Systematic Review And Meta Analysis Of Stroke And Bleeding Risk Scores For Patients With Nonvalvular Atrial Fibrillation

Background:For the evaluation of efficacy and safety of anticoagulant drug of atrial fibrillation（AF）patients,several embolism and bleeding risk score systems have beenestablished,includingCHADS2,CHA2DS2-VASc,HAS-BLED,HEMORR₂HAGES,ATRIA and ORBIT.However,there are still controversies about the predictive value and accuracy of correlation scores for embolism or major bleeding events.Objective:We performed three systematic review and meta-analysis respectively to compare CHADS₂ and CHA₂DS₂-VASc scores in predicting the risk of stroke and thromboembolism in patients with nonvalvular atrial fibrillation（NVAF）,the diagnostic accuracy between HAS-BLED score and HEMORR₂HAGES,ATRIA,CHADS₂,CHA₂DS₂-VASc score,and the predictive ability by using HAS-BLED and ORBIT scores,and in order to provide a basis for the selection of criteria for the risk of embolism and bleeding in anticoagulated patients with atrial fibrillation.Methods:We systematically searched the Cochrane Library,Elsevier,PubMed,Embase databases and Chinese databases for relevant studies.Data extraction and analysis are based on predefined clinical outcomes.Statistical analysis was performed with Review Manager software.We chose the C-statistic to reflect the diagnostic value.There are other statistics,such as Z-statistic,net reclassification improvement（NRI）,integrated discrimination improvement（IDI）.Results:1.Comparison of CHADS₂ and CHA₂DS₂-VASc scores in predicting the risk of stroke and thromboembolism in patients with NVAF.A total of eleven studies were included.The result showed that when analyzed as a continuous variable,the C-statistics for the CHADS₂ and CHA₂DS₂-VASc scores were 0.660（0.650-0.717）and 0.697（0.637-0.724）.The pooled C-statistics for the CHADS₂ and CHA₂DS₂-VASc scores were 0.66（0.66-0.67）and 0.67（0.66-0.68）,respectively.When analyzed as a categorical variable,the C-statistics for the CHADS₂ and CHA₂DS₂-VASc scores were 0.630（0.586-0.722）and 0.606（0.521-0.850）.The pooled C-statistics for the CHADS₂ and CHA₂DS₂-VASc scores were0.65（0.62-0.67）and 0.63（0.59-0.68）,respectively.Because of the high heterogeneity across studies,the results should be interpreted cautiously.The average ratio of adverse event in the low-risk group of CHA₂DS₂-VASc was less than that of CHADS₂（0.54%vs 1.40%,P<0.05）.The average proportion of the high-risk group of CHA₂DS₂-VASc was greater than that of CHADS₂（81.0%vs 46.0%,P<0.05）.2.Comparison of HAS-BLED and HEMORR₂HAGES,ATRIA,CHADS_2,CHA₂DS₂-VASc sorces in predicting the risk of major bleeding in anticoagulated patients with NVAF.A total of eleven studies were identified.The results showed that the C-statistics for HAS-BLED,HEMORR₂HAGES and ATRIA scores were 0.66（0.60-0.69）,0.63（0.60-0.67）and0.61（0.59-0.69）.TheC-statisticsforCHADS₂ and CHA₂DS₂-VASc scores were 0.53（0.51-0.59）and 0.56（0.53-0.58）.The pooled C-statistic were 0.65（0.61-0.69）,0.63（0.61-0.66）,0.63（0.56-0.72）,0.55（0.49-0.61）和0.56（0.53-0.59）,respectively.HAS-BLED vs HEMORR₂HAGES（Z=0.88,P>0.05）,HAS-BLED vs ATRIA（Z=0.49,P>0.05）,HAS-BLED vs CHADS₂（Z=2.72,P<0.01）,HAS-BLED vs CHA₂DS₂-VASc（Z=3.53,P<0.01）.The analysis of NRI and IDI demonstrates that the HAS-BLED score had better prediction accuracy for major bleeding risk than HEMORR₂HAGES or ATRIA bleeding scores,and CHADS₂ or CHA2DS2-VASc embolism scores.3.Comparison of ORBIT and HAS-BLED scores in predicting the risk of major bleeding in anticoagulated patients with NVAF.A total of seven studies were included.The results showed that the C-statistic of continuous variable of major bleeding was 0.65（0.60-0.69）for ORBIT score and0.63（0.60-0.66）for HAS-BLED score.Compared with HAS-BLED,more anticoagulated patients with atrial fibrillation（88.45%vs 32.59%）and major bleeding events（75.57%vs 25.57%）in the ORBIT score were categorized as low risk.The ORBIT score had 1.21,1.73 and 1.44-fold risk of major bleeding in the low,intermediate and high risk,respectively.The calibration analysis demonstrated that the ORBIT score under-predicted the risk of major bleeding.The odds ratio（OR）were0.64（0.37–1.10）,0.63（0.38–1.05）and 0.64（0.38–1.06）in the low,intermediate and high risk,respectively.Conclusion:1.The CHADS₂ and CHA₂DS₂-VASc scores have similar predictive value for stroke and TE risk in NVAF patients.2.The CHA₂DS₂-VASc score identifies truly"low risk"NVAF patients and classifies more patients into high risk,so as to direct anticoagulation more accurately.3.The accuracy of HAS-BLED score in predicting major bleeding in anticoagulated patients with NVAF was better than HEMORR₂HAGES,ATRIA bleeding scores,and CHADS₂,CHA₂DS₂-VASc embolism scores.4.The HAS-BLED score is currently the optimal choice for assess the risk of major bleeding in anticoagulated patients with NVAF in clinical practice.