On Arsenic, Diabetes, Creatinine, and Multiple Regression Modeling: A Response to the Commentaries on Our Reanalysis