This study adhered to the principles of the Declaration of Helsinki (2013) and was approved by an Institutional Review Board (IRB) at Seoul National University Hospital (IRB number: H-2403-107-1523). The requirement for written informed consent was waived due to the use of retrospective and anonymized data. The data underlying this article will be made available upon reasonable request to the corresponding authors.

The patient cohort included those diagnosed with HCM between 2007 and 2020 at two tertiary referral centers (Seoul National University Hospital and the Samsung Medical Center). Consecutive patients who met the standard diagnostic criteria for HCM were enrolled in the study. HCM was defined as a maximal wall thickness of ≥15 mm or a wall thickness of ≥13 mm with a family history of HCM.22) In patients with long-standing hypertension, the diagnosis of HCM was further evaluated based on clinical and imaging findings. These included a family history or genetic evidence of HCM, the evaluation of serial echocardiography examinations with information on blood pressure control, morphological patterns of LV hypertrophy, and the presence of systolic anterior motion of the mitral valve.2) We excluded patients younger than 19 years, as well as those with congenital heart disease, significant valvular heart disease (including moderate or greater mitral regurgitation), end-stage renal disease, or infiltrative cardiomyopathy, such as Fabry disease, mitochondrial cardiomyopathy, or cardiac amyloidosis. Patients with poor-quality echocardiographic images inadequate for LV-GLS measurement were also excluded.

Clinical information was collected at the time of HCM diagnosis. Hypertension was defined as a previously confirmed diagnosis meeting the standard diagnostic criteria, with repeated measurements of systolic blood pressure ≥140 mmHg or diastolic blood pressure ≥90 mmHg at rest.23) White-coat hypertension, defined as elevated office blood pressure but normal ambulatory or home blood pressure, was not classified as hypertension. The 5-year SCD risk score was calculated according to the 2014 European Society of Cardiology guidelines.24)

All patients underwent standard echocardiography at the time of HCM diagnosis. The maximum myocardial thickness at end-diastole was recorded. Left atrial volume was calculated using the area-length method and indexed to the body surface area. Septal tissue Doppler velocities were recorded at the medial portion of the mitral annulus. The maximal LV outflow tract gradient was reported as the largest value of pressure gradient measured by continuous wave Doppler, either at rest or during the Valsalva maneuver. An LV apical aneurysm was defined as an akinetic or dyskinetic thin-walled segment of the LV apex. In total, 639 (56%) patients had undergone cardiovascular magnetic resonance imaging at the time of HCM diagnosis. Late gadolinium enhancement was assessed 10–15 min after contrast injection and was defined as areas with a signal intensity exceeding six standard deviations of the remote myocardium. We reported the percentage of late gadolinium enhancement mass divided by the total myocardial mass.

Myocardial strain was measured using the speckle-tracking method with vendor-independent post-processing software (Imaging Arena 4.6; TomTec Imaging Systems, Munich, Germany), as described previously.11)12) LV-GLS was measured at the central core laboratory by experienced technicians who were blinded to patient information and clinical outcomes. LV endocardial borders were manually traced in 2-, 3-, and 4-chamber images at end-systole and tracked throughout the cardiac cycle. LV-GLS was calculated automatically across all segments. LV-GLS measurements were averaged over 3 cardiac cycles in patients with atrial fibrillation. To facilitate straightforward interpretation, we used the absolute value of LV-GLS.

The primary outcome was a CV event, defined as a composite of CV death, HF hospitalization, and ischemic or hemorrhagic stroke. CV death was defined as SCD, death from HF, myocardial infarction, major vascular disease, or stroke. The secondary outcome was SCD-related events, defined as either SCD, aborted SCD, or appropriate implantable cardioverter-defibrillator shock. Data on clinical events were collected based on electronic medical records and official national death records.

Continuous variables are expressed as median values with interquartile ranges, and categorical variables as frequencies with percentages. The Kruskal-Wallis test was used to compare continuous variables between groups, and the χ² test was used to compare categorical variables. The optimal cutoff of LV-GLS for predicting CV events was determined using maximally selected log-rank statistics (Supplementary Figure 1). Based on this analysis, patients were categorized according to an LV-GLS threshold of 13.1% (<13.1% vs. ≥13.1%). Kaplan-Meier survival curves were plotted, and group comparisons were performed using the log-rank test. Cox proportional hazard regression analysis was used to evaluate the association between LV-GLS and outcomes, which was reported as hazard ratios (HRs) and 95% confidence intervals (CIs). The variables included in the multivariable Cox analysis were selected based on their clinical significance and their significance in the univariable Cox analysis: 1) baseline characteristics (age and sex), 2) symptoms and cardiovascular comorbidities (New York Heart Association functional class, hypertension, diabetes mellitus, and atrial fibrillation), and 3) cardiovascular risk factors in the HCM population (syncope and LVEF).

All statistical analyses were performed using R software (version 4.3.0; R Foundation for Statistical Computing, Vienna, Austria), and a 2-tailed value of p<0.05 was considered statistically significant.

Of the 1,139 patients with HCM, 522 (45.8%) had underlying hypertension. The patients with hypertension were significantly older than those without (median 64 vs. 56 years, p<0.001), with a similar proportion of men (73.0% vs. 73.1%, p>0.99) (Table 1). A significantly higher prevalence of cardiovascular comorbidities, including diabetes mellitus, dyslipidemia, atrial fibrillation, and a history of percutaneous coronary intervention, was noted in patients with hypertension than in those without. Patients with hypertension received more frequent medications, including beta-blockers, dihydropyridine calcium channel blockers, renin-angiotensin receptor blockers, and thiazides. With regards to risk factors for SCD, patients with hypertension had a less frequent family history of SCD, and the 5-year SCD risk score was lower in patients with hypertension than in those without (median 1.8% vs. 1.9%, p=0.016).

Patients with hypertension had a significantly larger LV cavity size than those without, although the LVEF was similar between the groups (median 65.0% vs. 65.0%, p=0.619) (Table 1). There was no significant difference in LV maximal wall thickness between patients with and without hypertension (median 18.0 vs. 18.0 mm, p=0.337). Patients with hypertension had larger LA volume indices and higher E/e’ ratios, although the differences were of borderline statistical significance. The percentage of late gadolinium enhancement was higher in patients without hypertension (median 4.5% vs. 3.1%, p=0.005).

The median LV-GLS across the entire cohort was 14.0% (interquartile range, 11.3–17.3%). The distribution of LV-GLS across the hypertension subgroups is shown in Figure 1A and B. LV-GLS was significantly lower in patients with hypertension (median 13.7%) compared to those without (median 14.4%, p=0.001) (Table 1). In a linear regression analysis, hypertension was associated with a significantly lower LV-GLS after adjusting for multiple risk factors (standardized regression coefficient = −0.075±0.030, p=0.013) (Supplementary Table 1).

Clinical and echocardiographic characteristics of two LV-GLS groups (≥13.1% vs. <13.1%) are presented in Supplementary Table 2. Compared to patients with higher LV-GLS, those with lower LV-GLS more frequently had atrial fibrillation and exhibited lower LVEF, greater LV maximal wall thickness, and a larger LA volume index. In both patients with and without hypertension, those with a lower LV-GLS (<13.1%) had a significantly higher maximal wall thickness, lower LVEF, and higher LA volume index compared to those with a higher LV-GLS (≥13.1%) (Figure 1C-E).

Over a median follow-up of 6.6 years (interquartile interval 2.8–9.8 years), 155 CV events occurred, including 56 CV deaths, 74 hospitalizations due to HF, and 59 strokes. The annualized rate of CV events was 22.4 events per 1,000 patient years. Univariable Cox analysis showed that age, female sex, New York Heart Association class ≥III, syncope history, a higher 5-year SCD risk score, hypertension, diabetes mellitus, atrial fibrillation, a lower LVEF, a higher LV mass index, a higher LA dimension or volume index, a higher E/e’ ratio, a higher TR peak velocity, and a lower LV-GLS were significantly associated with an increased risk of CV events (Supplementary Table 3). However, in the multivariable Cox model, hypertension was not independently associated with a higher risk of CV event (p=0.761) (Supplementary Table 4).

CV events occurred more frequently in patients with hypertension than in those without (89 [17.0%] vs. 66 [10.7%], p=0.005) (Figure 2A). When stratified by LV-GLS, patients with a lower LV-GLS (<13.1%) had significantly higher CV events compared to those with a higher LV-GLS (≥13.1%) (80 [17.7%] vs. 75 [10.9%], p<0.001) (Figure 2B).

In the univariable Cox analysis, a lower LV-GLS was significantly associated with a higher risk of CV events in the entire cohort, as well as in each group of patients with hypertension (per 1% decrease in LV-GLS: HR, 1.09; 95% CI, 1.04–1.15; p<0.001) and without hypertension (per 1% decrease in LV-GLS: HR, 1.06; 95% CI, 1.00–1.13; p=0.041) (p-for-interaction=0.503) (Table 2). The 10-year CV events probabilities gradually increased as the LV-GLS decreased in both patients with and without hypertension, with a more pronounced increase observed in those with hypertension (Figure 3). After adjusting for multiple risk factors, the decrease in LV-GLS remained a significant predictor of CV events in patients with hypertension (per 1% decrease in LV-GLS: adjusted HR, 1.07; 95% CI, 1.01–1.13; p=0.013). However, the significance was lost in patients without hypertension, with no significant interaction observed (p-for-interaction=0.566). Similar findings were observed for LV-GLS categorization using a 13.1% cutoff value (Table 2).

When stratified by 4 groups according to hypertension and LV-GLS, CV events most frequently occurred in those with both hypertension and LV-GLS <13.1%, while the lowest rate of CV events occurred in patients without hypertension and LV-GLS ≥13.1% (Figure 4). Cox analysis demonstrated that patients with both hypertension and LV-GLS <13.1% had a significantly higher risk of CV events than those without hypertension and LV-GLS ≥13.1% (adjusted HR, 1.60; 95% CI, 1.01–2.54; p=0.044) (Table 3).

During the follow-up, there were 32 SCD-related events. There was no significant difference in the cumulative incidence of SCD-related events according to hypertension (Supplementary Figure 2A). However, patients with a lower LV-GLS had a significantly higher incidence of SCD-related events than those with a higher LV-GLS (Supplementary Figure 2B). In both hypertensive and non-hypertensive patients, those with a lower LV-GLS had a higher incidence of SCD-related events (Supplementary Figure 3).