Why Does Insulin Cost So Much?

At 11 o’clock on the morning of Monday,  January 23, 1922, a young Canadian boy, Leonard Thompson, made history.  He had been diagnosed with diabetes, which at that time was usually a death sentence; only rarely did children like Leonard survive more than a few weeks.  The only treatment was a  starvation diet.   Leonard took part in a bold experiment at the University of Toronto, and was the first person ever treated with the newly discovered pancreatic glucose-lowering extract, we now call insulin.  By the next morning, Leonard’s blood sugar had fallen from 520 to 120.

This seminal event has changed things forever for millions of people.  The scientists who discovered insulin, Frederick Banting, Charles Best, and another scientist, James Bertram Collip, who helped them purify the insulin, shared a patent for the discovery.  In a humanitarian gesture, these men  sold the patent to the University of Toronto for $1, with the intention of making insulin widely available as quickly as possible, and at low cost.  Eli Lilly and Company agreed to try to make the inventors’  dreams a reality.  By late 1923, that goal had been largely achieved, with insulin available on pharmacy shelves around the world.   This was truly an extraordinary accomplishment.  But, in 2016, almost 100 years later, a single bottle of insulin, which costs pennies to manufacture, can cost a patient as much as $300 or more?  How did this happen?

The Evolution of Insulin Therapy

The insulin used to treat Leonard Thompson was extracted from cow pancreases.  For commercialization of insulin production, Eli Lilly and Company obtained insulin from both cow and pig pancreases and called it “Iletin” (by the mid-1920s, the Danish manufacturers, Novo and Nordisk were also producing insulin, but only from pigs).  Iletin was what we would call animal regular insulin (R).  After subcutaneous injection, R had onset of action in about 30 minutes,  and was  gone within 6-8 hours.  Thus, insulin therapy usually consisted of 3-6 daily subcutaneous injections.

By the 1930s, scientists had learned how to lengthen the time course of insulin bioactivity by adding protein (generally protamine) and zinc to the insulin .  By the 1950s, the most widely use insulins were R, NPH, and the Lente insulins.  NPH (N) , which stands for neutral-protamine-Hagedorn, was an intermediate-acting insulin, lasting up to 24-hours, and named after Hans Christian Hagedorn, the scientist who developed it.  The Lente insulins were a family of insulins: Ultralente (UL). Lente (L), and Semi-lente (SL).  These consisted of R  to which varying concentrations of zinc were added, resulting in insulin crystals of different sizes and solubilities .  Mixing UL with SL in a 7:3 proportion, resulted in L  (not to be confused with 70/30 N/R mixtures), which was basically equivalent to N.  In 1978 when I arrived in Columbia, MO, to head up the pediatric diabetes division at the University-Columbia Medical School, those were the insulin options.

Back in the 1970s, the standard insulin regimen for a patient with what we now call type 1 diabetes, was 1 injection of either N or L.  In tough cases, one had to resort to 2 injections of N, L, or perhaps, one injection of N+R or L +R.

The next big step forward in insulin production occurred in 1978, when Genentech discovered how to manufacture human insulin using recombinant DNA techniques.  By 1982, human N and R became available.  So, cows and pigs were no longer needed to get adequate supplies of insulin!  Human insulin rapidly became the standard; it was  cheaper to make, associated with fewer allergic reactions, and had less fat build-up at injection sites than the animal insulins.  It should be noted that in 1981, the Danes succeeded in making human insulin by chemically, by converting pork insulin, which differs from human insulin by only 1 amino acid, into human insulin.

Since the 1990s, synthetic insulin analogues have provided another incremental  step forward in diabetes treatment.  These “designer” insulins are modifications of human insulin that alter its pharmacological properties.  For example, we now have ultra short-acting insulins, such as lispro and aspart.  These insulins have rapid on- and- off time- course of action; they act more like insulin that is normally secreted from the pancreas.  Similarly, we now have insulin analogues, such as glargine, which have an ultra long-acting profile, mimicking the way the pancreas secretes insulin between meals.

The Diabetes Control and Complications Trial (DCCT)

When insulin was discovered in the early 1920s, everyone thought that diabetes had moved from a killer disease to a nuisance disease, requiring only careful eating and insulin injections to maintain good health.  Unfortunately, we learned that insulin therapy had changed diabetes from a rapidly fatal disease, to one with devastating complications and shortened life-span.  All through the 1960s and 1970s there was a debate in the diabetes community about diabetes complications.  The debate was whether the complications were caused by high blood glucose levels or were just part of the genetic consequences of the disease.  Answering the question had been virtually impossible; we did not have the tools to objectively measure long-term glycemic control.  We didn’t even have easy ways for patients to measure their blood glucose levels.  Then with the development of fingerstick blood glucose testing and hemoglobin A1c in the mid-1970s,  it finally became possible to address the glucose control vs. complications question in a sound scientific manner.  Thus was born the DCCT, a long-term study in patients with type 1 diabetes.

The DCCT design was relatively simple: two groups of patients randomly assigned to one of two groups: intensive therapy (IT) or conventional therapy (CT).   IT was defined as treatment with either multiple daily injections (MDI) using pre-meal injections of R, and a bedtime injection of N or R delivered into the subcutaneous space, using an infusion pump.  The treatment goal was to achieve and maintain HbA1c and blood glucose levels as close to normal as possible.  CT was defined as no more than 2 daily injections of insulin, with no specific treatment goals.

Beginning in 1983, the DCCT recruited 1441 people in the U.S. and Canada with type 1 diabetes.  The study ended in 1993, with the exciting news that study volunteers treated with IT developed diabetes complications at a dramatically lower rate than did those treated with CT.  In addition, glycemic control in the IT group was far better than in the CT group: HbA1c levels in the IT group had been maintained at about 7% vs. 9% in the CT group.  Almost immediately, IT became the standard of care for patients with type 1 diabetes, with a goal of achieving the best possible glycemic control.

The Epidemiology Of Diabetes Interventions and Complications (EDIC)

The DCCT did not really end in 1993.  It has continued as a long-term observational study called the EDIC or DCCT-EDIC, which is still going strong.  Of the original 1441 DCCT study volunteers, 95% of those still alive (about 1300 people) have continued to participate.  The original DCCT results have been confirmed and extended.  While the study investigators no longer treat the study volunteers, most of the study volunteers are following IT regimens.  HbA1c levels in the original IT and CT groups have averaged about 8%, up about 1% in the original IT group, and down about 1% in the original CT group.

When insulin analogues became available, studies showed that they offered some advantages over human N and R, both for MDI and insulin pump use: the analogues offered more physiologic insulin delivery than did N and R, easier meal planning, and perhaps, fewer risks for hypoglycemia.  At present, most DCCT-EDIC study volunteers using IT are being treated with insulin analogues, as are most patients in the U.S. with type 1 diabetes who are treated with IT.  The insulin analogues are also being used increasingly in patients with  type 2 diabetes who become insulin-dependent.

So, What About The Insulin Cost Increases?

Back in 1978 when I came to the University of Missouri, things were pretty simple.  Pharmaceutical companies usually sold insulin directly to pharmacies or to pharmacy groups.  I had a good friend who worked at Eli Lilly and Company, and I was good friends with our pharmacy director.   I knew exactly how much it cost to manufacture insulin ($2, including packaging and marketing), how much it cost pharmacies to buy it ($5 per bottle), and how much patients paid for it ($10 per bottle).   A typical patient with type 1 diabetes goes through about  2 bottles of insulin per month.   So, at that price, a 3-month supply of insulin would have cost $60.  Ah, but how times have changed!

Let me give you a real life example of what has happened to some patients with diabetes.  MS  is a 27-year- old woman with type 1 diabetes, which was diagnosed at age 6 years.  She has been a model patient from the get go.  She is being treated with a MDI regimen using lispro and glargine insulins.  She has never had a HbA1c over 7% (note: the American Diabetes Association recommends that most patients with diabetes aim for HbA1c levels 7% or lower).  Until recently, her diabetes care costs had been covered through her parents’ health insurance, but all of a sudden she is on her own.  She is a college graduate with the usual big ticket college loans, but she has a good job.  Unfortunately, she has no health insurance yet.  She goes to see her doctor and gets prescriptions for her various diabetes care supplies, including insulin.  She then goes to her local pharmacy with prescriptions in hand and is informed that a 3-month supply of her insulin (3 bottles of lispro and 3 bottles of glargine) would cost her $1740.  She does not have $1740.  She walks out without filling the prescription and starts calling around, only to find that the prescription costs are about the same everywhere in town.  Oh yes, I forgot to mention that we haven’t even gotten to the rest of her diabetes care supplies- syringes, test strips, a glucagon injection kit (used to treat a serious hypoglycemic episode), glucose tablets, etc.  The cost for the glucagon injection kit alone would be $261.  In tears, she calls her doctor, who has heard this story over and over, and suggests she figure out how to get health insurance asap, but also tells her that if she cannot afford the glargine and lispro, there is a way to get her insulin at much lower cost, if she is willing to switch her insulin regimen to good old human N and R (e.g., R pre-meal and N at bedtime).   N and R are available at some pharmacies, including Walmart, for about $25 a bottle.  So, MS makes the change to N and R and has it down pretty well.  She now needs snacks between meals, and is likely to have more hypoglycemia that she did with the lispro and glargine, but at least she can afford her insulin.

So, would things be different if MS had health insurance?  The answer is “yes,” but with a catch.  Even with “good” insurance, insulin co-pays are up in the sky; depending on the fine print in the insurance policy, the insulin co-pay could be $500 or more for a 3-month supply.  And if she had a policy with a high deductible, she might have to spend $5000-$10,000 before the insurance coverage kicks in.  If she was unemployed, disabled, or had a very low income, she might qualify for Medicaid, and get her lispro and glargine at little or no cost.  Of course, someone would have to pay  for the insulin, in this case, the taxpayers.  If she has a decent income and is lucky enough to live in one of those states that accepted the federal government offer of Medicaid expansion in 2010 (about half the states in the U.S.), she might qualify for Medicaid.  Finally, if she were age 65 years or older, she would qualify for Medicare, and have most of her insulin and other diabetes care costs covered.

But Why Does Insulin Cost So Much?

In our semi-free market economy, competition can keep prices down.  Typically, when a generic version of a brand-name drug gets to the marketplace, the price of the drug drops by about 90%.  It is much the same thing as in other industries.  Remember when a wallet card calculator cost about $50, and computers with 64 K of memory cost several thousand dollars?  Competition generally brings prices down and down.

That has not happened for insulin; rather, insulin costs have gone up and up.  Why?  First, even though there are 3 companies that sell insulin in the U.S., there are no generic insulins.   Second,  pharmaceutical companies know that for a one-of-a-kind drug, the sky seems to be the limit, and pharmaceutical companies basically ask caregivers and patients the following: “How much is this drug worth to you?”  Even if  there is  potential competition among pharmaceutical manufacturers for comparable medications, companies know that it is better to compete in ways other than by lowering price.  In my opinion that is much of the problem today with the cost  of insulin.

The high cost of insulin does not seem to be much different from what has been happening in general in the pharmaceutical industry: steadily rising costs to patients for their medications.  Recently, some of the most striking examples have received considerable attention by the media and the federal government: take the case of Gilead Pharmaceuticals and the high cost of Solvaldi (sofosbuvir), a new drug for treating hepatitis C.  Or take the case of Valeant Pharmaceuticals International, Inc., and their pioneering business model of acquiring pharmaceutical companies or their drugs and drastically increasing prices; not a week goes by without another report in the NYT about these and other pharmaceutical companies.  For example, only a few days ago (April 28, 2016) there was an article in the NYT entitled: “Valeant made pricing mistakes, chief says,” written by Katie  Thomas.  The article summarized “a tense hearing on Capitol Hill,” in which the company chief executive J. Michael Pearson and others were accused of “favoring profits over patients’ needs.”

Why Do Medications Cost So Much More In The U.S. Than In Other Countries?

I am not an economist, but I know some card-carrying economists.  They tell me that one  important contributing reason for the very high drug costs in the U.S., is that, unlike most other countries around the world, the U.S. does not have a health care budget.  We generally  spend whatever health care costs.  Of course, we do in effect ration care, in that some people just cannot afford this or that treatment.  I have been told by a number of pharmacists, that about 20% of people who come in to fill prescriptions, do not fill them when they learn what the price tag is.  Shame on us.

If a country has a health care budget, it can weigh the benefits vs. costs for this or that treatment.  For example, if glargine insulin were offered to the British National Health Service (NHS) at a price equivalent to the cost in the U.S. (about $300 per bottle), the NHS could (and does) say, “no thanks,”  Then the manufacturer and/or distributor has to make a decision: negotiate or don’t sell glargine insulin to the British NHS.  Of course, the drug supplier is likely to negotiate, knowing that it costs next to nothing to manufacturer the glargine insulin.  A small profit is better than no profit.

But wait.  Next comes the argument that drugs cost so much because the drug development costs are so large.  No question.  But, with these incredible mark-ups between production costs and selling price, the manufacturers cover their costs in no time flat.  In my opinion, no one is trying to keep pharmaceutical companies from being profitable,  just from being totally unreasonable.

So, How Do We Bring Insulin Costs Down To A Reasonable Level?

Now we come to the sticky part.  How do we fix things?  This is very complicated stuff, and as I mentioned earlier, high costs affect almost everything in U.S. health care, not just insulin.  The U.S. spends about 3 trillion dollars annually for health care, far, far more than any other country on earth (that is about as much as all the oil reserves in Saudi Arabia are valued).  And what do we get for it?  Unfortunately, U.S. health care outcomes lag far behind those in many other countries.  Who is to blame?  In my opinion, we are all to blame: consumers, politicians, physicians and other health care professionals, hospitals, pharmaceutical companies, medical device manufacturers, and so forth.  There is plenty of blame to go around.  One thing is certain: the situation is deplorable.

I will just throw out a few suggestions.  First, all the key players need to sit down and discuss in an open and honest way, what the problems are and what can be done to fix them.  This is beginning to happen.  For example, there is a federal advisory panel that is calling for the U.S. Congress to force private insurers to rein in rapid increases in prescription drug costs.  There is also a U.S. Senate Special Committee on Aging that has been having hearings about drug pricing.  There are also some health professionals, such as Irl Hirsch, a diabetes physician at the University of Washington, who have finally had enough of the lunacy, and are “stirring the pot.”  It’s about time.

What About An Easy Fix?

Clearly there are no easy fixes, but I have one specific idea to consider.  The U.S. health care system cannot really be considered a single system; rather, it is a series of systems- private insurance, Medicaid, Medicare, the Veterans Administration, the Indian Health Service, and so forth.  Why don’t we  start with Medicaid and Medicare, and let them set prices for medications used by their enrollees?  They are already doing this to some to a great extent for services,  but generally not for medications.  The idea is that these government programs would determine what are reasonable charges for all medications and care supplies that they consider “medically necessary.”  So, the government probably won’t put  botox for wrinkles, tummy tucks, and such in the medically necessary category.  In the end, this pricing plan would need to apply to everyone- pharmaceutical companies, drug wholesalers, insurers, etc.  It ‘s just an idea.  I will let the economists and others figure out what to do.  But, for certain something needs to be done.  Patients are suffering every day we wait.  If  Banting, Best, and Collip knew what has happen to insulin costs since their ground-breaking discovery almost 100 years ago, they would be rolling over in their graves.

Want To Learn More?

You might want to check out the following:

  1. Diabetes Forecast March-April 2016 issue (www.diabetesforecast.org).  There is an excellent article entitled: “The rising cost of insulin,” written by Allison Tsai.
  2. Dr. Irl Hirsch American Diabetes Association (ADA) talk at the Scientific Sessions in Boston, June 2015 and entitled: “Changing cost of insulin therapy in the U.S..”  He also did an update at the ADA Postgraduate course in San Francisco, March 2016.  Dr. Hirsch’s meeting webcasts and other talks that may interest you can be found at http://professional.diabetes.org.
  3. Greene JA, Riggs KR: Why is there no generic insulin?  Historical origins of a modern problem.  N Engl J Med 2105;372:1171-75 (http://www.nejm.org).  A summary of this article can be found at http://www.medscape.com/viewarticle/841669.
  4. Grant RL: Certification of insulin.  This article can be found at http://www.ncbi.nim.nih.gov/pmc/articles/PMC2031238/pdf/pubhealthreporig00150-0086.pdf.
  5. Rutty CJ: “Couldn’t live without it” : Diabetes, the costs of innovation and th price of insulin in Canada, 1922-1984.  This article can be found at http://www.healthheritageresearch.com/Rutty-InsulinPrice-CBMH25-2008-407.pdf.



Leave a Reply

Your email address will not be published. Required fields are marked *