The National Portuguese Registry of Acute Coronary Syndromes is a continuous and prospective registry, promoted by the Portuguese Society of Cardiology, in which all Cardiology departments in the country are invited to participate actively. The registry began in 2002 and remains active today, with 28 cardiology departments contributing their data daily. Briefly, each site should include all patients admitted to the service with the diagnosis of ACS (with ST-segment elevation (STE) and non-ST-elevation (NSTE)) based on clinical evaluation, as well as the electrocardiographic and analytical changes found.

The registry focuses on collating several variables, in particular: 1) demographic and baseline characteristics of admitted patients; 2) laboratory data on admission and during hospitalization; 3) clinical evolution during hospital stay; 4) pharmacological and invasive strategy performed; 5) cardiovascular events and one-year follow-up, whenever possible. Patient identification remained anonymous at all times, and the registry has been authorized by national authorities and registered on the clinicaltrials.gov platform (NCT 01642329).

All ethical requirements contained in the Declaration of Helsinki 1975 have been met and no human and/or animal experiments have been carried out in this work. Written informed consent for the entry of patient data into the registry has been available since 2010 and has been applied following approval by the ethics committee of each hospital site.

A longitudinal, retrospective, multicenter, non-randomized study based on ProACS between 1st October 2010 and 31st December 2018. All patients with a diagnosis of ACS (STE and NSTE) during the defined time period were included, with patients without information regarding the admission hospital (with vs. without on-site PCI) excluded from the analysis. Subsequently, patients were divided into two groups (STE-ACS and NSTE-ACS), then further subdivided into 2 subgroups, according to their admission to hospitals with or without on-site PCI.

In order to make the two groups as homogeneous as possible, to infer conclusions on the occurrence of cardiovascular complications and mortality, a propensity score (PS) was calculated for each group, using a logistic regression model (Appendix A). A 1:1 pairing was performed, taking into account the following characteristics: age, body mass index, gender, hospital transport via ambulance, cardiovascular risk factors, personal cardiovascular and non-cardiovascular history, medication prior to hospitalization and physical examination on admission (Killip class, blood pressure and heart rate). Two patients were paired whenever their respective scores differed below 0.000001. Appendix A shows in detail which variables are included and their respective PS weighting. Variables common to both groups and the occurrence of death and major cardiovascular events during hospital admission were assessed. Figure 1 summarizes the consecutive stages of the study until the final population was obtained.

Figure 1.

Brief description of the study population.

ACS: acute coronary syndromes; NSTE-ACS: non-ST-elevation-acute coronary syndromes; PCI: percutaneous coronary intervention; STE-ACS: ST-elevation-acute coronary syndromes.

(0.16MB).

The authors focused on the acquisition of variables related to the population studied. The following variables were collected: 1) population baseline characteristics (age, body mass index, gender, hospital transport, hospital admission site, form of hospital admission, cardiovascular risk factors, cardiovascular history, non-cardiovascular history and previous medication); 2) clinical findings on hospital admission (hemodynamic profile and admission diagnosis); 3) electrocardiographic findings on hospital admission (rhythm, QRS morphology/duration, ST-T segment); 4) angiographic findings (number of coronary vessels with stenosis greater than 50% and culprit vessel, percentage of catheterization performed, PCI).

The variables were compared between the groups created to infer possible differences between them. Times until intervention were also assessed, with the authors focusing on “symptoms to reperfusion”, “first medical contact to reperfusion” and “door to reperfusion” times for the STE-ACS group and “admission to reperfusion” times for the NSTE-ACS group.

The primary endpoint was the occurrence of in-hospital mortality (defined as death from cardiac, vascular or non-cardiac causes) or the occurrence of major cardiovascular events (reinfarction, congestive heart failure (CHF), cardiogenic shock, mechanical complication, atrioventricular block with hemodynamic consequences, sustained ventricular tachycardia (VT) or cardiorespiratory arrest) during hospital admission.

Categorical variables were described by calculating the respective absolute and relative frequencies, while the continuous variables were described by determining the mean and standard deviation or median and interquartile range, depending on the level of normal distribution in the variables analyzed (assessed using the Kolmogorov-Smirnov test). Whenever comparisons were made between two groups, the chi-square test or Fisher's exact test were used for categorical variables, while the T-test or Mann-Whitney test were favored when comparing two continuous variables.

With regard to inferential statistical analysis, the prognostic impact of the presence of on-site PCI in relation to cardiovascular events and mortality was evaluated according to logistic regression models. These models considered the following variables: On-site PCI, gender, age, diagnosis on admission, cardiovascular risk factors, personal cardiovascular and non-cardiovascular history, Killip class, heart rate, blood pressure, cardiac rhythm, QRS morphology, ST-T segment morphology, coronary angiography (CCTA), coronary lesions and coronary intervention, left ventricular ejection fraction (LVEF) and medication prior to hospital admission.

The stepwise (forward) method, together with the likelihood ratio test, were considered for the selection of variables for inclusion in the present regression models. The adjusted odds ratio, as well as the respective 95% confidence interval (CI 95%), were estimated for each variable included in the regression model. The Hosmer-Lemeshow test was used for the calibration of the regression models. The goodness-of-fit of the logistic regression models was also assessed by determining the area under the curve and its sensitivity and specificity.

Statistical analysis was performed using the Statistical Package for the Social Sciences () 19.0® program, and a significance level of 5% was assumed for the hypothesis tests.

The mean age of the population was 64 ± 14 years, with a predominance of males (75.4%). Arterial hypertension (61.1%), dyslipidemia (51.6%) and smoking (37%) were the most frequent cardiovascular risk factors in this population. 7.7% had a family history of coronary disease and 10.4% had been hospitalized previously due to ACS. 39.3% of patients traveled to hospital by their own means, the most frequent place of admission being the emergency room (39.7%), followed by the cath lab (30.4%). We also found that pre-hospital transport, by specialized medical staff, was only observed in 27.4% of cases, with the coronary fast track system (FTS) activated in 28.9% of cases. It should also be noted that 3% of coronary FTS cases were transported to non-PCI hospitals. Table 1 displays the baseline characteristics of the population studied. The population comparison, taking into account the presence or absence of PCI capabilities at the admission hospital, shows that in PCI-capable hospitals, a higher proportion of patients arrived via ambulance, and there was a higher prevalence of coronary disease and previous coronary revascularization.

We found that the mean age was slightly higher (67 ± 13 years), with arterial hypertension (74.9%), dyslipidemia (63.8%) and diabetes (35%) being the main cardiovascular risk factors. 6.8% had a family history of coronary disease and 26.2% had been hospitalized previously due to ACS. 52.2% of patients traveled to hospital by their own means, the most frequent admission site being the emergency room (53.4%), followed by intermediate/intensive care units (44.9%)—see Table 2. Also, at PCI-capable hospitals, a higher proportion of patients traveled via ambulance, and there was a greater proportion of associated comorbidities (family history, dyslipidemia, chronic renal disease (CRD) and chronic obstructive pulmonary disease).

After applying the PS, we found results similar to those previously described as shown in Tables 3 and 4. Note that pre-hospital transportation by specialized medical staff was lower in STE-ACS after application of the PS.

• ✓

  Clinical, electrocardiographic and angiographic findings of the population

Most patients presented with Killip class I (85.3%). Inferior infarctions predominated (50.5%). Most patients presented a significant lesion of only one vessel (53.3%), with the most frequently found culprit lesion at the level of the anterior descending artery (45.9%). Table 5 summarizes in more detail the findings described here. In comparing the subgroups, the study found that in PCI-capable hospitals there is a greater proportion of coronary disease of between one and three vessels, and the common trunk is involved more frequently.

A predominance of Killip class I was observed (85.5%), as in the STE-ACS group. 90.9% of our patients presented sinus rhythm, with ST-segment depression being the most frequent finding (32.7%) in both groups. Most patients presented a significant lesion of only one vessel (35.4%), with the most frequently found culprit lesion at the level of the anterior descending artery (32.7%). Table 6 summarizes in more detail the findings described here. Once again, the proportion of common trunk involvement is greater in PCI-capable hospitals.

After applying the PS, we found results similar to those described for the general population, as shown in Tables 7 and 8.

• ✓

  Times until reperfusion in acute coronary syndrome

ST-elevation acute coronary syndrome The median time from first medical contact to reperfusion was 112 [68;175] min, with non-PCI hospitals result 69 min higher than PCI-capable hospitals (Table 9). In general, non-PCI hospitals presented a longer delay in the times assessed. However, there were still considerable delays in patients admitted to PCI-capable hospitals.

After applying PS, we observed a slight worsening of the median “first medical contact to reperfusion” time. We also found that the time was 72 min longer in non-PCI hospitals than PCI-capable hospitals (p<0.001). This is shown in Table 10, with a breakdown of the times under study.

92.7% of the population underwent CCTA, with 86.9% undergoing angioplasty. A slight predominance of catheterization (93.7% vs. 91%, p<0.001) and angioplasties (88.5% vs. 84.2%, p<0.001) is worthy of note in patients admitted to PCI-capable hospitals. A residual percentage (0.3%) of the population underwent coronary-aortic revascularization surgery. dRegarding medication during hospitalization and on discharge, we found that, despite a high adherence, drugs for secondary prevention in these patients are still not always routinely prescribed (Table 11), with beta-blockers being prescribed more on discharge in non-PCI hospitals (83.2% vs. 80.5%, p=0.004). Finally, we highlight the higher percentage of fibrinolysis (12.9% versus 2.1%, p-value <0.001) in patients admitted to non-PCI hospitals.

After applying PS, we found similar results to those previously described (Table 12), but we continue to see a higher number of angioplasties performed (87.8% vs. 81.8%, p-value <0.001). It should also be noted that prescription of acetylsalicylic acid (96.5% vs. 94.4%, p=0.018) and statin (96.1% vs. 93.8%, p=0.013) is slightly higher in patients admitted to PCI-capable hospitals.

84.2% of the population underwent CCTA, with 52.3% undergoing angioplasty. There is also a slight predominance of catheterization (86.8% vs. 82.2%, p<0.001) and angioplasties (54.6% vs. 50.4%, p<0.001) performed in non-PCI hospitals. A residual percentage (1%) of the population underwent aortic coronary revascularization surgery. As regards medication during hospitalization and discharge, we found that, as in STE-ACS, drugs for secondary prevention in these patients are still not always routinely prescribed. Note the higher prescription of beta-blockers (BB) (80.5% vs 77.1%, p<0.001) and angiotensin-converting enzyme inhibitors (ACEi)/angiotensin II receptor antagonists (ARA II) (85.1% versus 82.6%, p<0.001) in non-PCI hospitals (Table 13).

After applying PS, we found an absence of statistical significance between hospitals with and without PCI capabilities in relation to the number of catheterizations and angioplasties performed. We also observed an absence of differences regarding secondary prevention medication, namely ACEi/ARA and BBs (Table 14).

• ✓

  Cardiovascular complications and in-hospital mortality

Regarding this population (Table 17), we found that PCI-capable hospitals had more cases of CHF (14.6% vs. 11.7%, p<0.001), while non-PCI hospitals had more episodes of sustained VT (0.7% vs. 1.1%, p=0.020).

After application of PS, a higher incidence of CHF (12.3% vs. 10%, p=0.003) was observed in patients admitted to PCI-capable hospitals (Table 13). A longer stay in non-PCI hospitals was identified, but no differences were observed between the two groups relating to in-hospital mortality (Table 18).

The present study is an observational study based on data from a retrospective and non-randomized registry. Despite the inclusion of a significant number of sites, the current registry does not include all sites in Portugal with cardiology departments, which could lead to a potential selection bias of the data analyzed.

Two other potential sources of selection bias were the non-inclusion of patients admitted to departments other than cardiology, and patients who died before admission to a cardiology department, neither of which are included in the National Registry.

The impossibility of categorizing NSTE-ACS according to clinical indication (emergent, urgent or up to 72 h), due to the absence of some variables in the national registry, is an important limitation in this group. It does not allow the impact to be assessed for each subgroup, and only one overall assessment is possible in the present study.

Applying a PS is in itself a limitation, as by pairing the total sample, very seriously ill patients are removed from the most severe group and the healthier patients from the least severe group.

In conclusion, the objective of our study was to assess cardiovascular events and short-term mortality, and as such it is not possible to extrapolate data to the medium and long term.