Clinical characteristics and prognosis differences between isolated right and left ventricular myocardial infarction in the Chinese population: a retrospective study

Study design and patient population

This retrospective cohort study included 4,298 patients consecutive hospitalized with type 1 myocardial infarction (MI) in three hospitals in Hunan province from January 2013 to December 2017. There were 1,723 patients from the Third Xiangya Hospital of Central South University, 1,365 patients from the Xiangtan Central Hospital, and 1,210 patients from the First Affiliated Hospital of the University of South China. Patients with malignancy, severe and active autoimmune disease, trauma causing hemodynamic instability, active infection, severe renal impairment (estimated glomerular filtration rate (eGFR) <15 mL/min), and severe hepatic impairment (alanine transaminase (ALT) >400 U/L or total bilirubin (Tbil) >340 μmol/L) and whose age ≤18 years or ≥90 years and hospital stay less than 1 day were excluded. Patients were also excluded if the final diagnosis of RVMI or LVMI was unclear according to both coronary angiography and electrocardiogram and if the regions of infarction involved both the inferior or posterior wall of the left ventricle and right ventricular. There were 3,506 patients with type 1 MI in the final analysis (Fig. S1). The study was performed according to the 1975 Declaration of Helsinki and obtained the informed consent waiver that approved by the Third Xiangya Hospital of Central South University (NO: 2019-S489), the centre ethics committees of this study. This study was carried out according to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guidelines.

Clinical assessment

Clinical assessment of patients, including characteristics on admission and examination and treatment in hospital. Baseline characteristics on admission including (age, sex and health insurance (none, urban employee basic medical insurance, urban resident basic medical insurance, new rural cooperative medical scheme and others)), smoking status, body mass index (BMI), history of disease (diabetes, hypertension, stroke, angina, chronic heart failure, COPD (chronic obstructive pulmonary disease) and asthma), major presenting symptoms (chest pain, chest stuffiness, shortness of breath and syncope), cardiogenic shock on admission, laboratory test on admission (heart rate (HR), systolic blood pressure (SBP), diastolic blood pressure (DBP), hemoglobin (Hb), white blood cell (WBC), platelets count (PLT), peak cTnI, fasting blood glucose (FBG), alanine transaminase (ALT), low-density lipoprotein cholesterol (LDL-C), creatinine (Cr), serum potassium), and Killip classification indicating the severity of acute myocardial infarction. Findings and in-hospital procedures, including information on electrocardiography (ECG), echocardiography, coronary angiography, medicine treatment (angiotensin-converting enzyme inhibitors (ACEIs) or angiotensin receptor blockers (ARBs), beta-blockers, calcium channel blockers, statins, aspirin, clopidogrel, ticagrelor, intravenous nitrates and inotropes) and coronary revascularization.

Follow-up and clinical outcomes

Information about clinical outcomes, including all-cause and cardiovascular mortality, were obtained from the Chinese Center for Disease Control and Prevention (CDC), followed up from the time of hospitalization with AMI to 31 December 2021 or until death. The maximum follow-up time was 9.0 years (median and interquartile range (IQR): 5.7, 4.7–6.9 years). All-cause mortality meant death from any cause, and cardiovascular death were composed of cardiac death from myocardial infarction, hypertension and its complications, heart failure and malignant arrhythmia.

Statistical analysis

Continuous variables were expressed as medians and interquartile ranges (IQRs) and categorical variables as numbers and percentages. Baseline characteristics, findings, in-hospital procedures, and clinical outcomes between those with isolated RVMI and those with LVMI were compared by ANOVA for continuous variables and chi-squared tests or Fisher’s exact tests for categorical variables. All-cause mortality and cardiovascular mortality according to isolated RVMI and LVMI were plotted in inverse probability of treatment weighting (IPTW)-adjusted Kaplan-Meier curves, and the log-rank test was used for comparison of survival between the two groups.

Isolated LVMI was regarded as the reference group. Univariate and multivariable Cox proportional hazards models were used to evaluate the association of isolated RVMI vs LVMI with all-cause mortality and cardiovascular mortality. The multivariable models were adjusted for age, sex, health insurance, smoking status, BMI, history of disease (diabetes, hypertension, stroke, chronic heart failure and COPD or asthma), cardiogenic shock, ST-segment elevation myocardial infarction (STEMI) or Non-ST-segment elevation myocardial infarction (NSTEMI), laboratory test on admission (HR, SBP, DBP, FBG, LDL-C), LVEF, Killip classification, coronary angiography indicating multivessel stenosis, in-hospital medicine treatment (ACEIs or ARBs, beta-blockers, CCBs, statins, aspirin, clopidogrel, ticagrelor, intravenous nitrates and inotropes) and coronary revascularization. The multivariable models were adjusted for known confounders based on clinical consideration and previous research. The details for each abovementioned covariable were presented in Table S1. Besides, Cox proportional hazards models with an inverse probability of treatment weighting (IPTW) using the propensity score (PS) were applied to balance the difference in baseline characteristics and in-hospital findings and procedures between isolated RVMI and LVMI. In the IPTW approach, the PS was evaluated by logistic regression models, adjusting the same variables as the multivariable Cox proportional hazards model. Then each patient was weighted according to PS. To further minimize the selection bias, double robust models were conducted based on IPTW adjusted models through adjusting variables whose standardized mean difference (SDM) ≥0.05 (Table S2).

Subgroup analysis stratified according to sex (male or female), age (<60 years or ≥60 years), BMI (<24 kg/m² or ≥24 kg/m²), smoking status (yes or no), history of diabetes (yes or no) and history of hypertension (yes or no) were performed to characterize potential effect modifications. Multivariable and IPTW-adjusted Cox proportional hazards models were used in subgroup analysis.

In the sensitivity analysis, propensity score matching methods matched 394 patients in two groups, respectively, through the abovementioned variables with a 1:1 ratio and multivariable Cox proportional hazards model, mixed effects COX model and IPTW-adjusted Cox proportional hazards models further adjusted major presenting symptoms (chest pain, chest stuffiness, shortness of breath and syncope), echocardiography indices (left ventricular (LV) and right ventricular (RV)), arrhythmia (atrial fibrillation, atrioventricular block, bundle branch block, and bradycardia), and other laboratory tests (Hb, WBC, peak cTnI, ALT, PLT, Cr and serum potassium) were used to validate the stability of the results.

Model covariates with missing data were imputed before entering models using random forest with 10 iterations. All hypothesis testing was two-sided, and a p-value <0.05 was considered to indicate statistical significance. All statistical analyses were performed in R version 4.2.0 (The R Foundation, 2023).

Baseline characteristics in patients with isolated RVMI and LVMI

From January 2013 to December 2017, 3,506 patients with type 1 AMI were eligible for final analysis. The median (IQR) age was 64 (55–72) years, and 944 patients (26.9%) were females. Among 3,506 AMI patients, 406 individuals (11.6%) were diagnosed with isolated RVMI and 3,100 individuals (88.4%) were diagnosed with isolated LVMI. Most baseline characteristics, such as age, sex, and comorbidities, were comparable between isolated RVMI and LVMI. Differences included a higher proportion of patients have cardiogenic shock (11.6% vs 2.8%, p < 0.001) and atrioventricular blocker (10.3% vs 4.5%, p < 0.001) on admission in isolated RVMI group. In term of major presenting symptoms, patients with isolated RVMI tended to represent chest stuffiness (34.2% vs 28.8%, p = 0.029) and syncope (4.4% vs 1.5%, p < 0.001), while patients with isolated LVMI were more likely to had chest pain (79.1% vs 73.6%, p = 0.015). It is noteworthy that the median (IQR) heart rate (73 (62–85) beats/min vs 78 (69–89) beats/min, p < 0.001), systolic blood pressure (124 (110–140) mmHg vs 130 (116–146) mmHg, p < 0.001), diastolic blood pressure (77 (66–84) mmHg vs 80 (70–90) mmHg, p < 0.001) were significantly lower in patients with isolated RVMI than those in LVMI. The peak cTnI levels (1.72 [0.30–6.38] μg/L vs 1.19 [0.25–4.57] μg/L, p = 0.007) were higher in the isolated LVMI group than isolated RVMI group (Table 1).

In-hospital medications and procedures in patients with isolated RVMI and LVMI

As the patients with isolated RVMI present lower heart rate and blood pressure, these patients have a lower rate of receiving antihypertensive drugs and vasodilator agents, such as ACEIs or ARBs (65.5% vs 76.3%, p < 0.001), beta-blockers (62.3% vs 74.9%, p < 0.001), intravenous nitrates (61.3% vs 69.0%, p = 0.002), than patients with isolated LVMI. In comparison, patients with isolated RVMI have a higher rate of inotropes treatment (31.1% vs 18.4%, p < 0.001) than isolated LVMI patients. Compared to isolated LVMI patients, there is a higher proportion of patients with NSTEMI than patients with isolated RVMI (46.6% vs 30.6%, p < 0.001). In addition, patients in isolated RVMI group were more likely to complicate with atrioventricular block (10.3% vs 4.5%, p < 0.001) and bradycardia (20.0% vs 6.7%, p < 0.001). Coronary angiography suggested that multivessel coronary lesion was more common in patients with isolated RVMI (23.9% vs 14.7%, p < 0.001) and these patients have a higher rate of coronary revascularization (73.2% vs 67.5%, p = 0.025). Impairment of left ventricular ejection fractions (LVEF) was more severe in patients with isolated LVMI than that in patients with isolated RVMI (52 (44–60)% vs 56 (50–61)%, p < 0.001) (Table 2).

Long-term all-cause and cardiovascular mortality in patients with isolated RVMI and LVMI

The median (IQR) length of hospital stays of patients with isolated RVMI and LVMI were similar (9 (7–12) days vs 9 (7–12) days, p = 0.064). The median (IQR) duration of follow-up was 5.7 (4.7–6.9) years (Table 2). There were 37 all-cause deaths (9.1%) in the isolated RVMI and 680 (21.9%) all-cause deaths in the isolated LVMI. In the isolated RVMI group, a total of 21 (5.2%) patients died of cardiovascular disease, while a total of 427 (12.8%) patients died of cardiovascular in the isolated LVMI group (Table 3). After IPTW adjustment for age, sex, health insurance, smoking status, BMI, history of disease (diabetes, hypertension, stroke, chronic heart failure and COPD or asthma), cardiogenic shock, STEMI or NSTEMI, laboratory test on admission (HR, SBP, DBP, FBG, LDL-C), Killip classification, coronary angiography indicating multivessel stenosis, in-hospital medicine treatment (ACEIs or ARBs, beta-blockers, statins, aspirin, clopidogrel, ticagrelor, intravenous nitrates, and inotropes) and coronary revascularization, the multivariable Cox models revealed that patients suffered isolated RVMI has a lower risk of all-cause mortality(HR 0.36; 95% CI [0.24–0.54], p < 0.001) and cardiovascular mortality (HR 0.37; 95% CI [0.22–0.62], p < 0.001) compared to patients with isolated LVMI (Table 4; Fig. 1).

Subgroup analysis and sensitivity analysis

We further compared the all-cause and cardiovascular mortality between isolated RVMI and LVMI in subgroups with sex, age, BMI, smoking status, diabetes and hypertension. The all-cause mortality and cardiovascular mortality were lower in the subgroups of patients with isolated RVMI compared to the corresponding subgroups with isolated LVMI, except for patients with diabetes (Figs. S2 and S3). In patients with diabetes, all-cause mortality and cardiovascular mortality risks were similar in isolated RVMI and LVMI groups (Figs. S2 and S3). Obesity patients (BMI ≥ 24 kg/m²) and female patients had no significant difference in cardiovascular mortality between isolated RVMI and LVMI (Fig. S3).

Four sensitivity analyses further confirmed the robustness of the association between the regions of myocardial infarction and prognosis. In Model 1, we used propensity score matching the abovementioned confounders in the Cox model. In Model 2, we adjusted additional confounders, including laboratory findings on admission (Hb, WBC, peak cTnI, ALT, PLT and serum potassium), major presenting symptoms (chest pain, chest stuffiness, shortness of breath and syncope), arrhythmia (atrial fibrillation, atrioventricular block, bundle branch block and bradycardia), echocardiography indices (LV and RV) besides abovementioned confounders. In Model 3, we used IPTW method to adjust for the same confounders as model 2. In Model 4, we adjusted confounders with SMD ≥ 0.05 after IPTW adjustment in model 3, which were age, health insurance, Hb, WBC, peak cTnI, PLT, serum potassium, LDL-C, chest stuffiness, STEMI or NSTEMI, bundle branch block, bradycardia and multivessel lesion. In Model 5, we performed the mixed effects COX model using the three hospitals as a random effect. The associations between the regions of myocardial infarction and clinical outcomes remained consistent and statistically significant after sensitive analyses (Table S3).

Limitations and strengths

A previous study involving RVMI mainly focused on RVMI complied with posterior or inferior wall infarction, while our cohort study investigated the difference in clinical characteristics and prognosis in patients with isolated RVMI and LVMI. There are limitations in this study that are worthwhile to be mentioned. First, the retrospective nature of the study means that conclusion remains to be verified in further prospective studies. Second, the SYNTAX score II or Intermountain Risk Score accurately predict mortality in patients with coronary disease (Hayıroğlu et al., 2018; Cınar et al., 2022). Due to large number of patients lack of this index, we adjusted the “number of vessels with lesion” in the mortality risk estimation instead. Third, despite extensive adjustment of confounders, such as IPTW and PSM, there remain potential confounders did not sufficiently estimate and may bring bias to the results (i.e., infarct size, SYNTAX score/Intermountain Risk Score, work pressure, lifestyle and mental health). Fourth, the population of this study consisted of three hospitals in Hunan province, which limited the generalizability of our findings nationally and globally. Fifth, missing data of BMI and serum potassium was likely to result in bias in our findings, but we have conducted a multiple imputation strategy for missing data to minimized potential bias, Sixth, this study base on a small-scale population, which may reduce the strength and reliability of our hypothesis. Further research with a large-scale population derived nationally is warranted to be performed to address and validate the observation from this hospital.