Reverse causation and causality can make it look like LDL cholesterol caused increased death rates. Here is a brilliant study that corrected for all covariates and cofounders and found that shows that when you adjust correctly, low LDL-C is actually protective.
https://dralo.net/links