OBJECTIVE To report changes in liver function tests observed with canagliflozin, a sodium glucose co-transporter 2 inhibitor, across phase 3 studies in patients with type 2 diabetes, and to examine the relationship between changes in liver function tests and the weight loss and glycaemic improvements observed with canagliflozin. METHODS Data were pooled from four 26-week, placebo-controlled studies of canagliflozin 100 and 300mg (n=2313) and two 52-week, active-controlled studies of canagliflozin 300mg versus sitagliptin 100mg (n=1488). Analysis of covariance was performed to determine the contribution of changes in body weight and HbA1c to the changes in liver function tests. RESULTS Reductions in alanine aminotransferase, aspartate aminotransferase, alkaline phosphatase and gamma-glutamyl transferase, and increases in bilirubin were seen with canagliflozin 100 and 300mg versus placebo (nominal P<0.001 for alanine aminotransferase, aspartate aminotransferase and gamma-glutamyl transferase [both doses]; P<0.001 for alkaline phosphatase and P=0.015 for bilirubin [canagliflozin 300mg only]) at week 26 and with canagliflozin 300mg versus sitagliptin 100mg (nominal P<0.001 for alanine aminotransferase, aspartate aminotransferase, gamma-glutamyl transferase and bilirubin, and P<0.01 for alkaline phosphatase) at week 52. Few patients met predefined limits of change criteria for liver function tests, and none met Hy's law criteria. In both populations, alanine aminotransferase, aspartate aminotransferase and gamma-glutamyl transferase reductions were fully explained by HbA1c and body weight reductions. CONCLUSIONS Canagliflozin provided improvements in liver function tests versus either placebo or sitagliptin treatments that were fully explained by the combined effects of HbA1c and body weight reductions with canagliflozin.