Those 3,000 cases were actively monitored for cardiac issues, while the VA study is purely passive and observational (which the article itself raises as a cavea

Those 3,000 cases were actively monitored for cardiac issues, while the VA study is purely passive and observational (which the article itself raises as a caveat). It's not just a matter of comparing raw sample sizes.