The most common chronic disease after obesity, afflicting 84 million Americans and more than 1 billion people worldwide, was born as a public relations catchphrase. In 2001, the PR chief of the American Diabetes Association (ADA) approached Richard Kahn, then the group's chief scientific and medical officer, for help with a vexing problem, Kahn recalls. ADA needed a pitch to persuade complacent doctors and the public to take seriously a slight elevation in blood glucose, which might signal a heightened risk of type 2 diabetes. Raising the alarm wasn't easy, given the condition's abstruse name, impaired glucose tolerance, and lack of symptoms.
Kahn invited half a dozen diabetes thought leaders to brainstorm at a National Institutes of Health cafeteria in Bethesda, Maryland. Surrounded by hungry federal employees, many enjoying the kinds of fatty foods and sugary drinks tied to the diabetes epidemic, they landed on a then–little-used term that seemed sure to scare patients and doctors into action: prediabetes.
"We went back to the ADA office right after lunch and started the change. Within a relatively short period of time we … eliminated ‘impaired fasting glucose’ and ‘impaired glucose tolerance’ and substituted ‘prediabetes’ in all of our literature," Kahn says. Soon, the term was enshrined in the Arlington, Virginia, group's standards of care—widely regarded as the bible of diabetes. ADA and the Centers for Disease Control and Prevention (CDC) in Atlanta declared war against prediabetes, with CDC diabetes prevention chief Ann Albright, an ADA board member from 2005 to 2009, leading the charge. The two groups labeled prediabetes a first step on the road to diabetes, which can lead to amputations, blindness, and heart attacks.
In medicine, prevention is usually an unalloyed good. But in this case, other diabetes specialists argue, medical and epidemiological data give weak support, at most, for increasingly dire prediabetes admonitions. "Nobody really thought at the time, how ‘pre’ is prediabetes for all these people?" says Kahn, who left ADA in 2009 and is now at the University of North Carolina in Chapel Hill. The World Health Organization (WHO) in Geneva, Switzerland, and other medical authorities have rejected prediabetes as a diagnostic category because they are not convinced that it routinely leads to diabetes or that existing treatments do much good. John Yudkin, a diabetes researcher and emeritus professor of medicine at University College London, describes the ominous warnings about prediabetes from ADA and CDC as "scaremongering."
Yet ADA, a nonprofit that funds research, issues treatment standards, and raises public awareness, has gradually broadened its definition of prediabetes to encompass more people. "The public needs to know that right now, in the United States … one out of three may have some aspect of glucose abnormality," says William Cefalu, ADA's current chief scientific and medical officer. "A great percentage … particularly in select ethnic groups, may have an increased chance, or a higher rate of progressing [to diabetes]."
CDC has followed ADA's lead, because "they set the primary standards of care in the U.S.," Albright wrote in a statement to Science. (Albright declined interview requests, and CDC would not permit Edward Gregg, Albright's top epidemiologist, to comment for this story.) In the past, Albright and CDC have said repeatedly that 15% to 30% of untreated prediabetes patients progress to diabetes within 5 years—a claim that hospitals, professional organizations, and local and state health departments have embraced and publicized. She backed away from that number in response to a question from Science, saying, "We no longer use that statement to characterize risk." Indeed, CDC's own data show progression from prediabetes to diabetes at less than 2% per year, or less than 10% in 5 years. (Other studies show even slower rates.)
The push to diagnose and treat prediabetes has come at a cost. When told they have the condition, many people face psychological and financial burdens trying to address it. ADA, CDC, and other groups have spent billions of dollars on research, education, and health improvement programs—generally focused on weight loss and exercise—that have generated lackluster results, according to critics. Kahn makes the point with rhetorical bluntness: Spending vast sums of public money on such prevention programs "has nearly the same effect as burning it in a fire … overall, [it's] a terrible waste of money."
To lower blood sugar, ADA has increasingly advocated more aggressive measures, such as prescription drugs—a push that has opened it to charges of conflicts of interest. Science found that the group and its experts who promote aggressive treatment of prediabetes accept large amounts of funding from diabetes drugmakers. So far, no drugs have been approved specifically for prediabetes, meaning that doctors are limited to prescribing diabetes drugs or other medications "off label" to treat the condition. But drug companies are testing dozens of drugs aimed at prediabetes in hopes of tapping a potential worldwide market of hundreds of millions of people.
Given the avalanche of questionable spending and the wave of anxiety it has unleashed, Kahn now says he rues the day he helped promote the term prediabetes, calling it "a big mistake."
ADA's current definition of prediabetes was born in 2009, after the group, along with the European Association for the Study of Diabetes (EASD) and the International Diabetes Federation (IDF), convened an expert committee to review research on a diagnostic blood sugar test, A1c. It improves on prior tests because it requires no fasting. Hemoglobin A1c is a form of the red blood cell protein bound to glucose; its level indicates a person's average blood sugar over the preceding 3 months. The expert committee urged that people with A1c readings of 6% or higher should be considered for preventive interventions. But it unanimously rejected the term prediabetes, saying it implies that prediabetic patients eventually will get diabetes and everyone else won't—"neither of which is the case."
ADA went in the opposite direction. It kept the term and ratcheted down its prediabetes A1c threshold from 6.1% to 5.7%—a move its two partners in the expert report never embraced. Evidence favored the lower figure, Cefalu says, noting that prediabetes comprises a "continuum of risk," with higher A1c readings warranting more aggressive treatment. ADA's new A1c standard, combined with its adoption of a similarly broad standard on a different blood sugar test a few years earlier, created about 72 million potential new prediabetes patients in the United States alone—and could create hundreds of millions more if embraced worldwide.
One of those Americans was Nance Yuan, a surgical resident in Los Angeles, California. In 2018, amid a demanding career, Yuan decided to freeze some of her eggs in case she wanted to have a baby later. A blood test before the procedure led to unwelcome news: prediabetes. Her result was borderline, so normally she would have been urged simply to improve her diet and to exercise more. But in light of her egg harvesting plans, Yuan's gynecologist also prescribed metformin, a drug often given to patients with full-blown diabetes.
"It was a little bit of a shock," says the slim, 34-year-old physician. "I think of metformin as something that middle-aged obese patients are on." Using the drug for just a few months gave Yuan empathy for her own patients who don't take medicines as prescribed. "On random days, I would get these horrible side effects and feel really sick," she says. Even on good days, the diagnosis added a tinge of anxiety to her life.
In interviews and public statements over many years, Albright has suggested such anxiety is warranted. She has described prediabetes as a runaway train. Patients are "walking toward the cliff's edge," she said at a 2016 medical conference. The media has reported, widely and with little skepticism, that prediabetes as defined by ADA and CDC is a serious health threat. Even Nutrition Action, a popular newsletter that debunks outlandish health claims, depicted diabetes as the tip of an iceberg with the prediabetes behemoth lurking below.
The broadened definition of prediabetes has triggered far-reaching changes in the medical landscape. CDC's budget for diabetes prevention jumped from $66 million in 2010 to $173 million in 2017—a 123% rise, in constant dollars. (At the same time, the agency's cancer prevention budget plummeted.) Many people diagnosed with prediabetes visit doctors more often for blood sugar tests and advice on diet and exercise. And a major marketing opportunity has opened up. Companies have pressured the Food and Drug Administration (FDA) for a seal of approval on foods or supplements—such as coffee, dairy products, and sugar substitutes—that they say can help prevent diabetes. A cottage industry of specialty fitness coaches emerged to serve a multitude of worried prediabetes patients.
Several companies target prediabetes patients for sales of continuous glucose monitors that attach to the body and transmit data to smartphones. Market analysts say makers of the devices—which can cost thousands of dollars annually—could add 25 million customers in the next 12 years in the United States alone, with prospects of a much larger global market, including China. Simpler home glucose testers, such as those used in finger sticks, are a razors-and-blades proposition. A starter model costs just $11.55 on Amazon, but requires branded test strips. Writing recently in Psychology Today, a physician and nutrition consultant touted home glucose testing after every meal as the top medical necessity for everybody—regardless of health or blood sugar status.
As the market expands, individual patients and society as a whole face a steep price tag. In 2014, failing to note its own role in widening the patient pool, ADA warned of "alarming" U.S. prediabetes spending, estimated at $44 billion in 2012 (the most recent year for which data are available)—about 1.6% of all health care costs and 74% higher than in 2007. Even if that is an exaggeration, as some say, costs are substantial and rising.
Cefalu says lower-cost interventions are needed, but he describes many current programs as cost-effective, and a medical and moral necessity. In a recent article about diabetes prevention, he invoked Winston Churchill's wartime pronouncement: "It is no use saying, ‘We are doing our best.’ You have got to succeed in doing what is necessary."
Many scientists, however, question the need to identify and treat prediabetes as ADA defines it. All endorse healthy diets and regular exercise and say substantially elevated glucose levels can lead to diabetes. But researchers diverge over crucial questions: How often and how quickly does prediabetes progress to diabetes? Does prediabetes itself cause harm, particularly when a person's average glucose levels are at the low end of the test result spectrum as defined by ADA?
On one side are CDC and ADA, powerful financial gatekeepers and opinion leaders. ADA journals are the field's most influential, and the two organizations fund much of the nation's research and programs on diabetes prevention. On the other side are public health and primary care authorities, including WHO, the U.K. National Institute for Health and Care Excellence, EASD, and IDF. Those groups either do not use or do not emphasize the term prediabetes, and they normally advise treatment only when blood sugar levels approach those of frank diabetes.
Researchers skeptical of the ADA definition point to a comprehensive 2018 review of 103 studies by the Cochrane Library in London, which showed that most people who qualify as prediabetic never progress to diabetes over any period studied. People who do progress usually start out at the higher end of the ADA prediabetes test range. The review also noted that studies of people labeled prediabetic often fail to account for weight, age, and physical activity, which can all affect glucose, as can daily stress, inflammation, and other factors. According to the review, up to 59% of prediabetes patients returned to normal glycemic values over 1 to 11 years with no treatment whatsoever.
"Doctors should be careful about treating prediabetes because we are not sure whether this will result in more benefit than harm," the Cochrane authors concluded, "especially when done on a global scale."
Kahn adds that even people whose elevated blood sugar does lead to diabetes progress slowly. "If you screen for diabetes every 3 to 5 years, you're fine," he says. Prediabetes itself does not raise the risk of cardiovascular disease or other complications of diabetes, says former ADA President Mayer Davidson, a physician now at Charles R. Drew University of Medicine and Science in Los Angeles.
To support the value of aggressive treatment, ADA and CDC cite a landmark clinical trial of prediabetes, the Diabetes Prevention Program Outcomes Study (DPPOS), which began in 2002 and is still collecting data. Sponsored by the National Institute of Diabetes and Digestive and Kidney Diseases and conducted by researchers at 25 centers across the nation, it included nearly 3000 prediabetic subjects randomly separated into three groups. One received intensive diet and exercise interventions, one took metformin, and a control group got neither. Fewer people in the treatment groups progressed to diabetes over 6 years—5.3% per year in the lifestyle group and 6.4% for the metformin group, compared with 7.8% in the control group, DPPOS scientists reported in 2009 in The Lancet.
The study's patients were selected on the basis of a glucose tolerance test, now rarely used, whose results are inconsistent with those of today's more common A1c. On average, the people studied were less healthy than most people within ADA's broad definition of prediabetes—obese, with blood glucose levels near the cusp of diabetes. The sample also skewed to ethnicities more prone to diabetes. To critics such as Yudkin, those features of the study raise doubts about whether its results justify treating everyone diagnosed with prediabetes.
Doctors should be careful about treating prediabetes because we are not sure whether this will result in more benefit than harm.
David Nathan, a diabetes specialist at Harvard Medical School in Boston and chair of the DPPOS research group, explained to Science that the researchers chose a relatively high-risk cohort "so we could study a manageable/affordable-sized population over a realistic period of time." The data, he added, showed treatment benefits across all ethnicities.
Yet 38% of the lifestyle treatment group failed to maintain the strict regimen after just 6 months. That outcome was despite costly, intensive medical, psychological, and exercise support, which community-based programs can't provide. Henry Kahn (no relation to Richard Kahn), a former CDC physician who retired from Albright's staff in 2018, says the DPPOS demonstrated "efficacy—I won't say ‘effectiveness’—to delay diabetes through very intensive and expensive education for a selected group." He says the relevance of the DPPOS findings to the real world is unclear.
Richard Kahn goes further. "There's never been a study to show that a population can lose sufficient weight and keep it off over more than a few years in order to prevent diabetes over the long run. Never. Not one." Gojka Roglic, a physician responsible for WHO's diabetes program, says ADA's prediabetes definition sharply increases the number of people said to be at risk despite the lack of "a preventive intervention that has been shown to work for people identified by these criteria."
Cefalu acknowledges that "it's a different approach in the real world. There's a lot of additional hurdles—social-economic concerns, finances, adherence, compliance." But he and Albright told Science the DPPOS showed that, in principle, lifestyle changes can delay or prevent diabetes. "We need to find ways to ensure that these programs can be [successful]," Cefalu says.
The DPPOS has yet to report whether lifestyle changes or metformin reduced heart disease and stroke—key complications of full-blown diabetes—but those interventions have had no impact on mortality. Women in the control group did have more microvascular changes—seen in tiny blood vessels—which can eventually damage eyes, kidneys, and nerves, than those in the lifestyle intervention group. But that effect was barely statistically significant, and the opposite pattern emerged in men. "If after 15 years of follow-up, the rate of eye, kidney, and nerve damage is exactly the same in all three groups," Yudkin says, "then there are 2000 people being treated with no benefit."
Cefalu says that in other studies, researchers have found a relationship between having an average blood glucose level within the ADA prediabetes range and cardiovascular disease many years later. But many of the studies were observational, based on medical records, and such studies are generally considered weaker than controlled trials such as the DPPOS. Most of the studies did not show harm at statistically significant levels.
Cefalu and Albright cited another paper—a long-term Chinese study of people diagnosed with prediabetes published in 2014 in The Lancet—that Albright said "showed lifestyle changes improved cardiovascular outcomes." After 23 years, the study researchers found fewer deaths overall and fewer deaths from cardiovascular causes than in the control group. But the study was small, just 577 people, and the benefit was marginally significant and seen only for women.
While the scientific dispute simmers, drug companies are racing to fill the demand created by ADA's prediabetes definition. The federal registry ClinicalTrials.gov lists human tests for more than 100 drugs, supplements, and other prediabetes remedies, including devices such as gastric bands. Drugmakers, ADA, and others have funded trials for at least 10 classes of drugs.
So far, FDA has approved no drug or device for prediabetes. Approval timelines remain hazy because the agency has yet to define clear therapeutic goals, says Todd Hobbs, U.S. medical officer for Novo Nordisk, a leading diabetes drugmaker based in Bagsv?rd, Denmark. "Is it delaying of patients who have normal glucose from developing prediabetes? … Is it a lower percentage of people who have prediabetes going on to develop type 2 diabetes in a defined period of time?" he asks. Hobbs says FDA eventually will pick some criteria. Prediabetes "is already an important issue, and the obesity epidemic is clearly driving this," he says. "It will certainly only increase in focus."
Meanwhile, doctors can treat prediabetes with a rising number of drugs prescribed off label—most approved for diabetes or obesity. In 2007, ADA began to recommend metformin as a relatively safe, inexpensive, long-term option for prediabetes patients with other risk factors for diabetes, such as obesity. Then, in 2013, ADA joined other advocates of aggressive diabetes prevention in listing more powerful, costlier drug options.
Cefalu says ADA still discourages using any drug except metformin for prediabetes. But since 2013, ADA standards of care have consistently listed a variety of diabetes and obesity drugs that it says may decrease the incidence of diabetes among prediabetics. ADA urges doctors to consider "cost, side effects, and lack of persistence in effect" but does not advise against prescribing the drugs, as it did in the past.
It seems counterintuitive to take a medicine in order to prevent something for which you would take that medicine.
Some physicians are trying new drug options on ADA's list. For example, a 2018 observational study in The Lancet looked at 222 prediabetes patients in a Southern California community medical practice who tested, on average, at the very low end of ADA's A1c scale—barely prediabetic. They were given a cocktail of either two or three diabetes drugs. On average, their blood sugar decreased slightly.
Any drug for prediabetes would likely have to be taken for years, perhaps a lifetime, so such modest benefits must be weighed against the potential harms of continual treatment. And several prediabetes options described by ADA and others present serious hazards. Pioglitazone, a drug to lower blood sugar developed by the Tokyo-based Takeda Pharmaceutical Company under the brand name Actos, carries on its label a "black box" side effect warning of the risk of congestive heart failure. The drug can also raise the risk of bone fractures and cancer. Exenatide (branded Bydureon by Cambridge, U.K.–based AstraZeneca), which reduces blood glucose and suppresses appetite, has a black box warning for thyroid cancer. Liraglutide (sold as Victoza by Novo Nordisk) also carries warnings of potentially lethal side effects.
Yudkin says the growing emphasis on drugs for prediabetes reflects in part a false sense of urgency about its health risks. WHO's Roglic also doubts that the benefits of taking a diabetes drug for prediabetes outweigh the risks. "It seems counterintuitive to take a medicine in order to prevent something for which you would take that medicine. The reasoning is a bit twisted."
The push for drug treatments for prediabetes takes place in an environment rife with financial conflicts of interest. For years, ethicists have criticized ADA for financial dependence on diabetes drugmakers. In recent years, ADA says, it has received $18 million to $27 million annually from drug companies, including many donations of $500,000 to $1 million per year. The group also gets up to $500,000 annually from each of more than a dozen other firms in the diabetes and prediabetes markets, including makers of consumer and medical products, testing labs, insurance companies, and drug retailers.
The American College of Physicians recently reviewed the diabetes and prediabetes treatment guidelines from several authorities and assessed each for financial conflicts of interest with drug companies. It rated ADA's conflicts as among the most extreme. An ADA representative wrote in an email to Science that its funders have no influence over "the final outcome" of ADA's work.
Many prominent physicians who now recommend drugs for prediabetes also received large payments in recent years from companies whose sales could be affected by such endorsements. Science documented some of that compensation by examining data from mid-2013 through 2017 in Open Payments, a federal database that tracks money that drug- or devicemakers give to clinical professionals for consulting, teaching fees, travel, and other purposes. The database also shows funds funneled through their institutions for research.
Ralph DeFronzo, an endocrinologist at the University of Texas Health Science Center in San Antonio and a co-author of treatment guidelines for the American Association of Clinical Endocrinologists (AACE) and the American College of Endocrinology (ACE), is one of the most influential diabetes researchers. His papers discussing wider use of drugs for prediabetes—including the Southern California study in The Lancet—have been cited more than 3000 times. DeFronzo got about $5 million from diabetes drugmakers, including $1.3 million in fees paid directly to him for consulting, advice, and travel. About half of the total was for work associated with drugs he has cited as effective for some prediabetes patients, such as pioglitazone, exenatide, and liraglutide. DeFronzo declined to comment about whether those earnings might have influenced his views.
Many physicians who wrote the 2018 ADA standards of care for prediabetes, which recommend that doctors consider prescribing those same three drugs, also received large sums from drugmakers. Seven of the 14 ADA experts got between $41,000 and $6.8 million between mid-2013 and 2017 from makers of diabetes drugs or candidate prediabetes medicines or devices. The payments covered consulting, travel, and research and included an average of $276,000 for personal fees.
Guillermo Umpierrez, who served on the ADA standards panel and is an endocrinologist at Emory University in Atlanta (he was also a member of the AACE-ACE panel), received the most during that period, according to Open Payments: $6.8 million—including about $123,000 in consulting and travel fees. About $5.3 million came from the makers of drugs in the classes that ADA now recommends as options for prediabetes, for research or consulting on drugs in those classes. Umpierrez told Science that he plans to dispute some figures in Open Payments, which has sometimes included errors, but he gave no details. Umpierrez said drug company payments did not influence him.
ADA says it does not regard as conflicts of interest any drugmaker-provided research funding, funds received before the year its guidelines were written, or payments under $10,000. Even by those narrow criteria, 41% of the authors of its standards of care during the past 5 years were conflicted, according to ADA. The group can require such panelists to recuse themselves from discussion or voting, but in the past 2 years, none did so.
John Ioannidis, a physician and expert on evidence-based medicine at Stanford University in Palo Alto, California, calls prediabetes a classic example of how clinical guidelines from groups such as ADA can escalate costs to society, to the benefit of specialty practitioners and drug companies. "You have a combination of two forces. One is to expand the definition of disease and to get more people classified as being sick or in need of treatment. And second, direct endorsement of specific interventions that [guideline authors] have direct conflicts of interest with," he says. "It's really very worrisome."
ADA's expanded prediabetes definition is a financial win for physicians, companies that conduct lab tests, drug companies, device and app developers, clinics, and hospitals, says Victor Montori, a diabetes clinician at the Mayo Clinic in Rochester, Minnesota. "The people who lose are the people who go from being a healthy person to being a patient. Now, they have the sick role. They have to go for checkups and tests and treatments," often at considerable expense.
I just don’t think we [prevent diabetes] by making every healthy person a patient.
Yuan, the surgical resident, had her eggs frozen after a few months of prediabetes treatment. With great relief, she stopped taking metformin—which several diabetes and fertility experts say was overtreatment in her situation. But the prediabetes diagnosis and prescription remain on her medical record, which could affect her future insurance premiums. Other people given a prediabetes diagnosis face greater challenges: For example, according to a recent study, Medicare reimburses just a small fraction of costs for its own diabetes prevention program.
The insurance problem is one sign of what happens when a borderline medical test result becomes "medicalized," turning many basically healthy people into stressed-out patients, critics of ADA's prediabetes definition say. "Seventy to 80% of them are never going to get diabetes, so is it unnecessary stress?" asks Davidson, the former ADA president.
Many public health organizations believe a mainly clinical approach to diabetes prevention is ineffective. WHO, for example, favors societywide solutions, which aim to address the health impacts of social stratification and failures of urban planning. It endorses laws that help reduce consumption of sugary drinks and unhealthy foods.
A 2011 study in JAMA examined health outcomes in women living in stressful, low-income housing projects who had been randomized into three groups: One got a voucher for better housing and help moving, one a voucher for moving to any area without help, and a control group got neither. Women who got the most housing assistance suffered the least obesity and diabetes—over 20 years, about 15% had become diabetic, compared with 20% in the control group.
"None of those people got advice on lifestyle changes. None got advice on dietary changes or activities. None got diabetes drugs," Montori says. Diabetes can and should be prevented, he says. "I just don't think we do that by making every healthy person a patient."
Jia You and Meagan Weiland contributed reporting. This story was supported by the Science Fund for Investigative Reporting. If you want to contribute to similarly ambitious investigations, please give a tax-deductible gift to the fund today.