This is actually a topic that has received a lot of investigation. It's true that men who marry differ, before marriage, from men who don't. But it also seems to be true that marriage has a strong causal effect on male behavior.
There are a lot of instances where long married men lose their wives then die very soon afterwards. When there's a correlation of something with longevity and removing the something is soon followed by death, then there's good reason to believe the causation theory.
True for both partners -- usually after one dies the elderly widow/widower tend to have ~5 years or less.
Some of that is just timing, like if the average age of death per the actuary tables is like 79, and your spouse dies at 80, you're statistically going fairly soon as well.
As far as I recall, there were studies showing some causal effect in the behavioral difference of how early things like cancer were diagnosed.
In essence, the average man living with a female partner would get a "nuisance" investigated, diagnosed with cancer, and (sometimes) treated; while the average single man sought treatment only when the condition was truly disturbing, and by that time it's too late to treat the cancer.