Weight gain is often perceived as inevitable with insulin therapy, particularly as we strive for tight glycaemic control and are using increasingly proactive insulin titration regimes. The United Kingdom Prospective Diabetes Study documented that weight gain occurs most rapidly soon after insulin therapy is first initiated. The timing of this side effect is particularly undesirable, as weight gain may interfere with patients' adjustment to insulin therapy and may undermine appropriate diabetes self-management behaviours. Until recently, many patients had little alternative other than to accept unwanted weight gain if they were to achieve sufficient glycaemic control to reduce risk of chronic complications of diabetes. Insulin detemir is a novel basal insulin analogue that has consistently been shown in randomized, controlled trials to have a weight-sparing effect (i.e. weight loss or reduced weight gain compared with other insulins) in both type 1 and type 2 diabetes. Indeed, unlike neutral protamine Hagedorn (NPH) insulin, the weight-sparing effect of insulin detemir appears to be most prominent in people who are the most obese. The mechanisms behind the weight-sparing effect of insulin detemir are still being clarified. Reduced risk of hypoglycaemia with insulin detemir, coupled with a more consistent and reliable delivery of the desired dose than is available with traditional basal insulin, such as NPH, has been proposed to minimize defensive snacking by patients, and help to limit weight gain. However, even if this was proven, it would be unlikely to fully explain the weight-sparing effect of insulin detemir. Two additional theories have been put forward. One suggests that due to its novel method of prolonging action via acylation and albumin binding, insulin detemir may differentially influence hepatocytes more than peripheral tissues, thus effectively suppressing hepatic glucose output without promoting lipogenesis in the periphery. The second theory suggests that insulin detemir may be more effective than human insulin in communicating satiety signals within the central nervous system. Further clarification of these hypotheses is required.