In their recent meta-analysis, Wang and colleagues concluded that high-dose intravenous proton pump inhibitor (PPI) regimens were not superior to non–high-dose regimens for patients with recent ulcer bleeding.1 However, we consider these conclusions to be premature, possibly flawed, and potentially misleading.
Although Wang et al1 stated that they conducted their meta-analysis according to the recommendations of the Cochrane Collaboration, they graded the quality of trials according to the Jadad classification. This is contrary to the recommendations of the Cochrane Collaboration, which discourages the use of quality scales. Rather, it proposes the assessment of risk of bias separately in a number of domains, with concealment of allocation being the most important criterion in determining overall trial quality.2 The Jadad scale does not take concealment of allocation into account. Moreover, had the totality of evidence been assessed using the approach recommended by the Grading of Recommendations Assessment, Development, and Evaluation (GRADE) working group,3 it would have been graded as “low-quality evidence” rather than as “7 high-quality randomized studies.” Their literature search may have been incomplete, since recent major conference proceedings were not searched. Therefore, the results of this meta-analysis should be interpreted with great caution.