Misconceptions on the use of MR-Egger regression and the evaluation of the InSIDE assumption