However, sensitivity analyses involving truncation of
follow-up at the time of the event showed that drug exposure
of nsNSAIDs did not change after the event (ie, results
obtained were similar to those from the original analysis),
meaning that confounding by contraindication was unlikely
to explain the findings.