Rapid-acting insulin analogues in diabetes mellitus type 1: Superiority not proven

There is currently no evidence available of a superiority of rapid-acting insulin analogues over human insulin in the treatment of adult patients with diabetes mellitus type 1. The evidential value and design of studies available so far are inadequate and do not allow conclusions regarding most patient-relevant therapy goals, such as the reduction in long-term complications or overall mortality. Due to the lack of data, the benefit of rapid-acting insulin analogues in children and adolescents is unclear. Although one of the manufacturers conducted long-term comparative studies in this group of patients, it is withholding some of the results. This is the result of the final report of the Institute for Quality and Efficiency in Health Care (IQWiG), which was published on 6 June 2007.

The German Federal Joint Committee commissioned IQWiG to compare the benefit of rapid-acting insulin analogues versus human insulin, as well as to compare the benefit of various rapid-acting insulin analogues with each other. IQWiG assessed all 3 rapid-acting insulin analogues approved in Germany: insulin aspart (tradename in Germany: Novorapid), insulin lispro (tradenames in Germany: Humalog, Liprolog), and insulin glulisine (tradename in Germany: Apidra).

Effects determinable after 6 months at the earliest

The literature search retrieved a total of 9 published comparative studies, in which patients were observed for at least 24 weeks. Only studies with this minimum duration were included, as they give patients sufficient time to adjust to the new medication, and to observe the treatment effects under stable application. Eight of these studies compared either insulin aspart or insulin lispro with human insulin; no such study was available for glulisine. The only study available that compared two analogues referred to glulisine and lispro.

No long-term studies on insulin pump therapy available

Regarding insulin pump therapy, no study lasting at least 24 weeks was available. Therefore, it remains unclear whether patients would benefit and which advantage patients would have by using this form of administration. The same applies to children and adolescents, as only fully published short-term studies are available in this population so far. The company Novo Nordisk sponsored 2 completed long-term studies in children and adolescents. However, to date, both studies have only been partially published. In contrast to the manufacturers Sanofi and Lilly, Novo Nordisk was not prepared to provide the information needed for the report.

Studies not blinded

The conclusions of the 9 studies included in the evaluation are only of limited robustness. None of the studies was blinded, i.e. both the patient and the physician knew which type of insulin was being injected. Without blinding, there is a danger that patients, knowing their type of insulin, behave differently, which would subsequently lead to a bias in the results of the study. Moreover, inconsistent statements on important issues, which could not be clarified, were often made in the study documents.

No conclusions on important therapy goals possible

Even though patients have been treated with insulin analogues for 10 years, it is still unclear as to how these types of insulin affect long-term complications of diabetes type 1, mortality, and the necessity of hospital admissions. Regarding the reduction in blood glucose levels (measured by means of HbA1c), patients treated with insulin aspart had, on average, lower levels. However, these statistical differences were so small that an effect on patients' health is not to be expected.

Insulin lispro may prevent nocturnal hypoglycaemia better than insulin glulisine. The only study that compared these insulin analogues provided first indications, but no reliable evidence.

Rules of a fair comparison violated

In some studies, patients treated with insulin analogues assessed their quality of life as higher and were more satisfied with treatment than patients who injected human insulin. IQWiG did not evaluate this finding as evidence of an additional benefit of insulin analogues, as it was not based on a fair comparison: In the human insulin group, patients were requested to adhere to a fixed injection-meal interval; this was not the case in the insulin analogue group. It is therefore unclear whether the larger treatment satisfaction was caused by the drug class itself, or by the different forms of application prescribed by physicians.

Procedure of report production

IQWiG published the preliminary report on the Internet on 4 September 2006 and invited interested persons and parties to submit comments. A total of 397 comments were received by 2 October. However, only 37 were classified as "substantial”, i.e. comments that named additional studies, criticised the interpretation of studies already presented by IQWiG, or questioned the methodology of the report, and supported these arguments with citations from the literature. A large part of the remaining comments consisted of standard letters, which various organisations representing patients or their relatives had provided.

Open issues pertaining to the substantial comments were discussed with the authors of the comments on 16 November 2006 in Cologne. The meeting minutes of this oral debate, as well as the comments themselves, are documented in the appendices of the final report. The arguments presented in the comments are discussed in detail in the report, in particular in the discussion section (pp. 114-133). The final report was completed before the GKV-WSG, the "Act to promote competition of the statutory health insurance”, became operative on 1 April 2007, and was forwarded to the contracting agency, the Federal Joint Committee. If the Federal Joint Committee makes a directive decision based on the final report, this decision will also be debated.

The role of rapid-acting insulin analogues in the treatment of diabetes type 2 was assessed in a separate commission; the relevant final report had already been published by IQWiG in December 2005. This report also came to the conclusion that evidence of an additional benefit of insulin analogues has yet to be provided.

Contact: info@iqwig.de