February 17, 2017
5 min read

Should metformin be the first-line therapy choice in type 2 diabetes treatment?

You've successfully added to your alerts. You will receive an email when new content is published.

Click Here to Manage Email Alerts

We were unable to process your request. Please try again later. If you continue to have this issue please contact customerservice@slackinc.com.

Click here to read the Cover Story, "Beyond diabetes, metformin may prove to be a ‘wonder drug’."


Should metformin be the first-line therapy choice in type 2 diabetes treatment?


Metformin is recommended as the first-line drug of choice for several reasons, including its efficacy, tolerability, safety and cost-effectiveness.

Richard E. Pratley

Almost all current treatment guidelines now recommend initiating therapy for type 2 diabetes with metformin, followed by intensification with other agents in a stepwise fashion as necessary to achieve glycemic targets.

Metformin has been extensively studied as monotherapy in patients with type 2 diabetes. The ADOPT study compared initial monotherapy with metformin, glyburide and rosiglitazone (Avandia, GlaxoSmithKline). At 1 year, HbA1c reductions were equivalent in all treatment arms. By 5 years, HbA1c was lowest in patients treated with rosiglitazone, intermediate in those treated with metformin and highest in those treated with glyburide. Notably, the difference in mean HbA1c between patients assigned metformin and rosiglitazone was small, only about 0.1% to 0.2%. Moreover, metformin-treated patients lost on average 2.9 kg of body weight, whereas those randomly assigned rosiglitazone gained 4.8 kg, a difference of 7.7 kg, or nearly 17 lb.

Metformin has also been compared as monotherapy with exenatide once weekly (Bydureon, AstraZeneca), pioglitazone (Actos, Takeda) and sitagliptin (Januvia, Merck) in the DURATION-4 study. From a baseline of 8.4% to 8.6%, HbA1c was reduced by 1.53% with exenatide, 1.48% with metformin, 1.63% with pioglitazone and 1.15% with sitagliptin at 26 weeks. In this study, exenatide and metformin were both associated with a 2-kg weight loss vs. a loss of 0.8 kg with sitagliptin and a gain of 1.6 kg with pioglitazone. Thus, as monotherapy, metformin has equivalent or better efficacy compared with thiazolidinediones, a sulfonylurea, a long-acting GLP-1 receptor agonist and a DPP-IV inhibitor in the short term (26-52 weeks) and is comparable to a TZD and superior to a sulfonylurea in the long term (5 years).

In terms of safety and tolerability, rates of hypoglycemia — an important complication of some treatments — were low with metformin and comparable to those seen with the TZDs, exenatide and sitagliptin; hypoglycemia rates were high, as expected, with glyburide in the above studies. Although nausea (almost always mild) was more common with metformin than with the TZDs, glyburide or sitagliptin, it was less common than with exenatide. Additionally, although there are no robust cardiovascular outcomes studies with metformin, there are numerous prospective studies, meta-analyses and population studies indicating that metformin is safe and may be associated with a decreased risk for CV events. A recent review of several observational studies also concluded that metformin was safe and appeared to be associated with decreased CV and all-cause mortality in patients with chronic kidney disease, chronic heart failure and chronic liver disease. Finally, as metformin is available as a generic medication, costs are low to the health care system and patients, who can often receive the medication free.

Thus, metformin is effective and well-tolerated, leads to weight loss, is associated with low rates of hypoglycemia, decreases CV events and mortality, and is cheap. What’s not to love?


For a subset of patients with diabetes, other drugs should be considered.

Yehuda Handelsman

Generally, metformin should be considered first-line therapy, but there are many patients who have contraindications for metformin use. The American Association of Clinical Endocrinologists treatment algorithm recognizes this and suggests that, in these patients, other agents should be considered: GLP-1 receptor agonists, SGLT2 inhibitors and DPP-IV inhibitors.

Unlike 8 or 9 years ago, metformin is inexpensive today, and we have many years of experience with it. However, many patients, such as those with severe kidney disease, are contraindicated from taking it. Also, up to 20% of patients may experience side effects with metformin, such as nausea and diarrhea, that limit its use.

In such cases, the AACE/ACE 2017 treatment algorithm recommends using a GLP-1 receptor agonist or SGLT2 inhibitor as a first option, followed by DPP-IV inhibitors and other agents based on patient characteristics, comorbidities and baseline HbA1c.

In Japan, DPP-IV inhibitors are the recommended first-line therapy for diabetes because of a perceived better efficacy and better side-effect profile vs. metformin in that patient population. Therefore, one needs to consider whether a DPP-IV inhibitor should be first-line therapy in patients of Japanese ethnicity in the United States. The AACE recommendations focus on safety, specifically, preventing hypoglycemia and weight gain, hence sulfonylureas are not recommended early in treatment. In addition, these therapies that can offer multiple benefits beyond glucose control. When we treat diabetes, we like to manage beyond glucose control and reduce CV risks by encouraging weight loss and blood pressure control. When we prescribe a GLP-1 receptor agonist or SGLT2 inhibitor, patients generally experience sustained weight loss and reduced BP. This is something we ought to take into account. One drug may provide several benefits.

Had the cost of a generic GLP-1 receptor agonist or SGLT2 inhibitor been similar to that of metformin while providing similar glucose control but with benefits in weight loss and BP, I think that would be a strong argument for using an alternative to metformin. Of note, many patients may need more than one drug. In such a case, using combination therapy from the start, with or without metformin per the side effect and morbidity profile, is also indicated.

Addressing CV effects of antihyperglycemic medications, all diabetes medication outcome trials so far with long-acting insulin, DPP-IV inhibitors and GLP-1 receptor agonists have shown these agents to be safe. In the case of empagliflozin (Jardiance, Boehringer Ingelheim) and liraglutide (Victoza, Novo Nordisk), these two drugs had significant effects on reducing CV mortality. In fact, empagliflozin now has a CV death reduction indication from the FDA for people who have experienced a previous CV event. In the clinical trials, empagliflozin was given to people in addition to other background therapies. However, one could make the case that for the patient whose diabetes was managed with only diet and exercise and was not at HbA1c goal, and who experienced a CV event, the first drug of choice should be empagliflozin and not metformin. In other words, although not exactly used as in the trial, it is in accordance with the FDA indication. In anticipation of the results of the canagliflozin (Invokana, Janssen) and dapagliflozin (Farxiga, AstraZeneca) CV outcome trials — CANVAS and DECLARE — and the introduction of other GLP-1 receptor agonists with superior CV outcomes (eg, semaglutide), these drugs may become the first option for all patients who have had CV events. Should CANVAS and DECLARE also show primary CV prevention, we will need to reassess our approach to first-line therapy.