Reperfusion Therapy and Predictors of 30-Day Mortality after ST-Segment Elevation Myocardial Infarction in a University Medical Center in Western Iran

Received: January 31, 2021, Accepted: June 2, 2021, ePublished: November 1, 2021 Abstract Background: Considerable variability in survival rate after ST-segment elevation myocardial infarction (STEMI) is present and outcomes remain suboptimal, especially in lowand middle-income contraries. This study aimed to investigate predictors of 30day mortality after STEMI, including reperfusion therapy, in a tertiary hospital in western Iran. Methods: In this registry-based cohort study (2016–2019), we investigated reperfusion therapies – primary percutaneous coronary intervention (PPCI), pharmaco-invasive (thrombolysis followed by angiography/percutaneous coronary intervention), and thrombolysis alone – used in Imam-Ali hospital, the only hospital with a PPCI capability in the Kermanshah Province. We estimated hazard ratios (HRs) and 95% confidence intervals (CIs), using Cox proportional-hazard models, to investigate the potential predictors of 30-day mortality including reperfusion therapy, admission types (direct admission/referral from non-PPCIcapable hospitals), demographic variables, coronary risk factors, vital signs on admission, medical history, and laboratory tests. Results: Data of 2428 STEMI patients (mean age: 60.73; 22.9% female) were available. Reperfusion therapy was performed in 84% of patients (58% PPCI, 10% pharmaco-invasive, 16% thrombolysis alone). Only 17% of the referred patients had received thrombolysis at non-PPCI-capable hospitals. Among patients with thrombolysis, only 38.2% underwent coronary angiography/ percutaneous coronary intervention. The independent predictors of mortality were: no reperfusion therapy (HR: 2.01, 95% CI: 1.36–2.97), referral from non-PPCI-capable hospitals (1.73, 1.22–2.46), age (1.03, 1.01–1.04), glomerular filtration rate (0.97, 0.96–0.97), heart rate > 100 bpm (1.94, 1.22–3.08), and systolic blood pressure < 100 mm Hg (4.92, 3.43–7.04). Mortality was lower with the pharmaco-invasive approach, although statistically non-significant, than other reperfusion therapies. Conclusion: Reperfusion therapy, admission types, age, glomerular filtration rate, heart rate, and blood pressure were independently associated with 30-day mortality. Using a comprehensive STEMI network to increase reperfusion therapy, especially pharmacoinvasive therapy, is recommended.


Introduction
Ischemic heart disease is the single most common cause of death worldwide, accounting for 16.17% of global deaths. 1 In 2019, it was the main cause of 9.14 million deaths and 182 million disability-adjusted life years, with a global prevalence of 197 million cases. 2 In Iran, ischemic heart disease is the leading cause of death, accounting for 26.28% of total deaths. 1 Acute coronary syndrome including STsegment elevation myocardial infarction (STEMI) may be the first manifestation of ischemic heart disease with substantial morbidity and mortality. 3 Regional differences exist in treatment strategies and mortality rates attributed to STEMI within and across countries, suggesting opportunities for performance improvement. 4 Widespread use of modern reperfusion therapies including primary percutaneous coronary intervention (PPCI), thrombolysis, and pharmaco-invasive (i.e. thrombolysis followed by angiography and, if indicated, percutaneous coronary intervention [PCI]), has resulted in a fall in STEMI mortality, particularly in developed countries 5 ; however, in low-and middle-income countries (LMICs), where 80% of all cardiovascular deaths occur, lack of an appropriate care system is the most important barrier to implementing guidelinebased STEMI treatment. 6 A considerable number of STEMI patients, especially in LMICs, are not offered any reperfusion therapy, particularly PPCI which is the preferred reperfusion strategy in STEMI patients. 5 A study in China showed that the in-hospital mortality rate of STEMI in county-level hospitals was 3-times higher than province-level hospitals (10.2% vs. 3.1%, respectively), partly due to hospital facilities and reperfusion therapy. 7 There is no consensus on the predictors of short-term mortality of STEMI. 8 STEMI mortality may be influenced by many health-system level and individual level factors such as time delay to treatment, ambulance system efficacy, reperfusion strategy, in-hospital treatment, age, previous heart disease, renal function, number of diseased coronary arteries, and known risk factors including diabetes mellitus, hypertension, dyslipidemia, and tobacco smoking. 8,9 Little is known about the management and mortality predictors of STEMI in LMICs. Imam-Ali hospital is the only tertiary care academic cardiovascular center in the Kermanshah Province in western Iran. It is also the only hospital in the province with a 24 hours a day, 7 days a week (24/7) PPCI capability. The present study investigates the predictors of 30-day mortality of STEMI patients, including their received reperfusion therapies, in this hospital.

Study Setting, Design and Participants
The Kermanshah province in western Iran had 1 952 434 inhabitants in the 2016 census in its area of 25 009 km 2 . 10 There are 23 hospitals in all 14 counties of this province; 13 of them are in the city of Kermanshah, the province capital. 10 The geographical distribution of hospitals in the province is shown in Figure 1. Imam-Ali hospital is a cardiology training center in the city of Kermanshah, affiliated to the Kermanshah University of Medical Sciences. This is the only 24/7 PPCI-capable hospital in the province. Other hospitals in the province are non-PPCI capable but have been equipped to provide thrombolytic therapy. Ambulance services in this province are not equipped with electrocardiography machines and thrombolysis facilities.
This registry-based cohort study included all adult patients (> 18 years) who presented with STEMI to Imam-Ali hospital from July 1, 2016 to September 19, 2019. Diagnosis was made by cardiologists based on current guidelines. 11 The STEMI patients who were hospitalized more than 24-hours before referring to Imam-Ali hospital were excluded from this registry. In this study, we also excluded patients with out-of-hospital cardiac arrest. The eligible patients were followed up 30 days after STEMI events. All participants signed written informed consent.

Baseline Assessment
In the registry, trained physicians and nurses collected demographic and clinical data such as past medical history, the time of the onset of symptoms, and transfer to Imam-Ali hospital, from personal interviews with patients and/or their attendants. Direct admission (selfpresentation) to Imam-Ali hospital or referral from other hospitals was recorded. The time between symptom onset and arrival to hospital (onset-to-arrival time) were calculated. History of cardiovascular events (previous myocardial infarction, stroke, or chronic heart failure), coronary intervention (PCI or coronary artery bypass graft surgery), diabetes, and hypertension were recorded, based on self-reports of confirmed diagnoses by health care members. We recorded the reperfusion treatment used including: PPCI, pharmaco-invasive, thrombolysis alone, and none (no reperfusion). Thrombolysis was received either before or after admission at Imam-Ali hospital. Information about the admission process, hemodynamic status, electrocardiography data, medical treatment, laboratory tests, etc. was obtained from hospital medical records. Systolic blood pressure (SBP) and heart rate (HR) were measured on admission at Imam-Ali hospital and categorized into two groups (SBP: < 100/ ≥ 100 mm Hg and HR: > 100/ ≤ 100 bmp, respectively), based on the categories of thrombolysis in myocardial infarction (TIMI) risk score. 12 Body-mass index (BMI) was calculated as weight in kilograms divided by the square of height in meters. Lipid profile and creatinine level were measured on the first day of admission. We defined high low-density lipoprotein cholesterol (LDL-C) as LDL-C > 160 mg/dL and low high-density lipoprotein cholesterol (HDL-C) as HDL-C < 40 mg/dL in men and HDL-C < 50 mg/dL in women. 13 We estimated glomerular filtration rate (GFR) using the CKD-EPT equation, 14 based on patients' initial serum creatinine level. All recorded data were quality controlled by trained physicians.

Study Outcome and Follow-up
The main outcome was 30-day mortality after STEMI. Data of in-hospital mortality were collected using hospital records. Upon hospital admission, patients' mobile phone numbers and home phone numbers, as well as two phone numbers of their family members or their attendants were recorded. Trained nurses contacted each patient by phone and collected information about their vital status. If death occurred within 30 days after STEMI, the cause and date of death were recorded. Follow-up time extended from the date of STEMI diagnosis to the date of death, loss-tofollow up, or 30 days after STEMI, whichever came first.

Statistical Analysis
All statistical analyses were performed using Stata statistical software (Stata Corp, Release 12, College Station, TX). We used the registry data and included all eligible patients in the study, so we did not determine the sample size. Continuous variables were expressed as mean (SD) and discrete variables were presented as counts (percentages). Cox proportional-hazard modeling was performed to determine predictors of 30-day mortality. The proportionality of hazards was verified using Schoenfeld's global test and log-log plots. Candidate variables for inclusion in the models were selected according to previously reported mortality predictors 9,12 and variables of interest; including age (continues), sex (female/male), admission type (direct admission/ referral), onset-to-arrival time (< 4, 4-8, and > 8 hours), SBP (< 100/ ≥ 100 mm Hg), HR (> 100/ ≤ 100 bmp), history of cardiovascular events (yes/no), history of coronary intervention (yes/no), ever smoking (yes/no), diabetes (yes/no), hypertension (yes/no), type of MI (anterior wall or left bundle branch block/others), early reperfusion therapy (PPCI/pharmaco-invasive/thrombolysis/noreperfusion), GFR (continuous), BMI (continuous), high LDL-C (yes/no), and low HDL-C (yes/no). We reported HRs with 95% confidence intervals (95% CIs) using univariable, age-and sex-adjusted, and fully-adjusted Cox proportional-hazards models. We used change-inestimate strategies, with 10% change rule, 15 to identify the independent predictors of 30-day mortality in the final model. We also used stepwise selection to confirm the final model. We performed a subgroup analysis based on the admission typesdirect admission to Imam-Ali hospital and referral from other hospitals. In a sensitivity analysis, we included out-of-hospital cardiac arrests and re-evaluated the results. We performed all analyses on complete case data. P < 0.05 and 95% CIs not including one were considered as statistically significant.
The mean age of patients was 60.73 (SD: 12.44) years, and 22.86% were female. During 69,638 person-days of follow-up, 139 deaths were recorded, of which 123 (88.49%) occurred in hospital and 16 (11.51%) after discharge. The crude 30-day mortality rate was 5.72%. The baseline characteristics of the patients sorted by vital status are shown in Table 1.
In crude analyses, older age, female sex, diabetes, hypertension, history of cardiovascular events, referral from other hospitals, SBP < 100 mm Hg, HR > 100 bpm, low GFR, low HDL-C, high onset-to-arrival time, and no reperfusion therapy increased the risk of death significantly. High BMI and ever smoking were protective factors of mortality in crude analyses; however, after adjusting for age and sex, their significant associations disappeared. Pharmaco-invasive therapy tended to be more protective than PPCI, although this association was not statistically significant ( Table 2). Table 3 shows the independent predictors of 30day mortality based on the multivariable analysis. In this model, the association of age, HR, SBP, referral from other hospitals, no reperfusion therapy, and GFR remained statistically significant. Overall, patients who received reperfusion therapy had a 59% lower rate of 30-day mortality (HR: 0.41, 95% CI: 0.29-0.56). Subgroup analysis showed the same results in patients with direct admission and those who referred from other hospitals ( Table 4). In the sensitivity analysis, out-of-hospital cardiac arrests had the strongest association with 30-day mortality with an HR (95% CI) of 4.92 (2.95-8.21) in the final model. Age, referral from other hospitals, SBP < 100 mm Hg on admission, low GFR, type of MI (anterior wall or left bundle branch block), and no reperfusion therapy were also associated with increased mortality (see Table S1 of Supplementary file 1).
In Figure 2, we compared all the subgroups of patients, based on both reperfusion therapies and admission types, and considered PPCI in directly admitted patients as the reference group. Referred patients compared with those who admitted directly to Imam-Ali hospital tended to have poorer prognoses, in all the related categories of reperfusion treatments. Figure 2 also demonstrates lower mortality rates, although statistically non-significant, of the pharmaco-invasive therapy compared with PPCI, for both admission types.

Discussion
In the present study, 58% of patients received PPCI, 10% pharmaco-invasive, 16% thrombolysis alone, and 16% no reperfusion. Only 17% of patients referred to Imam-Ali hospital had received thrombolysis at non-PPCI-capable hospitals. Among patients with thrombolytic therapy, coronary angiography/PCI was reported in 38% of patients. The independent predictors of 30-day mortality after STEMI were lack of reperfusion therapy, referral from non-PPCI-capable hospitals, older age, lower GFR, and HR > 100 bpm or SBP < 100 mm Hg on admission. In our study, early reperfusion therapy was associated with a 59% lower rate of 30-day mortality. Other studies have also reported that implementation of reperfusion therapy, particularly prompt PPCI (< 120 minutes) which is the preferred reperfusion strategy in STEMI patients, is associated with a reduced mortality rate. 9 Although the guideline-based treatment in highincome countries has improved survival of STEMI patients, in LMICs, considerable healthcare resource and infrastructure constraints make implementing those guidelines unfeasible and thus, STEMI survival remains *LDL-C > 160mg/dL. **HDL-C < 40mg/dL in men and HDL-C < 50mg/dL in women. suboptimal. 6 We showed that there are approximately 219 PPCI per 1 000 000 inhabitants annually, in the only 24/7 capable hospital in the province with almost two million population, much lower than 600 PPCI per 1 000 000 inhabitants in developed European countries such as Austria, Germany, and the Netherlands. 5 The pharmacoinvasive strategy has been suggested as one of the most promising reperfusion strategies in LMICs. 6 In our study, among 631 patients with thrombolytic therapy, only 38.2% received subsequent angiography with or without PCI. In some developed countries, using well-designed STEMI networks, almost all patients treated with thrombolytic therapy are directly transferred to PPCI-capable hospitals and undergo pharmaco-invasive therapy. 16 In the current analysis, the mortality rate for the pharmaco-invasive approach was lower, although statistically non-significant, than other reperfusion therapy approaches. These results were seen in both subgroups of direct admission and referral to Imam-Ali hospital. Comparable survival rates between pharmaco-invasive and PPCI approaches have been reported in some other studies. 17,18 We revealed that referral from non-PPCI capable hospitals was associated with a 73% increase in 30-day mortality rate compared with direct admission to Imam-Ali hospital, independent of the reperfusion therapies. This may be due to the lack of advanced hospital facilities and insufficient capabilities in clinical expertise in non-PPCI-capable hospitals in the province. A large study of STEMI patients in 108 hospitals in China investigated the variations in care and mortality among various levels of hospitals. The rates of reperfusion therapy and inhospital mortality were 69.4% and 3.1% for province-level hospitals and 45.8% and 10.2% for county-level hospitals, respectively. More efforts to address the gaps in care and outcomes of STEMI for national quality improvement were recommended. 7 Our primary results also demonstrated that older age, female sex, history of cardiovascular events, hypertension, diabetes, SBP < 100 mm Hg and/or HR > 100 bpm on admission, lower GFR, low HDL-C, and high onsetto-arrival time were associated with increased 30-day STEMI mortality, but in multivariable analyses, the associations remained statistically significant for age, HR, SBP, and GFR. Likewise, the TIMI risk score for STEMI was created based on the adjusted logistic regression models to predict 30-day mortality. 12 Old age, female sex, diabetes, hypertension, HR > 100 bpm, SBP < 100 mm Hg, and previous cardiovascular diseases were among the final predictors of mortality. 12 However, in the ACTION (Acute Coronary Treatment and Intervention Outcomes Network) Registry-GWTG (Get With the Guidelines) study, 19 significant associations with mortality were not reported for sex, diabetes, hypertension, and previous coronary intervention, in multivariable models. Similar to the Action Registry-GWTG study, we found that renal function was an independent predictor of mortality. For every one mL/min/1.73 m 2 increase in GFR, the 30day mortality rate decreased by 3% in our study. Renal function was not analyzed in the TIMI risk score study. 12 In our study, 30-day mortality in never smokers was significantly higher than ever smokers; however, this association disappeared in the adjusted models. Smoking is one of the strongest risk factors for STEMI; however, some studies have shown that smokers have a favorable prognosis after acute myocardial infarction, known as "smoker's paradox". 20,21 For example, in a study among 985,174 patients with STEMI undergoing PPCI, smokers had lower in-hospital mortality than non-smokers (2.0% vs. 5.9%). 22 Another study showed a significant lower crude rate of death in smokers; however, after adjustment for age and other risk factors, smokers and non-smokers had similar mortality rates. Researchers revealed that smokers were 10 years younger than non-smokers with fewer cardiovascular risk factors. 20 In fact, smoking is the strongest behavioral risk factor for the premature onset of atherosclerosis. 23 Obesity is a known cardiovascular risk factor and nearly 70% of death related to high BMI are due to cardiovascular disease. 24 However, there are conflicting data on the association between BMI and mortality in established coronary artery disease patients. Some, 25,26 but not all, 27 studies reported the protective effects of obesity; a phenomenon often termed the "obesity paradox". We showed that higher BMI were associated with a decrease in 30-day mortality in our crude analysis, although these significant associations did not persist in age-and sexadjusted and the final multivariable model. In a large population-based cohort study in Iran, BMI had a weak association with cardiovascular mortality. 28 Researchers suggested using indicators of visceral adiposity, especially hip-adjusted waist circumference, as the best obesity indicator in that population. They argued that BMI cannot distinguish lean mass from fat mass and cannot recognize fat distribution in the body. 28 To our knowledge, this is the first large registry-based study with detailed data about the 30-day mortality of STEMI patients during more than 3 years in western Iran. This study collected data from all consecutive eligible patients to minimize selection bias. The availability of data to adjust for confounders and the extremely low loss to follow-up rate were among other strengths of this study. However, the results of this study should be interpreted in the context of the following limitations. Firstly, the study was conducted in the only 24/7 PPCI-capable hospital in the Kermanshah province after excluding out-of-hospital cardiac arrests. Therefore, the results of mortality rates may not be applicable to the entire province or country. Although we showed a 5.7% crude mortality rate in PPCIcapable Imam-Ali hospital, a nationwide study reported an in-hospital mortality rate of 12.1%, in 31 provinces of Iran. 29 In line with our results, in-hospital STEMI mortality rate in Tehran Heart Center (another PPCIcapable hospital in Iran) was 5.4% and reduced from 8% to 3.9% during 2006-2017, mainly due to the improvement of reperfusion therapies. 30 Secondly, we did not have data on all of the factors associated with 30-day mortality, as well as unmeasured confounding variables that may be present in any observational studies. Thirdly, as an observational study, we acknowledge information bias such as exposure identification bias and misclassification, especially about self-reported variables. Lastly, the small number of deaths in some reperfusion subgroups precluded further detailed analyses.

Conclusions and Implication
According to the findings of our study, no reperfusion therapy, referral from non-PPCI-capable hospitals, older age, lower GFR, and HR > 100 bpm or SBP < 100 mm Hg on admission were independently associated with increased 30-day mortality after STEMI. Our findings indicated the importance of immediate/management factors for the 30-day mortality of STEMI compared with the traditional long-term cardiovascular risk factors.
This study has important implications. Using an efficacious STEMI care network, efforts must be focused on widespread aspects of STEMI care in the entire province, such as improving hospital facilities, transportation systems, and intensive training in all parts of the health system, including PPCI-capable and non-PPCI-capable hospitals. Geographic limitations and resource constraints in the Kermanshah Province, as well as many other regions of LMICs, indicate that PPCI cannot be provided for most of the population; however, the pharmaco-invasive strategy can be considered as the best alternative and the most feasible pathway for STEMI treatment in such regions. 6