Intention to treat (ITT) is a gold standard strategy to analyze the results of randomized controlled trials (RCTs). ITT analysis has been considered a methodological indicator of the quality of clinical trials. The extent to which the use of ITT is related to the treatment effects observed in RCTs has not been rigorously explored. Therefore, the main objective of this study was to determine the association between biases related to attrition and missing data and the use of intention to treat principle, and changes in effect size estimates in RCTs.

This was a meta-epidemiological study. A random sample of RCTs included in meta-analyses was identified. Data extraction including assessments of the use of intention to treat principle, missing data and drop-outs was conducted independently by two reviewers. To determine the association between biases related to attrition, missing data, and the use of intention to treat and effect sizes, a two-level analysis was conducted using a meta-meta-analytic approach.

Three-hundred and ninety-three trials included in 43 meta-analyses, analyzing 44,622 patients contributed to this study. From these, 134 trials (34.1%) used ITT and 218 (55.5%) did not use ITT. Trials which did not use the ITT principle, or which were assessed as having an inappropriate control of incomplete outcome data (based on the Cochrane risk of bias tool) tended to underestimate the treatment effect when compared with trials with adequate use of ITT (ES= -0.13; 95%CI -0.26, -0.01) or trials which were assessed as having an appropriate control of incomplete outcome (ES= -0.18; 95%CI -0.29, -0.08).

Our results suggest that when evaluating risk of bias of primary RCTs, systematic reviewers should pay attention to these biases since they could underestimate treatment effects. Systematic reviewers should perform sensitivity analysis including trials with low risk of bias in these domains.