Hypertrophic cardiomyopathy (HCM) is a genetically transmitted form of cardiomyopathy with an estimated prevalence of 1/500 inhabitants in the general population.1–6 The disease can have a favorable outcome,7 especially with contemporary management strategies5; however, sudden cardiac death (SCD) remains a risk, and estimation of SCD risk is a rapidly evolving field of research.

The main strategy used to prevent SCD in high-risk HCM patients is the insertion of an implantable cardioverter-defibrillator (ICD).8–10 However, in most HCM patients the implanted ICD is never used. Furthermore, ICD insertion carries a risk of inappropriate shocks and other complications. There is therefore a need for better tools to stratify arrhythmia risk in HCM.

Ventricular arrhythmias leading to SCD in HCM are thought to develop from myocardial fibrosis.11 Some studies have linked ventricular arrhythmias to focal fibrosis, assessed from the presence and extent of late gadolinium enhancement on cardiac magnetic resonance (CMR).12–14 However, this relationship is not considered powerful enough to support ICD implantation as a primary prevention measure in American or European guidelines.15,16 Late gadolinium enhancement-CMR does not detect diffuse fibrosis; however, postmortem histology shows that diffuse fibrosis is more prevalent after SCD in HCM patients than in deaths not linked to a cardiovascular cause or in aths in patients with left ventricular hypertrophy of hypertensive origin, suggesting that it is a proarrhythmic substrate.11,17–20 To date, few studies have been designed to evaluate the association between malignant ventricular arrhythmias and diffuse fibrosis as detected noninvasively.21,22

Myocardial fibrosis is associated with increased extracellular volume (ECV), which can be quantified by CMR or computed tomography (CT).23,24 The aim of our study was to determine whether quantification of ECV by CT, as a surrogate measure of diffuse fibrosis, could distinguish between the presence or absence of malignant ventricular arrhythmias in HCM patients fitted with an ICD (ICD-HCM patients).

A retrospective case-control observational study was performed in ICD-HCM patients. Between November 2013 and February 2015, 78 ICD-HCM patients (> 18 years old without contraindications for contrast CT) were recruited at 5 Spanish cardiomyopathy units (Puerta de Hierro-Majadahonda, Madrid, n = 24; La Fe, Valencia, n = 15; Son Llatzer, Palma de Mallorca, n = 17; Clínico San Carlos, Madrid, n = 10; 12 de Octubre, Madrid, n = 12). The study was approved by the local ethics committees. All patients had been previously implanted with an ICD according to current risk stratification guidelines.16,25,26 All patients gave written informed consent.

Image Analysis

Image analysis was conducted at the CNIC core imaging laboratory by observers blinded to clinical data. Extended Work Station (Philips, Best, The Netherlands) was used to reconstruct 5-mm slices. The slice with the best quality, as defined by the absence of beam-hardening artifacts from the ICD lead, and enough myocardium and blood was selected for further analyses. A region of interest (ROI) was traced in the interventricular septum at the area of maximum myocardial thickness. When focal fibrosis was evident,28 it was included in the ROI with the rest of the septal tissue. Similarly, a ROI was traced inside the left ventricular blood pool (Figure 1). Finally, a lateral ROI was positioned on the lateral wall. Regions of interest were positioned identically for the precontrast and postcontrast acquisitions.

Figure 1.

Two computed tomography images with optimized window settings to highlight contrast between myocardial and blood attenuation are exhibited. In all images, the region of interest is positioned in the septum (white circle), and in blood pool (black circle). Precontrast acquisition image and acquisition 25minutes after initiation of pump infusion are shown.

(0.21MB).

Extracellular volume was estimated after 25minutes of pump infusion using the formula24:

A total of 78 patients were included (24 [30.8%] women, mean age 52 ± 16 years), 24 cases and 54 controls. Demographic characteristics are presented in Table 1 and Table 1 of the supplementary material. Most patients were in a good functional class at the time of inclusion: 71% were in New York Heart Association class I, with just 3% in New York Heart Association classes III-IV. Of the total population, 5% were on cardiac resynchronization therapy and 12% were on diuretics. Median maximal wall thickness was 20.7mm (IQR, 17.1-25.0mm) measured by CT, compared with 22.0mm (IQR, 18.0-28.3mm) on the last echo exam before inclusion. A total of 67% of patients were under treatment with beta-blockers and 13% were under treatment with amiodarone. Excluding the secondary prevention patients, the median 5-year SCD risk according to European Society of Cardiology (ESC) guidelines29 was 3.9% (IQR, 2.9%-6.3%) without statistical differences regarding the appearance of malignant ventricular arrhythmias after ICD implantation.

Within the cases population (N = 24), 14 patients had an ICD implanted for secondary prevention. The remaining 10 patients had an ICD implanted for primary prevention and documented malignant ventricular arrhythmias treated with appropriate ICD therapy during follow-up. Cases with ICD implemented on secondary prevention had any ICD therapy (antitachycardia pacing or shock) earlier than those on primary prevention. Data on ICD therapies in the whole study population are summarized in Table 2 of the supplementary material.

The mean volume of iodine contrast administered to patients was similar in cases and controls (146.7mL vs 138.4mL; P = .147). The median effective radiation dose was the same for both study groups (4.5 mSv vs 4.5 mSv for case patients and controls, P = .295). Attenuation values for each of the compartments are summarized in Table 2, with the noise level indicated by the standard deviation. Additional CT procedure-related data are shown in Table 3 of the supplementary material. The differences in septal ECV in patients with or without atrial fibrillation and type of atrial fibrillation are shown in Table 4 of the supplementary material and Table 5 of the supplementary material, respectively.

Mean ECV in the interventricular septum (primary outcome measure) was 31.3% ± 7.9% in the entire cohort, 29.8% ± 6.3% in cases, and 31.9% ± 8.5% in controls, P = .282 (Table 2, Figure 2). The ECV in the lateral wall was 27.1% ± 7.4%, 24.5% ± 6.8% in cases vs 28.2% ± 7.4% in controls, P = .043. Extracellular volume in the interventricular septum was significantly higher than that calculated in the lateral left ventricular wall: 31.3% ± 7.9% vs 27.1% ± 7.4%, P < .001. The ratio “septal ECV/lateral wall ECV” (a surrogate of left ventricular ECV asymmetry) was nonsignificantly higher in cases patients (1.32 ± 0.11) than in controls (1.17 ± 0.31), P = .133.

Figure 2.

Box plot and dot plot of ECV calculated at 25minutes after pump infusion showed nonsignificant differences between cases and controls. ECV, extracellular volume.

(0.06MB).

For the 64 participants with ICD in primary prevention, ECV was also compared between cases and controls. Extracellular volume was 31.9 ± 8.5% for controls (n = 54) and was 32.6 ± 6.8 (P = .81) for cases (n = 10).

For all parameters (myocardial and blood attenuation both precontrast and postcontrast), intraclass correlation coefficient values were between 0.8 and 0.9 with significant P values in all cases, demonstrating the good reproducibility of the CT data (Table 3).

The entire population was divided into septal and lateral ECV tertiles and ICD therapies were compared among groups. There was no linear trend in in the number of patients receiving appropriate shocks among septal ECV tertiles: 4 (15.1%), 5 (19.2%) and 3 (11.5%) in the lowest, intermediate and highest septal ECV tertiles respectively (P = .703). Conversely, we found a nonsignificant trend among ECV lateral tertiles with a higher incidence of appropriate ICD shocks in patients in the lowest lateral ECV tertile compared with those in the other tertiles (P = .056). A statistically significant linear trend among ECV lateral tertiles was found on comparison of cases vs controls, as previously defined (P = .037) (Table 4).

This is the first study to evaluate the association between diffuse myocardial fibrosis and malignant ventricular arrhythmias in high-risk HCM patients with ICD. We tested the hypothesis that CT-detected ECV, a surrogate for diffuse myocardial fibrosis, would be greater in high-risk HCM patients with malignant ventricular arrhythmic events than in those without. To do this, we examined HCM patients with a previously implanted ICD, dividing the population according to the presence or absence of arrhythmias (either preinsertion, ie, secondary prevention, or postinsertion, ie, appropriate ICD therapy in an individual with an ICD implanted as a primary prevention strategy). Extracellular volume in ICD-HCM patients with malignant ventricular arrhythmias was not increased compared with those without malignant ventricular arrhythmias. From another perspective, when the full cohort was divided by tertiles, we found a trend to a higher incidence of appropriate shocks in patients in the lowest ECV tertile. Our results thus do not support the hypothesis and suggest that quantification of ECV would not improve arrhythmia risk prediction in HCM patients.

Myocardial fibrosis, with increased myocardial collagen matrix deposition, is a recognizd phenomenon during the natural history of HCM, detected during autopsy of young victims of HCM-related sudden death.11 In HCM patients, increased collagen synthesis can be tracked from increased circulating levels of type I procollagen C-terminal propeptide.30 The advent of late gadolinium enhancement-CMR allowed the study of macroscopic fibrosis and led to suggestions of an association with lethal arrhythmic events in HCM patients. A study of 1293 HCM patients found the extent of late gadolinium enhancement to be a good predictor of SCD, showing a 40% increase in relative SCD risk for every 10% increase in late gadolinium enhancement over a median follow-up of 3.3 years.12 However, a meta-analysis including 1063 patients from 4 studies found the association between late gadolinium enhancement and sudden death to be nonsignificant over a mean follow-up of 3.1 years31; moreover, an analysis of 711 HCM patients similarly found no statistically significant association between the presence and quantity of fibrosis (late gadolinium enhancement + areas on CMR) and SCD over a median follow-up of 3.5 years.13 Thus, while there is a plausible link between fibrosis and arrhythmia risk, there is no definite association between focal macroscopic fibrosis detected by late gadolinium enhancement-CMR and malignant ventricular arrhythmias or SCD. Reflecting this situation, risk stratification in current guidelines for ICD implantation does not include the presence of late gadolinium enhancement only as a minor criterion, considering it potentially useful in patient selection,16 as it was included in the previous guidelines as a minor criteria.15 The recent advent of imaging techniques able to quantify diffuse microscopic fibrosis23 has allowed reassessment of the hypothesis that fibrosis is a risk marker in HCM patients.

In our population, 64 patients had an ICD implanted for primary prevention, 10 of whom (15.6%) had ≥ 1 appropriate therapy over a median follow-up of 4.8 years (3.3%/y). This is similar to the published figure of 3.6% per year,32 and clearly higher than the calculated median risk of 3.9% per 5 years (0.8%/y) in the ESC guidelines algorithm.29 However, it is important to mention that not all appropriate therapies would have been lifesaving treatments, because it is well known that some ventricular arrhythmias are self-terminating without any ICD intervention, and thus the percentage of appropriate ICD therapies is not directly equivalent to the risk of sudden death. However, there is a striking difference observed between the estimated risk of sudden death with the ESC algorithm and the much higher percentage of appropriate ICD therapies. In this regard, missing information should be considered. In 21.8% of our population, risk score was not calculated due to a lack of any of the necessary data. The percentage of appropriate therapies in our primary prevention population included in the cases group (3.3%/y) was slightly higher than the established risk of sudden death (6% 5-year risk, 1.2%/y) for ICD implantation according to current ESC guidelines.16 Of the 14 patients in our study who had an ICD implanted for secondary prevention, 8 (57.1%) had ≥ 1 documented appropriate therapy event over a mean follow-up of 5.0 years (11.4%/y), similar to the 10.6%/y recorded in other cohorts.32

As measured by CT in equilibrium, ECV in HCM patients with documented malignant ventricular arrhythmias is not increased compared with HCM patients without malignant ventricular arrhythmias. Longer follow-up studies are warranted to investigate the potential of ECV to improve risk prediction of malignant ventricular arrhythmias in HCM.

This work was funded by the RIC (Red de Investigación Cardiovascular) of the Spanish Ministry of Health, familial cardiomyopathy program (RD 12/0042/0054 to B. Ibáñez; RD 12/0042/066 to P. García-Pavía; RD 12/0042/0069 to T. Ripoll-Vera; RD12/0042/0036, RD06/0003/0009 to REDINSCOR [Red Española de Insuficiencia Cardiaca]). This work was supported by the Plan Estatal de I+D+I 2013-2016 – ERDF (European Regional Development Fund) “A way of making Europe”. This study forms part of a MRA (Master Research Agreement) between CNIC and Philips Healthcare. The CNIC is supported by the Spanish Ministry of Economy and Competitiveness and the Pro-CNIC Foundation and is a Severo Ochoa Center of Excellence (MINECO award SEV-2015-0505).

J. Sánchez-González is Philips employee.