What Comes After Metformin in Type 2 Diabetes?

Gregory A. Nichols, PhD


January 26, 2018

Does Choice of Second-line Therapy Affect Glycemic Control?

Until 2012, the American Diabetes Association (ADA) recommended lifestyle and metformin as first-line therapy for glycemic management in type 2 diabetes, followed by a sulfonylurea (SU) or insulin for additional A1c reduction.[1] The rationale was that SUs were well validated as a glucose-lowering therapy, even though the risk for hypoglycemia was recognized, along with concerns about increased cardiovascular risk,[2] especially in combination with metformin.[3]

Today, with many more second-line options available, the ADA's new guidelines[4] open the door to whichever therapy is best for the individual patient. But what exactly does that mean?

Khunti and colleagues[5] analyzed data from 10,256 patients who initiated a second-line glucose-lowering therapy after treatment with metformin monotherapy between 2011 and 2014 in Germany and the United Kingdom. The main outcome of interest was change in A1c at 6 months.

The researchers assessed the impact of various factors, including demographics, baseline A1c, time since diabetes diagnosis, and different types of second-line therapy, including SU or dipeptidyl peptidase-4 (DPP-4) inhibitor alone, SU or DPP-4 inhibitor with metformin, and insulin with or without other agents.

Most patients added a therapy to metformin, the most common of which was SU (41%) followed by DPP-4 inhibitor (31%). Baseline A1c was 8.7% overall and 9.2% and 8.4% for SU and DPP-4 inhibitor combined with metformin, respectively. Patients initiating insulin experienced the largest absolute 6-month change in A1c (-2.1%), followed by those adding SU to metformin (-1.7%).

However, after adjusting for baseline A1c, all therapies provided a drop in A1c of about 1% (although some of the differences were statistically significant). Variables associated with larger reductions were < 6 months between diabetes diagnosis and time to second-line therapy, older age, lower BMI, and male sex. But again, these effects were small.

How Might Choice of Therapy Affect Mortality?

Currie and associates[6] looked at the relative mortality risk for six treatment cohorts, including two associated with low risk for hypoglycemia:

  • metformin monotherapy

  • metformin plus any oral agent other than SU or meglitinides, or thiazolidinedione plus metformin or monotherapy

and four associated with a higher risk for hypoglycemia:

  • SU monotherapy

  • insulin monotherapy

  • SU or meglitinide monotherapy or in combination with metformin but excluding insulin

  • SU, meglitinide, or insulin as monotherapy or in combination with metformin

Because these cohort definitions overlap and could occur any time from 2004 to 2013, it was possible for patients to contribute data to more than one cohort.

Glucose control was initially categorized as A1c < 7%, 7%-8.4%, 8.5%-9.4%, and ≥ 9.5%, and then also according to A1c deciles. Patients were followed through January 2015 for the outcome of all-cause mortality. Comparisons of the effect of glycemic control on mortality were made within but not between cohorts.

For the two cohorts not associated with hypoglycemia, there was no difference in the risk for mortality between the lower and moderate A1c categories, but there was a significant increase in mortality risk in the high and very high A1c categories when compared with the moderate categories.

Conversely, in the cohorts associated with hypoglycemia, there was an increased mortality risk in the low versus moderate A1c categories, but not in the high and very high categories (versus the moderate category).

Thus, it appears that individualization of glycemic targets may need to include consideration of the therapy used to achieve the target to account for mortality risk.

Analysis and Commentary

Because diabetes is characterized by hyperglycemia, glycemic control is the cornerstone of disease management. Seminal research suggested that lower A1c reduced microvascular and (probably) macrovascular complications,[7,8,9] and that many patients who achieved early glycemic control would experience "legacy" benefits for many years to come.[10,11]

However, subsequent landmark trials were unable to show that intensive glucose lowering reduced cardiovascular risk; such therapy might even increase mortality.[12,13,14] Thus, while no one would argue that glycemic control is unimportant, the defining optimal A1c remains elusive.

With the exception of insulin, A1c reductions are about the same for all second-line agents. Indeed, Khunti and colleagues[5] show that SUs may provide the greatest A1c reduction, although differences were quite small. Strictly on the basis of A1c, then, SUs would remain the best option.

However, newer drug classes, specifically the sodium-glucose co-transporter 2 (SGLT-2) inhibitors and glucagonlike peptide-1 (GLP-1) receptor agonists, have demonstrated cardiovascular and/or mortality benefit, suggesting that how A1c is reduced may be more important than how much it is lowered.[15]

At least some of the confusion surrounding the relationship between A1c and cardiovascular/mortality risk probably arises from the fact that the relationship is not linear. Observational studies have demonstrated a U-shaped relationship whereby risk increases in both the upper and lower segments of the A1c continuum.[16,17,18,19,20]

The paper by Currie and colleagues[6] is a deep dive into the U-shaped relationship with special attention to second-line therapies. That their results suggest an increased mortality risk for A1c < 7% but no increased risk for A1c ≥ 8.5% for insulin and/or SU is reassuring for older patients with longer duration of diabetes who may have trouble attaining "tight" glycemic control.

Conversely, the fact that there was no increased mortality risk for A1c < 7% but an increased risk for A1c ≥ 8.5% for metformin alone and in combination with therapies not associated with hypoglycemia suggests that younger patients in relatively early stages of diabetes may be able to aggressively seek tight glycemic control with these drugs.

In summary, there is still no clear answer or "magic bullet" for treating hyperglycemia in diabetes. The studies reviewed above provide some understanding of the many anti-hyperglycemic treatment options that are now available. Nevertheless, if cardiovascular risk reduction is the goal, we should remember that the best way to achieve it is through blood pressure and LDL cholesterol control.[21]


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.