Is insulin the preferred treatment for HbA1c >9%?

The algorithms and guidelines of the American Association of Clinical Endocrinologists and the American Diabetes Association recommend that insulin administration be strongly considered for people with type 2 diabetes (T2D) with HbA1c levels exceeding 9.0% and 10%, respectively. Although the caveat is given in both sets of recommendations that this is particularly appropriate when patients are "symptomatic," referring to urinary frequency with increased thirst and appetite, weight loss, and ketosis, the clinical definition of such presentations may be ill-defined, and it is noteworthy that both documents consider insulin to offer particular benefit under such circumstances. However, with multiple options for glycemic treatment, it is of interest to reconsider this argument for insulin use. It should be recalled that in the UK Prospective Diabetes Study, diet alone was associated with a reduction in HbA1c from 9% to 7%. Drug-naïve people with T2D do often show surprisingly strong reductions in HbA1c with metformin-based dual-agent oral treatment approaches; a recent report showed that even with baseline HbA1c >11%, the combination of metformin with a sulfonylurea, pioglitazone, or sitagliptin was associated with reduction in HbA1c from 11.6% to 6.0%. A 32-week study of the combination of rosiglitazone with metformin in patients with mean baseline HbA1c 8.9% showed a mean HbA1c reduction of 2.3%, and an open-label cohort with baseline HbA1c 11.8% had a reduction in HbA1c to 7.8%. With metformin plus sitagliptin, a mean placebo-adjusted HbA1c reduction of 2.1% from a baseline of 8.8% was reported, with those patients with baseline HbA1c >9% having a 2.6% reduction in HbA1c, and an open-label cohort with baseline HbA1c 11.2% having a 2.9% reduction in HbA1c. Similar 2% HbA1c reductions from baseline levels of 9.1% were seen with metformin in initial combination with the sodium-glucose cotransporter 2 (SGLT2) inhibitor dapagliflozin. Although such dual oral agent approaches are more effective than monotherapy, with a combination regimen the HbA1c reduction will not be directly additive, because the expected reduction decreases at lower baseline HbA1c levels. As an example of this, administration of canagliflozin 300 mg daily to patients with baseline HbA1c >9% reduced levels from 9.6% by 1.8%, whereas at a baseline HbA1c of 10% either canagliflozin 300 mg or metformin 2 g/day reduced HbA1c by 2%; the addition of both agents led to an HbA1c reduction by somewhat less than 3%, which appears concordant with a reduction by the second agent from approximately 8% (10% to 2%). Similar less-than-additive effects of the addition of exenatide QW to dapagliflozin have been reported, with HbA1c reduction from a baseline of 10.0%-10.1% of 1.9% and 1.6% with the individual agents, respectively, and a reduction of 2.2% with their combination. However, one may consider these approaches inferior to the expected HbA1c reduction with insulin, suggesting that insulin should, indeed, be the preferred treatment for people with T2D and HbA1c >9%. Rather, studies comparing basal insulin directly with glucagon-like peptide-1 (GLP-1) receptor agonists (RA) suggest that the latter agents may offer superior benefit. The Diabetes Therapy Utilization: Researching Changes in HBA1C, Weight, and Other Factors Through Intervention with Exenatide Once Weekly (DURATION)-3 and Liraglutide Effect and Action in Diabetes (LEAD)-5 studies compared exenatide QW and liraglutide, respectively, with insulin glargine. Those study participants in the highest quartile of baseline HbA1c had levels ≥9.0% and ≥8.9%, with the GLP-1RA leading to 0.3% and 0.2% greater reductions in HbA1c, respectively, than insulin glargine. Another study comparing T2D patients receiving oral agents given liraglutide with those given insulin glargine showed that those in the highest baseline HbA1c quartile (mean 10.6%) had an HbA1c reduction of 3.1% with either agent. In the exenatide QW study, the reduction in HbA1c with this agent exceeded that with insulin glargine for those groups of study participants with HbA1c 9.0%-9.4%, 9.5%-9.9%, 10.0%-10.4%, 10.5%-10.9%, and even ≥11.0%. Similar superiority of the HbA1c-lowering effect of exenatide QW compared with that of insulin glargine was reported in a study with baseline HbA1c 8.5%. An individual-patient meta-analysis of six studies of another weekly GLP-1RA, namely dulaglutide, showed that at a baseline HbA1c of 10% the expected HbA1c reduction would be nearly 2.5%, and a study directly comparing dulaglutide with insulin glargine also showed a superior HbA1c-lowering effect of the former. Another advantage of the GLP-1RAs is their association with weight loss, rather than the weight gain associated with insulin treatment. An interesting potential combination is that of a GLP-1RA with a thiazolidinedione. In a study comparing the addition of exenatide QW and pioglitazone with the addition of basal-bolus insulin in 101 people receiving sulfonylureas and metformin with baseline HbA1c >10%, HbA1c fell from >11% by >4% compared with <4%, respectively, and the GLP-1RA plus thiazolidinedione treatment was associated with less weight gain and hypoglycemia. What can we conclude? Should HbA1c 11% be the new "use insulin" point? Insulin is an important part of our armamentarium for T2D, and is certainly needed for many patients, but with current therapeutic approaches including metformin, incretin-based treatments, SGLT2 inhibitors, and, possibly, thiazolidinediones, we can reconsider its use in many instances. Although there is no doubt that insulin is necessary for truly uncontrolled diabetes, we may wish to better define its correct indications.