Page 25 - Appetite_Stimulation_In_Dogs_Complete

Capromorelin for Appetite Stimulation in Dogs 1853

Ò
also completed the Owner Appetite Assessment they had previ- binary outcomes (the GLIMMIX procedure in SAS , assuming a
ously completed at day 0. A comparison of the total scores (range binomial distribution and logit link) was used to compare the
of 5–25) at day 0 and on day 3 1 was made to assess changes in capromorelin group to the placebo group. The model contained
this score over time. In addition, body weights on day 0 and day success/failure as the dependent variable, with treatment as a
3 1 were compared. fixed effect, and random effects for site and treatment by site
interaction.
Serum chemistry, hematology and urinalysis data, physical
Safety
examination data, and adverse event data were summarized and
evaluated by treatment group using descriptive statistics.
Safety was evaluated by physical examinations (day 0 and day
3 1), clinical pathology tests (day 0 and day 3 1), and moni-
toring for adverse events and owner observations during the Results
4 days of dosing. An adverse event was defined as any observa-
tion, undesirable experience, or reaction in animals that was unfa- Study Population
vorable and unintended and occurred after the use of
Out of the 259 dogs screened at the 24 sites, 244 cases
capromorelin or placebo, whether or not considered to be treat-
ment related. A serious adverse event was defined as any adverse were enrolled in the study. A total of 244 dogs were
event that resulted in death, was life-threatening, or resulted in included in the safety population (intention to treat pop-
persistent or substantial disability/incapacity and occurred after ulation; ITT) with 171 randomized to treatment with
the use of capromorelin or placebo, whether or not considered to capromorelin and 73 to treatment with placebo. Sixty-
be product related. seven dogs were considered unevaluable for the purpose
of assessing effectiveness. Therefore, the per protocol
Statistical Analysis population (PPP) included a total of 177 dogs with 121
dogs in the capromorelin treatment group and 56 in the
The primary effectiveness variable was the single owner ques- placebo treatment group (Fig 1). In the safety popula-
tion on appetite answered on day 3 1. Treatment success was tion, dogs ranged in age from 4 months to 18 years of
defined as an “increased” appetite as determined by the owner. If age, with the mean age of 8.0 years in the capromorelin
the owner reported that the dog’s appetite was “no change” or
“decreased” on day 3 1 compared to day 0, the dog was classi- group and 8.1 years in the placebo group. The majority
fied as a treatment failure. of dogs were spayed or castrated, with equal numbers of
The primary null hypothesis was there will be no difference in males and females in the population (122 females, 122
the proportion of treatment success between capromorelin and pla- males). Body weights ranged from 1.5 to 76.5 kg with
cebo-treated dogs. For the primary efficacy variable, the number the mean weights in the capromorelin and placebo
and percentage of success was presented for each site and overall group of 17.1 and 16.3 kg, respectively (Table 1).
by treatment group. For binary outcomes (“success” rates), the Enrolled dogs had various medical conditions includ-
Ò
GLIMMIX procedure in SAS (assuming a binomial distribution ing allergy, arthritis, cardiovascular disease, gastroin-
and logit link) was used to compare the capromorelin group to the
testinal disease, renal disease, and others. It was not
placebo group. The model contained success/failure as the depen-
required that the veterinarian identify the etiology of
dent variable, with treatment as a fixed effect and random effects
the inappetence for the dog to be enrolled in the study,
for site and treatment by site interaction. In addition, the scores
and in many dogs, the underlying condition was not
from the Owner Appetite Assessment were analyzed as follows:
specified.
Treatment success was defined as an increase in total score of ≥5
points from day 0 to day 3 1. This analysis was also completed
excluding dogs with day 0 scores of ≥13 because although owners Diet
reported these dogs had exhibited a reduced appetite for 2 days
before enrollment, they had only moderate appetite decreases. Dogs enrolled in the study were fed a variety of foods
Finally, mean percent changes in body weight over the treatment in the base diet, with owners in some cases listing more
period in the 2 groups were compared as well as the percent of than 10 foods and treats. Because adding a new or
dogs which experienced a change in body weight of >0% (defined highly palatable food during the study was prohibited,
as a “success” in that they gained weight during the study).
For Owner Appetite Assessment total score and body weight, twelve cases were considered unevaluable because dogs
descriptive statistics (number of subjects, mean, standard devia- were given a food during the 4 days of treatment that
tion, standard error of the mean, 95% confidence intervals, med- was not listed on their base diet. Several cases were con-
ian, minimum and maximum values) were calculated by treatment sidered unevaluable because the dog actively sought out
group for each study day and for the percent change from day 0 and consumed foods not on their base diet.
to day 3 1. Analysis of variance (ANOVA) modeling was uti-
lized to test for possible differences between treatment groups in
the percent change from day 0 by a mixed model (the MIXED Effectiveness
Ò
procedure in SAS ). The model contained treatment as a fixed Appetite Assessment. Using the single owner question
term with site and treatment by site interaction as random terms. on day 3 1, where owners could rate their dog’s
Assumptions of normality of residuals and homogeneity of vari-
ance were investigated before running the model. If it was deter- appetite as “increased,” “no change” or “decreased,”
mined that the distribution could not be approximated by a and where only the “increased” answer was defined as
normal curve, then values were ranked in ascending order with treatment success, 68.6% (95% CI, 59.7, 76.3) of dogs
tied values being given a mean rank before running statistical in the capromorelin-treated group were classified a
models. For each criterion, the number and percentage of success treatment success, compared to 44.6% (95% CI, 32.2,
was presented by treatment group. The method appropriate for 57.8) of the placebo-treated dogs (P = .008).