If you are going to downvote this comment, provide evidence. It's an empirical claim and the direction of causality OP is proposing is totally plausible. Did the authors of the original study control for this?
I mean the quote is literally "relationships at 50" and "outcome at 80." It's not "relationships at 80." Causality is inherent in the 30 year gap between factor and effect.
Only a partial control. Some bad effect @50 can cause both the divorce and the problems at @80
"Good dental hygiene at 50 correlates with cardiac health at 80, therefore brushing your teeth is good for your heart."