In reading these later papers, it occurred to me the paradox might be spurious?
High BMI is generally a risk factor for cardiovascular disease — thus the paradox — but it suggests that if someone has low risk by BMI and still has cardiovascular disease, there might be something more insidious or systemically advanced going on. In this case, it's not that adiposity protects, it's that the presence of disease in the absence of a risk factor suggests that other etiologic factors might be more advanced.
I guess collider bias is probably a different version of this explanation:
https://www.nature.com/articles/s41598-023-43493-0
It's a bit murky though and might reflect imprecision in BMI as an indicator of adiposity:
https://academic.oup.com/eurheartj/article/44/13/1136/707226...
In reading these later papers, it occurred to me the paradox might be spurious?
High BMI is generally a risk factor for cardiovascular disease — thus the paradox — but it suggests that if someone has low risk by BMI and still has cardiovascular disease, there might be something more insidious or systemically advanced going on. In this case, it's not that adiposity protects, it's that the presence of disease in the absence of a risk factor suggests that other etiologic factors might be more advanced.
I guess collider bias is probably a different version of this explanation:
https://journals.lww.com/epidem/fulltext/2013/05000/The__Obe...