The Italian Genetic Study on Early-onset Myocardial Infarction enrolled consecutive patients from 125 Italian Coronary Care Units between 1998 and 2002. Patients under the age of 45 years who had undergone coronary angiography during hospitalization for their index MI were eligible for participation. The initial diagnosis of the index event was based on a combination of: (1) symptoms suggesting myocardial ischemia; (2) electrocardiographic changes consistent with acute myocardial ischemia; and (3) an increase in CK-MB levels to more than twice the upper limit of normal with a typical rise and fall pattern.

The protocol was approved by the Ethics Committee of the coordinating center, and all of the study patients gave their written informed consent to enrolment. A standardized case report form (CRF) was completed for every patient suitable for enrolment in order to collect detailed information concerning their cardiovascular risk factors, lifestyle habits and medications, the history of any cardiovascular diseases affecting each member of their families (including first- and second-degree relatives), and the patients' own clinical, demographic and treatment characteristics. The definitions of all of these factors are described in detail in Supplementary Material 1.

The patients were followed up in order to track subsequent hospital admissions for cardiovascular or ischemic cerebrovascular events. The occurrence of these endpoints was assessed by means of scheduled outpatient visits and standardized telephone follow up. Every attempt was made to contact all of the patients or their primary care physicians; if this was not possible, a member of their family was contacted. If an event was reported by a patient, relative or general practitioner, the related medical record(s) were obtained.

All of the patients underwent coronary angiography at the time of the index event. Coronary arteries were considered normal in the absence of angiographic stenosis: a narrowing of <70% (or <50% in the case of the left main coronary artery) was considered non-significant coronary artery stenosis, and a narrowing of > 70% (or >50% in the case of the left main coronary artery) was considered significant coronary artery stenosis. Single-vessel disease was recorded when a significant stenosis was identified in only one major coronary artery, and multi-vessel disease when a significant stenosis was identified in two or more major epicardial coronary arteries. Left main coronary artery disease was considered multi-vessel disease. The total extent and complexity of the angiographic coronary artery disease burden was quantified using the Duke Coronary Artery Disease Index (14) and the Syntax score (15).

The primary composite endpoint was the occurrence of cardiovascular death, the re-occurrence of non-fatal MI, or the occurrence of a non-fatal ischemic stroke. All of the reported deaths were verified on the basis of death certificates specifying the cause(s) of death, and each reported event was investigated by means of source data verification. The outcome of cardiovascular death was defined as any death attributed to a cardiovascular cause on the death certificate. The re-occurrence of MI was defined as any hospitalization during follow-up ending with a discharge diagnosis of MI. The occurrence of stroke was defined as hospitalization ending with a discharge diagnosis of ischemic stroke.

The secondary study endpoint was any hospital admission that involved revascularisation by means of a percutaneous coronary intervention (PCI) or coronary artery by-pass graft (CABG) but did not end in a diagnosis of MI.

All of the documented events with verified source data were adjudicated by a dedicated Clinical Events Committee consisting of two cardiologists; in the case disagreement, the opinion of a third cardiologist was sought.

The study's clinical data were recorded using an off-line electronic CRF (e-CRF). The demographic, clinical, angiographic and treatment characteristics of all patients were collected at the time of the index event and recorded in the e-CRF along with events ascertained during follow up. Visual Basic for Applications (VBA) and SQL scripts were used to manage the database and make preliminary data checks during data entry. In the case of missing data, queries were generated and sent to study sites for resolution.

The study cohort was divided into females and males, and descriptive statistics were used to compare the baseline demographic and clinical characteristics of the two groups. Univariable logistic regression analysis was used to evaluate the relationship between covariates and outcomes. Variables included age, weight, height, body mass index, type of acute coronary syndrome at the time of the index event (non-ST elevation MI [NSTEMI] vs. ST elevation MI [STEMI]), family history of coronary artery disease (CAD), hypertension, dyslipidemia, diabetes, alcohol consumption, smoking, physical activity, cocaine use, estrogen replacement therapy, previous thromboembolic events, ejection fraction, pattern and extent of angiographic disease burden (Duke Coronary Artery Disease Index, Syntax score, presence of multivessel disease), presence and mode of coronary artery revascularisation, and pharmacological treatments. The variables with a p-value of <0.2 were selected for inclusion in a comprehensive logistic regression model of long-term outcomes that was subsequently simplified by means of the step-wise backward selection of independent predictors, which were then used to make an adjusted proportional hazard analysis of long-term outcomes.

The time to the first endpoint and the time to all endpoints by type were analyzed for their dependence on their putative predictors using Cox proportional hazard models. The proportional hazard assumption for the Cox regression model was confirmed by means of Schoenfeld's residuals test. The cumulative incidence of primary endpoints during follow-up was graphically depicted using Aalen-Johansen curves, and the significance of the differences between the hazard sub-distributions was tested using the Fine-Gray model. An Andersen-Gill intensity model analysis was not pre-specified, but was made in order to account for the repeated occurrence of all of the components of the primary and secondary endpoints during the study period using a time-dependent model.

All of the tests were two-sided, and a p-value of 0.05 was considered statistically significant. Given the consecutive nature of the enrolment, there was an uneven distribution of patients across sexes. For this reason, a post-hoc power calculation was performed based based on the number of patients by sex (222 females, 1,778 males), and incidence of primary events (25.7% vs. 37.0%), and alpha = 0.05. This yielded a power of 93.3%. The statistical analyses were made using R Statistical software, version 3.6.0 (R Foundation for Statistical Computing, Vienna, Austria).

The patients were followed up for a median 19.9 years (IQR 18.1–22.6) or 39,535 person-years. Follow-up was complete in 1,984 patients (99.2%), and vital status was ascertained in 1,988 (99.4%). A primary endpoint was experienced by 714 patients (36%): 153 died of cardiovascular causes, 479 experienced a recurrent non-fatal MI, and 82 an acute ischemic stroke. Table 2 and Figures 1A,B show the number and rate of first occurrences of the primary and secondary endpoints by sex.

During follow-up, the primary composite endpoint occurred less frequently among the women (26.1% vs. 37.2%; adjusted hazard ratio (aHR) 0.69, 95% CI 0.52–0.91; p = 0.01). Cardiovascular death was also less frequent among women although this did not reach statistical significance (4.1% vs. 8.1%, aHR 0.57, 95% CI 0.30–1.08, p = 0.082). Non-fatal myocardial re-infarction was significantly less frequent among the women (14.2% vs. 25.4%; aHR 0.53, 95% CI 0.37–0.77; p < 0.001), although women were more likely to experience a non-fatal ischemic stroke (7.8% vs. 3.7%; aHR 2.02, 95%CI 1.17–3.49; p = 0.012).

The occurrence of hospitalization for revascularisation was significantly less frequent among the women (15.6% vs. 25.5%; aHR 0.60, 95%CI 0.41–0.86; p = 0.006), driven largely by statistically less frequent PCI (9.6% vs. 15.9%; aHR 0.59, 95% CI 0.38–0.91; p = 0.018) and directionally consistent lower rates CABG (6.0% vs. 9.6%, aHR 0.61, 95% CI 0.35–1.07, p = 0.084).

The study population as a whole experienced a total of 973 composite primary events including repeated events. Table 3 and Figures 1C,D show the number of all occurrences of primary and secondary endpoints by sex. The total number of primary endpoints was significantly lower among the women (aHR 0.75; 95% CI 0.59–0.96; p = 0.02). With respect to the individual components of the composite, women were less likely to experience cardiovascular death (aHR 0.57, 95% CI 0.35–0.95; p = 0.032) or non-fatal myocardial re-infarction (aHR 0.62, 95% CI 0.45–0.86; p = 0.004). While not reaching statistical significance, there was a non-significant trend toward higher rates of ischaemic stroke in women (aHR1.6; 95% CI 0.97–2.7; p = ns) and a trend toward lower rates of admission for further revascularization (aHR 0.78; 95% CI 0.58–1.04; p = ns).

Table 4 shows the factors associated with experiencing the primary endpoint during follow-up. Multiple logistic regression analysis identified only two variables that were independently associated among women; hypertension (OR 2.56; 95% CI 1.01–6.58; p = 0.05) and a positive family history of CAD (OR 8.53, 95% CI 1.58–159, p = 0.04). Among men there were six variables with independent associations including the Duke Coronary Artery Disease Index (OR 1.01; 95% CI 1–1.01; p = 0.02), hypertension (OR 1.34; 95% CI 1.01–1.78; p = 0.04), dyslipidemia (OR 1.41; 95% CI 1.1–1.8; p = 0.01), diabetes (OR 2.1; 95% CI 1.33–3.36 p = 0.002), and previous thromboembolic event (OR 2.45; 95% CI 1.71–3.52; p < 0.001). An inverse association with the ejection fraction was noted (OR 0.98 95% CI 0.97–0.99 p = 0.006). The occurrence of a secondary endpoint was independently associated with the Duke Coronary Artery Disease Index (OR 1.02; 95% CI 1.0–1.04; P < 0.02) and diabetes (OR 6.17; 95% CI 1.2–36.5; p = 0.03) among women, and alcohol consumption (OR 0.72; 95% CI 0.55–0.93; p = 0.01), the Duke Coronary Artery Disease Index (OR 1.01; 95% CI 1.01–1.02; p < 0.001), a previous thromboembolic event (OR 1.64; 95% CI 1.15–2.34; p = 0.007), and diabetes (OR 2.88; 95% CI 1.81–4.67; p < 0.001) among men.