Introduction

Incidence and mortality of cardiovascular (CV) disease have decreased during recent decades1. Improved acute treatment as well as effective secondary prevention have contributed to this positive development2. However, ischemic heart disease and stroke still constitute the two leading causes of death, both in Sweden and worldwide3,4. There is limited data on long-term outcome after an acute coronary syndrome (ACS) in recent years. Most of our current knowledge on long-term risk after ACS is based on clinical trials, generally not mirroring the general population and with relatively short follow-up periods. Register based studies have study samples more representative of the general population, but they have an inherent weakness in the lack of adjudication of outcome events, which can lead to both under- and overestimation of endpoints, and thereby false risk estimates. Both clinical trials and register-based studies point at a substantial risk elevation for recurrent events in both the short and long terms after ACS5,6. Moreover, improved survival after ACS in combination with a growing and ageing population will lead to an increased burden of coronary heart disease7. These findings highlight the importance of optimized care, both in the acute phase and long term regarding secondary prevention of recurrent events. A reliable estimate of long-term risk in a contemporary post-ACS population is therefore important in order to be able to direct adequate secondary preventive measures. The aim of this study was to assess the risk for recurrent CV events in an unselected cohort of patients hospitalized for ACS, i.e. acute myocardial infarction (AMI) and unstable angina (UA), with the possibility of long-term follow-up.

Methods

Study population and data collection

We included all patients registered for potential inclusion in the NAILED-ACS Risk Factor Trial, which is a randomized, controlled trial conducted at Östersund Hospital, designed to investigate if intervention by a telephone-based, nurse-led secondary prevention programme could improve CV outcomes after ACS, compared with usual care. Traditional modifiable risk factors were monitored during follow-up for randomized patients in both groups. Patients registered, but not randomized were not monitored during follow-up. The study protocol has been published elsewhere8. In summary, the trial included all patients hospitalized for ACS at Östersund Hospital between 1 January 2010 and 31 December 2014. Suspected ACS patients were identified through review of patient records and included if they fulfilled the criteria for UA, defined as suspected ischemic chest pain and electrocardiogram changes suggestive of myocardial ischemia, such as ST-segment depression or T-wave alteration, or type 1 AMI according to the third universal definition of MI. A study flow chart is presented in Fig. 1.

Figure 1
figure 1

Study flow chart.

Full size image

Östersund Hospital is the only hospital in the county of Jämtland/Härjedalen, with a catchment area of about 126.000 people. During the inclusion period, the primary reperfusion treatment for ST-segment myocardial infarction was thrombolysis. In May 2015, Östersund hospital established a 24/7 primary percutaneous coronary intervention (PPCI) network in the county of Jämtland/Härjedalen, which resulted in a switch in reperfusion treatment for STEMI from thrombolysis to PPCI in the vast majority of cases.

Endpoints were identified and adjudicated by reviewing all study subjects discharge records from subsequent hospitalizations at the department of internal medicine and cardiology at Östersund Hospital. Cause of death was adjudicated by reviewing death certificates as well as patient records and classified according to prespecified criteria. To identify endpoints that took place in other departments, relevant International Statistical Classification of Diseases and Related Health Problems (ICD-10) diagnoses were extracted from the national patient register (NPR), which contains main and secondary diagnoses for all hospitalizations in Sweden. These were also adjudicated by review of patient records. The adjudication of events was performed by four experienced physicians (2 senior consultants and 2 senior residents) who were part of the study team. Each physician worked independently according to a standardized workflow algorithm and endpoints were adjudicated strictly according to a prespecified endpoint definition. Consecutive meetings were held where complicated cases were discussed and consensus reached.

Endpoints

The primary endpoint was major adverse cardiovascular events (MACE) defined as the composite of CV death, type 1 AMI and ischemic stroke, whichever happened first. The follow-up period was from the day after discharge from the index event hospitalization until 31 Dec 2017. All diagnoses were classified according to ICD-10 codes. Death was classified as CV, non-CV or unknown cause of death. CV death was defined as death from CV causes within 30 days after an AMI, sudden cardiac death (SCD), congestive heart failure, stroke, CV procedures, CV bleeding and other CV causes such as pulmonary embolism and peripheral artery disease9. AMI was defined according to the third universal definition of myocardial infarction. Only type 1 AMIs were included, where type 1 AMI is defined as an AMI classically caused by plaque rupture, erosion, ulceration or dissection followed by thrombosis at the lesion or distal embolization causing myocardial ischemia and necrosis10. Ischemic stroke was defined as an acute episode of focal or global cerebral, spinal or retinal dysfunction caused by infarction of central nervous system (CNS) tissue.

Statistical analysis

Baseline data for the cohort are presented as means for continuous variables and categorical variables as frequencies. Age is presented as a median and interquartile range (IQR). Baseline characteristics were tested for differences between subjects reaching the primary endpoint or not, with the independent samples T-test or Mann–Whitney U-test as appropriate for continuous variables and the Pearson Chi-square-test for categorical variables.

Kaplan–Meier analysis was used to estimate the cumulative probability of the primary endpoint. Cumulative Kaplan–Meier (KM) incidence was also estimated according to age stratification < 60, 60–69, 70–79 and ≥ 80 years. The individual components of the primary end-point were analyzed separately.

Risk factors for the combined endpoint were analyzed in a Cox proportional hazards model. The proportionality of hazards was verified with Schoenfeld residuals. Variables that previously have been shown to impact risk or otherwise deemed important, were analyzed in univariable analysis and included in the multivariable model if the p-value reached a significance level of < 0.1. Results are presented as hazard ratios (HR) and 95% confidence intervals (CI). Since there was a high overall death rate, and non-CV death might occur before the primary endpoint, a competing risk analysis according to Fine and Gray was performed. Statistical analysis was performed i SPSS (version 25.0; IBM Corp, Armonk, NY) and SAS software (version 9.4; SAS Institute Inc, Cary, NC, USA).

Ethics

This study was approved by the Regional Ethical Committee in Umeå on October 28, 2009 (Dnr: 09-142M), with supplements on June 10, 2013 (Dnr: 2013-204-32M) and January 13, 2015 (Dnr: 2014-416-32M). This study was conducted according to relevant guidelines and regulations. All study participants had to sign an informed consent form prior to randomization.

Results

Baseline characteristics

1379 ACS-patients were included (63.9% men and 36.1% women) and followed from the day after discharge from the index event hospitalization for ACS until 31 Dec 2017. Baseline characteristics are presented in Table 1. The median age was 71 years, with a considerably higher median age in the female population. At index hospitalization, 18.9% were current smokers and 40.8% were previous smokers. 28.4% reported a family history of CV disease defined as a first degree relative with CV disease before the age of 55, or a second degree relative with CV disease before the age of 65. Index ACS-type were NSTEMI in 62.9%, STEMI in 29.5% and UA in 7.6% of cases. 58.5% had hypertension, 22.1% diabetes and 8.6% atrial fibrillation. 21.4% had previously had an MI and 5.9% an ischemic stroke. 53.8% were revascularized during index hospitalization (43.8% PCI and 10% CABG). 60.1% of patients underwent echocardiographic examination during index hospitalization. Of these, 55.4% had a normal ejection fraction (EF), 24.5% were mildly reduced, 14.8% moderately reduced and 5.3% severely reduced. Generally, patients were pharmacologically well treated at discharge. 92.7% had aspirin, 78.5% a P2Y12-inhibitor, 8.2% oral anticoagulation (OAC), 88% betablocker, 87% lipid lowering therapy and 77.3% RAAS-blockade. Medication on admission and at discharge is presented in Tables 2 and 3 respectively.

Table 1 Baseline patient characteristics among 1379 patients discharged after hospitalization for ACS.
Full size table
Table 2 Medication on admission among 1379 patients discharged after hospitalization for AC.
Full size table
Table 3 Medication at discharge among 1379 patients discharged after hospitalization for ACS.
Full size table

Primary endpoint

In total, there were 327 events meeting the criteria for the primary endpoint during the entire study period, with a median follow-up time of 4.7 years, K–M cumulative incidence 28.6% (95% confidence interval (CI) 26.2–31.0) (Fig. 2). At 1 year, K-M cumulative incidence was 10.3% (95% CI 8.7–12.1). Outcomes including all-cause death at 1 and 4.7 years is presented in Table 4. The K–M incidence of the individual components of the primary endpoint are presented in Figs. 3, 4 and 5 respectively.

Figure 2
figure 2

Cumulative Kaplan–Meier estimate of survival without MACE among 1379 patients discharged after hospitalization for ACS between 1 Jan 2010 and 31 Dec 2014.

Full size image
Table 4 Outcomes at 1 and 4.7 years among 1379 ACS patients.
Full size table
Figure 3
figure 3

Cumulative Kaplan–Meier estimate of CV-death free survival among 1379 patients discharged after hospitalization for ACS between 1 Jan 2010 and 31 Dec 2014.

Full size image
Figure 4
figure 4

Cumulative Kaplan–Meier estimate of survival without type 1 myocardial infarction among 1379 patients discharged after hospitalization for ACS between 1 Jan 2010 and 31 Dec 2014.

Full size image
Figure 5
figure 5

Cumulative Kaplan–Meier estimate of survival without ischemic stroke among 1379 patients discharged after hospitalization for ACS between 1 Jan 2010 and 31 Dec 2014.

Full size image

Risk according to age

KM estimates for the primary end-point according to age stratification were analyzed at 3 years, due to few events in two youngest age strata towards the end of the studied time period. At 3 years, there were 19 events, KM estimate 8.1% (95% CI 5.1–12.5) among patients < 60 years (n = 237), 40 events, KM estimate 10.2% (95% CI 7.4–15.7) among patients 60–69 years (n = 396), 60 events, KM estimate 16% (95% CI 12.5–20.1) among patients 70–79 years (n = 377) and 126 events, KM estimate 34.2% (95% CI 29.4–39.3) among patients ≥ 80 years (n = 396), see Fig. 6.

Figure 6
figure 6

Cumulative Kaplan–Meier estimate of survival without MACE according to age groups among 1379 patients discharged after hospitalization for ACS between 1 Jan 2010 and 31 Dec 2014.

Full size image

Predictors of the primary endpoint

Baseline characteristics according to the occurrence of the primary endpoint are displayed in Table 5. Patients suffering a primary endpoint event were significantly older and more likely female. Index ACS-type was predominantly NSTEMI among patients reaching an endpoint, who also had significantly more comorbidities such as hypertension, diabetes mellitus and atrial fibrillation. There was also a clear overrepresentation of prior AMI, congestive heart failure (CHF), ischemic stroke and peripheral artery disease among patients experiencing an event. Event-free patients were also revascularized to a far greater extent.

Table 5 Patient Characteristics among 1379 patients discharged after hospitalization for ACS according to the occurrence of MACE during the entire follow-up period.
Full size table

Variables tested in a univariable cox regression model are presented in Table 6. In a multivariable model, age HR 1.03 (95% CI 1.02–1.05), prior CHF HR 1.56 (95% CI 1.11–2.21), prior diabetes mellitus HR 1.74 (95% CI 1.36–2.22), prior angina HR 1.41 (95% CI 1.08–1.85), prior PCI HR 1.41 (95% CI 1.00–2.00), prior CABG HR 1.82 (95% CI 1.36–2.49) and treatment with diuretics at discharge HR 1.32 (95% CI 1.04–1.68) were independently associated with an increased risk of the primary endpoint during the study period. Revascularization during index event hospitalization with PCI HR 0.58 (95% CI 0.44–0.77) and CABG HR 0.51 (95% CI 0.31–0.85) as well as lipid lowering therapy at discharge HR 0.69 (95% CI 0.51–0.94) were associated with a reduced risk for the occurrence of the primary endpoint. The result of the multivariable cox regression model is presented in Table 7.

Table 6 Risk factors associated with CV-death, MI and IS in 1379 patients discharged after hospitalization for ACS in a univariable analysis.
Full size table
Table 7 Risk factors associated with CV-death, MI and IS among 1379 patients discharged after hospitalization for ACS in a multivariable Cox regression analysis.
Full size table

Competing risk analysis according to Fine and Gray did not alter the result of the multivariable analysis.

Discussion

This was a retrospective observational study in a contemporary, population-based and unselected cohort of ACS-patients, with a long follow-up period, reporting long-term cardiovascular outcomes after ACS. To our knowledge, there are no comparable studies reporting outcomes in absolute numbers. Almost one out of three patients died for cardiovascular reasons or had a recurrent ischemic event during the study period. Over 40% of all events and more than half of all MI’s occurred during the first year. IS events were more evenly distributed, with nearly 1/3 occurring during the first year.

Compared with the Cardiovascular risk in post-myocardial infarction patients: nationwide real-world data demonstrate the importance of a long-term perspective study6, which was a Swedish nationwide register based study (n = 97,254) to assess the incidence and risk factors of subsequent cardiovascular events in the Swedish post MI population, we found a lower event rate both short and long term. That study included AMI-patients surviving the first week after discharge, with a total follow-up of 4 years and a mean follow-up time of 2.5 years. The 1-year rate of the composite of cardiovascular (CV) death, recurrent MI and stroke was estimated to 18.3%, and the risk remained high in patients without an event during the first year, in the subsequent 2 years at about 20%, which is almost twice as high as in our material. The patient data upon which these results are based is comparable with ours in terms of a high median age and a relatively low frequency of revascularization during hospitalization. On the other hand, preexisting heart failure was clearly more prevalent (26.3%) compared with our cohort (5.3%) which might contribute to a higher event rate in general and CV mortality in particular. Moreover, in our study all end-points were adjudicated and we only included type-1 MI’s, which likely contributes to the observed lower event rate. Type-2 MI has been associated with a clearly higher mortality in previous reports and the proportion of type-2 MI in SWEDEHEART has been estimated to 7.1%11. Despite these possible explanations, the results differ substantially which highlights the need for hiqh-quality long-term follow-up studies to estimate CV risk in a contemporary, unselected ACS population.

In a large study comparing long-term outcomes after MI (3 years) 2002–2011 in nationwide registers from England, France, Sweden and USA in more than 100,000 patients > 65 years, the cumulative incidence of the composite of all-cause death, MI and stroke varied between 26% (France) and 36.2% (USA)12. These numbers are more at level with our results when taking into account that the end-point included all-cause death and that the studied populations had a high burden of comorbidities and a very high mean age of 77.5–78.6 years.

These results indicate that the risk is high during the first year, and remains high also in a longer time perspective, which highlights the importance of adequate secondary prevention and revascularization procedures.

Older age, diabetes mellitus, prior ischemic stroke, heart failure (treatment with diuretics at discharge), established coronary heart disease (prior PCI, CABG or angina) independently predicted increased risk for the primary endpoint in our material, whereas revascularization during index hospitalization and lipid-lowering therapy were associated with favorable outcome. The most important established risk factors of CV events are family history of CVD, smoking, physical inactivity, hypertension, hyperlipidemia, diabetes mellitus, and abdominal obesity.

In accordance with previous publications, older age conferred increased risk for recurrent cardiovascular events in this study13. An ageing population and increasing proportion of survivors of ACS should urge health care systems to direct their attention regarding secondary prevention not only to younger patients. Diabetes Mellitus has repeatedly been shown to forcefully increase the risk for cardiovascular disease14. It has also been linked to excess mortality after MI compared with patients without diabetes15, and ESC guidelines for non-ST-segment elevation acute coronary syndromes (NSTE-ACS) recommend that a multifactorial approach, with treatment targets should be considered in this subset of ACS-patients (grade of recommendation IIa)16. In our material, diabetes was the comorbidity with the strongest association of elevated long term ischemic risk after ACS, which in conjunction with previous evidence points at the importance of good metabolic control and intensive lipid lowering therapy17. There is vast evidence from clinical trials that lipid-lowering statin therapy reduces cardiovascular morbidity and mortality in both chronic and acute coronary syndromes18,19. In our study, lipid-lowering therapy was strongly associated with a reduced risk for recurrent events in the long term, which confirm the results from clinical trials and contribute to generalizability to a real-world ACS-population.

Ischemic stroke and heart failure have, as in this study, been identified previously as markers of elevated risk for recurrent ischemic events and death6,13,20,21.

Revascularization during index-hospitalization for ACS, both CABG and PCI, were strongly associated with better outcome in this study. Jernberg et al. reported that absence of revascularization was associated with increased cardiovascular risk6. An early invasive strategy, with coronary angiography followed by PCI if appropriate, has been shown to reduce cardiovascular events long-term after NSTE-ACS, compared with an initial conservative strategy22,23. In ST-segment myocardial infarction (STEMI), there is solid evidence that prompt revascularization reduces mortality and subsequent ischemic events24,25. Elderly patients are generally considered to have a greater burden of coronary disease and thereby more myocardial ischemia and seem to derive a relatively greater benefit from revascularization26,27. Given the high median age in this cohort, this seems to hold true for our patient population and has important implications for the often difficult clinical decision-making regarding revascularization in the care of elderly patients with ACS.

Strengths and limitations

This unselected, single-center cohort, with the aim to study the long-term cardiovascular outcome after hospitalization for ACS, was relatively small, which is a limitation. 60.1% of patients were examined with echocardiography during hospitalization, which is a relatively low number and represents a limitation in terms of the multivariable analysis. However, the data collected can be considered to be of high quality as well as the adjudication of clinical endpoints. The outcome measures cardiovascular death, myocardial infarction and ischemic stroke are fairly standardized. We chose to only include type 1 MI’s, because we had the opportunity to evaluate every clinical event, and since a type 1 MI has a different prognostic impact than for example a type 2 MI. This might lead to lower event rates compared with register-based studies, in which it is hard to discriminate between infarction types. Due to the establishment of a PPCI-network in the region, patients received a different reperfusion treatment, with a clear dominance of PPCI from May 2015 until the end of follow-up, compared with a fibrinolysis dominated reperfusion treatment before May 2015, which might has affected outcomes.

Conclusion

In an unselected cohort of ACS-patients, with a median follow-up time of 4.7 years, the risk for the composite of CV-death, MI and IS was 10.3% at 1 year and 28.6% during the entire study period.

Risk factors associated with an increased risk of the primary end-point were Diabetes Mellitus, prior revascularization, heart failure at discharge and prior IS. Associated protective factors were revascularization during index hospitalization and lipid lowering therapy at discharge.