Gender disparities in ST-segment elevation myocardial infarction (STEMI) outcomes continue to be reported worldwide; however, the magnitude of this gap remains unknown. To evaluate gender-based discrepancies in clinical outcomes and identify the primary driving factors a global meta-analysis was performed. Studies were selected if they included all comers with STEMI, reported gender specific patient characteristics, treatments, and outcomes, according to the registered PROSPERO protocol: CRD42020161469. A total of 56 studies (705,098 patients, 31% females) were included. Females were older, had more comorbidities and received less antiplatelet therapy and primary percutaneous coronary intervention (PCI). Females experienced significantly longer delays to first medical contact (mean difference 42.5 min) and door-to-balloon time (mean difference 4.9 min). In-hospital, females had increased rates of mortality (odds ratio [OR] 1.91, 95% confidence interval [CI] 1.84 to 1.99, p <0.00001), repeat myocardial infarction (MI) (OR 1.25, 95% CI 1.00 to 1.56, p=0.05), stroke (OR 1.67, 95% CI 1.27 to 2.20, p <0.001), and major bleeding (OR 1.82, 95% CI 1.56 to 2.12, p <0.00001) compared with males. Older age at presentation was the primary driver of excess mortality in females, although other factors including lower rates of primary PCI and aspirin usage, and longer door-to-balloon times contributed. In contrast, excess rates of repeat MI and stroke in females appeared to be driven, at least in part, by lower use of primary PCI and P2Y12 inhibitors, respectively. In conclusion, despite improvements in STEMI care, women continue to have in-hospital rates of mortality, repeat MI, stroke, and major bleeding up to 2-fold higher than men. Gender disparities in in-hospital outcomes can largely be explained by age differences at presentation but comorbidities, delays to care and suboptimal treatment experienced by women may contribute to the gender gap.
Worldwide, acute myocardial infarction is the leading cause of death for males and females. Over the past 2-decades, worldwide initiatives to increase public awareness have led to the implementation of reliable medical systems and generated guidelines to reduce morbidity and mortality associated with ST-segment elevation myocardial infarction (STEMI). As the care for STEMI patients has continued to improve globally, gender disparities in quality of care and outcomes have become more apparent. While some studies have showed that the higher mortality in women is largely due to differences in age, comorbidities, treatment strategy, and reperfusion delays, others have found that the differences persist despite adjustment for these variables, especially in younger patients. Whether the gender disparities persist at a global level with the current widespread adoption of care systems for STEMI patients remains unknown. The aim of this global meta-analysis was to identify differences in patient characteristics, delays to care, treatment strategies, and outcomes by gender, in order to raise awareness of gender disparities and to inform future initiatives for improvement of care.
Methods
This meta-analysis was performed following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement guidelines and was prospectively registered in PROSPERO (CRD42020161469). PubMed was searched using the following search terms: STEMI, sex, gender, mortality, and outcome. The search was restricted to studies published in the English language between January 2000 and December 2019. Two authors (TS and SG) independently screened titles/abstracts for mention of patient characteristics and gender-based STEMI outcomes to identify eligible articles. The studies were selected if they included all-comer STEMI patients and reported patient characteristics and relevant prespecified clinical endpoints by gender. Studies were excluded if they reported data for only thrombolysis therapy, included less than 100 patients of any gender, or only patients from certain cohorts of STEMI patients (e.g., diabetics, cardiogenic shock, or narrow age ranges), or based on major shortcomings (e.g., improper data, different endpoints or timepoints). When overlapping data were identified only the study with the most relevant data was selected and the other publications were discarded. Bibliographies of selected studies were examined to identify potentially relevant studies.
Three authors (TS, SG, and RT) independently abstracted data on prespecified patient demographics/characteristics, delays to care, treatment strategies and outcomes from selected studies and crosschecked for accuracy. Data on treatment strategies included proportion of primary PCI (as the initial treatment strategy before fibrinolysis or GpIIb/IIIa inhibitor usage) and medication usage. Prespecified delays to care included time to first medical contact (FMC), door-to-balloon (DTB), and door-to-needle (DTN). Prespecified clinical outcomes included in-hospital mortality, repeat MI (site defined), stroke (site defined), and major bleeding (definition varied by study, generally including bleeding requiring transfusion or repeat procedure).
Odds ratios were calculated using the DerSimonian and Laird inverse variance random-effects model and were performed using the Review Manager version 5.3 (Nordic Cochrane Center, Copenhagen, Denmark). The I 2 statistic and heterogeneity p-value were used as a measure of variability in observed effect estimates attributable to heterogeneity between studies. For the I 2 statistic, heterogeneity was defined as low (25% to 50%), moderate (50% to 75%), or high (>75%). Estimates for pooled analyses per geographic regions and combined were displayed in Forest plots. R version 3.5.2 (R Foundation for Statistical Computing, Vienna, Austria) was used for demographic and clinical characteristics analyses. Two sample t -test was used for analyzing the continuous variable of age, expressed as mean and standard deviation. Categorical variables were evaluated by two-sample test for equality of proportions and reported as percentages. Simple and multiple logistic regression models were conducted using SAS software version 9.4 (SAS Institute Inc., Cary, North Carolina) to identify significant moderator variables.
Results
The literature search resulted in a total of 1,234 articles. After screening and eligibility assessment for inclusion criteria a total of 56 studies were selected and included in the global meta-analysis ( Figure 1 ). A total of 705,098 patients (31% females) from more than 30 countries, grouped into 6 regions: Asia, Australia, Europe, Middle East, mixed regional population, and North America were included in the meta-analysis ( Supplemental Table 1 ). The mixed regional population included patients for which regional separation was not possible. The studies consisted mostly of observational studies and registries, but also included one randomized controlled trial ( Supplemental Table 2 ).
The pooled baseline characteristics by gender demonstrated that females with STEMI were older than males (70.2 ± 3.1 vs 61.1 ± 2.2 years) and were more likely to have diabetes mellitus (27.4% vs 21.0%), hypertension (61.1% vs. 50.6%), a prior stroke (8.1% vs 7.4%), and cardiogenic shock at presentation (6.9% vs 5.5%) ( Table 1 ). Males were more likely to be smokers (44.2% vs 27.2%) and to have had a prior MI (16.2% vs 14.7%) ( Table 1 ).
| Variable | Women | Men | P-value | ||
|---|---|---|---|---|---|
| N | % of N | N | % of N | ||
| Mean Age (SD) | 187,764 | 70.2 (3.1) | 422,202 | 61.1 (2.2) | <0.0001 |
| Diabetes mellitus | 214,320 | 27.4 | 478,751 | 21.0 | <0.0001 |
| Hypertension | 200,018 | 61.1 | 438,495 | 50.6 | <0.0001 |
| Smoker | 164,239 | 27.2 | 360,768 | 44.2 | <0.0001 |
| Prior MI | 156,869 | 14.7 | 339,566 | 16.2 | <0.0001 |
| Prior stroke | 137,121 | 8.1 | 302,162 | 7.4 | <0.0001 |
| Cardiogenic shock | 156,510 | 6.9 | 344,128 | 5.5 | <0.0001 |
| Treatment | |||||
| Primary PCI | 213,296 | 59.5 | 478,335 | 68.2 | <0.0001 |
| Aspirin | 83,441 | 89.5 | 217,489 | 92.1 | <0.0001 |
| P2Y12 inhibitor | 73,926 | 67.6 | 193,405 | 75.4 | <0.0001 |
| GpIIb/IIIa inhibitor | 71,929 | 22.7 | 162,396 | 29.3 | <0.0001 |
| Beta Blocker | 79,099 | 75.1 | 202,210 | 76.1 | <0.0001 |
| ACE inhibitor | 55,072 | 55.6 | 144,422 | 59.4 | <0.0001 |
Gender differences in time to FMC and DTB were calculated for regions with available data. Time to FMC, determined for Asia, Australia and Europe was significantly longer for females, both per regions and combined, with a mean delay of 42.5 min (95% CI 28.4 to 56.6, p <0.00001) ( Supplemental Figure 1A ). In a combined analysis, DTB time was significantly longer for females with a mean delay of 4.9 min (95% CI 3.8 to 6.1, p <0.00001) ( Supplemental Figure 1B ). However, there was a high variability in reported DTB time per regions with longer delays, more than 5 min for females in Asia, Australia, North America and the mixed region. Shorter delays were in Europe (mean 2.1 min) and in the Middle Est (mean 1.1 min).
While it was not feasible to determine gender differences in DTN time and total ischemic time (defined from the symptom onset to balloon time) per regions due to limited data, it was possible by pooling all available data, regardless of region. As shown in Supplemental Table 3 , there were significant delays for females in DTN times (mean 0.9 min) and ischemic times (mean 21.3 min).
Female patients with STEMI received less optimal therapy during hospitalization, compared with their male counterparts, including: primary PCI (59.5% vs 68.2%), aspirin (89.5 vs 92.1%), P2Y12 inhibitors (67.6% vs 75.4%), GpIIb/IIIa inhibitors (22.7% vs 29.3%), beta blockers (75.1% vs 76.1%) and ACE inhibitors (55.6% vs 59.4%) ( Table 1 ). Of note, patients not receiving primary PCI often received PCI at some point during their hospital stay in most included studies. The proportion of primary PCI varied significantly across regions with >80% of patients from studies in North America and Europe receiving primary PCI while only 50% to 60% received primary PCI in studies from Asia and the Middle East.
In combined analysis, the unadjusted rate of in-hospital mortality was higher in females compared with males (N=669,358; OR 1.91, 95% CI 1.84 to 1.99, p <0.00001, I 2 =58%) ( Figure 2 ). Consistently, females had higher in-hospital mortality rates across all regions (OR varying from 1.54 to 2.57, p <0.00001, I 2 varying 0-65%) ( Figure 2 ). Furthermore, gender disparities in mortality rates did not significantly change in the last 20 years, since the implementation of care systems for STEMI ( Supplemental Figure 2 ).
For the other in-hospital outcomes, repeat MI, stroke and major bleeding, significant differences between genders were not identified in all regions, although the combined results showed worse outcomes for females. The unadjusted rate of repeat MI for females was higher compared with males in combined analysis (N=70,408; OR 1.25, 95% CI 1.00 to 1.56, p = 0.05, I 2 =57%), however there were no significant differences between females and males per regions ( Figure 3 ). The risk of stroke was higher in females compared with males (N = 60,881; OR 1.67, 95% CI 1.27 to 2.20, p <0.001, I 2 =41%) with significant differences disfavoring females in the Middle Est and North America ( Figure 4 ). The rate of major bleeding was higher in females (N = 208,201; OR 1.82, 95% CI 1.56 to 2.12, p <0.00001, I 2 =80%) with significantly increased rates in Asia, Europe, mixed region and North America ( Figure 5 ).
