There is good evidence supporting more extensive use of metformin in type 2 diabetes, in reducing morbidity and mortality. The evidence for a real problem from metformin-induced lactic acidosis is weak, and the risks of alternative agents are often overlooked. We have examined the available data regarding metformin that might cause concern in patients with kidney disease, and find it to be extremely limited. There is no good data on which to offer guidance, but it seems likely that metformin can be used in patients with GFR 60-90 ml/min but at reduced dose at lower levels of GFR, and can probably be safely used at GFRs from 30-60 ml/min but with the same caution as with any renally excreted drug. The risks (often overlooked) and benefits of alternative hypoglycaemic agents should be considered carefully. The overall evidence that metformin causes major harm is poor.