Women with hypertrophic cardiomyopathy have worse survival

Abstract

Aims

Sex differences in hypertrophic cardiomyopathy (HCM) remain unclear. We sought to characterize sex differences in a large HCM referral centre population.

Methods and results

Three thousand six hundred and seventy-three adult patients with HCM underwent evaluation between January 1975 and September 2012 with 1661 (45.2%) female. Kaplan–Meier survival curves were assessed via log-rank test. Cox proportional hazard regression analyses evaluated the relation of sex with survival. At index visit, women were older (59 ± 16 vs. 52 ± 15 years, P < 0.0001) had more symptoms [New York Heart Association (NYHA) Class III–IV 45.0% vs. 35.3%, P < 0.0001], more obstructive physiology (77.4% vs. 71.8%, P = 0.0001), more mitral regurgitation (moderate or greater in 56.1% vs. 43.9%, P < 0.0001), higher E/e′ ratio (n = 1649, 20.6 vs. 15.6, P < 0.0001), higher estimated pulmonary artery systolic pressure (n = 1783, 40.8 ± 15.4 vs. 34.8 ± 10.8 mmHg, P < 0.0001), worse cardiopulmonary exercise performance (n = 1267; percent VO₂ predicted 62.8 ± 20% vs. 65.8 ± 19.2%, P = 0.007), and underwent more frequent alcohol septal ablation (4.9% vs. 3.0%, P = 0.004) but similar frequency of myectomy (28% vs. 30%, P = 0.24). Median follow-up was 10.9 (IQR 7.4–16.2) years. Kaplan–Meier analysis demonstrated lower survival in women compared with men (P < 0.0001). In multivariable modelling, female sex remained independently associated with mortality (HR 1.13 [1.03–1.22], P = 0.01) when adjusted for age, NYHA Class III–IV symptoms, and cardiovascular comorbidities.

Conclusion

Women with HCM present at more advanced age, with more symptoms, worse cardiopulmonary exercise tolerance, and different haemodynamics than men. Sex is an important determinant in HCM management as women with HCM have worse survival. Women may require more aggressive diagnostic and therapeutic approaches.

Introduction

Patients with hypertrophic cardiomyopathy (HCM) are characterized by inherited cardiac hypertrophy, as opposed to hypertrophy secondary to another pathologic process.^1–3 On an organ level, there is diversity in the extent of myocardial hypertrophy, amount of left ventricular (LV) outflow tract obstruction and degree of diastolic dysfunction. The large phenotypic spectrum accounts for differences in clinical course including asymptomatic with normal lifespan, sudden cardiac death (SCD), and medically refractory heart failure symptoms requiring septal reduction therapy.

Key differences between men and women have been explored in a variety of cardiac conditions spanning heart failure, valvular heart disease, coronary artery disease, dilated cardiomyopathy, and others.^4–17 Sex differences in HCM are less well understood. While there have been studies describing demographic and clinical differences in HCM based upon sex, data assessing survival are limited and previously have shown no differences.^18,19 However, it is important to clarify sex differences in survival outcomes in order to better address diagnostic and treatment strategies for patients with this condition. We hypothesized that study of a large population of patients with HCM would demonstrate clinical and haemodynamic differences between men and women, and that these differences would affect survival. Therefore, we sought to characterize sex differences in symptoms, phenotypes, treatments, and outcomes in a large, single-centre HCM referral population.

Methods

Patient selection

The studied population comprised adult patients with HCM who underwent index evaluation at Mayo Clinic (Rochester, MN, USA) between January 1975 and September 2012. The diagnosis of HCM was determined both clinically and by echocardiography, on the basis of the presence of myocardial hypertrophy in the absence of local or systemic oetiologies.^1,20,21 Informed consent was obtained for use of patient medical records for research purposes in accordance with Minnesota law and an Institutional Review Board approved study.

The diagnosis of atrial fibrillation (AF) was based on a prior history of paroxysmal or chronic AF or demonstration of AF on index electrocardiogram. It was noted if patients had a history of systemic hypertension, although clinical evaluation in all cases determined that hypertension did not obviate the diagnosis of HCM. Survival data were collected from the electronic health records and the national death and location database (Accurint, Lexisnexis for Mayo Clinic patients). Assuming no lag in data, a date of last follow-up of January 2017 was used as the last follow-up in those patients still assumed to be alive. Plasma B-type natriuretic peptide (BNP) assessed via fluorescence immunoassay (Biosite Diagnostics, San Diego, CA, USA)²² and plasma N-terminal pro BNP (NT-Pro BNP) measured with electrochemiluminescence immunoassay (Roche Diagnostics, Indianapolis, IN, USA)²³ were obtained when clinically indicated. Symptom-limited graded exercise testing was performed using an institutionally designed incremental exercise testing protocol as described previously.^24,25 Continuous electrocardiographic monitoring (Marquette Electronics, Milwaukee, WI, USA) and breath-to-breath metabolic measurement (Medical Graphics, St. Paul, MN, USA) were utilized as previously described.²⁶

Echocardiographic evaluation

The presence of a transthoracic echocardiographic examination constituted an inclusion criterion. Left ventricular cavity size, LV mass, wall thickness, and ejection fraction were determined in standard fashion.²⁷ Left atrial volume index was calculated by either the prolate ellipse or biplane area-length methods,²⁸ with left atrial enlargement defined as a left atrial volume index ≥34 cm³/m². Mitral regurgitation was graded as none to mild, moderate, moderately severe or severe after analysing jet area and width, spectral Doppler intensity, as well as regurgitation quantitation with the continuity equation and/or PISA method as appropriate.²⁹ Not all mitral regurgitation could be quantitated given jet eccentricity and LV outflow tract flow turbulence merging with the regurgitant jet flow convergence. Mitral inflow velocity curves [peak early diastolic velocity (E) and peak late diastolic velocity (A)] and annular tissue Doppler signals (medial e′ and lateral e′) were obtained as previously described.³⁰ Right ventricular systolic pressure (RVSP) was estimated in a standard fashion per ASE guidelines.³¹ Right ventricular systolic pressure cutoffs of 35 and 50 mmHg for assessment of pulmonary hypertension (PH) were assessed, as previously.³² LV outflow tract gradients were obtained by continuous-wave Doppler interrogation of the LV outflow tract from an apical window and calculated using the modified Bernoulli equation (i.e. gradient = 4v_LVOT², where v_LVOT is peak LV outflow tract velocity). Resting obstruction was defined as a LV outflow tract gradient of ≥30 mmHg. The presence of labile obstruction (provoked gradient ≥50 mmHg) was determined by provocative manoeuvres (Valsalva manoeuvre, amyl nitrite inhalation, and exercise). Because of the long timespan of this study, some patients were evaluated prior to the availability of certain echocardiographic parameters (mainly Doppler-based), which were unavailable for incorporation in this analysis.

Data analysis

Normality of distribution was determined via histogram visual assessment. Continuous variables are expressed as mean ± standard deviation ( SD) when normally distributed; non-normally distributed variables are reported as median with IQR. Comparison of sex differences in variables were performed using Student’s t-tests, Fischer exact tests, and Wilcoxon rank sum tests as appropriate. All-cause mortality was the endpoint of interest in this analysis. Survival was assessed using Kaplan–Meier survival curves, with log rank test P-values reported. Survival was also compared with expected survival determined via an age- and sex-matched population derived from US Census data.

Univariate and multivariable Cox proportional hazard regression analyses were performed to determine the relation of sex with study endpoints, with hazard ratios (HR) and 95% confidence intervals (CI) reported. To assess for the influence of date of referral on survival, visit date was added to the multivariable model and assessed as both a linear and non-linear variable. Logistic regression analysis was used to construct a propensity score, which was used as an adjustment variable in the overall survival models. Using sex as the x variable in the model, additional clinical variables were included. The score from this model was used as a potential adjusting variable in the multivariable Cox proportional hazards model with sex.

Any missing data were handled by omission from the respective analyses. Statistical significance was set a priori at P < 0.05. All analyses were performed using JMP version 9.0.2 and SAS version 9.4 (SAS Institute Inc., Cary, NC, USA).

Results

Study population

A total of 3673 patients were included in the study, with male predominance [2012 men (55%), P < 0.0001 compared with the expected sex distribution of 50% given an autosomal dominant inheritance pattern]. Eight hundred and sixty-eight (23.6%) patients underwent index evaluation at our institution before 1990, 936 (25.5%) patients in the period 1990 to 2000, and 1869 (50.9%) patients after 2000. The average age was 55 ± 16 years. Patient demographics are shown in Table 1. This was primarily a referral population since only 197 of 3673 (5.4%) patients resided in Olmsted County or one of the nearby counties with 559XX zip codes (Mower, Steele, Houston, and Dodge). Reflecting the referral composition of our practice, 40% of patients presented with New York Heart Association (NYHA) Class III–IV symptoms. Compared with men, women were older at index evaluation (59 ± 16 vs. 52 ± 15 years, P < 0.0001), more likely to carry a prior diagnosis of systemic hypertension (49.1% vs. 43.4%, P = 0.0010), and were more symptomatic (NYHA Class III–IV symptoms in 45.0% vs. 35.3%, P < 0.0001). When assessed by age quartiles, systemic hypertension was more common in men in the first age quartile (P = 0.002) and women in the last quartile (P = 0.01). There were no age-dependent trends in other comorbidities across age quartiles. Women were more likely to present NYHA Class III–IV symptoms across all age quartiles (P ≤ 0.01 for all). Women were slightly less likely to be taking beta receptor antagonists but were more likely to be utilizing calcium channel antagonists (Table 1).

Table 1.

Baseline characteristics

Variables	Women n = 1661 (45%)	Men n = 2012 (55%)	P-value	All patients n = 3673
Age (years)	59 ± 16	52 ± 15	<0.0001	55 ± 16
Referral patient, n (%)	1560 (94)	1916 (95)	0.09	3476 (95)
Systolic blood pressure (mmHg)	123 ± 21	124 ± 17	0.40	123 ± 19
Diastolic blood pressure (mmHg)	70 ± 11	73 ± 11	<0.0001	72 ± 11
Heart rate (beats/min)	68 ± 13	64 ± 12	<0.0001	66 ± 12
Body mass index (kg/m2)	28.3 ± 6.7	28.9 ± 5.0	0.0053	28.6 ± 5.8
Body surface area (m2)	1.78 ± 0.21	2.10 ± 0.21	<0.0001	1.95 ± 0.26
NYHA Class III–IV, n (%)	748 (45)	710 (35)	<0.0001	1458 (40)
Coronary artery disease, n (%)	271 (16)	336 (17)	0.89	607 (17)
Systemic hypertension, n (%)	816 (49)	874 (43)	0.0010	1690 (46)
Syncope, n (%)	236 (14)	309 (15)	0.24	545 (15)
Atrial fibrillation, n (%)	283 (17)	367 (18)	0.36	650 (18)
Non-sustained ventricular tachycardia, n (%)a	41 (21)	74 (25)	0.33	115 (23)
ICD, n (%)	99 (6)	149 (7)	0.086	248 (7)
Resuscitated SCD, n (%)	7 (0.4)	10 (0.5)	0.63	17 (0.5)
Family history of HCM, n (%)	392 (24)	422 (21)	0.085	814 (22)
Family history of SCD, n (%)	242 (15)	302 (15)	0.64	544 (15)
BNP (pg/mL)b	250 (IQR 115–526)	126 (IQR 50–304)	<0.0001	173 (IQR 71–383)
NT-Pro BNP (pg/mL)c	1022 (IQR 372–1796)	476 (IQR 237–994)	<0.0001	649 (IQR 276–1340)
VO2 (mL/kg/min)d	16.8 ± 5.2	22.2 ± 6.7	<0.0001	20.1 ± 6.7
Percentage VO2 predicted (%)d	63 ± 20	66 ± 19	0.0073	65 ± 20
Medications, n (%)
Beta receptor antagonist	1088 (66)	1370 (68)	0.042	2458 (67)
Calcium channel blocker	664 (40)	706 (35)	0.0031	1370 (37)
ACE-inhibitor/ARB	169 (10)	251 (12)	0.76	420 (11)
Disopyramide	146 (9)	156 (8)	0.27	302 (8)
Amiodarone	90 (5)	104 (5)	0.77	194 (5)
Future alcohol septal ablation	81 (5)	61 (3)	0.0044	142 (4)
Future septal myectomy	497 (30)	566 (28)	0.24	1063 (29)
Future septal reduction therapy	572 (34)	619 (31)	0.02	1191 (32)

Continuous variables expressed as mean ± SD or median and IQR.

ACE, angiotensin converting enzyme; ARB, angiotensin receptor blocker; BNP, B-type natriuretic peptide; HCM, hypertrophic cardiomyopathy; ICD, internal cardiac defibrillator; NT-Pro BNP, Amino-terminal pro B-type natriuretic peptide; SCD, sudden cardiac death; NYHA, New York Heart Association; IQR, Interquartile range; SD, standard deviation.

^aHolter data available in n = 491 (13%).

^bBNP assessed in n = 763 (21%).

^cNT-Pro BNP assessed in n = 294 (8%).

^dCardiopulmonary exercise testing performed in n = 1267 (35%).

Natriuretic peptide assessment was performed in a minority of patients (29%). Plasma B-type natriuretic peptide was higher in women compared with men [n = 763, 250 (IQR 115–526) vs. 126 (IQR 50–304) pg/mL, P < 0.0001] as was NT-Pro BNP [n = 294, 1022 (IQR 372–1796) vs. 476 (IQR 237–994) pg/mL, P < 0.0001]. It should be acknowledged, however, that BNP and NT-pro BNP tend to be higher in women and with increasing age, even in the absence of cardiac disease.^22,33 In patients with clinically driven cardiopulmonary exercise testing, women had worse exercise capacity, both with regards to VO₂ (n = 1267, 16.8 ± 5.2 vs. 22.2 ± 6.7 mL/kg/min, P < 0.0001) and percent VO₂ predicted (63 ± 20 vs. 66 ± 19%, P = 0.0031). Men were more likely to undergo cardiopulmonary exercise assessment compared with women (40.5% vs. 31.7%, P < 0.0001).

Echocardiographic assessment

Index echocardiographic evaluation is shown in Table 2. The median LV outflow tract gradient at rest was 29 (IQR 8–70) mmHg. There was evidence of resting obstruction in 1830 patients (50%), with an additional 900 patients (25%) demonstrating labile obstruction. Women had a higher LV outflow tract gradient when compared with men [36 (IQR 10–81) vs. 23 (IQR 0–61) mmHg] and were more likely to demonstrate resting obstruction when compared with men (54% vs. 46%, P < 0.0001). Consistent with these results, women had more mitral regurgitation (moderate or greater in 28% vs. 18%, P < 0.0001).

Table 2.

Pre-septal reduction echocardiographic characteristics

Variables	Women n = 1661 (45%)	Men n = 2012 (55%)	P-value	All patients n = 3673
LV outflow tract gradient (mmHg)	36 (IQR 10–81)	23 (IQR 0–61)	<0.0001	29 (IQR 8–70)
Resting obstruction, n (%)	905 (54)	925 (46)	<0.0001	1830 (50)
Labile obstruction, n (%)	380 (23)	520 (26)	0.038	900 (25)
Non-obstructive, n (%)	376 (23)	567 (28)	0.0001	943 (25)
LV ejection fraction (%)	71 ± 8	69 ± 9	<0.0001	70 ± 9
LV end diastolic dimension (mm)	43 ± 6	47 ± 6	<0.0001	45 ± 6
Anteroseptal wall thickness (mm)	17.8 ± 5.5	18.3 ± 5.5	0.016	18 ± 6
Wall thickness >30 mm, n (%)	41 (2)	50 (2)	0.83	91 (2)
Posterior wall thickness (mm)	12.6 ± 3.1	13.2 ± 3.1	<0.0001	13 ± 3
LV mass (g)	265 ± 101	319 ± 110	<0.0001	295 ± 109
LV mass index (g/m2)	150 ± 57	152 ± 51	0.16	151 ± 54
Moderate or greater mitral regurgitation, n (%)a	329 (28)	257 (18)	<0.0001	586 (22)
Mitral E velocity (cm/s)b	0.9 ± 0.3	0.8 ± 0.3	<0.0001	0.9 ± 0.3
Mitral A velocity (cm/s)b	0.9 ± 0.4	0.7 ± 0.3	<0.0001	0.8 ± 0.3
E/A ratiob	1.2 ± 0.8	1.4 ± 0.7	0.0002	1.3 ± 0.7
Medial e’ (cm/s)c	0.05 ± 0.02	0.06 ± 0.02	<0.0001	0.05 ± 0.02
Lateral e’ (cm/s)c	0.06 ± 0.03	0.07 ± 0.03	<0.0001	0.07 ± 0.03
Medial E/e’ ratioc	21 ± 10	16 ± 7	<0.0001	17.7 ± 8.4
Lateral E/e’ ratioc	16 ± 8	12 ± 6	<0.0001	13.9 ± 7.3
Left atrial volume index (cm3/m2)d	48 ± 18	49 ± 25	0.23	48 ± 23
Left atrial enlargement, n (%)d	537 (76)	733 (73)	0.18	1270 (75)
Estimated RVSP (mmHg)e	36 (IQR 30–48)	32 (IQR 28–39)	<0.0001	34 (IQR 28–43)
RVSP ≥35 mmHg, n (%)e	472 (56)	364 (39)	<0.0001	836 (47)
RVSP ≥50 mmHg, n (%)e	175 (21)	87 (9)	<0.0001	262 (15)

Continuous variables expressed as mean ± SD or median and IQR.

LV, left ventricular; RVSP, pulmonary artery systolic pressure; IQR, Interquartile range; SD, standard deviation; E, peak early diastolic velocity; A, peak late diastolic velocity.

^aSemi-quantitative assessment of mitral regurgitation severity available in n = 2606 (71%).

^bMitral inflow assessment available in n = 2757 (75%).

^cMitral annular tissue Doppler assessment available in n = 1671 (46%).

^dLeft atrial volume index available in n = 1706 (46%).

^eRVSP available in n = 1783 (49%).

While measurement of maximal LV wall thickness was statistically different between women and men, there was little practical difference (17.8 ± 5.5 vs. 18.3 ± 5.5 mm, P = 0.016) and no difference in indexed LV mass (150 ± 57 vs. 152 ± 51, P = 0.16). Owing to evolution of echocardiographic diastolic function assessment over time, mitral inflow, tissue Doppler, and left atrial volume assessments were available in many but not all patients (Table 2). Both medial and lateral E/e′ ratios were higher in women (21 ± 10 vs. 16 ± 7 and 16 ± 8 vs. 12 ± 6, respectively, P < 0.0001 for both). In patients with RVSP assessment at echocardiography (n = 1783, 49%), women demonstrated higher estimated RVSP (36 [IQR 30–48] vs. 32 [IQR 28–39] mmHg, P < 0.0001). Women more frequently exceeded RVSP thresholds of 35 and 50 mmHg compared with men (56 vs. 39% and 21 vs. 9%, respectively, P < 0.0001 for both). There was no difference in indexed left atrial volume or presence of left atrial enlargement.

Septal reduction therapy

Myectomy was performed in 1063 patients (29%) at median of 6 (IQR 2–49) days after index evaluation. Although patients who underwent myectomy were younger than those who did not (51.1 ± 14.9 vs. 57.2 ± 15.9 years, P < 0.0001), women (n = 497, 29.9% of all women) and men (n = 566, 28.9% of all men) were equally likely to undergo myectomy (P = 0.24). Women who underwent myectomy were older (53 ± 16 vs. 49 ± 14 years), had higher estimated RVSP (43 ± 16 vs. 35 ± 11 mmHg) and higher resting gradient (67 ± 45 vs. 52 ± 38 mmHg) compared with men who underwent myectomy, but did not differ with regards to septal thickness or provoked gradient. Alcohol septal ablation was performed with greater frequency in women (n = 81, 4.9% of all women) compared with men (n = 61, 3.0% of all men; P = 0.0044), although overall it was utilized less frequently than myectomy. Patients undergoing alcohol septal ablation were older than those undergoing myectomy (62 ± 14 vs. 51 ± 15 years). Women who underwent alcohol septal ablation were significantly older compared with men (66 ± 12 vs. 55 ± 12 years, P < 0.0001) and had higher estimated RVSP (46 ± 22 vs. 34 ± 13 mmHg) but did not differ with regards to septal thickness, resting gradient, or provoked gradient. Fourteen patients (0.4%) underwent both alcohol septal ablation and myectomy. In looking at all septal reduction therapies following referral, women were more likely to undergo intervention (34.4% vs. 30.8%, P = 0.02).

Survival analysis

Median follow-up was 10.9 (IQR 7.4–16.2) years (mean 12.7 ± 8.1 years), encompassing 46 647 patient-years, with 3569 patients (97%) followed beyond 1 year. During follow-up, 1477 deaths occurred. Median survival in patients who died was 9.0 (IQR 4.3–15.0) years after index evaluation. Women had worse survival than men (Figure 1, P < 0.0001). Five- and 10-year survival estimates of women vs. men were 85.1% (CI 83.4%–86.8%) vs. 91.1% (CI 89.9%–92.4%) and 70.8% (CI 68.6%–73.1%) vs. 82.0% (CI 80.3%–83.8%), respectively.

Figure 1.

Survival in all patients stratified by sex. Kaplan–Meier survival curves grouped by sex demonstrate worse survival in women (red) as opposed to men (blue).

When stratified by subjective criteria of symptoms (NYHA class at time of referral), women had worse survival both when more symptomatic (NYHA Class III–IV, n = 1458, P < 0.0001) and less symptomatic (NYHA Class I–II, P < 0.0001). When stratified by objective cardiopulmonary exercise performance (n = 1176, using a cutoff of 75% VO2 predicted), there was no difference in survival between men and women in either group, although only n = 224 deaths were noted within this cohort. Women had worse survival than men following septal reduction therapy (P < 0.0001).

Comparison of HCM patients with an age and sex-matched US population (Take home figure) demonstrated worse survival than expected for both men and women (P = 0.005). Five-year survival estimates demonstrated a difference between expected and observed survival of 5.4% in women (90.5% vs. 85.1%), as compared with 0.9% in men (92.0% vs. 91.1%); 10-year survival estimates revealed a difference of 8.1% in women (78.9% vs. 70.8%), as compared with 0.7% in men (82.7% vs. 82.0%).

Take home figure 2.

Survival comparison of patients with hypertrophic cardiomyopathy to an age- and sex-matched US population. Kaplan–Meier assessment of men with hypertrophic cardiomyopathy (solid blue line) compared with an age- and sex-matched US population (dashed blue line). Women with hypertrophic cardiomyopathy (solid red line) compared with an age- and sex-matched US population (dashed red line). HCM, hypertrophic cardiomyopathy.

Hazard modelling

Female sex was a strong univariate predictor of mortality (Table 3). Age, NYHA Class III–IV symptoms, history of AF, coronary artery disease, and systemic hypertension were also predictors of mortality, with ICD implantation and beta receptor antagonist use predictors of lower mortality. Echocardiographic predictors of mortality were LV outflow tract gradient and moderate or greater mitral regurgitation. Female sex remained independently associated with mortality using multivariable modelling inclusive of all statistically significant univariate predictors other than echocardiographic variables: age, NYHA Class III/IV symptoms, and history of AF, coronary artery disease, hypertension, ICD implantation, and beta receptor antagonist use (Table 4). Echocardiographic variables were excluded from main multivariable analysis given reduction of sample size. However, the inclusion of LV outflow tract gradient, LV ejection fraction, and moderate or greater mitral regurgitation to multivariable modelling did not significantly alter the relationship of female sex to mortality. When a propensity score including age, NYHA Class III–IV, ICD implantation, medical therapies, history of AF, coronary artery disease, and hypertension was used to adjust the multivariable Cox proportional hazards model, female sex continued to be associated with mortality; HR 1.17 (95% CI 1.07–1.25), P = 0.0009.

Table 3.

Univariate predictors of mortality

Variables	HR	95% CI	P-value
Female sex	1.40	1.34–1.46	<0.0001
Age (per year)	1.05	1.05–1.06	<0.0001
NYHA Class III–IV	1.14	1.02–1.27	0.02
Atrial fibrillation	1.79	1.59–2.03	<0.0001
Coronary artery disease	1.26	1.09–1.44	0.0001
Hypertension	1.31	1.19–1.46	<0.0001
ICD implantation	0.57	0.41–0.79	0.0006
Beta receptor antagonist use	0.75	0.68–0.83	<0.0001
Calcium channel antagonist use	0.94	0.85–1.05	0.3
Resting LV outflow tract gradient (per mmHg)	1.002	1.000–1.003	0.04
Obstructive physiology	0.98	0.87–1.10	0.7
Moderate or greater mitral regurgitation	1.35	1.15–1.59	0.0003
Maximal wall thickness (per mm)	0.99	0.98–1.00	0.2
LV ejection fraction (per %)	0.99	0.99–1.00	0.04
Estimated RVSP (per mmHg)	1.03	1.03–1.04	<0.0001

Univariate predictors of all-cause mortality in a large referral HCM population.

HR, hazard ratio; ICD, internal cardioverter-defibrillator; LV, left ventricular; RVSP, right ventricular systolic pressure; HCM, hypertrophic cardiomyopathy; NYHA, New York Heart Association; CI, confidence interval.

Table 4.

Multivariable predictors of mortality

Variables	HR	95% CI	P-value
Female sex	1.13	1.03–1.22	0.01
Age (per year)	1.05	1.05–1.06	<0.0001
NYHA Class III–IV	1.21	1.08–1.35	0.001
Atrial fibrillation	1.57	1.38–1.78	<0.0001
Coronary artery disease	0.95	0.83–1.10	0.5
Hypertension	0.92	0.83–1.02	0.1
ICD implantation	0.93	0.67–1.30	0.7
Beta receptor antagonist use	0.83	0.74–0.92	0.0007

Multivariable predictors of all-cause mortality.

HR, hazard ratio; ICD, internal cardioverter-defibrillator; NYHA, New York Heart Association; CI, confidence interval.

To assess for the influence of date of referral on survival, visit date as a continuous linear variable was added to the multivariable model shown in Table 4. When adjusted for referral date, female sex remained an independent predictor of mortality [HR 1.13 (1.03–1.22), P = 0.01]. Utilizing a non-linear time adjustment did not alter statistical significance.

In patients with NYHA Class I–II symptoms, the HR for female sex was 1.72 (95% CI 1.52–1.96, P < 0.0001) and after adjustment for age was still borderline significant (HR 1.15, 95% CI 1.00–1.31, P = 0.05), suggesting that age differences within this cohort may account for differences in survival.

Discussion

This large study of patients with HCM from a single tertiary referral centre helps to clarify sex differences in HCM and demonstrates that women with HCM have worse all-cause mortality when compared with men. Women with HCM were referred at an older age with more symptoms, more LV outflow tract obstruction, more PH, higher natriuretic peptides, and worse measures of cardiopulmonary exercise tolerance compared with men. Compared to an age and sex-matched US population, women with HCM had worse relative survival than men. Sex remained predictive of mortality in multivariable modelling.

Male predominance was seen within the study population. Skewing of study populations towards men is a consistent finding amongst studies of HCM.^18,34–40 Whether this finding represents decreased disease penetrance in women or reflects clinical recognition/referral practices remains unclear.

Despite similar degrees of hypertrophy, women were more likely to demonstrate obstructive physiology, a finding also consistent with prior studies.^18,41,42 This may have contributed to symptomatic differences at presentation, although symptoms may also have been influenced by abnormalities of diastolic function. Women had higher E/e′ ratios, which is known to correlate modestly with higher filling pressures in HCM.⁴³ Women demonstrated more PH, a finding noted in general populations^44,45 and in HCM patients undergoing myectomy,³² which may reflect elevated filling pressures, more obstruction, or more reactive pulmonary vasculature. Accordingly, women were more symptomatic in the present study, both on the basis of NYHA class and objective measures of cardiopulmonary exercise tolerance, consistent with other recently reported findings.⁴⁶

Why do women present with such a different phenotype? This phenomenon may relate to the genetic and endocrine differences between males and females. In experimental models of hypertrophy⁴⁷ and HCM,⁴⁸ there are differences in gene expression and associated phenotype between males and females. There is also evidence in humans and experimental animals that sex hormones may modulate the disease phenotype in HCM.^49,50 We have previously studied sex differences in patterns of hypertrophy in HCM, albeit in a much smaller cohort, with no significant differences demonstrated.⁴¹

Beyond sex differences, gender differences, referring to the psychosocial context in which men and women interface with the healthcare system, are critical.^17,51 Women are less aware of their risk of cardiovascular disease and frequently do not have cardiovascular disease discussed with them by their physician.⁵² In addition, populations of young athletes subjected to cardiovascular disease screening are predominantly male.^53–55 It is plausible that women with HCM are under-recognized and their diagnosis delayed until an advanced phenotype is present. This may be in part because of a lack of patient and physician suspicion for the condition and in part because of decreased incidental diagnosis through cardiovascular screening tests.

Previous studies have shown that women with HCM have more major cardiovascular events⁵⁶ and greater progression to NYHA Class III/IV symptoms,¹⁸ but unlike this study have not shown a sex discrepancy in survival.^18,19 Our study demonstrates that women with HCM have worse survival than men, a relationship which remained significant in multivariable analyses and propensity score analyses. We believe that these results are not discordant with prior studies, but rather are a clarification of the relationship between sex and survival in a larger, sicker cohort. In comparison to the largest prior study with focus on sex differences within HCM, our population was nearly four times greater in size, included over eight times as many deaths, demonstrated more obstructive physiology, and had a much larger proportion of NYHA Class III or IV patients.¹⁸ Our findings are in accord with a prior study of patients undergoing surgical myectomy which demonstrated that female sex was an independent predictor of mortality in long-term post-operative follow-up.⁵⁷

Women are less likely to be diagnosed with HCM at routine medical examination.¹⁸ Furthermore, when diagnosed, women are more likely to carry a diagnosis of system hypertension,^41,58 a finding reflected in our data. Despite this, objective measures of systolic blood pressure did not differ at the time of referral evaluation. This may suggest that women with HCM have LV hypertrophy misattributed to systemic hypertension preceding determination of their true underlying pathology. Supporting a hypothesis of misdiagnosis, women with HCM are more likely to be on non-cardiac medications as compared with men,⁵⁹ which could imply treatment of alternative causes of non-specific symptoms such as dyspnoea. Consistent with prior studies, women were older than men at the time of referral.^60,61 Furthermore, women with HCM have been repeatedly shown to be more symptomatic than men,^18,59 a finding confirmed in our investigation, which would also support persistent, progressive symptoms eventually unmasking the correct diagnosis.

Beyond symptom assessment, since women have worse survival, is a more aggressive therapeutic approach needed? Beta receptor antagonist at time of referral correlated with improved survival; however, women were less likely to be on beta receptor antagonists. Women were slightly more likely to undergo future reduction therapies, driven by a higher rate of alcohol septal ablation in women, a finding we have previously shown⁶² but which has not present in all series.⁶³ The 2014 European Society of Cardiology guidelines on diagnosis and management of HCM discuss management of HCM in specific clinical cohorts (athletes, hypertensive patients, and patients with concomitant valvular heart disease) but, apart from discussion of reproduction, do not specifically comment on sex or gender specific approaches to HCM.³ Data herein highlight the clinical importance of differences between men and women with HCM, which we hope future guidelines will emphasize. Our data suggest the need for a different approach to women with HCM: a higher index of diagnostic suspicion and lower threshold for referral for advanced therapies are warranted.

Study limitations

This is a retrospective, single-centre study and is subject to inherent limitations associated with retrospective analyses. All-cause mortality is reported, as the specific cause of death was often unavailable. Whether women have differing rates of cardiac mortality (including SCD) could not be determined by this study and requires further investigation. While a prior study did not find any sex differences in cardiac death, the study population was significantly smaller and a difference in all-cause mortality was not demonstrated,¹⁸ unlike the present study.

Our practice represents a high-volume referral centre with long-standing experience in the care of patients with HCM, as demonstrated by both the highly symptomatic patient population and high rate of surgical myectomy. Accordingly, the study may demonstrate referral bias. However, the large patient population herein also includes many patients with less symptomatic status on index evaluation, which provides applicability across HCM. Changes in medication therapy following referral were not quantitatively assessed and have potential to confound results; however the fact that women more frequently underwent septal reduction therapy may imply that less aggressive therapeutic strategies are unlikely to be the major determinant of worse outcomes in women compared with men. Finally, similar to any other HCM cohort, this study may have included patients with hypertrophic heart disease that is difficult to distinguish from hypertrophic changes of the elderly. However, this population had a mean age of 55 years and all included patients were deemed most likely to have true HCM at the time of the comprehensive index evaluation at our institution.

Conclusions

This large HCM cohort revealed significant demographic, haemodynamic, and prognostic differences between men and women with HCM. Women present at an older age and with more symptoms, a phenomenon that may represent delayed diagnosis. Women also demonstrate more echocardiographic evidence of diastolic dysfunction and obstructive physiology despite similar LV mass index. Women with HCM have worse survival compared with men. These numerous significant clinical differences underscore a need for recognition of the sex and gender differences in the diagnosis and care of patients with HCM. Women with HCM may benefit from more intensive surveillance and medical therapies, as well as earlier consideration of septal reduction therapies.

Conflict of interest: J.B.G—consultant (moderate): MyoKardia. B.J.G.—consultant (moderate): MyoKardia. M.J.A.—consultant (moderate): Audentes Therapeutics, Boston Scientific, Giliead Sciences, Medtronic, MyoKardia, and St Jude Medical; other/royalty/intellectual property (significant): AliveCor and Transgenomic (FAMILION). All other authors have no disclosures, financial or otherwise.