Mendelian randomisation (MR) overcomes classical confounding issues that beset observational studies. “Horizontal” but not “vertical” pleiotropy can bias MR study findings but we found no evidence that it was present in our MR analysis of GORD on IPF.