Natural history of hypertrophic cardiomyopathy in cats from rehoming centers: The CatScan II study

Abstract Background The natural history of hypertrophic cardiomyopathy (HCM) in cats has been mainly studied in cats referred for suspected heart disease, which can skew the results towards cats with clinical signs. Few data are available on factors associated with development of HCM in cats. Hypotheses (1) Clinical variables can predict which cats will develop HCM; (2) HCM in cats not referred for suspected heart disease is associated with a low rate of cardiovascular events. Animals One hundred seven cats from rehoming centers without a history of clinical signs of cardiac or systemic disease at the time of adoption. Methods Prospective longitudinal study. After rehoming, shelter cats were reexamined for serial echocardiograms. Cox regression analysis was used to identify predictors of development of HCM in cats that were normal at baseline. Adverse cardiovascular events including heart failure, thromboembolism, or sudden death were recorded. Results Cats were monitored for a median of 5.6 [1.2‐9.2] years. At baseline, 68/107 cats were normal, 18/107 were equivocal and 21/107 had HCM. Nineteen cats developed HCM during the study period. The factors at baseline associated with increased hazard of developing HCM were lower left atrial fractional shortening, higher left ventricular fractional shortening, and higher body weight. Cardiovascular events were observed in 21% of cats with HCM. Conclusions and Clinical Importance Cardiovascular events were common in cats with HCM from a rehoming center study sample. Lower left atrial systolic function appears to precede overt HCM.


| INTRODUCTION
Most reports of the natural history of hypertrophic cardiomyopathy (HCM) in humans 1-3 and cats [4][5][6][7][8] have been based on studies performed at referral centers. In people, referral cohorts can be biased towards patients suspected of heart disease, skewing the sampling to higher risk patients and more severe forms of HCM, which can lead to overestimation of disease malignancy. [9][10][11] Human studies of HCM in nonreferral (general) centers, which are more likely to reflect the true natural history of this disease, consistently have lower rates of morbidity and death and a more benign prognosis than people diagnosed with HCM at at referral centers. [9][10][11][12][13] While some human HCM patients experience serious complications, the majority will have an uneventful clinical course and normal longevity. 12,13 The REVEAL study 8 evaluated the natural history of subclinical HCM in a large study sample of cats from referral centers worldwide. This is the largest veterinary study to date evaluating the prognosis of cats with subclinical HCM, describing an overall incidence of cardiovascular events and death of approximately 30%.
In people, changes in left ventricular (LV) function and geometry precede LV hypertrophy (LVH), suggesting that LVH is not the first manifestation of HCM. 14 Certain clinical characteristics, such LV ejection fraction, elongated mitral leaflets and ECG abnormalities, are suggested to predict development of LVH (overt HCM). 15,16 The natural history and prognosis of HCM in cats not referred for suspected heart disease have not been reported. Furthermore, there are scarce data assessing clinical or echocardiographic factors that might predict the development of HCM in cats. In the present longitudinal study, we serially examined a study sample of cats from rehoming centers in the UK. Our hypotheses were: (1) Clinical variables can predict which cats will develop an HCM phenotype; (2) HCM in cats not referred for a suspected heart disease is associated with a low rate of cardiovascular events. The primary aims of this study were: (1) to determine whether any clinical variables are associated with increased risk of developing HCM by studying cats that were normal at baseline and developed HCM during the study period; (2) to assess the incidence of cardiovascular events including congestive heart failure (CHF), arterial thromboembolism (ATE) and sudden cardiac death (SD) in cats with HCM from rehoming centers.

| MATERIALS AND METHODS
This was a prospective, longitudinal study of the natural history of HCM in cats from rehoming centers (CatScan II study). This study was the longitudinal arm of a cross-sectional study (CatScan I study). 17 The cross-sectional study collated clinical and echocardiographic data were included in the present study.

| Cats lost to follow-up
There were 459/570 cats initially enrolled in the longitudinal study that were subsequently lost to follow-up. A summary of the reasons for losing cats to follow-up is described in Table 1.

| Cats enrolled in the longitudinal study
One hundred eleven cats were reexamined after rehoming. The scan and clinical data were lost from 3/111 cats, and scan quality was inadequate in 1 cat, thus the study sample comprised 107 cats. At each visit, physical examination, blood pressure measurement, and echocardiography were performed. Additionally, N-terminal pro B-type natriuretic peptide was measured in cats with echocardiographic abnormalities, and an ECG was performed if an arrhythmia was detected. A summary of the cats enrolled in the longitudinal study is described in Figure 1

| Cats with HCM at initial evaluation
Baseline and follow-up characteristics of cats with HCM at baseline are summarized in Table 3 Figure 3). In multivariable analysis, initial LAFS% ≤25%, higher LVFS% and higher BW at baseline were independently associated with increased hazard of developing HCM (Figure 4). Graphical and statistical analysis showed no violation of the proportional hazard assumptions, there was no multicollinearity, and the multivariable model was a good fit to the data.
When HCM was defined as LVWT ≥5.5 mm, none of the assessed variables at baseline was associated with a greater risk of HCM on Cox proportional hazard analyses, LAFS% was lower at baseline in cats that later developed HCM but not significantly (P = .05).

| HCM with SAM (HCM + SAM)
The proportion of cats with HCM that had SAM was similar at base-

| Cats that remained free of HCM during the study (normal/equivocal-normal/equivocal)
With a median follow-up time of 5.6 years, characteristics of the 64 cats with a normal/equivocal cardiac phenotype throughout are summarized in Table 5. Weight, body condition score, systolic blood pressure, LVWT, LVIDd and LA size increased over time.

| Change in LVWT over time in cats with LVWT <6 mm at baseline
In 67/83 (81%) cats the LVWT increased over the study period ( Figure 5). There was no association between %ΔLVWT/year and age at baseline, BW change or systolic blood pressure change over time.

| Echocardiographic measurement variability
Intraclass correlation coefficient (ICC) was 0.90 and coefficient of variation (%CV) was 5% for both LVWT and LA%FS indicating excellent intraobserver agreement and low measurement variability.

| DISCUSSION
In this study, we followed a study sample of cats that were originally In this study, 40/107 (37.4%) cats had HCM at the final followup. The prevalence of subclinical HCM in cats not referred for suspected heart disease is 14.6% to 34%, increasing with age. 26,27 The higher prevalence could be related to the older age of our study sample. 17,28 The results confirm that the prevalence of HCM in cats appears to be high in the geographic region, environment, and genetic pool where this study took place.
In people, changes in LV function and geometry have been identified before the development of overt HCM (prephenotypic HCM), including increased LV ejection fraction, 14 36,37 Prephenotypic changes in cats include diastolic dysfunction 38 and elongated anterior mitral valve leaflets. 39 We assessed baseline factors in cats that subsequently developed HCM to look for predictors of HCM. Multivariable analysis showed that lower LAFS%, higher LVFS%, and higher BW independently predicted an increased hazard of developing HCM.
Cats with a baseline LAFS% >25% were much less likely to develop HCM, and reduced LA function has been suggested as an early phenotypic change in people. 30 It has been speculated in humans that LA dysfunction might reflect early myocardial fibrosis, 30,40 which could be a marker for early HCM. 40 A recent study showed that LAFS% in cats with subclinical HCM was correlated with LV extracellular volume, a measure of interstitial fibrosis, but not with LVWT. 41 Abnormal collagen metabolism has been also suggested in Ragdoll cats carrying MYBPC3 R820W mutation without LVH, 42 and interstitial fibrosis has been described on histopathology in cats diagnosed antemortem with mild subclinical HCM. 43 Thus, our findings of LA dysfunction might, as in humans, reflect abnormal myocardial properties that precede overt HCM. LA dysfunction is recognized to be an important prognostic factor in overt HCM, 44,45 and we suggest that LA dysfunction might be an important abnormality even before manifestation of the HCM phenotype.
Cats with higher LVFS% were also more likely to develop HCM.
In people, higher LV ejection fraction has been documented before the development of HCM. 14,16,29,46 Hyperdynamic LV function detected by speckle-tracking echocardiography and cardiac magnetic resonance imaging is present in prephenotypic HCM. 47,48 It has been proposed that a hyperdynamic state in HCM might reflect intrinsic myocardial properties associated with sarcomere mutations 29 that increase sarcomere power output. 49 Targeting sarcomere hypercontractility with a cardiac myosin inhibitor [49][50][51] halts the development and progression of HCM, 49 suggesting enhanced sarcomere function might be an important mechanism in the development of HCM (LVH). The higher LV%FS identified in our study might parallel the prephenotypic hyperdynamic systolic function described in humans.
Alternatively, the thicker LV walls and smaller LVIDd at baseline ( There is a progressive increase in LVWT in human patients with HCM. 63 A large number of cats were lost to follow-up, as might be expected in a large longitudinal study following cats after rehoming.
At baseline, longitudinal and cross-sectional study samples had similar demographic characteristics, and the prevalence of HCM was similar.
However, there was a higher prevalence of cats with heart murmurs and cats with equivocal cardiomyopathy in the longitudinal study sample. This might reflect a selection bias, where owners of cats known to have a heart murmur and, or with a suspected cardiomyopathy, or both, were more likely to enroll their cats into the study or to participate in the follow-up examinations.
We defined HCM as LVWT ≥6 mm, but there is no single cut-off value for LVWT that will reliably differentiate cats with and without HCM. 18,55,64 We used a cut-off of 6 mm to err on the side of high specificity for identifying LVH. Our findings might have been different if we had used a lower cut-off, or normalized LVWT to BW, 55,59,60 but there is currently no consensus on cut-off values for defining HCM using allometric scaling. We repeated the Cox regression analysis using a 5.5 mm cut-off to diagnose HCM, and none of the baseline variables were associated with a higher hazard of HCM, although LAFS% was also lower at baseline in cats that later developed HCM albeit not significantly (P = .05).
Our study had some limitations. We evaluated a sample of cats originating from 2 rehoming centers, thereby avoiding potential referral biases. Nevertheless, other types of bias might have been present, such as selection or survivor bias. Owners of cats known to have heart disease or a heart murmur might have been more likely to enroll their cats in the longitudinal arm of the study (selection bias).
Additionally, the large number of cats lost to follow-up could have skewed the data, impacting the results (survivor bias). 65 Cats lost to follow-up (Table 1) were not included in the outcome analysis, and therefore our cardiovascular event rate and death rate might be underestimated. In