• Open Access

Prognostic Indicators in Cats with Hypertrophic Cardiomyopathy

Authors


  • Meeting, if any, at which the paper was presented: ECVIM-CA, Toulouse, September 2010 (diastolic function in part of this cohort); 2012 ACVIM Forum, New Orleans, LA (overall survival in this cohort).

Corresponding author: V. Luis Fuentes, Clinical Sciences and Services, Royal Veterinary College, Hawkshead Lane, North Mymms, Hatfield, Hertfordshire, AL9 7TA, UK; e-mail: vluisfuentes@rvc.ac.uk.

Abstract

Background

Left atrial (LA) enlargement, congestive heart failure (CHF), and aortic thromboembolism (ATE) are associated with decreased survival in cats with hypertrophic cardiomyopathy (HCM), but the prognostic value of echocardiographic variables has not been well characterized.

Hypothesis/Objectives

We hypothesized that LA echocardiographic variables and assessment of left ventricular (LV) diastolic and systolic function would have prognostic value in cats with HCM.

Animals

Two hundred eighty-two cats diagnosed with HCM.

Methods

Clinical and echocardiographic records of affected cats seen at the Royal Veterinary College from 2004 to 2009 were retrospectively analyzed. Only cats with echocardiographic confirmation of LV diastolic wall thickness ≥6 mm were included. Outcomes were obtained from clinical records or referring veterinarians and owners.

Results

Deaths occurred in 164 cats, of which 107 were believed to have been cardiac deaths. Univariable predictors of an increased risk of cardiac death included older age, absence of a murmur, presence of a gallop sound or arrhythmia, presentation with either CHF or ATE, extreme LV hypertrophy (≥9.0 mm), LV fractional shortening (FS%) ≤30%, regional wall hypokinesis, increased left atrial size, decreased left atrial function, spontaneous echo-contrast/thrombus or both, absence of left ventricular outflow tract obstruction, and a restrictive diastolic filling pattern. Cox's proportional hazard analysis identified LA dysfunction, low LV systolic function, and extreme LV hypertrophy as independent predictors of decreased cardiac survival time.

Conclusions and Clinical Importance

Echocardiographic measurement of LA function, extreme LV hypertrophy, and LV systolic function provides important prognostic information in cats with HCM.

Abbreviations
Adur:ARdur

ratio of the duration of the late filling transmitral wave (Adur) to the duration of the pulmonary venous atrial reversal wave (ARdur)

ATE

arterial thromboembolism

CHF

congestive heart failure

CVs

coefficients of variation

E:A

ratio of mitral inflow peak early filling (E) to late filling (A) velocities

E:E′

ratio of mitral inflow peak early filling to early diastolic tissue Doppler mitral annular velocity

E′:A′

ratio of early diastolic tissue Doppler mitral annular velocity to the late diastolic mitral annular velocity

FS%

left ventricular fractional shortening

HCM

hypertrophic cardiomyopathy

IVRT

isovolumic relaxation time

LA:Ao

ratio of diastolic left atrial diameter to aortic root diameter measured from a short axis image

LAA

left auricular appendage velocity

LAD

the diameter of the left atrium measured parallel with the mitral annulus in the last frame before mitral valve opening

LA-EF%

total left atrial ejection fraction

LA-FS%

total left atrial fractional shortening

LA

left atrial

LV

left ventricle

LVWd

end-diastolic left ventricular septal or free wall thickness

MR

mitral regurgitation

QMHA

Queen Mother Hospital for Animals

S:D

ratio of the pulmonary venous peak systolic (S) to peak diastolic (D) velocity

SAM

systolic anterior motion of the mitral valve

SEC

spontaneous echo-contrast

Hypertrophic cardiomyopathy (HCM) is defined as a hypertrophied, nondilated left ventricle (LV) in the absence of other systemic or cardiac disease capable of producing a similar degree of hypertrophy.[1] It is the most common familial heart disease in humans[2] and the most commonly diagnosed myocardial disease in cats.[3]

Hypertrophic cardiomyopathy is a heterogeneous disease, both in terms of presentation and outcome. Some affected cats are presented with signs of congestive heart failure (CHF) or thromboembolism, and some cats die suddenly. Other affected cats can have long survival times and die of noncardiac causes.[4-8]

Prognostic factors have been reported in a number of retrospective studies of HCM in cats.[5, 7-11] Factors suggested to be associated with a worse outcome have included the presence of clinical signs at diagnosis,[5, 8, 10, 11] left atrial (LA) enlargement,[7-10, 12] and increased age at diagnosis.[8, 10, 12] Certain breeds may have shorter survival times, including the Ragdoll[10] and Maine Coon.[11] An early study found that a heart rate ≥200 bpm at initial diagnosis was negatively associated with survival,[5] but this has not been demonstrated in subsequent studies.[8, 10]

Systolic anterior motion (SAM) is a negative prognostic indicator in people,[13-15] but the prognostic effect in cats remains unclear,[7, 8, 10] because there are no prospective longitudinal studies of SAM in cats, and the long-term relationship between outcome and the presence or absence of dynamic LV outflow tract obstruction has not been established.

Maximal LV wall thickness was found to be greater in cats that later died because of their cardiac disease compared with survivors in another study.[7] Prognosis has not been found to be linked to sex[5, 10] or body weight.[5] In asymptomatic cats, the administration of atenolol has not been found to influence survival to 5 years post diagnosis.[12]

In HCM of humans, and other cardiac conditions, echocardiographic measures are used routinely as prognostic indicators. These include measures of LA function,[16] LV systolic function,[17-19] and diastolic function.[20-28] Although LA enlargement is considered a very important prognostic indicator in cats with HCM,[7-10] little has been done to investigate other echocardiographic markers of prognosis.

We hypothesized that although LA echocardiographic variables would provide prognostic information in cats with HCM, assessment of left ventricular diastolic and systolic function would provide additional prognostic value. The aim of this study was, therefore, to investigate the prognostic value of echocardiographic variables in cats with HCM.

Materials and Methods

Cases of HCM were identified by searching the electronic patient records of cats seen at the Royal Veterinary College, Queen Mother Hospital for Animals (QMHA) between June 2004 and August 2009. Cats were included if a diagnosis of HCM had been made by a board-certified cardiologist or a cardiology resident supervised by a board-certified cardiologist based on 2D or M-mode echocardiography or both1 and if follow-up information was available. Cats were excluded from the study if they had a known diagnosis of hyperthyroidism or hypertension, defined as either systolic blood pressure ≥180 mmHg; systolic blood pressure ≥160 mmHg with retinal changes suggestive of hypertension; medically controlled hypertension; or where renal disease was present and blood pressure had not been determined. The first visit date was defined as the date of the first echocardiographic examination. Cats with a prior diagnosis of HCM were not excluded, but data were taken from the first visit with an echocardiographic examination within the study period rather than the original diagnosis date.

The medical records for each cat were reviewed for date of birth, sex, breed, and date of first visit to the QMHA, with age calculated at that visit. Physical examination findings were recorded, as well as systolic arterial pressure2 and serum total thyroxine concentrations, where available. Clinical status at presentation was recorded (Table 1) and grouped into “asymptomatic” (cats presenting asymptomatically, with concurrent disease, with concurrent respiratory disease, or with noncardiac chylothorax), “syncope,” “CHF” (cats presenting with CHF or chylothorax of a cardiac cause), and “ATE” (cats presenting with ATE with or without CHF). Cats with exertional open-mouth breathing were not analyzed because of small numbers. Cats with HCM with “non-cardiac chylothorax” were defined as having persistently normal LA size and function despite progressive or persistent chylothorax. Presence of arrhythmias was determined from available ECG recordings (6-lead, Holter, or both) and by reviewing the concurrent ECG recorded during echocardiography. Arrhythmias were classified as “bradyarrhythmias,” “supraventricular arrhythmias,” “ventricular arrhythmias,” or both “supraventricular and ventricular arrhythmias.”

Table 1. Definitions of categorization of clinical signs at presentation.
Clinical Status at PresentationDefinition
AsymptomaticReferred for investigation of a murmur, gallop, or arrhythmia, but without clinical signs
Concurrent diseaseReferred for investigation of noncardiac disease, but cardiac disease suspected during hospitalization. No clinical signs referable to cardiac disease
Concurrent respiratory diseaseReferred for investigation of respiratory disease, but cardiac disease suspected during hospitalization. Primary respiratory disease subsequently confirmed on further investigations
SyncopeEpisodes described as syncopal or seizure-like activity. No noncardiac cause for syncope identified
Exertional open-mouth breathingOpen-mouth breathing only during periods of excitement, stress, or exercise
Congestive heart failureAny of the following: present or previous radiographic evidence of pulmonary edema; present or previous radiographic/ultrasonographic evidence of pleural or pericardial effusion; or severe tachypnea that showed a clear response to furosemide administration
ChylothoraxMilky white-colored pleural effusion, with triglyceride concentrations exceeding serum concentrations and the absence of infectious organisms identified on fluid analysis
Noncardiac chylothoraxChylothorax with normal atrial chamber dimensions and normal left atrial function
Arterial thromboembolism

Limb: Sudden onset of lower motor neuron deficits in one or more limbs diagnosed as arterial thromboembolism by attending clinician

Brain: Magnetic resonance imaging findings of a well-demarcated lesion, hyperintense on T2-weighted image

Mesenteric: History of acute onset abdominal pain, high creatine kinase levels, and echocardiographic evidence of spontaneous echo-contrast in the left atrium

All echocardiographic examinations were reviewed and remeasured by 1 board-certified cardiologist (VLF) or a resident (KB). Measurements were obtained only from images with optimal image quality with average values reported from at least 3 cardiac cycles except for maximal 2D LV end-diastolic wall thickness, for which the maximum measurement of at least 3 cardiac cycles was reported.

Left ventricular hypertrophy was defined as LV septal or free wall end-diastolic thickness ≥6 mm,[7] whether measured from 2D or M-mode images.[29] Otherwise, M-mode images were used only to obtain LV fractional shortening (FS%) values, and 2D images were used for measuring maximal LV end-diastolic wall thickness from either the septum or free wall. The presence of any left ventricular regional wall hypokinesis was noted, based subjectively on 2D images and defined as a region of LV wall that was thinner than the rest of the LV, with minimal excursion or moving asynchronously with the rest of the ventricle.

Left atrial size was assessed by 2 separate methods, LA:Ao and LAD. LA:Ao was obtained from a 2D right parasternal short-axis view to calculate the ratio of LA diameter to aortic root diameter measured at the onset of the QRS.[28] LAD was measured from a right parasternal long-axis view in which the diameter of the left atrium in the last frame before mitral valve opening was measured parallel with the mitral annulus, bisecting the left atrium.[30] The presence of either spontaneous echo-contrast (SEC) or a thrombus in the left atrium was recorded. Maximal LA appendage velocities (LAA) were recorded where available.[31] LA systolic function was assessed by both 2D and M-mode methods. In the 2D method (LA-EF%), “total” LA ejection fraction was calculated from maximum and minimum LA volumes measured by a single plane modified Simpson's method from right parasternal long-axis images.[16, 32, 33] Anatomic M-mode was used to calculate LA-FS% based on minimum and maximum LA diameter in a short-axis image at the level of the aortic valve and left atrium.[34]

If systolic anterior motion of the mitral valve (SAM) or mitral regurgitation (MR) was present on a 2D image or color flow Doppler, respectively, this was recorded. SAM was identified on 2D echo by identifying any abnormal motion of the anterior mitral leaflet from cine-loops played at slow speed, and confirmed by demonstrating variance in the LV outflow tract on color flow Doppler.

Transmitral flow patterns, pulmonary venous flow patterns, and septal tissue Doppler imaging variables all were recorded where available. Calculated variables included E:A, the ratio of mitral inflow peak early filling (E) to late filling (A) velocities in cats with E wave velocities ≤0.2 m/s at E and A wave separation[35]; E:IVRT, the ratio of the mitral E wave (in m/s) to the isovolumic relaxation time (in ms); E′:A′, the ratio of the early diastolic tissue Doppler septal mitral annular velocity to the late diastolic septal mitral annular velocity; E:E′, the ratio of the mitral E wave to the early diastolic tissue Doppler septal annular velocity (E′); S:D, the ratio of the pulmonary venous peak systolic (S) to peak diastolic (D) velocity; and Adur : ARdur, the ratio of duration of the late filling transmitral wave (Adur) to duration of the pulmonary venous atrial reversal wave (ARdur).[31, 36] For E:IVRT, E′:A′ and E:E′ values only were calculated if the E or E′ components were adequately separated from the A or A′ components, respectively.

The diastolic filling pattern was classified according to transmitral flow patterns. An E:A ratio of <1 was classified as a delayed relaxation pattern, an E:A ratio of 1–2 was either a normal or a pseudonormal pattern and an E:A ratio of >2 was a restrictive filling pattern.[36] Normal and pseudonormal filling patterns were differentiated by a sequence of factors. If LA enlargement was present (LA:Ao >1.5[37] or LAD >16 mm[36]) or the cat had presented with CHF, the pattern was considered pseudonormal. For cats with normal LA size and without CHF, differentiation of normal from pseudonormal filling was based on assessment of E:E′, Adur : ARdur, and E′:A′, depending on availability. Diastolic filling pattern in these cats was considered pseudonormal in the presence of any of the following: E:E′ >10, Adur : ARdur <1, and E′:A′ <1.[31, 35, 36]

Information was gathered on whether cats survived to discharge and which cardiac medications, if any, had been prescribed at discharge. Mortality was determined by reviewing QMHA medical records and contacting referring veterinarians. Date of death, whether the cat died naturally or because of euthanasia, and whether death was related to cardiac disease (sudden death, CHF, ATE) were recorded. Cats that were euthanized because of a new episode of CHF, becoming refractory to CHF medication, a new episode of ATE, or failing to recover limb function all were considered cardiac deaths. Where insufficient evidence was available, owners were contacted and asked to complete a questionnaire. Sudden death was defined as being found dead without obvious cause at home, or as a witnessed event where the cat had been apparently well in the preceding 24 hours and death was assumed to be cardiac related. For cats still alive, survival was calculated up until either the date of interview with the owner or the last date of veterinary examination.

Statistical analysis was performed by commercially available software3 and values are reported as mean ± SD (normally distributed data) and median (interquartile range [IQR]) for non-normally distributed data. Repeatability for the 2 observers was assessed by coefficients of variation (CVs) for continuous variables and kappa statistics for categorical variables. Kaplan-Meier survival curves were generated, survival times reported as median (range), and differences between groups analyzed by the Logrank (Mantel-Cox) test. Noncardiac deaths (with survival included up to the point of death) and cats still alive at the time of analysis were right-censored. A composite end-point of all cardiac-related deaths (CHF, ATE, sudden death) was analyzed. Hazard ratios and 95% confidence intervals (CI) for the univariable factors were calculated by Cox proportional hazards. To evaluate the effects of multiple variables on survival, multivariable analysis was performed considering variables significant at P < .10 by a manual backward selection stepwise Cox regression model. Hazard ratios and 95% CI were calculated. Final model variable pair-wise interactions were evaluated, the proportional hazard assumption was assessed by log cumulative hazard plots, and Schoenfeld residuals and overall model fit via graphical assessment of Cox-Snell residuals.[38]

Continuous variables initially were analyzed as quartiles (or halves for variables with n < 80) by the Kaplan-Meier method to assess whether a single cut-off value or a continuous variable should be taken forward to the multivariable model. Variables where there was an ordinal increase or decrease in the hazard with each quartile were taken forward as continuous. Variables where 1 or more quartiles had significantly different survival curves from the other quartiles were taken forward as categorical data using cut-offs based on the values of the differing quartile(s). A value of P < .05 was considered statistically significant.

Results

Between June 2004 and August 2009, 282 cats were diagnosed with HCM and included in the study. At initial presentation, median (IQR) age was 6.2 (2.8–9.7) years, with a range of 0.4–17.4 years, and mean weight 4.71 ± 1.11 kg. The majority of cats were male (75.5%), neutered (95.0%), and nonpedigree (80.5%). In total, 17 breeds were represented, of which the most common pedigree breeds were British Shorthair (n = 18), Persian (n = 15), Ragdoll (n = 6), and Sphynx (n = 4).

At initial presentation, CHF was present in 93 cats (33.0%), including 88 (31.2%) in the CHF group (CHF or cardiac chylothorax) and 5 (1.8%) with ATE and CHF (Table 2). Of these, 49 had pericardial effusion, 53 had pleural effusion and 65 had pulmonary edema. Of the 17 cats presenting with a thromboembolic event, 7 were affected in 1 limb, 7 affected in 2 limbs, 1 was affected in 3 limbs, 1 embolus was to the brain, and 1 was mesenteric.

Table 2. Clinical status of cats with HCM at first presentation.
Clinical Statusn (%)
  1. HCM, hypertrophic cardiomyopathy; CHF, congestive heart failure; ATE, arterial thromboembolism.

Asymptomatic88 (31.2)158 (56.0)
Concurrent disease54 (19.1)
Concurrent respiratory disease13 (4.6)
Noncardiac chylothorax3 (1.1)
Syncope16 (5.7)16 (5.7)
Exertional open-mouth breathing3 (1.1)3 (1.1)
CHF81 (28.7)88 (31.2)
Chylothorax7 (2.5)
ATE12 (4.3)17 (6.0)
ATE + CHF5 (1.8)

Two hundred and six (73.0%) cats presented with cardiac murmurs, whereas only 67 (23.8%) presented with gallop sounds (Table 3). Blood pressure was measured in 178 (63.1%) cats. Serum total thyroxine concentration was assessed in 87 (30.9%) cats overall, and in 55 (58.5%) cats aged ≥9 years.

Table 3. Physical examination, auscultation, and ECG findings at first presentation in all cats with HCM.
 Present (%)/Median (IQR)Assessed in
  1. HCM, hypertrophic cardiomyopathy.

Heart rate (/min)180 (160–200)272
Respiratory rate (/min)48 (36–60)262
Temperature (°C)38.2 (37.7–38.7)203
Murmur206 (73.3)281
Gallop67 (23.8)281
Arrhythmia (auscultation)50 (17.8)281
Arrhythmia (ECG)76 (31.1)244
Bradyarrhythmia6 (2.5) 
Supraventricular15 (6.1) 
Ventricular49 (20.1) 
Supraventricular and ventricular6 (2.5) 

All echocardiographic variables were measured within a 2-month period, with 77 echocardiographic studies being measured by the board-certified cardiologist (VLF) and 205 studies measured by the resident (KB). Interobserver repeatability of echocardiographic measurements was assessed in 12 randomly selected cats. 2D and M-mode continuous measurement variables had CVs of <10% and Doppler variables had CVs of <15%, except IVRT and septal tissue Doppler imaging variables, which had CVs <20%. All categorical measurement variables had kappa values of 0.75–1.00.

Systolic anterior motion of the mitral valve (SAM) was present in 179 (64.9%) cats (Table 4). Cats with SAM were younger (median [IQR]: 4.5 [2.2–8.6] years) than cats without SAM (median [IQR]: 9.0 [5.2–11.6] years) (P < .001). Only 15 of the 282 cats in the study were sedated for their echocardiographic examination. Sedation protocols consisted of butorphanol alone; butorphanol with either midazolam or acepromazine; or midazolam with ketamine.

Table 4. Echocardiographic variables in all cats with HCM at first presentation.
Echo AssessmentPresent (%)/Mean ± SD/Median (IQR)Assessed in
  1. HCM, hypertrophic cardiomyopathy; LVWd, end-diastolic left ventricular septal or free wall thickness; FS%, left ventricular fractional shortening; LA:Ao, short-axis ratio of diastolic left atrial diameter to aortic root diameter; LAD, the diameter of the left atrium measured parallel with the mitral annulus in the last frame before mitral valve opening; LAA, left auricular appendage velocity; LA-EF%, left atrial ejection fraction; LA-FS%, left atrial fractional shortening; SEC, spontaneous echo-contrast; MR, mitral regurgitation; SAM, systolic anterior motion of the mitral valve; E:A, ratio of mitral inflow peak early filling (E) to late filling (A) velocities; S:D, ratio of the pulmonary venous peak systolic (S) to peak diastolic (D) velocity; IVRT, isovolumic relaxation time; Adur : ARdur, ratio of the duration of the late filling transmitral wave (Adur) to the duration of the pulmonary venous atrial reversal wave (ARdur); E:E′, ratio of mitral inflow peak early filling to early diastolic tissue Doppler mitral annular velocity.

Max 2D LVWd thickness (mm)7.8 (7.0–8.9)276
FS%50.3 ± 13.4265
LV Regional wall hypokinesis15 (5.5)273
LA:Ao1.44 (1.22–1.92)241
LAD (mm)16.6 (14.8–21.0)240
LAA (m/s)0.26 (0.17–0.43)47
LA-EF% (%)50.4 (27.2–61.9)221
LA-FS% (%)18.1 ± 8.6220
SEC/thrombus48 (17.5)275
MR188 (77.0)244
SAM179 (64.9)276
Diastolic filling pattern
Normal7 (6.1)114
Delayed31 (27.2)
Pseudonormal44 (38.6)
Restrictive32 (28.1)
E:A (≤0.2 m/s separation)1.31 (0.98–2.28)114
S:D1.33 ± 0.57138
E:IVRT16.5 (12.2–21.3)95
Adur : ARdur1.05 ± 0.5270
E:E′14.2 (11.1–19.0)100

More than a quarter of the population (n = 76, 27.0%) did not receive any cardiac medications (Table 5). Of the 94 cats that received furosemide, 74 cats received furosemide before the echocardiographic examination used for this study and 20 cats received furosemide treatment after echocardiography. Of the 75 cats that received beta blockers, 12 were already receiving beta blockers at the time of the echo, having been diagnosed on a previous occasion.

Table 5. Medications dispensed to all cats with HCM.
Medicationn (%)
  1. HCM, hypertrophic cardiomyopathy; ACE, angiotension-converting enzyme.

No treatment76 (27.0)
Beta-adrenergic antagonists75 (26.6)
ACE-inhibitors97 (34.4)
Diltiazem4 (1.4)
Furosemide94 (33.3)
Pimobendan18 (6.4)
Spironolactone16 (5.7)
Aspirin45 (16.0)
Clopidogrel4 (1.4)
Low-molecular-weight heparin3 (1.1)

During the follow-up period, 164 cats died, of which 57 were considered noncardiac and 107 were considered cardiac deaths. Sixteen cats died or were euthanized during their initial visit to the QMHA, of which 8 deaths were caused by cardiac causes. Cardiac deaths (either spontaneous death or euthanasia) included sudden death (n = 17), CHF (n = 56), and ATE (n = 34). Median (range) follow-up time was 729 (0–2,755) days and median (range) survival time (all-cause mortality) was 1,007 (0–2,755) days. Median (range) survival time (cardiac mortality) was 2,153 (0–2,755) days. Age at death ranged from 0.5 to 19.9 years (Fig 1).

Figure 1.

Kaplan–Meier curve—age at death—cardiac and all-cause mortality.

Univariable predictors of increased risk of cardiac death (Table 6) included older age; absence of a murmur; a gallop sound or arrhythmia; CHF or ATE; extreme LV hypertrophy (≥9.0 mm); decreased LV systolic function (FS% <30); regional wall hypokinesis; increased LA size; decreasing LA function; SEC, thrombus or both; absence of left ventricular outflow tract obstruction; S:D ≤0.5; E:IVRT ≥20; Adur : ARdur <1.0; E:E′ >19.0; and an overall restrictive diastolic filling pattern. Factors not significantly associated with survival included sex (P = .617), breed (pedigree versus nonpedigree, P = .270), weight (P = .162), heart rate (P = .335), respiratory rate (P = .056), and the presence of MR (P = .569). The effect of the multiple treatment regimes on survival was not analyzed in the overall population, but the subpopulation of asymptomatic cats that did not receive cardiac medications had a median (range) survival time of 2,171 (8–2,755) days.

Table 6. Univariable predictors of outcome (cardiac mortality) in all cats with HCM.
FactorHR95% CIP valueNumber with Factor/Assessed in
  1. ATE, arterial thromboembolism; CHF, congestive heart failure. For other abbreviations, see Table 4.

Age (per year of life)1.071.02–1.12.003282
No murmurReference <.00175/281
Murmur0.340.23–0.51 206/281
No gallopReference .009214/281
Gallop1.771.16–2.70 67/281
No arrhythmiaReference  231/281
Arrhythmia3.252.15–4.90<.00150/281
No arrhythmia on ECGReference <.001168/244
Arrhythmia on ECG2.381.57–3.61 76/244
Clinical presentation  <.001 
AsymptomaticReference3.17–12.77<.001158/279
ATE6.364.3–10.55<.00117/279
CHF6.740.88–5.86.09288/279
Syncope2.27  16/279
No extreme hypertrophy (6.0–8.9 mm)Reference <.001210/276
Extreme hypertrophy (≥9.0 mm)3.312.22–4.95 66/276
FS% >30%Reference <.001245/265
FS% ≤30%6.593.70–11.73 20/265
No regional wall hypokinesisReference <.001258/273
Regional wall hypokinesis6.453.44–12.09 15/273
LA:Ao (for a unit change of 1.0)4.133.13–5.44<.001241
LAD  <.001 
≤16.0 mmReference1.41–4.70.002102/240
16.1–21.0 mm2.585.97–19.30<.00179/240
≥21.1 mm10.73  59/240
LA-EF% >30%Reference <.001164/221
LA-EF% ≤30%9.726.13–15.40 57/221
LA-FS% (for a unit change of 1%)0.870.84–0.90<.001220
No SEC/thrombusReference <.001227/275
Presence of SEC/thrombus7.334.78–11.25 48/275
LAA >0.26Reference .00224/47
LAA ≤0.263.911.62–9.43 23/47
No SAMReference <.00197/276
Presence of SAM0.480.33–0.71 179/276
No restrictive filling patternReference <.00182/114
Restrictive filling pattern3.081.68–5.66 32/114
S:D >0.5Reference <.001128/138
S:D ≤0.55.972.65–13.44 10/138
E:IVRT <20Reference .00964/95
E:IVRT ≥202.531.25–5.09 31/95
Adur : ARdur ≥1.0Reference .01135/70
Adur : ARdur <1.03.131.30–7.57 35/70
E:E′ ≤19.0Reference .00874/99
E:E′ >19.02.491.26–4.89 25/99

Several multivariable models were generated, all of which included a measure of LA size or function and LV systolic function as well as 1 or 2 other factors. The model with the best fit showed that extreme hypertrophy (≥9.0 mm), FS% ≤30%, and decreasing LA-FS% were independently associated with increasing hazard of cardiac death in this population of cats with HCM (Fig 2). The variable LA-FS% was entered in the model as a continuous variable, so that the hazard of a cardiac death increased per unit decrease in LA-FS% (a 1% increase in LA-FS% changed the hazard of dying a cardiac death by 0.89). For example, the hazard of a cardiac death for a cat with an LA-FS% of 18.1% was less than one-third (0.31) the hazard for a cat with an LA-FS% of 8.1%. Graphical assessment of log cumulative hazard plots and statistical analysis of the Schoenfeld residuals showed no violation of the proportional hazards assumptions and assessment of the Cox-Snell residual plots showed that the model was a good fit to the data.

Figure 2.

Cox's proportional hazard model of factors independently associated with survival (n = 210) *for a unit increase of 1%. LVWd, end-diastolic left ventricular septal or free wall thickness; FS%, left ventricular fractional shortening; LA-FS%, left atrial fractional shortening.

Discussion

Several retrospective HCM studies in cats have reported survival times,[3, 5, 7, 8, 10, 11] but to the authors' knowledge, this study presents the effects on prognosis of a larger number of echocardiographic variables than have been published previously.

Although median (range) survival time for the entire population (2,153, IQR, 0–2,755 days) was longer than previously reported (709–1,276 days),[5, 8, 10] this may partly reflect the high percentage of asymptomatic cats in our population (56.0%) compared with previous studies (33.5–46.4%),[8, 10] or may reflect differences in treatment protocols or euthanasia recommendations.

Decreasing LA systolic function (in this study, measured using LA-FS%), decreased LV systolic function (FS% ≤30%), and extreme hypertrophy (≥9.0 mm) all were independent predictors of an increased hazard of cardiac death in our population of cats.

Left atrial enlargement is consistently reported as an indicator of decreased survival times in cats with HCM,[7-10, 12] although in humans with HCM, the use of LA function to predict cardiovascular outcome is considered as good as, if not better than, LA size.[16] Our study shows an increasing hazard of cardiac death at the univariable level with increasing LA size, whether measured as LA:Ao or LAD. A recent study demonstrated worse LA function in cats with CHF compared with healthy controls.[39] Using Cox proportional hazards analysis, models involving LA-FS% showed a better fit than models involving measures of LA size.

Human HCM patients with decreased LV systolic function (ie, with left ventricular ejection fraction <50%) are considered to be “end-stage HCM” patients. These patients are more likely to progress to advanced CHF and die sooner than those without systolic dysfunction.[17-19] An end-stage HCM phenotype has been reported in 1 family of cats,[40, 41] but systolic dysfunction has not been widely evaluated as a prognostic indicator for cats with HCM. Considering the prognostic importance of systolic dysfunction in human HCM, it is not surprising that low FS% was an important prognostic indicator in the cats presented here.

Presence of extreme hypertrophy (≥9.0 mm) has not been described as a prognostic indicator in previous studies of cats, but was an independent predictor of cardiac mortality in our population. The relationship between LV wall thickness and increased risk of cardiac death was not linear, with similar survival times for cats with diastolic LV wall thicknesses from 6.0 to 8.9 mm.

The age at presentation (6.2 years) was similar to previously reported studies,[3, 5, 8, 10] with a wide age range at both presentation (0.4–17.4 years) and death (0.5–19.9 years). Although not an independent predictor in the multivariable analysis, older age at first diagnosis was shown in our study and in previous studies[8, 10, 12] to be associated with an increased hazard of cardiac death at the univariable level (HR [95% CI]: 1.07 [1.02–1.12]).

A male predisposition in HCM was observed in this study and is consistently reported,[3, 5, 7, 8, 10, 11] even in the human literature,[42-45] despite the apparent autosomal dominant inheritance of HCM.[46, 47] Despite the male predisposition, males did not appear to have significantly different survival times after diagnosis than females, in either our study or previous studies.[5, 10]

Breed variation in outcome has been reported, with Ragdolls[10] and Maine Coons[11] believed to have shorter survival times than other breeds. The low number of cats representing individual breeds in our study prevented specific comparisons, but a comparison of nonpedigree and pedigree cats did not show a statistical difference in survival time, in common with a previous study.[5]

Clinical status at diagnosis is well-recognized as a prognostic indicator.[5, 8, 10, 11] In our study, cats presenting with CHF or ATE had the shortest survival times, with no significant difference between the 2 groups, as also seen in previous studies.[5, 8] In a previous study, cats with syncope had similar survival times to CHF and ATE cats.[8] In our study, cats with syncope did better than cats with CHF or ATE, and were not significantly worse than asymptomatic cats, although the small number of syncopal cats (5.7% of our population, 3.8% of the population in the paper by Rush et al[8]) makes this difficult to evaluate.

As found in other recent studies,[8, 10] the presence of tachycardia (HR >200/min) was not associated with a worse outcome, but the presence of a gallop sound or an arrhythmia was associated with worse survival. In contrast, the presence of a murmur was associated with longer survival times. Gallop sounds and arrhythmias are likely to be associated with severe disease, and therefore an increased risk of cardiac death. This has important clinical relevance because access to expensive equipment or specialist expertise is not necessary for detection of a gallop sound or arrhythmia. The presence of a murmur often is attributable to the presence of SAM causing dynamic left ventricular outflow tract obstruction. In contrast with human HCM patients,[13-15] SAM has not been reported to be a negative prognostic indicator in cats.[7, 8, 10] In our study, cats with SAM were younger (P < .001) and had longer survival times (P < .001) than cats without SAM.

Regional hypokinesis was associated with a very high hazard of cardiac death (HR [95% CI]: 6.45 [3.44–12.09]) in our population. Although we suspected that regional wall hypokinesis might be associated with infarction of the LV free wall, we did not obtain necropsy confirmation of this speculation in any cat.

Various measures of diastolic function were assessed in our study population. Measures of diastolic function are used routinely for prognostication in a variety of cardiac diseases in humans,[20-26] with E:A ratios, S:D ratios, and, most recently, E:E′ providing important information. In people with HCM, E:A ratios have poor correlation with filling pressures,[48] but a restrictive filling pattern is associated with exercise intolerance, development of end-stage HCM, sudden death, and CHF.[27, 28] Although these measures of diastolic function all were important at the univariable level in our study, no measure of diastolic function was a significant predictor of outcome in multivariable analysis. This may either be because other factors had a greater impact on survival, or may be because not enough cats had diastolic function documented. None of these variables was recorded in more than 50% of our population, which affected the validity of the derived models.

Because this was a retrospective study, we were unable to account for the multiple factors influencing choice of treatment, which could confound any possible effects on survival. Factors influencing treatment choice may include differences in owner attitudes to cost and euthanasia, differences in preferences among clinicians and in a given clinician over time, drug availability, variations in dosage and dosing frequencies, comorbidities and their treatments, and the use of multiple drug combinations. Some cats with HCM had a very good prognosis, even without treatment, but accurate identification of this group is constrained by the retrospective nature of the study.

Ours was a retrospective study conducted in a referral environment and as such, there are limitations. Multiple clinicians were involved in investigating the cases. Based on their clinical experience and the wishes and financial constraints of the owners, not every case had every cause of LV hypertrophy excluded. Blood pressure was not recorded for every cat, so it is possible that hypertrophy was secondary to systemic hypertension in some cats. However, we made deliberate efforts to exclude cats with factors predisposing to hypertension, such as underlying renal disease. Serum thyroxine concentrations were not measured in every older cat. It is thus possible that not every cat had true “idiopathic” HCM. Arrhythmias were evaluated with 24-hour ambulatory monitoring in only a few cases and consequently, accuracy in characterizing the type and frequency of arrhythmias was limited. Each cat only had 1 echocardiographic examination measured and used for the study. Some cats that had unstable cardiac disease were included and these cats may have had different measurements a few hours later (eg, LA size or function). Some cats already were receiving furosemide or a beta blocker at the time of the echocardiographic examination and this may have had some effect on the variables measured. Progression of cardiac disease and response of the heart to cardiac medications would be a potentially interesting area of investigation. Cases described as noncardiac chylothorax because of the presence of normal LA dimensions may have had constrictive cardiac disease rather than noncardiac disease. Measurements were not available for all indices for every case, especially for some of the echocardiographic variables, because of clinician preference, stability of patient symptomatic status, or patient temperament. Our assessment of whether LV regional wall hypokinesis was present was subjective. Reliance was placed on owner description of the outcomes for cats that died at home and on the clinical records of multiple practices for those that were euthanized. Veterinary survival studies are always confounded by variability in selection of timing for euthanasia versus survival attributable to spontaneous death, and this is also true of our study.

The prognosis for cats with HCM is very variable, with some experiencing normal life spans and dying of noncardiac disease, whereas others experienced sudden death, CHF, or ATE. Measures of LA systolic function, LV systolic function, and LV maximal wall thickness are all 2D or M-mode measures, and should be included in diagnostic investigations wherever possible to provide prognostic information. Presence of a gallop or arrhythmia on auscultation and measures of diastolic function also can provide valuable prognostic information.

Acknowledgments

JRP's PhD was funded by Everts Luff Feline Endowment and IDEXX laboratories.

Grant support: JRP's PhD was funded by Everts Luff Feline Endowment and IDEXX laboratories.

Conflict of Interest Declaration: Authors disclose no conflict of interest.

Footnotes

  1. 1

    Vivid 7; GE Medical Systems Ltd, Hatfield, Hertfordshire, UK

  2. 2

    Model 811-B Doppler Ultrasonic Flow Detector; Parks Medical Electronics Inc, Aloha, OR

  3. 3

    GraphPad Prism 5; GraphPad Software, 2007, La Jolla, CA; and PASW Statistics 20, 2011, Armonk, NY

Ancillary