Meta-meta-analysis: A paradigm in the case of surgical publications

Meta-analyses are considered to provide level I-II evidence. Based on this premise, several statements have been developed to standardize guidelines and optimize results. The purpose of this study was to investigate the quality of the information delivered by meta-analyses.

Meta-analyses published in Annals of Surgery during an 11-year period were reviewed whereas individual publications of each meta-analysis were assessed. An Excel database encompassing 29 parameters was constructed based on the Quality of Reporting of Meta-analyses (QUOROM) statement.

The present study included 31 consecutive meta- analyses. The number of meta-analyses conforming with each of the parameters considered was as follows: information obtained from more than 2 databases 23/31; language of publication exclusively English 25/31; defined population, intervention, and principal outcomes 31/31; study design encompassing review of randomised controlled trials (RCTs) 10/31; quality assessment of contributing publications 10/31; handling of missing data 10/31; assessment of statistical heterogeneity 30/31; subgroup analysis 23/31; assessment of publication bias 26/31; agreement on selection and validity assessment 22/31; simple summary results 28/31; data available to calculate effect size and confidence interval 27/31; key findings summarized 30/31; clinical inferences based on internal and external validity 24/31; description of potential biases in the review process 23/31; future research agenda suggested 18/31.

Evidence derived from meta-analyses must be interpreted with caution. Although QUOROM guidelines were observed, quality assessments showed considerable variability.