Temporal changes in clinical and radiographic variables in dogs with preclinical myxomatous mitral valve disease: The EPIC study.

BACKGROUND
The Evaluation of pimobendan in dogs with cardiomegaly caused by preclinical myxomatous mitral valve disease (EPIC) study monitored dogs with myxomatous mitral valve disease (MMVD) as they developed congestive heart failure (CHF).


OBJECTIVES
To describe the changes in clinical and radiographic variables occurring as dogs with MMVD and cardiomegaly develop CHF, compared to similar dogs that do not develop CHF.


ANIMALS
One hundred and thirty-five, and 73 dogs that did or did not develop CHF, respectively.


MATERIALS AND METHODS
The following variables were evaluated in 2 groups of dogs (dogs that did or did not develop CHF): Heart rate (HR), clinic respiratory rate (RR), home-measured resting respiratory rate (RRR), rectal temperature (RT), body weight (BW), and vertebral heart sum (VHS). Absolute value and rate of change of each variable were calculated for each day a dog was in study. Daily means were calculated and plotted against time. The onset of CHF or last visit before leaving the study were set as reference time points.


RESULTS
The most extreme values and rate of change occurred in variables immediately before onset of CHF. Vertebral heart sum increased earliest. Heart rate, RR, and RRR also increased. Rectal temperature and BW decreased. Increases in RR and RRR were most extreme and occurred immediately before CHF.


CONCLUSIONS AND CLINICAL IMPORTANCE
Dogs with MMVD and cardiomegaly experience increases in HR, RR, RRR, and VHS, and decreases in BW and RT as they develop CHF. The variables with highest absolute change and rate of change were RR and RRR. These findings reinforce the value of RR and RRR as indicators of impending or incipient CHF.

RR and RRR. These findings reinforce the value of RR and RRR as indicators of impending or incipient CHF. Previous studies have identified not only the absolute changes in clinical, radiographic, and echocardiographic variables in patients with progressive MMVD, but the time course over which these changes occur. One study documented the change in vertebral heart sum (VHS) that occurred in Cavalier King Charles Spaniels before the onset of CHF. 3 The change in heart size occurred over approximately 1 year and was maximal immediately before the onset of CHF. 4 Longitudinal changes in echocardiographic indicators of heart size, 5 heart rate (HR), and heart rate variability 6 before patients succumb to MMVD indicate a similar pattern with an increasing rate of change of the monitored variables as the disease progresses. The increase in heart size before the onset of CHF occurs so consistently that heart size has been validated as a component of a regression equation that can be used with reasonable accuracy to predict the onset of CHF. 7 Dogs developing CHF in the EPIC study showed increased heart size and HR at the time of onset of CHF compared to measurements made at baseline. 1 Respiratory rate (RR), measured in the clinic, and client-measured resting respiratory rate (RRR), measured in the dog's home environment, have been shown to be higher in dogs in CHF 8,9 returning to more normal values once signs of heart failure are medically controlled. 10 The longitudinal change in RR and RRR in a large population of dogs as they develop CHF has not been described previously. Dogs with preclinical left-sided heart disease have been shown to have a RRR that usually is <25 breaths per minute. 11 Respiratory rate measured in the clinic was significantly higher at the time of onset of CHF in dogs in the EPIC study compared to results obtained at baseline. The RRR measured by owners in the home environment was not available at the baseline visit for comparison with results obtained immediately before CHF. 1 As well as the previously reported changes in HR, heart size, and RR, dogs in the EPIC study that developed CHF experienced decreases in rectal temperature and body weight as they developed CHF. 1 In the EPIC study, over 100 dogs were monitored as they developed CHF. 2 Data from this population therefore provide a unique opportunity to determine not only the magnitude of the change in clinical and radiographic variables as CHF develops, but the time course over which those changes occur.
The aim of our current study is to describe, in a group of dogs with stage B2 12 MMVD, the temporal changes in clinical and radiographic variables that occur before development of CHF and to contrast the changes in dogs that developed CHF to a population of dogs known not to have developed CHF over the duration of the study.

| Trial design
The EPIC trial was a prospective multicenter, blinded, randomized, placebo-controlled study. Complete and detailed description of the study and longitudinal changes in measured variables have been published. 1,2 The study was approved by an ethical review committee at each site where this was required.

| Exclusion criteria
Dogs were excluded from the study if they had any of the following: known clinically important systemic or other organ-related disease that was expected to limit the dog's life expectancy or required chronic administration of cardiovascular medication precluded as part of the trial. Dogs with hypothyroidism could be included provided the investigator deemed them clinically stable on treatment. Dogs with current or previous evidence of cardiogenic pulmonary edema, pulmonary venous congestion or both, cardiac disease other than MMVD, clinically significant supraventricular, ventricular tachyarrhythmias or both (ie, requiring antiarrhythmic treatment), or evidence of pulmonary hypertension considered to be clinically relevant (right ventricular: right atrial pressure gradient >65 mmHg) were excluded. Dogs with a history of chronic or recent administration (>14 days of duration or within 30 days of intended enrollment) of any precluded medication were excluded. In the event that, before study enrollment, a dog had received short-term treatment (<14 days) with a precluded agent, but was no longer receiving treatment and had not received it within 30 days of intended enrollment, then the dog was eligible for inclusion. Dogs that were pregnant or lactating were not eligible for enrollment.
Details of study sites, randomization and blinding, trial medication, concomitant treatment, and data management have been described previously. 2

| Schedule of events
Before inclusion, a case history was taken for each dog. At a baseline visit, dogs underwent physical examination, echocardiography, thoracic radiography, and routine hematology and blood biochemistry. Reexaminations were scheduled at 35 days after the baseline visit, approximately 4 months after baseline and every 4 months thereafter. Details of examinations that were undertaken on each visit are provided in Table 1.

| Clinical evaluation
At inclusion, dog characteristics such as breed, age, sex, and neuter status were noted. The BW, HR, RR, and rectal temperature (RT) were measured at each visit. Cardiac size was assessed by the VHS method, 15 and pulmonary edema and congestion were recorded, when considered to be present, by the attending cardiologist.

| Resting respiratory rate
Owners also were asked to measure RRR before every reexamination.
Instruction was provided to owners who were advised to count the T A B L E 1 Schedule of procedures undergone by animals remaining in the per-protocol population of the study at different examinations

Baseline visit
Baseline +35 days (±7 days) Baseline +4 months a Baseline +8 months b Event Physical examination X X X X X Thoracic radiograph X X X Owner measured respiratory rate X X X X a Four months from baseline visit and every 8 months thereafter. b Eight months from baseline visit and every 8 months thereafter.
The method of calculation of imputed values of continuous variables and the rate of change of variables respiratory rate over 1 minute, in the dog's home environment, within the week before each reexamination. Ideally, determination was performed on several days and the average of the measurements obtained was recorded as a single value at the corresponding visit.

| Primary endpoint
The primary endpoint was a composite of the development of leftsided CHF verified by an endpoint committee, 2 euthanasia for a cardiac reason, or death presumed to be cardiac in origin.

| Statistical methods
For those dogs with verified CHF (Group CHF), the day of confirmed CHF was considered day 0. The day of measurement of different variables was expressed as the number of days before day 0. For example, if a measurement was taken 7 days before the onset of CHF, this day would be considered day −7.
For comparison, a group of dogs that were known not to have developed CHF (Group no-CHF) for at least 4 months after their final examination were included. This group consisted of dogs that were censored at the time of study closure that were known not to have experienced CHF by that time (March 1, 2015), and for which sufficient data were available to contribute to the analyses. Day 0 for no-CHF dogs was defined as the day of the last visit on which all radiographic and clinical examination data were obtained (see Table 1) that was at least 4 months before closure of the study.
The values for BW were indexed to the baseline value of this variable because of the large range of values seen in the population.   The baseline characteristics of the 2 groups are summarized in Table 2  show that RR and RRR are comparably stable during the progression into CHF until late stages. Both RR and RRR only appeared to change within the 4 months immediately before the onset of CHF and showed little variation before that time. It is also interesting to note that a lower proportion of values for RRR were excluded from analyses because of presumed panting. This may suggest that measures of RRR are less likely to be confounded by panting and, being obtained by averages obtained over several days, are less prone to being affected by outlying observations.

| Continuous variables and their rate of change
A nonlinear increase in VHS previously has been described in dogs and the findings of the present study agree with these studies 3,4 (ie, the changes were characterized by a gradual increase over time Heart rate previously has been shown to increase over time in MMVD dogs as they progress into CHF. 6,16 In our study, small increases in HR could be seen from 10 months before confirmed CHF, but no such change could be observed in dogs not progressing to CHF within 4 months of the last examination. Mean heart rate increased from approximately 125 to 150 beats per minute, representing a change of 20% from the baseline value. The increase in HR probably is brought about by a response to worsening MR causing a decrease in forward stroke volume, which is compensated by increased sympathetic tone leading to increased HR. 17,18 Body weight and RT both decreased over time as dogs progressed into CHF. The decrease in RT is likely to be a consequence of impaired peripheral perfusion. 19 The decrease in BW may be a consequence of the loss of skeletal muscle mass, but we did not record body condition or muscle condition scores at each visit and therefore cannot be sure of this conclusion. Loss of muscle mass may occur as a consequence of cardiac cachexia. 20 Cardiac cachexia is a complex syndrome that may be, at least in part, a consequence of systemic inflammation 21 and intestinal underperfusion. 22 Furthermore, during the development of CHF, dogs may show evidence of a decreased appetite, which will decrease calorie intake and predispose to weight loss. 20 Decreased appetite previously has been associated with worse outcome, both in the present study population of EPIC trial dogs 2 and other studies. 23 Cardiac cachexia has been associated with shorter survival times in dogs with heart disease. 21,24 The magnitude of the change observed in both BW and RT were comparably small in relation to the natural variability observed; an approximately 3% change in BW and 0.2 C change in RT. This observation indicates that, although these changes were evident on a population level, they are very unlikely to be detectable in an individual dog. A previous study reported a difference in RT of 0.3 C between dogs in stages B2 and C of MMVD, however the difference observed was not significant, perhaps because of the low number of dogs in the study. 19 Although some definitions of cachexia require the loss of 5% of a dog's BW for it to be considered cachexic, 21 our findings suggest that smaller decreases in BW are common in dogs developing CHF and not seen in dogs that do not develop CHF. It seems likely that similar mechanisms underlie the observed weight loss in these dogs. The most important findings of our study illustrate the temporal evolution of the observed changes in the studied variables in MMVD stage B2 dogs before the onset of confirmed CHF. For all of the 6 variables evaluated, the rate of change of the variable was maximal in the period immediately before the onset of CHF (Figures 5 and 6).
Although rate of change increased in VHS the closer in time the dogs were to CHF, the change was gradual over time compared to changes in the other variables and seemed to be evident at least 1 year before the onset of CHF. By contrast, the changes in BW, HR, RT, RR, and RRR occurred late in the process (ie, within the last few months before confirmed CHF). The magnitude and timing of the increase in respiratory rate variables reinforces the value of these variables as indicators of impending or incipient CHF. In the present study, the maximum duration of time over which variables were described was 18 months. Many dogs, particularly those in the no-CHF group, spent longer in the study, but the reason for selecting 18 months was that larger numbers of dogs could contribute observations over the entirety of this period leading to more precise estimates of variables.

| Limitations
Our study was limited to characterizing the changes in only 6 continuous variables. The reason for this limitation was that these were the only continuous variables measured at least once in every 8-month period throughout the study. Accordingly, the study cannot characterize changes in other potential variables of interest, such as heart murmur intensity, echocardiographic findings, blood pressure, and biomarker concentrations.
The frequency of measured variables was limited to every 4 months for the clinical variables and every 8 months for VHS measurement.
The changes occurring between these points of measurements therefore are unknown. Furthermore, not all intervals were standardized, because some visits at which variables were assessed were emergency visits rather than visits scheduled in the protocol.
We described the mean value and mean rate of change in the 6 variables we have evaluated. Within a population of dogs, variability will exist in the absolute values and their rates of change. We have not attempted to perform subanalyses or to identify populations that may have different rates of change such as those in which ruptured chordae tendinae may cause sudden development of CHF in the absence of a marked increase in heart size.
As shown in Figure 1, our method of imputing values of variables assumed a linear rate of change of those variables between true observations, but our analyses indicate that the rate of change was not linear.
This situation may have resulted in an underestimate of the rate of change of variables at the end of periods between observations and an overestimate of the rate of change at the beginning of periods between observations. The overall effect on the absolute values of variables calculated may have been to conclude that values increased sooner than would have truly been observed had they been measured more frequently. This may mean that changes in the variables that were only seen to occur late in the disease process in reality occurred even later, and more rapidly, than are illustrated in our figures.
Finally, on any given day, the mean values of the variables illustrated are based on an average of both observed and imputed values.
This makes it impossible to calculate appropriate confidence intervals for the mean values illustrated in our graphs because calculation of such would require independent real measurements, which was not the case in the present study, causing an underestimation of the interval sizes.

OFF-LABEL ANTIMICROBIAL DECLARATION
Authors declare no off-label use of antimicrobials.