borderline significantP = .0503
When adjusting for potential confounders, this was borderline significant ( P = .0503) but significant when adjusting only for likely confounders ( P = .025; Table S2 ).
When adjusting for potential confounders, this was borderline significant ( P = .0503) but significant when adjusting only for likely confounders ( P = .025; Table S2 ).
First, there was “only” borderline significance in one model and a slightly lower P -value in another, very similar model.