BACKGROUND: Patients with type 2 diabetes (T2DM) have an increased mortality rate primarily because of macrovascular disease. Where T2DM patients cannot be managed sufficiently through diet, exercise and peroral antidiabetic drugs, that is when haemoglobin A1c (HbA1c) is above 7.0%, it is yet unknown whether a combination of metformin and insulin analogues is superior to insulin analogues alone. Nor is it known which insulin analogue regimen is the optimal. OBJECTIVE: The primary objective of this trial is to evaluate the effect of an 18-month treatment with metformin vs. placebo in combination with one of three insulin analogue regimens, the primary outcome measure being carotid intima-media thickness (CIMT) in T2DM patients. DESIGN: A randomized, stratified, multicentre trial having a 2 x 3 factorial design. The metformin part is double masked and placebo controlled. The insulin treatment is open. The intervention period is 18 months. PATIENT POPULATION: Nine hundred and fifty patients with T2DM and HbA1c > or = 7.5% on treatment with oral hypoglycaemic agents or on insulin treatment and deemed able, by the investigator, to manage once-daily insulin therapy with a long-acting insulin analogue. RANDOMIZATION: Central randomization stratified for age (above 65 years), previous insulin treatment and treatment centre. INTERVENTIONS: Metformin 1 g x two times daily vs. placebo (approximately 475 patients vs. 475 patients) in combination with insulin detemir before bedtime (approximately 315 patients) or biphasic insulin aspart 30 before dinner with the possibility to increase to two or three injections daily (approximately 315 patients) or insulin aspart before the main meals (three times daily) and insulin detemir before bedtime (approximately 315 patients). Intervention follows a treat-to-target principle in all six arms aiming for an HbA1c < or = 7.0%. OUTCOME MEASURES: Primary outcome measure is the change in CIMT from baseline to 18 months. Secondary outcome measures comprises the composite outcome of death, acute myocardial infarction, stroke or amputation assessed by an adjudication committee blinded to intervention, other cardiovascular clinical outcomes, average postprandial glucose increment from 0 to 18 months, hypoglycaemia and any inadvertent medical episodes. In addition, change in plaque formation in the carotids, HbA1c, cardiovascular biomarkers, body composition, progression of microvascular complications and quality of life will be assessed as tertiary outcome measures. TIME SCHEDULE: Patient enrolment started May 2008. Follow-up is expected to finish in March 2011. CONCLUSION: CIMT is designed to provide evidence as to whether metformin is advantageous even during insulin treatment and to provide evidence regarding which insulin analogue regimen is most advantageous with regard to cardiovascular disease.