This study assessed the incidence of severe hypoglycaemia with two insulin glargine titration algorithms: Algorithm 1 (increments of at least 10%, but not exceeding 4U) versus Algorithm 2 (1-6U increments). In this multicenter (n=409), multinational (n=54), open-label, 24-week randomized trial in 2442 subjects with sub-optimally controlled Type 1 diabetes (T1DM), mean prior insulin therapy duration was 14.6+/-10.3 years. The incidence of severe hypoglycaemia was similar with Algorithms 1 and 2 (16.6events/100 patient-years versus 14.4events/100 patient-years). There were similar rates of both symptomatic and nocturnal hypoglycaemia. HbA(1c) and fasting blood glucose (FBG) decreased significantly (baseline to endpoint; p<0.001), and comparably with Algorithms 1 and 2 (HbA(1c): -0.64% versus -0.72%; FBG: -57mg/dL versus -59mg/dL). Mean basal insulin dose increased with both algorithms (+5.7U versus +5.9U). In a diverse population with longstanding T1DM, transfer from any insulin regimen, including basal-bolus or premixed insulin to an insulin glargine-based regimen resulted in significant improvements in glycaemic control, with low rates of severe hypoglycaemia, irrespective of the titration algorithm used.