The quality of reporting of systematic reviews or meta-analyses on peripartum antiretroviral therapies for preventing the risk of mother-to-child transmission of HIV

Article type
Authors
Wang L, Li Y, Peng L, Wen J, Sun X
Abstract
Objectives: To assess the efficacy of different regimens of peripartum antiretroviral therapies (ART) for preventing the risk of mother-to-child transmission (MTCT), and identify the risk factors for MTCT of HIV.
Methods: We searched the databases of PubMed, EMBASE, CINAHL, AID Search, AIDS LINE, AIDSTRIALS, AIDSDRUGS, AIDS info and the Chinese Databases (CBM, CNKI, VIP) from the beginning of the databases to April 30, 2007. The Cochrane Library (2007 issue 1) was also searched. The systematic reviews (SRs) or meta-analysis of randomized controlled trials (RCTs) with pooled results to assess the efficacy of ART for preventing MTCT were included. Quality assessment of the included studies was undertaken using the two methods: the quality of reporting of meta-analysis (QUOROM) statement and the Oxman and Guyatt index.
Results: Six systematic reviews were eligible, including 2 Cochrane reviews (Brocklehurst 2002 & 2005) and 4 meta-analyses using individual patient data (IPD Meta-analysis). The quality of the included studies were relatively high with Oxman and Guyatt scores varied from 5 to 7 and QUOROM score from 12-16. The quality of 2 Cochrane reviews and 2 IPD meta-analyses (Leroy 2005 and BHITSG 2004) is higher than another 2 meta-analyses (Leroy 2002 & 2003). Using the Oxman & Guyatt index, Cochrane reviews were more likely than paper-based systematic reviews to report within the review the search methods used to find the evidence (Item 1), eligibility criteria for selecting studies (Item 3), criteria used for assessing the validity (Item 5), and assessed appropriately (Item 6). But all the included studies were weaker in inadequate searching strategy for the evidence (Item 2) and selection bias could not be avoided (Item 4). Using the QUOROM checklist, the proportion of Cochrane and paper-based systematic reviews that met each of the 18 items ranged from 77.8% to 83.3% for Cochrane reviews and 66.7% to 88.9% for paper-based systematic reviews. The Cochrane reviews were significantly more likely than the published systematic reviews to adequately report the data sources, review methods in abstract and searching, selecting and assessing the quality in methods. Cochrane reviews were weaker in reporting quantitative data synthesis in methods and trial flow in results. All of them were less likely to report as a meta-analysis or systematic review of RCTs in title and the extensive literature search. All included studies failed to explicit cite the QUOROM statement. There is some limitation of the QUOROM in reporting IPD meta-analysis, especially the central collection, validation and re-analysis of 'raw' data obtained from those responsible for the original trials.
Conclusions: The methodological limitations existed in all included systematic reviews or meta-analyses, especially for searching bias, and validity assessment. The Cochrane reviewers handbook should be updated following the QUOROM statement. A QUOROM statement for IPD meta-analysis should be made or revised.