Thank you for the comment. Adding +0 to have the regression through the origin works for the deming package but the deming package relies on the maximum likelihood estimation and not on least squares.