Study design and population

This observational study was based on the data from the Ogolnopolski Rejestr Procedur Kardiologii Inwazyjnej (ORPKI registry—Polish National Registry of Invasive Cardiology Procedures). ORPKI is an electronic database operated by the Jagiellonian University Medical College in Krakow and endorsed by the Polish Association of Cardiovascular Interventions of the Polish Cardiac Society that collects data on all percutaneous procedures in interventional cardiology performed in Poland [7,15]. Data in the ORPKI registry have been gathered via electronic case report forms. Data on patients with a diagnosis of ACS (ST-segment elevation MI, non-ST-segment elevation MI or unstable angina) treated in the Swietokrzyskie region of Poland and included in the ORPKI registry between January 1, 2014, and December 20, 2017, were analyzed in this study. Swietokrzyskie is a region in southeast Poland with over 1.2 million inhabitants. In the region, seven centers with catheterization labs actively participated in the ORPKI initiative and recruited patients throughout the study period. Patients with ACS INOCA were identified based on non-obstructive coronary arteries visualized on angiography. In case any significant lesion (>50%) was found, a diagnosis of ACS OCAD was established. All patients provided informed consent for the procedures. All medical procedures were performed according to the current medical standards. The study complied with the ethical principles for clinical research based on the Declaration of Helsinki with its later amendments.

The patients were followed up for up to 36 months to assess mortality and major adverse cardiovascular events (MACE), which were defined as the composite of death, cardiac arrest, MI, stroke, or heart failure hospitalization. Data regarding all-cause mortality, hospitalization for revascularization, and MACE were obtained from the official records of the National Health Fund.

Statistical analysis

Categorical variables were presented as percentages. Continuous variables were presented as mean ± standard deviation or median with quartiles (IQR, interquartile range). Normality was assessed using the Kolmogorov–Smirnov–Lilliefors (KSL) test for samples of over 2000 observations or the Shapiro–Wilk test, as appropriate. Equality of variances was assessed using the Levene’s test. Differences between groups were compared using the Student’s or Welch’s t-test based on the equality of variances for normally distributed variables. The Mann–Whitney U test was used for non-normally distributed continuous variables or ordinal variables. Categorical variables were compared using the Pearson’s chi-squared test or Fisher’s exact test if 20% of cells had expected count <5. The survival function was estimated using the Kaplan–Meier estimator and comparisons between subgroups were performed using the log-rank test.

To avoid the potential influence of non-randomized design and reduce bias, a propensity score was calculated using a multivariate logistic regression model with ACS INOCA diagnosis as the dependent variable and the following variables: site, age, sex, weight, diagnosis, diabetes, previous stroke, previous MI, previous percutaneous coronary intervention (PCI), previous coronary artery bypass graft surgery (CABG), active smoking status, hypertension, kidney disease, cardiac arrest at baseline, access site during angiogram, chronic obstructive pulmonary disease, and Killip class on admission. Covariate balance was assessed using standardized differences that were less than 0.05. Pairs of patients with ACS INOCA and ACS OCAD were formed using 1:1 caliper matching with caliper width of 0.03. Unpaired patients were excluded from the analysis. Continuous parameters were compared for pairs using paired Student’s t-test if differences between pairs were normally distributed or using Wilcoxon signed-rank test otherwise. For nominal variables, the McNemar’s or Bowker’s test was used. For lifetime data, the matched pairs approach of the win ratio method was applied [16]. The level of statistical significance was set at p<0.05. Statistical analyses were performed using JMP^® v14.0.0 (SAS Institute INC., Cary, NC, USA) and R v3.4.1 (R Foundation for Statistical Computing, Vienna, Austria, 2017).

Clinical characteristics of patients with ACS INOCA

There were notable differences in the initial clinical characteristics of patients in the two groups. Patients with ACS INOCA tended to be younger and had a different sex distribution than those with ACS OCAD. The mean age of patients with ACS INOCA was 65.6 years, which was close to the upper limit reported in previous MINOCA studies [5,17]; however, it was still lower than that in patients with ACS OCAD (67.5 years).

Some studies have demonstrated that the prevalence of traditional cardiac risk factors is lower in patients with MINOCA than in those with MI and OCAD [7,18]. In contrast, among the most common CAD risk factors, a recent systematic review has demonstrated that only hyperlipidemia was significantly less frequent in patients with MINOCA [5]. In our patients with ACS INOCA, traditional cardiovascular risk factors including diabetes, active smoking, previous MI, or kidney disease were significantly less frequent than in those with ACS OCAD. Additionally, their hemodynamic presentation on admission was more favorable; more patients had no heart failure symptoms (Killip class I) and fewer patients presented with cardiogenic shock (Killip class IV) or cardiac arrest at baseline.

The definition of MINOCA includes patients with either normal coronary arteries (no stenosis or stenosis <30%) or mild coronary atherosclerosis (stenosis >30% and <50%) [2]. In our study, the majority of patients with ACS INOCA (78.5%) had at least some evidence of coronary atherosclerosis on angiogram, which was higher than that reported in a recently published meta-analysis of patients with MINOCA in which non-obstructive CAD was present in 54% of patients [13]. This suggests that elevated biomarkers do not necessarily predict the burden of coronary atherosclerosis.

Clinical profiles of patients with ACS INOCA were different from those of patients with ACS in whom OCAD would be subsequently identified; therefore, adjustments were necessary to compare outcomes between the groups that differed only in baseline angiography findings. It is not clear if the risk assessment tools used for predicting adverse clinical outcomes in patients with MI and OCAD should also be used in patients with INOCA [10]. At least they appear to be more useful in predicting the clinical outcomes in patients with MINOCA [17]. Multiple factors have been proposed as independent predictors of adverse outcomes in patients with MINOCA, and most of them are related to patient characteristics on admission [17,19]. These factors were also available for propensity score estimation in this study.

In-hospital outcomes

Mortality rates reported in patients with OCAD are generally higher than those in patients with MINOCA; however, the difference does not reach statistical significance [18]. To some extent, this may be linked to a relatively low incidence of death and limited sample sizes. Data from a specific population of a “super-aging society” in Japan indicated that the short-term prognosis is worse in patients with MINOCA than in those with MI and CAD [20]. A recent meta-analysis of patients with MINOCA reported in-hospital mortality of 0.9% [5]. In the present study, in-hospital mortality rate of approximately 0.5% was observed in patients with INOCA and was lower than that in patients with ACS OCAD both before and after propensity matching.

Follow-up mortality and MACE after propensity score matching

After adjusting for baseline risk factors, the prognosis tended to be similar in patients with MI irrespective of the presence of obstructive CAD [14,21]. Our study partially confirms these observations in a broader group of patients with ACS, including those without elevated cardiac biomarkers. Despite better in-hospital outcomes in patients with ACS INOCA, the difference in MACE attenuated over time. In terms of the incidence of MACE during the follow-up, ACS carried a prognosis which was only numerically better (p = 0.07) in patients with INOCA than in those with OCAD after propensity score matching. The win ratio for all-cause death remained higher in patients with OCAD, while that for MI and HF hospitalizations was similar between the groups. Rather surprisingly, in a recent Chinese study, the incidence of non-fatal MI or HF was extremely low even in patients with MINOCA over a 1-year follow-up [9]. Elevated cardiac biomarkers are known to be associated with worse prognosis in a variety of clinical scenarios. However, ACS presentation that convinces a treating physician to refer the patient for urgent angiography should not be regarded as a trivial incident, even if no significant lesions are identified and cardiac biomarkers remain stable. In a group of patients with ST-elevation ACS triaged for urgent coronary angiography, patients with INOCA had a similar or higher long-term risk of death than those with obstructive CAD, irrespective of the cardiac troponin levels [8].

The risk of subsequent revascularizations

Patients with ACS INOCA and OCAD are anatomically very different from each other; therefore, the risk of subsequent revascularization during the follow-up period was, as expected, higher in patients with ACS OCAD. This is in accordance with the findings of a previous study that reported an over 6-fold higher risk of any repeat revascularization [14]. The magnitude of this risk was even higher in the current analysis (Fig 2) that included mostly patients with ACS but without a diagnosis of MI. The possible explanation is that CAD could be likely diagnosed at an earlier stage in these patients with ACS INOCA, and one could not determine if they would have undergone revascularizations later after the completion of follow-up. In fact, approximately half of patients with reinfarction after MINOCA who underwent coronary angiography had progression of coronary stenosis [22]. The higher win ratio for any cardiac hospitalization could be explained by the higher rate of repeated revascularizations, whereas the ratios of hospitalizations for HF were not different between the groups.

Options for secondary prevention

Optimal secondary prevention strategies are not yet established for ACS INOCA, and treatment guidelines do not specifically address this population. MINOCA should be considered as a working diagnosis that requires further diagnostic workup to identify possible coronary or extra-coronary pathologies [2,6,23]. This approach is likely to be true for INOCA as well. However, additional coronary artery assessment is rarely performed in everyday clinical practice [7] or not reported in large registries [22]. This was also true in this study group where only 1.6% of patients with INOCA underwent additional coronary artery assessments while data on any further diagnostic workout outside the catheterization lab were not available. Given the many possible underlying causes of INOCA and lack of systematic evaluation of the underlying mechanisms, treatment recommendations remain empiric [2]. Factors predicting unfavorable outcomes in patients with MINOCA appear similar to factors previously demonstrated to predict new cardiovascular events in patients with OCAD [17]. The comparison of Kaplan–Meier curves for unadjusted and propensity score matched cohorts (Fig 1) suggest the effects of baseline characteristics—which included most traditional CAD risk factors—on the magnitude of differences in outcomes. Management guided by an additional assessments of functional coronary arteries [24] could be the optimal strategy; however, such an approach is currently underutilized. Additionally, myocarditis may account for approximately 33% cases of MINOCA [6]; therefore, additional tests including cardiac magnetic resonance imaging with late gadolinium enhancement could entirely change the management strategy. Nevertheless, we believe that in patients with ACS INOCA in whom no overt diagnosis for ACS background can be established, the management should focus on the known CAD risk factors irrespective of results of angiography because, apart from the rates of subsequent revascularization procedures, patients with ACS INOCA had similar outcomes in terms of subsequent MI or heart failure hospitalizations over 3 years of follow-up as patients who initially presented with ACS OCAD. This strategy is also supported by the fact that the frequency of non-significant lesions in patients with ACS INOCA in our study (78.5%) was relatively high. However, the 3-year all-cause mortality was still slightly higher in those with ACS OCAD than in those with ACS INOCA. It appears likely that the cardiac risk of patients with INOCA changes over time and the ACS incident should warrant appropriate cardiovascular risk management rather than reassurance of the benign nature of this condition manifesting as negative cardiac biomarkers and non-obstructive coronary arteries on angiography.

Limitations

Several limitations of our study should be acknowledged. In the ORPKI registry, all data about the patients are entered by the local physicians according to the registry protocol. Therefore, there were some potentially important baseline data missing, especially regarding the laboratory and echocardiographic findings. This may affect the quality of our propensity score model. Importantly, data on the diagnostic workup for possible non-cardiac causes of an acute presentation were missing in the registry. The diagnosis of ACS and referral for coronary angiography was performed in each participating institution according to the current medical standards. Furthermore, additional tests to continue the diagnostic workup of patients with INOCA during and after coronary angiography were limited. This may be an important source of bias because some of the included patients with INOCA might have had non-ischemic or even non-cardiac causes of their acute presentations. However, given the retrospective nature of the registry data, we could only emphasize the problem that the current utilization of these tests in the real world is far from optimal. Furthermore, the ORPKI registry does not include data on pharmacotherapy after discharge, and many patients enrolled late during the study period were followed up for less than 3 years. The study included patients treated in one region in Poland; therefore, caution is warranted before generalizing our findings to other populations. Additionally, patients with INOCA are a heterogeneous group and their prognoses may differ depending on the actual underlying mechanism of the acute presentation. Unfortunately, this is not routinely investigated in real-life settings, and our data also reflect this problem.