Radiation and fat cell distribution in the body.
You've got a lot of options to accomplish your goal. Indeed, as other answers have pointed out the gap you're hoping to explain already exists, though it isn't as wide as in your setting. In that spirit, I'll suggest a potential agent which also has some real-world basis: radiation.
There is some evidence that some forms of radiation are more hazardous to men than to women. The explanation I've seen is that there are some sex-linked differences in how fat is stored in the body, and arrangements more common in women tend to physically shield vital organs. So radiation has the same effects in women as in men, but the tissues damaged by the radiation are less likely to be those comprising vital organs. Men, on the other hand, tend not to have such protection.
You can fiddle with the details to suit your setting better (especially if those are the average life expectancies you want), but the vital elements are: persistent, low-acuity environmental hazard which is invisible to casual detection; lethal effects of that hazard only occurring after a long period of exposure causing damage to accrue until a "noticeable" threshold is reached; and an unrelated physical structure largely restricted to women (for other reasons which exist without the hazard) which happens to also have a protective effect.
This kind of setup also suits the scenario well. Radiation sickness and various cancers can certainly seem supernatural or divine, even to trained physicians, until enough relevant research is undertaken and understood. And the nature of randomness and average effects among groups would make the different outcomes very difficult to understand before discovery of the specific mechanisms underlying those outcomes.
If you'd like to look at a novel in which this idea is explored, and with similar relevance, I can recommend
Seveneves by Neal Stephenson. I recommend it all on its own, but it also has some interesting descriptions on how radiation hazards exist and operate on human bodies, and also gives some decent scale and scope for how much of an "advantage" women as a group might have against this specific hazard