Adding insulin to oral agents in poorly controlled type 2 diabetes - trial evidence to guide practice

Interim results from a controlled trial in patients with type 2 diabetes and poor glycaemic control showed that adding one of three simple insulin regimens to maximally tolerated oral antidiabetic drugs benefited some patients; regimens giving greater control also caused more adverse events.

The 4-T (Treating to Target in Type 2 Diabetes) study aims to clarify the use of insulin in patients with type 2 diabetes poorly controlled despite maximum tolerated doses of oral antidiabetic drugs. While the addition of insulin is widely practised in such patients, there is little evidence from large clinical trials to guide such use: the first year of the three-year 4-T study, reported here, compared three simple insulin regimes - more complex regimens are being compared in the remaining two years. Patients were adults with type 2 diabetes not previously treated with insulin, and had poor glycaemic control (glycated haemoglobin, HbA1c, 7-10%) with maximally tolerated doses of sulphonylurea plus metformin, or individual drugs if the other was not tolerated. They were randomised to open-label treatment with one of three regimens: basal insulin (insulin detemir at bedtime), prandial insulin (insulin aspart immediately before meals), or biphasic insulin (NovoMix 30 twice daily). Primary outcome was glycaemic control as HbA1c at one year; secondary outcomes included weight gain and hypoglycaemia as well as other measures of control.

A total of 936 patients were screened for inclusion, and 708 were randomised; 235 to biphasic, 239 to prandial, and 234 to basal insulin. The most common reason for exclusion was a baseline HbA1c outside the range 7-10%. Withdrawals were not statistically different across the three groups (5.5%, 7.1%, and 4.3% respectively). Maximum fall in HbA1c levels was achieved by week 24 in all groups, however at week 52 the biphasic and prandial groups had significantly lower mean levels than the basal group (7.3% and 7.2% vs. 7.6% respectively, p<0.001 for both comparisons): most differences were in patients with poorest baseline control - there was less difference between the three regimens in patients with a baseline HbA1c of <8.5%. Patients in the prandial group were most likely to reach HbA1c of 6.5%, however even in this group only a quarter did so (prandial 25%, biphasic 17.0%, basal 8.1%). Patients in the prandial group were most likely to have hypoglycaemic events and also gained the most weight (basal least for both).

The authors conclude that adding a single insulin to maximal oral therapy in these patients achieved target glycaemic control, measured as HbA1c level, in only a minority of patients. Biphasic and prandial insulin gave better control but at the cost of more hypoglycaemic events and more weight gain. They conclude that the results support basal insulin therapy for a first line option, as it has lowest levels of adverse effects, and is simple and convenient for patients; however many will need rapid intensification of therapy. The final phase of the study is examining more complex regimens in this patient group.

According to an accompanying editorial, “the 4-T study provides a clear indication that prandial and biphasic insulin formulations are suboptimal choices for insulin initiation and probably expose patients to an unnecessarily high risk of hypoglycemia without clinically important benefit.” Results are awaited from the second phase of the study to define the best next step for patients who do not reach their glucose target whilst on basal insulin alone, since they are most likely to benefit from additional prandial insulin.

For now, the author suggests that the recommendations for starting insulin therapy need not change as a result of this study: for patients with a HbA1c > 7% on maximal doses of two oral agents, the best approach is to continue metformin and add a basal insulin; sulfonylureas are not synergistic with insulin and should generally be stopped. He adds that choice of strategies for insulin initiation is probably less important than taking steps to start insulin in patients who need it. Furthermore, he stresses that though it is important to focus on glucose levels, clinicians should be aggressive with BP management since hypertension contributes at least as much as glucose to overall cardiovascular risk. In addition, aspirin, lipid-lowering therapies, smoking cessation, exercise and weight-loss programmes should be initiated when appropriate. He concludes that “achieving these integrated goals saves lives, whatever insulin formulation is chosen.”

New Engl J Med, published early online 21 September 2007; doi: 10.1056/NEJMoa075392 (link to abstract); New Engl J Med, published early online 21 September 2007; doi: 10.1056/NEJMe078196