Improving the Reporting of Pharmacogenetic Studies to Facilitate Evidence Synthesis: Anti-Tuberculosis Drug-Related Toxicity as an Example

Abstract

Background In pharmacogenetic studies, researchers explore how genetic variants impact individuals’ responses to drugs. Implementation of pharmacogenetic tests in clinical practice can improve treatment efficacy and reduce toxicity. For health service providers to implement pharmacogenetic testing in clinical practice, the pharmacogenetic association of interest must be supported by strong evidence. Performing meta-analyses of pharmacogenetic studies increases sample size and power, and is therefore an indispensable tool to researchers striving to improve the strength of evidence for pharmacogenetic associations. The aim of this thesis is to identify and resolve challenges that reviewers might encounter when synthesising evidence from primary pharmacogenetic studies. Methods We explored methods of evidence synthesis for pharmacogenetic studies and applied them to undertake a systematic review and meta-analysis of associations between genetic variants and anti-tuberculosis drug-related toxicity. We applied both standard methods of meta-analysis, and more complex methods of meta-analysis that account for correlation between related effect sizes for each genetic variant. Conducting this systematic review and meta-analysis enabled us to identify that key information was often poorly reported in the primary pharmacogenetic studies. In order to improve the reporting of pharmacogenetic studies with a view to facilitating the evidence synthesis process, we used consensus methodology to develop a reporting guideline for pharmacogenetic studies, known as the STROPS (Strengthening The Reporting Of Pharmacogenetic Studies) guideline. Results Our systematic review of the association between genetic variants and anti-tuberculosis drug-related toxicity included 70 studies. Slow acetylators are more likely to experience anti-TB drug-induced hepatotoxicity than intermediate/rapid acetylators. We also observed associations between the CYP2E1 RsaI and GSTM1 null polymorphisms and hepatotoxicity. Key information, such as the ethnicity of included patients, methodological quality, and patient cohort overlap, was poorly reported. We also found that improvements in the reporting of outcome data would give systematic reviewers greater freedom in terms of their analysis approach. As part of the development of the STROPS guideline, 52 individuals from key stakeholder groups participated in two rounds of a Delphi survey. A total of eight individuals participated in a consensus meeting, before the 54-item STROPS guideline was finalised. Conclusions Our systematic review showed that pharmacogenetic testing may be useful in clinical practice in terms of risk stratification for hepatotoxicity during TB treatment. More studies are needed to overcome methodological limitations of the existing studies and to assess the feasibility and cost-effectiveness of a stratified medicine approach. It is currently challenging to synthesise pharmacogenetic evidence, due to poor reporting of primary studies. We encourage authors to adhere to the STROPS guideline when publishing pharmacogenetic studies. The STROPS guideline will not only improve the transparency of reporting of pharmacogenetic studies, but will also facilitate the conduct of high-quality systematic reviews and meta-analyses, and thus improve the power to detect pharmacogenetic associations.