This retrospective study was conducted in accordance with the principles of the Declaration of Helsinki. The study was approved by the ethics committee of the hospital of which all patients provided written informed consent [code number: Ethic(2022)246]. The inclusion criteria were patients aged > 18 years admitted to the department of cardiology at the hospital between January 2018 and December 2020 with a diagnosis of ST-segment elevation myocardial infarction (STEMI) or non–ST-segment elevation myocardial infarction (NSTEMI) defined according to the latest guidelines (12). The exclusion criteria included those with any comorbid condition that limited HPA (e.g., leg amputation, advanced cancer, disabling stroke, Parkinson's disease) or severe mental illness that would interfere with the ability to engage in HPA.

Patients were allocated to the HPA or non-HPA group based on whether they engaged in HPA before the index admission. HPA is defined as moderate to vigorous activity, including walking, swimming, using a treadmill, cycling, and using an ergometer, tennis, and other kinds of physical activity, with a duration of no less than 150 min/weak (13).

Participants were asked to complete a questionnaire at baseline during their index admission. Clinical characteristics were extracted from medical charts, including demographic characteristics [age, sex, body mass index (BMI)], cardiovascular risk factors (smoking, alcohol consumption, hypertension, diabetes, family history of MI and interventional therapy), and blood biochemical index [cardiac troponin T (cTnT), low-density lipoprotein (LDL) and high-density lipoprotein (HDL)]. The Killip class during the index admission was ascertained by the cardiologist in charge according to the guideline as follows (14):

• Class I: No signs of heart failure, but PCWP (pulmonary capillary wedge pressure) may be elevated, with a fatality rate of 0%–5%

• Class II: mild to moderate heart failure, lung rates are less than 50% of the two lung fields, third heart sound, persistent sinus tachycardia or other arrhythmias, elevated venous pressure, x-ray manifestations of pulmonary congestion, with a mortality rate of 10%–20%

• Class III: severe heart failure with acute pulmonary edema, rates in more than 50% of both lungs, and a fatality rate of 35%–40%

• Class IV: cardiogenic shock, systolic blood pressure less than 90 mmHg, urine output less than 20 ml/h, cold and clammy skin, cyanosis, accelerated breathing, pulse rate more than 100 beats/min, and a case fatality rate of 85%–95%

All patients were followed up by phone and/or clinical visits to collect clinical outcome information, including 1-year MACEs (defined as death from any CV diseases, cardiac readmission for nonfatal MI, or nonfatal heart failure or arrhythmia within 1 year following the index date of admission), 1-year CV mortality (defined as death from any CV diseases within 1 year following the index date of admission), and 1-year cardiac-related readmission rate (defined as hospital admissions for nonfatal cardiac causes within 1 year following the index date of admission, which included reinfarction, heart failure, and arrhythmia).

All analytical tasks were completed by using SPSS v24.0 (IBM Corp.). For follow-up assessments, the baseline was defined as the date the patient was first admitted to the hospital. To compare the differences in baseline demographics between those with and without HPA, continuous variables and categorical variables of the two groups were analyzed by Student's t-test and the chi-square (χ²) test, respectively. The age, BMI, LDL, and HDL results are described as the means and standard deviation (means ± SD), and the cTnT results are presented as the medians and interquartile ranges [medians (IQRs)], whereas the cardiovascular risk factors like smoking, alcohol consumption, hypertension, diabetes, family history of MI, and interventional therapy are presented as number and percentage. p ≤ 0.05 denoted significant differences between the two groups.

For analysis of the influence of HPA on the Killip class, we performed univariate and multivariate binary logistic regression analyses. The quantitative variables (age, BMI, and cTnT) and dichotomous variables (HPA, sex, smoking, alcohol consumption, hypertension, diabetes, family history of MI, and interventional therapy) identified in the univariate analysis as being significantly associated with a lower Killip class (p ≤ 0.05) were further analyzed with a multivariate binary logistic regression model. For analysis of the influence of HPA on 1-year MACEs, 1-year CV mortality, and 1-year cardiac readmission rate, we performed univariate and multivariate binary logistic regression analyses. The quantitative variables (age and BMI) and dichotomous variables (HPA, sex, smoking, alcohol consumption, hypertension, diabetes, family history of MI, interventional therapy, and higher Killip class) identified in the univariate analysis as being significantly associated with 1-year MACEs, 1-year CV mortality, and 1-year cardiac readmission rate (p ≤ 0.05) were further analyzed with a multivariate binary logistic regression model. Effect sizes are given as odds ratio (OR) and 95% confidence interval (95% CI). Interaction analysis was used to investigate the comprehensive effects of HPA and subgroup factors on the probability of adverse outcomes. Effect sizes are demonstrated as OR and 95% CI. Student's t-test was used for the comparison between continuous variables, and the χ² test was used for the comparison between categorical variables. p ≤ 0.05 denoted significant differences between the two groups.

The study flow diagram is shown in Figure 1. Between January 2018 and December 2020, 1,575 patients were admitted for MI, and all gave written consent. Based on the information in the questionnaire and the definition of HPA mentioned above, 723 participants were considered to engage in HPA, while the rest of the 852 participants were allocated to the non-HPA group. Information about key outcomes, including 1-year MACEs, 1-year CV mortality, and 1-year cardiac-related readmission rate, was collected for all except 309 patients (152 patients with HPA and 157 patients without HPA) who were lost to follow-up during the 1-year follow-up period. Finally, 571 patients with HPA and 695 patients without HPA completed this 1-year follow-up study.

Baseline characteristics are shown in Table 1. We recruited an MI population of elderly, predominantly male patients (age, 63.4 ± 12.6 years; sex, 72% are male). Among the population, participants who engaged in HPA were younger (61.4 ± 11.5 years vs. 64.9 ± 13.3 years, p = 0.001), more likely to be male (77% vs. 68%, p = 0.001), consume alcohol (30% vs. 20%, p < 0.001), have a family history of MI (43% vs. 32%, p < 0.001), and receive interventional therapy (77% vs. 51%, p < 0.001) than those who did not engage in HPA. For the prevalence of smoking, hypertension, and diabetes, as well as BMI, cTnT, LDL, and HDL values, there were no differences between the HPA and non-HPA groups at baseline.

According to the Killip classification, 732 (58%) patients were in class I, 362 (29%) were in class II, 129 (10%) were in class III, and 43 (3%) were in class IV upon admission. The distribution of participants in different Killip classes among patients who engaged in HPA and those who did not are depicted in Figure 2. Patients who engaged in HPA were less likely to be in Killip class III (5% vs. 15%) and Killip class IV (2% vs. 5%) upon admission to MI than those who did not engage in HPA.

The relationship between a high Killip class (Killip classes III and IV) and HPA is shown in Table 2. Univariate analysis demonstrated that a higher Killip class was negatively associated with HPA (OR = 0.35: 95% CI, 0.24–0.50; p < 0.001), which remained significant following multivariate adjustment for sociodemographic factors (age, sex, smoking, and alcohol consumption) and clinical demographic factors (hypertension, diabetes, family history of MI, interventional therapy, BMI, and cTnT) (adjusted OR = 0.48: 95% CI, 0.32–0.71; p < 0.001).

Variables related to 1-year MACEs, including HPA, age, sex, smoking, alcohol consumption, hypertension, diabetes, family history of MI, interventional therapy, BMI, and higher Killip class, are described in Table 3. Univariate analysis demonstrated that patients who had HPA had significantly fewer 1-year MACEs than those who were inactive (OR = 0.57: 95% CI, 0.44–0.74; p < 0.001), which remained significant following multivariable adjustment (adjusted OR = 0.74: 95% CI, 0.56–0.98; p = 0.038).

To investigate the influence of HPA on 1-year MACEs in different subgroups, patients were stratified according to age (<60 or ≥60 years), sex (male or female), history of smoking, alcohol consumption, hypertension and diabetes (presence or absence), family history of MI (presence or absence), interventional therapy (presence or absence), obesity (BMI < 28 or ≥28 kg/m²), and Killip class (higher or lower) (Figure 3). The association of HPA and a lower risk of 1-year MACEs following MI was consistent for both sexes, irrespective of history of hypertension, diabetes, or interventional therapy (p for interaction >0.05). The effect of HPA was more pronounced among those who did not smoke or drink, were not obese, and had a higher Killip class (p for interaction <0.05).

We further investigated the effect of HPA on 1-year CV mortality and 1-year cardiac readmission rate. As shown in Table 4, participants with HPA had a lower risk of 1-year CV mortality after multivariate adjustment (OR = 0.50: 95% CI, 0.28–0.88; p = 0.017). Nonetheless, there was no significant association between HPA and the 1-year cardiac readmission rate (OR = 0.87: 95% CI, 0.64–1.17; p = 0.346) (Supplementary Table S1).

To investigate the influence of HPA on 1-year CV mortality among patients in different subgroups, participants were stratified according to age (<60 or ≥60 years), sex (male or female), history of smoking, alcohol consumption, hypertension and diabetes (presence or absence), family history of MI (presence or absence), interventional therapy (presence or absence), obesity (BMI < 28 or ≥28 kg/m²), and Killip class (higher or lower) (Figure 4). Patients who engaged in HPA had a lower rate of 1-year CV mortality irrespective of age, sex, history of smoking, hypertension and diabetes, family history of MI, interventional therapy, and Killip class (p for interaction >0.05). The effect of HPA on lower 1-year CV mortality was more apparent among non-obese patients than among those who were obese (p for interaction <0.01).