Prevalence and risk factors for atrial fibrillation in dogs with myxomatous mitral valve disease

Abstract Background Atrial fibrillation (AF) is a common supraventricular arrhythmia more frequently observed in large breed dogs. Objectives Estimate the prevalence of AF in dogs with myxomatous mitral valve disease (MMVD) and identify risk factors for developing AF. Animals A total of 2194 client‐owned dogs with MMVD, including 1280, 588, 290, and 36 dogs in ACVIM stages B1, B2, C, and D, respectively. Methods Retrospective, cross‐sectional study. The medical databases of 3 veterinary teaching hospitals were reviewed. Inclusion criteria were a diagnosis of MMVD after complete cardiovascular evaluation and cardiac rhythm assessment using routine 2‐minute ECG or good quality ECG tracing during echocardiographic examination. Results Atrial fibrillation was diagnosed in 59 dogs with a prevalence of 2.7%. Univariate analysis showed that mixed breed, male sex, advanced ACVIM stage, left atrial and ventricular enlargement, fractional shortening (FS), and presence of pulmonary hypertension were significantly associated with development of AF. According to 2 multivariable models, the left atrium (LA)‐to‐aorta ratio (odds ratio [OR] = 14.011, 7.463‐26.304), early trans‐mitral velocity (OR = 2.204, 1.192‐4.076), body weight (OR = 1.094, 1.058‐1.130), and FS (OR = 0.899, 0.865‐0.934) and LA (OR = 5.28, 3.377‐8.092), advanced ACVIM stage (OR = 4.922, 1.481‐16.353), and FS (OR = 0.919, 0.881‐0.959) were significant predictors of AF for models 1 and 2, respectively. Conclusions and Clinical Importance Atrial fibrillation is an uncommon complication of MMVD and is significantly associated with the more advanced stage of the disease, increased LA dimension and body weight, and decreased FS.

K E Y W O R D S canine, cardiac arrhythmia, echocardiography, electrocardiography, epidemiology

| INTRODUCTION
Atrial fibrillation (AF) is the most common supraventricular arrhythmia in both humans and dogs. [1][2][3][4] Although some dogs can develop AF in the absence of recognizable cardiac disease (ie, primary or lone AF), AF secondary to cardiac diseases associated with left atrial enlargement (LAE) is more commonly observed. 3,4 In humans, the prevalence and risk factors associated with development of AF are well known. 1,2,[5][6][7] In particular, heart failure (HF) and AF are linked by similar risk factors in people and share a common pathophysiology. 1,2,[5][6][7] Thus, it is widely accepted that HF and AF can cause and exacerbate each other through mechanisms such as structural cardiac remodeling, activation of neuro-hormonal mechanisms, and rate-related impairment of left ventricular function. 1,2,[5][6][7] In dogs, AF usually is observed in animals with dilated cardiomyopathy, myxomatous mitral valve disease (MMVD), or congenital cardiac disease associated with LAE. 3,4 Among these diseases, MMVD is the most frequently diagnosed, accounting for approximately 70% of acquired cardiac disease in dogs. [8][9][10] Progressive LAE is a common sequela during MMVD progression, and guidelines provided by the American College of Veterinary Internal Medicine (ACVIM) established a specific classification scheme of the different stages of HF associated with MMVD. 9,10 Although some epidemiologic data and associations for AF development have been described in dogs, 3,11 no study has thoroughly examined the prevalence rate and risk factors for this arrhythmia in dogs with MMVD. In addition, results of a recent study showed that the presence of AF is associated with a worse prognosis in medium to large-sized dogs with MMVD. 12 Thus, knowledge of the risk factors for developing AF is useful in the clinical evaluation of dogs with MMVD.
We estimated the prevalence of AF in a large population of dogs with MMVD and identified the risk factors for developing AF in these animals. We hypothesized that AF is associated with worsening of MMVD and that some clinical variables and echocardiographic parameters indicative of cardiac remodeling can be useful predictors of developing AF in dogs with MMVD.

| Animals
For this retrospective study, data were collected from the internal database of the veterinary teaching hospitals (VTH) of the Universities of Padova, Bologna, and Curitiba. Dogs were recruited to the study from those undergoing cardiac diagnostic evaluation and final diagnosis of MMVD between January 2012 and December 2018. Diagnosis of MMVD was based on clinical and echocardiographic findings including a left apical systolic murmur, thickened or prolapsing mitral valve leaflets or both on 2-dimensional (2D) echocardiography associated with mitral valve regurgitation on color flow Doppler interrogation. 9,10 Disease severity was classified according to the ACVIM guidelines. 9,10 The presence or absence of AF was based on at least 1 of the following methods: routine surface ECG recording with dedicated machines of at least 2 minutes duration or good quality ECG recordings during the echocardiographic examination. In particular, the echocardiographic diagnosis of AF was based on the combined presence of the following findings: irregularly irregular cardiac rhythm with narrow QRS complexes, isoelectric trace without recognizable P waves, and absence of A wave on mitral inflow on Doppler interrogation.
For dogs with more visits available during the period of observation, only data obtained during the most recent visit and those obtained during the first visit documenting AF were analyzed for dogs without AF and those developing AF, respectively. Dogs with sinus rhythm and dogs with cardiac arrhythmias other than AF were merged together and only 2 groups were considered for analysis: dogs with AF and dogs without AF. The absence of a final rhythm diagnosis was considered an exclusion criterion. Dogs with equivocal cardiac diagnosis or other concomitant congenital or acquired cardiac disease also were excluded from this study.

| Echocardiographic examination
At each VTH, an experienced operator performed the echocardiographic examination. The left ventricular diameter at diastole (LVDD) and systole (LVSD) were measured from M-mode short-axis echocardiographic images at the level of the chordae tendinae and fractional shortening (FS) then was calculated according to the formula (LVDD-LVSD)/LVDD. Normalized left ventricular (LV) dimensions (ie, LVDSn and LVDDn) were calculated according to reported allometric scaling. 13 Left atrial diameter (LA) and aortic root diameter (Ao) were measured at early diastole from 2D echocardiographic short axis images obtained at the level of the heart base and the LA/Ao ratio then was calculated. 14,15 Pulsed-wave Doppler interrogation of trans-mitral blood flow was obtained from the left apical 4-chamber view and the peak of the early diastolic wave (E max) was measured. Multiple views were used to evaluate trans-tricuspid blood flow and the maximal value of any tricuspid regurgitation (TR) velocity was measured using continuous-wave Doppler. In the absence of right ventricular outflow obstruction, the presence of pulmonary hypertension (PH) was considered when the TR maximal velocity was ≥3 m/s, corresponding to a Doppler-derived estimated systolic pressure gradient of 36 mm Hg. 16    Atrial fibrosis is 1 of the most important abnormalities contributing to the development of AF in humans. 34 Potential mediators known to promote fibrosis in the atrium include pressure and volume overload, aging, atrial stretch, inflammation, and oxidative stress. 34 As atrial myopathy progresses, it can lead to atrial dysfunction. 34 Few studies have investigated the microscopic changes of atrial myocytes in dogs with MMVD or AF or both. [35][36][37] In dogs with MMVD, the observed changes included myocardial fatty replacement, immune cell infiltration, and interstitial fibrosis, 35,36 whereas atrial histological changes in dogs with AF were investigated mainly in animals with experimentally induced arrhythmias. 37 Evaluation of left atrial function recently has been investigated in dogs with MMVD using speckle-tracking echocardiography, and the results suggested the presence of left atrial dysfunction in more advanced stages of the disease. [38][39][40][41][42] Furthermore, in addition to absolute cardiac diameters, evaluation of LA function with speckle-tracking echocardiography can be useful to predict development of AF in dogs with MMVD. 43 Decompensated HF is a risk factor for AF in dogs with an OR (4.9) similar to that observed in humans (4.5-5.9) in the classical Framingham heart study. 44 Heart failure and AF are linked by similar risk factors in humans and share a common pathophysiology. Thus, it is widely accepted that HF and AF can cause and exacerbate each other through mechanisms such as structural cardiac remodeling, activation of neuro-hormonal mechanisms, and rate-related impairment of LV function. 44 Our results and those of another recent study suggest that the same relationship between AF and HF also could be present in dogs. 29 However, the precise cause-effect relationship between HF and AF remain elusive in our study as well as in studies of humans, 45 because the temporal relationship of each condition could not be determined. In addition, the prevalence of AF in HF is known to increase with increased severity of pump failure in humans. 46 Evaluation of the systolic function is challenging in dogs with MMVD and was not specifically addressed in our study. However, the observed risk factor of decreased FS suggests that a similar pathophysiological mechanism also might be present in dogs, but results of FS measurements in subjects with MMVD can be misleading. 28 Our study had some limitations because of its retrospective