Unfortunate as that is it's not at all surprising. Comparing the median with the top 1% would be more interesting. The gap there is still quite significant:
(for 40 year old men, unadjusted by race):
100th inc. prct : ~ 88 years:
75th inc. prct : ~ 84
50th inc. prct : ~ 82.5
25th inc. prct : ~ 79
5th inc. prct : ~ 76
1st inc. prct : ~ 72.5
(had to infer the values visually from charts because I wasn't to find a table including all the groups...)
However (I assume the data is very limited though) there is almost no difference in life expectancy (for men or women) when your household income is above >$200k (back in 2014, so probably quite a bit higher now). So I don't think there are any efficient treatments available only for the ultra-rich, just being rich or upper-middle class should be enough to get access the best(ish) treatment there is.
For the bottom income quartile when comparing local areas: the % of people with not insurance, medicare spending per enrolled person and 30-day hospital mortality rate seem to have the highest correlation with life expectancy. Which all should be trivial to fix for a relatively extremely-rich country like the US...
Looking at the appendices one interesting point I noticed (assuming I understood it correctly) is that people at the 50th percentile are more likely to reach 77 years than those in the top 75th or 100th prcts. But after that point income seems to matter a whole lot more.
Another seemingly very weird correlation (page 43): higher inequality in local area seems to be correlated with lower life expectancy for all income quartiles except the bottom one (so basically poorer people tend to liver longer in high inequality areas even though the difference in years is not very big).