After controlling for covariates in the multivariate model, only the group reporting coffee consumption between 0 and 4 cups per day remained statistically significant, while the group reporting more than four cups per day was of borderline significance, despite the large sample size.
1
—
—