Introduction: Advantages of the meta-analysis of individual patient data are the possibility to: verify the randomization and the comparability of follow-up in different treatment groups, carry out time to event analyses and subgroup analyses, and use new updated information. Analysis by intention to treat (ITT) using all randomized patients regardless of their eligibility is the gold standard in meta-analysis of individual patient data. This approach can however lead to, at first sight, absurd situations.

Results: In superficial bladder cancer trials, the histological results confirming the patients eligibility may only be known to the investigator after randomization. Due to this, ineligible patients may be entered in the trials and the clinician might administer a less effective or even harmful treatment during the period that the histological results are not available. In a trial testing local treatments of the bladder, ITT analyses might lead to including ineligible patients that have to be cystectomized immediately after the histological results are available. Although ITT analyses may seem absurd in this case, it provides a more valid assessment of treatment efficacy as related to actual clinical practice. Summary data coming from publications that do not analyze by intention to treat would not allow such analyses at the meta-analysis level. This issue, together with other advantages of individual patient data meta-analyses, will be illustrated based on superficial bladder cancer data.

Discussion: Individual patient data meta-analyses allow a more flexible approach to the analysis of the data. For example with regards to ITT, the advantage is that both analyses with and without ineligible patients can be done to get more insight into biases that might occur due to the exclusion of ineligible patients.