Abstract
Ventricular septal rupture (VSR) is a mechanical complication of acute myocardial infarction (AMI), and its mortality has not decreased significantly in recent decades. However, no clinical model has been developed to predict short-term mortality in patients with post-infarction VSR (PIVSR). This study aimed to develop a nomogram to predict the 30-day mortality by using the clinical characteristics of hospitalized patients with PIVSR. The least absolute shrinkage and selection operator (LASSO) and multivariate logistic regression analysis was used to construct a nomogram by R. The model was evaluated by the area under the curve (AUC), calibration curve and decision curve analysis (DCA). The bootstrap method was used to validate the model internally. As a result, a nomogram was constructed by using six variables, including CRRT, mechanical ventilation, PPCI, WBC, PASP and methods of treatment. The AUC of the prediction model was 0.96 (0.93, 0.98). The prediction model was well calibrated. The DCA showed that if the threshold probability was between 15% and 95%, the nomogram model would provide a net benefit. The well-constructed and evaluated nomogram can be beneficial to clinicians to predict the risk of death within 30 days in patients with PIVSR.
Keywords: Ventricular septal rupture, Acute myocardial infarction, Nomogram, Prediction model, 30-day mortality
Subject terms: Cardiovascular diseases, Interventional cardiology
Introduction
Ventricular septal rupture (VSR) is a rare but potentially fatal mechanical complication that occurs after acute myocardial infarction (AMI). The popularity of early reperfusion therapy for AMI has significantly reduced the incidence of post-infarction VSR (PIVSR). Before the era of thrombolysis, VSR had an incidence of approximately 1–2% in patients with AMI1,2. During the period of thrombolysis reperfusion, the incidence of mechanical complications after AMI decreased, but delayed thrombolysis was a common cause of VSR after AMI3,4. Owing to early percutaneous coronary intervention (PCI), the incidence has decreased to 0.17%–0.31%5,6. Patients with PIVSR are usually severely ill and many of them died while waiting for surgical treatment. The mortality of PIVSR remains high. According to a report by Elbadawi and colleagues based on the National Inpatient Sample Database of the United States from 2003 to 2015, the mortality of patients without mechanical complications after AMI was about 12.7%, while that of patients with mechanical complications was as high as 42.4%1, which has not changed significantly over the past decades. The results of the American College of Chest Physicians database showed that the 30-day mortality of PIVSR was 43%5.
Studies have shown that some factors, such as no history of angina, cardiogenic shock, anterior AMI, delayed reperfusion of the coronary artery, no formation of ventricular aneurysm may be predictors of early death in patients with PIVSR7–11. With the improvement of surgical methods and the rapid development of percutaneous catheter techniques, the repair of PIVSR is of great help to improve the prognosis of patients with PIVSR2. Meanwhile, the effect of inflammation on AMI and its related complications is receiving increasing attention12,13. The healing process after AMI can be roughly divided into three phases: inflammation phase, proliferation phase of inflammation regression, and repair phase. After 30 days of inflammation phase and inflammation regression, it has been relatively stabilized.
A nomogram is a graphical tool for determining the probability of an individual experiencing the clinical event based on a statistical prediction model14. However, the nomogram for predicting 30-day mortality in patients with PIVSR has received little attention.
Therefore, this study intended to include some clinical indicators that were previously considered to be important for the prognosis of PIVSR, inflammation-related laboratory tests such as white blood cell counts, and methods of treatment, in order to develop a nomogram model that can help predict the risk of death within 30 days in patients with PIVSR.
Materials and methods
Population
Patients with PIVSR hospitalized at Zhengzhou University Central China Fuwai Hospital from January 1, 2018, to April 30, 2023, were selected as study participants in our retrospective study. The Ethics Committee of Fuwai Central China Cardiovascular Hospital (approval number:2019–87) approved the research protocol, and the need for informed consent was waived because of the retrospective nature of the study. All procedures involving human participants were conducted according to the Declaration of Helsinki (as revised in 2013). All methods were performed in accordance with the relevant guidelines and regulations. The data were accessed for research purposes between June 1, 2023 and October 31, 2023. The authors had no access to information that could identify individual participants during or after data collection. The inclusion criteria were as follows: (1) conformance to the diagnostic criteria of AMI and (2) echocardiography showing the interruption of the echo continuity of the interventricular septum. The exclusion criteria were as follows: (1) congenital ventricular septal defect, (2) malignant tumors, (3) autoimmune diseases, and (4) missing clinical data. Figure 1 shows a flowchart for screening the included patients.
Figure 1.
Flowchart for screening the included patients. AMI, acute myocardial infarction.
Clinical data and outcome information collection
The clinical indicators were obtained from the hospital’s electronic medical records system and the results of all the tests and examinations were obtained within 24 h after the patient was diagnosed with PIVSR. The collected data included gender, age, hypertension, diabetes, history of angina, smoke, anterior wall infarction, Killip III-IV, use of intra-aortic balloon pump (IABP), use of extracorporeal membrane oxygenation (ECMO), use of continuous renal replacement therapy (CRRT), mechanical ventilation, primary percutaneous coronary intervention(PPCI), admission heart rate (HR), admission systolic blood pressure (SBP), laboratory examination—white blood cell(WBC) counts, neutrophil counts, lymphocyte counts, monocyte counts, hemoglobin, platelet counts, C-reactive protein(CRP), creatinine, glucose, alanine aminotransferase (ALT), aspartate aminotransferase (AST), total bilirubin, fibrinogen, D dimer, N-terminal pro-B-type natriuretic peptide (NT proBNP), lactic acid; ultrasonic indicators—left ventricular end-diastolic diamension(LVEDD), left ventricular ejection fraction(LVEF), site of ventricular septal rupture(apex of VSR), rupture size, pulmonary artery systolic pressure (PASP), ventricular aneurysm, and methods of treatment including percutaneous closure, surgical repair and medical management. The methods of treatment were discussed between a cardiac surgeon, cardiologist, and cardiac intensivist and the patient and families were consulted. The principle was that according to the patient’s condition, the first consideration was whether or not to accept surgical repair. For patients who couldn’t undergo surgical repair, if the diameter of the rupture is not more than 24mm, percutaneous closure could be considered. If both invasive treatments couldn’t be accomplished or were not considered, then only medical management will be used. Information on whether the patient died within 30 days was partly obtained from the patient’s hospitalization information and partly followed up by telephone. The patients were divided into deceased and survival groups according to whether they died within 30 days.
Statistical analysis
Measurement data conforming to the normal distribution were expressed as mean ± standard deviation, measurement data not conforming to the normal distribution were expressed as median (interquartile range), and the count data were expressed as frequency and percentage (%). Two independent sample mean t-test was used to compare the measurement data with normal distribution and equal variance between the 2 groups. Non- parametric Mann–Whitney U test was used to compare the measurement data without normal distribution or equal variance between the 2 groups. Person’s chi-square test was used to compare the enumeration data between groups. This was analyzed using the CBCgrps package (version 2.8.2)15. The Pearson method of the COR function was used to detect correlations between measurement data. The Corrplot package (version 0.92) was used to generate the heat map of the correlation coefficients. Least absolute shrinkage and selection operator (LASSO) analysis was used to screen the variables by using the glmnet package (version 4.1-7). A multivariate logistic regression analysis was performed using the glm function in the stats package. A nomogram was constructed using the nomogram function of the rms package (version 6.6-0). The discrimination ability of the nomogram was determined by calculating the area under the curve (AUC). A receiver operating characteristic (ROC) curve was drawn using the pROC function in the rms package. The Hosmer–Lemeshow goodness of fit test was used to evaluate the calibration using the ResourceSelection package (version 0.3-6), and the calibration curve was drawn using the rms package. The decision curve analysis (DCA) curve was used to evaluate the clinical practicability and was performed using the rmda package (version 1.6). The bootstrap method (resampling = 1000) was used for internal validation. All statistical analyses were conducted using R software (version 4.2.2), and statistical significance was set at p < 0.05, except for the Hosmer–Lemeshow goodness of fit (p > 0.05).
Results
Of the 213 patients who met the inclusion criteria, 7 patients met the exclusion criteria, including 1 patient with congenital VSR, 3 patients with malignant tumors, and 3 patients with missing laboratory examination data. Finally, 206 patients were included in the study, including 108 patients (52.4%) in the deceased group and 98 patients in the survival group. Table 1 shows a comparison of the clinical data between the 2 groups. There were significant differences (p < 0.05) in age, Killip III-IV, IABP, ECMO, CRRT, mechanical ventilation, PPCI, admission HR, admission SBP, as well as WBC, neutrophil, monocyte, CRP, creatinine, glucose, ALT, AST, total bilirubin, D dimer, NT proBNP, lactic acid, PASP, and methods of treatment.
Table 1.
Comparison of clinical data between decreased group and survival group.
| Variables | Overall (N = 206) | Deceased group (N = 108) | Survival group (N = 98) | p value |
|---|---|---|---|---|
| Female (%) | 109 (53) | 59 (55) | 50 (51) | 0.71 |
| Age (years) | 68 (63, 73) | 70 (64, 75) | 67 (61, 71) | 0.04 |
| Hypertension (%) | 113 (55) | 55 (51) | 58 (59) | 0.29 |
| Diabetes (%) | 71 (34) | 38 (35) | 33 (34) | 0.94 |
| History of angina (%) | 78 (38) | 35 (32) | 43 (44) | 0.12 |
| Smoke (%) | 66 (32) | 30 (28) | 36 (37) | 0.22 |
| Anterior wall infarction (%) | 164 (80) | 89 (82) | 75 (77) | 0.38 |
| Killip III-IV (%) | 147 (71) | 92 (85) | 55 (56) | < 0.01 |
| IABP (%) | 143 (69) | 84 (78) | 59 (60) | 0.01 |
| ECMO (%) | 27 (13) | 24 (22) | 3 (3) | < 0.01 |
| CRRT (%) | 84 (41) | 65 (60) | 19 (19) | < 0.01 |
| Mechanical ventilation (%) | 79 (38) | 59 (55) | 20 (20) | < 0.01 |
| PPCI (%) | 153 (74) | 68 (63) | 85 (87) | < 0.01 |
| Admission HR (beat/min) | 96 (85, 110) | 102 (90, 117) | 90 (83, 100) | < 0.01 |
| Admission SBP (mmHg) | 107 ± 16 | 104 ± 17 | 109 ± 16 | 0.03 |
| WBC (109/L) | 12.4 (8.7, 16.0) | 14.6 (11.2, 17.8) | 9.7 (6.9, 13.1) | < 0.01 |
| Neutrophil (109/L) | 10.1 (6.6, 14.0) | 12.0 (9.1, 15.1) | 7.7 (5.1, 10.9) | < 0.01 |
| Lymphocyte (109/L) | 1.3 (1.0, 1.7) | 1.3 (0.9, 1.7) | 1.4 (1.0, 1.8) | 0.16 |
| Monocyte (109/L) | 0.8 (0.6, 1.0) | 0.9 (0.6, 1.1) | 0.7 (0.5, 0.9) | < 0.01 |
| Hemoglobin (g/L) | 120.5 ± 16.9 | 119.5 ± 17.2 | 121.6 ± 16.6 | 0.36 |
| Platelet (109/L) | 228 (169, 299) | 220 (164, 289) | 252 (179, 312) | 0.16 |
| CRP (mg/L) | 47.4 (16.0, 95.2) | 68.6 (26.4, 110.5) | 31.2 (7.7, 62.9) | < 0.01 |
| Creatinine (μmol/L) | 96 (75, 139) | 111 (83, 158) | 88 (71, 113) | < 0.01 |
| Glucose (mmol/L) | 8.6 (6.7, 11.8) | 9.3 (7.2, 13.1) | 7.3 (6.2, 9.9) | < 0.01 |
| ALT (U/L) | 54 (25, 397) | 133 (47, 1110) | 29 (17, 80) | < 0.01 |
| AST (U/L) | 82 (30, 533) | 284 (71, 1201) | 32 (20, 114) | < 0.01 |
| Total bilirubin (μmol/L) | 16.0 (9.7, 23.2) | 16.6 (10.6, 27.6) | 14.8 (8.8, 20.2) | 0.01 |
| Fibrinogen (g/L) | 3.7 (3.0, 4.8) | 3.7 (2.7, 5.0) | 3.7 (3.1, 4.8) | 0.69 |
| D dimer (mg/L) | 1.9 (1.0, 4.8) | 2.4 (1.2, 5.6) | 1.5 (0.9, 4.0) | 0.02 |
| NT proBNP (pg/ml) | 2050 (885, 3253) | 2555 (1194, 3870) | 1505 (807, 2638) | < 0.01 |
| Lactic acid (mmol/L) | 2.1 (1.4, 3.5) | 2.9 (2.0, 5.1) | 1.5 (1.1, 2.2) | < 0.01 |
| LVEDD (mm) | 49 (45, 53) | 48 (44, 52) | 49 (45, 55) | 0.28 |
| LVEF (%) | 54 (47, 59) | 54 (46, 59) | 54 (48, 59) | 0.7 |
| Apex of VSR (%) | 141 (68) | 74 (69) | 67 (68) | 0.99 |
| rupture size (mm) | 15 (10, 18) | 13 (10, 18) | 15 (10, 18) | 0.44 |
| PASP (mmHg) | 48 ± 14 | 54 ± 12 | 43 ± 14 | < 0.01 |
| Ventricular aneurysm (%) | 90 (44) | 41 (38) | 49 (50) | 0.11 |
| Methods of treatment | < 0.01 | |||
| Percutaneous closure (%) | 81 (39) | 22 (20) | 59 (60) | |
| Surgical repair (%) | 50 (24) | 17 (16) | 33 (36) | |
| Medical management (%) | 75 (36) | 69 (64) | 6 (6) |
ALT, alanine aminotransferase; AST, aspartate aminotransferase; CRP, C-reactive protein; CRRT, continuous renal replacement therapy; ECMO, extracorporeal membrane oxygenation; HR, heart rate; IABP, Intra-aortic balloon pump; LVEDD, left ventricular end-diastolic dimension; LVEF, left ventricular ejection fraction; NT proBNP, N-terminal prohormone of brain natriuretic peptide; PASP, pulmonary artery systolic pressure; PPCI, primary percutaneous coronary intervention; SBP, systolic blood pressure; VSR, ventricular septal rupture; WBC, white blood cell.
Correlation analysis for numerical variables of clinical data
To evaluate whether there were collinear variables in the clinical data, Spearman’s correlation analysis between variables was performed on multiple numerical variables in Table 1, and the results are shown in Supplementary Figure S1. From the heat map, the depth of the color in the picture represents the size of the Spearman correlation coefficient value. Dark colors indicate high correlations, and light colors indicate low correlations. The redder the color, the closer the Spearman correlation coefficient is to 1, and the bluer the color, the closer the Spearman correlation coefficient is to − 1. The asterisk represents P < 0.05. As shown in Supplementary Figure S1, there was a significant correlation between the multiple numerical variables.
LASSO analysis screening characteristic variables
Owing to the collinearity between several numerical variables, LASSO analysis was conducted to solve the multicollinearity problem and avoid overfitting. Figure 2A shows the LASSO coefficient path diagram for screening the variables, and Fig. 2B shows the LASSO cross-validation curve. The dashed line on the right is λ_1SE, corresponding to non-zero coefficient variables were considered for further multivariate analysis. The 10 selected variables were Killip III-IV, ECMO, CRRT, mechanical ventilation, PPCI, admission HR, WBC, glucose, PASP and methods of treatment.
Figure 2.
Variables selection using the LASSO analysis. (A) LASSO coefficient profiles of the radiomic features. Each colored line represented the coefficient of each variables. (B) Optimal parameter (λ) selection in the LASSO analysis used tenfold cross validation. The λ_1SE on the right side of the dotted line was selected as the final equation screening criterion. LASSO, least absolute shrinkage and selection operator; SE, standard error.
Multivariate logistic regression analysis of independent risk factors
On the basis of the variables selected by the LASSO analysis, binary logistic regression analysis was performed to establish a regression model to predict the risk of death within 30 days in patients with PIVSR. Finally, the selected variables with statistical significance were CRRT, mechanical ventilation, PPCI, WBC, PASP and methods of treatment (Table 2).
Table 2.
Multivariate logistic regression analysis on the risk of death within 30 days for patients with PIVSR.
| Variables | Coefficient | OR (95% CI) | p value |
|---|---|---|---|
| Killip III–IV | 0.56 | 1.74 (0.54, 5.96) | 0.36 |
| ECMO | 2.56 | 12.97 (0.92, 315.65) | 0.10 |
| CRRT | 1.76 | 5.80 (1.82, 20.62) | < 0.01 |
| Mechanical ventilation | 1.35 | 3.86 (1.24, 13.00) | 0.02 |
| PPCI | − 2.55 | 0.08 (0.02, 0.28) | < 0.01 |
| HR | 0.02 | 1.02 (0.98, 1.05) | 0.30 |
| WBC (109/L) | 0.15 | 1.17 (1.04, 1.32) | 0.01 |
| Glucose (mmol/L) | 0.12 | 1.14 (0.99, 1.32) | 0.07 |
| PASP (mmHg) | 0.07 | 1.07 (1.03, 1.12) | < 0.01 |
| Methods of treatment | |||
| Medical management | Reference | ||
| Percutaneous closure | − 4.27 | 0.01 (0.00,0.06) | < 0.01 |
| Surgical repair | − 3.92 | 0.02 (0.00,0.09) | < 0.01 |
OR, odds ratio; CI, confidence interval; ECMO, extracorporeal membrane oxygenation; CRRT, continuous renal replacement therapy; PPCI, primary percutaneous coronary intervention; HR, heart rate; WBC, white blood cell; PASP, pulmonary artery systolic pressure.
The nomogram to present the prediction model
A nomogram was drawn according to the risk factors determined by multivariate analysis, and the results are shown in Fig. 3. The 4 categorical variables of CRRT, mechanical ventilation, PPCI, methods of treatment corresponded to “0 points” or “scores” according to “yes” or “no”. The numerical variables of WBC and PASP were based on different measured values corresponding to different scores according to different scales on the number axis of nomogram. The scores of the 6 variables were summed to obtain the total score, and the predictive probability of death within 30 days in patients with PIVSR was obtained. For example, a patient with PIVSR: if CRRT was used during hospitalization, scored 35 points; mechanical ventilation was performed, scored 21 points; PPCI was operated, scored 0 point; the WBC count was 10 × 109/L, scored 27 points; the PASP was 30mmHg, scored 20 points; the treatment was medical management, scored 66 points. The total score of the patient was 169. According to the nomogram, the 30-day mortality corresponding to the total score of 169 was 0.95, which meant that the probability of death within 30 days of the patient was 95%.
Figure 3.
30-day mortality nomogram for patients with PIVSR. The nomogram model included CRRT, mechanical ventilation, PPCI, WBC, PASP and methods of treatment. The patient was evaluated according to the specific data of each variable, and a total score was added according to the nomogram. Based on these total points, the risk of death within 30 days could be predicted. For example, a patient with PIVSR: if CRRT was used during hospitalization, scored 35 points; mechanical ventilation was performed, scored 21 points; PPCI was operated, scored 0 point; the WBC count was 10 × 109/L, scored 27 points; the PASP was 30 mmHg, scored 20 points; the treatment was medical management, scored 66 points. The total score of the patient was 169. According to the nomogram, the 30-day mortality corresponding to the total score of 169 was 0.95, which meant that the probability of death within 30 days of the patient was 95%. CRRT, continuous renal replacement therapy; PASP, pulmonary artery systolic pressure; PIVSR, post-infarction ventricular septal rupture; PPCI, primary percutaneous coronary intervention; WBC, white blood cell.
Meanwhile, we constructed a dynamic nomogram of the web page. The corresponding website was https://zhaeng.shinyapps.io/DynNomapp/. Figure 4 shows the content of the web page. In the web page, the value of the 6 predictive variables can be freely selected, and then click the Predict button to make a prediction. The results will be automatically calculated and present graphically in the Graphical Summary and Numerical Summary window. The Model Summary window shows the fitting results of the model.
Figure 4.
A web-based dynamic nomogram to predict the 30-day mortality for patients with PIVSR. In the dynamic nomogram, there were 6 predictive variables whose values could be freely selected and then clicked the Predict button on the left side, the results will be automatically calculated. The results were shown on the right side. In the Graphical Summary window, the horizontal lines represented the prediction and when the mouse pointer was close to the square in the line, the window would display the value of each variable and the specific probability it predicted and 95% confidence interval. The Numerical Summary window showed the results of specific values. The Model Summary window showed the fitting results of the model. The website corresponding to the dynamic nomogram was https://zhaeng.shinyapps.io/dynnomapp/. CRRT, continuous renal replacement therapy; PASP, pulmonary artery systolic pressure; PIVSR, post-infarction ventricular septal rupture; PPCI, primary percutaneous coronary intervention; WBC, white blood cell.
Evaluation of the nomogram model
AUC to evaluate the accuracy of the model
The ROC curve of death within 30 days in patients who underwent PIVSR was drawn (Fig. 5A) to evaluate the accuracy of the nomogram model. The area under the ROC curve was 0.96 (95% confidence interval 0.93–0.98), thus suggesting that the prediction model had high accuracy and good discrimination.
Figure 5.
Evaluation of the nomogram model. (A) ROC curves of the nomogram. The area under the ROC curve was 0.96 (95% CI: 0.93–0.98). (B) The calibration curve of the nomogram. It showed that the apparent curve and the bias-corrected curve were close to the ideal curve, which indicated that the predicted probability of death was highly consistent with the actual probability. (C) DCA curves of the nomogram. The threshold probability is 15%–95%, and the nomogram for predicting 30-day mortality has a high net benefit. AUC, area under the curve; ROC, receiver operating characteristic; DCA, decision curve analysis; CI, confidence interval.
Calibration curve to evaluate the calibration degree of the model
The calibration curve was used to evaluate the calibration of the nomogram model to predict 30-day mortality in patients with PIVSR, as shown in Fig. 5B. The Hosmer–Lemeshow goodness of fit test (χ2 = 10.62, P = 0.22) had good consistency. The calibration curve shows that the apparent and bias-corrected curves were close to the ideal curve, thus indicating that the predicted probability was highly consistent with the actual probability.
DCA curve to evaluate the clinical practicability of the model
Figure 5C shows the DCA curve of the nomogram model for 30-day mortality in patients with PIVSR. It can be observed from the figure that in a large range of abscissa intervals, the curve of the prediction model was above the 2 invalid lines, and the model effect was acceptable. Specifically, when the threshold probability was 15%– 95%, the nomogram for predicting 30-day mortality had a higher net benefit.
Validation of the nomogram model
To avoid a reduction in sample size caused by cross-validation, we used the bootstrap method to perform 1000 instances of resampling to validate the prediction model internally. Supplementary Figure S2 shows the ROC curve for 1000 instances of resampling. The dark blue curve in the figure represents the ROC curve corresponding to the original predicted value and outcome event, and the light blue curve represents the set of ROC curves of the samples generated by bootstrap resampling. Given that the pROC package in the R project also uses the bootstrap method when calculating the confidence interval, the AUC confidence interval in Supplementary Figure S2 was consistent with the confidence interval in Fig. 5A, which had high accuracy.
Discussion
PIVSR is a catastrophic mechanical complication with high mortality rate. Although the incidence of PIVSR has further decreased since the era of PCI, there has been no significant decrease in the associated mortality rates over the past 2 decades6. The number of deaths from PIVSR still accounted for 5% of all deaths from AMI16. As an important mechanical complication after AMI, the analysis of the risk factors for early death due to PIVSR has an important influence on the overall prognosis of AMI. However, only a few studies have described risk prediction models for death within 30 days in patients with PIVSR.
In this study, LASSO and multivariate logistic regression analysis were used to screen out factors to construct a nomogram prediction model, which is considered superior for selecting predictors by only univariate analysis11. The results of the nomogram were scored on these 6 risk factors, including CRRT, mechanical ventilation, PPCI, WBC, PASP and methods of treatment. The points of the factors were added to obtain the total points, and the prediction probability of 30-day mortality in patients with PIVSR was obtained corresponding to the total points. Or, it could be automatically calculated in the web-based dynamic nomogram after the values of 6 predictive variables were selected. Meanwhile, the prediction model was comprehensively evaluated. It showed that there was good performance in the accuracy, calibration and clinical practicality of the nomogram model. This study established a simple and intuitive nomogram, which was more convenient for clinical use and quickly obtained the specific risk probability.
The use of CRRT and mechanical ventilation are important factors in evaluating patients with PIVSR. In a recent systematic review of PIVSR, a total of 8579 patients were included in the analysis, and it was concluded that renal failure was proved to be a predictor of mortality. The best treatment for patients with renal failure is renal replacement therapy17. An analysis from Spain’s national database showed that VSR occurred in 126 patients hospitalized for AMI between 2010 and 2015, with an in-hospital mortality rate of 59.5%. Cardiopulmonary failure is one of the most important risk factors for in-hospital death18. The vast majority of these patients with respiratory dysfunction require mechanical ventilation support. CRRT and mechanical ventilation were also screened as significant predictors of death in patients with PIVSR in the prediction model of this study, which was consistent with these studies.
PPCI is currently the main treatment for patients with AMI and greatly reduces patient mortality. Patients with AMI underwent PPCI, which indicated that these patients could reach the hospital for reperfusion therapy in time to open the infarcted blood vessels after the onset of symptoms. PIVSR usually occurs 3–5 days after AMI19. Patients with PIVSR after reperfusion therapy have theoretical advantages, which may improve the blood supply in the watershed area around the infarcted myocardium edge, accelerate the hardening of the ventricular septal tissue around the infarct, and increase the possibility of subsequent successful closure of the defect. Emergency surgery may be performed to repair VSR along with coronary artery bypass grafting, but mortality rate is high. Our study showed that PPCI can reduce the 30-day mortality risk of PIVSR patients. PIVSR may have been found at the same time as the diagnosis of AMI, which is before PPCI; or may occur within 1 week after AMI, which is after PPCI. The effect of PPCI on PIVSR is worthy of further study.
Elevated WBC count is a possible indicator of inflammation. In recent years, an increasing number of studies have clarified that inflammation plays an important role in the development of AMI14,15, and clinical trials on anti-inflammatory treatments for AMI are in full swing. A retrospective observational study that included patients who underwent PIVSR from June 2012 to July 2021 showed that elevated WBC counts were significantly associated with mortality20. Another summary study involving authors with more than 10 years of experience in the percutaneous closure of PIVSR identified all patients who underwent percutaneous closure between 2003 and 2016 in 8 participating centers, thus indicating that WBC count had an effect on 30-day mortality21. Considering the role of inflammation in the development of mechanical complications after AMI, it is not surprising that total WBC count may be a predictor of PIVSR. This study also included WBC count as one of the important factors to predict the 30-day mortality in PIVSR patients. As the white blood cell count increased, the corresponding score in the nomogram increased and the probability increased.
After PIVSR occurs, there is an acute left-to-right shunt in cardiac blood flow and a significant increase in blood flow of pulmonary circulation, which reflexively causes pulmonary vasospasm and abnormal contraction and diastole of the pulmonary vasculature, leading to dynamic pulmonary hypertension. Meanwhile, excessive right ventricular volume, excessive right ventricular expansion, and increased right ventricular wall tension will reduce the right ventricular contractility, aggravate tricuspid regurgitation, and increase the related constraints of the left and right ventricles, leading to a restriction of left ventricular filling and output ultimately. In severe cases, pulmonary congestion and edema may occur, thus aggravating the heart burden caused by AMI and leading to low cardiac output and cardiogenic shock. Therefore, it is not difficult to understand that the higher the PASP, the greater the risk of death within 30 days. Chen et al.22 indicated that the pulmonary artery pressure in patients after percutaneous closure was reduced, and the prognosis of patients improved. This study also showed that as PASP decreased, the points decreased, the cumulative total points decreased, and the 30-day mortality decreased.
The invasive treatment of PIVSR includes surgical repair and percutaneous closure. The current guidelines of the American Heart Association and European Society of Cardiology also recommend early surgical repair for hemodynamically unstable patients with PIVSR21,23. In practice, the treatment of PIVSR is determined by cardiac surgeons, cardiologists, and cardiac intensivists according to the patient’s condition, and in consultation with the patients and their families. Several studies have shown that improvements in surgical and catheter techniques decrease VSR mortality. A multicenter retrospective study in France collected the clinical data of patients with VSR admitted to 3 university-affiliated hospitals from 2008 to 2019 and compared the characteristics of patients who underwent surgery and those who received only drug therapy. The results showed that although surgery had considerable risks, it significantly improved the 30-day and 1-year survival rates of the patients24. One study on patients with PIVSR from the UK National Registry concluded that both surgical repair and percutaneous closure are feasible options for the treatment of VSR25. There was no significant difference in long-term mortality after discharge between the 2 treatment methods, although the in-hospital mortality of the surgical repair group was relatively low. In our nomogram, 3 different methods of treatment correspond to different scores. The scores of percutaneous closure and surgical repair are lower than those of only medical treatment, and the predicted probability is lower, which is consistent with the actual situation. When communicating with the patient and their families, doctors can use this nomogram model to clearly inform them of 30-day motility for the patient with different methods of treatment in combination with the other current clinical features.
The nomogram based on the clinical characteristics of patients with PIVSR eventually had the above 6 variables, including CRRT, mechanical ventilation, PPCI, WBC, PASP and methods of treatment. The evaluation of nomogram showed that it had good discriminatory ability, highly consistent with the actual probability and clinical usefulness. Therefore, this nomogram can help clinicians predict the risk of death within 30 days in patients with PIVSR in a comprehensive and convenient way.
This predictive model has a few limitations. First, the cases in this study were obtained from a single center, and the sample size was small owing to the low incidence of VSR after AMI. Second, the nomogram model was established via a retrospective study. Finally, the prediction model was not validated by the external data. It requires external validations to enter the clinical scenario.
Conclusion
In conclusion, a nomogram model including 6 clinical features of CRRT, mechanical ventilation, PPCI, WBC, PASP and methods of treatment was well constructed to help predict the risk of death within 30 days in patients with PIVSR. This model can help clinicians comprehensively and conveniently predict the probability of short-term mortality in patients with PIVSR, and early identify high-risk patients with poor prognosis. Future studies would be warranted to validate the potential clinical benefits of this model.
Supplementary Information
Acknowledgements
We are grateful to the Information Center of Zhengzhou University Central China Fuwai Hospital (Fuwai Central China Cardiovascular Hospital) for providing the list of patients. We would also like to thank Editage (www.editage.cn) for the English language editing.
Author contributions
C.Y.G and Z.Z. designed the study and wrote the manuscript. Z.Z. and Y.H.L. were responsible for collecting, sorting and statistical data. Q.Q.C supervised the overall study and contributed to editing and review of the manuscript. J.Z. is the guarantor of this work and takes responsibility for the integrity of the data and the accuracy of the data analysis. All authors read and approved the final manuscript.
Data availability
The datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher's note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Supplementary Information
The online version contains supplementary material available at 10.1038/s41598-024-68792-y.
References
- 1.Elbadawi, A. et al. Temporal trends and outcomes of mechanical complications in patients with acute myocardial infarction. JACC Cardiovasc. Interv.12, 1825–1836 (2019). 10.1016/j.jcin.2019.04.039 [DOI] [PubMed] [Google Scholar]
- 2.Alfonso, F., Aguilar, R. & Reyes, G. Management of post-infarction ventricular septal defects: Are we moving forward?. Eur. Heart J.43, 5033–5036 (2022). 10.1093/eurheartj/ehac532 [DOI] [PubMed] [Google Scholar]
- 3.Honan, M. B. et al. Cardiac rupture, mortality and the timing of thrombolytic therapy: A meta-analysis. J. Am. Coll. Cardiol.16, 359–367 (1990). 10.1016/0735-1097(90)90586-E [DOI] [PubMed] [Google Scholar]
- 4.Zalewski, J., Nowak, K., Furczynska, P. & Zalewska, M. Complicating acute myocardial infarction. Current status and unresolved targets for subsequent research. J. Clin. Med.10, 5904 (2021). 10.3390/jcm10245904 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Arnaoutakis, G. J. et al. Surgical repair of ventricular septal defect after myocardial infarction: Outcomes from the Society of Thoracic Surgeons National Database. Ann. Thorac. Surg.94, 436–443 (2012). 10.1016/j.athoracsur.2012.04.020 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Gong, F. F., Vaitenas, I., Malaisrie, S. C. & Maganti, K. Mechanical complications of acute myocardial infarction: A review. JAMA Cardiol.6, 341–349 (2021). 10.1001/jamacardio.2020.3690 [DOI] [PubMed] [Google Scholar]
- 7.Aijaz, S. et al. Management and outcome of post-myocardial infarction ventricular septal rupture—A low-middle-income country experience. PLoS One17, e0276615 (2022). 10.1371/journal.pone.0276615 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Wang, L. et al. Clinical characteristics and contemporary prognosis of ventricular septal rupture complicating acute myocardial infarction: A single-center experience. Front. Cardiovasc .Med.8, 679148 (2021). 10.3389/fcvm.2021.679148 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Malhotra, A. et al. Techniques, timing & prognosis of post infarct ventricular septal repair: A re-look at old dogmas. Braz. J. Cardiovasc. Surg.32, 147–155 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Ni, Y. Q. et al. Risk factors for early death in acute myocardial infarction patients complicating with ventricular septal rupture. Zhonghua Xin Xue Guan Bing Za Zhi.46, 981–986 (2018). [DOI] [PubMed] [Google Scholar]
- 11.Crenshaw, B. S. et al. Risk factors, angiographic patterns, and outcomes in patients with ventricular septal defect complicating acute myocardial infarction. GUSTO-I (Global Utilization of Streptokinase and TPA for Occluded Coronary Arteries) Trial Investigators. Circulation101, 27–32 (2000). 10.1161/01.CIR.101.1.27 [DOI] [PubMed] [Google Scholar]
- 12.Swirski, F. K. & Nahrendorf, M. Cardioimmunology: The immune system in cardiac homeostasis and disease. Nat. Rev. Immunol.18, 733–744 (2018). 10.1038/s41577-018-0065-8 [DOI] [PubMed] [Google Scholar]
- 13.Prabhu, S. D. & Frangogiannis, N. G. The biological basis for cardiac repair after myocardial infarction: From inflammation to fibrosis. Circ. Res.119, 91–112 (2016). 10.1161/CIRCRESAHA.116.303577 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Navarese, E. P. et al. Development and validation of a practical model to identify patients at risk of bleeding after TAVR. JACC Cardiovasc. Interv.14, 196–1206 (2021). 10.1016/j.jcin.2021.03.024 [DOI] [PubMed] [Google Scholar]
- 15.Zhang, Z., Gayle, A. A., Wang, J., Zhang, H. & Cardinal-Fernández, P. Comparing baseline characteristics between groups: an introduction to the CBCgrps package. Ann. Transl. Med.5, 484 (2017). 10.21037/atm.2017.09.39 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Bajaj, A., Sethi, A., Rathor, P., Suppogu, N. & Sethi, A. Acute complications of myocardial infarction in the current era: Diagnosis and management. J. Investig. Med.63, 844–855 (2015). 10.1097/JIM.0000000000000232 [DOI] [PubMed] [Google Scholar]
- 17.Rashid, H. et al. Delayed ventricular septal rupture repair on patient outcomes after myocardial infarction: A systematic review. Curr. Probl. Cardiol.48, 101521 (2023). 10.1016/j.cpcardiol.2022.101521 [DOI] [PubMed] [Google Scholar]
- 18.Sanmartín-Fernández, M. et al. In-hospital outcomes of mechanical complications in acute myocardial infarction: Analysis from a nationwide Spanish database. Cardiol. J.28, 589–597 (2021). 10.5603/CJ.a2020.0181 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Damluji, A. A. et al. Mechanical complications of acute myocardial infarction: A scientific statement from the American Heart Association. Circulation144, e16–e35 (2021). 10.1161/CIR.0000000000000985 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Zhang, X. Y., Bian, L. Z. & Tian, N. L. The clinical outcomes of ventricular septal rupture secondary to acute myocardial infarction: A retrospective, Observational Trial. J. Interv. Cardiol.2021, 3900269 (2021). 10.1155/2021/3900269 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Sabiniewicz, R. et al. Percutaneous closure of post-infarction ventricular septal defects-an over decade-long experience. J. Interv. Cardiol.30, 63–71 (2017). 10.1111/joic.12367 [DOI] [PubMed] [Google Scholar]
- 22.Chen, T. F. et al. Efficacy and safety of percutaneous closure of ventricular septal rupture after acute myocardial infarction: A clinical study of 69 cases. Zhonghua Xin Xue Guan Bing Za Zhi.49, 1094–1101 (2021). [DOI] [PubMed] [Google Scholar]
- 23.Jones, B. M. et al. Ventricular septal rupture complicating acute myocardial infarction: A contemporary review. Eur. Heart J.35, 2060–2068 (2014). 10.1093/eurheartj/ehu248 [DOI] [PubMed] [Google Scholar]
- 24.Piriou, P. G. et al. Management and outcomes of ventricular septal defects after acute myocardial infarction: A multicenter retrospective study. J. Card. Surg.37, 5019–5026 (2022). 10.1111/jocs.17151 [DOI] [PubMed] [Google Scholar]
- 25.Giblett, J. P. et al. Post-infarction ventricular septal defect: Percutaneous or sugical management in the UK national registry. Eur. Heart J.43, 5020–5032 (2022). 10.1093/eurheartj/ehac511 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request.