Sex-based differences in adverse left ventricular remodelling and clinical outcomes after ST-segment elevation myocardial infarction

Abstract

Aims

The impact of sex on adverse left ventricular remodelling (LVR) after ST-elevation myocardial infarction (STEMI) is unclear due to conflicting results. This study sought to establish sex-based differences in adverse LVR using cardiovascular magnetic resonance (CMR) among STEMI patients and their impact on clinical outcomes.

Methods and results

The study included patients with a first STEMI who underwent primary percutaneous coronary intervention (PCI). Cardiovascular magnetic resonance was performed at 6 days (interquartile range [IQR]: 4–9 days) and after 6 months (6.42 months; IQR: 5.98–7.47 months). Follow-up was 6.94 years (IQR: 4.48–9.32 years). The primary endpoint was the presence of adverse LVR (>15% of LV end-diastolic volume and a decrease of >3% in LV ejection fraction) at 6 months. The secondary endpoint was major adverse cardiac events (MACEs), defined as a combined variable: cardiovascular death, heart failure admission, or ventricular arrhythmias. One thousand sixty-seven patients were included (17.5% women; mean age: 58.71 ± 11.85 years). Women were older and had more cardiovascular risk factors (CVRF). There was no association between sex and adverse LVR [OR: 0.80; 95% confidence interval (CI), 0.39–1.64; P = 0.536]. Major adverse cardiac events occurred in 177 patients (16.7%) and was more frequent in women (22.6% vs. 15.4%; P = 0.017). However, after adjusting for baseline differences and CVRF, the female sex was not associated with MACE (hazard ratio: 1.21; 95% CI, 0.81–1.81; P = 0.343).

Conclusion

The higher rate of MACE after STEMI in women compared to men appears to be associated with a higher prevalence of CVRF and comorbidities rather than a more significant occurrence of adverse LVR.

Graphical Abstract

Open in new tabDownload slide

Introduction

Sex differences in clinical care and mortality after ST-segment elevation myocardial infarction (STEMI) in the current primary percutaneous coronary intervention (PCI) era remain a matter of constant debate and have significant clinical as well as therapeutic implications.^1–5 Several studies have shown that women have a higher post-STEMI mortality compared with men.^6–10 Age differences, comorbidities, delay in reperfusion, or medical treatment could justify possible explanations for these sex-related findings.^1,6,11 However, other authors have suggested that additional factors, such as biological sex and hormonal differences, might contribute to these worse outcomes.^12–16

Post-STEMI adverse left ventricular remodelling (LVR) is a complex process that can lead to adverse cardiovascular outcomes. It occurs in a significant proportion of patients (approximately 10–50%) within the first year, primarily influenced by factors such as infarct size (IS) and the inflammatory response. The incidence of adverse LVR can vary depending on its definition, but this should not affect the comparison between sexes. Understanding the relationship between sex, myocardial salvage (MS), IS, and post-infarction LVR is crucial since post-STEMI outcomes are closely tied to tissue perfusion after revascularization and the final IS. Cardiovascular magnetic resonance (CMR) offers valuable prognostic information in STEMI by assessing the area at risk (AAR), IS, microvascular obstruction (MVO), MS, and post-infarction LVR. While earlier studies on sex-based differences in adverse remodelling may have limited applicability to modern clinical settings, CMR offers a more precise assessment of these factors, making it suitable for studying sex differences in infarct features.^17–19 Therefore, we hypothesize that potential differences in adverse LVR might contribute to the worse post-STEMI prognosis observed in women.

Thus, we sought to investigate potential sex disparities in the incidence of adverse LVR and its impact on clinical outcomes post-STEMI in patients reperfused by primary PCI.

Methods

Study population

The study included 1720 patients with a first STEMI who underwent a pre-discharge CMR study. These patients were part of a multicentre registry combining data from three ongoing registries conducted at three teaching Hospitals, of which the results have been published previously.^20–23 To evaluate adverse LVR as a potential mechanism of cardiovascular events, the study included only those patients from the three registries with two consecutive CMRs at least 6 months apart (n = 1067). The exclusion criteria comprised patients who experienced death (n = 18), reinfarction (n = 28), or severe clinical instability during admission (n = 9) before the follow-up CMR. Additionally, patients with unavailable or incomplete CMR studies, inadequate image quality, or any contraindications to CMR were excluded (including severe claustrophobia, existing non-conditional CMR pacemakers or implantable cardioverter-defibrillator, patient's refusal, and prior contrast reactions).

Baseline characteristics were collected prospectively in all cases. The primary PCI technique was left to the discretion of the interventional cardiologists. Patients were managed both in-hospital and after discharge at the discretion of the corresponding clinician. However, clinicians were encouraged to adhere to STEMI management guidelines.²⁴

The investigation conforms with the principles of the Declaration of Helsinki, and it was approved by the respective local Ethics Committees on Human Research PR(AG)412/2019. Patient informed consent was obtained. Additionally, the same Committees approved compiling the three databases in an extensive multicentre registry. The data supporting this study's findings are available from the corresponding author upon reasonable request.

The primary endpoint was the occurrence of adverse LVR, defined as an increase of >15% in LV end-diastolic volume (EDV) and a decrease of >3% in LV ejection fraction (EF) within 6 months post-STEMI as previously defined.²⁰ The secondary endpoint was the presence of major adverse cardiac events (MACEs) described as a composite variable: cardiovascular mortality, hospitalization due to heart failure (HF), or malignant ventricular arrhythmias. As mentioned, MACE events were collected following the 6-month CMR scan, excluding patients who died before the follow-up. This approach was used to avoid the effect of a second infarction on remodelling and exclude patients who could not undergo a second CMR. Other definitions of LVR, such as the presence of an LVEDV increase of > 20% or >15%,^17,25 have also been evaluated, and their results are described in Supplementary data online, Table S1.

Cardiovascular magnetic resonance

All CMR studies were performed with a 1.5-T clinical scanner (Sonata or Avanto scanner, Siemens, Erlangen, Germany) or 3-T (Trio, Siemens, Erlangen, Germany or Signa Architect, General Electric, Chicago, IL, USA). Further details on the technical aspects of CMR acquisition, sequences, and quantification can be found in the Supplementary Material (Section 1.1).

The following measurements were estimated: LVEF (%), LVEDV index (mL/m²), left ventricular end-systolic volume (LVESV) index (mL/m²), LV mass index (g/m²), IS (% of LV mass), MVO (number of segments with MVO), myocardial edema or myocardial AAR (% of LV mass), and myocardial salvage index (MSI; % of LV mass). Cardiovascular magnetic resonance studies were analysed offline by an experienced observer blinded to all patient data using customized software (QMASS MR 6.15, Medis, Leiden, the Netherlands; or Cvi42, Circle Cardiovascular Imaging Inc., Calgary, Alberta, Canada). See Supplementary Material (Section 1.2).

Statistical analysis

Continuous variables were expressed as mean ± standard deviation (SD) or median (IQR), as appropriate for their distribution. Differences between groups for continuous parameters were assessed by Student's t-test when assuming a normal distribution. Mann–Whitney U test was used when this assumption was not met. Categorical variables were described as absolute values and percentages. Chi-square and Fisher's exact tests were used as appropriate for the comparison between categorical variables.

Univariable and multivariable logistic regression analyses were used to assess the association of sex with adverse LVR, adjusting for potential confounding clinical variables. We first evaluated the crude association between all clinical and CMR variables with the presence of LVR. A robust, non-parsimonious multivariable logistic regression model included all variables demonstrating a statistically significant association in the univariable analysis. Then, sex was also included to assess the adjusted association with LVR. Estimates of the associations between the different variables and LVR are presented as odds ratios (OR) and a 95% confidence interval (CI).

Survival analyses were employed to assess the association between sex and MACE. Kaplan–Meier curves and log-rank tests were used to compare the survival patterns across different subgroups of patients. Cox regression modelling was used to estimate the adjusted association between sex and MACE. Again, we assessed the crude association between all clinical and CMR variables with MACE. A robust, non-parsimonious multivariable Cox model included all variables demonstrating a statistically significant association in the univariable analysis. Thus, sex was also included in this non-parsimonious model to assess the adjusted association with MACE. Estimates of the associations between the different variables and MACE are presented as hazard ratios (HR) and a 95% CI.

All analyses used STATA version 15.1 (StataCorp, TX, USA). All tests were two-sided, and a P-value of <0.05 was considered statistically significant.

Results

Among 1720 patients with a first STEMI who underwent primary PCI, 653 patients were excluded because a second CMR was not performed at 6 months per protocol. Also, patients presenting cardiovascular events before the second CMR (18 patients who died and 28 patients due to reinfarction) were excluded. An analysis of how the exclusion of these patients impacts the study outcomes is detailed in Supplementary data online, Table S2. This table shows no differences in baseline characteristics between the two groups (population excluded, n = 653, vs. population with a follow-up CMR, n = 1067), except for slightly older age and incidence of dyslipidemia in patients excluded. However, there were no differences in the infarct location or the reperfusion time. Therefore, 1067 patients (187 women; 17%) were included in the analysis (Flowchart, Figure 1).

The baseline data of the study population are shown in Table 1. Women were older and had a significantly higher incidence of diabetes, hypertension, and more delayed time-to-reperfusion compared with men. However, smoking was more common in men. No other significant differences were observed between sexes regarding angiographic, procedural characteristics, Killip class, or medical treatment at discharge (Table 1).

Patients in this study underwent several forms of acute reperfusion following STEMI. Most (94.1%) were treated with primary PCI, including facilitated PCI. A small proportion (0.8%) received no revascularization, which was the only significant difference observed between groups, likely due to the higher incidence of spontaneous coronary artery dissection (SCAD) in women (P = 0.001). Additionally, 2.0% of patients received thrombolysis, 2.9% required rescue PCI following unsuccessful thrombolysis, and 0.2% underwent coronary artery bypass grafting (CABG). Among patients who were revascularized, there were no differences in the type of revascularization received between sexes (Table 1).

The median time between the primary PCI procedure and the first CMR was 6 days (IQR: 4–9 days) for both sexes. After discharge, a second CMR was performed at a median time of 6.47 months (IQR: 5.98–7.47 months), and patients were followed up for a median of 6.94 years (IQR: 4.48–9.32 years).

Incidence of adverse left ventricular remodelling

Area at risk and IS did not differ significantly between sexes, and no differences in MSI were observed (Table 1) in the baseline CMR. However, the occurrence and extent of segments with MVO were higher in males (1.39 vs. 0.85, P < 0.001) despite longer ischaemic time and worse cardiovascular risk profile in women. No differences in LVEF at baseline were observed between sexes. After 6 months, 82 patients (7.7%) experienced adverse LVR with no differences in men and women (8.0% vs. 6.4%, P = 0.473). We investigated alternative definitions of LVR following STEMI, including criteria such as a > 20% or >15% increase in LVEDV from baseline, without observing any sex differences (see Supplementary data online, Table S1). Additionally, while LVEDVi increased similarly in both women and men (2.2 vs. 2.3 mL/m², P = 0.939), there was a significantly greater reduction in LVESVi between baseline and 6 months in women compared to men (−3.0 vs. −0.7 mL/m², P = 0.028; see Supplementary data online, Table S3).

Following univariable analysis (see Supplementary data online, Table S4), we adjusted a non-parsimonious multivariable model. This analysis revealed that only the Killip II-IV on admission (OR = 1.75; 95% CI, 1.00–3.08; P = 0.049) and IS (OR = 1.02; 95% CI, 1.01–1.04; P = 0.024) remained associated with adverse LVR after adjustment. The association between sex and adverse LVR was not statistically significant in the multivariable model (Figure 2).

Sex and clinical outcomes (MACE)

After a follow-up of 6.94 years (IQR: 4.48–9.32 years), women had a significantly higher incidence of MACE compared to men [42 events (22.5%), incidence rate 3.35 events per 100 patients per year (PYs) vs. 135 events (15.4%), incidence rate 2.10 per 100 PYs, P = 0.011]. The increased risk of MACE in women was primarily driven by a higher risk of HF hospitalization (2.26 events per 100 PYs vs. 1.10 events per 100 PYs; P = 0.002, respectively). No significant differences were found in terms of cardiovascular mortality (1.55 events per 100 PYs vs. 1.16 per 100 PYs; P = 0.243, respectively) or ventricular arrhythmias (0.00 events per 100 PYs vs. 0.14 per 100 PYs; P = 0.197, respectively; Figure 3 and Supplementary data online, Table S5).

The presence of adverse LVR was associated with a higher incidence of MACE [29 (35.80%) vs. 148 (15.15%); log-rank P < 0.001]. Men with adverse LVR had a significantly higher cumulative event rate compared to those without adverse LVR [21 (30.43%) vs. 114 (14.16%); log-rank P = 0.001]. Similarly, significant differences were found between women with and without adverse LVR in terms of cumulative event rates [8 (66.67%) vs. 34 (19.54%); log-rank P < 0.001; Figure 4].

In a univariable analysis (see Supplementary data online, Table S6), several clinical and CMR imaging characteristics were significantly associated with MACE (age, sex, diabetes mellitus, hypertension, Killip ≥ II, time to reperfusion from symptom onset, culprit vessel, LVEF at admission, IS, and number of segments with MVO). However, in a non-parsimonious multivariable model including all these variables, only age at admission [hazard ratio (HR): 1.08; 95% CI, 1.06–1.09; P < 0.001], LVEF at admission (HR: 0.98; 95% CI, 0.97–1.00; P = 0.028), and the number of segments with MVO (HR: 1.16; 95% CI, 1.06–1.28; P = 0.002) remained significantly associated with MACE. After adjustment in the multivariable model, the female sex was not associated with MACE (HR: 1.21; 95% CI, 0.81–1.81; P = 0.343; see Table 2). To confirm that sex alone was not a predictor of MACE, a multivariable analysis was conducted in the unselected cohort of 1720 patients, using sex as the primary independent variable and MACE as the dependent variable. The analysis showed no significant association between sex and MACE (HR: 1.23; 95% CI, 0.88–1.73; P = 0.225), indicating that sex did not influence primary or secondary outcomes in this study.

Discussion

In this prospective multicentre study, we documented that the incidence of post-STEMI adverse LVR was not influenced by sex. Although women had worse clinical outcomes after adjusting for other factors, such as clinical and CMR variables, female sex was not associated with MACE. Remarkably, our population was comprised of unselected and consecutive patients with a first STEMI, mainly reperfused by primary PCI (97.0%).

To date, conflicting data exist regarding sex differences in LV post-infarct remodelling. While some authors describe a more significant occurrence in women,^25–27 other reports suggest a higher frequency of adverse LV post-infarct remodelling in males,^28,29 driving a more favourable LV post-infarct remodelling process in women.²⁹ In our study, Killip class ≥ II and IS were the only factors associated with the adverse LVR appearance in the overall population, but the influence of MVO was not evident. This lack of significance of MVO in adverse LVR occurrence could be attributed to the number of affected segments rather than the overall mass (extent as a % of left ventricular mass). However, post-ischaemic LVR is a complex multifactorial process in which myocardial hypertrophy, fibrosis, ventricular dilation, apoptosis, and autophagy at the cellular level play a complex role.^30,31

Furthermore, we found no sex differences in the IS nor AAR, although there was a reperfusion delay between sexes, as reported in previous studies.^32,33 Some factors could justify this protective effect in females. First, oestrogen has been suggested to have a protective effect on the heart that may contribute to better preservation of LV function and smaller IS, leading to less late remodelling.^16,29,34 However, this was not the case in our study because women had a mean age of 63 years, similar to previous studies.^1,19,32,35 Finally, MVO has also been associated with LV post-infarct remodelling,^36,37 and we found that women have less MVO compared to men, which also agrees with previous studies.^16,19,38

In our study, women had higher unadjusted rates of admission due to HF than men. However, a multivariable analysis revealed that these differences were likely because of baseline cardiovascular risk factors disparities. Women with STEMI were older and had a higher incidence of hypertension and diabetes, consistent with previous studies investigating sex differences in STEMI.^1,9,19,35 As expected by those risk factors, women had a worse prognosis than men, particularly concerning admission due to HF, despite no differences were found in terms of receiving guideline-medical-directed therapy in both sexes, including rates of primary PCI, which also showed no significant differences, as supported by previous studies.^{6–8,10,11,14} Although the need to explore the presence of LVR over basal CMR parameters seems controversial,²⁰ given that HF is the most common MACE and there is a known link between HF and LVR, we wanted to include this variable in the analysis. For this purpose, we excluded patients who did not have a second CMR in the clinical study and patients who experienced an event (excluded due to death, reinfarction, or clinical instability) between the two CMR studies (see Supplementary data online, Tables S2 and S7). This exclusion strategy allowed us to concentrate on a more homogeneous population, thereby enhancing the reliability of our conclusions regarding the prognostic significance of LVR. Despite this, the statistical analysis of the overall population does not show significant differences compared with the selected population, as seen in Supplementary data online, Table S2, where there are no differences in the percentage of sex, infarct location, or reperfusion time.^39,40

The role of a second CMR in STEMI risk stratification remains controversial, as Masci et al.³⁹ found its timing may not impact prognostic implications for MACE. Traditional volumetric definitions of remodelling may overlook critical prognostic factors, while alternative definitions have shown correlations with worse outcomes. Frantz et al.⁴⁰ highlight that LVR is key in assessing morbidity and mortality. Our study, with a larger sample than Masci's, provides a more robust analysis of remodelling's prognostic value post-MI. Furthermore, while Van der Bijl et al.³⁵ found no sex-based differences in LV remodelling, our use of advanced CMR and a more prognostically relevant definition offers new insights into potential sex disparities in adverse remodelling and outcomes post-STEMI.

In summary, in our multicentre study of patients with a first STEMI treated by primary PCI, the incidence of post-infarction adverse LVR was similar between men and women. Although women present worse clinical outcomes than men, this cannot be explained by differences in adverse LVR but by baseline risk factors and clinical characteristics.

Study limitations

In accordance with earlier studies in STEMI patients,^17,41,42 women were underrepresented in our work,^1,29 reflecting the known challenges and barriers to female participation in RCTs. Essentially, our population primarily consists of participants from clinical trials, where, as commonly known, fewer women usually participate. Additionally, the number of women who developed an adverse LVR was lower than in other studies,^16,35 possibly due to a more stringent definition. However, we have also conducted the analysis using other classical definitions of remodelling and found no differences between the sexes (see Supplementary data online, Table S1).

Furthermore, it should be noted that CMR at 6-month follow-up could not be conducted in all patients, introducing a potential risk of selection bias. In this sense, although the baseline characteristics of these patients were mainly comparable to those who underwent the second CMR, the mean age was higher in the cohort who underwent only one CMR (see Supplementary data online, Table S4). Although this could have introduced a selection bias in the sense of an over or underestimation of LVR incidence, it does not likely affect the main conclusions of the study since age was not associated with LVR.

Additionally, patients excluded due to death, reinfarction, or clinical instability between the first and second CMR scans may represent a more severe subset, potentially impacting the study results, had these been included. Therefore, these patients were excluded from the analysis to maintain the integrity and reliability of the findings. However, there are no differences in sexes between these two populations (P = 0.638, Supplementary data online, Table S7), so we do not expect this exclusion to impact the main results of our study.

Conclusions

The incidence of adverse LVR is comparable between men and women in STEMI patients treated by primary PCI using a comprehensive CMR approach. However, women experienced worse clinical outcomes than men (mainly due to hospitalization for HF), which may be more closely related to a higher prevalence of risk factors and comorbidities rather than differences in adverse LVR.

Clinical perspectives

This study found no gender difference in adverse LVR in patients post-STEMI. This suggests that the higher rate of MACE after STEMI in women compared with men appears to be associated with a higher prevalence of CVRF and comorbidities rather than a greater occurrence of adverse LVR.

To address the observed gender-based variations in clinical outcomes, a tailored approach is required, focusing on the mitigation of cardiovascular risk factors and customizing interventions specifically for older and comorbid female patients.

Future studies should focus on creating gender-specific risk assessment tools to improve prognosis accuracy in women after STEMI by investigating novel, sex-specific biomarkers since traditional models may not predict outcomes effectively.

Supplementary data

Supplementary data are available at European Heart Journal - Cardiovascular Imaging online.

Funding

V.B. acknowledges financial support from ‘Instituto de Salud Carlos III’ and ‘Fondo Europeo de Desarrollo Regional FEDER’ (grant numbers PI20/00637, PI23/01150, CIBERCV16/11/00486), by Conselleria de Cultura Educacion y Ciencia – Generalitat Valenciana (PROMETEO/2021/008) and by Sociedad Española de Cardiología (grant SEC/FEC-INV-CLI 21/024). J.T.O.-P. acknowledges partial financial support from Fundació La Marato de TV3 (grant number 20153030-31-32), Siemens Healthcare, and La Caixa Banking Foundation (HR17-00527).

Data availability

The corresponding author will share the article's data at a reasonable request.

References

Author notes