Question special

The majority of the patients belong to the group who woke up with a stroke of unknown onset. This might be ranging from within few minutes into falling asleep to few minutes from waking up. If the real onset was within 1-2 hours from waking up and patients got thrombolytics within the next 1-2 hours they would still fall within the guidelines range of TPA (<4.5 hours) and one may argue that those subgroup of patients may have confounded the results.
Understanding that there is no practical way to tell when the stroke happened during sleep, and that the percentage of patients who would fall into this particular category is unknown, when we were reading the main article before the official release, we were interested in the results of the subgroup analysis of primary outcome between the three groups (waking up with stroke, 4.5-6 and 6-9 hours) to see If it was still significantly better in the confirmed >4.5 hours onset groups or was it just in the waking up from the sleep group.
This data is presented in the supplementary appendix and it appears that there is still a trend for benefit across all three subgroups (less in the 6-9 groups) but statistical significance was not achieved which is due to low power, likely from the lower sample size due to the early termination of the study. I believe this further supports the validity of the trial and I felt it was a point worth mentioning for our audience.