Author(s)


Reviewed By


Citation
Eng C, Kramer CK, Zinman B, Retnakaran R. Glucagon-like peptide-1 receptor agonist and basal insulin combination treatment for the management of type 2 diabetes: a systematic review and meta-analysis. Lancet 2014; 384: 2228-34

The treatment options for type 2 diabetes mellitus (T2DM) have increased dramatically over the past decade. Despite these advancements, treatment recommendations remain largely unchanged. The American Diabetes Association (ADA) recommends that patients newly diagnosed with T2DM modify their diet, exercise, and take oral medication (i.e. metformin).  Only later should injectable therapies be considered. Until recently, the ADA’s treatment algorithm recommended basal-bolus insulin strategies only when three-drug oral combinations didn’t work or when the A1c was greater than 10%.1

Glucagon-like peptide-1 receptor agonists (GLP-1 RAs) are generally well tolerated, lead to significant improvements in HbA1c, stimulate glucose-dependent insulin release, and potentially preserve beta-cell mass.2,3 There are five FDA-approved GLP-1 RAs available and they are currently recommended as second-line therapy. Should GLP-1 RAs be used with basal insulin?   Is this combination safer or more effective than basal-bolus insulin regimens?4,5

A recent meta-analysis examined the efficacy and safety of the GLP-1 RA plus basal insulin combination.5  Eligible studies were published between January 1, 1950 and July 29, 2014, enrolled adults with T2DM, were at least 8 weeks in duration, and provided data regarding changes in HbA1c, the proportion achieving an HbA1c less than 7.0%, or hypoglycemic events. All trials included an active comparator. Retrospective and observational studies were excluded. Lastly, results were reported consistent with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement.

The meta-analysis examined changes in HbA1c, goal achievement (HbA1c < 7%), weight, and hypoglycemia using a random-effects model to account for heterogeneity between trials. Measures of heterogeneity including Cochran’s Q and I2 were used to identify significant heterogeneity. Trials were evaluated for publication bias.  Lastly, a pre-specified sensitivity analysis comparing only GLP-1 RA + basal to basal/bolus regimens was performed.

Fifteen trials (n=4348 subjects) were included in the analysis with an average duration of 24.8 weeks (range: 12-36 weeks). Six studies compared a GLP-1 RA to placebo with concomitant basal insulin ± oral anti-hyperglycemic medications (OAMs), while the other trials compared bolus insulin directly to a GLP-1 RA (4 studies), combination insulin therapies with a GLP-1 RA (2 studies), or a fixed-combination product GLP-1 RA plus insulin glargine (1 study). Although trial designs differed, all trials randomized treatment allocations and most (80%) were multi-center.  A majority of trials didn’t provide sufficient evidence that participants or researchers were blinded to allocations. Liraglutide was the GLP-1 RA comparator in 7 studies. Lixidenatide (4 studies), exenatide (3 studies), and albiglutide (1 study) were used in the other eight studies.

The pooled baseline HbA1c for the study subjects included in the analysis was 8.13% (7.4-8.8%), duration of DM since diagnosis was 12.2 years (7.9-17.1 years), and BMI was 32.9 kg/m2 (25.2-39.6 kg/m2).  The results of analysis are detailed in Table 1.

Table 1:  GLP-1 RA plus Basal Insulin Effects on HbA1c and Weight

Primary Endpoints

Weighted mean difference
(95% CI)

Relative Risk (RR)
(95% CI)

Cochran-Q
(p-value)

I2

Egger
(p-value)

NNT§

Reduction in HbA1c vs. any treatment

-0.44%
(-0.60 to -0.29)

<0.0001

96.6%

0.0047

Reduction in weight vs. any treatment

-3.22 kg
(-4.90 to -1.54)

<0.0001

99.6%

0.81

Proportion achieving HbA1c < 7% vs. any treatment

1.92
(1.43-2.56)

<0.0001

93.3%

0.16

6

Incidence of hypoglycemia vs. any treatment

0.99
(0.76-1.29)

<0.0001

77.1%

0.81

Sensitivity Analysis

 

Reduction in HbA1c vs. basal-bolus insulin

-0.10%
(-0.17 to -0.02)

 

0.470

0.0%

NS

Reduction in weight vs. basal-bolus insulin

-5.66 kg
(-9.80 to -1.51)

 

<0.0001

98.7%

Proportion achieving HbA1c less than 7% vs. basal-bolus insulin

1.07
(0.91-1.26)

0.287

12%

Incidence of hypoglycemia vs. basal-bolus insulin

0.67
(0.56-0.80)

0.526

0.0%

3

NS = not statistically significant; CI = confidence interval; NNT = number needed to treat

 

The combination of a GLP-1 RA plus basal insulin significantly lowered HbA1c and a larger proportion of patients achieved an HbA1c less than 7% when compared to any other treatment without increasing the risk of hypoglycemia.  The combination also promoted significant but modest weight loss (-3.22 kg). It should be noted that even modest weight loss may impart significant improvements in cardiovascular risk in patients with T2DM9 and appears to be sustained with continued therapy.10  In the sensitivity analysis which included three trials comparing GLP-1 RA plus basal insulin to a basal-bolus insulin regimen, the GLP-1 RA combination resulted in a statistically lower, but clinically similar, HbA1c reduction (-0.1%) with no difference in the proportion achieving an HbA1c less than 7%. Although similarly effective, the GLP-1 RA combination was less likely to cause hypoglycemia (33% RR reduction) and positively impacted weight (-5.66 kg difference).

In the primary analysis there was significant heterogeneity indicating the treatment effects were markedly different among the included studies. However, when comparing studies of similar design in the sensitivity analysis (GLP-1 RA combination vs. basal/bolus only), heterogeneity was lower. Lastly, trim-and-fill testing indicate that publication bias is unlikely to have impacted the pooled estimate.

Strengths of this meta-analysis include adherence to the PRISMA statement, inclusion of only well-designed control trials (randomized, multicenter, minimized selection bias), and a pre-specified sensitivity analysis. Results were consistent despite differences in GLP-1 RA formulations, concomitant treatments, and sequence of initiation of GLP-1 vs. basal insulin.  Heterogeneity was reasonable when comparing the results of studies using similar treatments. Lastly, these findings are consistent with a prior systematic review that used broader study inclusion criteria.8

There are several limitation to this analysis.  The long-term safety and durability of the GLP-1 RAs was not examined.  Most studies were open-label and thus unblinded.  Open-label designs may exaggerate the effects compared to double-blinded designs.11  There was no differentiation between short and long-acting GLP-1 RAs.  And there was no data included in this analysis regarding the use of this combination in newly diagnosed T2DM who likely have more residual beta-cell mass.  The heterogeneity between trials in the primary analysis are undoubtedly due to differences in trial design particularly with regard to the use of active comparators versus placebo. Trials evaluating the non-inferiority of a GLP-1 RA combination to an active comparator were designed to show no differences in efficacy but this may have inadvertently minimized other favorable attributes.

Lastly, GLP-1 RAs differ in terms of their pharmacokinetic and pharmacodynamic properties.  Assuming there is a class effect is unwise.  There are several important distinctions between GLP-1 RAs including differences in efficacy, adverse effects, timing and frequency of administration, and effect on post-prandial (PPG) versus fasting plasma glucose (FPG).6,7  Exenatide, for example, seems to affect PPG more and FPG less than longer-acting formulations due to higher GLP-1 RA concentrations which may have a greater impact on gastric emptying and satiety.7,8  Therefore, GLP-1 RAs with a shorter biologic half-life may be better suited to replace bolus insulin.7  Clearly more data are needed to determine if the differences between GLP-1 RAs are clinically important.

The combination of basal insulin plus GLP-1 RA is effective, reduces the risk for hypoglycemia, and has a favorable impact on weight compared to bolus insulin in patients with T2DM. These results question reliance on basal-bolus insulin as the “less step” in the treatment paradigm.   What do you think?

  1. American Diabetes Association. 7. Approaches to Glycemic Treatment. Diabetes care 2015; 38(Supplement 1): S41-S48.
  2. Drucker DJ, Nauck MA. The incretin system: glucagon-like peptide-1 receptor agonists and dipeptidyl peptidase-4 inhibitors in type 2 diabetes. Lancet 2006; 368: 1696-1705.
  3. Egan JM, Clocquet AR, Elahi D. The insulinotropic effect of acute exendin-4 administered to humans: comparison of nondiabetic state to type 2 diabetes. J Clin Endo Metab 2002; 87: 1282-90.
  4. Young LA, Buse JB. GLP-1 receptor agonists and basal insulin in type 2 diabetes. Lancet 2014; 384: 2180-1.
  5. Eng C, Kramer CK, Zinman B, Retnakaran R. Glucagon-like peptide-1 receptor agonist and basal insulin combination treatment for the management of type 2 diabetes: a systematic review and meta-analysis. Lancet. 2014; 384: 2228-34.
  6. Meier JJ. GLP-1 receptor agonists for individualized treatment of type 2 diabetes mellitus. Nature reviews. Endocrinology 2012; 8: 728-42.
  7. Trujillo JM, Nuffer W. GLP-1 Receptor Agonists for Type 2 Diabetes Mellitus: Recent Developments and Emerging Agents. Pharmacotherapy 2014; 34: 1174-86.
  8. Balena R, Hensley IE, Miller S, Barnett AH. Combination therapy with GLP-1 receptor agonists and basal insulin: a systematic review of the literature. Diab Obesity Metab 2013; 15: 485-502.
  9. Wing RR, Lang W, Wadden TA, et al. Benefits of modest weight loss in improving cardiovascular risk factors in overweight and obese individuals with type 2 diabetes. Diabetes care 2011; 34: 1481-6.
  10. Klonoff DC, Buse JB, Nielsen LL, et al. Exenatide effects on diabetes, obesity, cardiovascular risk factors and hepatic biomarkers in patients with type 2 diabetes treated for at least 3 years. CMRO 2008; 24: 275-286.
  11. Schulz KF, Chalmers I, Hayes RJ, Altman DG. Empirical evidence of bias. Dimensions of methodological quality associated with estimates of treatment effects in controlled trials. JAMA 1995; 273: 408-412.