On the "lack of evidence in patients" ... does "testing" this in humans mean decades long trials? Given that the effects in their animal model was 25%, it seems like measuring if there was such an increase in humans (and beginning taking it in late middle age) it would be a generation before you had results.
On the "unthinkable to treat every 50-year-old" comment, I think we as a society would need to change as a function of our changed lifespans -- making all the middle-aged people live to 100+ would exacerbate issues around social benefits, housing stock, etc. But if we all had longer, more productive lives (i.e. worked longer, added our skills and expertise to the economy for longer), could that more than offset the costs? If a drug raised your life expectancy by 20 years, and most people worked another 15, that's still pretty good.