Critical quality appraisal of randomized controlled trials with traditional Chinese medicines for the coronavirus disease 2019

The methodological reporting quality in strictly randomized controlled trials for COVID-19 and precise reporting of Chinese herbal medicine formula intervention

Background

Chinese herbal medicine (CHM) formulas played an important role during the pandemic of coronavirus disease 2019 (COVID-19). Many randomized controlled trials (RCTs) on CHM for COVID-19 were quickly published. Concerns have been raised about their quality. In addition, inadequate detailed information on CHM formula intervention may arouse suspicion about their effectiveness. We aim to assess the most recent evidence of the methodological reporting quality of these RCTs with strict randomization, and the precise reporting of the CHM formula intervention.

Methods

RCTs on CHM formulas for COVID-19 were searched from nine databases. The CONSORT 2010, CONSORT-CHM Formulas 2017, and risk of bias were the guidelines used to assess the included RCTs. The checklist of sub-questions based on CONSORT-CHM Formulas 2017 was used to evaluate the precise reporting of CHM formula intervention. A comparison was made between RCTs that enrolled participants during and after the first wave of the pandemic (defined here as December 2019 to March 2020).

Results

The average score for 66 studies evaluated based on three guidelines, the CONSORT 2010, the CONSORT-CHM Formulas 2017, and the checklist of sub-questions based on the CONSORT-CHM Formulas 2017, is 16.4, 15.2, and 17.2, respectively. The reporting rate of sample size calculation, allocation concealment, and blinding is less than 30%. The checklist of sub-questions based on the CONSORT-CHM formulas 2017 can help report and assess CHM formula intervention more precisely. Most studies assessed an "unclear risk of bias" due to insufficient information. RCTs published in English and recruited subjects during the first wave of the pandemic have a higher risk of participant blinding bias than the studies recruited subjects after that (P < 0.05).

Conclusion

The methodological reporting quality in strictly randomized RCTs on CHM formulas for COVID-19 is inadequate-the reporting of sample size calculation, allocation concealment, and blinding need to improve especially. The checklist of sub-questions based on CONSORT-CHM formulas 2017 can help report and assess CHM formula intervention more precisely. The methodological reporting quality of RCTs published in English and enrolled participants during the first wave of the pandemic is worse than the studies that recruited subjects after the first wave of the pandemic.

The efficacy of Chinese herbal drugs for adults with angina pectoris: Bayesian network meta-analysis of 331 RCTs involving 36,467 individuals

ETHNOPHARMACOLOGICAL RELEVANCE

Hundreds of randomized controlled trials (RCT) on Chinese herbal drugs (CHDs) including Shexiang baoxin pill (BXP), compound Danshen dripping pill (DSP), compound Danshen tablet (DST), Suxiao jiuxin pill (JXP), Naoxintong capsule (NXT), Tongxinluo capsule (TXL), and Di'ao xinxuekang capsule (XXK) and conventional chemical drugs, such as isosorbide dinitrate (ISDN), for angina pectoris are available but have not been evaluated by a PRISMA-compliant network meta-analysis (NMA).

AIM OF THE STUDY

This study aimed to compare the efficacy of nine anti-anginal drugs through NMA on RCTs.

METHODS

RCTs of drug treatment for adult patients with angina pectoris for improvements in symptoms and electrocardiography were retrieved. Odds ratios and 95% credible intervals were computed to measure effect sizes. RCT quality was evaluated with the Cochrane risk of bias tool. Evidence synthesis was performed with Bayesian NMA. Essential analyses including subgroup analysis, sensitivity analysis, meta-regression analysis, publication bias analysis, and ranking analysis were conducted to assess the robustness of efficacies. Evidence strength was assessed with the GRADE approach.

RESULTS

A total of 331 RCTs with 36,467 participants were eligible. The overall quality of all included RCTs was low. Overall efficacy estimates from different approaches of evidential synthesis found that BXP, TXL, and DSP were more efficacious than DST and ISDN. Essential analyses indicated consistent efficacy estimates, insignificant publication bias, and corroborative ranking results. The overall GRADE evidence strength was low.

CONCLUSION

This comprehensive Bayesian NMA found BXP, TXL, and DSP to be the top three candidates among the seven tested CHDs for treating adults suffering from angina pectoris. However, the quality and the evidence strength of eligible RCTs were low. Further high-quality RCTs with more outcome measures and their NMAs are warranted.

REGISTRATION

PROSPERO CRD42014007035.