Canine hypoadrenocorticism: Insights into the Addisonian crisis

Abstract

Objective

Primary objectives of this study were to determine presenting complaints, physical examination, clinicopathologic findings, and hospitalization time of dogs with spontaneous hypoadrenocorticism presenting with critical disease; and to compare those end points to dogs with a more stable presentation. Secondary objectives were to evaluate the shock index and to identify precipitating stressors.

Animals

Eighty-four dogs at the Western College of Veterinary Medicine between 1998 and 2018 were included.

Procedure

Data were retrieved from the medical records.

Results

Collapse and depression were more common among critically ill dogs. Hyperlactatemia was rare despite a diagnosis of hypovolemic shock, and a shock index was ineffective in this patient subset. Isosthenuria, total hypocalcemia, and more severe acidosis were more common (P < 0.05) in critical dogs. Owner separation was the most common precipitating stressor.

Conclusion and clinical relevance

We concluded that the critical Addisonian dog has unique characteristics that may aid in early disease identification.

Introduction

Canine spontaneous hypoadrenocorticism (HoAC), or Addison’s disease, is a potentially life-threatening endocrine disease characterized by deficiencies in cortisol and aldosterone produced by the adrenal cortex (1). Addisonian patients have been reported to present either with chronic vague and non-specific clinical signs, or in a state of severe debilitation (i.e., an Addisonian crisis) (1).

Due to its rarity and broad range of presentations, HoAC can present as a diagnostic challenge. Various studies have examined the clinical presentation and hematological and biochemical parameters of the Addisonian population (2–6). However, dogs with critical disease are often reported to represent a subset of all dogs with undiagnosed HoAC (2,3,7,8). Primary objectives of this study were to determine presenting complaints, physical examination, and clinicopathologic findings of dogs with spontaneous HoAC presenting with critical disease; and compare those end points to dogs with a more stable presentation. Secondary objectives were to evaluate the efficacy of a shock index and to identify precipitating stressors in all dogs. We hypothesized that the critical Addisonian dog has unique characteristics when compared to the non-critical dog; our goal was to highlight these differences to aid in early identification of canine HoAC.

Materials and methods

Case selection

A retrospective analysis was conducted on 84 cases of spontaneous canine HoAC diagnosed at the Western College of Veterinary Medicine (Saskatoon, Saskatchewan) from the years 1998 to 2018. Inclusion criteria required confirmation of the diagnosis with an ACTH-stimulation test (pre- and post-ACTH cortisol concentrations < 55 nmol/L) using a solid phase, competitive chemiluminescent enzyme immunoassay (IMMULITE 1000 Cortisol; Siemens Canada, Oakville, Ontario) conducted by Prairie Diagnostic Services (PDS; Saskatoon, Saskatchewan). Dogs were excluded if they were diagnosed with iatrogenic HoAC; had a confirmed diagnosis of HoAC before presentation at the teaching hospital; or were given antifungals or corticosteroids, except for a single dose of dexamethasone, before receiving an ACTH-stimulation test.

Procedures

Data retrieved from the medical records included signalment, presenting complaints, physical examination findings, clinicopathologic findings, hospitalization time (interval from admission to discharge), and a history of any potential precipitating stressors associated with the onset of clinical signs. Bradycardia and tachycardia were defined as a heart rate below or above the normal range of 70 to 160 bpm, respectively (9). Precipitating causes were defined as acute stressors thought to be associated with the onset of clinical signs related to HoAC. The shock index (SI), defined as the ratio of heart rate to systolic blood pressure (SBP) on presentation, was calculated for dogs that were presented in hypovolemic shock (HVS) (10,11). Data pertaining to peripheral pulses, capillary refill time, blood glucose concentrations, and blood pressure measurements were excluded from the results due to their central role in defining critical versus non-critical groups.

Critical case definition

To meet the criteria of the critical case definition, the dog had to present in HVS, in a hypoglycemic crisis (HGC), or with severe hyperkalemia (> 7.5 mmol/L) and documented atrial standstill on an electrocardiogram. Dogs deemed to be in HVS met at least 1 of the following criteria: i) average indirect blood pressure measurements on triage examination indicative of hypotension (mean arterial pressure < 60 mmHg or systolic blood pressure < 90 mmHg); ii) fluid resuscitation that required more than 1/4 of a shock bolus (> 22.5 mL/kg) of intravenous (IV) crystalloids with concurrent weak peripheral pulses or a prolonged capillary refill time (i.e., > 2 s); or iii) required the use of colloids or vasopressors for blood pressure stabilization. Dogs deemed to be in an HGC met at least 1 of the following criteria: i) severe hypoglycemia (< 1.5 mmol/L) on initial emergency panel, or ii) requirement for IV dextrose administration to treat hypoglycemic neurological disturbances. Cases with mild to moderate hypoglycemia without neurological signs were excluded from the critical case definition based on an anticipated high prevalence of this finding in otherwise stable dogs.

Clinicopathologic data

Dogs were evaluated on initial presentation via a combination of an emergency panel [BUN via Azostix Reagent Strips (Siemens Canada), PCV via centrifugation, TP by refractometry, and a blood glucose measurement via portable glucometer]; venous blood gas (VBG) analysis; serum biochemistry panel; complete blood (cell) count (CBC); and urinalysis. All emergency panels were performed patient-side, whereas all serum biochemistry panels, CBCs, and urinalyses were conducted at PDS. Data from 34 healthy dogs were used to define an arterial blood gas reference interval for the blood gas analyzer that was used for most of the samples in this study. Current 2018 reference intervals at PDS were used to evaluate the proportions of critical and non-critical dogs that exhibited abnormalities on their emergency panel, serum biochemistry, and CBC findings on presentation. Azotemia was defined, as per the emergency panel Azostix BUN reading, as BUN ≥ 10.7 to 14.3 mmol/L (≥ 30 to 40 mg/dL) (12). Samples were included in the study if they were collected on presentation to the referring veterinary hospital before referral to the university teaching hospital, and samples were sent directly to PDS. Samples were excluded if they were collected after initiation of treatments, including fluid therapy, or if they were conducted internally by the referring veterinarian before referral.

Statistical methods

The X² test was used to compare proportions of the dichotomous outcome variable for the critical and non-critical populations when the cell counts in the 2 × 2 contingency table were > 5; otherwise, Fisher’s Exact test was used. Proportions were described using the count and percentage. The Shapiro-Wilk normality test was used to examine the distribution of continuous outcome variables. Mean and standard deviation were used to describe parametric data, whereas median and range were used to describe nonparametric data. The median was used to describe all clinicopathologic data for comparative purposes. A 2-sample independent Student’s t-test was used to compare critical versus non-critical population means when the data were normally distributed. However, when these normality assumptions were not met, the Mann-Whitney U test was used. The Bonferroni correction was applied to account for multiple comparisons. All statistical calculations were conducted using a commercial software package (R; R Foundation for Statistical Computing, Vienna, Austria: http://www.R-project.org/). For all comparisons, P-values < 0.05 were deemed significant.

Results

Classification

The critical case definition applied to 28 (33%) cases, whereas 56 (67%) cases were defined as non-critical. At presentation, 16 (57%) dogs were in HVS, 6 (21%) were in an HGC, and 6 (21%) were in both HVS and an HGC. Five dogs had severe hyperkalemia with evidence of atrial standstill on an electrocardiogram; 4 (80%) of these were in concurrent HVS and 1 (20%) was in an HGC.

Signalment

Out of 84 cases, there were 8 (10%) intact females, 1 (1%) intact male, 40 (48%) spayed females, and 35 (42%) neutered males. Mean age at diagnosis was 4.8 ± 2.9 y for the critical group and 5.0 ± 2.4 y for the non-critical group (P = 0.741). There were 37 (44%) mixed-breed dogs, which was the predominant breed in both groups. Other notable breeds represented were the standard poodle (6%), Labrador retriever (5%), Rottweiler (5%), and West Highland white terrier (4%). Before diagnosis, the median number of visits to either the regular veterinarian or the Western College of Veterinary Medicine over the animal’s entire lifespan with symptoms that could be consistent with HoAC was 1 visit for the group with non-critical disease (range: 0 to 3 visits) and 0 for the group with critical disease (range: 0 to 4 visits) (P = 0.127).

Presenting complaints

Common presenting complaints are summarized in Table 1. Dogs in the critical group were more commonly presented with collapse (P < 0.001). Non-critical dogs were presented with the complaint of polydipsia, whereas this complaint was absent among dogs in the critical group (P = 0.611). Signs of gastrointestinal bleeding were rarely reported in either group.

Table 1.

Common presenting complaints of critical and non-critical dogs diagnosed with spontaneous hypoadrenocorticism.

Presenting complaint	Critical dogs (n = 28) N (%)	Non-critical dogs (n = 56) N (%)	P-value
Lethargy	25 (89)	46 (82)	1.000
Anorexia	24 (86)	47 (84)	1.000
Vomiting	20 (71)	35 (63)	1.000
Weakness	12 (43)	8 (14)	0.117
Collapse	11 (39)	1 (2)	< 0.001
Diarrhea	8 (29)	14 (25)	1.000
Weight loss	7 (25)	14 (25)	1.000
Seizure-like episodes	7 (25)	3 (5)	0.182
Tremors	6 (21)	9 (16)	1.000
Hematochezia	2 (7)	3 (5)	1.000
Melena	1 (4)	0 (0)	1.000
Polydipsia	0 (0)	8 (14)	0.611
Polyuria	0 (0)	5 (9)	1.000

n — Number in group; N — Number of dogs with each presenting complaint.

Bonferroni correction for multiple comparisons, α < 0.004.

Both critical and non-critical dogs had a median duration of illness before presentation of 4 d. In critical dogs, the duration ranged from 1 to 90 d, whereas non-critical dogs had a wider range of 1 to 640 d. Four non-critical dogs had durations of illness > 90 d (120, 180, 365, and 640 d, respectively) associated with chronic waxing and waning gastrointestinal disease.

Physical examination

Critical and non-critical dogs were evaluated for the presence of various physical exam findings on presentation (Table 2). The most common findings in both groups were depression and dehydration, though the former was more common in the critical group (P = 0.042). Critical and non-critical groups had similar proportions presenting with weakness; recumbency; hypothermia; tacky, injected, and pale mucous membranes; tachypnea; tachycardia; abdominal pain on palpation; ataxia; and muscle wasting. More than 1/2 of the critical dogs were presented with depression, dehydration, weakness, recumbency, and hypothermia, which would be expected given the parameters used to identify shock in this population.

Table 2.

Proportions of critical and non-critical dogs diagnosed with spontaneous hypoadrenocorticism presenting with various physical examination findings.

Physical exam finding	Critical dogs N (%)	n	Non-critical dogs N (%)	n	P-value
Depression	25 (89)	28	31 (55)	56	0.042
Dehydrated (≥ 5%)	22 (79)	28	29 (52)	56	0.462
Weakness	18 (64)	28	13 (23)	56	0.084
Recumbent	15 (54)	28	1 (2)	56	0.266
Hypothermic (rectal temperature < 37.4°C)	15 (54)	28	15 (27)	56	0.21
Tacky mucous membranes	11 (39)	28	21 (38)	56	1.000
Tachypnea (RR > 30 bpm)	6 (23)	26	20 (36)	55	1.000
Tachycardia (HR > 160 bpm)	1 (4)	28	0 (0)	56	1.000
Bradycardia (HR < 70 bpm)	6 (21)	28	5 (9)	56	1.000
Abdominal pain on palpation	6 (21)	28	10 (18)	56	1.000
Injected mucous membranes	5 (18)	28	3 (5)	56	1.000
Ataxia	4 (14)	28	3 (5)	56	1.000
Muscle loss	3 (11)	28	6 (11)	56	1.000
Pale mucous membranes	3 (11)	28	2 (4)	56	1.000

HR — Heart rate; N — Number of dogs with each physical examination finding; n — Number in group; RR — Respiratory rate.

Bonferroni correction for multiple comparisons, α < 0.004.

Clinical pathology

Proportions from both groups were evaluated for clinically relevant findings on the emergency panels (Table 3). Based on these results, there were no significant differences between the proportions of dogs in each group presenting with azotemia, erythrocytosis, anemia, hyperproteinemia, or hypoproteinemia. Furthermore, 10% of all cases had a blood glucose concentration < 1.5 mmol/L at presentation.

Table 3.

Emergency panel findings on presentation of critical and non-critical dogs diagnosed with spontaneous hypoadrenocorticism.

Parameter	RD	Critical dogs N (%)	n	Non-critical dogs N (%)	n	P-value
Azotemia	BUN 10.7 to 14.3 mmol/L	16 (57)	27	11 (28)	39	0.135
	(30 to 40 mg/dL) or 17.9 to 28.6 mmol/L (50 to 80 mg/dL)
Erythrocytosis	PCV > 0.56	17 (63)	27	26 (63)	41	1.000
Anemia	PCV < 0.39	1 (4)	27	4 (8)	41	1.000
Hyperproteinemia	TP > 74 g/L	6 (22)	27	12 (30)	40	1.000
Hypoproteinemia	TP < 56 g/L	5 (19)	27	7 (18)	40	1.000

N — Number of dogs; n — Number in group; RD — Reference definition. Bonferroni correction for multiple comparisons, α < 0.01.

Venous blood gas analyses were evaluated in both groups (Table 4). Following comparison of continuous variables, dogs in the critical group had significantly lower median blood pH [critical median (CM) = 7.253, range: 7.134 to 7.390; non-critical median (NCM) = 7.296, range: 7.108 to 7.451; P = 0.03]; HCO₃ concentration (CM = 14.7 mmol/L, range: 12.5 to 19.9 mmol/L; NCM = 17.6 mmol/L, range: 13.9 to 21.4 mmol/L; P < 0.001); and base excess (CM = −11.0 mmol/L, range: −16.5 to 4.2 mmol/L; NCM = −7.2 mmol/L, range: −14.8 to −1.7 mmol/L; P < 0.001) on VBG reports (Bonferroni correction for multiple comparisons, α < 0.01). Based on the VBG results, there were no significant group differences in median sodium and potassium concentrations or the proportion of dogs presenting with hyponatremia and hyperkalemia. Furthermore, the Na/K ratio did not differ significantly between the critical and non-critical groups. A peripheral Na/K ratio < 27 (1,3,6) was present in at least 80% of dogs in both critical and non-critical groups; however, only 14 (54%) critical and 13 (37%) non-critical dogs met the more rigorous criteria of a Na/K ratio < 20 (1,13). Ionized hypocalcemia was reported more commonly than ionized hypercalcemia in both groups of dogs. The overall prevalences of ionized hypocalcemia and ionized hypercalcemia in all dogs were 41% (24 dogs) and 24% (14 dogs), respectively (n = 58). Twelve (36%) non-critical dogs and 12 (48%) critical dogs had an ionized hypocalcemia on presentation, whereas only 3 (13%) critical dogs and 6 (18%) non-critical dogs were hyperlactatemic (P = 1.000). Venous blood gas lactate concentrations > 5.0 mmol/L were only present in 1 (4%) critical case.

Table 4.

Venous blood gas findings on presentation of critical and non-critical dogs diagnosed with spontaneous hypoadrenocorticism.

Parameter	RD	Critical dogs N (%)	n	Non-critical dogs N (%)	n	P-value
Acidosis	pH < 7.395	26 (100)	26	31 (86)	36	1.000
Hypercapnea	pCO2 > 37.4 mmHg	12 (46)	26	15 (42)	36	1.000
Low bicarbonate	HCO3− < 17.3 mmol/L	20 (77)	26	17 (47)	36	0.592
Low base excess	BE < −5.3 mmol/L	24 (92)	26	31 (86)	36	1.000
Hyponatremia	Na < 145 mmol/L	25 (96)	26	33 (92)	36	1.000
Hyperkalemia	K > 4.3 mmol/L	22 (85)	26	30 (83)	36	1.000
	K > 6.5 mmol/L	12 (46)	26	13 (36)	36	1.000
	K > 8.0 mmol/L	4 (15)	26	1 (3)	36	1.000
Low Na/K ratio	Na/K < 27	22 (85)	26	29 (80)	36	1.000
	Na/K < 23	20 (77)	26	24 (66)	36	1.000
	Na/K < 20	14 (54)	26	13 (37)	36	1.000
Hypochloremia	Cl < 113 mmol/L	20 (83)	24	30 (83)	36	1.000
Ionized hypocalcemia	Ca2+ < 1.31 mmol/L	12 (48)	25	12 (36)	33	1.000
Ionized hypercalcemia	Ca2+ > 1.40 mmol/L	6 (24)	25	8 (24)	33	1.000
Hyperlactatemia	Lactate > 2.33 mmol/L	3 (13)	24	6 (18)	33	1.000
Lactate	> 5.00 mmol/L	1 (4)	24	0 (0)	33	1.000

N — Number of dogs; n — Number in group; RD — Reference definition. Bonferroni correction for multiple comparisons, α < 0.003.

Serum biochemistry findings are summarized in Table 5. Total hypocalcemia was identified in the critical group but was absent in the non-critical group (P = 0.042). Of the 5 critical dogs with total hypocalcemia, 4 (80%) had concurrent ionized hypocalcemia and hypoalbuminemia and 1 (20%) had normal ionized calcium and serum albumin concentrations. The overall prevalences of total hypocalcemia and total hypercalcemia in all dogs were 7% (5 dogs) and 32% (23 dogs), respectively (n = 73), with no difference between groups (P = 0.238) in the prevalence of total hypercalcemia. The proportion of dogs with azotemia on the serum biochemistry panel was approximately 75% in both groups.

Table 5.

Serum biochemistry findings on presentation of critical and non-critical dogs diagnosed with spontaneous hypoadrenocorticism.

Parameter	RD	Critical dogs N (%)	n	Non-critical dogs N (%)	n	P-value
Total hypercalcemia	Ca2+ > 3.03 mmol/L	3 (13)	24	20 (41)	49	0.238
Total hypocalcemia	Ca2+ < 1.91 mmol/L	5 (21)	24	0 (0)	49	0.042
Elevated urea	Urea > 11.4 mmol/L	17 (71)	24	33 (67)	49	1.000
Elevated creatinine Creatinine	> 121 μmol/L	13 (54)	24	35 (71)	49	1.000
Azotemia	Urea > 11.4 mmol/L or Creatinine > 121 μmol/L	18 (75)	24	37 (76)	49	1.000
Elevated amylase	Amylase > 1375 U/L	0 (0)	15	7 (25)	28	1.000
Hypercholesterolemia	Cholesterol > 5.94 mmol/L	1 (4)	24	6 (12)	49	1.000
Hypocholesterolemia	Cholesterol < 2.70 mmol/L	14 (58)	24	13 (26)	49	0.238
Elevated total bilirubin	T. Bili > 4 μmol/L	6 (25)	24	5 (10)	48	1.000
Elevated ALT	ALT > 59 U/L	12 (50)	24	28 (57)	49	1.000
Elevated GLDH	GLDH > 7 U/L	7 (35)	20	15 (42)	36	1.000
Elevated creatinine kinase	CK > 418 U/L	18 (75)	24	24 (49)	49	0.882
Hypoproteinemia	Total protein < 55 g/L	13 (54)	24	12 (24)	49	0.350
Hypoalbuminemia	Albumin < 32 g/L	17 (71)	24	28 (57)	49	1.000

N — Number of dogs; n — Number in group; RD — Reference definition. Bonferroni correction for multiple comparisons, α < 0.004.

The CBC did not differ significantly between dogs in the critical (n = 24) and non-critical (n = 52) groups. Approximately 40% of dogs in both groups were presented with an erythrocytosis. Anemia was present in 13% of both critical and non-critical dogs based on hematocrit values. Twenty-one (88%) critical dogs and 41 (79%) non-critical dogs had a normal absolute lymphocyte count. Lymphopenia was present in 1 (4%) critical case and 7 (13%) non-critical cases. An absolute lymphocytosis was present in 8% of both critical (n = 2) and non-critical (n = 4) dogs. Approximately 80% of all dogs had a normal absolute monocyte count. Eleven (46%) critical dogs and 14 (27%) non-critical dogs were presented with an absolute eosinophilia; overall, 25 of 84 (30%) total dogs had an absolute eosinophilia on presentation.

Urine samples were collected on presentation from 17 dogs in the critical group and 37 in the non-critical group. Nine (53%) critical dogs were presented with isosthenuria, compared to only 3 (8%) non-critical dogs (P < 0.001). A total of 37 dogs (67%) were presented with a urine specific gravity < 1.030 (n = 55).

Shock index

Dogs in the critical group that were defined as presenting in HVS were evaluated for the applicability of a shock index (SI) (Table 6) (11,12). Shock index values of both > 0.9 (10) and > 1.0 (11) were reported to be fairly sensitive and specific indicators for diagnosing shock in dogs in an emergency setting (10,11). Of the 22 cases that met the HVS criteria, 16 had a systolic blood pressure recording available in the medical record. Three (19%) dogs were presented with bradycardia (< 70 bpm) and 13 (19%) dogs were presented with a normal heart rate (70 to 160 bpm); no dogs were presented with tachycardia (> 160 bpm). Among the dogs with bradycardia or a normal heart rate, only 11 (50%) had blood potassium concentration > 6.5 mmol/L. The overall mean SI for HVS cases was 1.02 ± 0.35; however, for cases presenting with bradycardia, the mean SI was 0.65 ± 0.23. Eight (50%) of the total HVS cases were presented with an SI greater than 1.0, and all 8 were presented with a normal heart rate. Only 1 additional dog had an SI > 0.9 but < 1.0.

Table 6.

Evaluation of the shock index (HR/SBP) in critical dogs diagnosed with spontaneous hypoadrenocorticism and presenting in hypovolemic shock.

	Total hypovolemic shock cases	Hypovolemic shock cases presenting with bradycardia	Hypovolemic shock cases presenting with a normal heart rate
Mean ± SD shock index (HR/SBP)	1.02 ± 0.35	0.65 ± 0.23	1.10 ± 0.32
n	16	3	13

HR — Heart rate; n — Number in group; SBP — Systolic blood pressure; SD — Standard deviation.

Precipitating stressors

Of the 84 total cases of spontaneous HoAC in this study, 22 (26%) had a suggestive precipitating cause associated with the onset of clinical signs, whereas the remaining 62 (74%) cases did not. Precipitating causes were distributed into 4 categories: 8 (36%) cases were associated with the owner leaving the dog for multiple days under the care of another individual, 4 (18%) cases were associated with the introduction of a new animal or baby into the household, 2 (9%) cases were associated with a visit to a dog groomer, and 8 (36%) cases were associated with various other stressful events (e.g., veterinary visits, strenuous exercise).

Hospitalization time

Dogs in the critical group had a median hospitalization time of 41 h (range: 24 to 216 h), and those in the non-critical group had a median hospitalization time of 48 h (range: 0 to 335 h; P = 0.910) (Figure 1). All critical dogs were hospitalized for a minimum of 24 h, whereas 5 (9%) non-critical dogs were not hospitalized. Euthanasia was elected by the owner for 1 non-critical dog following initial assessment. No other deaths occurred at the time of diagnosis and initial treatment.

Figure 1.

Box plot of hospitalization times for critical and non-critical dogs diagnosed with spontaneous hypoadrenocorticism (P = 0.910). The box plot displays the first quartile, median, and third quartile hospitalization times. Whiskers display the minimum and maximum hospitalization times that do not exceed 1.5 × the interquartile range (IQR), represented by the vertical axis of the box plot. Points plotted beyond the maximum whisker value represent single hospitalization times that exceeded 1.5 × the IQR.

Discussion

The Addisonian dog with critical disease is typically characterized by the presentation of HVS, development of an HGC, or severe hyperkalemia and bradyarrhythmias. Unsurprisingly, collapse and depression were more commonly identified among dogs classified with critical illness in this study. However, dogs in HVS rarely exhibited hyperlactatemia, a laboratory finding that may be unique to this patient subset. In the face of hyperkalemia and cortisol deficiency, an SI was also deemed ineffective in identifying HVS. Therefore, a lack of tachycardia despite other signs of inadequate tissue perfusion at triage may raise suspicion for HoAC. Additionally, a greater degree of illness was associated with total hypocalcemia rather than hypercalcemia, which may be important to consider when interpreting clinicopathologic data.

Although hypoglycemic seizures have been associated with canine HoAC (3,14–16), it is important to note that the presenting complaint of seizure-like activity was also noted among dogs that were non-critical at presentation. Despite no indication in the medical record that seizure-like episodes were attributed to factors other than hypoglycemia, this possibility cannot be ruled out.

Canine HoAC has classically been associated with both total and ionized hypercalcemia (17). Elevated total and ionized calcium concentrations were reported in 30 to 40% (2,4,18) or 18% (13) of dogs with HoAC. Although the mechanism is not completely understood, reduced glomerular filtration rate, increased gastrointestinal calcium absorption, hypovolemia, and reduced urinary calcium excretion have been implicated (19). It was reported that cases of severe HoAC were more likely to develop hypercalcemia than those with mild to moderate disease (2), and that the severity of hypercalcemia was correlated with the severity of HoAC (18). In this study, 24% of dogs in both the critical and non-critical groups were presented with ionized hypercalcemia (no difference between groups). Ionized hypercalcemia has also been related to acidosis, where a decrease in blood pH decreases protein-bound calcium (13). Dogs in the critical group were presented with a lower median VBG pH compared to those in the non-critical group (P = 0.006); however, despite this difference, critical and non-critical dogs were still presented with similar proportions of ionized hypercalcemia. Regarding serum biochemistry results, only 13% of critical cases had total hypercalcemia, compared to 41% of non-critical cases, although this difference did not reach significance (P = 0.238). Of further interest, total (n = 5, 21%) and ionized (n = 12, 48%) hypocalcemia was a more common finding than total (n = 3, 13%) and ionized (n = 6, 24%) hypercalcemia among the critical population. Hypocalcemia of critical illness has been reported in dogs (20), although ionized hypocalcemia has only been associated with canine HoAC in 1 other study, affecting 25% of dogs (13). Proposed mechanisms in critically ill animals include aggressive IV fluid therapy, hypovitaminosis D, hypomagnesemia, hypoparathyroidism, and changes in the degree of protein-bound or chelated calcium (20). Total hypocalcemia has classically been attributed to concurrent hypoalbuminemia (19) and was reported as a rare finding in approximately 9% of dogs with HoAC in another study (2). In this study, hypoalbuminemia was present in 80 and 91% of critical dogs with total and ionized hypocalcemia, respectively.

The suspicion of canine HoAC is facilitated by the lack of a stress leukogram, which has been reported as the most common leukogram abnormality in Addisonian dogs (19). Whereas the presence of lymphopenia was previously shown to rule out a diagnosis of HoAC (21), an absolute lymphopenia was present in 1 (4%) critical and 7 (13%) non-critical dogs in this study.

Similar to the findings of this study, it was reported that ~60% of dogs with HoAC present with a urine specific gravity < 1.030 (2). More specifically, in the current study, isosthenuria was common in critical patients. Isosthenuria, polyuria, and polydipsia in dogs with HoAC have been attributed to hyponatremia resulting in loss of the normal medullary concentration gradient and impaired capacity for water resorption associated with aldosterone insufficiency (1).

The SI has been used as an effective means of diagnosing acute hypovolemic shock in humans (22), and more recently has been applied to canine populations in veterinary medicine (10,11). Due to the key roles of corticosteroids in catecholamine production and adrenergic receptor sensitivity (1), it was questioned whether dogs with spontaneous HoAC would present with an SI reflective of their degree of hypovolemia. An SI > 0.9 was demonstrated (10) to be a fairly sensitive and specific marker for shock in dogs, but in this study, only 69% of dogs that had a normal heart rate and were characterized as being in HVS on presentation had an SI > 0.9. None of the bradycardic dogs characterized as being in HVS in this study were presented with an SI > 0.83. This was not surprising, given the method by which SI is calculated (heart rate divided by SBP), but it highlighted the unique presentation of HVS in cases of HoAC. This uniqueness was further demonstrated by the fact that tachypnea and tachycardia — physical exam findings classic of compensated HVS (22,23) — were not more common in the critical Addisonian cases compared to the non-critical cases.

A plasma venous lactate > 5 mmol/L has also been used as an inclusion criterion for identifying shock (11). Interestingly, despite various other clinical indicators of HVS, such as hypotension, weak peripheral pulses, collapse, and a prolonged capillary refill time, only 1 critical case in this study had a VBG lactate measurement > 5 mmol/L. The cause for the lack of hyperlactatemia in the critical group is unknown, but hyperlactatemia was reported in dogs with increased blood corticosteroid concentrations (24). This link is postulated to be related to increased glucose utilization secondary to an upregulation in gluconeogenesis (25). Perhaps the inverse of this effect may have contributed to the normal blood lactate concentrations we observed in the critical Addisonian dogs presenting in HVS.

Although it has often been reported that development of clinical signs associated with HoAC may be associated with a precipitating cause (17), studies have yet to investigate these potential stressors. Potential triggers cited in the literature include boarding at a kennel or visiting the veterinarian (17); in this study, 26% (22/84) of cases were presented after an identified potentially stressful event. The most common potential stressor, identified in 36% (8/22) of these cases, was absence of the owner and subsequent pet care by another individual, albeit not specifically at a boarding facility. Based on these findings, precipitating stressful events are common in dogs diagnosed with HoAC, and identification of potential stressors is a vital component of HoAC management. To reduce the risk of a crisis, a higher glucocorticoid dose should be recommended for dogs with HoAC during periods when the owner is absent for extended intervals or during the introduction of a new baby or animal, in addition to other potentially stressful events.

Limitations of this study were primarily related to its retrospective nature. In addition, the determination of a critical patient definition is a point of open criticism. Given the large number of comparisons conducted, Bonferroni corrections were applied to reduce the chances of Type 1 errors. However, insufficient power may have also resulted in Type 2 errors, particularly among proportionate comparisons. Due to inconsistent classification of HoAC cases within the medical records (i.e., primary versus secondary, typical versus atypical), we were unable to appropriately report the role that these classifications may have in cases of critical and non-critical Addison’s disease. Furthermore, it is important to recognize that nearly all dogs in this study were sufficiently ill to warrant hospitalization and the majority were presented through the Emergency Department. Therefore, patients with more chronic waxing and waning signs may not have been sufficiently captured and the authors caution extrapolation of these findings to populations with a lesser degree of illness. There is a margin of error associated with the clinical pathology reference intervals in the study. Reference definitions were based on current reference intervals; however, due to the time frame of the study, both analyzers and reference intervals varied over time at the diagnostic laboratory.

Hypoadrenocorticism often presents as chronic, stable debilitation; however, it also has the potential to present as an acute, life-threatening disease characterized by HVS, an HGC, and life-threatening hyperkalemia. It is important to recognize key factors suggesting the presentation of a critical Addisonian, such as a history of collapse, total hypocalcemia, and isosthenuria. Patients in HVS may not have a significant hyperlactatemia on VBG and sinus tachycardia may be absent at triage despite other signs of poor tissue perfusion. Although retrospective data are useful for identification of the critical Addisonian, further prospective studies are needed to evaluate the repeatability and clinical usefulness of the characteristics identified in this study. CVJ