In 1997, a group of experts convened by the American Diabetes Association changed the definition of type 2 diabetes, lowering the blood sugar threshold, and instantly as many as 1.9 million more Americans had the condition.
The same pattern played out in 2003, in an even bigger way, when the association changed the definition of a condition known as pre-diabetes and -- overnight -- 25 million more Americans were affected.
In the decade that followed, the diabetes industry boomed -- thanks in part to a 2008 declaration by two endocrinology groups that pre-diabetes could be treated with drugs if diet and exercise didn't lower blood sugar.
Last year, sales of diabetes drugs reached $23 billion, according to the data from IMS Health, a drug market research firm. That was more than the combined revenue of the National Football League, Major League Baseball, and the National Basketball Association.
But from 2004 to 2013, none of the 30 new diabetes drugs that came on the market were proven to improve key outcomes, such as reducing heart attacks or strokes, blindness, or other complications of the disease, an investigation by MedPage Today and the Milwaukee Journal Sentinel found.
The U.S. Food and Drug Administration approved all of those drugs based on a surrogate endpoint: the ability to lower blood sugar. Many of the new drugs have dubious benefit; some can be harmful.
"We have an entire industry -- a diabetes economy -- that revolves around glycemic control," said David H. Newman, MD, director of clinical research at Icahn School of Medicine at Mount Sinai Hospital in New York.
But the disease is not well understood, is difficult to control, and can't be cured, said Newman, an emergency room physician who founded theNNT, an independent group of physicians who analyze the effectiveness of medical therapies.
"We've called a sign of the disease the disease, but there are no rigorous studies that prove we understand how to treat the illness rather than its symptoms," Newman said.
The Diabetes Epidemic
To be sure, type 2 diabetes -- once known as adult-onset diabetes -- is a serious problem in America, often requiring drugs. The condition can lead to cardiovascular disease, kidney failure, amputations, and more immediate symptoms, especially in those with very poor glycemic control.
The surging number of cases has closely paralleled increasing rates of obesity, sedentary living, and an aging population.
But the number of people with diabetes or pre-diabetes and who are candidates for drugs has been magnified by organizations and doctors with financial ties to drug companies, the MedPage Today/Milwaukee Journal Sentinel investigation found.
The Money Trail
The groups that made the changes in diagnostic thresholds each receive several million dollars a year from companies that make diabetes drugs. And each of the panels that moved to increase the numbers, or said more patients could be put on the drugs, had majorities of members who received money from such companies.
In one case, 13 of 19 members of a committee received more than $2 million in speaking and consulting fees since 2009 from companies that make diabetes drugs, according to a MedPage Today/Milwaukee Journal Sentinel analysis of drug company data.
Many of the new drugs approved by the FDA can cause serious side effects, including heart problems, cancers, and overdoses leading to an estimated 100,000 emergency room visits each year by people with dangerously low blood sugar, according to published research, interviews, and other data reviewed for this story.
Separately, a MedPage Today/Milwaukee Journal Sentinel analysis found about 3,300 cases in which diabetes drugs approved since 2004 were the "primary suspect" in a patient's death, according to adverse event reports to the U.S. Food and Drug Administration. In another 20,000 cases, the drugs were believed to be responsible for hospitalizations.
"We may be using drugs that are great for lowering glucose, but don't do anything for cardiovascular risk," said Clifford Rosen, MD, an endocrinologist and professor of medicine at Tufts University School of Medicine.
It also is unproven whether many of the newer drugs are preventing other diabetes complications such as blindness and amputations, said Rosen, who also is an associate editor of the New England Journal of Medicine and has served on FDA advisory committees.
"Why haven't we done that?" he asked. "There is a lot of pressure from companies to go with the simplest and easiest approach."
In recent years, the FDA has increasingly relied on "surrogate" measures when approving new drugs. In heart disease, that can mean relying on better numbers on a cholesterol test rather than reductions in actual heart attacks.
In cancer, it can mean relying on the shrinkage of tumors, rather than the drug actually increasing survival.
An earlier MedPage Today/Milwaukee Journal Sentinel investigation found that 74% of 54 new cancer drugs allowed on the market between 2004 and 2013 were approved based on surrogate measures.
Surrogate endpoints are attractive to both industry and regulators because they provide a faster, less expensive pathway to marketing approval than clinical trials that ask if a drug decreases the number of heart attacks, kidney failure, or death. The reason is simple: clinical trials that rely on those hard endpoints take years longer and require many thousands of patients.
For example, IMPROVE-IT a recent study investigating two cholesterol-lowering drugs using heart attack, stroke, and death as endpoints enrolled 18,000 patients and took 9 years from conception to conclusion.
Friend or Foe?
In 2008, based on concerns that diabetes drugs could lead to cardiovascular problems, the FDA started requiring companies whose drugs were approved on surrogate measures to do follow-up studies. But those studies are not required to show they actually improve heart health.
Rather, companies are required to show they do not increase the risk of cardiovascular problems by more than 1.3 times. Most of those follow-up studies are still going on, but the one that has been completed found a serious problem.
The study of the drug saxagliptin (Onglyza), approved in 2009, was designed to show saxagliptin was better at reducing heart attacks, strokes, and cardiovascular deaths than a placebo, but it failed to do so. A 2013 paper in the New England Journal of Medicine showed the drug actually increased the rate of hospitalization for heart failure by 27%.
The drug remains on the market, though the FDA says it is investigating the matter.
In response to questions about its approval of diabetes drugs, the FDA pointed to two landmark 1990s clinical trials that showed lowering glucose over a 10-year period reduced the so-called microvascular complications of diabetes such as eye, nerve, and kidney damage.
But one of those studies from 1993 was done in people with type 1 diabetes, which can't be prevented, and typically is an autoimmune disorder that occurs in youth. Type 2 diabetes is a different disease, usually occurring in adulthood, that can be prevented or delayed with a healthy lifestyle.
The participants in the study also had extremely high glucose levels and were put on intensive therapy using insulin pumps or three or more insulin injections a day, a much more critical circumstance than for the vast majority of people diagnosed with type 2 diabetes today.
Even the authors of that study cautioned against using its results in making treatment decisions, other than diet, for people with type 2 diabetes.
The other study cited by the FDA, a 1998 study in the U.K., found significant reductions in both large and small blood vessel problems such as heart or eye damage -- but more recent large clinical trials have not replicated those findings.
One of them, a 2008 trial, even found a 22% higher rate of death from any cause and a 35% higher rate of death from cardiovascular causes among people with type 2 diabetes whose blood sugar was brought under intensive control.
FDA spokesman Jeff Ventura acknowledged a lack of proof that treating high glucose reduces macrovascular complications. In an email, he noted: "Requiring that all manufacturers establish CV (cardiovascular) benefit as a condition for approval of glucose-lowering medications would set an unreasonably high hurdle for the development of new therapies for a condition affecting over 26 million Americans."
But when it comes to causes of death, cardiovascular disease is at the top of the list for those with type 2 diabetes. It ultimately kills up to 80% of those with the condition.
In contrast, there is a lower lifetime risk of blindness or end-stage kidney disease in people who develop type 2 diabetes in midlife, when it's most often diagnosed.
For instance, in someone age 45 whose blood sugar is relatively high, the absolute lifetime risk of blindness is 2.6% and the risk of end-stage kidney disease is 3.5%, according to a 1997 paper in the Annals of Internal Medicine. It's mainly for these risks that physicians treat pre-diabetes and type 2 diabetes with drugs to lower blood sugar.
Mantra
For years, a mantra in the diabetes community has been a concept known as tight or intensive glycemic control. It means using drugs to drive blood sugar levels down to more acceptable levels.
But recent research has shown that attempting to tightly control glucose can increase the risk of hypoglycemia and its complications -- seizure, unconsciousness, or death.
Often, it is older people who are the most susceptible to hypoglycemia.
A 2014 study in JAMA Internal Medicine of people with an average age of 77 showed that 404,000 patients were admitted to the hospital because of hypoglycemia between 1999 and 2011, compared with 280,000 for hyperglycemia. Five percent, or 20,000 people, died within 30 days of their hypoglycemia admission. The study was based on Medicare data.
Newman, the emergency physician at Mount Sinai Hospital, said he's seen more and more cases of hypoglycemia.
"When people listen to doctors talk about diabetes they assume scientists understand it, but it's a veneer of science," he said. "Treating a sign or a symptom rather than the disease is not the science patients believe their treatments are based on."
Doctors say the convergence of new drugs and the expanded definitions for diabetes and pre-diabetes has led to an over-medicalization of blood sugar disorders.
The 2003 change by the ADA in the pre-diabetes definition means that today an additional 46 million adults, at least, have the condition, according to an analysis done for this story by Lisa Schwartz, MD, MS, and Steven Woloshin, MD, MS, researchers with the Dartmouth Institute for Health Policy and Clinical Practice.
In 2010, the association further expanded the definition of pre-diabetes by saying that a blood sugar test known as hemoglobin A1C could be used to diagnose it and set the level for that measure. The test is easier because it does not require fasting.
Of the 14 outside experts on that panel, nine worked as speakers, consultants, and advisers to companies that make diabetes drugs, a MedPage Today/Milwaukee Journal Sentinel analysis found.
Between 2009 and 2013, committee members as a group received more than $600,000 from the drug companies.
The association would not provide the financial disclosures of its 2003 panel members nor a list of drug companies that it received funding from at the time.
However, a report by the Center for Science in the Public Interest found the association received more than $6 million from diabetes drug companies in 2002. And one document from the ADA itself shows it received more than $7 million from such companies in 2004 alone.
In an emailed statement, Robert Ratner, MD, chief medical and scientific officer of the association, said its committees relied on the most current epidemiological data to establish cutoffs for diabetes.
"No consideration of therapy entered into the discussion, as lifestyle modification has always been the first line intervention for type 2 diabetes," he said.
A similar scenario played out in 2008, when a joint panel of two other groups -- the AACE and American College of Endocrinology -- said diabetes drugs can be used to treat pre-diabetes with "careful judgment."
Of the 17 members on that panel, 13 worked as speakers and consultants to companies that make diabetes drugs.
In 2013, the same groups went further and said drugs could be used to treat pre-diabetes if lifestyle changes don't work.
That year, the association received more than $8 million from companies that market or develop diabetes drugs and related products, according to its tax form.
Thirteen of the 19 doctors on the panel that made the recommendation, including its chairman, worked as consultants, speakers, or advisers to companies that make diabetes drugs.
As a group, those experts got $2.1 million from the companies since 2009, according to a MedPage Today/Milwaukee Journal-Sentinel analysis of drug company data.
Alan Garber, MD, co-chairman of 2008 panel, chairman of the 2013 panel and past president of the association, said the integrity of the panel members prevented them from being influenced by financial concerns.
All of the financial relationships were disclosed, he added.
Garber said that treating pre-diabetes with diabetes drugs can reduce the risk that it will develop into diabetes.
"We need to do something different," he said. "Everything we've done so far has done nothing to deter the epidemic of obesity and type 2 diabetes in the U.S."
But Garber acknowledged and other doctors interviewed for this story said there is no proof that treating pre-diabetes with drugs reduces important health outcomes such as heart attacks or amputations.
What's more, none of the drugs it recommended were approved to treat pre-diabetes. Using them that way amounts to off-label use.
"The logic of treating pre-diabetes with drugs is basically nuts," said James McCormack, PharmD, a professor of pharmaceutical sciences at the University of British Columbia, who has studied diabetes drug research."We treat 100% of people with pre-diabetes with a drug to prevent roughly 15% of people from getting diabetes -- a condition for which they may need a drug."
Diet and exercise are proven to be much more effective at preventing pre-diabetes from becoming diabetes.
A major study published in 2002 found that the combination of diet and exercise reduced the odds of pre-diabetes becoming diabetes by 58% compared with 31% among those using the common diabetes drug metformin.
Viewed another way, only 5% of those in the diet and exercise group went on to develop diabetes each year compared with 7.8% of those taking metformin and 11% of those who got a placebo.
"There is a question whether that category deserves a name and deserves attention," said Victor Montori, MD, a diabetes expert with the Mayo Clinic. "What's the purpose of it? These people are too low risk to deserve clinical attention."
Today, 86 million Americans -- 37% of the adult population -- meet HbA1C criteria for pre-diabetes. More than 26 million Americans have type 2 diabetes. Combined, that is more than 45% of adults.
Me Too, Me Eleven
As with the cancer drugs, the FDA has approved some diabetes drugs despite concern about serious side effects -- and even over objections from its own advisers.
Consider the drug liraglutide (Victoza), an injectable drug approved in 2010 based on studies showing it lowered blood sugar levels.
Three FDA doctors recommended against its approval, citing concerns about an increased risk of thyroid cancer and questions about cardiovascular safety, agency records show. In addition, two cardiologists on an FDA advisory committee voted against approving the drug.
FDA doctor Karen Murry Mahoney, MD, noted there was no pressing need to approve a drug with safety concerns because there already were 11 different classes of diabetes drugs on the market.
"The need for new therapies for type 2 diabetes is not so urgent that one must tolerate a significant degree of uncertainty regarding serious risk concerns," she said in her report.
Yet the FDA approved liraglutide -- and went on to approve three more drugs in the same class, all of which carried serious risks. The four drugs, known as GLP-1 receptor agonists, all carry the agency's most stringent black box warning because the drugs cause thyroid tumors in rodents.
Liraglutide and some other diabetes drugs also have been linked to serious inflammation of the pancreas and pancreatic cancer.
FDA spokesman Ventura said the opinions of the FDA doctors who opposed approving Victoza don't capture the complex scientific issues that were considered by the FDA's advisory committee, which voted to approve the drug.
He said practitioners, professional societies, and patient advocacy groups all want options for treating diabetes.
In the case of liraglutide, though, the independent patient advocacy group Public Citizen petitioned the FDA in 2012 to remove the drug from the market.
"This is not cancer where there are no other choices for treatment option," said Michael Carome, MD,, director of the group's health research group. "There's no evidence of benefit and the FDA too often relies on surrogate markers and ignores significant risks."
Earlier this year, the FDA denied the petition.
Novo Nordisk, the maker of liraglutide, is now seeking approval for a higher-dose of the injectable drug to be used in weight loss, including in those with pre-diabetes.
In an email, Michael Bachner, a spokesman for Novo Nordisk, said liraglutide is the leading drug in its class in the world, and it has a proven track record of helping U.S. patients with type 2 diabetes control their blood sugar levels.
John Fauber is a reporter with the Milwaukee Journal Sentinel. Coulter Jones and Elbert Chu are reporters with MedPage Today. This story was reported as a joint project of the Milwaukee Journal Sentinel and MedPage Today.