We performed a single-center retrospective study. The study screened 1,705 patients with a diagnosis of AMI admitted to the cardiology department of the Affiliated Hospital of Jining Medical University between October 2013 and May 2020. A total of 1,699 patients were included in this analysis, after excluding 6 patients with missing data. The flow diagram of the case selection is shown in Figure 1. In the present study, data were compiled from the hospital's electronic medical record system. The study was approved by the Human Ethics Committee of the Affiliated Hospital of Jining Medical University, with a waiver for informed consent.

A total of 1,699 patients hospitalized with AMI were recruited. Fifty-one patients (3.0%) developed CR, including 19 with FWR, 25 with VSR, and 7 with PMR.

The diagnostic criteria of AMI, including ST-segment elevation myocardial infarction (STEMI) and non-STEMI, were as follows: (1) elevation of cardiac biomarkers [creatine kinase-MB (CK-MB) or troponin I or T], (2) typical chest pain with a sudden onset and long persistent period, (3) a new regional wall motion abnormality with imaging evidence, and (4) ST-segment elevation of ≥0.2 mV or depression, prominent T-wave inversion, or pathological Q waves with an ECG.

The diagnostic criteria of CR were as follows: (1) FWR was proven by echocardiography or a large pericardial effusion with pericardiocentesis once AMI patients experienced sudden cardiogenic shock or electrical mechanical dissociation, (2) PMR was confirmed by new mitral regurgitation with echocardiography, and (3) VSR was identified by new significant left-to-right shunting on echocardiography and a cardiac systolic murmur. We collected the clinical data of patients from the hospital's electronic medical record system, including characteristics, medical history, and laboratory parameters, etc.

The Kolmogorov–Smirnov goodness-of-fit test was performed to test the normality of continuous data. Normally distributed continuous variables are represented by mean ± SD and were obtained using an independent sample t-test. Continuous variables with skewed distributions are expressed by medians (with interquartile ranges) and were obtained using rank-sum testing. Categorical variables are expressed as percentages, and comparisons were performed using Pearson’s chi-squared test or Fisher's exact test. We randomly selected 204 AMI patients without CR in a 1:4 ratio as the control group. Factors associated with CR were assessed using binary logistic regression, and the univariate model was further adjusted for the variables age and sex. Subsequently, multivariate analysis was carried out to analyze risk factors with p < 0.1 in the univariate analysis results or clinically significant factors to assess independent predictors of CR. Univariate and multivariable Cox regression hazard analysis was used to estimate the factors associated with in-hospital mortality from CR. Stratifying analysis based on sex was further used to eliminate the interference of confounding factors. All p-values were two-tailed; p < 0.05 was considered statistically significant. SPSS v25.0 statistical software was used.

The clinical characteristics of the study population are presented in Supplementary Table 1. Significant differences in age, gender, the body mass index (BMI), systolic blood pressure (SBP), the left ventricular ejection fraction (LVEF), the cardiac troponin I (cTnI) level, the myoglobin (MYO) level, the glucose level, white blood cell (WBC) counts, red blood cell (RBC) counts, the hemoglobin (Hb) level, the neutrophil percentage (neut%), the albumin level, the creatinine (Cr) level, the lactate dehydrogenase (LDH) level, the triglyceride (TG) level, the site of AMI, and the Killip class were detected in the CR and no-CR groups (p < 0.05). We further separated by gender and examined the difference in clinical characteristics between the CR and no-CR groups. The significant results are detailed in Table 1 (see Supplementary Table 2 for full results). Among the female patients, compared with the no-CR group, CR patients had lower diastolic blood pressure (DBP), lower SBP, and lower total protein (TP) levels (p < 0.05) but higher blood glucose levels (p < 0.05). Among the male patients, CR patients were older and had a lower BMI, lower RBC counts, a lower Hb level, and a lower TG level than no-CR (p < 0.05) but a higher Cr level (p < 0.05). We established that male CR patients had higher WBC counts, neut%, cTnI, and MYO levels but lower albumin levels (p < 0.05) than male no-CR patients, with similar results established in females. Combined with Table 2, we found that primary AMI patients were more common in the CR patient group, accounting for 95.5% of female patients and 93.1% of male patients, but this difference was not significant. The majority of infarction sites were anterior in the CR group, making up 63.6% of female patients and 62.1% of male patients (p < 0.05). Echocardiography showed a lower LVEF in the CR group with the worst Killip class on arrival. The difference in the LVEF between CR and no-CR was statistically significant in males (p < 0.05) but not in females. The Killip class between CR and no-CR had significant differences in both males and females (p < 0.05). The time delay from AMI symptom onset to hospital admission was more common; therefore, only 13 (25%) CR patients received percutaneous coronary intervention (PCI) treatment. The mean days from AMI symptom onset to CR was 5.07 ± 7.18 days in males and 2.39 ± 2.56 days in females. CR was most common in the first 3 days after AMI symptom onset (Figure 2). Only four (7.8%) patients underwent surgical repair. Of the 51 patients with CR, 29 (57%) died in the hospital.

The clinical CR data were analyzed in terms of gender; the results are shown in Supplementary Table 3. Compared with male CR patients, female CR patients had a higher TG level and prevalence of diabetes (p < 0.05) but lower levels of Cr and spent a shorter time in-hospital (p < 0.05).

The clinical CR data were analyzed in terms of age; the results are shown in Supplementary Table 4. More than 50% of female patients were more than 65 years old. CK-MB, high-density lipoprotein (HDL), very low-density lipoprotein (VLDL), RBC counts, Hb level, in-hospital mortality, the prevalence of diabetes, and time spent in hospital were statistically significant (p < 0.05). Next, we carried out a stratification analysis based on gender to assess the difference between the more than 65 years and under 65 years groups. Compared with male CR patients under 65 years of age, male CR patients more than 65 years of age had lower CK-MB, CK, HDL, VLDL, LDH, and hydroxybutyrate dehydrogenase (HBDH) levels (p < 0.05). Furthermore, male CR patients more than 65 years of age had a shorter in-hospital time and time of admission (p < 0.05) (see Supplementary Table 5). However, we did not find significant differences in female CR patients.

Univariate logistic regression identified that the cTnI level, WBC counts, the neut%, the Cr level, anterior MI, a Killip class of >II, SBP, LVEF, the albumin level, and the TG level were associated with CR (p < 0.05). The results after adjusting for age and sex were the same (Supplementary Table 6). Multivariate logistic regression discovered that high WBC counts [odds ratio (OR) = 1.258, 95% confidence interval (CI) 1.104–1.443, p = 0.001], a high neut% (OR = 1.063, 95% CI 1.004–1.125, p = 0.036), anterior MI (OR = 8.406, 95% CI 3.036–23.273, p < 0.001), and a Killip class of >II (OR = 4.517, 95% CI 1.399–14.582, p = 0.012) were risk factors for CR. In contrast, high albumin (OR = 0.869, 95% CI 0.773–0.978, p = 0.020) was protective against CR (shown in Table 3). We further examined the risk factors with CR in different genders. When conducting stratifying analysis by gender, we found that age (OR = 1.096, 95% CI 1.002–1.197, p = 0.044), SBP (OR = 0.936, 95% CI 0.899–0.975, p = 0.002), and bicarbonate (OR = 0.775, 95% CI 0.612–0.983, p = 0.035) were independently associated with female CR patients but not with male patients. The cTnI level (OR = 1.051, 95% CI 1.023–1.079, p < 0.001) and TG level (OR = 0.169, 95% CI 0.032–0.896, p = 0.037) were independently associated with male CR patients but not with female CR patients. Neut% (female, OR = 1.139, 95% CI 1.029–1.260, p = 0.012; male, OR = 1.108, 95% CI 1.025–1.199, p = 0.010), anterior MI (female, OR = 5.977, 95% CI 1.025–34.838, p = 0.047; male, OR = 11.761, 95%CI 2.651–52.177, p = 0.001), and a Killip class of >II (female, OR = 13.400, 95% CI 1.146–156.652, p = 0.039; male, OR = 11.901, 95% CI 2.273–62.314, p = 0.003) were independently associated with CR between females and males (see Figure 3, see more details in Supplementary Table 7).

Univariate and multivariate Cox regression analysis was used to estimate the OR and 95% CIs for in-hospital mortality. Univariate analysis revealed that age, an elevated lipoprotein(a) level, and previous cerebral infarction were related to a high in-hospital mortality risk (p < 0.05). Nevertheless, the site of CR (VSR) and a longer CR time (the time from symptom to CR) were connected with a low in-hospital mortality risk (p < 0.05) (see Supplementary Table 8 for details). In multivariate analysis, predictors of in-hospital mortality included the CR time (≥3 days) (OR = 0.157, 95% CI 0.040–0.622, p = 0.008), the site of CR [PMR (OR = 0.129, 95% CI 0.019–0.861, p = 0.035) and VSR (OR = 0.063, 95% CI 0.011–0.360, p = 0.002)]. The results are shown in Figure 4. We established that, in addition to CR time (p < 0.05) and the site of CR (p < 0.05), higher platelet counts (p < 0.05) were associated with a low in-hospital mortality for male CR patients according to stratification analysis, which was the same in female CR patients (see Table 4).