Interesting read. I must say I was not very surprised by the finding that mothers had a greater risk of cardiac arrests with 4-5 sons than with 4-5 girls.
Just looking at my own experience being two boys growing up, and with a mum that would get nervous by anything(!), I feel like I can relate to it. I could easily call my dad and say "hi dad, you know what? Tomorrow I will be jumping out of an airplane in a parachute!" To which his first reply would be "That's awesome! But don't tell your mother until after you've done it, or she will freak out!" That difference in mentality seems likely to be the explanation. If it is true, as some evidence suggests, that meditation can help reduce the risk of coronary heart disease, then being more calm about what your kids are up to could have a similar effect.
It's interesting though that it is only with 4 sons (or 3 in some cases according to Figure 1) that the mother's risk is greater than with 0 children. I guess if we were to then closely examine the boys of the parents in the sample. We could try and split them into two groups: Troublemakers and non-troublemakers, where troublemakers are defined based on measures such as # of fights he got into at school. Number of times he got into trouble with the police. Etc. Then we could get a measure of the percentage of boys that were troublemakers by that definition. As the mother's risk goes up with the number of boys, perhaps it only reflects the fact that she is then statistically more likely to have at least 1 son that is a troublemaker. So that it is not a biological effect, nor boys in general, but whether or not she has a son that is a troublemaker that makes her stay awake at night, or spend time worrying about him. It would be interesting to do such a split of the boys, and then run the analysis again where you control for whether there is a troublemaker among the children, to both observe it's effect on the parents health, but also if that then eliminates the negative effect of boys in general.
What do you think?
I think that you are right and that would be interesting. Any sort of added experimentation that can help to eliminate more of a psychological, or biological component can better hone in on what results specifically in this effect in women. However that's a lot of extra information and very difficult to obtain for a study like this where the sample size was so big like this one. They would need to cross reference school records or something for the children of the participants.