for dogs with a body weight of 15 kg or more IV, and cortisolg was measured before and 1 hour after ACTH stimulation.21 Follow-up. Follow-up for clinicopathologic findings and outcome was recorded from the time of the last recheck of each dog at MJR-VHUP. Imaging. All imaging studies were performed at MJR-VHUP, and results were reviewed by a board-certified radiologist at the time of hospitalization. The left adrenal gland was defined as slender if its width was ,3.0 mm and as normal if its width was 3.0–5.2 mm. The right adrenal gland was defined as slender if its width was ,3.1 mm and as normal if its width was 3.1–6.0 mm.14 Comparison Dogs. Two hundred dogs were randomly selected from the population of all other dogs examined at the MJR-VHUP between June 2004 and June 2005. This group served to compare the signalment and serum sodium and potassium concentrations in the general hospital population with the signalment and serum sodium and potassium concentrations in the study dogs with naturally occurring HA.

1169

Data analysis. Median, minimum, and maximum were used to describe nonparametric variables. Unpaired t-test or Wilcoxon rank sum test was used to compare continuous variables among groups depending on whether data had a Gaussian distribution, respectively. Categoric variables were described with frequencies, proportions, or percentages. If the expected cell count in a 2 3 2 table was ,5, Fisher’s exact test was used to compare categoric variables between groups, otherwise Pearson’s x2 test was used. Correlations between continuous variables were evaluated by visual inspection of scatter plots and Spearman rank correlation or Pearson’s correlation, depending on whether the variables were nonparametric or parametric, respectively. When sensitivity and specificity were calculated for a certain Na:K ratio, the sensitivity and specificity describe the characteristics of this Na:K ratio cutoff for all dogs with a Na:K ratio less than the cutoff. Optimal cutoffs of sensitivity and specificity for the Na:K ratio in diagnosing HA were determined by receiver operating characteristic curve analysis. Ninety-five percent confidence intervals (CIs) were calculated for proportions that described natural characteristics of the study population (not for treatment decisions that clinicians made) using the binomial exact method; 97.5 CIs were generated when the proportion was 100%. CIs for continuous variables were not generated as all of the reported continuous variables were not normally distributed. For all comparisons, P , .05 was considered significant. The 200 comparison dogs were used to determine signalment characteristics and sensitivity and specificity of the Na:K ratio. All statistical analyses were performed using a statistical software package.h

Results The computer search identified 256 dogs with a coded medical diagnosis of HA. However, upon review of medical records, only 76 dogs met the criteria for inclusion in the study.

Comparison Dogs Among 200 randomly selected comparison dogs, the primary reasons for examination were neoplasia (62 dogs); orthopedic disease (18 dogs); gastrointestinal disease (15 dogs); endocrine disease (14 dogs, not including HA); urinary tract disease (11 dogs); dental disease (10 dogs); infectious disease (9 dogs); neurologic disease (8 dogs); cardiac, dermatologic, or behavioral disease (7 dogs each); immune-mediated disease (6 dogs); trauma or ophthalmologic disease (5 dogs each); respiratory disease (4 dogs); or other less represented conditions (12 dogs). Hypoadrenocorticism was excluded on the basis of results of a normal ACTH stimulation test in 4 of the 200 dogs.

Signalment and Historical Findings Median duration of clinical signs before examination was 10 days (range, 2–450 days). The median age of dogs with HA was 3.2 years (range, 2 months to 11.6 years). Of the 76 dogs, 6 (8%) were ,1 year old and 2 (3%) were .10 years old. Median age of dogs with HA was significantly lower than that of the 200 comparison dogs (8 years, range, 1–19 years; P , .001). Thirty-nine dogs (51%; CI, 39–63%) were neutered female dogs, 21 (27%; CI, 18–39%) were neutered male dogs, 8 (11%; CI, 5– 20%) were intact male dogs, and 8 (11%; CI, 5–20%) were

1170

Adler, Drobatz, and Hess

Table 1. Sensitivity and specificity of sodium to potassium (Na:K) ratio as a diagnostic cutoff in dogs with hypoadrenocorticism. Na:K Ratio ,32 ,30 ,29 ,28 ,27 ,24 ,23 ,15

% Sensitivity (95% CI) 100 100 97 93 89 79 74 8

(95–100)a (95–100)a (91–100) (85–98) (80–95) (67–86) (61–82) (3–16)

% Specificity (95% CI)

Percent Dogs Correctly Classified as Diseased or Not Diseased

62 (54–68) 84 (80–90) 90 (86–95) 96 (92–98) 97 (93–99) 100 (97–100) 100 (97–100) 100 (98–100)a

72 88 92 95 95 94 92 75

CI, confidence interval; Na: K ratio, sodium to potassium ratio. 97.5% one-sided CI.

a

intact female dogs. Female dogs were overrepresented in the group of dogs with HA compared with the 200 comparison dogs (P 5 .047). Great Danes and Standard Poodles were overrepresented in the group of dogs with HA compared with the 200 comparison dogs (P 5 .001 and P 5 .039, respectively). There were 7 Great Danes in the study population and 1 Great Dane in the comparison population. There were 5 Standard Poodles in the study population and 3 in the comparison population. Of the 76 dogs, 27 (36%; CI, 25–47%) had received fluids intravenously or subcutaneously within 48 hours of referral, and 5 (7%; CI, 2–15%) had received 1 dose of dexamethasone within 24 hours of referral; these 5 dogs had a Na:K ratio that ranged from 15.4–22.7.

Clinicopathologic Findings Results of biochemical profiles performed at MJRVHUP were available for review in 71 dogs, and 70 (99%; CI, 92–100%) had either a low serum sodium concentration or a high serum potassium concentration on serum biochemistry profile performed at MJRVHUP. In addition, 64 of 71 dogs (90%; CI, 81–96%) had both a low serum sodium concentration and a high serum potassium concentration on serum biochemistry profile performed at MJR-VHUP. Median Na:K ratio on biochemical profile was 21.4 (range, 12.9–35.1). To calculate the Na:K ratio, the lowest Na:K ratio for each dog was recorded from the biochemical profile (58/76 dogs, 76%), the initial blood gas analysis (16/76 dogs, 21%), or referral biochemical profile (2/76 dogs, 3%). Na:K ratios of 27 or 28 correctly identified 95% of dogs as diseased or not diseased (Table 1). The sensitivity of a Na:K ratio of 28 was 93% (CI, 85– 98%), and the sensitivity of a Na:K ratio of 27 was 89% (CI, 80–95%). The specificity of a Na:K ratio of 28 was 96% (CI, 92 - 98%) and that of a Na:K ratio of 27 was 97% (CI, 93–99%). Venous blood gas analysis was performed in 73 dogs (Table 2) at the time of initial examination. Serum iCa2+ concentration was significantly and inversely associated with venous pH (Spearman’s rho 5 2.2439, P 5 .039). Serum iCa2+ concentration was not associated with serum creatinine concentration (Spearman’s rho 5 .2262, P 5 .064), or iMg2+ concentration (Spearman’s

rho 5 2.0427, P 5 .75) but was associated with serum phosphorus concentration (Spearman’s rho 5 .2417, P 5 .047). The iMg2+ concentration was not significantly associated with venous pH or serum creatinine concentration (Spearman’s rho 5 2.0380, P 5 .77 and Spearman’s rho 5 20.2550, P 5 0.056, respectively). Serum Na:K ratio was also calculated based on values obtained from the venous blood gas analysis performed at the time of initial examination (73 dogs). Median Na:K ratio on blood gas analysis was 22.9 (range, 13.7–33.1). Serum Na:K ratio was directly associated with venous pH (Pearson’s correlation coefficient 5 .4173, P , .001). Similarly, serum potassium concentration was significantly and inversely associated with venous pH (Pearso’s correlation coefficient 5 2.4253, P , .001). Serum sodium concentration was not significantly associated with venous pH (Spearman’s rho 5 .0072, P 5 .95).

Imaging Studies Abdominal ultrasonography was performed in 26 dogs. Abnormal ultrasonographic findings included mild to moderately enlarged mesenteric lymph nodes (6 dogs, 23%; CI, 9–44%), gastric wall thickening (5 dogs, 19%; CI, 7–39%), hypoechoic liver (2 dogs, 8%; CI, 1–25%), mild abdominal effusion (2 dogs, 8%; CI, 1–25%), and hyperechoic liver (2 dogs, 8%; CI, 1–25%). One dog had a liver mass. The ultrasonographic appearance of the adrenal glands was noted in 25 dogs. The size of the left adrenal gland was described in 20 of 25 dogs (80%; CI, 59–93%) and was not seen in 5 of 25 dogs (20%; CI, 7– 41%). The left adrenal gland was described as slender in 14 of 20 dogs (70%; CI, 46–88%) and normal in 6 of 20 dogs (30%; CI, 12–54%). The size of the right adrenal gland was described in 18 of 25 dogs (72%; CI, 51–88%) and was not seen in 7 of 25 dogs (28%; CI, 12–49%). The right adrenal gland was described as slender in 13 and normal in 5 of 18 dogs. Both adrenal glands were described as small in 12 and normal in 3 of 15 dogs.

Outcome and Follow-Up All 76 dogs survived to discharge from hospital (97.5% CI: 95–100%). Median time from initial exam-

Hypoadrenocorticism in Dogs

1171

Table 2. Venous blood gas in dogs with hypoadrenocorticism. Variable Venous pH Sodium (mmol/L) Potassium (mmol/L) Na:K ratio Chloride (mmol/L) iCa2+ (mmol/L) iMg2+ (mmol/L) HCO3 (mEq/L) Base Deficit (mmol/L) Lactate (mmol/L) Glucose (mg/dL)

N

Median (Range)

Above RR

73 73 73 73 69 72 61 73 73 66 73

7.335 (7.178–7.463) 134 (107–151) 5.8 (3.9–8.6) 22.9 (13.7–33.1) 106 (81–126) 1.24 (0.77–1.97) 0.37 (0.15–0.78) 19.5 (11.8–27.5) 26.3 (2.6 to 215.0) 1.5 (0.5–5.7) 89 (27–263)

0 1 (1%) 62 (85%) — 1 (1%) 13 (18%) 22 (36%) 7 (10%) 0 9 (14%) 13 (18%)

Normal 29 12 11 16 21 41 36 24 19 55 53

(40%) (17%) (15%) (22%) (30%) (57%) (59%) (33%) (26%) (83%) (72%)

Below RR

RR

44 (60%) 60 (82%) 0 57 (78%) 47 (68%) 18 (25%) 3 (5%) 42 (57%) 54 (74%) 2 (3%) 7 (10%)

7.35–7.47 140–150 3.9–4.9 .27 109–120 1.13–1.33 0.25–0.41 20–24 24 to +4 0.60 – 2.5 65–117

RR, reference range; Na:K, sodium to potassium; iCa2+, ionized calcium; iMg2+, ionized magnesium.

ination for HA to last recheck examination at MJRVHUP in 55 dogs was 121 days (range, 8–2240 days). Two dogs were not supplemented with mineralocorticoid therapy because of a Na:K ratio greater than 27 (Na:K ratio was 28 and 29.5, respectively) at the time of initial hospitalization. Both dogs developed clinical significant hyperkalemia and hyponatremia 1 and 41 days after discharge from the hospital, respectively.

Discussion A Na:K ratio of less than 27 has been used as a diagnostic aid for HA in dogs, although this has been done without knowledge of the specificity and sensitivity of the Na:K ratio as a diagnostic tool.1–3,7–9,17–19,22 Although Na:K ratios of 27 or 28 identify the same percentage of dogs correctly, these Na:K cutoff ratios differ in their sensitivity and specificity. The sensitivity of a Na:K ratio cutoff of 28 is 93%, while the sensitivity of a Na:K ratio cutoff of 27 is 89%. Conversely, the specificity of a Na:K ratio cutoff of 28 is 96%, while the specificity of a Na:K ratio cutoff of 27 is 97%. A clinician may therefore choose to use a Na:K ratio of 27 or 28 as a cutoff, depending on whether a higher sensitivity or specificity is desired. The specificity of a Na:K ratio #24 was 100%. Although a diagnosis of HA must ultimately be confirmed with an ACTH stimulation test, these results suggest that owners of dogs with a Na:K ratio #24 can be advised that HA is very likely and that further diagnostic testing and treatment could result in an excellent outcome. This may be of particular importance in dogs suspected of having another disorder, such as renal or liver disease, that has a guarded to poor prognosis. The sensitivity of a Na:K ratio of 30 was 100% and the specificity was 84%, which suggests that few dogs with HA will be excluded if this cutoff is used but that there will be a substantial number of false-positive tests. We conclude that HA should not be excluded as a possible differential diagnosis in dogs with a Na:K ratio greater than 27, and that dogs with clinical signs consistent with HA should undergo adrenal axis testing

if their Na:K ratio is #30. It is also concluded that the likelihood of diagnosing mineralocorticoid-deficient HA in a dog with a Na:K ratio .30 is small. Given that the prognosis of treated canine HA is excellent, performance of ACTH stimulation tests in dogs with a Na:K ratio between 27 and 30 could be worthwhile, despite the fact that most dogs with HA have a lower Na:K ratio. This point is illustrated by 2 dogs in which a mineralocorticoid deficiency was not apparent at the time of initial examination because their Na:K ratio was 28 and 29.5, respectively. However, these dogs developed Na and K abnormalities that required mineralocorticoid therapy 1 and 41 days after discharge from the hospital. Five dogs received one dose of dexamethasone within 24 hours of referral, which could have altered results of the ACTH stimulation test. However, cortisol concentration returns to normal 24 hours after administration of a single dexamethasone injection to dogs.23 Excluding these 5 dogs that had an acute need for treatment and referral to an emergency department would have skewed the data in that dogs that were compromised to the degree that they needed immediate therapy would have been excluded. Results of this study suggest that venous pH is directly associated with the Na:K ratio and is inversely associated with serum potassium concentration, but it is not associated with serum sodium concentration. Hyperkalemic acidosis has been described in association with decreased aldosterone function.24,25 One of the actions of aldosterone is to facilitate sodium reabsorption and promote potassium excretion from renal epithelial cells into the lumen of the renal collecting ducts.24 Lack of aldosterone, therefore, results in hyponatremia and hyperkalemia.24 Dehydration and hypotension also develop with lack of aldosterone function because water follows sodium.24 One of the theories regarding the acidosis that develops with lack of aldosterone action is that hyperkalemia inhibits ammonium (NH4+) production resulting in reduced ability to excrete acid in the renal tubules.25 The association between decreased venous pH and hyperkalemia identified in this study suggests that the theory regarding hyperkalemia-induced inhibition of NH4+

1172

Adler, Drobatz, and Hess

production might apply to dogs with HA. Although other factors, such as dehydration, can contribute to acidosis, the normal serum lactate concentration identified in most dogs suggests that hyperkalemia is an important factor determining acid base status in dogs with HA (Table 2). Inclusion criteria for this study were results of an ACTH stimulation test that confirmed a diagnosis of HA and a serum sodium concentration below the reference range or a serum potassium concentration above the reference range. The Na:K ratio was not an inclusion criterion as it was one of the variables examined in the study. Dogs were included in the study even if serum sodium and potassium concentrations were normal at the time of admission to MJR-VHUP, as long as one of these electrolyte concentrations was abnormal on earlier biochemical profile. This was done because approximately 40% of dogs received fluid therapy before examination at MJR-VHUP, and fluid therapy could have corrected serum sodium and potassium concentrations. For this reason it is also possible that in a primary care facility the Na:K ratio of dogs with HA is actually lower than the values identified in this study. However, the values recognized in this study are accurate for a tertiary care facility. To reflect Na:K alterations most accurately, the lowest Na:K ratio, either at the time of admission or before referral, was used for sensitivity and specificity calculations Dogs with HA in which sodium and potassium concentrations were normal before and at the time of referral were excluded from the study. An additional 6 dogs were excluded because they did not require mineralocorticoid therapy during the study period. These dogs were excluded to ensure that the findings of the study were clinically useful. One of the goals of the study was to provide clinicians with a screening tool that would enable differentiation of dogs with HA from other dogs that have abnormal sodium or potassium concentrations. It is unlikely that clinicians would seek to use a Na:K ratio in dogs in which both electrolyte concentrations were normal. Two hundred randomly selected hospitalized dogs were chosen to compare the Na:K ratio in dogs with HA to the Na:K ratio in dogs with other diseases. Dogs randomly selected from the hospital population were chosen because a comparison of the Na:K ratio in dogs with HA to healthy dogs would not be clinically useful. Results of this study therefore provide clinicians with a valuable tool for comparing dogs with HA to dogs that have other disease. One limitation of this study design is that some of the diseases of dogs in the comparison group could have influenced serum sodium or potassium concentration. Another limitation is that some of the dogs in the control group may have had HA. HA was excluded on the basis of results of a normal ACTH stimulation test in 4 of the 200 comparison dogs. Future prospective studies that set out to investigate the Na:K ratio in dogs with HA may want to have an ACTH stimulation test performed in control dogs to definitively determine that control dogs do not have HA.

Another limitation of this study is that the ACTH stimulation test was performed by giving ACTH gel intramuscularly at 2.2 U/kg and cortisol was measured 2 hours later. This widely used protocol17 has not been specifically validated for dogs with HA. However, the protocol has been reported in clinically normal dogs in which cortisol concentrations before, and 2 hours after IM injection of an ACTH gel were greater than 3.6 mg/ dL.21 Therefore, documentation of cortisol concentrations below 1.0 mg/dL before and after IM injection with ACTH gel in all dogs included in the present study is consistent with a diagnosis of HA. Abnormal Na+ and K+ concentrations and the need for mineralocorticoid therapy strengthen the diagnosis of HA in these dogs. Elevated iCa2+ concentration was documented in 18% of the dogs. Although elevations in total calcium concentration have been previously reported in dogs with HA,1–3,5,9,11,26 these data confirm that ionized hypercalcemia develops in dogs with naturally occurring HA. Furthermore, serum iCa2+ concentration was significantly and inversely associated with venous pH. We therefore conclude that acidosis might be involved in the pathophysiology of elevated iCa2+ concentration observed in dogs with HA. Total hypocalcemia has been reported in dogs with HA,1–3,5,9 ionized hypocalcemia in dogs with HA is reported here for the first time. Low iCa2+ concentration is associated with a poor prognosis in critically ill human patients with sepsis or pancreatitis.27,28 All of the dogs in this study survived to discharge, indicating that ionized hypocalcemia is not associated with a poor prognosis in dogs with HA. Future studies may further the understanding of decreased iCa2+ concentration noted in dogs with HA in this study. Thirty-five percent of dogs had elevated iMg2+ concentration. Ionized magnesium abnormalities have not been previously reported in dogs with spontaneous HA.1–3,5–11,12 However, experimental hyperaldosteronism in rats reveals decreased plasma iMg2+ and increased urinary and fecal loss of iMg2+.29 In vitro studies of human lymphocytes have also demonstrated that exposure to excess aldosterone results in increased intracellular iMg2+ concentration.30 The effect of aldosterone on iMg2+ homeostasis is incompletely understood; however, it might be mediated by the effect of aldosterone on the Na-Mg antiporter, which is the principal known mechanism of magnesium flux.30 It is possible that lack of aldosterone action in dogs with spontaneous HA would result in elevated plasma iMg2+, as noted in some of the dogs in this study. Ionized magnesium concentration was not significantly associated with venous pH or serum creatinine concentration, suggesting that acidosis or decreased renal function may not be important contributors to alterations in iMg2+ concentration. Abdominal ultrasound was performed in 26 dogs, and adrenal gland appearance was noted in 25 dogs. Abdominal ultrasound supported a diagnosis of HA in 12 dogs in which both adrenal glands were noted to be slender. Both adrenal glands were considered normal in 3 dogs. Therefore, normal appearance of adrenal glands

Hypoadrenocorticism in Dogs

on abdominal ultrasound should not preclude further testing, such as an ACTH stimulation test, in dogs suspected of having HA.

Footnotes a

Cortigel-40, Savage Laboratories, Melville, NY Cortrosyn, Amphstar Pharmaceuticals, Rancho Cucamonga, CA c Synacthen, Alliance Pharmaceuticals Limited, Chippenham, Wiltshire, UK d Antech Diagnostics, Los Angeles, CA e Chemistry analyzer, Kodak Ektachem 250, Eastman Kodak Co, Rochester, NY f Stat Profile, NOVA Biomedical Corporation, Waltham, MA g Immulite, DPC, Los Angeles, CA h Intercooled Stata 8.0 for Windows, Stata Corporation, College Station, TX b

Acknowledgments This work was done at the Matthew J. Ryan Veterinary Hospital, School of Veterinary Medicine, University of Pennsylvania. The findings were reported in part as an abstract in J Vet Intern Med 2006;20:772.

References 1. Peterson ME, Kintzer PP, Kass PH. Pretreatment clinical and laboratory findings in dogs with hypoadrenocorticism: 225 cases (1979–1993). J Am Vet Med Assoc 1996;208:85–91. 2. Melia´n C, Peterson ME. Diagnosis and treatment of naturally occurring hypoadrenocorticism in 42 dogs. J Small Anim Pract 1996;37:268–275. 3. Willard MD, Schall WD, McCaw DE, Nachreiner RF. Canine hypoadrenocorticism: Report of 37 cases and review of 39 previously reported cases. J Am Vet Med Assoc 1982;180:59–62. 4. Kintzer PP, Peterson ME. Treatment and long-term followup of 205 dogs with hypoadrenocorticism. J Vet Intern Med 1997;11:43–49. 5. Rakich PM, Lorenz MD. Clinical signs and laboratory abnormalities in 23 dogs with spontaneous hypoadrenocorticism. J Am Anim Hosp Assoc 1984;20:647–649. 6. Schaer M, Chen CL. A clinical survey of 48 dogs with adrenocortical hypofunction. J Am Anim Hosp Assoc 1983;19: 443–452. 7. Ahlgren M, Bamberg-Thale´n B. Hypoadrenocorticism in the dog. Eur J Comp Anim Pract 1993;3:62–68. 8. Sadek D, Schaer M. Atypical Addison’s disease in the dog: A retrospective survey of 14 cases. J Am Anim Hosp Assoc 1996;32: 159–163. 9. Lifton SJ, King LG, Zerbe CA. Glucocorticoid deficient hypoadrenocorticism in dogs: 18 cases (1986–1995). J Am Vet Med Assoc 1996;209:2076–2081. 10. Langlais-Burgess L, Lumsden JH, Mackin A. Concurrent hypoadrenocorticism and hypoalbuminemia in dogs: A retrospective study. J Am Anim Hosp Assoc 1995;31:307–311.

1173

11. Peterson ME, Feinman JM. Hypercalcemia associated with hypoadrenocorticism in sixteen dogs. J Am Vet Med Assoc 1982;181:802–804. 12. Schaer M, Riley WJ, Buergelt CD, et al. Autoimmunity and Addison’s disease in the dog. J Am Anim Hosp Assoc 1986;22:789–794. 13. Melia´n C, Stefanacci J, Peterson ME, et al. Radiographic findings in dogs with naturally-occurring primary hypoadrenocorticism. J Am Anim Hosp Assoc 1999;35:208–212. 14. Hoerauf A, Reusch C. Ultrasonographic evaluation of the adrenal glands in six dogs with hypoadrenocorticism. J Am Anim Hosp Assoc 1999;35:214–218. 15. Javadi S, Galac S, Boer P, et al. Aldosterone-to-renin and cortisol-to-adrenocorticotropic hormone ratios in healthy dogs and dogs with primary hypoadrenocorticism. J Vet Intern Med 2006; 20:556–56. 16. Rose BD. Effects of hormones on renal function. In: Rose BD, ed. Clinical Physiology of Acid-Base and Electrolyte Disorders, 4th ed. New York: McGraw-Hill; 1994:150–215. 17. Feldman EC, Nelson RW. Hypoadrenocorticism (Addison’s disease). In: Feldman EC, Nelson RW, eds. Canine and Feline Endocrinology and Reproduction, 3rd ed. St Louis, MO: WB Saunders; 2004:394–439. 18. Herrtage ME. Hypoadrenocorticism. In: Ettinger SJ, Feldman EC, eds. Textbook of Veterinary Internal Medicine, 6th edition. Philadelphia, PA: Elsevier; 2005:1612–1622. 19. Pak SI. The clinical implication of sodium-potassium ratios in dogs. J Vet Sci 2000;1:61–65. 20. Reimers TJ, Salerno VJ, Lamb SV. Validation and application of solid-phase chemiluminescent immunoassays for diagnosis of endocrine diseases in animals. Comp Haematol Int 1996;6(3):170–175. 21. Kemppainen RJ, Behrend EN, Busch KA. Use of Compounded adrenocorticotropic hormone (ACTH) for adrenal function testing in dogs. J Am Anim Hosp Assoc 2005;41:368–372. 22. Roth L, Tyler RD. Evaluation of low sodium: Potassium ratios in dogs. J Vet Diagn Invest 1999;11:60–64. 23. Kemppainen RJ, Sartin JL. Effects of single intravenous doses of dexamethasone on baseline plasma cortisol concentrations and responses to synthetic ACTH in healthy dogs. Am J Vet Res 1984;45:742–746. 24. Connell JM, Davies E. The new biology of aldosterone. J Endocrinol 2005;186:1–20. 25. Unwin RJ, Capasso G. The renal tubular acidoses. J R Soc Med 2001;94:221–225. 26. Elliott J, Dobson JM, Dunn JK, et al. Hypercalcaemia in the dog: A study of 40 cases. J Small Anim Pract 1991;32:564–571. 27. Desai TK, Carlson RW, Geheb MA. Prevalence and clinical implications of hypocalcemia in acutely ill patients in a medical intensive care setting. Am J Med 1988;84:209–214. 28. Ariyan CE, Sosa JA. Assessment and management of patients with abnormal calcium. Crit Care Med 2004;32:S146– S154. 29. Chhokar VS, Sun Y, Bhattacharya SK, et al. Hyperparathyroidism and the calcium paradox of aldosteronism. Circulation 2005;111:871–878. 30. Delva P, Pastori C, Degan M, et al. Intralymphocyte free magnesium in patients with primary aldosteronism: Aldosterone and lymphocyte magnesium homeostasis. Hypertension 2000;35: 113–117.