Construction and validation of a nomogram prediction model for the risk of new-onset atrial fibrillation following percutaneous coronary intervention in acute myocardial infarction patients (2025)

Objective

The objective of this study was to investigate risk factors for new-onset atrial fibrillation (NOAF) post-percutaneous coronary intervention (PCI) in patients with acute myocardial infarction (AMI), aiming to develop a predictive nomogram for NOAF risk.

Methods

A retrospective cohort study involving 397 AMI patients who underwent PCI at a tertiary hospital in Anhui, China, from January 2021 to July 2022 was performed. Patients were divided into NOAF (n = 63) and non-NOAF (n = 334) groups based on post-PCI outcomes. Clinical data were extracted from the hospital information system (HIS) and analyzed using univariate and multivariate logistic regression to identify independent risk factors. A nomogram was generated utilizing R software (version 3.6.1), with its performance evaluated through receiver operating characteristic (ROC) curves, calibration curves, decision curve analysis (DCA), and Bootstrap resampling.

Results

Independent risk factors for NOAF included age, left atrial diameter (LAD), Gensini score, N-terminal pro-B-type natriuretic peptide (NT-proBNP), alanine transaminase (ALT), low-density lipoprotein cholesterol (LDL-C), left ventricular end-systolic diameter (LVESD), and ventricular rate (P < 0.05). The nomogram’s ROC curve demonstrated an area under the curve (AUC) of 0.925 (95% CI: 0.887–0.963), supported by a Bootstrap-verified AUC of 0.924 (95% CI: 0.883–0.954), reflecting strong discriminative capability. The calibration curve indicated a mean absolute error (MAE) of 0.031 and 0.017 prior to and following Bootstrap verification, respectively, signifying robust calibration. The DCA curve illustrated that the nomogram offered optimal clinical net benefit for patients with a threshold probability of NOAF ranging from 0.01 to 0.99.

Conclusion

The nomogram developed from independent risk factors for NOAF exhibits significant predictive accuracy and clinical relevance for evaluating the risk of NOAF in AMI patients following PCI, thereby enabling the identification of high-risk individuals for targeted interventions.

Study patients

This study employed a retrospective cohort design involving 397 AMI patients who received PCI at a tertiary hospital in Anhui Province, China, from January 2021 to July 2022. Inclusion criteria encompassed individuals over 18 years old, without a history of AF, who underwent emergency PCI within 24h of symptom onset and provided comprehensive clinical data during hospitalization. Exclusion criteria included patients with malignant tumors, non-obstructive coronary artery disease or primary cardiomyopathy, symptoms of clinical infection, autoimmune disorders, significant hepatic or renal impairment, a history of severe trauma, blood transfusions or surgeries within the six months preceding emergency PCI, and those diagnosed with anemia. Patients were categorized based on the development of NOAF post-PCI: the NOAF group (n = 63) and the non-NOAF group (n = 334).

Study methods

Data collection

Clinical data from AMI patients during hospitalization were retrospectively gathered using the Hospital Information System (HIS). This dataset encompassed gender, age, smoking and drinking histories, admission heart rate, systolic and diastolic blood pressure at admission, complications (including hypertension, diabetes, stroke, and dyslipidemia), color Doppler echocardiography results (including left ventricular ejection fraction (LVEF), left atrial diameter (LAD), left ventricular end-systolic diameter (LVESD), and left ventricular end-diastolic diameter (LVEDD)), myocardial infarction type, Gensini score, culprit vessels, number of diseased vessels, postoperative ventricular rate, and various laboratory indicators (including hemoglobin, white blood cell count, high-sensitivity C-reactive protein (hs-CRP), cardiac troponin I (cTn I), N-terminal pro-B-type natriuretic peptide (NT-proBNP), alanine transaminase (ALT), fasting blood glucose, glycated hemoglobin, triglycerides, cholesterol, high-density lipoprotein cholesterol (HDL-C), and low-density lipoprotein cholesterol (LDL-C)), totaling 32 distinct indicators. Hypertension is defined by the diagnostic criteria outlined in the Chinese Guidelines for the Prevention and Treatment of Hypertension (2018 Revised Edition) [10]: a systolic blood pressure of ≥ 140mm Hg (1mm Hg = 0.133kPa) and/or a diastolic blood pressure of ≥ 90mm Hg. Diabetes is identified according to the Guidelines for the Prevention and Treatment of Type 2 Diabetes in China (2020 Edition) [11], with diagnostic criteria including fasting blood sugar ≥ 7.0 mmol/L or postprandial random blood glucose ≥ 11.1 mmol/L. Dyslipidemia encompasses hypercholesterolemia, hypertriglyceridemia, and low levels of high-density lipoprotein cholesterol. The Gensini score, utilized for assessing coronary artery disease severity, is derived from coronary angiography results [12]. Approval for this study was granted by the Medical Research Ethics Committee of the First Affiliated Hospital of University of Science and Technology of China (approval number: 2024-RE-386). Additionally, the study has been registered in the Chinese Clinical Trial Registry under the registration number ChiCTR2400091026. In light of the study’s retrospective design, informed consent from participants was waived.

Diagnostic criteria for NOAF after PCI

NOAF was defined as the absence of a prior history of AF and the occurrence of first AF during hospitalization following PCI. AF assessment utilized data from in-hospital cardiac monitoring, Holter recordings, or 12-lead electrocardiograms (ECGs), characterized by the absence of P waves, irregular R-R intervals, and an irregular rhythm persisting for over 30s [13]. All patients identified with NOAF in this study exhibited paroxysmal atrial fibrillation, marked by sudden onset and resolution, typically lasting no longer than 48h, with episodes that may self-terminate and recur at irregular intervals.

Statistical methods

Data analysis employed SPSS 22.0 (SPSS Software, IBM) and R software (version 3.6.1, https://www.r-project.org/). Categorical data were presented as counts and percentages, with chi-square tests applied for inter-group comparisons. Measurement data adhering to a normal distribution were reported as (x̅±s), with independent sample t-tests utilized for comparisons between groups. The least absolute shrinkage and selection operator (LASSO) regression method facilitated the selection of characteristic variables for NOAF in AMI patients following PCI, while both univariate and multivariate logistic regression models were applied to identify independent risk factors associated with NOAF in this population. The “rms” package in R was employed to create a nomogram and construct its calibration curve, while the Hosmer-Lemeshow deviation test assessed the model’s calibration. Receiver operating characteristic (ROC) curves were generated using the “pROC” package, with the area under the ROC curve (AUC) serving as a metric for the nomogram’s discrimination capability. The “rmda” package facilitated the plotting of the decision curve analysis (DCA) curve to evaluate the model’s clinical applicability. To mitigate overfitting, the Bootstrap method conducted 1000 re-sampling iterations for internal validation of the nomogram’s predictive performance. Statistical significance was determined at P < 0.05.

Univariate analysis of the risk of NOAF after PCI in AMI patients

Comparison of clinical data between the NOAF and non-NOAF groups revealed statistically significant differences in the distribution of 18 variables, including age, systolic blood pressure at admission, LVEF, LAD, Gensini score, hs-CRP, cTnI, NT-proBNP, ALT, fasting blood sugar, LDL-C, LVEDD, LVESD, ventricular rate, smoking history, diabetes history, dyslipidemia, and stroke history (P < 0.05). These variables served as potential risk factors for postoperative NOAF in AMI patients after PCI, as detailed in Table1.

Screening of characteristic variables for the occurrence of NOAF after PCI in AMI patients

Eighteen potential risk factors identified in Table1 may exhibit multicollinearity among variables, while the limited number of NOAF cases necessitated dimensionality reduction of these factors through LASSO regression. This approach isolated the most representative risk factors for NOAF occurrence. Optimal lambda parameters were determined using 10-fold cross-validation, with the Lambda.1se value corresponding to the minimum cross-validation error selected as the model’s optimal value. The count of variables with non-zero regression coefficients at this optimal value was conducted. LASSO regression indicated a Lambda.1se value of 0.056 at the minimum cross-validation error. The identified characteristic variables for NOAF occurrence in AMI patients post-PCI included LAD, ALT, ventricular rate, Gensini score, LVESD, age, LDL-C, and NT-proBNP, as illustrated in Fig.1.

Multivariate Logistic Regression Analysis results of the occurrence of NOAF after PCI in AMI patients

Eight characteristic variables identified through LASSO regression underwent further evaluation via univariate and multivariate logistic regression to determine independent risk factors for NOAF in AMI patients post-PCI. Both analyses revealed that age, LAD, Gensini score, NT-proBNP, ALT, LDL-C, LVESD, and post-PCI ventricular rate served as independent risk factors for NOAF occurrence in this patient cohort (P < 0.05), as illustrated in Table2.

Construction of a Nomogram Model for the risk of NOAF after PCI in AMI patients

Utilizing the logistic regression analysis of eight indicators presented in Table2, a nomogram prediction model was developed to assess the risk of NOAF in AMI patients post-PCI, employing the “rms” package in R software, as illustrated in Fig.2. The nomogram comprised eight key variables: age, LAD, Gensini score, NT-proBNP, ALT, LDL-C, LVESD, and ventricular rate. For each patient, a vertical line was drawn from each predictive variable on the horizontal axis to its corresponding score on the “Points” axis, yielding a total of eight individual scores. The aggregate of these scores provided the total score value. This total score was then located on the “Total Points” axis, from which a vertical line was drawn to intersect the “Risk of AF” axis, indicating the predicted risk of NOAF for that patient. Besides, among the eight indicators, including age, LAD, Gensini score, NT-proBNP, ALT, LDL-C, LVESD, and ventricular rate, the top three important indicators are ALT, LAD, and ventricular rate. These three indicators have higher ratings in the nomogram compared to the other indicators.

Clinical applicability analysis of the Nomogram Prediction Model for the risk of NOAF

The DCA curve served as an essential instrument for evaluating the clinical utility of prediction models, such as the presented nomogram. Figure3 illustrated the DCA curve’s capacity to assess net clinical benefits from various intervention strategies across a spectrum of threshold probabilities for NOAF occurrence in AMI patients post-PCI.

The DCA curve indicated that when the threshold probability of NOAF occurrence after PCI in AMI patients fell between 0.01 and 0.09, the nomogram-guided intervention strategy offered a markedly greater net clinical benefit than both the “Treat All” and “Treat None” approaches. This range was particularly noteworthy as it highlighted a probability interval where the nomogram enhances the number of patients receiving beneficial interventions, thereby reducing the risks linked to unnecessary treatments.

The “Treat All” strategy posited that all patients would receive the intervention irrespective of risk, while the “Treat None” strategy implied that no patients would receive the intervention. The nomogram’s performance exceeding these two extreme strategies within the 0.01 to 0.99 probability range demonstrated its effectiveness in identifying patients at moderate risk of NOAF, where intervention yielded the greatest benefit. Specifically, the nomogram’s ability to accurately predict the risk of NOAF within this range allowed for a more targeted approach to intervention, optimizing resource allocation and improving patient outcomes. Additionally, the 0.01 to 0.99 probability range encompasses a broad spectrum of risk levels, enabling healthcare providers to make informed decisions about intervention based on individual patient characteristics and risk profiles.

Discrimination ability analysis of the nomogram prediction model

The ROC curve assessed the discrimination capability of the nomogram model. To mitigate the risk of overfitting, the Bootstrap resampling method conducted internal validation of the model through 1000 iterations. The ROC analysis revealed an AUC of 0.925 (95% CI: 0.887–0.963) prior to internal validation and 0.924 (95% CI: 0.883–0.954) afterward, indicating robust discrimination capability, as illustrated in Fig.4.

The calibration ability analysis of the nomogram prediction model

The calibration curve assessed the nomogram model’s accuracy. The mean absolute error (MAE) between predicted probabilities and actual occurrences of NOAF was 0.031 before internal validation and 0.017 afterward. The Hosmer-Lemeshow deviation test indicated no statistically significant prediction deviation (P > 0.05), demonstrating the model’s strong calibration performance, as illustrated in Fig.5.

The incidence of NOAF in AMI patients post-PCI in this study was 15.87%, aligning with prior research reports [14,15,16]. In AMI patients, postoperative NOAF can lead to worsened cardiac function and heightened ischemic stroke risk, adversely affecting prognosis. Research by MAAGH et al. indicates that AF increases hospital mortality among AMI patients [17]. Additionally, BEN et al. highlighted that NOAF frequently complicated acute coronary syndrome (ACS), resulting in both short- and long-term negative outcomes [18]. WORME et al. demonstrated that NOAF significantly elevated the risk of death and major adverse cardiovascular events (MACE) compared to patients without AF in the context of ACS [19]. Consequently, employing scientifically validated risk screening tools is vital for identifying high-risk AMI patients with NOAF post-PCI and implementing effective prevention strategies.

Both univariate and multivariate logistic regression analyses in this study identified age, LAD, Gensini score, NT-proBNP, ALT, LDL-C, LVESD, and ventricular rate after PCI as independent risk factors for postoperative NOAF in AMI patients (P < 0.05). Among the eight indicators, the top three important indicators are ALT, LAD, and ventricular rate. These three indicators have higher ratings in the nomogram compared to the other indicators. These three indicators are critical due to their direct association with cardiac function and stress. Elevated ALT levels may reflect myocardial injury or ischemia, increasing the risk of NOAF. LAD diameter is a marker of coronary artery disease severity, with larger diameters indicating higher risk. Faster ventricular rates can exacerbate myocardial ischemia and electrical instability, promoting NOAF. Together, these indicators provide a comprehensive assessment of the patient’s cardiac condition, making them more predictive than other factors. Age has been consistently recognized as a risk factor for NOAF in prior research [20, 21]. Increasing age correlates with myocardial degeneration and fibrosis, resulting in reduced myocardial autonomy, excitability, and conductivity, thereby elevating the risk of atrial fibrillation. Additionally, LVESD and LAD have been validated as predictors for the progression and recurrence of atrial fibrillation following catheter ablation [22]. An increased end-systolic diameter of the left ventricle necessitates greater cardiac power to propel blood into the aorta, elevating pressure load on the heart and contributing to myocardial hypertrophy and dilation. This condition restricts left ventricular systolic function, impedes blood flow, and elevates cardiac volume load, resulting in myocardial damage and electrophysiological abnormalities that heighten the risk of AF [23, 24]. A meta-analysis indicates that an LAD greater than 50mm serves as a significant predictor of AF recurrence [25]. Left atrial size is a critical indicator of atrial structural remodeling and fibrosis, with the primary consequence of atrial fibrosis being the induction of abnormal electrical activity, thereby triggering AF. Research demonstrates that an increased LAD in AMI patients correlates with a heightened risk of NOAF [26]. An observational study found that patients with cardiovascular disease and AF exhibited higher ALT levels compared to those without AF (P < 0.05) [27]. Elevation of ALT is typically regarded as an indicator of liver injury or disease, including conditions like hepatitis and cirrhosis. The liver plays a central role in drug metabolism; impaired liver function may reduce the metabolism and clearance of medications, leading to increased blood concentrations of these drugs. Consequently, patients using specific medications, such as bisphosphonates and opioids, may face a heightened risk of AF [28, 29]. The Gensini score serves as a standard tool for assessing the severity of coronary artery disease; a higher score correlates with more advanced disease, poorer collateral circulation, and increased myocardial ischemia. Thus, cardiac function in these patients is more susceptible to damage, resulting in more pronounced activation of the renin-angiotensin-aldosterone system and an elevated risk of AF [30, 31]. NT-proBNP, a fragment of brain natriuretic peptide, primarily resides in myocardial and brain tissue cells. In AMI patients, myocardial ischemia and hypoxia induce vascular remodeling, coronary artery spasm, and elevated blood viscosity, resulting in increased atrial pressure, atrial stretch, and subsequent NT-proBNP release, thereby raising its serum levels [32]. Literature [33] indicates that elevated NT-proBNP significantly heightened the risk of AF. Furthermore, NT-proBNP expression in AF patients markedly exceeds that in individuals with sinus rhythm. A prospective cohort study [34] demonstrated that higher LDL-C levels correlated with increased NOAF risk in patients experiencing acute ST-segment elevation myocardial infarction. Elevated LDL-C contributes to atherosclerosis, compromising cardiac structure and function, which further destabilizes atrial electrical activity. Additionally, elevated LDL-C levels may indirectly influence AF occurrence by promoting inflammation and oxidative stress. Conversely, some studies indicate that high LDL-C levels serve as independent protective factors against NOAF [35, 36], possibly due to the increased risk of AF onset and recurrence associated with low LDL-C levels, which enhance inflammatory responses [37]. This study also identified a higher postoperative ventricular rate as a significant predictor of NOAF. The primary strategies for AF management include anticoagulant therapy, heart rate control, rhythm control, and lifestyle modifications. Among these, ventricular rate control remains the foremost treatment approach and a fundamental goal in AF management. For all AF patients, prioritizing ventricular rate control is essential for symptom improvement [38].

The nomogram model serves as an effective visual prediction tool to guide clinical decision-making [39]. Prior studies demonstrate its efficacy in predicting myocardial infarction risk within the Chinese Han population [40], assessing heart injury risk in patients with ST-segment elevation myocardial infarction [41], and evaluating coronary artery disease risk among elderly patients with AMI [42]. Additionally, there have been reports of studies on predictive models for acute kidney injury and mortality risk after cardiac surgery [43, 44]. However, research on the application of the nomogram model for predicting NOAF risk in AMI patients following percutaneous coronary intervention (PCI) remains limited. This study identified eight independent risk factors from LASSO and multivariate logistic regression analyses to construct a nomogram for predicting NOAF risk in AMI patients following PCI. Evaluation through ROC curve, calibration curve, and DCA demonstrated that the nomogram exhibits strong discrimination ability, robust calibration, and significant clinical applicability.

The clinical value of the nomogram model constructed in this study is substantial. It provides a practical and visual tool for predicting the risk of postoperative NOAF in AMI patients following PCI, incorporating eight independent risk factors identified through LASSO and multivariate logistic regression analyses. These factors include age, LAD, Gensini score, NT-proBNP, ALT, LDL-C, LVESD, and ventricular rate after PCI, which are relatively easy to collect through routine clinical assessments such as admission evaluation forms, laboratory tests, electrocardiograms, echocardiograms, and coronary angiography. By utilizing this nomogram, clinicians can make more informed decisions about patient management, such as identifying high-risk patients who may benefit from more aggressive monitoring or prophylactic treatment strategies. This could potentially lead to improved patient outcomes, reduced hospitalizations, and a better quality of life for patients at risk of NOAF. Furthermore, the nomogram can serve as a valuable research tool for future studies investigating the prevention and treatment of NOAF in AMI patients, providing a standardized method for risk assessment and facilitating comparisons across different patient populations and treatment approaches.

This study presents several limitations. First, the focus on internal testing and the limited sample size restrict the generalizability of the findings. While internal validation employed Bootstrap resampling within the dataset, the lack of verification against independent datasets may hinder the nomogram’s broader applicability. Future investigations should prioritize multi-center collaborations to validate the model’s predictive efficacy and stability. Second, the number of AMI patients included and the factors influencing NOAF occurrence are restricted. The conclusions drawn from this study may be influenced by random variations due to the limited sample size and selected research indicators, potentially overlooking relevant factors. To achieve a more comprehensive understanding of NOAF risk factors, future studies should involve larger sample sizes and consider additional variables. Ultimately, subsequent research must build on these results, further validating conclusions through large-scale, multi-center investigations. Such studies will enhance the reliability of findings and offer valuable guidance for clinical practice.

Authors and Affiliations

1. Department of Cardiology, The First Affiliated Hospital of USTC, Division of Life Science and Medicine, University of Science and Technology of China, No. 1, Swan Lake Road, Hefei, Anhui Province, 230001, China

   Li-Xiang Zhang,Jiao-Yu Cao&Xiao-Juan Zhou

Contributions

L.Z. and J.C. conceived the study, X.Z. designed and supervised the study. L.Z. contributed materials and analysis tools. J.C. analyzed the data. L.Z. drafted the manuscript. X.Z. and J.C. supervised the study and revised the manuscript. All authors reviewed and approved the final version of the manuscript.

Corresponding author

Correspondence to Xiao-Juan Zhou.

Ethics approval and consent to participate

The Medical Ethics Committee at the First Affiliated Hospital of the University of Science and Technology of China approved the study protocol (Approval number: 2024-RE-386), ensuring compliance with the principles outlined in the Declaration of Helsinki. Additionally, the study has been registered in the Chinese Clinical Trial Registry under the registration number ChiCTR2400091026. Given the retrospective design of the study, informed consent from participants was waived.

Competing interests

The authors declare no competing interests.

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions