Original research•

Sex-specific aspects on prognosis after aortic valve replacement for aortic stenosis: a SWEDEHEART registry study

Abstract

Objective To compare long-term cardiovascular (CV) outcomes between men and women with aortic stenosis (AS) undergoing aortic valve replacement (AVR) by the type of valve implant.

Methods The study population consisted of 14 123 non-selected patients with AS undergoing first-time AVR and included in the Swedish Web system for Enhancement and Development of Evidence-based care in Heart disease Evaluated According to Recommended Therapies (SWEDEHEART) registry during 2008–2016. Comparisons were made between men and women and type of valve implant (ie, surgical implantation with a mechanical (mSAVR) (n=1 966) or biological valve (bioSAVR) (n=9 801)) or by a transcatheter approach (TAVR) (n=2 356). Outcomes included all-cause mortality, ischaemic stroke, major bleeding, thromboembolic events, heart failure and myocardial infarction, continuously adjusted for significant comorbidities and medical treatment.

Results In the mSAVR cohort, there were no significant sex differences in any CV events. In the bioSAVR cohort, a higher risk of death (HR: 1.14; 95% CI: 1.04 to 1.26, p=0.007) and major bleeding (HR: 1.41; 95% CI: 1.18 to 1.69, p<0.001) was observed in men. In the TAVR cohort, men suffered a higher risk of death (HR: 1.24; 95% CI: 1.07 to 1.45, p=0.005), major bleeding (HR: 1.35; 95% CI: 1.00 to 1.82, p=0.022) and thromboembolism (HR: 1.35, 95% CI: 1.00 to 1.82, p=0.047).

Conclusion No significant long-term difference in CV events was noted between men and women undergoing AVR with a mechanical aortic valve. In both the bioSAVR and TAVR cohort, mortality was higher in men who also had an increased incidence of several other CV events.

What is already known on this topic

  • In patients with aortic stenosis, few studies have examined sex differences on the long-term prognosis of important cardiovascular events (CV) related to different valve implants.

What this study adds

  • In this study, we found no significant long-term difference in CV events between men and women undergoing aortic valve replacement (AVR) with a mechanical aortic valve.

  • Mortality and major bleeds were higher in men undergoing surgical AVR with a biological prosthesis. Mortality, major bleeds and thromboembolic events were more frequent in men undergoing transcatheter AVR.

How this study might affect research, practice or policy

  • Our results suggest a need for an increased awareness among physicians for sex-specific differences in the risk of CV events during long-term follow-up after AVR.

Introduction

In severe aortic stenosis (AS), aortic valve replacement (AVR) improves symptoms and survival. Current treatment options for AS include surgical AVR with a mechanical (mSAVR) or biological valve (bioSAVR) or a transcatheter (TAVR) approach. The choice of intervention is often based on careful medical assessment taking into consideration patient characteristics, patient preference and the risks associated with the different valve types. In current guideline recommendations for intervention in AS, women receive attention as a special patient population.1 This cautiousness regarding recommendations is largely based on reports of excess female mortality, foremost in those undergoing surgical valvular replacement for severe AS.2 3

Sex-specific differences in presentation, pathophysiology and anatomy of valvular disease are proposed as contributory factors in the worse prognosis in women after valvular surgery.4 It is not uncommon that women with severe AS are diagnosed later than men and are therefore older and have more disabling comorbidities putting them at a higher surgical risk. Also, each of the different aortic valve implants carry different risks for subsequent cardiovascular (CV) events including thromboembolic, bleeding and heart failure.5 Sex-specific differences in long-term CV outcomes in patients treated for AS with AVR, related to different valve implants, remain unclear.

The aim of the present study is to describe and compare important long-term CV outcomes by sex and type of AVR in a contemporary and all-inclusive population with severe AS. By identifying the long-term risks associated with each type of valve implant a more optimal selection of patients and clinical follow-up can be ensured for both men and women.

Methods

Study cohort and data sources

All patients undergoing cardiac surgery or percutaneous valve interventions who are residents in Sweden are continuously included in the Swedish Web system for Enhancement and Development of Evidence-based care in Heart disease Evaluated According to Recommended Therapies (SWEDEHEART) registry.6 The present study cohort included all patients (n=14 123) in the SWEDEHEART registry undergoing AVR due to stenosis, with or without coronary artery bypass grafting or percutaneous coronary intervention, between 1 January 2008 and 31 December 2016 and who were alive at discharge from the index intervention. Both mechanical and biological valve prostheses were included. Patients with missing information about previous coronary intervention or left ventricular ejection fraction at index intervention and, any patients undergoing concomitant surgery of the aorta, other heart valves, other cardiac or pulmonary vasculature defects were excluded.

Data collection

The SWEDEHEART registry contains detailed information on the procedures and concomitant diseases. Baseline information from SWEDEHEART was enriched with information from the National Patient Register (NPR), which includes the International Code of Disease, Tenth Revision (ICD-10) of all hospital admissions in Sweden since 1987 and the Swedish Cancer Registry (Swedish National Board of Health and Welfare) with information of a new cancer diagnosis since the mid-1980s.7 Linkage was based on the unique 10-digit personal identification number assigned to all Swedish residents at birth or immigration. The National Board of Health and Welfare approved merging of the registries.

All patients were followed up through computerised linkage between the database and the updated census register, the Swedish Cause of Death Register, the NPR and the Swedish Cancer Registry, all managed by the National Board of Health and Welfare. The start date was 1 day after discharge from the index intervention. The study population was followed until death or the end of follow-up (FU) (31 December 2017), whichever occurred first.

Baseline characteristics and comorbidities

Information on baseline characteristics and previously diagnosed comorbidities were collected from the SWEDEHEART registry and enriched with data from NPR and the Swedish Cancer Registry up to 3 years before the valve intervention. Predefined comorbidities at baseline were diabetes, hypertension, congestive heart failure (CHF), atrial fibrillation (AF), myocardial infarction (MI), any thromboembolism (TE) (ie, ischaemic stroke, systemic embolism, pulmonary embolism and venous thromboembolism), cancer and major bleeding (ie, haemorrhagic stroke and/or hospitalisation for any bleeding event). Comorbidities were continuously monitored and updated during the FU and included in the adjusted models. ICD-10 codes were applied to identify comorbidities and outcome events (online supplemental eTable 1).

Medical treatment

Information on the dispensation of medical treatment was collected by computerised linkage with the Dispensed Drug Register. This register contains information on all prescribed drugs dispensed from all pharmacies in Sweden. The patient was considered exposed to the corresponding dispensed drug for 120 days after each dispensation. Information on dispensation was continuously updated, and patients could be exposed to different pharmaceutical treatments during FU.

Outcome events

Outcome events after discharge from the index AVR were all-cause mortality, new hospitalisation for ischaemic stroke, any TE, major bleeding, MI and CHF (online supplemental eTable 1).

Statistical analysis

Categorical variables were expressed as frequencies and percentages and continuous variables as median and IQRs (Q1–Q3). Exposure to a pharmaceutical agent was assessed and updated at every dispensation. Consequently, each patient could be exposed to different pharmaceutical categories during FU. The number of outcome events are presented as incidence rate/100 person-years. Kaplan-Meier plots for mortality, TE and major bleeding events describe the time to event separated by sex.

The associations between sex and repeated outcome events were analysed using multivariable Cox regression models. To account for data dependency due to recurrent events, robust Huber-White standard errors were computed. A crude model that included only sex was performed for each outcome event: all-cause mortality, ischaemic stroke, any TE, major bleeding, MI and CHF. In addition, adjusted models were fitted, to include age, diabetes, hypertension, prior coronary intervention, coronary intervention at index valve replacement, and was continuously updated for the comorbidities CHF, AF, MI, ischaemic stroke, for all outcome events. The following comorbidities and medical treatments were also included in the adjusted models for specific outcomes; renal failure and cancer for all-cause mortality; antiplatelet and oral anticoagulant (OAC) treatment for ischaemic stroke; previous TE, antiplatelet and OAC treatment for TE; previous major bleeding, antiplatelet and OAC treatment for major bleeding event; antiplatelet and OAC treatment for MI; ACE inhibitors (ACE-I) or angiotensin II receptor blockers (ARB) for CHF. The results of unadjusted and fully adjusted Cox analysis are presented as HR with 95% CIs.

Patient and public involvement

Neither patients nor the public were involved in the design, or conduct, or reporting, or dissemination plans of our research.

Results

Clinical characteristics

The study population consisted of 14 123 patients with AS who underwent mSAVR (13.9%), bioSAVR (69.4%) or TAVR (n=16.7%). No patients were lost to follow-up. The median (max) follow-up time in years was 5.4 (9.97), 4.7 (9.97) and 2.75 (9.71) in the mSAVR, bioSAVR and TAVR cohorts, respectively.

The baseline characteristics for the different cohorts are shown in table 1. The largest proportion of women was found among patients undergoing TAVR compared with surgical intervention. In general, women were slightly older than men. Except for hypertension, the frequency of CV comorbidities was higher among men than women in all three cohorts. Concomitant coronary artery disease was the standout comorbidity, being much more common in men than women. Coronary artery revascularisation was performed in conjunction with the index AVR more often in the surgical cohorts compared with TAVR, both in men and women. A diagnosis of CHF at study inclusion was much more common in the TAVR cohort, in both men and women, compared with both the mSAVR and bioSVAR cohorts. In patients undergoing TAVR the transfemoral approach was used in 96% of the patients. EuroSCORE was reported in 23.1% of mSAVR, 23.2% in bioSAVR and 40.6% in patients undergoing TAVR. The majority of the patients had a EuroSCORE of 0–4 in the mSAVR cohort, and 5–9 in bioSAVR cohort and 10–14 in the TAVR cohort (online supplemental eTable 2).

Table 1
|
Baseline characteristics grouped by prothesis type and sex

Pharmaceutical exposure during follow-up

As expected, warfarin exposure was highest in the mSAVR cohort compared with the bioSAVR and the TAVR cohorts. Men were more commonly exposed to warfarin and antiplatelet drugs when treated with TAVR. A higher proportion of men were exposed to ACE-Is/ARBs, whereas women were more often exposed to diuretics (online supplemental eTable 2).

Association of outcome to prosthesis type

The total number of outcome events and the corresponding incidence rates of death and specific CV outcomes by prothesis type and sex are presented in table 2. Of all the CV events, incidence rates were highest for CHF, followed by death, in both men and women.

Table 2
|
Outcome events at follow-up by prosthesis type and sex*

Outcome in the mSAVR cohort

At 1-year follow-up, trends existed for a lower risk of death (HR: 0.57; 95% CI: 0.28 to 1.17) in men compared with women, whereas men tended to have a higher risk of TE (HR: 2.49; 95% CI: 0.96 to 6.41), major bleeding (HR: 1.02; 95% CI: 0.50 to 2.05) and CHF (HR: 1.08; 95% CI: 0.64 to 1.80). After adjustments, a borderline lower rate of death was seen in men compared with women (HR: 0.50; 95% CI: 0.24 to 1.03). At long-term follow-up the crude risk for all outcomes tended to be higher in women. However, no significant differences were noted after adjustments. (figures 1A–C, and 2A).

Figure 1
Request permissionrequest permission icon
Figure 1

Incidence of outcome by sex in mSAVR cohort. Kaplan-Meier estimates for (A) all-cause mortality, (B) thromboembolism and (C) bleeding. mSAVR, mechanical prothesis.

Figure 2
Request permissionrequest permission icon
Figure 2

The relative risk for all-cause mortality and cardiovascular events for men relative to women at long-term follow-up by valve type (A) mSAVR, (B) bioSAVR and (C) TAVR. bioSAVR, biological surgical prosthesis; CHF, congestive heart failure; MI, myocardial infarction; mSAVR, mechanical prothesis; TAVR, transcatheter aortic valve replacement.

Outcome in the bioSAVR cohort

At 1-year follow-up the crude risk of major bleeding (HR: 1.08; 95% CI: 0.82 to 1.41), MI (HR: 1.07; 95% CI: 0.71 to 1.62) and CHF (HR: 1.13; 95% CI: 0.95 to 1.35) tended to be higher in men compared with women, whereas their risk of death (HR: 0.99; 95% CI: 0.79 to 1.26), ischaemic stroke (HR: 0.70; 95% CI: 0.50 to 0.98) and TE (HR 0.87; 95% CI 0.70 to 1.07) was lower. After adjustments no statistically significant differences in outcome were noted at 1 year. At long-term follow-up crude risks for death, major bleeding, MI and CHF tended to be higher in men than women. After adjustments there remained a higher risk of death and bleeding in men (figure 3A–C and figure 2B).

Figure 3
Request permissionrequest permission icon
Figure 3

Incidence of outcome by sex in bioSAVR cohort. Kaplan-Meier estimates for (A) all-cause mortality, (B) thromboembolism and (C) bleeding. bioSAVR, biological surgical prosthesis.

Outcome in the TAVR cohort

At 1 year and long-term follow-up crude risks for all outcome events were higher in men than women. At 1 year, after adjustments all risks tended to be higher in men, but only the risk of TE and CHF remained higher in men. At long-term follow-up the higher risk of death, TE and major bleeding in men remained significant (figure 4A–C and figure 2C).

Figure 4
Request permissionrequest permission icon
Figure 4

Incidence of outcome by sex in TAVR cohort. Kaplan-Meier estimates for (A) all-cause mortality, (B) thromboembolism and (C) bleeding. TAVR, transcatheter aortic valve replacement.

Discussion

In this large study, based on Sweden’s complete national patient registries, we found that men undergoing bioSAVR or TAVR, had higher all-cause death rates and major bleeding events as compared with women, and in those undergoing TAVR, men also had higher rates of TE after adjustments for significant comorbidities and medical treatment. There was no difference between men and women in CV events in the mSAVR cohort.

Our findings contrast those from previous studies in which women have shown higher rates of mortality compared with men after cardiac surgery, including AVR.2 3 8 9 The reasons for higher mortality in women are multifactorial.10 Studies reporting higher death rates in women post SAVR have found women to be older, to have more comorbidities such as hypertension, diabetes, peripheral artery disease and renal dysfunction which negatively impact on outcomes. A further explanation may be the urgency of the procedure with a late diagnosis more common in women than men.11 Surgery may be more technically challenging in women due to anatomical reasons, such as smaller valvular orifices, and thereby heightening the risk of complications.12 In our study, we observed a tendency towards a higher death rate in women receiving a mechanical prothesis at 1-year follow-up which disappeared at long-term follow-up. Compared with other studies citing a sex-specific difference in death rates, the women in our study undergoing mSAVR were of a similar age and had, apart from hypertension, fewer comorbidities compared with men. Also, we did not include in-hospital mortality. Patients in the bioSAVR and TAVR cohorts were much older than the mSAVR group and the higher death rate in men may be explained by the longer longevity of women.

Among the CV events, the incidence rates for CHF were the highest after all types of replacement (ie, mSAVR, bioSAVR and TAVR) with no difference observed between the sexes at long-term outcome. These findings corroborate findings in other studies of a significant incidence and prevalence of cardiac decompensation even after AVR. In our study, all pre-existing CHF significant comorbidities and medical treatment was adjusted for. The mechanisms of heart failure, specifically post AVR, are complex and not well understood but may include valvular complications as well as the myocardium’s response to haemodynamic overload.13 The sex difference observed at 1 year in our TAVR cohort may partially be explained by prosthesis mismatch with subsequently higher rates of paravalvular regurgitation which is more common in men undergoing TAVR.14 Late development of heart failure post AVR, in all cohorts, may have several causes and there is data suggesting sex-specific differences in pathophysiological mechanisms. Oestrogen appears to favourably influence cardiac remodelling; maladaptive remodelling (ie, eccentric hypertrophy and focal fibrosis) occurs more frequently in men.15 Heart failure due to reduced ejection fraction is more common in men compared with women who more often suffer from CHF with preserved ejection fraction (HFpEF). The evidence for treatment of HFpEF has lagged behind CHF due to reduced ejection fraction and may well explain why, in our study, a higher proportion of men were exposed to treatment with renin–angiotensin antagonists and women diuretics.16

Whereas no sex difference in major bleeding was seen for the mSAVR cohort, major bleeding was seen more often in men than women in the bioSAVR and TAVR cohorts. In patients receiving mechanical valves tight warfarin controls may explain the lower risk of bleeding in these patients and the lack of difference between the sexes. The bleeding risks noted for men in the bioSAVR and TAVR groups were not apparent at follow-up at 1 year suggesting bleedings other than procedure related. The cause of bleeds is unknown to us but, may be explained by the higher combined exposure in men to antiplatelet and warfarin treatment.

Interestingly and paradoxically, we also found an increased risk of TE in men compared with women in the TAVR cohort, both at 1 year and at long-term follow-up. This finding contrasts with reports of a similar incidence in venous TE for women and men, despite differences in specific provoking factors.17 Although continuous adjustments were made for OAC and antiplatelet medications, a possible mechanism may be the interruption of these medications due to bleeds, thereby opening a window for TE events.

Strengths and limitations

This study is based on national registries which enables generalisation to European AVR populations. The SWEDEHEART registry is a national registry in which all patients undergoing AVR are enrolled and is continuously subjected to on-site monitoring and validation resulting in high levels of agreement between the registry and data from electronic health records.6 The NPR contains data on all hospital admissions in Sweden since 1987 for which the validity of CV diagnoses approaches 95%.7 A major limitation is the retrospective use of registry data without access to more detailed operational information. Although we included most clinically relevant factors, we did not have available data on the extent of chronic obstructive pulmonary disease, endocarditis or valve sizes. Our definition of drug exposure does not guarantee actual intake of prescribed medications. Furthermore, we had no information on the international normalised ratio of patients prescribed warfarin. However, the anticoagulant treatment in Sweden is generally of high quality.18 Pharmaceuticals are involved in a comprehensive reimbursement programme in Sweden, and usage of pharmaceuticals outside this programme can be expected to be marginal. Finally, our study must be interpreted with caution because of inherent limitations to all observational studies related to confounding by indication.

Conclusions

Our findings suggest that women and men who undergo AVR at long-term suffer from different CV events which vary according to the type of valve implanted. In patients surviving to discharge, no sex difference was noted for CV events in patients receiving a mechanical prothesis, whereas men receiving a bioprosthetic valve fared worse. Mortality and major bleeds were higher in men undergoing bioSAVR. Thromboembolic events were more frequent in men undergoing TAVR. Heart failure was the most common complication following AVR, of any valve type, in both men and women. To improve outcomes in men and women undergoing AVR more research is needed to identify the causes of these sex-specific differences.

Article metrics
Altmetricopen-url
Dimensionsopen-url