OBJECTIVES: To investigate the effect on glycaemic control of switching from a premix-based regimen to a glargine-based regimen in 528 patients with type 1 (n=183) or type 2 (n=345) diabetes, using unselected primary care data from a UK database (The Health Improvement Network). METHODS: Patients were required to have at least 12 months of available data, before and after switching. The principle analysis was the change in HbA1c; secondary analyses included change in weight, bolus usage and insulin dose. Hypoglycaemia could not be assessed due to lack of consistency in the recording of episodes. Multivariate analyses were used to adjust for baseline characteristics and confounding variables. RESULTS: Mean HbA1c levels at baseline were similar in the T1 and T2 cohorts (9.4% and 9.3% respectively). After adjustment both cohorts showed significant reduction in mean HbA1c 12 months after the switch: by -0.67% (p<0.001) in the T1 cohort and by ‑0.53% (p<0.001) in the T2 cohort. Patients with a baseline HbA1c ≥10% had the greatest reduction in HbA1c, by -1.7% (p<0.001) and -1.2% (p<0.001), respectively. Proportion of patients receiving a co-bolus prescription went from 24.6% on premix to 95.1% on glargine in the T1 cohort, p<0.001 and from 16.2% to 73.9% in the T2 cohort, p<0.001. There was no significant change in weight in the T2 cohort with a moderate increase in T1 patients (0.3 kg and 3.7 kg respectively, p>0.05). Total insulin use (±SD) increased in T2 patients (0.67 ± 1.35 IU/Kg to 0.88 ± 1.33 IU/Kg, p<0.001) with no significant increase in T1 diabetes patients.  CONCLUSIONS: In everyday practice, patients with T1 or T2 diabetes inadequately controlled by premix insulins experienced significant improvement in glycaemic control over 12 months after switching to a glargine-based regimen. These findings support the use of glargine plus boluses in patients poorly controlled on premix.