An abstract of this work was presented at the 24th Annual ACVIM Forum in Louisville, KY, June 2006.
Corresponding author: L.A. Cohn, DVM, PhD, DACVIM (Small Animal Internal Medicine), 379 E. Campus Drive, Columbia, MO 65211; e-mail: firstname.lastname@example.org.
Background: Inhaled glucocorticoids reduce airway inflammation while minimizing systemic effects in several species.
Hypothesis: Inhaled fluticasone suppresses the hypothalamic–pituitary–adrenal axis (HPAA), modifies immune function, and induces clinical signs to a lesser extent than PO-administered prednisone in dogs.
Animals: Seven healthy adult pet dogs.
Methods: Dogs were randomized to 1 of 3 treatment groups in a crossover design: fluticasone propionate (220 μg actuation of a metered dose inhaler delivered via a spacer and mask, q12h), placebo (spacer and mask alone, q12h), or prednisone (1 mg/kg PO q24h). Each treatment was administered for 3 weeks followed by a 4-week washout. Appetite, attitude, and water consumption were recorded during the last week of each treatment period. Urine cortisol : creatinine ratios, ACTH stimulation tests, white blood cell counts, lymphocyte phenotype, and serum IgM and IgA concentrations were recorded at each baseline and after the last day of each treatment. Clinical observations were expressed descriptively. Friedman's test was applied to all data comparisons. Pairwise comparisons were made with a mixed model analysis when data were normally distributed, whereas signed rank tests were used otherwise (significance P-value <.01).
Results: Appetite and water consumption increased during prednisone treatment. Peak serum cortisol concentrations post-ACTH were significantly decreased in prednisone- and fluticasone-treated dogs compared with placebo (prednisone > fluticasone). Serum IgM concentrations were significantly decreased in dogs treated with prednisone.
Conclusions and Clinical Importance: As used, fluticasone suppresses the HPAA to a lesser extent than prednisone and may avert systemic signs associated with PO-administered glucocorticoids in dogs.
Chronic bronchitis, a neutrophilic inflammatory airway disease, is a common respiratory disorder in dogs.1,2 Although mortality from chronic bronchitis is rare, morbidity related to cough can be severe. Traditional treatment for dogs with chronic bronchitis has relied on the anti-inflammatory effects of PO-administered glucocorticoids (GC) in combination with cough suppressants, bronchodilators, and the periodic use of antimicrobials.1,2 Unfortunately, not all dogs can tolerate PO-administered GC, especially when given long term. Adverse effects include polyuria, polydipsia, behavioral changes, polyphagia, weight gain, hair loss, excessive bruising, hepatomegaly, and serum biochemical and hematologic changes.3,4 GC are relatively contraindicated in dogs with active concurrent infections (including respiratory infections), predisposing factors for infection, certain types of heart disease, and diabetes mellitus, and in those receiving concurrent treatment with nonsteroidal anti-inflammatory drugs.3 Additionally, GC treatment can result in suppression of the hypothalamic–pituitary–adrenal axis (HPAA).4–7
Inhalational administration of drugs is designed to minimize systemic effects while preserving local respiratory efficacy. Topical delivery of GC by metered dose inhalers (MDI) has largely replaced the use of PO-administered GC in people with asthma, chronic bronchitis, and other inflammatory airway diseases.8 Through the use of spacer devices and face masks, these same MDI delivery systems for GC have been successfully adapted for use in both cats and horses with inflammatory airway disease.9–12 Similarly, inhalant GC could be adapted for use in dogs with chronic bronchitis.13 Although inhalant delivery may minimize systemic absorption of GC, it does not eliminate systemic effects altogether. Adrenal suppression has been reported in humans, horses, and cats receiving inhaled GC.10,14–16 Likewise, alterations in systemic immunologic parameters have been described in these same species.17–19
There is little published information regarding the effects of clinically relevant inhalant GC administration to dogs,13 and we are aware of only a single prospective study examining the effects of any inhalant GC in this species.20 The purpose of the study reported here was to compare systemic clinical, endocrine, and immunologic effects of inhaled GC (220 μg fluticasone q12h delivered through a spacer device) with PO-administered GC (1 mg/kg prednisone q24h) and placebo. We hypothesized that in healthy dogs the inhalant GC would result in fewer observable clinical effects and less HPAA and immunologic suppression than the PO-administered GC. Appetite, water consumption, and attitude were observed. Endocrine effects were evaluated by means of ACTH stimulation testing and urine cortisol : creatinine ratio (UCC). Immunologic parameters included determination of absolute white blood cell (WBC) counts, flow cytometric determination of peripheral blood mononuclear cell phenotype, and measurement of serum immunoglobulin (IgM and IgA) concentrations.
Materials and Methods
Seven privately owned dogs were enrolled in the study with informed owner consent. Dogs included 5 Border Collies (1 intact male, 2 castrated males, and 2 intact females) and 2 spayed female coonhounds. Border Collie dogs did not carry the multidrug sensitivity mutation known as MDR1. Estimated or actual ages ranged from 2 to 11 years. Body weights ranged from 14 to 36 kg (mean ± SD: 23.9 ± SD 7.7). All dogs were from a single household and were housed both indoors and in individual outdoor kennels with separate water containers. Dogs were determined to be healthy by complete physical examination as well as screening CBC, serum biochemical profile, urinalysis, and thoracic radiographs. All experimental procedures were reviewed and approved by the University of Missouri Animal Care and Use Committee.
A randomized, crossover, placebo-controlled prospective study design was used. Dogs were assigned to treatment groups by means of a random number table. Treatment groups included PO-administered GC (prednisone 1 mg/kg PO q24h), inhaled GC (fluticasonea 220 μg q12h), or placebo. Fluticasone was delivered by attachment of the MDI to a 1-way valved spacer deviceb attached to an anesthetic face mask. The mask was fitted over the dog's face, the MDI was activated, and the dog was allowed to take 12 breaths into the mask. The placebo consisted of a separate mask and spacer device but without an attached MDI. Each dog was treated for 3 weeks, followed by a 4-week washout period. The washout period was followed by the 2nd assigned treatment for 3 weeks and then a final 4-week washout period. The final washout period was followed by a final 3-week treatment. Each dog received all 3 treatments during the course of the study. Dogs underwent evaluation at each baseline (before the 1st treatment or on the last day of each washout period) and again after the last day of each treatment period; drug was not administered on the morning of the evaluation. Treatments were not administered by the same people who made the subjective observations; personnel who completed both subjective and objective assessments remained unaware of assigned treatment.
Attitude, appetite, and water consumption for each dog were assessed during the final week of each 3-week treatment period. Assessments were made by a single owner who was unaware of which treatment each dog received. During the final week of administration for each of the 3 treatments, dogs were housed in individual kennels approximately 23 h/d. Water and food consumption were measured daily in an approximately quantitative manner. Water buckets were marked in increments of 250 mL and refilled each morning. The following morning, the amount of water consumed during the previous 24 hours was recorded. Similarly, the dogs were fed a specified amount each morning and the remaining food was measured the following morning. Attitude was assessed in an entirely subjective manner; the owners simply noted any changes in the dogs' usual activity. Specific gravity (SG) of an early morning urine sample was obtained at each baseline and again at the end of each treatment period.
The HPAA was evaluated at the baseline of each treatment and at the end of each treatment period. Early morning urine was collected at home by free catch and frozen at −20 °C until analysis. Urine cortisol was assayed by radioimmunoassayc validated for use in canine urine21,22 and urine creatinine was measured in a routine manner in the same laboratory. UCC ratio was calculated from the 2 results (reference range, 8–24). Additionally, ACTH stimulation was performed at the dogs' homes before 10:00 am. Two milliliters of whole blood was collected aseptically from the jugular vein for cortisol measurement, followed by administration of 5 μg/1 kg synthetic ACTH IV.d A postinjection blood sample was obtained 60 minutes later. Samples were stored and transported in a cooler and centrifuged at 800 ×g for 20 minutes within 2 hours of collection. Serum was frozen at −20 °C until analysis at the University of Missouri Veterinary Medical Diagnostic Laboratory by a chemiluminescent assaye validated for use in canine serum.23
CBC, flow cytometric evaluation of lymphocyte phenotype, and serum IgM and IgA concentrations were assessed at each baseline and at the end of each treatment. Two milliters of blood was collected into a clot tube and centrifuged at 1720 ×g for 20 minutes, and the serum was frozen at −20 °C until immunoglobulin analysis. Serum total IgMf and IgAg concentrations were measured with commercially available ELISA test kits according to the manufacturer's instructions. An additional 4 mL of blood was collected by jugular venipuncture into an EDTA tube for CBC and flow cytometry.
Phenotypic analysis of lymphocytes was performed with flow cytometry according to a standard protocol. Briefly, 100 μL of anticoagulated blood (EDTA) was placed in 12 × 75 mm tubesh for each antibody/antibody pair as well as for cells only (unstained control). The following antibodies were used to identify subpopulations of lymphocytes in 3 separate tubes per dog: anticanine CD3/CD4 10 μL,i CD3/CD8 10 μL,j and CD21 5 μLk (pan-B cell marker). Samples were incubated in the dark at room temperature for 15 minutes and the RBCs were lysed with 1 mL of lysing reagent for 5 minutes (8.26 g NH4Cl, 1.0 g KHCO3, 0.037 g Na2EDTA in 1.0 L deionized distilled H2O, pH 7.2). Samples were washed twice with FACS buffer (phosphate-buffered saline [PBS] with 1% fetal bovine serum and 0.09% sodium azide) and pelleted (300 ×g for 6 minutes). Pellets were resuspended in 200 μL of FACS buffer and 200 μL of formalin in PBS (pH 7.4) at 4 °C until read (within 5 days). The lymphocyte populations recognized on the forward and side scatter plots were gated to identify the fluoresceinated antibodies listed above on FL1 versus FL2 plots. Results were determined both as a percentage of lymphocytes with each marker and by using values determined on CBC, as an absolute number of lymphocytes with each marker.
Statistical analysis was accomplished by SAS v9 software.l Data from each of the 3 treatments were subjected to Friedman's test. When treatment differences were apparent for normally distributed data, pairwise comparisons were made. A mixed model analysis procedure in which the subject was considered a random effect and the treatment a fixed effect was used; the baseline for each treatment was used as a covariate. Pairwise comparison of the least squares means was then employed. When parameteric analysis was not appropriate because the data were not normally distributed, signed rank tests were used for comparisons. Values >2.5 SD from the mean were deemed outliers. A P value of <.01 was chosen to represent statistical significance because multiple variables were compared; when .01<P <.05, marginal significance was said to exist.
All dogs accepted inhalational treatment without difficulty. As judged subjectively, the only alterations in attitude observed during the study were that 2 dogs became less active during their daily exercise period while receiving prednisone. Appetite was assessed as increased in 4 of 7 dogs during prednisone treatment, but no alterations in appetite were observed in any dog during placebo or fluticasone treatments. Body weight did not change significantly after any treatment and no differences were detected among treatments. The mean water consumption (Fig 1) was greater after treatment with prednisone (mean ± SD: 1.99 ± 0.25 L/d; P <.0001) as compared with water consumption in placebo (1.23 ± 0.22 L/d) or fluticasone-treated dogs (1.11 ± 0.27 L/d). There was no difference in water consumption after fluticasone treatment versus treatment with placebo (P=.41). During the 3 baseline evaluations, only 1 of 21 measurements of urine SG was <1.035 (ie, 1.034). After treatments, urine SG of <1.035 was documented in no placebo-treated dogs, in 4 prednisone-treated dogs, and in 1 fluticasone-treated dog. However, no statistically significant differences were detected between urine SG at baseline or after completion of any 3-week treatment (Fig 2).
No differences were detected among early morning, unstimulated serum cortisol concentrations at any of the 3 baselines. Likewise, no differences were detected among ACTH-stimulated serum cortisol concentrations at any of the 3 baselines. Early morning serum cortisol concentrations before ACTH administration were marginally lower after treatment with prednisone (0.34 ± 0.30 μg/dL) than after placebo treatment (1.61 ± 0.95 μg/dL; P=.0129) but neither differed from concentrations found in fluticasone-treated dogs (0.91 ± 0.50 μg/dL; P=.1 and .091 placebo or prednisone, respectively). Significant differences were detected among all 3 treatments in peak cortisol concentrations after ACTH administration (Fig 3; P <.0004 in each case). Peak cortisol after ACTH stimulation was significantly lower in prednisone-treated dogs than in fluticasone- or placebo-treated dogs and was also lower in fluticasone-treated dogs than in placebo-treated dogs. UCC ratios did not differ significantly among the drug treatments nor were there differences detected between baseline and posttreatment samples for any of the treatments.
No differences were detected among any baseline immunologic variables. Absolute numbers of WBC, neutrophils, monocytes, lymphocytes, and eosinophils were not statistically different in any treatment group. Although statistically significant differences in eosinophil numbers among treatments were not detected, 4 of 7 dogs had >250 eosinophils/μL after placebo treatment, eosinophils were not detected in any dog after prednisone treatment, and eosinophils (79 eosinophils/μL) were identified in only 1 dog after fluticasone treatment. No significant differences among treatment groups were detected in lymphocyte phenotype (data not shown). Serum IgM concentrations decreased significantly (Fig 4) in dogs treated with prednisone as compared with placebo-treated dogs (P=.0013) or to fluticasone-treated dogs (P= .0032). Marginally significant decreases in serum IgA (Fig 5) were detected in prednisone-treated dogs as compared with placebo-treated dogs (P=.0434) and in fluticasone-treated dogs as compared with placebo-treated dogs (P=.0146).
We hypothesized that a commonly used inhalant GC formulation (ie, fluticasone) used at a clinically relevant dose would have less systemic impact on healthy dogs than PO-administered GC (ie, prednisone) used at an anti-inflammatory dosage. We chose what is commonly described as an anti-inflammatory dosage of prednisone for this study (1 mg/kg/d); this dosage has been suggested specifically for the treatment of chronic bronchitis in dogs.1,2,13 Although the dosage chosen was typical of an initial dosing scheme, prednisone generally is tapered to the lowest effective dose (often administered on an every other day basis). Because the adverse effects of GC are dose-dependent, lower dosages of PO-administered prednisone would be likely lead to fewer or milder responses than those documented in our study.
Fluticasone is designed for topical respiratory use either as a nasal spray or for inhalant delivery via MDI or dry powder inhaler. The pharmacodynamics and pharmacokinetics of inhalant GC are complex, and do not readily allow comparison of dose potency between these formulations and traditional oral formulations.24 However, fluticasone is considered a potent GC with approximately 20-fold greater GC receptor affinity than dexamethasone.25 Despite its very potent anti-inflammatory effects in the airway, systemic absorption of fluticasone is limited.24,26,27 Although as much as 80% of the drug dose delivered by MDI is deposited in the oral cavity of humans, any swallowed fraction of fluticasone is inactivated by first-pass hepatic metabolism.24 A dose of 100–200 μg per actuation of fluticasone per day is appropriate for the treatment of inflammatory airway disease in most people.28 To the authors' knowledge, neither the pharmacokinetics nor the pharmacodynamics of fluticasone have been evaluated in dogs.
Common systemic effects of GC include polyuria, polydipsia, and polyphagia.3 Water consumption was significantly increased in prednisone-treated dogs but not in dogs treated with placebo or fluticasone. At the dosages used here, fluticasone was associated with less polydipsia, and subjectively perhaps less polyuria, than prednisone. Appetite was increased in 4 of 7 dogs during prednisone treatment but not in any dog during fluticasone treatment. Although body weights were unchanged during the 3-week treatment period, it is likely that chronic prednisone administration would have resulted in sustained increase in appetite and resultant weight gain.
The magnitude of suppression of the HPAA after treatment with GC depends on the biologically available GC dose and dose interval. The degree of HPAA suppression was significantly greater in dogs treated with prednisone than in dogs treated with fluticasone based on peak cortisol concentration post-ACTH stimulation. Additionally, marginal suppression in early morning cortisol concentration was documented after prednisone treatment as compared with placebo, further supporting greater suppression of the HPAA by prednisone as compared with fluticasone. Although differences were not detected in UCC ratio after any of the treatments, this test is used primarily as a screening tool for hyperadrenocorticism and is likely to be insensitive for the detection of HPAA suppression.29 As with these dogs, HPAA suppression has been documented in cats and horses treated with inhalant GC.10,16 The lesser degree of suppression of the HPAA after treatment with inhaled fluticasone as opposed to PO-administered GC agrees with what is generally documented in people.14,30–32
Although it was beyond the scope of this study to extensively investigate the immunologic effects of inhaled and PO-administered GC in the dog, we were able to evaluate and compare changes in peripheral WBC counts, lymphocyte phenotype, and serum IgM and IgA concentrations. Differences in WBC counts were not detected after treatment with either PO-administered prednisone or inhaled fluticasone. Classically, administration of GC to dogs is associated with a stress leukogram characterized by increases in neutrophil and monocyte numbers and decreased numbers of lymphocytes and eosinophils.33,34 In these dogs, mean lymphocyte counts remained within reference range and no statistically significant decreases in lymphocyte counts were detected after either fluticasone or prednisone treatment. Variable decreases in lymphocyte number have been documented in other studies in which healthy dogs were administered similar dosages of prednisone PO.34–36 Although changes in eosinophil counts in these dogs did not reach the level of statistical significance, type II statistical error in this small study group may have prevented us from identifying the expected GC-associated eosinopenia.34,35
We were unable to detect differences in the phenotype of peripheral lymphocytes after treatment with either PO-administered prednisone or inhaled fluticasone. Although decreases in CD4+, CD8+, and B lymphocytes were documented after treatment with 2 mg/kg/d prednisone in healthy dogs, we did not observe similar changes in the dogs receiving 1 mg/kg/d prednisone.37
The effects of GC on humoral immunity generally are indirect (ie, immunoglobulin concentrations are affected primarily by the effects of GC on antigen presentation, T-helper lymphocytes, and protein catabolism).38 Our study documented a significant decrease in serum IgM concentrations after treatment with prednisone but did not detect a similar decrease in serum IgM after fluticasone (or placebo). Similar but marginally significant (.01<P<.05) decreases in serum IgA were also detected after treatment with prednisone or fluticasone as compared with placebo. The decrease in IgM might imply a greater effect on systemic humoral immunity after treatment with prednisone as compared with fluticasone, but results from only 7 dogs must be interpreted cautiously.
At the dosages used in our study, prednisone but not fluticasone resulted in increased water consumption, increased appetite, and decreased serum immunoglobulin concentrations. Although both prednisone and fluticasone resulted in suppression of the HPAA, the degree of suppression was greater in prednisone-treated dogs. We did not attempt to assess the efficacy of fluticasone for the treatment of any disease in these healthy dogs, but inhaled GC are successfully used to decrease airway inflammation in several other species.8,17,39,40 If the efficacy of inhaled fluticasone is comparable to that of PO-administered prednisone for the treatment of inflammatory airway disease in dogs, our study demonstrates that the use of fluticasone would lessen the systemic GC response and might minimize adverse reactions associated with GC treatment.
This work was supported by a grant from the University of Missouri, College of Veterinary Medicine Committee on Research.
The authors would like to acknowledge the technical contributions of Laura Davis, Dewayne Davis, and Eli Davis. They would also like to thank Dr Katrina Mealey for assays of MDR1 mutation.
aFlovent HFA 220 μg inhalation aerosol, GlaxoSmithKline, Research Triangle Park, NC
bOptiChamber Advantage Valved Holding Chamber, Respironics Inc, Murrysville, PA
cDPC Coat-A-Count Cortisol RIA, Siemens AG, Tarrytown, NY, as run by Michigan State University Diagnostic Center for Population and Animal Health, Lansing, MI
dCortrosyn, Amphastar Pharmaceuticals Inc, Rancho Cucamonga, CA
eDCP Immulite chemiluminescent assay, Siemens AG, Tarrytown, NY