The SAE rate increases 2 percentage points on average, as I understand it - not necessarily uniformly across interventions. It could be the case that medicine A has 4% SAE in healthy patients, and 5% in unhealthy* ones, whereas medicine B has 3% SAE in healthy and 6% in unhealthy - and without testing on unhealthy patients, you don't know that medicine B is riskier for those patients than A.

It could be that I'm totally misunderstanding, and that every medicine has the same elevation of risk of SAE for unhealthy patients, but that seems unlikely to me. You do have 'doctor' in your username though, so I'm probably embarrassing myself here.

*apologies for the healthy/unhealthy terminology, I don't know the right lingo to use here.