BACKGROUND Short acting insulin analogue use for diabetic patients is still controversial, as reflected in many scientific debates. OBJECTIVE To assess the effects of short acting insulin analogues versus regular human insulin. METHODS The Cochrane Library (Issue 3, 2005), MEDLINE, EMBASE until September 2005. METHODS Randomised controlled trials with an intervention duration of at least 4 weeks. METHODS Trial selection and evaluation of study quality was done independently by two reviewers. RESULTS Altogether 8274 participants took part in 49 randomised controlled studies. Most studies were of poor methodological quality. In patients with type 1 diabetes, the weighted mean difference (WMD) of HbA1c was -0.1% (95% CI: -0.2 to -0.1) in favour of insulin analogue, whereas in patients with type 2 diabetes the WMD was 0.0% (95% CI: -0.1 to 0.0). In subgroup analyses of different types of interventions in type 1 diabetic patients, the WMD in HbA1c was -0.2% (95% CI: -0.3 to -0.1) in favour of insulin analogue in studies using continuous subcutaneous insulin injections (CSII), whereas for conventional intensified insulin therapy (IIT) studies the WMD in HbA1c was -0.1% (95% CI: -0.1 to 0.0). The WMD of the overall mean hypoglycaemic episodes per patient per month was -0.2 (95% CI: -1.1 to 0.7) and -0.2 (95% CI: -0.5 to 0.1) for analogues in comparison to regular insulin in patients with type 1 diabetes and type 2 diabetes, respectively. For studies in type 1 diabetes patients the incidence of severe hypoglycaemia ranged from 0 to 247.3 (median 21.8) episodes per 100 person-years for insulin analogues and from 0 to 544 (median 46.1) for regular insulin, in type 2 the incidence ranged from 0 to 30.3 (median 0.3) episodes per 100 person-years for insulin analogues and from 0 to 50.4 (median 1.4) for regular insulin. No study was designed to investigate possible long term effects (e.g. mortality, diabetic complications), in particular in patients with diabetes related complications. CONCLUSIONS Our analysis suggests only a minor benefit of short acting insulin analogues in the majority of diabetic patients treated with insulin. Until long term efficacy and safety data are available we suggest a cautious response to the vigorous promotion of insulin analogues. For safety purposes, we need a long-term follow-up of large numbers of patients and well designed studies in pregnant women to determine the safety profile for both the mother and the unborn child.