Efficacy of Chinese herbal drugs for angina pectoris: frequentist network meta-analyses

Abstract

BackgroundMany randomised controlled trials have compared Chinese herbal drugs with conventional chemical drugs for treatment of angina pectoris, but these studies have not been evaluated by network meta-analysis. We did a frequentist network meta-analysis to assess whether Chinese herbal drugs are efficacious in treating angina pectoris. MethodsThis study was designed and reported in accordance with the PRISMA guideline for network meta-analyses (PRISMA-NMA). We searched for randomised controlled trials of monotherapy with Chinese herbal drugs versus conventional chemical drugs for angina pectoris in major bibliographic databases, including PubMed, China National Knowledge Infrastructure, WanFang Data, MEDLINE, Cochrane Library, and Science Direct. The search was last updated on Sept 4, 2013. Outcomes were symptomatic and electrocardiography (ECG) improvement. We evaluated the quality of randomised controlled trials with the Cochrane Collaboration's tool for assessing risk of bias. We used odds ratios (ORs) and 95% CIs to measure effect sizes. Two independent network meta-analyses were conducted based on a combined pairwise meta-analysis model and a linear mixed-effects (LME) model. The strength of evidence was assessed with the GRADE approach. FindingsThe database search found 7456 records, of which 7225 were excluded; 231 records of randomised controlled trial with 25 180 patients were included according to the eligibility criteria. The overall quality of the included studies was deemed moderate. Overall ORs comparing Chinese herbal drugs with conventional chemical drugs were 2·68 (95% CI 2·31–3·10) for symptomatic improvement and 2·06 (1·83–2·31) for ECG improvement with a combined pairwise network meta-analysis. Relative to isosorbide dinitrate, LME-based network meta-analysis found that naoxintong capsule (OR 2·18 [95% CI 1·58–3·00] for symptoms; 1·57 [1·22–2·01] for ECG), baoxin pill (2·17 [1·65–2·86] for symptoms; 1·62 [1·32–1·97] for ECG), and tongxinluo capsule (2·12 [1·60–2·80] for symptoms; 1·57 [1·28–1·92] for ECG) were more efficacious. Relative to nifedipine in improving symptoms, LME-based network meta-analysis found that naoxintong capsule (OR 1·49 [95% CI 1·01–2·18]), baoxin pill (1·48 [1·05–2·10]), and tongxinluo capsule (OR 1·45 [1·02–2·08]) were more efficacious. However, relative to nifedipine in improving ECG results, naoxintong capsule (OR 1·21, 95% CI 0·89–1·65), baoxin pill (1·25, 0·94–1·67), and tongxinluo capsule (1·22, 0·90–1·63) were not more efficacious. The strength of evidence according to the GRADE assessment was low to moderate. InterpretationThese frequentist network meta-analyses of randomised controlled trials indicated that specific Chinese herbal drugs are more efficacious than some conventional chemical drugs (ie, nifedipine and isosorbide dinitrate) in treating angina pectoris. Further larger-scale network meta-analyses of registered randomised controlled trials, including all antianginal drugs and outcome measures, are warranted. FundingUniversity of Macau (MYRG190[Y1-L3]-ICMS11-LSW and MYRG2014-00117-ICMS-QRCM).