Validation of CHADS2 and CHA2DS2, and CCI scores for predicting mortality in hemodialysis patients: A retrospective study from a resource-limited setting

Background: Limited tools are available to predict the mortality outcome in chronic kidney disease (CKD) patients. Recently, the Charlson Comorbidity Index (CCI), CHADS2, and CHA2DS2 scores were used to evaluate mortality among these patients. However, the effectiveness of these scoring systems for predicting mortality in hemodialysis (HD) patients remains unclear, and comparative data are limited. This study aimed to assess the utility of the CHADS2, CHA2DS2, and CCI scores in predicting mortality among HD patients in a resource-limited setting. Patients and Methods: This retrospective study, conducted from March 2018 to September 2023, included 447 HD patients from the Nephrology Center of Al-Thora General Hospital in Yemen. The Kaplan–Meier and log-rank tests were used to evaluate and compare survival curves, and the proportional Cox hazard model was used to investigate the factors associated with mortality. CCI, CHADS2, and CHA2DS2 scores were calculated for each patient and categorized into three groups based on their scores: 0–1, 2–3, and >4. The correlation between these scores and mortality was analyzed using receiver operating characteristic (ROC) curves and multivariate analysis. Results: The mean age of the patients was 48.9 ± 16.3 years, with 73 (16.3%) over the age of 65 and 21 (4.7%) over the age of 74. Among the cohort, 258 (57.7%) were male, and 189 (42.3%) were female, with the majority ( n = 393, 87.9%) residing in Ibb city. The mean follow-up duration was 40.2 ± 13.7 months, and 97 patients (22%) had died by the end of the study. The median survival duration was 48 months, with 1-, 3-, and 5-year survival rates of 94%, 82%, and 69%, respectively. The mean scores for CHA2DS2, CHADS2, and CCI were 2.50 ± 1.45, 1.63 ± 1.03, and 3.98 ± 3.73, respectively. On multivariate analysis, residence outside Ibb city (HR: 0.38; 95% CI: 0.19–0.79, p = 0.009), positive HBV infection (HR: 3.46; 95% CI: 2.10–5.71, p < 0.001), history of diabetes mellitus (HR: 12.47; 95% CI: 2.86–54.27, p = 0.001), and history of the rheumatologic disease (HR: 518.05; 95% CI: 117.10–2291.84, p < 0.001) were the predictors of all-cause mortality. The areas under the ROC curve (AUC) for CCI, CHA2DS2, and CHADS2 scores were 0.991 (95% CI: 0.977–0.9), 0.829 (95% CI: 0.791–0.863), and 0.737 (95% CI: 0.694–0.777), respectively, indicating high predictive value for the CCI score (sensitivity: 100%, specificity: 93.4%), moderate predictive value for CHA2DS2 (sensitivity: 82.47%, specificity: 74.57%), and low predictive value for CHADS2 (sensitivity: 51.55%, specificity: 93.71%). Conclusion: Our findings suggest that CCI, CHADS2, and CHA2DS2 scores can help predict mortality in newly admitted HD patients. Furthermore, factors such as residence outside Ibb city, positive HBV infection, history of diabetes mellitus, and history of rheumatologic disease, were the predictors of all-cause mortality in these patients.