The PRo-HCM registry was conceived by the Working Group on Myocardial and Pericardial Diseases of the Portuguese Society of Cardiology, directed by an executive and a scientific committee, and managed in the Portuguese National Center for Data Collection in Cardiology (CNCDC). This study was formulated and conducted in compliance with the principles of the declaration of Helsinki, and approved by the National Center for Data Protection. It was an observational, multicenter, voluntary, non-mandatory study, with a two-year enrollment period (April 2013-April 2015), retrospective but including a prospective update.

A direct invitation was made to cardiology departments nationwide, central and regional, public and private, academic and non-academic, covering rural and urban, coastal and inland areas. Additionally, the registry was advertised in the Portuguese Journal of Cardiology, meetings and newsletters. If the invitation was accepted, the principal investigator received detailed instructions, a center identification number and a unique username and password to gain access to the electronic case report form (CRF) (http://www.spc.pt/RegistosMiocardiopatia/Public/Login.aspx?ReturnUrl=%2fRegistosMiocardiopatia%2f). The CRF contained seven sections: (1) patient identification and demographic/epidemiological data; (2) past history and baseline clinical data; (3) mortality and risk stratification; (4) diagnostic tests; (5) genetic testing, family screening, and genetic counseling; (6) treatment; (7) last assessment (clinical course, follow-up, and outcomes). In the diagnostic tests section the investigators were asked to enter the exams performed at the time of first assessment, including electrocardiogram (ECG), echocardiogram, ambulatory ECG, exercise test, exercise echocardiogram, cardiac magnetic resonance (CMR), and cardiac computed tomography.

Centers were asked to include all patients with a diagnosis of HCM followed at the center currently or in the past (no retrospective time limit), including those already deceased at the time of enrollment. Written informed consent was obtained from living patients and from a proxy of deceased patients.

Inclusion criteria were age >18 years at the time of enrollment, and unexplained left ventricular hypertrophy (LVH): wall thickness ≥15 mm by imaging techniques (in first-degree relatives10 ≥14 mm in the inferior interventricular septum (IVS) or lateral wall or ≥13 mm in the anterior IVS or inferior wall).

Exclusion criteria were secondary LVH (grade ≥2 hypertension11), moderate or severe aortic stenosis,12 previously diagnosed cardiac or systemic disease, and metabolic or multi-organ syndrome associated with LVH.

After the inclusion period, extra time was provided to complete the CRFs and to clean the database. The final date of registry closure was December 31, 2015. CRFs were reviewed to confirm consistency of data. Whenever necessary, queries were sent to investigators. In the event of repeated patients (same initials, gender and birth date), the one with the longer follow-up time was included.

Throughout the study, most data are relative to the time of first visit. When clinically relevant, data at the time of diagnosis of HCM are also shown.

Follow-up time was defined as time from initial assessment at the center to last assessment or death.

Sudden cardiac death (SCD) was defined as unexpected death occurring within one hour of symptom onset in patients who had previously experienced a relatively stable or uneventful clinical course. Resuscitation from cardiac arrest or appropriate implantable cardioverter-defibrillator (ICD) therapies for primary prevention were considered as equivalents of SCD.

HF-related death was defined as that occurring in the context of progressive cardiac decompensation, with decline in left ventricular function.13 Heart transplantation was considered as equivalent to HF-related death.

Stroke-related deaths in the setting of paroxysmal, persistent or permanent AF were classified as AF-stroke related deaths. Stroke-related deaths in the absence of documented AF were not included in this group.

Thromboembolic events, defined as stroke, transient ischemic attack (TIA) or systemic peripheral embolism, were recorded.14

The classification of identified genetic variants was assigned to the investigators, as pathogenic/probably pathogenic, of unknown significance or benign/probably benign, according to current knowledge of their pathogenicity,15,16 as provided by genetic laboratories (these data were not centrally reviewed or corrected by the coordinators of the registry). A genetic study was defined as negative if no pathogenic/probably pathogenic mutation was detected and as in progress if no result was provided at inclusion.

Continuous variables were expressed as mean and standard deviation or as median and interquartile range (IQR) (P25-P75). Categorical variables were given as total number and percentages. Chi-square or Fisher tests were used for comparisons of categorical variables and Student's t tests for continuous variables. Survival was assessed by Cox proportional hazard regression. Survival curves were constructed according to the Kaplan-Meier method, and comparisons were performed using the log-rank test. All p-values were two-sided and considered significant when <0.05. All analyses were performed using SPSS 19.0®.

Of the 62 institutions contacted, 37 accepted, and the final number of participating centers was 29 (Figure 1). The total number of patients was 1042. Figures were compared with other national registries17,18 (Table 1).

HCM registry: distribution by regions and by center. Left: participating centers (n=29); top right: distribution of the 1042 patients by regions of Portugal: the Lisbon region included the highest number of patients and the central region of Portugal the lowest; bottom right: note the heterogeneity in terms of patients enrolled per center.'> HCM registry: distribution by regions and by center. Left: participating centers (n=29); top right: distribution of the 1042 patients by regions of Portugal: the Lisbon region included the highest number of patients and the central region of Portugal the lowest; bottom right: note the heterogeneity in terms of patients enrolled per center.' title='Participating centers in the Pro-HCM registry: distribution by regions and by center. Left: participating centers (n=29); top right: distribution of the 1042 patients by regions of Portugal: the Lisbon region included the highest number of patients and the central region of Portugal the lowest; bottom right: note the heterogeneity in terms of patients enrolled per center.'/>

Figure 1.

Participating centers in the Pro-HCM registry: distribution by regions and by center. Left: participating centers (n=29); top right: distribution of the 1042 patients by regions of Portugal: the Lisbon region included the highest number of patients and the central region of Portugal the lowest; bottom right: note the heterogeneity in terms of patients enrolled per center.

(0.31MB).

Almost half of the patients (n=514, 49%) came from the four major centers with specific interest in HCM (the other 25 centers enrolled 528 patients, 51%) (Figure 1). The Lisbon region included the largest number, followed by the North region, the South and Islands, and the Central region. Of the 29 centers, only three included more than 100 patients and eight more than 50 patients. Twenty-one centers included fewer than 50 patients each, 13 centers fewer than 10 and six centers fewer than five.

The patient cohort showed a slight preponderance of males. Mean age at diagnosis was 53±16 years and more than one quarter were diagnosed over the age of 65 years. The disease was classified as familial in one third. At first consultation most patients were symptomatic19 (Table 2).

The ECG was abnormal in 964 individuals (93%). AF was recorded in 117 (11%).

Echocardiographic assessment at enrollment showed that HCM was non-obstructive (instantaneous peak Doppler intraventricular pressure gradient <30 mmHg) in 613 (59% of patients) and obstructive in 365 (35%) (Table 2). Of these, 323 (88%) had obstruction at rest and 42 (12%) had exercise-induced obstruction only, during exercise echocardiography. Obstruction was at the left ventricular outflow tract in 89%. An apical aneurysm was present in 23 patients (2%).

On ambulatory Holter ECG monitoring, AF was present in 118 patients (11%). An exercise test was carried out in less than half of the population and exercise echocardiography in approximately one fifth (Table 2).

CMR was performed in almost half of the cohort. Its incremental value over echocardiography was the assessment of fibrosis (59%), diagnosis in false-negative echocardiograms (6%) and detection of massive LVH (4%).

Based on the American Heart Association model for SCD2,20 (Supplementary Table 1), half of the patients had no risk factors, one third had one risk factor and 15% more than one risk factor. Our data also showed that according to the European Society of Cardiology SCD risk score,1,21 the majority of patients had a five-year risk lower than 4%.

In total, 51% of the patients had undergone genetic testing and in 40% of these a pathogenic/probably pathogenic mutation was found (Table 3). In this group, when the causative gene mutation was reported, the two most frequent genes were MYBPC3 and MYH7.

Most patients (n=909; 87%) received medical treatment (Table 4). Septal reduction therapy was performed in 8% of the cohort, 23% of the obstructive group. Cardiac surgery was performed 2.6 times more frequently than ASA. Surgery was performed in 11 centers (of these, only two performed more than 10 surgeries). ASA was performed in four centers (only one reached 10 procedures).

An ICD was implanted in 13% of the population, mainly for primary prevention. A pacemaker was implanted in 9%, usually for conduction disorders.

Mean follow-up was 5.3±6.1 years, median 3.3 years (IQR [P25-P75] 1.3-6.5 years). At last assessment, most patients were symptomatic (Figure 2), usually with mild to moderate symptoms. A small number (n=42, 4%) developed systolic dysfunction.

Figure 2.

Follow-up data: symptoms at last assessment. At the last assessment most patients were symptomatic (left), and the majority had mild to moderate symptoms (right).

(0.12MB).

All-cause mortality was 6.2% (Table 5). Cardiovascular mortality was 3.4%, most frequently due to HF, followed by SCD and by stroke-related death.

In univariate analysis, 16 of the predefined variables were significantly related to mortality. Multivariate analysis showed four major risk indicators of cardiovascular mortality: late diagnosis (>60 years), family history of SCD, progressive systolic dysfunction and obstructive HCM (Supplementary Table 2).

Of the 12 patients with SCD, seven were between 40 and 65 years old, three were older than 65, and only two were aged under 40 years. In a number of patients SCD was aborted by appropriate ICD shocks in the setting of primary prevention or documented successful in- and/or out-of-hospital resuscitation. Therefore, actual plus aborted SCD occurred in 29 patients.

The incidence of all-cause and cardiovascular mortality was 1.19%/year and 0.65%/year, respectively; the incidence of HF-related death was higher than that of SCD, and the latter was higher than that of stroke. However, the incidence of SCD death plus equivalents was higher than the incidence of HF death, with or without equivalents (Table 5 and Figure 3).

SCD: sudden cardiac death mortality; SCD+Equiv: sudden cardiac death mortality + equivalents.'> SCD: sudden cardiac death mortality; SCD+Equiv: sudden cardiac death mortality + equivalents.' title='Kaplan-Meier estimates of the cumulative hazard function for mortality during follow-up. Left: cumulative hazard function for mortality; right: cumulative hazard function for mortality, including sudden cardiac death and heart failure equivalents. See text for explanation. CV: cardiovascular mortality; HF: heart failure mortality; HF+Equiv: heart failure mortality+ equivalents; Stroke: stroke related mortality; SCD: sudden cardiac death mortality; SCD+Equiv: sudden cardiac death mortality + equivalents.'/>

Figure 3.

Kaplan-Meier estimates of the cumulative hazard function for mortality during follow-up. Left: cumulative hazard function for mortality; right: cumulative hazard function for mortality, including sudden cardiac death and heart failure equivalents. See text for explanation. CV: cardiovascular mortality; HF: heart failure mortality; HF+Equiv: heart failure mortality+ equivalents; Stroke: stroke related mortality; SCD: sudden cardiac death mortality; SCD+Equiv: sudden cardiac death mortality + equivalents.

(0.19MB).

Thromboembolic events occurred in 65 patients (6%) (stroke n=52, TIA n=11, peripheral embolism n=2). Of these, half had documented AF.

Compared with low-volume centers (<15 patients included, n=16), high-volume (>100 patients, n=3) centers had younger patients and more familial HCM and performed more genetic testing, family screening and exclusion of phenocopies (Supplementary Table 3). Additionally, despite the higher number of diagnostic tests and of drug prescriptions of high-volume centers, no major differences in outcomes were found.

The total number of patients included represents about 5% of the estimated prevalence in Portugal,5,7 but up to one third of the Portuguese population with ‘clinical’ HCM.6 Accordingly, this is, to our knowledge, the most comprehensive national HCM registry published.17,18,22 This national effort provides credibility to our data as representative of the real Portuguese scenario. The distribution of patients between referral and community-based centers (four centers included half of the patients and 25 centers the other half) shows that a significant number of patients are followed in non-referral centers. Of note, however, was the low proportion of reported familial HCM, probably reflecting a low rate of systematic family screening programs and/or a referral center bias in another registry.22

Over a decade since the publication of another national registry,17 the clinical spectrum of HCM appears very similar, suggesting that its clinical profile is not undergoing major changes in the Western world. The major difference is the older age at diagnosis, with more than one fourth of patients diagnosed over the age of 65 years. This finding may reflect delayed disease penetrance, lack of systematic family screening, and – potentially – an increased diagnostic yield in older patients.1,2 By contrast, the association found in our cohort of a low rate of familial HCM, later age of presentation, and low risk profile, may more closely mirror the real-world disease scenario, reflecting the inclusion of these unselected lower-risk HCM patients in the cohort. Recent reports have in fact identified a lower-risk cohort of HCM patients, with later onset and lower rate of familial disease,23,24 which may explain our findings.

The proportion of obstructive forms in our cohort, about one third, basically reflects patients with obstruction at rest, and is consistent with the existing literature for resting obstruction.1,2 Accordingly, due to the low rate of exercise echocardiography performed,25 many patients with labile obstruction were probably not detected and were classified as non-obstructive, which at first sight suggests a deviation from the guidelines. However, as the recommendations1,2 for the use of exercise echocardiography in non-obstructive HCM at rest are relatively recent, some of these patients, assessed earlier, have not undergone exercise echocardiography and were diagnosed as non-obstructive in this observational study.

Our data show a relatively limited penetration of CMR, despite the evidence of its incremental value.1–4 These results reflect its high costs, limited availability, and relatively recent introduction in clinical practice.1–4 By contrast, despite the factors that limit the dissemination of genetic testing1,2 (price, lack of co-payment, low availability), half of the patients underwent genetic study, which in many cases is already part of routine practice.26 The proportion of tests in which a variant was found15,16,27,28 and the relative prevalence of the disease-causing genes is mostly similar to what has been described.1,2,15,16,27,28 However, according to the results provided by the investigators, an unexpectedly high prevalence of pathogenic/probably pathogenic mutations15,16,27,28 was found in the TPM1 and CSRP3 genes.17 These results should be interpreted with caution, because they are derived from CRF raw data that were not centrally reviewed or corrected by the PRo-HCM coordinators.

Both of the contemporary models for SCD risk1,2,20,21 show that our cohort was, at baseline, a low-risk population for SCD, which partially explains the low rate of SCD and of ICD implantations.

Invasive septal reduction was offered to almost one fourth of obstructive patients, including those who were mildly symptomatic. Although it cannot be excluded that this rate is biased by the low number of patients detected with labile obstruction, it probably also results from knowledge of the adverse long-term effects of obstruction, as well as from the safety of invasive procedures, which may impact future HCM guidelines.

Of note, the number of surgical myectomies was much higher than the number of ASA procedures, which is partially explained by the late introduction of the latter in Portugal (2009).29 The fact that the two procedures were performed in different centers deserves reflection, taking into account the known effect of expertise on results.1,2

Finally, fewer than 15% of patients received an ICD during follow-up, reflecting the low risk profile of our non-selected population.

Overall, our data suggest that in Portugal, in the era of better diagnostic and therapeutic techniques, HCM has low mortality but high morbidity.

Additionally, despite greater use of diagnostic tests and differences in medical treatment, outcomes of high-volume centers are similar to those of low-volume centers, calling into question the value of HCM centers and of the ‘hub and spoke’ model.7

Outcome data show that the SCD rate in HCM patients in Portugal is very low. Even though this finding may be partially explained by lives saved by successful resuscitation and ICD implantation, the incidence of SCD is still low after including these SCD equivalents in the SCD rate. As a consequence of the efficacy of these preventive measures, HF has become the leading cause of death in HCM patients in Portugal.

Our figures are in overall agreement with those from other groups,30 showing that overall mortality in treated HCM in Portugal is 1.19%, similar to that of the general Portuguese population (around 1.1% year).31 Importantly, at follow-up most patients were symptomatic, confirming that disease morbidity represents a significant burden to patients, health care services and providers. Accordingly, the “contemporary treatable disease”30 has became, at least in Portugal, a “contemporary chronic treatable disease” in which, side by side with ICDs, the role of chronic medical treatment is increasing.

Despite their inherent limitations, registries provide realistic geographical data on disease course and management.

The inclusion of mostly symptomatic patients with advanced, established disease (mainly included by HCM referral centers) is a limitation of this registry, providing a biased view of the disease (selection bias, a common limitation of many HCM studies).

Additionally, disease-related mortality is underestimated, as patients who died before diagnosis were not included. This survival bias partially explains the low rate of events, especially the low rate of SCD.

Children were excluded because of important clinical differences.1,2

The identification, at a national level, of discrepancies between our data and the guidelines is an important finding, warranting a national effort to correct them (for instance to include exercise echocardiography as a standard initial assessment of non-obstructive HCM at rest, to better detect labile obstruction).

Because of the large volume of data, we were unable to cover some important topics in depth. Accordingly, further work will be directed at comparisons between subgroups, addressing family screening, genetic testing (including founder effects, differences in phenotype between genes, and analysis of specific mutations considered as pathogenic/probably pathogenic by the investigators), awareness of phenocopies (for instance Fabry disease), and detailed assessment of clinical HCM profiles.