Diabetes, Inc.

Today, there is a growing campaign in the United States waged by people demanding that manufacturers lower the price of analogue insulins, which began to enter the market in the mid-1990s. These people say that analogues are preferable to Humulin, a product of recombinant DNA technology, because analogues provide better control and support a better quality of life. But most independent studies comparing recombinant human insulins (RHI) and analogues say that differences in the safety and effectiveness profiles between the two are so minimal that the higher price of analogues cannot be defended. Not surprisingly, manufacturers have long claimed that the higher cost of analogues is justified because they are far superior, provide greater flexibility, and have action profiles that more closely mimic the way that non-diabetics produce insulin, assertions that are contested. 

 The clash between patients and their families, on the one hand, and the science, on the other, is distressing. But it’s not the first time there’s been division in the world of insulin. Beginning in the 1980s, patients who had successfully used animal insulin were transferred to more expensive recombinant human insulin – the first “switch campaign” – and found themselves thrust suddenly and unsuspectingly into the business plans for higher profits being engineered by Eli Lilly and Novo Nordisk. Many who were seriously harmed by Humulin and/or Novolin were challenged by those who relied on weak scientific evidence, which either claimed RHI was better than animal insulin or that there was no difference between the two. 

Image: “Patients, doctors at odds over medications.” The Baltimore Sun, 11 August 2006.

Recombinant human insulins – referred to variously as synthetic, biosynthetic, genetically engineered, genetically modified and (erroneously) human insulin – were first introduced in countries around the world in the early 1980s, by Eli Lilly (Humulin) and Novo Nordisk (Novolin), beginning in the United Kingdom. In May 1982 , Lilly submitted its application for approval of Humulin to the US Food & Drugs Administration, a process that usually took between 28 and 34 months.[1] But Humulin N and R were approved with lightning speed five months after the FDA had received the company’s application.[2] The same pattern was seen in Canada, West Germany and the Netherlands over the next several months. According to Henry Miller, the Medical Review Officer who led the FDA’s review of Lilly’s application, the “quality of the [company’s] submission…was unsurpassed and evidence of safety and efficacy was unequivocal and copious.”[3]  

What exactly was this unsurpassed evidence to support Eli Lilly’s and Novo Nordisk’s claims that recombinant human insulin was safe and effective and a major advance over natural, animal-sourced insulins that had been used since 1922? 

Clinical trials of Eli Lilly’s Humulin insulin were initiated in December 1980[4], and two years later it was available in drug stores in a number of countries. Although clinical testing of new molecular entities in humans at the time averaged 40 to 60 months in the United States[5], Eli Lilly spent 18 months testing the first product of biotechnology ever to be introduced. When it entered the market, it not only was poorly tested, but it also failed to meet the high standards of evidence that are important to patients and their families. 

According to the Cochrane Collaboration, which conducts reviews of scientific and clinical evidence, the studies of RHI were generally of “poor methodological quality.” Both Eli Lilly and Novo Nordisk had failed to investigate essential endpoints such as mortality, morbidity and health-related quality of life issues– the things that matter most to patients. The companies were unable to show any therapeutic or clinical advantage of RHI compared with animal insulins; and only 40% of the studies reviewed provided information about adverse side effects. And while Miller claimed the FDA approval was an “epochal event”[6], Cochrane concluded that the introduction of recombinant human insulin should serve as a warning of “pharmaceutical and technological innovations that are not backed up by sufficient proof of their advantages and safety.”[7]

So, the clinical evidence – 70% of which was funded by manufacturers – to support the use of RHI was nothing to get excited about, especially since the best that could be said was that it was no better than animal insulin but cost up to three times more. Most independent reports suggested there was no difference between recombinant human and animal insulins in regard to the severity, incidence, or warning signs of hypoglycaemia. But no long-term, large-scale studies were undertaken, in spite of the fact that small trials of 50 people or less would likely not detect any problems. 

But, there’s more to evidence than clinical trials. By 1989, only six years after they were introduced, the number of reports of harm linked to Humulin and Novolin began to mount around the world. There was a virtual war of words between those who asserted they had experienced serious harm when using RHI and the manufacturers who claimed it was safer because it was chemically and structurally identical to “real” human insulin made by people without diabetes.[8] What was remarkable was the consistency, from one country to the next, in the reports of adverse side effects, including unexpected and severe hypoglycaemia, the loss of vital warning signals of hypoglycaemia, memory loss, insulin shock (comas), blackouts, arthritis/arthralgia/myalgia symptoms, extreme weight loss or gain and dead-in-bed syndrome.[9]

The response of manufacturers, who were facing lawsuits from harmed patients in three countries between 1992 and 2002, can only be described as ruthless. Company representatives launched a propaganda campaign to protect their products (and themselves) which included funding studies that seemed specifically designed to refute the reports of harm. But by 1995, Humulin was listed among the top 10 drugs linked to serious harm by the FDA. Nonetheless, doctors, regulators and manufacturers told people who relied on insulin that RHI was safer and more effective than animal sourced insulins. The experiences of those who used the new insulins were dismissed in a callous and brutal fashion. 

Image: “Diabetics not told of insulin risk.” The Guardian, 9 March 1999.

There continues to be a complete absence of post-market surveillance regarding insulin products – that is, studies of how people are doing in the real world and outside the confines of a controlled clinical trial.[10] The actual experience that people have with medicines, including insulin, is often dismissed as “anecdotal” by the medical establishment, manufacturers and regulators. Thus, reports of serious harm or lack of efficacy often are not given credence. But every decision to withdraw a drug begins with patient reports of serious harm. Regulators and manufacturers are supposed to – but usually don’t – monitor these experiences closely in order to ascertain whether reports of harm are accumulating and whether they are consistently similar. Patients act as sentinels in the real world of medicine; that is, when they report harm, they are sending warning signals to regulators, manufacturers and the public. 

So, what is the truth? In most studies of insulin, including those that compare RHI to analogues and regardless of their conclusions, investigators look at two endpoints and only two endpoints: HbA1c levels and the incidence of severe hypoglycaemia, including during the night. Many studies have found that, while nocturnal hypoglycaemia occurs slightly less frequently in those using analogues, “the overall difference in hypoglycaemia is small, and the difference in HbA1c level almost undetectable.” 

Most of the studies reviewed by the Cochrane Collaboration used the same endpoints: HbA1c and hypoglycaemia. Those who reported that their early warning signals of low blood sugar had disappeared when they were switched to Humulin or Novolin could find validation in very few studies– because hypoglycaemia unawareness, now a distressingly common problem, was not part of the investigation prior to approval. After approval, a number of independent investigators who verified the increased incidence of hypoglycaemia unawareness struggled to have their concerns heard. So, the “evidence” was on the side of the manufacturers, who had funded and sponsored the overwhelming majority of studies.

Cochrane warned in 2002 that the introduction of recombinant human insulin was approved by regulators around the world in spite of a complete absence of information about whether these much more expensive brands improved the quality of life, the incidence of diabetes-related complications and longevity. Nothing has changed. 

One problem is that most of the evidence relies heavily on a surrogate marker- HbA1c. A surrogate market is defined as an “endpoint that is a laboratory measurement or physical sign that is used in therapeutic trials as a substitute for a clinically meaningful end point that is a direct measure of how a patient feels, functions, or survives and that is expected to predict the effect of the therapy.” In clinical trials comparing analogue and human insulin, the difference in HbA1c – the laboratory measure – is minimal, and these studies are being utilized to support an argument that people should use “Walmart insulin”. 

Yet many patients report meaningful and relevant differences in the real world outside of the trial. Evidence is important, and should never be dismissed. But for those using insulin, evidence must be used to support informed choice and include how we experience the essential and basic tool that enables us to live healthy lives.

Want to support the authors of Diabetes, Inc. and help bring this important history into the world? Make a Patreon contribution.

[1] Spivey, R. N., Lasagna, L. & Trimble, A. G. (1985). New drug applications: How long to gain approval? Clin. Pharmac. Ther.,37,361-366.

[2] Junod, Suzanne White. “Celebrating a Milestone: FDA's Approval of First Genetically Engineered Product.” U.S. Food and Drug Administration (FDA), September–October 2007. Available at http://tinyurl.com/ybdj3sn. 

[3] Thayer, A. 2005. “Insulin,” in Top Pharmaceuticals: a look at drugs that changed the world. Chemical & Engineering News 83(25):3.

[4] News Release, Eli Lilly & Co. “Clinical evaluation of biosynthetic human insulin in diabetic patients.” PR Newswire, 12 February 1981. The news release indicated that human trials began in Witchita, Kansas in December 1980. Dr Fred Whitehouse, chief of metabolic diseases division at Ford Hospital, Detroit, was one of the investigators and reported in JAMA in 1982 that trials had been conducted over a two-year period. See “Future of Human Insulin,” Fred W. Whitehouse, MD in JAMA (Letters). JAMA 1983 August 26; 250(8): 1027-1028

[5] Miller, HI. Designer genes for producing drugs: Will they wash? (Editorial), JAMA 1982; 248:1503 

[6] U.S. Food and Drug Administration: NDA 18,780. Medical Officers’ Review, 20 July 1982. 

[7] Richter B, Neises G. 'Human' insulin versus animal insulin in people with diabetes mellitus (Cochrane Review). In: The Cochrane Library, Issue 3, 2002. Oxford: Update Software.

[8] See, for example, letter to Helga Kellner from Mark Fleming, Manager, Boehringer Mannheim/Lilly Canada Diabetes Care, July 8, 1996. Ms Kellner was writing about the harm experienced by her son when he was switched to Humulin. 

[9] Three main organizations, none funded by the pharmaceutical industry, were involved in documenting adverse reactions to rDNA human insulin: the Insulin Dependent Diabetes Trust (UK) - https://www.iddt.org/here-to-help/iddt-international; the Diabetics International Foundation (USA) - http://members.tripod.com/diabetics_world/;  and the Society for Diabetic Rights (Canada). 

[10] Hogg C, (1999). Patients, Power and Politics: From Patients to Citizens, London: Sage Publications.