This retrospective observational study recruited all consecutive patients diagnosed with STEMI and treated with a primary PCI in our center from January 2014 to October 2021. Our center is the referral hospital for a population of 1 million inhabitants. Our team has more than 10 years’ experience and treats more than 250 patients annually. Exclusion criteria were age <18 years and fibrinolysis treatment.

We constructed the reference population using mortality tables provided by the National Institute of Statistics of Spain (Instituto National de Estadística [INE]). This institute publishes mortality data stratified by age ranges, sex, and autonomous regions in Spain on its official website. To compare our sample of patients with STEMI with the reference population, we matched our data with the mortality tables provided by the INE with participants of the same age, sex, geographical area, and year of the event. All participants were followed up for at least 1 year after the main event and the study ended in October 2022.9

Data on baseline characteristics, procedure, hospitalization, and discharge were obtained from a prospectively collected STEMI database. The follow-up data to confirm if patients were alive and the causes of death were collected from electronic medical records and telephone interviews. All the hospitals in our regions are connected through an intranet, and all data are readily available for consultation.

Patient treatment during admission and optimal medical therapy at discharge were selected following the latest European Society of Cardiology guidelines. Standardized endpoints and causes of death were defined according to the Academic Research Consortium-2 consensus document.10

As a convention, most studies and clinical guidelines issued by European and American societies refer to elderly patients as those older than 65 years.11,12 Because our sample was similarly divided at this age and had a sufficient number of events to power this study, we separated participants into 2 groups: a) women aged ≤ 65 years, and b) women aged> 65 years.

To compare survival in women with STEMI treated with primary PCI with that in the general population of the same age and geographical region surviving the first 30 days.

• a)

  To compare survival in women aged ≤ 65 years with STEMI treated with primary PCI and surviving the first 30 days with that in the general population of the same age and geographical region.

• b)

  To compare survival in women aged> 65 years with STEMI treated with primary PCI and surviving the first 30 days with that in the general population of the same age and geographical region.

Quantitative variables are expressed as mean±standard deviation or median [interquartile range] and were compared using the Student t test and Wilcoxon rank test, respectively. Categorical variables are expressed as numbers (percentages) and were compared using the chi-square test and, if there were several categories, the ANOVA test.

We performed a multivariable Cox regression analysis to determine the factors that predicted mortality. The variables introduced were defined using backward stepwise elimination with a cutoff of P <.10. Associations are expressed as hazard ratios (HR) with 95% confidence intervals (95%CI). The predictive capacity of the model was evaluated with the Harrel c test and survival was assessed using Kaplan-Meier curves. Statistical significance was set at P <.05.

To compare women with STEMI undergoing primary PCI with the general population of the same age and geographical area, we made the following estimates:

• 1.

  Observed survival (OS): this is the probability of surviving all causes of death in the group of women undergoing primary PCI, and is estimated using the life table method. OS is represented with a 95%CI.

• 2.

  Expected survival (ES): The survival probability of a population similar to the study sample but free from the underlying disease (STEMI in this case). In other words, ES represents the expected mortality of individuals in this sample if they had not had STEMI. In this case, we obtained the data from the INE, which provided mortality tables from individuals of the same age and geographical region. Since the entire population of the region is included, ES had no sampling error and therefore no confidence intervals were estimated.9 The Ederer II method works under the assumption that when ES is modeled, deaths from a specific cause like STEMI represent a negligible proportion of all the deaths in the reference population.13

We calculated ES with the Ederer II method, which considers individuals at risk until the corresponding patient dies or is censored.14 There is growing consensus, especially among cancer registries when the follow-up time is less than 10 years, that Ederer II is the method of choice, given that other methods tend to overestimate survival.15,16 If ES was included within the 95%CI of the OS, no differences were considered to exist.

• 3.

  Relative survival (RS): Calculation of RS was derived from the ratio of the OS in our sample of patients during a specific time interval to the ES of the reference group, which was free from the disease.14,17,18 In other words, RS is the survival expected in patients if they could only die from the main event or its consequences. In our sample, an RS of 100% would indicate that STEMI had no effect on survival in the study group. However, an RS of 90% would indicate that 10% of the patients died from the main event or its consequences. If the 95%CI of the RS included 100%, there would be no evidence of mortality due to STEMI. Therefore, the treatment provided effectively restored life expectancy to a level similar to that in the general population.

• 4.

  Excess mortality (EM): EM refers to mortality among our patients due to the main event or its consequences under the premise that they could only die from STEMI. Its calculation is 1 − RS. A 10% EM indicates that 10% of the patients died from STEMI, and 0% EM indicates that none of the patients died from the event or its consequences. If its 95%CI includes 0, no statistically significant differences are considered to exist.7,19

All analyses were performed using STATA /IC 15.1 (STATA Corp, United States). The “strs” command by Dickman et al.14 was used to estimate the OS, ES, and RS.

From January 2014 to October 2021, 2501 patients were diagnosed with STEMI and referred to our hospital. After exclusion of 307 patients because they did not undergo primary PCI, 2194 patients were included in the analysis. The mean age was 64.8±13.2 years, and 528 patients were female (23.9%); 194 (36.7%) were aged ≤ 65 years and 334 (63.3%)> 65 years. Compared with men, women were older (70.9±13.1 vs 62.8±12.7; P <.001) and had higher rates of hypertension (57.9% vs 42%) and chronic kidney disease (8% vs 5.5%; P=.036). Conversely, women had lower rates of previous myocardial infarction (8.5% vs 11.7%; P=.001) and smoking (36.1% vs 59.7%; P=.001). The remaining baseline characteristics are shown in table 1.

Significant differences were found in the procedure. Women more frequently required femoral access (34.5% vs 25.4%; P <.001), had higher rates of vascular complications (2.7% vs 1.3%, P=.026), moderate or severe valvular disease (15% vs 7.1%; P <.001), and consulted with worse hemodynamic status (Killip-Kimball ≥ II 27% vs 20.8%, P <.001). However, women less frequently had multivessel disease (37.5% vs 45%; P=.003) with a lower peak of high-sensitivity cardiac troponin T (HS-cTnT) (4676 vs 5409 pg/mL; P=.08). The remaining characteristics are shown in table 1.

Medical treatment at discharge was similar in both sexes and followed the recommendations of the latest clinical guidelines (table 1 of the supplementary data).

Almost one fifth of mortality occurred in the first 30 days. In this period, crude mortality was lower in women (7.4% vs 10.4; P=.03). After this period, mortality was more frequent in women, with 10.7% at 1 year and 26.5% at the end of the follow-up; the percentages for men were 10.7% and 19.7%, respectively. The most frequent cause of death was acute myocardial infarction, representing almost half of the events. Details on mortality and causes of death are shown in table 2 of the supplementary data.

Despite these differences in crude mortality, when adjustment was performed by age and risk factors, sex was not a predictor of death in our multivariable Cox regression analysis (HR, 0.98; 95IC%, 0.8-1.3) with an adequate model with a Harrel c index of 0.83. The factors significantly predicting death were age, chronic kidney disease, HS-cTnT level, left ventricular ejection fraction, contrast-induced acute kidney injury, and Killip-Kimball classification. Univariate and multivariable analyses are shown in table 3 of the supplementary data.

The mean follow-up was 4.1±2.5 years, with a total of 140 deaths (26.5%) in women and 329 (19.7%) in men (figure 1 of the supplementary data). Cumulative survival in women at 1, 5, and 7 years of follow-up was 88.5% (95%CI, 85.4-90.9), 75.2% (95%CI, 70.6-79.1), and 67.6% (95%CI, 61.6%-72.8) and the ES for the general population was 97%, 85.1%, and 75.6%, respectively. After exclusion of women not surviving the first 30 days, cumulative survival at 1, 5, and 7 years was 95.6% (95%CI, 93.3-97.1), 81.2% (95%CI, 76.6-85), and 73% (95%CI, 66.8-78.3) and the ES for the general population was 97.1%, 85.2%, and 76.6%, respectively. The OS and ES for all women are shown in figure 1A shows and those in women surviving the first 30 days are shown in figure 1B.

Figure 1.

A: observed mortality with 95%CI and expected survival in the reference population in women of all ages. B: observed mortality with 95%CI and expected survival in the reference population of women of all ages who survived the first 30 days. C: excess mortality in women of all ages. D: excess mortality in women of all ages who survived the first 30 days. 95%CI, 95% confidence interval.

(0.33MB).

The cumulative EM in women of all ages surviving the first 30 days was 1.6% (95%CI, 0.3-4) in the first year, 4.7% (95%CI, 0.3-10.1) in the fifth year, and 7.2% (95%CI, 0.5-15.1) in the seventh year. The EM in all women is shown in figure 1C and that in women surviving the first 30 days is shown in figure 1D. Details are shown in table 2.

Among women aged ≤ 65 years surviving the first 30 days, there was a slight cumulative EM of 0.7% (95%CI, −0.1 to 4) in the first year, 3.1% (95%CI, 0.4 to 8.9) in the fifth year, and 3.8% (95%CI, 0.2 to 10.9) in the seventh year. The EM for all women is shown In figure 2C and that in women surviving the first 30 days is shown in figure 2D. Details on OS and ES are shown in table 3.

Figure 2.

A: observed mortality with 95%CI and expected survival in the reference population of women aged ≤ 65 years. B: observed mortality with its 95%CI and expected survival in the reference population of women ≤ 65 years who survived the first 30 days. C: excess mortality in women aged ≤ 65 years. D: excess mortality in women aged ≤ 65 years who survived the first 30 days. 95%CI, 95% confidence interval.

(0.34MB).

In women aged> 65 years surviving the first 30 days, there was no cumulative EM (95%CI, −0.3 to 5.9) at 1 year, 5.7% (95%CI, −1.3 to 14.1) at 5 years, and 9.6% (95%CI, −1.49 to 22.6) at 7 years. figure 3C shows the EM for all women and figure 3D shows that in women surviving the first 30 days. Details on OS and ES are shown in table 4.

Figure 3.

A: observed mortality with 95%CI and expected survival in the reference population of women aged> 65 years. B: observed mortality with 95%CI and expected survival in the reference population of women aged> 65 years who survived the first 30 days. C: excess mortality in women aged> 65 years. D: excess mortality in women ≤ 65 years who survived the first 30 days. 95%CI, 95% confidence interval.

(0.36MB).

Because the incidence of STEMI is low in women aged ≤ 65 years, they are the most underrepresented group in clinical trials and large-scale registries. ES in the reference group after 7 years was less than 3%. Therefore, any life-changing event, like STEMI, modifies prognosis. Cumulative EM at 7 years was almost 8%. After inclusion of patients surviving the first 30 days, the OS and ES curves tended to converge but showed an EM of less than 4% by the seventh year of follow-up. Of interest, mortality was lower in the first 4 years, at less than 2.5%, and there was a tendency toward higher EM after the fifth year. These results are similar to those of modern-era registries and analyses showing scarce mortality after the implementation of emergent reperfusion therapies and guideline-directed secondary prevention.23,24

The findings in women aged ≤ 65 years are arguably the most important of our study. Because of the potential years of life lost in younger patients and the detrimental effect of a life-changing event such as STEMI, it is crucial to compare mortality in these patients with that in a reference population in order to determine the most vulnerable periods to intensify treatments and to instigate early attention. In this sample of patients, the first 30 days carried a dismal prognosis, while 30-day survivors showed a peak in mortality in the fifth year.

Mortality was higher in this group than in younger patients, with almost 50% mortality during follow-up. Mortality in the reference group was also elevated, with an ES of 62.8% during this period. Among survivors of the main event, the OS and the ES curves converged from the first year of follow-up. When we analyzed EM from the first year, the 95%CI included 0, and no statistical differences were considered to exist. These noteworthy results could indicate that life expectancy in this sample of women is similar to that in the general population, as if STEMI had no effect on survival. After the first year, RS remained stable until the end of the follow-up.

Like other reports, this study found that crude mortality was higher in women than in men. However, when we included age and risk factors, sex did not appear to be a predictor of mortality in our sample. Similar results were reported by Cenko et al.5 for short-term mortality and by Piackova et al.25 for long-term mortality, in which, in absolute numbers, sex was associated with mortality. Nevertheless, after correction for age and comorbidities, this difference disappeared.

This observational retrospective analysis is only applicable to the study population from a single center of a region in Spain. Caution should be exercised when extrapolating its results. The study was powered for the primary endpoint and when we stratified the sample by women aged ≤ and> 65 years, the sample size was relatively small. However, a factor to consider is the low incidence of STEMI in women.