The life expectancy of people with type 1 diabetes (T1D) is reduced by an average of 11 to 13 years.1 Tight glycaemic control using intensive insulin therapy is essential to reduce microvascular complications in T1D.2 However, insulin therapy requires a high degree of support and may cause hypoglycaemia and weight gain.3 Metformin is commonly prescribed as adjuvant therapy to reduce insulin dose requirement and prevent weight gain.4 In recent years, there has been an increased focus on macrovascular complications of T1D, in particular cardiovascular disease.5 The rates of cardiovascular events in patients with T1D are more than double those in the general population and account for around 45% of deaths.6 Metformin has been associated with reduced cardiovascular mortality in patients with type 2 diabetes,7 as well as patients with ST-segment elevation myocardial infarction (STEMI) without diabetes.8 Investigators therefore wondered whether the cardiovascular impact of metformin would be similar in patients with T1D.
The double blind, placebo controlled REMOVAL (REducing with MetfOrmin Vascular Adverse Lesions in type 1 diabetes) study was undertaken at 23 hospital diabetes clinics in five countries (Australia, Canada, Denmark, the Netherlands, and the UK).9 Individuals aged 40 years or over with T1D of at least five years’ duration were eligible if they had at least 3 of 10 specified CV risk factors. A total of 428 adults were randomized to daily metformin 1000 mg twice daily or placebo as adjuncts to insulin. Investigators measured the thickness of the wall of the carotid artery by ultrasonography; this is a strong indicator of heart disease progression. They also measured cholesterol, weight, insulin dose requirement and glucose levels over three years.
Results of the REMOVAL study were presented on June 11 at the American Diabetes Association (ADA) 2017 Scientific Sessions in San Diego, California, and simultaneously published in Lancet Diabetes & Endocrinology.10 The primary end point, significant reduction in progression of mean far-wall carotid artery intima-medial thickness (cIMT) at 3 year, was not met, although metformin reduced maximal far-wall cIMT, which includes plaque. Metformin also reduced glycated haemoglobin (HbA1c) levels over 3 years (-0•13%, 95% CI -0•22 to -0•037; p=0•0060), but this was accounted for by a reduction in HbA1c in the first 3 months that was not sustained.
However, results for the secondary endpoints of the study were more positive. Metformin reduced body weight (-1•17 kg, 95% CI -1•66 to -0•69; p<0•0001) and LDL cholesterol -0•13 mmol/L, (-0•24 to -0•03; p=0•0117), both of which play a role in reducing atherosclerosis. Estimated glomerular filtration rate (eGFR) was sharply increased on starting metformin. This needs further study to determine if it has any clinical significance. Around a quarter (27%) of participants stopped taking metformin because of nausea or abdominal pain.10
According to lead author, John R Petrie, professor of diabetic medicine at the University of Glasgow, Scotland, this study “changes the way we think about metformin”. Since the study showed no evidence of glucose reduction beyond 3 months, Prof. Petrie suggested: “Don't use it for glucose lowering. But consider using it for reducing weight and LDL cholesterol and possibly atherosclerosis prevention." Prof. Petrie also suggested that guidelines should be updated to reflect that metformin does not improve blood glucose regulation in T1D.
Other experts are less positive in their view of these findings. In an accompanying editorial, Eberhard Standl of Munich Diabetes Group at Helmholz Center, Neuherberg, Germany, said, “the primary core question of whether metformin has a practically relevant role in the treatment of (specific) patients with type 1 diabetes has yet to be substantiated.” He did, however, acknowledge the useful findings that metformin did not show an increased risk for hypoglycaemia, but rather suggested the need for regular monitoring for vitamin B12 deficiency.11
Other glucose-lowering agents have shown varied success as adjuncts to insulin in T1D. Dipeptidyl peptidase-4 (DPP4) inhibitors reduce daily insulin dosage significantly while having no impact on hypoglycaemia, but do not reduce HbA1c.12 In two randomised controlled trials, the addition of liraglutide to insulin therapy reduced HbA1c and body weight but increased the risk of hypoglycaemia and hyperglycaemia with ketosis.13,14 The sodium-glucose cotransporter 2 (SGLT-2) inhibitors appear more promising A recent literature review found that all 3 SGLT-2 inhibitors (anagliflozin, empagliflozin, dapagliflozin), when used as adjuncts to insulin, reduced HbA1c, lowered total daily insulin doses, and reduced body weight. The combination of insulin and SGLT-2 inhibitors also resulted in a lower incidence of hypoglycaemia.15 However these were small, short-term studies (study duration varied from 2 to 18 weeks). Until the results of ongoing Phase III trials are available, there is insufficient evidence to recommend their use, and safety concerns persist around ketoacidosis.
While the results of the REMOVAL trial can best be described as mixed, they have provided evidence to recommend the use of metformin as an adjunct to insulin to help control weight and LDL in patients with T1D. Metformin also has the advantages of being inexpensive and tolerated by the majority of patients.4 With three years of follow-up, the REMOVAL trial is already one of the longest studies to date of metformin use for people with T1D. There is now a need for evidence on the long-term effects of metformin on cardiovascular events rather than intermediate markers of cardiovascular health.