The study included 2772 consecutive patients admitted to the CNR Institute of Clinical Physiology, Pisa, Italy due to IHD over a 12-year period (from 1995 to 2007), who fulfilled the following criteria: sudden unheralded MI without history of previous AP (group MI), patients with chronic AP (group AP), without history of previous MI or congestive heart failure; documentation of transient myocardial ischemia; age < = 70; submitted to coronary angiography at the time of hospitalization; available follow-up information. At discharge, all demographic, history, clinical and instrumental data were collected in the Institute’s dedicated cardiovascular database- IMAGE [13]. For this study, data on the ‘onset type’ of IHD (which could precede hospitalization in the Institute and thus enrollment in the present study, even years before), and its progression features, risk factors, clinical features at the time of hospitalization and diagnosis. The definition of chronic AP was based on the ACC/AHA clinical statement and documentation of inducible ischemia on ECG and/or imaging stress test. Traditional risk factors were defined for each patient: basal obesity: BMI > 30; diabetes mellitus: fasting plasma glucose > 126 mg/dl out of treatment; arterial hypertension: systolic blood pressure > 140 mmHg and/or diastolic pressure > 90 mmHg out of treatment; hypercholesterolemia: fasting plasma total cholesterol > 200 mg/dl out of treatment, hypertriglyceridemia: fasting plasma triglycerides > 160 mg /dl out of treatment. Coronary artery disease (CAD) was defined by the number of major coronary vessels affected by > 50% lumen reduction (left main coronary branch being considered equivalent to two vessels) at invasive coronary angiography. Patients with normal coronary vessels were included when diagnosis of angina and ischemia at rest was present at discharge, and /or vasospastic angina was suspected and /or documented. Patients presenting with unstable angina as well as patients with both MI and angina were excluded. Moreover patients who had congenital heart disease, hypertrophic cardiomyopathy, or significant valvular heart disease other than mitral insufficiency thought to be secondary to IHD were discarded.

A total of 1419 patients were selected in group MI (1101 STEMI, 318 NSTEMI) and 1353 in group AP with an even distribution (51 and 49% respectively).

For each patient, follow-up began at discharge and was planned for a minimum period of 10 years. For the purposes of the present study it was concluded in December 2013. Follow-up data were obtained in at least one of the following ways: from the patient’s hospital record, by contacting the patient’s physician, and by telephone interview conducted by trained personnel, during periodic scheduled visits at the outpatient clinic. The only endpoints considered were major events such as non-fatal MI and death classified as cardiac death and non-cardiac death. The diagnosis of MI during the follow-up period required a hospital discharge chart documenting and defining, when possible, the acute event (Q or non-Q MI and site). Medical records or death certificates were used to obtain cause of death. The documentation of life-threatening arrhythmias, myocardial infarction or cardiac arrest, or death attributable to congestive heart failure was required to make diagnosis of cardiac death, in the absence of any other precipitating factor. Coronary revascularization procedures during the follow-up were also reported.

Student’s t-test or analysis of variance as appropriate was used to test for significant difference in quantitative measures and the χ ² test for the qualitative parameters. The Cox proportional hazard regression model was used to explore the possible independent association between the two phenotypes of IHD and gender and the major endpoints such as non-fatal MI and death. The regression model included known potential confounders such as age and traditional cardiovascular risk factors. Hazard ratios (HR) with 95% Confidence interval (CI) were calculated on the entire study population. Kaplan Meyer survival curves were analyzed for the two groups and for the two genders. Analyses were performed with the statistical package SPSS (version 20).

During the average follow-up period of 115 ± 58 months, the clinical course was more favorable in Group AP. A total of 628 deaths (23%), including 269 (43%) cardiac deaths, 239 new non-fatal MI were recorded. The distribution of outcome events and the type of death in the two groups is reported in Table 2. The rate of coronary revascularizations was significantly higher in the MI group (1063, 76% vs 892, 65% respectively, p < 0.001).

Our study expands the investigation to a relatively large cohort of well characterized patients with IHD with a mean follow-up period of 10 years, and in particular provides additional evidence regarding the follow-up events in the two IHD phenotypes which did never overlap until the index hospitalization and database inclusion. It is important to consider, that the overlap between the two phenotypes is quite common in the populations enrolled in many studies which makes quite difficult to compare the results [14‐16].

In the present study the proportion of MI and AP patients was quite similar, which is in contrast with the reported lower prevalence of MI as first presentation of IHD as compared to angina without infarction [6]. A recent report from American Heart Association reports a prevalence around 2.8% for MI and 3.8 for AP [17]. The discrepancy with our population could be due to different selection criteria. Patients in our study were consecutive and < = 70 years old, most admitted for special investigation and treatment as coronary angiography and or PCI, without history of congestive heart failure (group AP). Another reason could be the different prevalence of AP and MI in different countries. In the Hemingway meta-analysis, angina prevalence varied approximately 20-fold between countries, ranging from 0.73 to 14.4% in women and 0.76 to 15.1% in men [18]. In Italy the reported prevalence of the two phenotypes is very similar and this could explain our results [19, 20].

The overall prognosis was significantly more benign in AP and in particular the incidence of new MI during follow-up (0.9/100 patient years vs 1.6/100 patient years, p < 0.001), which was lower than the one reported by Daly et al. (3.2/100 patient years, CI 2.3–4.4) [21] and in other trials (ranging from 1.1 to 1.5%) [21‐24]. The re-infarction rate was also in the lower range of the one reported in the literature (range 1.8–6.9% year in different studies, 95% CI 0.97–1.77, P = 0.08.) [25]. The difference might reflect the duration of follow-up, age, inclusion and exclusion criteria, coronary atherosclerosis involvement, type of treatment. This study included in both groups, with a prevalence in AP (25% vs 7%), patients with normal coronary vessels who have a well known good prognosis. However, the Kaplan-Meier estimates of MI occurrence performed excluding patients with normal coronary vessels obtained similar results (Additional file 2: Figure S3a). Our study was a single-center study and all patients received similar treatment at discharge as secondary prevention and revascularization procedure when required.

In the literature, MI databases report on the significance of angina before and/or after MI [3, 5, 16, 26], and stress the association of angina with higher likelihood of readmission to the hospital, and higher risk of events [26] but can hardly be compared with our results due to different criteria of patients’ selection, modalities and duration of follow-up.

The severity and location of the coronary lesions are among the main factors that have been compared in the two ischemic syndromes in the literature and that have been taken into account to explain the differences in clinical outcomes [27, 28]. A number of studies reported that patients with AP have more severe and extensive coronary lesions than patients with acute unheralded MI [29]. According with more recent papers [16] we found a higher coronary atherosclerotic involvement (expressed as average number of main coronary diseased vessels at target coronary angiography) in MI compared to AP patients, and the extent of coronary atherosclerosis was predictive of new infarct during the follow-up.

A major result of the study was that the different association of sex with the two manifestations of IHD does not wane over time. In contrast with previous studies [6, 18, 30, 31], we compared the two IHD phenotypes of similar age and followed them for a longer time period to investigate the persisting long term effect of sex on MI occurrence independently by age. In the AP group, the incidence of new MI was still significantly lower in females, while no significant difference between sexes was found in the MI group (Fig. 3). On the contrary, sex did not have any significant effect on cardiac and non-cardiac death at approximately 20 years (Additional file 4: Figure S4 and Additional file 5: Figure S5). To the best of our knowledge, this long-term different effect of sex on the two phenotypes of IHD and on (re)infarction rate has never been reported in a large population.

The mechanisms through which sex participates in the pathogenesis of IHD remains speculative [32]. It is assumed that exposure to endogenous estrogens during the fertile period of life delays the development of atherosclerosis in women. In the Women’s Ischemia Syndrome Evaluation (WISE) study it was shown that young women with endogenous estrogen deficiency have a more than sevenfold increase in coronary event risk [33]. Our results showed that the protective effect in females was evident in the AP group and persisted after menopausal age, preventing acute cardiac instability and MI. On the other hand the female sex was not protective against cardiac and total mortality in both IHD phenotypes (MI or AP). This was in accord with previous report on the adverse IHD prognosis in women [33]. Adverse coronary reactivity and microvascular dysfunction have been suggested as contributory factors to a female-specific myocardial ischemia pathophysiology [32]. Moreover differences in medical strategy in the follow-up comparing to men cannot be excluded. Finally, new MI in females could have been underestimated particularly if considered that sex did not have any significant effect on cardiac and non-cardiac death. It’s known that women compared to men are more frequently underdiagnosed for acute MI and they are less frequently studied with coronary angiography.

Regarding the other conventional risk factors no consistent pattern has emerged from this study as well as from the literature [34‐37]. AP patients had a history of hypertension more often than did MI patients as well as higher levels of total cholesterol and LDL at entry. Differently, MI patients were prevalently smokers (66% vs 49%) with slightly higher levels of glycaemia at enrollment.

Our study has several potential limitations. It was a single-center study, which, on the other way, meant similar treatment at discharge.

The type and number of events occurring in the two groups from disease onset to the time of enrollment in the study were carefully taken into account. However, it should be considered that the patients enrolled in the study were only those newly affected by IHD who survived from the onset of the disease up to the time of enrollment. In other words, the incidence of death in our study as well as the final disparity in outcome between the two groups were underestimated in proportion to the time elapsed between onset and enrollment (average 4 years in our study).

During the follow-up we had no information either on the site of the new infarct in all patients or on the potential progression of coronary disease in the follow-up in those not undergoing a new coronary angiography. As such, we cannot definitively draw conclusions on the coronary atherosclerotic burden leading to new MI.

We did not use time-dependent covariates in the Cox model and we cannot exclude the effect of changes in life style, such as smoking cessation on mortality curves.

Moreover, we did not have detailed information on the interim changes in medical treatment as compared to time of enrollment. At discharge from the hospital, medical treatment and lifestyle recommendations did not differ substantially in the two groups and were generally comparable in women and men. During the follow-up medical therapy was left to the clinical judgment of the referring physician-cardiologist. Some chronic medical therapy are associated to a reduced risk of events, such as aspirin or statins or beta-blockers and angiotensin-converting enzyme inhibitors. Although we did not have detailed information regarding medical treatment during follow-up, we feel confident that patients adhered to the best therapy according to physician recommendations.