Great, now multiply the usage expectations by 1000 and include that people in that pool will have each possible additional liver related factor. (Now "should" being 99% isn't even good enough.)
When we analyze how safe a vaccine is, a few dozen people who might have very specific health problems having potentially fatal reactions is a problem if we can't preidentify that group. When we allow a contagion to spread to endemic and therefore eventually encompass the entire population a treatment needs a closely proportional safety level to a vaccine and will eventually need to be far safer than any long acting vaccines.
If 10% of the population are going to take any treatment much more active than placebo, it has to be safer than most OTC medications for it not to kill a hundred times more people than giving everyone the least safe vaccine.