A, Unadjusted sex-specific estimates of survival compared with corresponding age-matched, sex-matched US population rates. Blue indicates male; orange, female; dotted lines, expected values; solid lines, mean observed values; shaded areas, 95% CIs. B, Adjusted (Cox-Kalbfleisch-Prentice) estimates of survival for women and men in a combined sample of 2328 patients (92.9% of the entire cohort) who had complete information with respect to the baseline adjustment factors. Survival curves are adjusted to the median levels of continuous covariates and modal categories of categorical covariates from the combined sample; the resulting difference is nonsignificant. Blue indicates male; orange, female; solid lines, mean values; shaded areas, 95% CIs.
This illustrates how the partial effect on risk changes when the factor plotted is varied, with all other factors not being plotted held fixed to constants based on their median value or modal category. ACE indicates angiotensin-converting enzyme; ARB, angiotensin II receptor blockers; HCM, hypertrophic cardiomyopathy; ICD, implantable cardioversion defibrillator.
eTable. Baseline Characteristics over the Study Period.
eFigure. Temporal trend in baseline characteristics by sex. Curve estimates trend over time stratified by sex using a locally weighted scatterplot smooth (loess).
Customize your JAMA Network experience by selecting one or more topics from the list below.
Meghji Z, Nguyen A, Fatima B, et al. Survival Differences in Women and Men After Septal Myectomy for Obstructive Hypertrophic Cardiomyopathy. JAMA Cardiol. 2019;4(3):237–245. doi:10.1001/jamacardio.2019.0084
Are there any survival differences in women and men undergoing septal myectomy for hypertrophic cardiomyopathy?
In this study of clinical data collected from 1961 to 2016, women were significantly older at the time of surgery, but adjusted statistics of survival were similar between women and men.
This study suggests that clinicians should focus on early identification of disease in both women and men and promptly refer patients who do not respond to medical treatment for surgical evaluation.
Recent data indicate that women with hypertrophic cardiomyopathy (HCM) are older and more symptomatic at presentation and have worse clinical outcomes than men. However, to our knowledge, there are no large studies of the association of patient sex with outcomes after surgical myectomy.
To analyze preoperative characteristics and overall survival of women and men undergoing septal myectomy for obstructive HCM.
Design, Setting, and Participants
This retrospective, single-center study included the clinical data of adult patients who underwent septal myectomy from January 1961 through April 2016. Data analysis occurred from December 2017 to December 2018.
Main Outcomes and Measures
A total of 2506 adults were included; 1379 patients (55.0%) were men. At the time of surgery, women were older, with median (IQR) age of 59.5 (46.6-68.2) years vs 52.9 (42.9-62.7) years in men (P < .001). Women were more likely to have New York Heart Association class III or IV status at presentation (women, 1023 [90.8%]; men, 1169 [84.8%]; P < .001) and more severe obstructive physiology, as reflected in higher resting left ventricular outflow tract gradients (women, 67.0 [36.0-97.0] mm Hg; men, 50.0 [23.0-81.0] mm Hg; P < .001). Women also had a greater likelihood of having moderate or severe mitral regurgitation (606 [55.2%]) than men (581 [43.1%]; P < .001) and higher right ventricular systolic pressure (women, 36.0 [30.0-46.0] mm Hg; men, 33.0 [28.0-39.0] mm Hg; P < .001). The unadjusted overall survival was lower in women, corresponding to a median 3.9-year shorter survival than men (median [IQR] survival time: women, 18.2 [12.1-27.2] years; men, 22.1 [15.1-32.5] years; P < .001). In a multivariable Cox regression analysis, however, the association between sex and mortality was attenuated and not significant after controlling for other baseline variables (hazard ratio, 0.98 [95% CI, 0.76-1.26]; P = .86). Among the covariates in the model, older age at surgery (adjusted hazard ratio [aHR], 3.09 [95% CI, 2.12-4.52]; P < .001), higher body mass index (aHR, 1.22 [95% CI, 0.90-1.66]; P < .001), greater NYHA class (aHR, 2.31 [95% CI, 1.03-5.15]; P = .04), and presence of diabetes prior to surgery (aHR, 1.57 [95% CI, 1.10-2.24]; P = .01) were each independently associated with increased mortality. Operations performed later in the study period (2013 vs 2004) were associated with decreased mortality (aHR, 0.82 [95% CI, 0.55-1.22]; P = .001).
Conclusions and Relevance
In this large cohort of surgical patients with obstructive HCM, we observed significant differences at clinical presentation between women and men, in that women were older and more symptomatic. However, after adjustment for important baseline prognostic factors, there was no survival difference after septal myectomy by sex. Improved care of women with obstructive HCM should focus on early identification of disease and prompt surgical referral of appropriate patients who do not respond to medical treatment.
Hypertrophic cardiomyopathy (HCM) is the most common monogenic heart disease, with a prevalence that may be as high as 1:200.1,2 Hypertrophic cardiomyopathy can be managed effectively both medically and surgically and should now be considered as a treatable genetic disease associated with a normal life span and low disease-associated mortality.3 Septal reduction therapy is useful for patients with medically refractory heart failure symptoms owing to obstructive HCM. Extended transaortic septal myectomy and alcohol septal ablation reduce left ventricular outflow tract (LVOT) obstruction and improve functional status.4
There has been accumulating interest in sex-specific differences and outcomes in a variety of cardiovascular diseases, including HCM, and most studies have shown that women experience worse outcomes than men.5-16 Women with HCM are diagnosed at an older age, tend to be more symptomatic at presentation, have more obstructive physiology, and have reduced survival compared with men.5,17 However, the delay in presentation may not necessarily result in worse survival, and survival differences after myectomy to provide durable relief of LVOT obstruction are less well understood. We hypothesized that patients with obstructive HCM undergoing extended transaortic septal myectomy would demonstrate sex-specific differences in clinical characteristics preoperatively, but that outcomes after surgery would be similar between women and men. We therefore sought to analyze preoperative characteristics, postoperative outcomes, and overall survival of both women and men undergoing extended transaortic septal myectomy for obstructive HCM.
This study included adult patients (≥18 years) who underwent extended transaortic septal myectomy at Mayo Clinic in Rochester, Minnesota from January 1961 through April 2016. We reviewed available clinical and echocardiographic data from all eligible patients and stratified the study cohort by sex. Informed consent was obtained for the use of patient medical records for research purposes in accordance with Minnesota law, and the study was approved by the Mayo Clinic Institutional Review Board.
Survival data were collated from the electronic medical record and multiple national death databases. This involved linking our database with the Social Security Death Index, state death records, and other data sources via Accurint (LexisNexis). Patients not indicated as deceased in these records were assumed to still be alive and were censored 1 month prior to the date last searched (predominantly from December 2017 to January 2018). For a limited number of patients in whom a link could not be established, their vital status was determined by use of the electronic medical records and censoring at last follow-up visit. In addition, the cause of death was ascertained by merging the study subset of deceased patients with the National Death Index.
Information from transthoracic echocardiographic examination was available in patients from November 1974. Because of the long time frame of this study, some patients were evaluated prior to the availability of certain echocardiographic parameters. Left ventricular outflow tract gradients were obtained by continuous-wave Doppler interrogation of the LVOT from an apical window and calculated using the modified Bernoulli equation (gradient = 4VLVOT2, where VLVOT is peak LVOT velocity). Provoked gradients were determined by provocative maneuvers (ie, Valsalva maneuver, amyl nitrite inhalation, and exercise). Ejection fraction, left ventricular (LV) cavity size, LV mass, and wall thickness were determined as previously described.18 Mitral regurgitation was graded as none to trivial (0), mild (1), moderate (2), moderately severe (3), or severe (4) after analyzing jet area and width, and spectral Doppler intensity, as well as regurgitation quantitation with the continuity and/or proximal isovelocity surface area method, as appropriate.19 Mitral regurgitation could not be quantitated in all patients due to jet eccentricity and LVOT turbulence merging with the regurgitant jet flow convergence. Right ventricular systolic pressure was estimated in a standard procedure as per American Society of Echocardiography guidelines.20 Information from the most recent Doppler echocardiograms performed before septal myectomy were compared with those obtained postoperatively prior to hospital dismissal.
Descriptive statistics used to summarize baseline data included median (interquartile range [IQR]) for continuous variables and number (percentage) for categorical or ordinal variables. Statistical tests for assessing baseline differences by sex were based on the Wilcoxon-Kruskal-Wallis test for continuous or ordinal variables and the Pearson χ2 test for categorical variables. Unadjusted survival time was summarized with Kaplan-Meier survival probabilities and quartile estimates; follow-up time was described with quartiles by the same estimator, except that the codes for the event or censoring indicator were reversed. Multivariable Cox proportional hazards modeling was used to assess the partial effect of sex on time until death in the presence of known important baseline prognostic factors. Model assumptions of proportional hazards and additivity, specifically for nonadditivity of effects by sex, were assessed with global (overall) tests and further by partial tests, if warranted. Independent variables that violated proportional hazards (atrial fibrillation and left ventricular ejection fraction) were entered into the model as stratification variables rather than as covariates as a way to adjust for outcome without the need to assume proportional hazards. The Cox model was also used to analyze postoperative length of stay, with in-hospital deaths censored and tied responses handled exactly in the model. In addition, analysis of covariance ordinal regression based on the proportional-odds model was applied to postoperative echocardiographic data to compare baseline-adjusted responses between men and women. In all these regression models, continuous covariates were fitted using restricted cubic splines with 4 knots to relax linearity assumptions and to allow nonlinear effects. Effects of variables are estimated with adjusted hazard ratios (and 95% CIs). For continuous variables, hazard ratios were calculated comparing the 75th percentile with the 25th percentile. Since these variables are modeled nonlinearly, general tests of association can easily generate significant P values, even when the confidence interval of a hazard ratio contains 1.0. Statistical significance was defined as P ≤ .05, and analyses were performed using SAS version 9.4 (SAS Institute Inc) and R version 3.2.3 (R Foundation for Statistical Computing) from December 2017 to December 2018.
A total of 2540 patients underwent septal myectomy at the study institution from January 1961 to April 2016. Thirty-four patients who did not consent for research were excluded from this analysis. The study group was therefore 2506 patients, of whom 1379 patients (55.0%) were men. The number of patients increased markedly over the course of the study (n = 49 in the 1960-1970s, 76 in the 1980s, 267 in the 1990s, 985 in the 2000s, and 1129 from 2010 to 2016), with 84.4% of patients having their surgery in or after 2000. The median (IQR) age was 55.6 (44.5-65.6) years. A total of 2192 patients (87.4%) presented with New York Heart Association (NYHA) class III or IV symptoms. (Patients with class I or II symptoms preoperatively had troublesome syncope or extremely high gradients that led to activity limitations.21)
Patient demographics are shown by sex in Table 1. At the time of surgery, women were older, with a median (IQR) age of 59.5 (46.6-68.2) years vs 52.9 (42.9-62.7) years for men (P < .001), and were more symptomatic, with higher likelihood of being in NYHA class III or IV at presentation (women, 1023 [90.8%]; men, 1169 [84.8%]; P < .001). Women had greater prevalence of systemic hypertension (586 [52.0%]) than men (652 [47.3%]; P = .02), were more likely to have a family history of HCM (239 [21.3%]) than men (236 [17.2%]; P = .008), and were more likely to have had previous septal reduction therapy with either alcohol septal reduction or transaortic septal myectomy (women 62 [5.5%]; men, 41 [3.0%]; P < .002). There were no sex differences in the other comorbidities, nor were there differences in the use of any preoperative medical therapy between women and men.
The distribution of women vs men undergoing myectomy has remained fairly stable through time (for example, 1960-1979: 22 of 1127 women [2.0%]; 27 of 1379 men [2.0%]; 2010-2016, 514 of 1127 women [45.6%]; 615 of 1379 men [44.6%]; P = .81 across all 5 decades; Table 1), and additional demographic data stratified by decade are shown in the eTable and the eFigure in the Supplement. In more recent years, the median age of both women and men has increased, and preoperative resting gradient has decreased (eTable in the Supplement).
Preoperative echocardiographic data are shown in Table 1. Women had more severe obstructive physiology as reflected in greater median (IQR) resting LVOT gradients (women, 67 [36-97] mm Hg; men, 50 [23-81] mm Hg; P < .001) and median (IQR) provoked LVOT gradients (women, 81 [64-208] mm Hg; men, 74 [55-100] mm Hg; P < .001) among those subsets of patients with available measurements. The proportion of individuals found to have moderate or severe mitral valve regurgitation was greater in women (606 [55.2%]) than men (581 [43.1%]; P < .001).
Septal hypertrophy, measured as median (IQR) absolute anteroseptal thickness (women, 19 [17-23] mm; men, 20 [18-23] mm) and median (IQR) posterior wall thicknesses (women, 13 [11-15] mm; men, 14 [12-15] mm), was less pronounced in women (both P < .001). The proportion of women with extreme anteroseptal wall thickness (≥30 mm) was slightly lower (women, 47 [4.4%]; men, 85 [6.4%]; P < .001). Women also had significantly smaller LV mass (women, 262 [213-325] g; men, 328 [259-400] g; P < .001) and LV end-diastolic dimensions (women, 43 [40-47] mm; men, 47 [43-51] mm; Table 1). In patients with right ventricular systolic pressure assessment (n = 1704; 68.0%), women demonstrated higher estimated right ventricular systolic pressure (36.0 [30.0-46.0] mm Hg) than men (33.0 [28.0-39.0] mm Hg; P < .001).
The median (IQR) interval from operation to the predischarge echocardiogram was 4 (3-5) days. Table 2 shows observed changes in echocardiographic parameters, as well as expected mean changes for women and men with baseline measurement and other covariates, adjusted to their median levels. After extended transaortic septal myectomy, there was a significant reduction in LVOT gradient in both women and men, with a similar magnitude of expected mean change after baseline adjustment (−50.3 [95% CI, −51.5 to −49.1]; −50.4 [95% CI, −51.4 to −49.4]; P = .88). There was no significant difference in reduction of anteroseptal wall thickness, but adjusted difference in reduction of posterior wall thickness was greater in women than men (−0.9 [95% CI, −1.2 to −0.6; −0.5 [95% CI, −0.7 to −0.2]; P = .006).
All patients underwent septal myectomy. Interestingly, women were less likely to undergo concomitant coronary artery bypass grafting (women, 80 [7.2%]; men, 140 [10.3%]; P = .008), but the percentage who underwent concomitant valve operations was not significantly different (women, 451 [40.7%]; men, 509 [37.4%]; P = .10).
Early postoperative deaths (within 30 days of operation) were observed in 21 participants (0.8%) over the course of the study. As time progressed, early mortality decreased markedly, from 4 of 49 (8.2%) in the 1960s and 1970s to 3 of 1129 (0.3%) in the present decade, with declining percentages throughout the study period. Early mortality occurred in 13 of 1127 women (1.2%) and 8 of 1379 men (0.6%; P = .12). Postoperative median (IQR) length of hospital stay was longer in women than men (6 [5-7] days vs 5 [4-7] days; P < .001 after adjusting for age, year of operation, and diabetes status). Risk of postoperative stroke was slightly but nonsignificantly higher in women (12 of 1125 [1.1%] vs 6 of 1377 [0.4%]; P = .07), whereas early arrhythmias after surgery, including atrial fibrillation and ventricular tachycardia, were similar among the 2 sexes (atrial fibrillation: women, 334 of 1125 [29.7%]; men, 419 of 1377 [30.4%]; P = .70; ventricular tachycardia: women, 26 of 1125 [2.3%]; men, 35 of 1377 [2.5%]; P = .71).
During a median (IQR) follow-up of 8.2 (3.1-13.2) years, a total of 445 patients died. The 1-year, 5-year, and 10-year survival estimates were 98.3% (95% CI, 97.7%-98.8%), 94.2% (95% CI, 93.1%-95.2%), and 84.6% (95% CI, 82.6%-86.4%), respectively, and median (IQR) survival time was 20.8 (13.6-30.8) years. The cause of mortality was known in 371 of the 445 deaths (83.4%). Deaths associated with cardiovascular conditions accounted for 1124 of 197 deaths (58.2%) in women and 99 of 174 deaths (56.9%) in men (P = .81).
In the unadjusted analysis (Figure 1A), women had lower survival than men after myectomy, corresponding to an apparent 3.9-year shorter median survival time (median [IQR] survival time: men, 22.1 [15.1-32.5] years; women, 18.2 [12.1-27.2] years). These estimates were significantly worse than those of age-matched women in the general US population (238 observed vs 152.3 expected deaths; P < .001), whereas observed and population-based survival estimates for men varied considerably over time but did not have significantly different means (207 observed vs 184.7 expected deaths; P = .10) (Figure 1A). In the multivariable-adjusted Cox analysis, the association between sex and adjusted mortality was attenuated and not significant after controlling for all other covariates included in the model (hazard ratio [HR], 0.98 [95% CI, 0.76-1.26]; P = .86) (Table 3 and Figure 1B). Furthermore, a global test of all interaction terms between sex and the other covariates suggested no important sex-specific effects on mortality risk (χ224 = 27.2; P = .29). Independent factors associated with mortality (Table 3 and Figure 2) were older age at surgery (adjusted HR, 3.09 [95% CI, 2.12-4.52]; P < .001), higher body mass index (adjusted HR, 1.22 [95% CI, 0.90-1.66]; P < .001), greater NYHA class (adjusted HR, 2.31 [95% CI, 1.03-5.15]; P = .04), and presence of diabetes prior to surgery (adjusted HR, 1.57 [95% CI, 1.10-2.24]; P = .01). In addition, myectomy performed more recently was independently associated with decreased mortality (adjusted hazard ratio, 0.82 [95% CI, 0.55-1.22]; P = .001).
In this review of 2506 patients undergoing transaortic septal myectomy for obstructive HCM, we observed significant differences in preoperative characteristics between women and men. Importantly, women were older and more symptomatic than men at the time of surgery. However, there were no differences between the 2 sexes in early postoperative outcomes, including mortality and relief of LVOT obstruction. Factors associated with overall mortality included older age at surgery, greater body mass index, higher NYHA class, diabetes mellitus, and surgery performed earlier in the study period.
Overall survival after septal myectomy was worse in women than men. Indeed, outcome of women who underwent surgery was reduced compared with that expected in a matched US population. In contrast, for men, observed and expected survivals were similar (Figure 1A). In a previous study of patients who underwent septal myectomy at the Mayo clinic, 10-year survival of observed vs expected populations seemed comparable.22 The present investigation, however, extends follow-up beyond 10 years, and it is in this later postoperative period that survival appears reduced in surgical patients compared with a matched US population; these differences were observed in both sexes.
Although the unadjusted analysis demonstrated worse survival in women (Figure 1A), this difference was attenuated in an adjusted model (Figure 1B). These results differ from a report by Geske et al,5 who studied more than 3600 patients with HCM, most of whom had no LVOT obstruction. In that analysis, women had poorer survival than men, and septal reduction was performed in only 32%. These results also contrast with the investigation by Olivotto et al,17 who observed worse chances of survival in women with HCM; however, in their study, only 10% of patients had advanced (NYHA class III or IV) symptoms, and less than 30% had important LVOT obstruction.
In this large study of patients with HCM who underwent transaortic septal myectomy, female sex was not a factor independently associated with mortality (Table 3). This differs from the findings of Woo et al23 who reported that, among 338 patients with HCM undergoing septal myectomy, female sex was an important factor associated with late survival, with a hazard ratio for death of 2.5 in their multivariable analysis. The explanation for the different results is not clear. In both the Woo et al study23 and this analysis, older age was associated with reduced late survival, but other risk factors associated with survival differ from the present investigation. For example, left atrial diameter was independently associated with late mortality in the study from Toronto, Ontario, Canada, but in a recent investigation from the study clinic, left atrial size, as measured by left atrial volume index, was not associated with late death.24 Another difference in the studies was the finding in the patients that diabetes is an important factor associated with overall mortality.
Risk of 30-day mortality was 0.8% in this review, and there were no differences in early outcomes between women and men. These results are consistent with reports from other centers, where early operative mortality rates of less than 1% can be achieved when patients with HCM are treated by experienced teams.25,26 Indeed, some clinicians argue that HCM should be treated at experienced referral centers to achieve the best patient outcomes.27 Other early postoperative complications, including atrial fibrillation and stroke, were similar among the 2 sexes.
Hypertrophic cardiomyopathy is a genetic disease in which inheritance is not linked to sex chromosomes, and thus women and men should be affected equally. Lower proportions of women vs men in contemporary medical and surgical series of HCM may be because of underdiagnosis, which can be explained by multiple factors.5,17,28,29 First, studies suggest that women are less aware of their risks for cardiovascular disease and may not seek medical attention during earlier stages of the disease.30 Second, direct comparisons between women and men with similar cardiovascular risks have demonstrated that women are less likely to receive aggressive treatment.31 This may be owing to some degree of physician bias, and it is possible that surgical referral for septal reduction is less common in women with similar disease profile compared with men. Early and late results from the present study, however, show that regardless of potential referral bias, surgical myectomy is an excellent treatment option for women with obstructive HCM.
Another consideration is that healthy women have slightly thinner interventricular septa compared with healthy men,32 but the diagnostic criteria for HCM are identical for both sexes.33,34 The variations in septal thickness are especially important for patients undergoing echocardiographic screening for HCM. Thus, women may have relatively more severe hypertrophy when they first meet criteria for diagnosis of HCM. This potential lead-time bias could also contribute to survival differences observed in studies of women and men with HCM.
This is a retrospective study from a single tertiary center. The analysis is limited by inherent selection bias. The results reported may not be generalizable to other centers.
In this large cohort of surgical patients with obstructive HCM, women were older and more symptomatic at clinical presentation compared with men. Women also had more severe obstructive physiology and diastolic dysfunction in association with smaller LV mass index on echocardiography. However, survival after septal myectomy was not associated with female sex after adjustment for important baseline prognostic factors. Improved care of women with obstructive HCM should focus on early identification of disease and prompt surgical referral of appropriate patients who do not respond to medical treatment.
Corresponding Author: Hartzell V. Schaff, MD, Department of Cardiovascular Surgery, Mayo Clinic, 200 First St SW, Rochester, MN 55905 (firstname.lastname@example.org).
Accepted for Publication: December 28, 2018
Published Online: February 27, 2019. doi:10.1001/jamacardio.2019.0084
Author Contributions: Dr Meghji had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Concept and design: Meghji, Nguyen, Fatima, Nishimura, Ommen, Dearani.
Study concept and design: Schaff.
Acquisition, analysis, or interpretation of data: Meghji, Nguyen, Fatima, Geske, Lahr, Schaff.
Drafting of the manuscript: Meghji, Nguyen, Fatima, Schaff.
Critical revision of the manuscript for important intellectual content: Meghji, Nguyen, Geske, Nishimura, Ommen, Lahr, Dearani, Schaff.
Statistical analysis: Lahr.
Obtained funding: Schaff.
Administrative, technical, or material support: Nguyen, Fatima, Nishimura, Ommen, Schaff.
Supervision: Schaff, Geske, Ommen, Dearani.
Conflict of Interest Disclosures: None reported.
Funding/Support: This work was supported by the Paul and Ruby Tsai Family.
Role of the Funder/Sponsor: The funder had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Create a personal account or sign in to: