Article ID: CJ-17-0845
Background: There have been few studies on the clinical course of hypertrophic cardiomyopathy (HCM) in a community-based patient cohort in Japan.
Methods and Results: In 2004, we established a cardiomyopathy registration network in Kochi Prefecture (the Kochi RYOMA study) that consisted of 9 hospitals, and finally, 293 patients with HCM were followed. The ages at registration and at diagnosis were 63±14 and 56±16 years, respectively, and 197 patients (67%) were male. HCM-related deaths occurred in 23 patients during a mean follow-up period of 6.1±3.2 years. The HCM-related 5-year survival rate was 94%. In addition, a total of 77 cardiovascular events that were clinically severe occurred in 70 patients, and the HCM-related 5-year event-free rate was 80%. Multivariate Cox proportional hazards model analysis showed that the presence of NYHA class III at registration was a significant predictor of HCM-related deaths and that the presence of atrial fibrillation, lower fractional shortening and presence of left ventricular outflow tract obstruction in addition to NYHA class III were significant predictors of cardiovascular events.
Conclusions: In our unselected registry in an aged Japanese community, HCM mortality was favorable, but one-fifth of the patients commonly suffered from HCM-related adverse cardiovascular events during the 5-year follow-up period. Careful management of HCM patients is needed, particularly for those with the above-mentioned clinical determinants.
Hypertrophic cardiomyopathy (HCM) is a primary myocardial disorder with heterogeneous morphological, functional, and clinical features.1–3 Recent observations in community-based cohorts in Western countries suggest a more favorable prognosis than that suggested by the results of previous studies.4–10 However, there have been few studies of the clinical features of HCM in a prospectively assembled, community-based cohort in Japan.11 Therefore, in 2004, we established the Kochi Cardiomyopathy Network, named the Kochi RYOMA (registry of myocardial diseases) study, to provide detailed information on the clinical features of HCM in an unselected regional Japanese population.12 For the present study, we obtained follow-up data in order to assess the clinical course and to clarify the prognostic factors in a regional Japanese HCM cohort.
In 2004, we established the Kochi Cardiomyopathy Network consisting of 9 hospitals serving as primary, secondary, and tertiary referral medical centers for cardiovascular patients in Kochi Prefecture, Japan, which has approximately 800,000 inhabitants. Between February 2004 and December 2013, 305 patients with a diagnosis of HCM were registered. Shortly afterward, 3 of them were diagnosed as having a specific cardiomyopathy: 1 with cardiac amyloidosis and 2 with cardiac involvement of Fabry disease. Finally 302 patients were registered. In this longitudinal study, we excluded 9 patients with no follow-up data and so the final study population consisted of 293 patients. The diagnosis of HCM was based on echocardiographic demonstration of unexplained left ventricular hypertrophy (LVH, i.e., maximum LV wall thickness ≥15 mm). Informed consent was given by all patients or their parents in accordance with the guidelines of the Ethics Committee on Medical Research of Kochi Medical School.
Evaluation of the patients included medical history, clinical examination, 12-lead ECG, M-mode, 2-D and Doppler echocardiography, and ambulatory 24-h Holter ECG analysis. The severity and distribution of LVH were assessed in the parasternal short-axis plane at the mitral valve and papillary muscle levels. Left ventricular end-diastolic diameter (LVEDD) and end-systolic diameter (LVESD) were measured from M-mode and 2-D images obtained in the parasternal long-axis view, and LV fractional shortening (LVFS=(LVEDD−LVESD)/LVEDD×100%) was calculated. The LV outflow tract (LVOT) gradient was calculated from continuous-wave Doppler using the simplified Bernoulli equation.
Based on morphologic and hemodynamic assessments by echocardiography, we divided the patients into the following 5 groups: (1) hypertrophic obstructive cardiomyopathy (HOCM), defined as the presence of basal LVOT obstruction (LVOTO: gradient ≥30 mmHg at rest), (2) midventricular obstruction (MVO), defined as the presence of systolic LV cavity obliteration at the midventricle creating midventricular obstruction with a peak systolic gradient ≥30 mmHg at rest, (3) dilated phase of HCM (D-HCM), defined as LV systolic dysfunction of global ejection fraction (EF) <50% (global EF was determined from apical 2- and 4-chamber views by modified Simpson’s method), (4) apical HCM, defined as hypertrophy confined to the LV apex, and (5) others: HCM without obstruction other than D-HCM and apical HCM.13
For survival analysis, 3 types of HCM-related death were defined: (1) sudden death, in which unexpected sudden collapse occurred in patients with a relatively stable or uneventful clinical course; (2) heart failure (HF)-related death, which was in the context of progressive cardiac decompensation ≥1 year before death, particularly if complicated by pulmonary edema or evolution to D-HCM; and (3) stroke-related death, which occurred as a result of probable or proven embolic stroke.7–10,13 Other morbid events included hospitalization for HF, embolic stroke admission, and spontaneous sustained ventricular tachycardia (VT) associated with hemodynamic instability or appropriate implantable cardioverter-defibrillator (ICD) discharge. All cardiovascular events were a composite of HCM-related death and the morbid events described. Arrhythmic events were a composite of sudden death, spontaneous sustained VT associated with hemodynamic instability, and appropriate ICD discharge. Composite HF events included HF-related death and hospitalization for HF. Data on the survival and clinical status of patients were obtained during serial clinic visits or from records on their clinical charts, including information from other institutes. The study closed on December 31, 2014.
All data are expressed as mean±SD or frequency (percentage). Differences in continuous variables were assessed using Student’s t-test or the Mann-Whitney U test. Pearson’s chi-square test was used for comparisons between non-continuous variables, and Fisher’s exact test was used when the expected frequency was <5. Event-free estimates curves were calculated by the Kaplan-Meier method, and the log-rank test was used for comparison. The multivariate Cox proportional hazards model was used to analyze the relationships between HCM-related deaths or all cardiovascular events and prognostic parameters including some variables with P values ≤0.1 in the univariate analysis and variables considered to be clinically important. Statistical significance was defined as P≤0.05. Statistical analysis was performed using SPSS version 14.0J (SPSS Inc., Chicago, IL, USA).
The clinical characteristics of the 293 patients with HCM at registration are summarized in Table 1. The ages at registration and at diagnosis were 63±14 (range: 7–88) and 56±16 (range: 6–87) years, respectively, and 197 patients (67%) were male; 40 patients (14%) were diagnosed as HCM at less than 40 years old. Nearly half (46%) of the patients were diagnosed symptomatically; 76 patients (26%) had proven familial HCM and 52 patients (18%) had a family history of sudden death. Most of the 293 patients were completely asymptomatic or mildly symptomatic at registration: 163 (56%) were NYHA functional class I, 109 (37%) were NYHA class II, and only 21 (7%) were NYHA class III. At registration, 86 patients (29%) had documentation of paroxysmal or chronic atrial fibrillation (AF). Of the 293 patients, there were 40 patients in the HOCM group, 8 patients in the MVO group, 13 patients in the D-HCM group, 52 patients in the apical HCM group, and 180 patients in the other group.
*Holter ECG was performed in 213 patients during followup period. ACEI, angiotensin-converting enzyme inhibitor; AF, atrial fibrillation; ARB, angiotensin II receptor blocker; D-HCM, dilated phase of hypertrophic cardiomyopathy; HCM, hypertrophic cardiomyopathy; HOCM, hypertrophic obstructive cardiomyopathy; ICD, implantable cardioverter-defibrillator; LVOTO, left ventricular outlet tract obstruction; MVO, midventricular obstruction; NYHA, New York Heart Association; VT, ventricular tachycardia.
The mean follow-up period from registration of the patient cohort was 6.1±3.2 years, and 44 patients died. HCM-related deaths occurred in 23 of the 44 patients: sudden deaths in 9, HF deaths in 9 and embolic stroke deaths in 3 patients. The other 21 patients had non-HCM-related deaths: cancer in 8, infectious disease in 5, respiratory disease, renal failure and intestinal bleeding in 1 patient each, suicide in 2 patients and unknown causes in 3 patients. Among the 9 cases of sudden death, unexpected death occurred during physical labor in only 1 patient. The remaining 8 patients died in situations of no physical stress. Holter ECG monitoring was performed in 8 patients during the follow-up period and 3 showed non-sustained VT. Concerning the medications in these 9 patients, 6 took a β-blocker and none were taking amiodarone. Of the 11 patients with HF death, 5 had D-HCM. Angiotensin-converting enzyme inhibitor or angiotensin II receptor blocker was used for 6 patients, β-blocker for 9 patients and diuretics in all 11 patients. In the group analysis for the study population, the distribution of age at HCM-related death differed for the 3 types (Figure 1): mean age of 62±17 years for sudden deaths, 76±10 years for HF deaths and 75±7 years for embolic stroke deaths. Patients in whom sudden death occurred were younger than patients who died of HF or embolic stroke (mean age: 62±17 years vs. 76±9 years, P=0.017).
Distribution of age at HCM-related death. HCM, hypertrophic cardiomyopathy.
Figure 2 shows the Kaplan-Meier curves for HCM-related deaths. The HCM-related annual mortality rate was 1.3% and the HCM-related 5-year survival rate was 94%.
Kaplan-Meier curves for HCM-related deaths. HCM, hypertrophic cardiomyopathy.
Table 2 shows the clinical characteristics at registration of patients with and without HCM-related death. A sex difference was not seen between the 2 groups. The prevalence of both NYHA class III and AF in patients with HCM-related deaths was significantly higher than in patients without HCM-related deaths. Echocardiography showed that the prevalence of HOCM was comparable between the 2 groups. Patients in the HCM-related death group had a higher prevalence of D-HCM and a lower prevalence of apical HCM. Patients with HCM-related deaths showed larger LVEDD and left atrial size and lower LVFS than did patients without HCM-related deaths.
Abbreviations as in Table 1.
During the follow-up period, a total of 77 cardiovascular events in 70 patients occurred: arrhythmic events in 19 patients, including 9 sudden deaths; HF events in 35 patients, including 11 HF deaths; and embolic events in 23 patients including 3 embolic stroke deaths. Among the 23 patients with embolic events, 12 patients had AF prior to the embolic complications and all 12 patients took warfarin; 6 patients had documented AF after the embolic events and in the remaining 5 patients, AF was not detected during the follow-up period. Figure 3 shows the incidence of HCM-related adverse cardiovascular events. The HCM-related 5-year event rate was 20%.
Incidence of HCM-related adverse cardiovascular events. HCM, hypertrophic cardiomyopathy.
Table 3 shows the clinical characteristics at registration of patients with and without adverse cardiovascular events. The prevalence of both NYHA class III and AF was significantly higher in patients with adverse cardiovascular events than in patients without adverse cardiovascular events. Echocardiography showed that patients in the cardiovascular events group had higher prevalence of HOCM and of D-HCM and a lower prevalence of apical HCM. As was found in the HCM-related death group, patients with cardiovascular events showed larger LVEDD and left atrial size and lower LVFS than patients without cardiovascular events.
Abbreviations as in Table 1.
Multivariate Cox proportional hazards model analysis including NYHA class III, presence of AF and LVFS showed that the presence of NYHA class III at registration was a significant predictor of HCM-related death (Table 4). NYHA class III, presence of AF, lower FS and presence of LVOTO were significant predictors of adverse cardiovascular events by similar multivariate analysis including age at registration, sex, NYHA class III, presence of AF, maximum LV wall thickness, LVFS, and presence of LVOTO (Table 4).
CI, confidence interval; LVFS, left ventricular fractional shortening; MLVWT, maximum left ventricular wall thickness. Other abbreviations as in Tables 1,2.
HCM is a primary myocardial disorder with heterogeneous clinical presentation and course.1–3 Although the natural history of HCM varies from an asymptomatic and benign clinical course to sudden premature death, there have been very few studies on the clinical course of HCM in a community-based patient cohort in Japan. The Kochi Cardiomyopathy Network, named the Kochi RYOMA study, was established as a community-based and unselected Japanese patient cohort to provide detailed information on the clinical features of HCM and to establish a strategy for better management of this disease.12 To the best of our knowledge, this is the first report of a prospective multicenter investigation showing the clinical course of HCM in a regional Japanese cohort.
Our regional cohort in the Kochi RYOMA study was older than HCM cohorts in previous studies,7–11,14 probably because Kochi Prefecture, where our study was performed, is located far from urban areas and is one of the most aged communities in Japan. We believe that, at least in Japanese rural regions, many patients with this disease are middle-aged or elderly despite the fact that HCM is regarded as a genetic disorder. The percentage of male patients in our cohort (67%) was similar to the percentages in previous studies in Western countries and Japan.7–11,14 Although half of the patients were diagnosed by symptoms only, severe functional disability (NYHA class III/IV) was not common in our study. The reported prevalence of AF in several HCM populations is approximately 20%.8,15,16 In our study, the prevalence of AF was 29% and this relatively high prevalence is probably related to our aged cohort. As for the subtypes of HCM, 40 patients (14%) had HOCM. This prevalence of outflow obstruction was lower than in cohorts in Western countries.7–10,17 The reasons for the lower prevalence of HOCM might be that (1) in some patients at registration, LVOTO had been improved by treatment or had disappeared through the natural long-term course, and (2) there was a large percentage of patients with apical HCM (18% in the whole cohort), which is higher than in Western countries.18,19
Our prospective community-based HCM study showed a favorable prognosis. The HCM-related annual mortality rate was 1.3%, which is in accordance with previously reported data obtained from community-based cohorts for HCM in Western countries.10 Compared with the results from referral centers for HCM, a significant portion of patients with HCM in regional cohort populations die from progressive HF or stroke, in addition to premature sudden death.7,20
In our study, we also focused on HCM-related adverse cardiovascular events, which are clinically severe, including hospitalization for HF or embolic stroke, and successful recovery from sustained VT associated with hemodynamic instability such as cardiopulmonary arrest, and appropriate ICD discharge. For the overall study group, the HCM-related 5-year event rate was 20%, which means that almost 1 in 5 HCM patients had a HCM-related cardiovascular event within 5 years. From these results, we realized that clinically important morbid events frequently occurred in our community-based cohort, although the survival rate was good.
Regarding the risk factors for HCM-related cardiovascular events in our study, NYHA class III was an independent predictor of HCM death. Furthermore, the presence of AF, lower %FS and presence of LVOTO in addition to NYHA class III were significant predictors of adverse cardiovascular events. These factors are known to be associated with prognosis.8,15,21–25 Firstly, NYHA class is known to be associated with HCM death. Maron et al reported that some patients (9%) showed severe, limiting symptoms consistent with NYHA class III or IV and these patients had worse survival rate in a HCM cohort in regional USA.8 There was a recent report that advanced HF (NYHA class III or IV) not associated with outflow obstruction portended a severely unfavorable prognosis in an Italian cohort.26 Secondly, AF represents the most common sustained arrhythmia in patients with HCM. Olivotto et al reported that HCM patients with AF had increased risk for HF-related death, stroke, and severe functional disability.15 Thirdly, lower %FS reflects D-HCM. Although LV systolic function is supernormal or preserved in most cases of HCM, progression to systolic impairment occurs in approximately 5–10% of patients when they are followed for long enough. After the dilated phase becomes established, the prognosis for these patients has been reported as poor.21–23 In our study, 10 of 13 patients with D-HCM at registration suffered from HCM-related cardiovascular events but none had a HCM-related death. Finally, LVOTO is an important pathophysiological component of HCM, initially described in the late 1950s. LVOTO is known to be an independent determinant of limiting symptoms and adverse clinical consequences such as progressive HF and cardiovascular death.24,25 On the other hand, MLVWT previously identified as a risk factor for sudden death in HCM was not associated with HCM mortality and morbidity in our regional cohort.27 There are several positive and negative reports on the relationship between MLVWT and HCM-related cardiac events.28–30 It seems better to consider MLVWT in the context of a multifactorial approach to risk stratification.30
There are several limitations to be acknowledged. First, the number of HCM-related events, particularly HCM-related deaths, was relatively small and some of the statistical analyses might have been affected. Second, Holter monitoring should have been performed in all patients at registration. Almost 70% of the patients had Holter ECG monitoring during the follow-up period. Therefore, the clinical significance of non-sustained VT is not fully understood.
In our unselected registry in an aged Japanese community, HF and embolic events causing disability were very important. Although the HCM-related mortality rate was favorable, one-fifth of the patients suffered from HCM-related adverse cardiovascular events during the 5-year follow-up period. Careful management of HCM patients is needed, particularly those with severe symptoms judged by NYHA class, AF, LVOTO or LV systolic dysfunction, which are independent risk factors for complications.
Participating investigators from the study hospitals were Kazuya Kawai, MD, Masahiko Fukatani, MD, Shoichi Kubokawa, MD, Yoko Nakaoka, MD (Chikamori Hospital); Takashi Yamasaki, MD (National Hospital Organization Kochi National Hospital); Yoko Hirakawa, MD (Tosa Municipal Hospital); Masanori Kuwabara, MD (Kochi Prefectural Aki General Hospital); Katsutoshi Tanioka, MD, Tatsuya Noguchi, MD, Kazuya Miyagawa, MD (Kochi Medical School). We would like to thank the physicians who made this study possible.