A Nomogram Predicting Atrial Fibrillation in Patients with Dilated Cardiomyopathy
Dilated cardiomyopathy (DCM) is a primary myocardial disease characterized by left ventricular dilation and systolic dysfunction, often resulting from genetic and environmental factors. The clinical manifestations of DCM include progressive heart failure, arrhythmias, thromboembolism, and even sudden death. The morbidity and mortality associated with DCM have been steadily increasing, making it one of the leading causes of death in cardiomyopathy. Among the various arrhythmias associated with DCM, atrial fibrillation (AF) is particularly significant due to its potential to increase the risk of vascular embolism. Studies have shown that patients with AF have a five-fold increased incidence of stroke and a two-fold increased mortality rate. Approximately 15% to 20% of ischemic strokes are attributed to AF. The hemodynamic changes and increased risk of thromboembolic events caused by AF in patients with DCM significantly elevate their disability and mortality rates. Therefore, early prediction and intervention of AF in patients with DCM are crucial for improving prognosis and quality of life.
The development of an accurate risk classification model for AF in patients with DCM is essential for guiding clinical interventions. Nomograms, which are based on multivariate logistic regression analysis, integrate multiple prediction indices and visualize the correlation among various variables in the form of a graph. This study aimed to develop a nomogram to identify patients with DCM who are likely to develop AF.
The study collected data from consecutive DCM patients admitted to the First Affiliated Hospital of Nanjing Medical University from September 2009 to November 2015. The inclusion criteria were as follows: (1) left ventricular end-diastolic dimension (LVEDd) greater than 5.5 cm in males and greater than 5.0 cm in females, or LVEDd greater than 117% of the predicted value corrected for age and body surface area; (2) left ventricular fractional shortening less than 25% or left ventricular ejection fraction (LVEF) less than 45%; (3) no history of AF in previous medical records or electrocardiogram (ECG) data. Patients with hypertensive heart disease, valvular heart disease, congenital heart disease, ischemic heart disease, severe hepatic and renal insufficiency, or those under 18 years of age and alcoholics were excluded. Basic clinical characteristics were collected from electronic medical records, and patients were followed up every three months. The primary endpoint was the occurrence of AF, as confirmed by ECG or 24-hour dynamic ECG.
Data were presented as mean ± standard deviation (SD) or median (interquartile range). Categorical variables were presented as numbers with percentages. Multivariable logistic regression analysis was performed to predict the probability of AF using a forward stepwise method that included all variables with a probability (P) value less than 0.20 in the univariable analysis. Variables with P values less than 0.05 in the multivariable logistic regression were entered into the prediction model. The main outcome of this study was the risk of AF based on baseline characteristics. The multivariate logistic regression model was used to estimate the odds ratio (OR) and 95% confidence intervals (CIs) of the risk of AF in the model.
The area under the receiver operating characteristic (ROC) curve (AUC) was calculated to evaluate the predictive performance of the model. The calibration of the model was validated using the Hosmer-Lemeshow goodness-of-fit test and calibration plots. The clinical value of the predictive model was tested using decision curve analysis (DCA).
Between September 2009 and December 2015, a total of 243 consecutive patients with DCM were admitted to the First Affiliated Hospital of Nanjing Medical University. After excluding 30 patients with missing data and 16 patients lost to follow-up, 197 patients were eligible for analysis. The demographics and clinical characteristics for the training set (n = 138, mean age = 56.4 ± 14.9 years, 80.4% male) and the test set (n = 59, mean age = 58.1 ± 14.7 years, 74.6% male) are listed in Supplementary Table 1. In the training set, 30 patients (21.7%) developed AF, whereas in the test set, 13 patients (22%) developed AF. The median follow-up was 92.9 (66.5-172.6) months.
Multivariate analyses demonstrated that age (OR: 3.91, 95% CI: 1.71–8.93, P < 0.01), weight (OR: 5.11, 95% CI: 2.23–11.70, P < 0.01), thyroid stimulating hormone (TSH) (OR: 1.55, 95% CI: 1.11–2.16, P = 0.01), d-dimer (D-D) (OR: 1.47, 95% CI: 1.06–2.03, P = 0.02), left atrial diameter (LAD) (OR: 2.34, 95% CI: 1.04–5.25, P = 0.04), and LVEF (OR: 0.33, 95% CI: 0.15–0.75, P < 0.01) were independent risk factors for AF in patients with DCM. The nomogram was developed by assigning a graphic initial score to each of the six independent prognostic factors (age, weight, TSH, D-D, LAD, and LVEF), with a point range from 0 to 100. The scores for all variables were then added to obtain the total score, and a vertical line was drawn from the total points row to indicate the estimated probability of AF being present. It was predicted that a higher total score in the nomogram was associated with a higher likelihood of AF, whereas a lower total score was associated with a lower likelihood of AF.
The AUC-ROC was 0.931 (95% CI: 0.86–0.99) in the training set and 0.90 (95% CI: 0.80–0.95) in the test set. The nomogram model was calibrated using the Hosmer-Lemeshow goodness-of-fit test and a calibration plot. The Hosmer-Lemeshow test revealed high concordance between the predicted and observed probabilities for both the training set (x2 = 7.83, df = 8, P = 0.45) and the test set (x2 = 8.51, df = 8, P = 0.49). The calibration plot also showed good agreement between the predicted and observed outcomes for the training and test sets. DCA was applied to assess the clinical validity of the nomogram, which corroborated good clinical applicability of the nomogram in predicting AF because the ranges of threshold probabilities were wide and practical for the training and test sets.
The main findings of this study were as follows: (1) DCM patients were susceptible to AF (21.8%). (2) Age, weight, TSH, D-D, LAD, and LVEF were independent predictors for AF occurrence in DCM patients. (3) The nomogram was feasible for predicting AF in DCM patients at our hospital and showed good predictive performance.
Patients with DCM are susceptible to AF due to cardiac degenerative changes, such as enlargement of cardiomyocytes, reduced atrioventricular compliance, and degeneration of fibrous tissue. The non-regeneration of myocardial cells often leads to atrial enlargement, which is accompanied by cardiac chamber fibrosis, forming focal or patch-like scars. This results in slowing of local conduction speed, heterogeneity of electrical impulse conduction, and irregular impulse conduction of ectopic lesions with rapid discharge, contributing to the occurrence of AF. AF leads to uncoordinated atrioventricular contraction, reduced effective atrial contraction, decreased ventricular filling volume, and ultimately reduced perfusion of vital organs, leading to multi-organ damage and failure.
The nomogram developed in this study provides an individualized and highly estimated AF risk by combining six independent variables and assigning an appropriate weight to each variable based on its prognostic value. This makes it easy to use and facilitates management-related decision-making for doctors. The prediction model showed that older patients and those with severe or critical DCM had more frequent AF, which was associated with cardiac degenerative changes, such as cardiac enlargement, cardiomyocyte degeneration, and increased fibrous tissue resulting in electrical remodeling. Left atrial enlargement and fibrosis are concomitant in DCM patients, leading to atrial structural remodeling and AF. The lower the LVEF, the more frequent the AF, which may be associated with the occurrence of heart failure in patients with DCM, resulting in reduced effective circulation of blood volume, reduced renal blood flow, and activation of the renin-angiotensin system. Elevated angiotensin II directly causes myocardial cell apoptosis and interstitial fibrosis, further accelerating atrial structural remodeling. Abnormal hemorheology in patients with DCM, caused by the enlargement of the heart and weakened ventricular pulse, makes the blood hypercoagulable and increases the D-dimer level, which is associated with an increased risk of AF. Patients with DCM typically have concomitant hyperthyroidism and heart disease, wherein TSH is decreased and thyroid hormone is increased, which can directly affect the cardiac muscle, decrease the sinoatrial node action potential time, increase atrial muscle excitability, and decrease the refractory period, leading to AF. However, this study found that higher TSH levels in patients with DCM were associated with an increased risk of AF, which may be due to the small sample size and requires further investigation.
This study had some limitations. First, it was a single-center retrospective study with a small sample size, which might have limited the statistical power of the results. Second, the model has not been validated in external cohorts. Third, the model cannot distinguish various subtypes of AF, such as persistent and paroxysmal AF. Further studies addressing these limitations are necessary to improve the nomogram and enhance its predictive performance.
In conclusion, the nomogram developed in this study provides a practical and effective tool for predicting the risk of AF in patients with DCM. By integrating easily obtainable clinical and biomarker data, the nomogram offers an individualized risk assessment that can guide clinical decision-making and improve patient outcomes. Future studies with larger sample sizes and external validation are needed to further refine and validate this predictive model.
doi.org/10.1097/CM9.0000000000001926
Was this helpful?
0 / 0