A systematic review of the quality of publications reporting Coronary Artery Bypass Grafting (CABG) trials

PurposeSeveral studies have shown that the quality of reporting randomized controlled trials (RCT) is variable and often poor. Since the results of meta-analyses on the comparison of off-pump (OFCAB) and on-pump (ONCAB) techniques are controversial, we aimed to assess the reporting quality of the RCTs on comparing OFCAB and ONCAB surgery.MethodsRCTs published (2000 to 2004) comparing OFCAB and ONCAB were identified from electronic searches of the Medline, Cochrane, Cinahl, HealthStar, and EMBASE. Report quality was assessed with 35 items from the CONSORT statement and 53 additional indicators relevant to CABG surgery. Some of the indicators were composed of several small parts making the maximum possible total score greater than 88. Two authors independently reviewed to assess the reporting quality of each RCT and the level of agreement was assessed using kappa statistics. Disagreements were resolved upon consensus. Descriptive analyses are expressed as median [M] and (minimum [min]-maximum [max]). Pearson, spearman and partial correlation coefficients were used for data analysis.ResultsForty-three trials were included with a total of 4544 patients. The Kappa was greater than 0.6 for 67 of 104 (64.4%) indicators. The overall report quality varied between 35 (33.6%) and 80 (76.9%) of 104, with a median score of 51. The CONSORT score reporting quality varied between 17 (40.5%) and 37 (88.0%) of 42, with a median score of 25. The quality of reporting was poor and insufficient for methods (M: 17, min-max: 8–32 of 46) and discussion (4:2–7 of 9) sections. The reporting quality of CABG trials had a strong and positive association with CONSORT score, allocation/blinding, results, protocol and statistical methods. The reporting quality had a strong and positive correlation with the size of trial but it was not correlated with the year of publication.ConclusionThe quality of publications reporting for methods and discussion sections was suboptimal. It is critical that in reporting surgical trials, authors follow CONSORT guidelines as well as consider the surgical factors.