The case for basal analogue insulins as first‐line insulins: back to the future?

Since the launch of glargine insulin (Lantus) at the start of this century and the arrival of detemir insulin (Levemir) insulin in 2004, there has been a reduction in the amount of neutral protamine Hagedorn (NPH) insulin used. The majority of new prescriptions for basal insulins over the last few years in the UK have been for glargine or detemir and these insulins now account for between 38 and 97% of basal insulin prescriptions (excluding biphasic insulins) in the UK.1 Is this justified on clinical or monetary grounds? The claims made of the newer basal insulin analogues are that they are better than NPH both in terms of predictability and duration of action as well as in lowering the risk of hypoglycaemia;2,3 and in the case of detemir, as well as this, there seems to be less weight gain associated with its use compared with glargine.3-5 Both glargine and detemir are licensed in the European Union for once-a-day administration. What is the utility of these attributes in the real world of diabetes management? Type 2 diabetes accounts for the majority of insulin use due to the greater numbers of patients with this condition as compared with type 1; NICE, in drawing up its guideline for the management of glycaemia in type 2 diabetes,6,7 looked at the available evidence, and concluded that, while there is support for glargine and detemir to lower the incidence of hypoglycaemic events, the use of these insulins as first-line insulins in type 2 diabetes could not be justified on health economic grounds. NICE did recommend their use in certain circumstances, such as in those who developed problematic hypoglycaemia on NPH, and in those needing once-daily insulin, e.g. those dependent on carers7 and/or where other complicated regimens would be required. One can ask the question: ‘Are those patients in the studies demonstrating reduction in hypoglycaemia representative of the type 2 diabetic population for whom we are considering insulin prescription?’ To answer this, one must understand the fundamental difficulty in studying glycaemic control in insulin-treated type 2 diabetes, which is that it is a very heterogeneous condition, with some requiring insulin only once a day and others requiring complex multi-dose regimens; this is because it is progressive.8 NICE recognised this heterogeneity and discussed the need for other regimens after once-daily insulin, such as fixed-mix and variable regimens.6 As studies, by their nature, are forced to have exclusion and inclusion criteria we can, unfortunately, not pick one set of studies and conclude that a particular insulin, or insulins, is, or are, the first choice for all individuals. Moreover, until lately the assumption has always been that aggressive lowering of the HbA1c is appropriate in itself; however, the publication of the ACCORD9 and ADVANCE10 studies has undermined this and there seems to be a mood, at least among professionals in diabetes care, that we should be cautious and perhaps concentrate on anti-lipid and anti-hypertensive therapies with less focus on glycaemia.11,12 In type 1 diabetes the use of glargine and detemir has become almost universal as the dogma that basal insulins with long duration of actions are de rigueur has taken hold over the last 10 years. Nevertheless, those of us who manage type 1 patients have become increasingly aware that this dogma is not always applicable. We now manage this group of people with carbohydrate counting and insulin dose adjustment, and also have a significant minority on insulin pump therapy; the whole rationale for this approach is flexibility, not rigidity.13-15 Often, without necessarily thinking too deeply about things, we have moved on occasions to twice-daily detemir or even twice-daily glargine; which is counter-intuitive. Why do we not use twice-daily NPH? A study by Lawrence et al.16 looked at what happened to Dose Adjustment For Normal Eating (DAFNE)-trained type 1 patients and simply observed what happened to their HbA1c upon completing the course and six months later. They found in this retrospective analysis that those patients who chose to stay on twice-daily NPH for their basal insulin regimen, rather than taking once-a-day detemir or glargine, fared better in terms of glycaemic control and had no differences in hypoglycaemic episodes. The need for a randomised controlled trial of twice-daily NPH versus both once- and twice-daily detemir and glargine is clear. The cost of these basal insulin analogues is about double that of NPH; to give just one example from the same manufacturer comparing disposable pens, the cost of Insuman Basal in the SoloStar pens is £19.80 (€21.78) per pack of five and for glargine in the same pen is £40.36 (€44.40).17 Similar differences exist for Levemir and Insulatard, although the latter is not available at the moment in disposable pens. Does this matter? After all, the cost of diabetes to the health economy is mainly from complications. A recent report1 showed that the cost of diabetes drugs from the British National Formulary section 6.1 (i.e. glycaemic lowering) has reached 8.4% of the NHS drug bill in 2010–11, up from 6.6% in 2005–6. This equates to £725m spent on diabetes drugs in 2010–11 and, of this, £269m were on basal analogue insulins. Moreover, a study from Cardiff18 by Holden et al. supports this, estimating that if guidelines had been followed between 2000 and 2009 the UK NHS would have saved £625m. To put this spending in perspective, taking the 2010–11 figures quoted above: if NICE's guidance had been followed in type 2 diabetic patients alone and, for illustrative purposes, imagine 50% of those on insulin were taking the basal analogues and the other 50% were on NPH at roughly the same dosages, then the NHS would have saved about 25% of this spend, which is approximately £67m. This equates to about 2233 middle-increment Band 6 diabetes specialist nurses or dietitians, and still leaves 50% of patients on newer basal analogues. Assuming that the move to more NPH usage would indeed result in more nocturnal hypoglycaemia, we need to try and account for this; sophisticated health modelling is required. After the (English) NICE type 2 diabetes guideline was published, a Canadian research group19 used one such mathematical model (Centre for Outcomes Research Diabetes Model) to compare cost effectiveness of NPH insulin versus long-acting insulin analogues. The input into this model included: age; HbA1c; patient characteristics; type of diabetes; and history of diabetes-related complications. The clinical effects of therapy (HbA1c level and hypoglycaemia) were also used as inputs for the model and were derived from a meta-analysis of randomised controlled trials. The primary outcome measure was the quality-adjusted life-year (QALY). In type 1 diabetes, insulin glargine and detemir did generate more QALYs than did NPH but at an increased cost; and the probability that, as compared to NPH, glargine and detemir were more cost-effective (at cost-effectiveness threshold of Can$50 000 per QALY), was 42.5% and 29.2% respectively. In type 2 diabetes, glargine was more effective but more costly than NPH, and detemir was actually more costly and less effective than NPH; and the probability that, as compared to NPH, glargine and detemir were more cost-effective, was 25.1% and 10.8% respectively. The incremental costs per QALY decreased when ‘fear of hypoglycaemia’ was included in the model and increased when ‘no increase in HbA1c between the comparators’ was incorporated. In an earlier cost effective evaluation by NICE20 (TAG53), the incremental cost effectiveness ratios for insulin glargine when compared with NPH were around £32 000 per QALY for type 1 diabetes and around £120 000 for type 2 diabetes; and when ‘fear of hypoglycaemia’ was included in cost-effective analysis, the incremental cost per QALY for insulin glargine compared with NPH reduced from £32 000 to £3500 for type 1 diabetes and from £120 000 to £32 500 for type 2 diabetes, suggesting therefore that their use could be justified in type 1 diabetes but less so in type 2. In order to further refine these models it would be entirely justified, and indeed proper, when doing such modelling to also include the beneficial effects of an extra 2000+ diabetes health care professionals available for patient education and to pay particular attention to their role in educating patients on how to avoid hypoglycaemia. Should not our strategy therefore be to start NPH in type 1 and type 2 diabetes and move to the basal analogues if indicated in the individual patient? After all, as health care professionals in the clinic, we are treating individuals, not populations; but, on the other hand, do we not have a duty to spend taxpayers' money judiciously? As the Scottish legal system has it, the case for the use of basal analogue insulins as first-line insulins is clearly ‘not proven’, neither in type 2 nor in type 1 diabetes. Despite this, their usefulness is clear as second-line insulin therapies and they will continue to be used. MWS was a member of the guideline development group for the type 2 diabetes guideline published by NICE in 2008. He has also received honoraria for giving talks at educational meetings from Novo Nordisk, Lilly, Aventis, Merck, GlaxoSmithKline, and Bristol-Myers Squibb. CND has received honoraria for giving talks at educational meetings from Lilly, Sanofi-Aventis, and Bristol-Myers Squibb. IM has nothing to declare.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.