Intended for healthcare professionals

Feature Diabetes

How small changes led to big profits for insulin manufacturers

BMJ 2010; 341 doi: (Published 15 December 2010) Cite this as: BMJ 2010;341:c7139
  1. Deborah Cohen, investigations editor1,
  2. Philip Carter, journalist2
  1. 1BMJ, London WC1H 9JR, UK
  2. 2London
  1. dcohen{at}

In the first of a series of investigations by the BMJ and Channel 4 News, Deborah Cohen and Philip Carter discover why more expensive analogue insulins are increasingly prescribed instead of cheaper human insulin despite lack of evidence of benefit for patients with type 2 diabetes

Diabetes is a market worth fighting for—it’s the fourth biggest global therapeutic class, generating total sales in 2009 of $30.4bn (£19bn; €23bn).1 In the UK alone, diabetes is estimated to cost the NHS some £9bn—about 10% of its entire budget.2

In May 2010, Sanofi-aventis briefed investors in London on the emerging opportunity.3 The company stated that diabetes remains one of the largest growth opportunities in healthcare and one of the fastest growing markets—outstripping growth in the wider drug market by 5% a year for the past six years. It was, said the company, destined to grow to €50bn by 2015. Sanofi laid out the scale of its ambitions clearly: to be the number one firm in the diabetes market.3 4

The growth in diabetes drug revenue has been caused in part by rising numbers of people with diabetes; partly by more aggressive approaches to glucose lowering; but mainly by increasing unit costs for insulin5—a cost that can be attributed to the use of more expensive recombinant insulin analogues.

Sanofi’s ambition to be, in effect, the dominant force means wresting control from the traditional heavyweights of the insulin world, Novo Nordisk and Eli Lilly. However, health services and individuals across the world are paying the price for this market drive, which has seen more expensive analogue insulins taking the place of cheaper human insulin.

Added value?

Insulin comes in different forms: short acting, basal (intermediate and long acting), and a mixture of the two types called biphasic insulin. Each of these types also comes in animal, human, and analogue forms, with their own price differentials. Analogue insulins are molecularly tweaked to make rapid acting versions act a little more rapidly and long acting ones deliver insulin to the body more slowly than conventional human insulins. Analogue insulins have already taken the market by storm. In the UK, analogues now account for some 80% of insulin prescription items.6 But where the battle for market share has really been contested is in the basal insulin market. This is the form that people with type 2 diabetes are given if oral antidiabetic control is not working. Currently, two analogues compete—Novo Nordisk’s insulin detemir (Levemir) and Sanofi’s insulin glargine (Lantus).

The basal market accounts for 44% of all insulin use, rising from 37% in 2005. Insulin glargine now accounts for 66% of the global basal market, with the balance split equally at 17% between determir and the older and cheaper isophane insulin—which has seen its share fall from 41% of the basal market in 2005.4

But this boom does not seem to be supported by the scientific literature. A raft of meta-analyses, systematic reviews, and other papers come to some strikingly consistent conclusions.7 For most people with type 2 diabetes the extra cost does not correspond to the equivalent extra benefit. In the UK, a defined daily dose of insulin—the assumed maintenance dose in adults—costs over twice as much for the insulin analogues than for isophane insulin.6

It is something that concerns the World Health Organization. An unpublished position paper on global access to insulin, written by Hans Hogerzeil, director of WHO’s Essential Medicines and Pharmaceutical Policies, at the behest of director general, Margaret Chan, states: “The global insulin market is dominated by two giant companies who are pushing a new type of insulin analogue at 3-5× the cost of human insulin, while its marginal cost-effectiveness is not fully established.”

In the UK the view is clear. A health technology assessment earlier this year concluded that in type 2 diabetes, analogue insulins weren’t worth what the NHS was being charged in most cases.8

These findings were echoed by other healthcare organisations. In 2006, the German Institute for Quality and Economic Efficiency in Health Care (IQWIG), assessed analogue insulins. It concluded that analogues would not be reimbursable as their cost effectiveness was unproved.

“The result of this report was that insulin analogues have not shown superiority over human insulin; hence no higher price is justifiable,” says Peter Sawicki, former head of IQWIG. The price of short acting insulin analogues for type 2 diabetic patients in the German healthcare system was subsequently reduced.

Meanwhile NICE guidelines recommend use of human insulin as first choice, with insulin analogues to be considered in certain circumstances.9 And yet somehow the analogues have shrugged off this body of evidence and their sales continue to climb rapidly. Amanda Adler, chair of the NICE guidance committee,9 says, “I would estimate that around 90% of people with type 2 diabetes would probably do quite well on these human insulins compared with the long acting insulin analogues.” Of these, she says, two thirds would remain on human insulin. Figures from the University of Cardiff collated for the BMJ and Channel 4 News suggest that if only 50% of those using analogue insulins had been put on human insulin instead the NHS would have saved close to £250m in the cost of insulin over the past five years.

A spokesperson for the Department of Health said: “We would expect clinicians to take into account NICE’s clinical guidelines on type 2 diabetes when prescribing decisions are made.

“The National Prescribing Centre published advice in the summer on 15 key areas of medicines use that local NHS organisations should scrutinise in taking forward work on quality, innovation, improvement, and productivity (QIPP). This specifically drew attention to NICE’s guidance about targeting of long acting insulins and the need for doctors to assess the underlying causes of poor blood sugar control before considering whether to prescribe these newer, more expensive insulins.”

But it’s not just in the UK that analogues have taken hold. According to Novo Nordisk, analogues now account for about 70% of insulin use in the United States and 60% in Europe.10

Reasons for success

So given the evidence and the costs, why are analogues so popular? According to Nick Freemantle, an epidemiologist from Birmingham University, the demand seems to have come from patients and doctors. Indeed, Aventis didn’t realise what an important product they had. “I believe it to be related to the once a day dosage, which can be conducted in the privacy of the patient’s own home, and the perceived relative safety. Certainly Aventis underestimated it, having to put off launch in some countries because of higher than expected initial demand. Since launch, of course Aventis (now Sanofi-aventis) has vigorously promoted the product, but it has always been seen to be useful by many physicians,” he said.

Dr Adler suggests that this market share is largely down to clever marketing. During the production of the NICE guideline on type 2 diabetes, Norman Waugh, professor of public health at Aberdeen University, and other members of the NICE team asked general practitioners why there had been such a big shift to the long acting analogues. There were several reasons given, he said. There is a notion that insulin glargine—the first analogue—with a once daily injection made it easier to switch people in whom oral treatment was not working, on to insulin. “But isophane insulin can be used once daily in type 2 diabetes mellitus,” he said. There was also a “belief that hypoglycaemia was less common—true but the difference is less than was hyped,” he added. Indeed, one funded five year randomised trial by Sanofi-aventis—the longest conducted on glargine—showed a smaller drop in haemoglobin A1c in patients using glargine than in those using isophane insulin (mean change from baseline −0.55% v −0.76%; mean difference 0.21, 95% confidence interval 0.06 to 0.35).11

And a Sanofi-aventis funded and coauthored meta-analysis published this year suggests that severe hypoglycaemia was seen in 1.2% of patients using glargine compared with 1.1% using isophane insulin.12

“The only real differences seen were in nocturnal hypoglycaemia within the first six months of therapy,” Edwin Gale, a Bristol University diabetologist and editor of Diabetologica, said. Yet it is this aspect that the conclusion in the abstract emphasised.

“This meta-analysis of open label studies provides confidence that reductions of around 50% of risk for nocturnal hypoglycaemia can be achieved with using glargine instead of NPH [isophane insulin],” it said. However, according to Professor Gale: “Such episodes are usually easily avoidable with simple clinical management without change of insulin.”

Some of the claims about basal analogues have caught the eye of the regulators too. In October this year, claims made in advertisements for insulin detemir were challenged in the European Court of Justice. The case was brought by the Estonian medicines regulator, Ravimiamet, who questioned three claims made in an advert in the medical journal Lege Artis.

These were that detemir offered effective control of blood sugar with a low risk of hypoglycaemia; 68% of patients who took the drug did not gain, but rather lost weight; and 82% of patients in clinical studies injected once a day. Ravimiamet argued that the claims were not borne out by the summary of product characteristics—how a drug can be used and prescribed as agreed by the regulators.

The summary states that the reduction in hypoglycaemia is at night and that weight gain is less than with other types of basal insulin but does not mention weight loss; there is no mention that only 82% of patients required once daily injections.13 In ruling on the case the judge said that Novo Nordisk had gone beyond what was written in the summary of product characteristics. However, there could be situations where this was justifiable if the company could provide evidence of its claims.14 This may broaden the boundaries of what can be said in drug advertisements.

Another case brought to the Prescription Medicines Code of Practice Authority (PMCPA) by Eli Lilly in 2007 centred on an advert for insulin detemir in Diabetes Update featuring overweight adults in a swimming pool. “Levemir is changing figures,” the advert claimed, implying that the drug was associated with weight loss. Weight gain is a recognised side effect of insulin therapy, the PMCPA ruled, and while determir caused less weight gain than other insulins, there was no evidence to substantiate weight loss.

The panel considered that the advertisement was misleading as alleged even though boxed text contained the claim “Less weight gain than NPH and insulin glargine.” Breaches of the code were ruled and upheld on appeal.15

But another change in favour of the analogues was the devices used to deliver the insulin given. “The pen devices for giving the new insulins were thought to be easier to use,” Professor Waugh said.

Patients are often given a choice about which pen they like best—and this provides another area for market manipulation. According to Dr Adler, the new insulins have much fancier packaging—the “Mont Blanc of insulin devices”—than older human types. And much like the way razor blades can be used only with a certain manufacturer’s razor, refills are made to fit a specific pen device.

Professional support

But perhaps more alarmingly, Professor Waugh said that companies provided nurses. “[There is] the provision of nice nurses from manufacturers of long-acting analogues to GPs to get people started on insulin. No company will provide nurses to start people on generic isophane insulin,” he said.

In the UK, much diabetes care is general practice based and delegated to nurses, who make the decisions about the type of insulin and injection device. At the time Sanofi was rolling out insulin glargine in the early 2000s, the primary care sector for diabetes was hugely overloaded. It was a fraught period for GPs and practice nurses with limited skills in starting patients on insulin. There were backlogs of patients who needed to be started on insulin.

Sanofi seized on the opportunity, producing a marketing masterstroke to support insulin uptake and, by extension at least, the adoption of insulin glargine in the UK. The company funded specialist diabetes nurses to train primary care staff in managing diabetes and insulin initiation. It worked with Warwick University to create a training scheme for health professionals working in diabetes care. The scheme was branded Insulin for Life. Although the training was not dependent on doctors prescribing insulin glargine, according to Martin Hadley-Brown, chair of the Primary Care Diabetes Society, healthcare professionals became familiar with it and many patients would end up using glargine as a result. “I think it was a masterful piece of marketing. The company came in with a product that we felt we needed at a time we felt ready to use it, and with the training to use it. So in marketing terms, I think it’s probably used around the drug industry as an example of good marketing that they’d all die to mimic,” he said.

Sanofi has maintained that Insulin for Life was not explicitly linked to the promotion of insulin glargine. But a simultaneous marketing drive for the drug muddied the waters. Glargine was seen by some to be too closely allied to the scheme. Complaints were received by the PMCPA and Medicines and Healthcare Products Regulatory Authority (MHRA) about the campaign. The PMCPA stated that while the scheme was not in breach of the code the briefing material linked the provision of the service to the use and promotion of glargine, which in effect meant that the service was unacceptable.16

In another judgment, the MHRA examined a complaint about the role of the nurses funded by Sanofi to provide supervision and practical advice to primary care staff taking an extended course element of the scheme known as a ‘Statement of extended practice’’ certificate. The MHRA did not uphold the complaint about the nurses’ incentive scheme—which gave financial rewards for results such as the number of practice based diabetes audits, the number of courses arranged, the number of clinics run, and the number of people started on insulin—because the drug company funded nurses were not responsible for prescribing decisions and the scheme was designed to provide an incentive for successful training practices not the prescription of medicines. However, the MHRA also stated that “relating the remuneration given to individual nurses to the number of insulin starts made in the practices they support and for changes in dose was very ill-advised.” At the time of the MHRA’s findings, Sanofi stated that the incentive scheme was no longer running and that it would not be undertaking a similar exercise in the future.17 A spokesperson said: “The concerns raised in 2004 about Insulin for Life were fully addressed at the time. The programme, which teaches clinicians how to start, not switch insulin, is highly regarded by those who have taken part, and has been shown to positively improve the health of their patients.”

Sanofi, of course, is not alone in providing educational and practical assistance to primary care staff dealing with diabetes. Not to be outdone, Novo Nordisk also funds a training programme, known as MERIT (meeting educational requirements, improving treatment) in Diabetes. Discussing the scheme at the Pharmaceutical Marketing Society in November 2008, Jean Woodwar, the head of Novo Nordisk’s education programmes set out the interplay between the educational and commercial objectives of educational schemes such as MERIT.

“Chronic disease management (including diabetes) is currently being moved from secondary to primary care. Unfortunately primary care is expected to take this on but without the upskilling and education being provided to go with it.”

Noting the shift in diabetes management from secondary to primary care and the need for greater skills among primary care staff, she explained that “MERIT was designed to fill this gap. It is a win for the NHS and also for Novo Nordisk, 80% of whose business comes from diabetes. Another objective was to improve care for patients on insulin—by appropriate dose titration or changes in insulin regimen. The third objective was to identify customers in primary care and build relationships with them enabling rep access.”18

A year earlier Watermeadow Medical, a medical communications company, announced it had been nominated for a pharmaceutical marketing effectiveness award for its work on behalf of Novo Nordisk in developing MERIT.19 It has also provided writing assistance for numerous journal papers on insulin determir.19

Cutting choice

The push towards analogues is supported by the corresponding withdrawal of cheaper human insulins by drug firms. This has been a steady feature of the insulin market for some years, and one critics claim is designed to encourage the switch to analogue insulins. At the end of December, Novo Nordisk will withdraw Mixtard 30, a popular and well tolerated biphasic human insulin from the UK. Its 90 000 or so users will be forced to seek alternatives, some no doubt moving to analogues. However, Mixtard 30 will remain on the market in other parts of Europe under the brand name Actrophane 30.

What may lie behind this global switch campaign was set out by Jesper Brandgaard, chief financial officer of Novo Nordisk, in an interview on CNBC in 2006. He said that generic competition provided a threat to the human insulin market.

“We are about to convert the market from human insulin on to insulin analogue,” he said. Adding: “If we look at it globally, we have more than 40% of the market now converting to analogue. So we are actually in a different situation from most other companies. We are taking our portfolio from being generic product, human insulin, on to a patent protected insulin analogue. So we are actually getting our portfolio on-patent, not off-patent.” According to Craig Currie, reader in diabetes pharmacoepidemiology in Cardiff, Novo Nordisk also needs extra capacity to manufacture a newer insulin it is developing called degledec.

But this debate on human versus analogue insulin is just one part of the debate around funding of diabetes care. A 10 year report by Dr Currie and colleagues on diabetic patients using insulin finds that average HbA1c remains at 8.6%. The paper says: “Over the 10 years to 2007, diabetes related primary care adjusted costs increased considerably, whereas glycated haemoglobin values did not improve at all over the same period.”20 In which case, are we really getting value for money out of diabetes care?

How Sanofi-aventis took on the competition

A decade ago, Sanofi’s currently stated ambition to be the number one business in the insulin market is unlikely to have rattled the traditional dominant players—Eli Lilly and Novo Nordisk.

But the development of the recombinant insulin analogue C267H408N72O77S6 changed everything.21 Originally developed by Novo Nordisk, it was abandoned because it caused localised reactions at the injection site.22 But the technology was picked up by Hoechst and then fell into the hands of Aventis (later Sanofi-aventis) after its acquisition of the Germany company. The compound eventually emerged on to the market as Lantus (insulin glargine), first in Germany in 2000.23

Today Lantus is the number one diabetes brand in the world with sales well over €3bn in the past year. More importantly, according to Sanofi, glargine has changed the way diabetes is managed, leading to the “treatment paradigm shift towards basal insulin.”4

UK Health Technology Assessment8

The assessment concluded that there was no difference in HbA1c concentrations between patients using glargine and isophane insulin. And the difference between detemir and isophane insulin was “only a small non-significant” one.

The frequency of severe hypoglycaemia was similar with both analogues and isophane insulin, but, overall, hypoglycaemia was less frequent with both glargine (odds ratio 0.74, 95% confidence interval 0.63 to 0.89) and detemir (0.51, 0.35 to 0.76). Many of the hypoglycaemic episodes were nocturnal, and the odds ratios for those were 0.47 (0.37 to 0.59) for glargine and 0.48 (0.37 to 0.63) for detemir.

The meta-analyses showed that patients using glargine gained slightly less weight than those using isophane insulin (0.28 kg; 95% confidence interval–0.72 to 0.15), but this difference was neither clinically nor statistically significant. The weight difference was slightly greater for detemir (1.2 kg; 90% confidence interval –1.6 to –0.8). However this difference was eliminated when detemir was used twice daily.


Cite this as: BMJ 2010;341:c7139


  • Competing interests: All authors have completed the unified competing interest form at (available on request from the corresponding author) and declare: no support from any organisation for the submitted work; no financial relationships with any organisations that might have an interest in the submitted work in the previous three years; no other relationships or activities that could appear to have influenced the submitted work.

  • Provenance and peer review: Not commissioned; externally peer reviewed.


View Abstract