Increased risk of select glucocorticoid adverse events in dogs of higher body weight

Abstract

There are limited data on glucocorticoid treatment in dogs. The purpose of this study was to investigate whether dogs of higher body weight experienced more adverse events when receiving glucocorticoid therapy. Data pertaining to glucocorticoid therapy was abstracted from the records of 61 dogs that were prescribed glucocorticoids for treatment of immune-mediated thrombocytopenia or hemolytic anemia from 2014 to 2019. The odds of developing muscle atrophy and polyphagia during therapy were increased by 30% for each 5 kg of additional body weight. Almost half of the dogs (44.3%) fluctuated > 15% from baseline weight during therapy. Dogs whose body condition scored as above ideal were at increased risk (odds ratio = 4.2) for being diagnosed with urinary tract infection. Our findings suggest that standard linear glucocorticoid dosing may place higher body weight dogs at increased risk of developing adverse events. Accelerated glucocorticoid tapering and/or alternative dosing schemes in dogs with higher body weights may be prudent in efforts to improve tolerance and client compliance.

Introduction

Glucocorticoids (GC), steroid hormones that bind to GC receptors, are associated with a vast array of dose-dependent biological effects (1). Prednisolone, prednisone, methylprednisolone, and dexamethasone are the most widely used GC in veterinary medicine for dogs, with dose-dependent biological activities that include hormone replacement or supplementation as well as anti-inflammatory, immunosuppressive, and antineoplastic effects (2–9). Due to widespread distribution of GC receptors in virtually all mammalian cells, GC therapy is indicated for a variety of conditions, but systemic administration is associated with an array of adverse effects (2). In dogs, commonly reported adverse events of systemic GC, exogenous or endogenous, include polyphagia, polyuria, polydipsia, restlessness, panting, obesity, muscle atrophy and weakness, ligament degeneration, susceptibility to infection, insulin resistance, hepatopathy, and impaired wound healing (10,11).

In dogs, hepatic conversion of prednisone to prednisolone is rapid and complete, allowing these drugs to be used interchangeably at equivalent dosages (11). The broad dosing range for prednisolone varies based on indication and reference (3,11). The 2019 American College of Veterinary Internal Medicine (ACVIM) consensus statement for management of immune-mediated hemolytic anemia (IMHA) suggested the initial dose of prednisone or prednisolone should not exceed 2.0 mg/kg/d in dogs weighing > 25 kg; however, the recommendation was based on the clinical experience of panel members and not on peer-reviewed data (6).

Multiple studies have investigated the efficacy of GC therapy for various conditions, with a majority documenting incidence of adverse effects. One prospective study of dogs with inflammatory bowel disease that were prescribed an immunosuppressive dose of prednisone (2.0 mg/kg/d) reported that 80% developed polydipsia, 70% polyuria, 55% excessive panting, and 30% lethargy (9). Muscle weakness or atrophy, urinary incontinence, and behavioral changes were less commonly reported (9). Another retrospective study investigating immunosuppressive regimens for IMHA indicated that 28.6% of animals receiving prednisone at 2.0 to 2.5 mg/kg/d developed polyuria or polydipsia (12). In a study investigating prednisone-induced adverse effects with 2.0 mg/kg/day dosing for management of immune-mediated polyarthritis, all dogs (100%) developed polydipsia and polyphagia, with the majority also experiencing polyuria (91%) and excessive panting (81%) (13). Other studies documented increased risk of dogs developing urinary tract infections (UTI) with protracted GC administration (14–16).

Adverse effects associated with immunosuppressive GC treatment occurs with significant frequency and can impact not only an animal’s quality of life but also owner compliance and willingness to continue therapy (17). Although higher GC doses carry greater risk and severity of adverse effects, there is a paucity of data investigating methods to refine GC therapy for dogs in the veterinary literature (2,6). Anecdotally, dogs > 25 kg that are prescribed currently accepted immunosuppressive doses experience increased adverse effects severity compared to dogs of lower weight (6). However, to our knowledge, no study has investigated this relationship. The purpose of the current study was to investigate the hypothesis, based on our clinical experience, that dogs with higher body weight experience a higher incidence of adverse effects, compared to dogs with lower body weights, when receiving currently accepted immunosuppressive doses of GC. If a significant relationship does exist, investigating alternative dosing regimens to optimize treatment and minimize adverse effects would be warranted.

Materials and methods

This was a retrospective study, with dogs’ status compared before and after pharmacologic intervention. Medical records of client-owned dogs diagnosed with immune-mediated thrombocytopenia (ITP) or IMHA from 2014 to 2019 at Pittsburgh Veterinary Specialty and Emergency Center (PVSEC) and prescribed immunosuppressive doses of GC followed by a gradual medication taper were reviewed. The period of data collection was selected based on the implementation of the authors’ current institutional electronic medical record system and the start of manuscript data collection. To optimize completeness and availability of medical records within the context of accessible records, dogs were excluded if the primary care veterinarian prescribed GC more than 1 mo prior to presentation to PVSEC or if medical records from the time when first examined were unavailable. Dogs were also excluded if they had fewer than 4 recheck appointments during GC taper, if initial rechecks did not fall within 10 mo of initiating therapy, or if they did not receive GC during these visits. Data were gathered from standardized client intake questionnaires, imaging reports, laboratory results, and doctor examination notes. Collected information included breed, body weight, body condition score, sex, diagnosis, age, date of initiation of therapy, adverse events, concurrent medications, and initial GC dose and tapering schedule. Adverse events included laboratory values outside of the reference range, physical examination abnormalities, and reported clinical signs. Reported clinical signs and physical examination abnormalities included polyphagia, polyuria, panting, muscle loss/atrophy, cranial abdominal organomegaly, dermatologic changes (pyoderma, pododermatitis, otitis externa, calcinosis cutis), gastrointestinal disturbances (ulceration, vomiting, diarrhea), UTI, coagulopathy, or other infections (Table 1). Urinary tract infection was diagnosed by microbial culture based on clinical suspicion. Clinical signs reported at the time of initial examination were excluded from analysis, to minimize inclusion of adverse events that may have been related to the underlying disease process rather than the result of therapeutic intervention.

Table 1.

Adverse events during the treatment period.

Adverse event	Number of individuals experiencing	Mild	Moderate	Severe
Polyuria	41/61 (67.2%)	41/41 (100%)	0	0
Cranial abdominal organomegaly	34/61 (55.7%)	34/34 (100%)	0	0
Polyphagia	34/61 (55.7%)	34/34 (100%)	0	0
Panting	28/61 (45.9%)	28/28 (100%)	0	0
Muscle atrophy/loss	25/61 (41.0%)	25/25 (100%)	0	0
Pyoderma	18/61 (29.5%)	0	18/18 (100%)	0
Diarrhea	18/61 (29.5%)	4/18 (22.2%)	14/18 (77.8%)	0
Urinary tract infection	17/61 (27.9%)	0	17/17 (100%)	0
Lethargy	13/61 (21.3%)	13/13 (100%)	0	0
Vomiting	11/61 (18.0%)	0	11/11 (100%)	0
GI bleeding/ulceration	4/61 (6.6%)	0	1/4 (25%)	3/4 (75%)
Otitis externa	3/61 (4.9%)	0	3/3 (100%)	0
Pododermatitis	3/61 (4.9%)	0	3/3 (100%)	0
Digit infarction, requiring amputation	1/61 (1.5%)	0	0	1/1 (100%)
Cranial cruciate ligament rupture	1/61 (1.5%)	0	1/1 (100%)	0
Cough	1/61 (1.5%)	0	1/1 (100%)	0
Insulin resistance (known diabetic)	1/61 (1.5%)	0	1/1 (100%)	0
Calcinosis cutis	1/61 (1.5%)	0	1/1 (100%)	0
Keratoconjunctivitis sicca	1/61 (1.5%)	0	1/1 (100%)	0
Subcutaneous abscess requiring surgery	1/61 (1.5%)	0	0	1/1 (100%)
Sublingual granuloma formation	1/61 (1.5%)	1/1 (100%)	0	0
Ascites secondary to portal thrombus	1/61 (1.5%)	0	0	1/1 (100%)
Ascites of unknown etiology	1/61 (1.5%)	0	0	1/1 (100%)
Gingival hyperplasia	1/61 (1.5%)	1/1 (100%)	0	0
Tapeworm infestation	1/61 (1.5%)	0	1/1 (100%)	0
Flea infestation	1/61 (1.5%)	0	1/1 (100%)	0
Laryngeal paralysis	1/61 (1.5%)	1/1 (100%)	0	0

The “number of individuals experiencing” column represents the number of individuals experiencing the events over the total number of individuals, with the percentage in parentheses. The remaining columns show individuals experiencing each event at the given severity. Adverse event severity was categorized as mild (requiring no change in therapy beyond continued GC taper), moderate (requiring additional outpatient therapy), or severe (requiring hospitalization or surgical intervention).

Statistical analysis

Measurements are reported as mean ± standard deviation, unless otherwise noted. Statistical analyses were performed using SAS 9.4 (Cary, North Carolina, USA). The threshold of statistical significance was P < 0.05. Different types of gastrointestinal adverse effects were grouped together, and dermatological adverse effects were combined to have more than 10 cases to increase statistical power. Individuals experiencing multiple adverse effects within the combined categories were counted only once for analysis. Coagulopathies and infections other than UTI or pyoderma were not analyzed due to low incidence. Clinical sign outcome variables consisted of the following 9 adverse effects: polyphagia, polyuria, panting, muscle loss/atrophy, cranial abdominal organomegaly, lethargy, dermatological signs, UTI, and gastrointestinal signs.

Initial GC dose and starting weight were explored as potential predictors of likelihood of adverse effects, and body condition category and diagnosis (ITP versus IMHA) were considered as possible confounders. Available body condition scores were on 5-point or 9-point scales; to allow for combined analysis, animals were classified either as overweight or not overweight, based on the corresponding scale. Nine separate logistic regressions were used to test whether starting weight affected odds of each adverse effect. Eighteen separate logistic regressions were used to test if either potential confounder affected the odds of development of each adverse effect. Log-likelihood ratio values of P were used to assess the regression analyses. Student’s t-tests were used to compare initial GC dose between dogs that did or did not experience each adverse effect.

To explore the severity of adverse effects, 3 possible outcome variables were constructed. Adverse effects were classified as mild if they required no change in therapy beyond continued GC taper, moderate if they required additional outpatient therapy, or severe if they required hospitalization or surgical intervention (Table 1). Separate Poisson regressions were used to test whether the number of mild, moderate, or severe events were affected by starting weight or potential confounders.

If P < 0.10 for either potential confounder and any outcome and data for the adverse effects and confounder were available for at least 20 dogs, then multiple logistic regression models with starting weight and confounder were constructed. Multiple comparisons (9 adverse effects plus 3 severity categories) were corrected by the Adaptive Hochberg method (adjusted P).

The Veterinary Cooperative Oncology Group Common Terminology Criteria for Adverse Events v1.1 (VCOG-CTCAE) served as the guideline for conversion of continuous variables into categorical variables for laboratory data analysis and comparison between different reference ranges (Table 2) (18). The VCOG-CTCAE guidelines were applied for grading derangements in alkaline phosphatase (ALP), alanine aminotransferase (ALT), weight, neutrophil, and platelet counts. Starting values greater than the values during therapy were assigned a Grade 0. Due to a low number of observations in each group, Grades 0 and 1 were combined for derangements in ALP and ALT and for weight change. For neutrophil and platelet count derangements, Grades 3 and 4 were combined. Abnormal platelet count was analyzed separately for ITP and IMHA dogs. Starting weight was compared for ALP, ALT, neutrophil, and platelet changes using the Kruskal-Wallis test, with Dunn’s test for post-hoc analysis.

Table 2.

Criteria for weight change and laboratory value grading.

VCOG-CTCAE
Grade	Weight change	ALP	ALT	Neutrophil	Platelet
0	< 5%	≤ ULN	≤ ULN	≤ ULN	≤ ULN
1	5 to 10%	> ULN to 2.5× ULN	> ULN to 1.5× ULN	> ULN to 1.5× ULN	> ULN to 1.5× ULN
2	10 to 15%	> 2.5 to 5.0× ULN	> 1.5 to 4.0× ULN	> 1.5 to 2.5× ULN	> 1.5 to 2.0× ULN
3	> 15% from baseline	> 5.0 to 20× ULN	> 4.0 to 10× ULN	> 2.5 to 3.5× ULN	> 2.0 to 2.5× ULN
4	—	> 20× ULN	> 10× ULN	> 3.5× ULN	> 2.5× ULN

Weight change, ALP, and ALT were adopted directly from VCOG-CTCAE v1.1. Neutrophil and platelet ranges were modified to accommodate the dataset. ULN — Upper limit of the normal reference range.

Results

Patient population

Seventy-five patient medical records were reviewed. Fourteen of the 75 dogs were excluded due to insufficient baseline data, such as initial GC dose, intake physical examination, presenting clinical signs, or starting body weight or because of insufficient follow-up information. Data from 61 dogs were included. Mixed breeds represented the largest proportion of dogs (n = 15), followed by shih tzu (n = 6), Labrador retriever (n = 5), and cocker spaniel (n = 4); the remainder of breeds were represented by 1 or 2 individuals. The population included 1 intact female, 30 spayed females, 3 intact males, and 27 neutered males. The mean weight at the initial examination was 22.4 ± 13.9 kg (range: 3 to 68.4 kg); 26 dogs were > 25 kg, and 35 dogs were < 25 kg. Patient weight was recorded at a minimum of 2 visits during therapy. The greatest fluctuation of weight from baseline occurred a median of 102 d after initial examination (range: 16 to 267 d). Body condition score at the initial examination was available for 42 dogs: 20 dogs were classified as overweight, 20 as ideal weight, and 2 as underweight, with body condition scores unavailable for 19 dogs. The mean age was 6.7 y. Median time to first follow-up was 3.1 wk (range: 0.9 to 13.1 wk); second follow-up 7.0 wk (range: 2.3 to 21.1 wk); third follow-up 10.7 wk (range: 4.1 to 31.1 wk); and fourth follow-up 14.7 wk (range: 4.6 to 39.1 wk). Nineteen dogs were prescribed GC therapy for ITP and 42 dogs for IMHA. Comorbidities at initial presentation were recorded, but there were insufficient data to control for comorbidities in the statistical analysis. Twenty-three dogs were reported to have no previous medical history beyond routine wellness care or sterilization. Documented comorbidities included hypothyroidism (n = 2) and diabetes mellitus (n = 1); these 3 dogs were all undergoing therapy for pre-existing endocrinopathy before presentation.

Medications

The mean initial oral dose of prednisone/prednisolone was 2.04 ± 0.30 mg/kg/d (range: 0.87 to 2.73 mg/kg/d). The mean prednisone/prednisolone dose at the first recheck was reduced to 1.61 ± 0.44 mg/kg/d (range: 0.43 to 2.54 mg/kg/d); 1.00 ± 0.39 mg/kg/d (range: 0.27 to 2.57 mg/kg/d) after the second recheck; 0.62 ± 0.37 mg/kg/d (range: 0.12 to 2.4 mg/kg/d) after the third recheck; and 0.35 ± 0.25 mg/kg/d (range: 0 to 1.27 mg/kg/d) after the fourth recheck. Biologically equivalent doses of oral dexamethasone were substituted for initially prescribed prednisone/prednisolone in 4 dogs, due to lack of control of the targeted disease process with prednisone (n = 2), suspected gastrointestinal ulceration (n = 1), and undetermined reason based on medical record (n = 1). The patient with suspected gastrointestinal ulceration that was treated with dexamethasone was ultimately euthanized due to lack of control of the primary disease process.

All dogs were prescribed additional medications during GC therapy, including immunomodulatory agents, antibiotics, analgesics, and gastrointestinal support medications. At initial diagnosis, 85% (52/61) of dogs were prescribed mycophenolate mofetil (MMF), 2% (1/61) were prescribed azathioprine, and 13% (8/61) were not prescribed a secondary immunomodulatory agent. At some point during GC therapy, 53 dogs were prescribed MMF, 12 were prescribed cyclosporine, 3 were prescribed azathioprine, and 3 were not prescribed secondary immunomodulatory agents. Sixty-nine percent (42/61) were prescribed clopidogrel, 61% (37/61) were prescribed omeprazole, 52% (32/61) were prescribed doxycycline, 48% (29/61) were prescribed famotidine, 41% (25/61) were prescribed metronidazole, and 28% (17/61) were prescribed maropitant. Forty-eight percent (29/61) were prescribed antibiotics other than metronidazole or doxycycline, with enrofloxacin prescribed most often (26%; 16/61). Twenty percent (12/61) of dogs received tramadol or gabapentin; no dogs received nonsteroidal anti-inflammatory agents.

Adverse events

Higher body weight dogs were more likely to develop muscle atrophy [odds ratio (OR): 1.3 per 5 kg body weight (BW) increase; 95% confidence interval (CI): 1.1 to 1.7; adjusted P = 0.0175] and polyphagia (OR: 1.3 per 5 kg BW increase; 95% CI: 1.1 to 1.7; adjusted P = 0.0201) (Table 3). Categorization as overweight increased the odds of being diagnosed with a UTI (OR: 4.2; 95% CI: 1.0 to 22.0; P = 0.0495). Overweight dogs had an increased risk of developing cranial abdominal organomegaly nearing statistical significance (OR: 3.4; 95% CI: 1.0 to 13.0; P = 0.0564). Dogs diagnosed with UTI were prescribed a higher initial starting dose (mean: 2.18 mg/kg/d) than dogs not diagnosed with UTI (mean: 1.95 mg/kg/d; P = 0.03). No significant relationships with the initial dosing were noted for other evaluated adverse effects. Starting weight, body condition score, and initial diagnosis were not significantly associated with development of the remaining clinical signs or the number of moderate or severe adverse effects.

Table 3.

Starting weight as a predictor of adverse events during glucocorticoid therapy.

Adverse event	Yes/No	Individuals/Total (percentage)	Mean weight (kg) of affected	Weight (kg) Std Dev	P-valuea	Adj P-valueb	Odds ratio (95% CI)
Cranial abdominal organomegaly	No	27/61 (44.2%)	19.5	13.9	0.1473	0.4418
	Yes	34/61 (55.7%)	24.6	13.8
Dermatologic signs	No	39/61 (63.9%)	20.8	15.5	0.2445	0.6979
	Yes	22/61 (36.1%)	25.1	10.4
Gastrointestinal signs	No	38/61 (62.3%)	24.7	14.4	0.0748	0.2245
	Yes	23/61 (37.7%)	18.4	12.4
Lethargy	No	48/61 (78.7%)	23.5	14.5	0.2146	0.6438
	Yes	13/61 (21.3%)	18.3	11.3
Muscle atrophy/loss	No	36/61 (59.0%)	18.4	12.7	0.0058*	0.0175*	1.3 (1.1 to 1.7)
	Yes	25/61 (41.0%)	28.1	13.9
Panting	No	33/61 (54.1%)	19.5	12.5	0.0767	0.2302
	Yes	28/61 (45.9%)	25.7	15
Polyphagia	No	27/61 (44.3%)	17.2	10.6	0.0067*	0.0201*	1.3 (1.1 to 1.7)
	Yes	34/61 (55.7%)	26.4	15.1
Polyuria	No	20/61 (32.8%)	18.2	13.4	0.0868	0.2604
	Yes	41/61 (67.2%)	24.4	13.9
Urinary tract infection	No	44/61 (72.1%)	21	12.6	0.2118	0.6354
	Yes	17/61 (27.9%)	25.9	16.8

^aValues of P from univariate analyses.

^bValues of P, adjusted for multiplicity with the Adaptive Hochberg method.

^*P < 0.05.

A total of 95.1% of dogs (n = 58) experienced weight change exceeding VCOG Grade 0: 4.9% (n = 3) experienced Grade 0 change, 19.7% (n = 12) Grade 1, 31.1% (n = 19) Grade 2, and 44.3% (n = 27) Grade 3. The greatest magnitude of weight change from baseline was negative (weight loss) for 25 dogs and positive (weight gain) for 36 dogs. Dogs experiencing Grade 3 weight change were lower body weight compared to dogs with Grade 1 weight change (P = 0.01; Table 4).

Table 4.

VCOG-CTCAE grade distribution for weight change and laboratory data.

Laboratory Value/Weight	VCOG grade	Individuals/Total (percentage)	Mean weight (kg) of affected	Weight (kg) Std Dev	P-value
ALP	0	3/54 (5.6%)	12.0	2.0	0.33
	1	6/54 (11.1%)	20.9	18.8
	2	7/54 (13.0%)	25.2	15.9
	3	23/54 (42.6%)	22.6	11.2
	4	15/54 (27.8%)	27.0	16.0
ALT	0	8/55 (14.5%)	19.2	9.4	0.5
	1	2/55 (3.6%)	24.7	19.7
	2	19/55 (34.5%)	25.1	11.9
	3	22/55 (40.0%)	25.1	16.5
	4	4/55 (7.3%)	13.9	11.2
Neutrophil count	0	14/61 (23.0%)	21.1	10.8	0.81
	1	8/61 (13.1%)	18.8	9.3
	2	24/61 (39.3%)	23.6	18.0
	3	13/61 (21.3%)	22.9	12.3
	4	2/61 (3.3%)	27.3	10.0
Platelet count	0	22/61 (36.1%)	24.8	12.8	0.06
	1	22/61 (36.1%)	21.1	15.1
	2	13/61 (21.3%)	25.0	13.5
	3	3/61 (4.9%)	7.9	1.5
	4	1/61 (1.6%)	4.6	0.0
Weight change	0	3/61 (4.9%)	24.4	12.7	0.01*
	1	12/61 (19.7%)	30.5	15.3
	2	19/61 (31.1%)	24.1	13.3
	3	27/61 (44.3%)	16.7	12.7

VCOG grades were calculated by comparing laboratory values just prior to initiation of GC therapy with the highest laboratory values observed at follow-up in patients receiving GC therapy.

^*P < 0.05.

During the course of GC therapy, polyuria was reported in 67.2% of dogs (n = 41), polyphagia in 55.7% (n = 34), cranial abdominal organomegaly in 55.7% (n = 34), panting in 45.9% (n = 28), muscle atrophy in 40.9% (n = 25), gastrointestinal signs in 37.7% (n = 23), dermatologic signs in 36.1% (n = 22), UTI in 27.9% (n = 17), and lethargy in 21.3% (n = 13) (Table 3). Gastrointestinal bleeding and/or ulceration were noted in 6.6% of cases (n = 4: 1 duodenal ulceration confirmed by endoscopy, 1 gastric ulcer based on abdominal ultrasound, 2 based on presence of melena). Other adverse effects reported in single dogs are included in Table 1. Considering all adverse effects, the median number of mild adverse effects for each patient was 3 (range: 0 to 7); median 1 for moderate adverse effects (range: 0 to 3); and median 0 for severe adverse effects (range: 0 to 2).

Starting weight remained statistically insignificant (P > 0.15) in multivariable models, so final analyses were unadjusted for confounders. Overweight status was not significantly associated with muscle atrophy (P = 0.83) or polyphagia (P = 0.72). Cranial abdominal organomegaly developed in 70% (n = 14/20) of overweight dogs, lethargy was noted in 45% (n = 9/20), and 55% (n = 11/20) were diagnosed with a UTI. Moderate severity adverse effects was the only dependent variable that met the minimum criteria for number of dogs to assess in the multivariable model with body condition score and starting body weight as independent variables (P = 0.75). A diagnosis of ITP was associated with lethargy (P = 0.05), but the limited number of dogs with lethargy (n = 13) precluded its consideration as a potential confounder using multivariable modeling.

Laboratory data

There were no significant differences in starting weight or starting dose of steroid between VCOG grades for aberrant ALP, ALT, neutrophil counts, or platelet counts (Table 4). There were sufficient data to include all dogs in analyses of neutrophil and platelet counts; 7 dogs had insufficient data to be included in an analysis of ALP; 6 dogs had insufficient data to be included in an analysis of ALT. Platelet counts were analyzed separately for dogs with IMHA and ITP, and no additional statistically significant differences in VCOG grades were elucidated.

Discussion

Dogs with higher body weights were at increased risk for select adverse effects attributable to GC using standard linear dosing, compared to dogs with lower body weights. The odds of developing muscle atrophy or polyphagia during therapy each increased by 30% per 5 kg of additional body weight. Although not typically life-threatening, these adverse effects can negatively impact patient quality of life, owner compliance, and the human-animal bond. In addition, these adverse effects may signal systemic implications not fully elucidated by the data collected in this study. The increased risk of polyphagia and muscle atrophy in higher body weight dogs warrant attention when constructing treatment protocols for these dogs, to minimize disruptions of therapy or decreased quality of life.

The findings of this study echo other studies in recognizing polyuria, polyphagia, and panting as 3 of the most reported adverse effects in dogs receiving GC therapy (Table 1) (9,12,13). Urinary tract infection was diagnosed in 27.9% of dogs, compared to 39% and 18.1% in the aforementioned studies. Low urine osmolarity and immunosuppression are considered major factors for development of UTI, but in dogs receiving GC, it is possible that additional factors, such as alterations in mucosal immune response, may contribute to development of UTI. In this study, dogs classified as overweight were more likely to be diagnosed with UTI. This may reflect a failure to adjust GC dosing, resulting in greater immunosuppression than would happen at an ideal body weight. Another potential explanation is the anatomic impact of obesity, with adipose deposition trapping moisture and bacteria near the urethral opening, which would predispose for ascending infection. Previous studies identified an increased risk of UTI with long-term GC therapy, often with minimal or no clinical signs (15,16).

With standard linear dosing, higher body weight dogs may be prone to greater incidence of adverse effects compared to lower body weight dogs. Cancer chemotherapeutic drugs are commonly prescribed based on body surface area (BSA) rather than weight because BSA is thought to better correlate with physiologic variables (19–22). One pharmacokinetic study compared plasma drug concentrations in large breed dogs (mean: 29 ± 2.9 kg) and small breed dogs (mean: 9.9 ± 1.7 kg) when dosing prednisolone by body weight versus dosing by BSA (23). Nam et al (23) revealed that plasma concentrations were significantly greater in large breed dogs dosed at 2.0 mg/kg/d compared to smaller breed dogs dosed at 2.0 mg/kg/d or large breed dogs dosed at 40 mg/m² BSA, indicating that mg/kg dosing may cause larger dogs to receive doses higher than needed for therapeutic effect. In contrast, large breed dogs receiving 40 mg/m² dosing had plasma drug levels significantly less than those of small breed dogs given 2 mg/kg dosing, indicating that dosages based on BSA may result in subtherapeutic concentrations in larger breed dogs (23). The differences in dosing by body weight versus BSA become more apparent as body weight increases (Figure 1).

Figure 1.

Comparison of mg/kg versus mg/m² dosing schemes for administration of prednisone or prednisolone to dogs.

All but 3 dogs in our study experienced weight changes > 5% of their original weight, with 44.3% of dogs fluctuating > 15% from baseline. Dogs experiencing Grade 3 weight changes were of lower body weight compared to those experiencing Grade 1 weight change. This observation may be due to differences in body composition and caloric requirements of large versus small dogs receiving GC therapy. Regardless of cause, the weight fluctuations noted in most dogs suggest that close monitoring of weight change and appropriate adjustment of dose are indicated.

Although there is evidence in support of BSA-based dosing in veterinary medicine, its use has not effectively prevented toxicosis in dogs. Administration of melphalan, cisplatin, carboplatin, vinblastine, and doxorubicin based on BSA-based calculations of dose resulted in greater toxicosis in small dogs than in large dogs (24). The BSA-based dose calculation has been criticized for excluding size-independent factors capable of influencing drug distribution, metabolism, and excretion (21). Accounting for breed-associated conformation and metabolic differences poses additional challenges to calculating dosages according to the conventional formula for body surface (24). For GC dosage, calculation of the appropriate dose using BSA may be more appropriate in heavier dogs, as suggested in an ACVIM consensus statement crafted by Swann et al (6), who recommended a dose of 50 mg/m², rather than the 40 mg/m² dose evaluated by Nam et al (23). The current study, coupled with the previously published reports, supports the need for additional investigation to guide the establishment of a GC dosing strategy, including determination of dose-related alterations in lymphocyte activation and proliferation, as well as cytokine production, coupled with pharmacokinetics and pharmacodynamics.

Limitations of the present study are those typical of retrospective investigations, including incomplete patient data and lack of standardization of treatment and follow-up. Recording of clinical signs requires reporting by client and/or veterinarian for entry into the medical record, which may result in under-reporting of perceived mild clinical signs. The reporting of muscle atrophy was subjective in the present study. Further studies with objective muscle mass data, such as quantification with a tape measure, may be warranted, particularly because hair coat and dog breed can influence subjective assessments. Laboratory testing and imaging were not standardized, which could influence the identification of adverse effects, such as subclinical UTI, gastrointestinal bleeding, or thromboembolic events. Sample size may also be a factor in the low number of adverse effects that showed statistically significant relationships to starting weight, and an expanded cohort may allow identification of additional significant relationships.

Treatment of immune-mediated conditions typically involves an initially high dosage of GC, followed by gradual tapering based on disease process, coupled with response to and tolerance of therapy. Variability in steroid tapering protocols was an important limitation of the present study, because receiving higher GC dose for a longer duration likely affects development of adverse effects. Although most dogs in this study were tapered over the course of several months, we cannot rule out that variability in tapering duration affected incidence and severity of documented adverse effects. The authors’ institutional bias is thought to be toward more rapid GC taper for higher body weight dogs. Controlling for this bias would likely help illuminate the adverse effects differences between higher and lower body weight dogs. Administration of multiple medications in addition to GC therapy was another important study limitation. Polyphagia and muscle atrophy are not typically associated with the other medications prescribed in this dataset, but relationships between starting weight and other adverse effects may have been obscured by overlapping side effects.

The data collected suggest that, with standard linear GC dosing, dogs with higher body weight may be at increased risk for developing adverse effects. The authors aim to increase clinician awareness of this problem during construction of treatment protocols and to prompt further investigation into optimal GC dosing strategies, particularly for heavier body weight dogs. Dosing of GC based on BSA or other novel strategies may be more appropriate, and this will best be guided by additional investigation of the downstream impacts of dose-related GC administration on lymphocyte activation, proliferation, and cytokine production, coupled with pharmacokinetic and pharmacodynamic data.