The Paradox of How We Treat Diabetes

Understanding diabetes today requires holding two conflicting realities in your head simultaneously.

First, diabetes therapy has been revolutionized by a world of new drugs that have become available since the turn of the century—most notably, drugs of the same class as Wegovy and Ozempic that began their existence as diabetes medications and are now hailed as wonder drugs for treating obesity. These drugs do the best job yet of controlling blood sugar and, of course, body weight, which is critical for those Type 2 diabetes, the common form of the disease that constitutes over 90 percent of cases and is associated with age and obesity. For type 1 diabetes, the acute condition that typically strikes in childhood and adolescence, new devices—continuous blood sugar monitors and automated insulin delivery systems—make blood sugar control easier than ever. Still more advanced devices and better drugs are in the pipeline.

But then there’s the flip-side. It’s why the pharmaceutical industry has invested so heavily in new therapies: Once a relatively rare condition, diabetes is now so common that drugstores dedicate entire aisles to it and television commercials for diabetic medications are common fare. In 1960, when the first concerted federal surveys were quantifying prevalence, two million Americans were living with a diabetes diagnosis. Today that number is 30 million; almost nine million more have diabetes but don’t yet know it. Each year, 1.4 million new cases are diagnosed and at ever younger ages.  

Diabetes puts all of these individuals at increased risk of heart disease, strokes, cancer, blindness, kidney failure, nerve damage, gangrene, and lower limb amputation. It increases cognitive impairment and dementia risk as patients age. Living with diabetes still comes with a decrease in life expectancy of six years.

For those with Type 1 diabetes, despite the remarkable new drugs and devices, blood sugar control is seemingly getting worse, on average, not better. As of 2018, fewer than one in five individuals diagnosed with Type 1 diabetes were achieving even the relatively generous blood-sugar goals set by the American Diabetes Association (ADA); this was a smaller proportion than a decade earlier.

More From TIME

Despite the remarkable advances in therapy, both Type 1 and Type 2 diabetes are still considered progressive chronic diseases, meaning the patient’s condition is expected inevitably to deteriorate as they live with the disease. The greatest challenge to better therapy, as one recent analysis suggested, is the hesitation of physicians to continue prescribing more or newer drugs and increasing dosages as the diseases progress.

All of this comes with a staggering financial burden. In November, the ADA estimated that the total annual cost of diabetes in the U.S. is over $400 billion; over $300 billion is direct medical costs. This was up $80 billion from 2017 when an editorial commenting on a similar accounting characterized these costs as the “elephant in the room” of the diabetes epidemic. Patients with diabetes are likely to spend over $12,000 a year just for medical care, almost three times that of healthy individuals of equivalent age. It does not help that the drugs themselves—whether insulin or Ozempic and its ilk —are expensive, costing many thousands of dollars a year. One in every four health care dollars spent in America goes to treating diabetic patients.

And the U.S. is by no means unique. The World Health Organization estimates that diabetes prevalence worldwide increased four-fold between 1980 and 2014, from 108 million to over 400 million, with the greatest rise coming, paradoxically, in the poorest countries. In 2016, Margaret Chan, then WHO director general, described the situation as a “slow-motion disaster” and predicted with near absolute certainty that these numbers would only get worse. They have.  

So how do we reconcile these conflicting realities: Unprecedented advances in medical therapies for an out-of-control disease epidemic in which patients, at least in general, are doing poorly and can expect to do worse as time goes on? Confronted with such a dismal state of affairs shouldn’t we be asking how we got to this point? Were mistakes made in how we think about this disease? Were questionable assumptions treated as facts, and could those assumptions be wrong?

Asking the Right Questions

These are the kinds of questions you would hope health organizations worldwide would be asking, but surprisingly they have no mechanisms or protocols to do so. Diabetes associations like the ADA will regularly convene expert panels to address revisions in the latest standard of care guidelines to accommodate the latest research, but not whether the guiding principles underlying those guidelines should be rethought entirely. Independent investigators are not recruited to analyze and to provide an unbiased assessment of where progress might have gone off the rails. That job instead has been left to physicians in their clinics, those confronted with ever more diabetic patients and willing to take the risk of thinking independently, and to investigative journalists like myself, whose obligation when confronted with such conflicting realities is to ask just these kinds of questions.

Among the revolutions that changed medical practice over the past half century, one in particular is very relevant here. Beginning in the 1970s, health-care analysts began to confront quite how little physicians really knew about the risks and benefits of what they were doing for their patients. Not only had clinical trials demonstrated that some standard medical practices resulted in far more harm than good—the surgical procedure known as a radical mastectomy, most infamously, for breast cancer—but researchers were documenting wide variations in medical practices from physician to physician, hospital to hospital and state to state. This, in turn, resulted in a wide variation of benefits, harms and costs to the patients, depending on which physicians they might visit, and so which treatments they might get.

Read More: Should We End Obesity?

The revolution that followed became known as the Evidence-Based Medicine (EBM) movement, founded on the principle that medical interventions should be rigorously tested in clinical trials— double-blind, randomized, placebo-controlled—before they be used or prescribed. This would be necessary whenever physicians were faced with a choice between multiple options, and whenever the harms of an intervention might outweigh the benefits. David Sackett of McMaster University, a founder of the movement, would describe the EBM process as beginning with the fact that half of what aspiring doctors learn in medical school is “dead wrong,” and then trying to establish thoughtfully and critically which half that is. David Eddy of Duke University, another EBM pioneer, later described his motivation and that of his colleagues as the revelation that “medical decision making was not built on a bedrock of evidence or formal analysis, but was standing on Jell-O.”

It would be nice to think that this situation has been widely resolved by evidence-based guidelines, but that’s not the case. Journalists or physicians looking for the evidence base in decision making about diabetes therapies, will likely find themselves, as I did, with the same revelation. Clearly it, too, was standing on Jello-O in the 1970s, but the problem neither began nor ended there. A remarkable history emerges, with three clear observations.

First, we’ve been here before. We have had miracle drugs for diabetes. Most notably, the hormone insulin itself, when University of Toronto researchers led by Frederick Banting and Charles Best purified it and put it to use in 1922 treating patients with severe cases of diabetes. We then had better insulins, slower-acting and longer-lasting, and then, in the post-World War 2 years, drugs (oral hypoglycemic agents) that could lower blood sugar without having to be injected, as insulin did. We have had revolutionary advances in diabetes technology, beginning in the 1970s with devices that allowed patients to monitor their own blood sugar, and then insulin pumps that automated the process of insulin therapy. All contributed to easing the day-to-day burden of diabetes. None had any influence in controlling the epidemic, nor did they eradicate or meaningfully reduce the long-term complications of the disease. Put simply: diabetes plus drug therapy and devices, even the best drug therapy and devices, does not equate to health.

Secondly, diabetes researchers have not been averse to testing their fundamental assumptions. They‘ve done so in ever more ambitious clinical trials. But a disconcerting proportion of those trials failed to confirm the assumptions, despite the fact that it was these assumptions that constituted the rationale for therapeutic approaches. The $200 million Look AHEAD Trial, for example, tested a foundational belief in the field: that weight loss in those with Type 2 diabetes would lengthen lives. The trial was ended for “futility” in 2012. ”We have to have an adult conversation about this,” as David Nathan, a Harvard diabetes specialist, said to The New York Times. The 10,000-patient ACCORD trial had also been ended prematurely just four years earlier. “Halted After Deaths,” in the words of The New York Times headline. “Medical experts were stunned,” the 2008 article said. ACCORD was one of three trials testing the assumption that intensive blood sugar control by medications would reduce the macrovascular complications of Type 2 diabetes—particularly heart disease—and premature death. All three trials failed to confirm it.

Third, the remarkable aspect of all these trials is that they all assumed an approach to dietary therapy that itself had never been tested. This is the “standing on Jell-O” problem. For well over a century, diabetes textbooks and chapters in medical texts invariably included some variation on the statement that diet is the cornerstone of treatment. The most recent guidelines from the ADA refer to dieting as “medical nutrition therapy” (MNT) and say MNT is “integral” to therapy.

But what constitutes MNT—the dietary advice given—has been determined not by any meaningful research comparing different dietary approaches. Rather it has been assumed that individuals with diabetes should eat the same “healthful eating pattern” that health organizations recommend for all of us—“non-starchy vegetables, fruits, legumes, dairy, lean sources of protein… nuts, seeds, and whole grains”—albeit with the expectation, if weight control is necessary, that they should eat fewer calories.

Read More: Are Weight Loss Drugs From Compounding Pharmacies Safe?

Controlling the symptoms and complications of the disease is left to insulin and the pharmacopeia of drugs that work to maintain blood sugar levels near enough normal that the specter of diabetic complications may be reduced as well. Diabetes associations have assumed that this approach is easiest on the patients, allowing them to balance the burden of insulin injections or multi-drug therapy, against the joy of eating as their non-diabetic friends and family do. But this assumption has never been tested to see if it is true, nor whether a better approach exists that might truly minimize the disease burden of diabetes, extend lives and make the trade-off of restrictive eating vs. health worthwhile.

History of Diet and Diabetes

This is where understanding the history of the diet-diabetes relationship can be vitally important. What has been known for certain about diabetes since the 19^th century is that it is characterized by the inability to safely metabolize the carbohydrates in our diet. This observation led to two divergent approaches/philosophies to dietary therapy. Beginning in 1797, when a British physician named John Rollo wrote about curing a diabetic patient using a diet of fatty (rancid) meat and green vegetables, through the early 1900s, diabetes therapy was based on the assumption that since individuals with diabetes could not safely metabolize the sugary and starchy foods in their diet, they should abstain from eating them. In this pre-insulin era, the only meaningful advice physicians could give their patients was dietary, variations on Rollo’s approach: sugars, grains, starches, even legumes were prohibited because they are carbohydrate-rich: meats, ideally as fatty as possible, butter and eggs, along with green leafy vegetables (boiled three times to remove the digestible carbohydrates) could be eaten to satiety.

Throughout Europe and America, this was known was “the animal diet,” endorsed by virtually every major diabetes specialist of the 19th Century. Physicians believed that the more calories their diabetic patients consumed, and ideally the more fat (because protein is composed of amino acids, some of which the liver converts to carbohydrates), the healthier they would be.  “Patients were always urged to take more fat,” is how this was described in 1930 by the Harvard physician Elliot Joslin, who was then, far and away, the most influential diabetes authority worldwide. “At one time my patients put fat in their soup, their coffee and matched their eggs with portions of fat of equal size. The carbohydrate was kept extraordinarily low….”

This thinking only changed in the years before World War I, when Joslin embraced and disseminated the idea promoted by a Harvard colleague, Frederick Allen, that diabetic patients, still without insulin, were best served if they were semi-starved—avoiding carbohydrates and fat. In short, patients suffering from a disease in which one characteristic symptom is ravenous hunger would be treated by making them go even hungrier than otherwise. The approach was unsurprisingly controversial. Joslin and others, though, came to believe they could keep their young Type 1 patients alive longer with Allen’s starvation therapy, even while the high fat, animal-based diet seemed more than adequate for their older Type 2 patients. Allen’s starvation therapy was in turn challenged between 1920 and 1923, when University of Michigan physicians Louis Newburgh and Robert Marsh reported in a series of articles that it was simply unnecessary, that even young patients with severe diabetes could thrive on the high-fat, carbohydrate-abstention approach if properly administered. By then, though, it was too late.

Insulin therapy had arrived in the winter of 1922. It launched what medical historians would call a “therapeutic revolution,” as close as medicine had ever come, and maybe ever has, to a miracle. Patients, often children, on the brink of death, horribly emaciated by the disease and the starvation therapy, would recover their health in weeks, if not days on insulin therapy. They were resurrected, to use the biblical terminology, which physicians of the era often did.

Diabetes specialists realized that insulin therapy was not a cure of the disease, but it allowed their patients to metabolize carbohydrates and held the promise of allowing them to eat whatever and however they wanted. “Were I a diabetic patient,” wrote Frederick Banting in 1930, by then a Nobel Laureate. “I would go to the doctor and tell him what I was going to eat and relieve myself of the worry by demanding of him a proper dose of insulin.”

That thinking, for better or worse, has governed diabetes therapy ever since.

While diabetes specialists still had no conception of the long-term complications of living with diabetes—the damage to large and small blood vessels that results in heart disease, strokes, kidney disease, neuropathy, amputations, blindness, dementia—they would advocate for ever more liberal carbohydrate diets and ever higher insulin doses to cover them. Patients would be taught to count the carbohydrate content of each meal, but only so they could properly dose their insulin. Diets would be prescribed, and still are, to allow for the drugs to be used freely, not to minimize their use. Patients, in turn, were allowed to eat anything, which physicians assumed they would do anyway.

Whether the patients lived longer, healthier lives because of it, would never be tested.  As diabetes specialists began to understand the burden of the disease they were treating, the wave of microvascular and macrovascular complications that set in after 10 or 20 years, they would rarely, if ever, ask the question, whether these complications were mitigated by their dietary approach or perhaps exacerbated by it. They would only test drug therapy.

In 1971, the American Diabetes Association institutionalized this philosophy with dietary guidelines that would commit the organization to this approach ever after: diabetic patients would be told to restrict dietary fat—by then thought to cause heart disease—rather than carbohydrates, the one macronutrient they could not metabolize safely without pharmaceutical help. “Medical Group, in a Major Change, Urges a Normal Carbohydrate Diet for Diabetics,” was the headline in The New York Times. By taking the ADA’s advice, diabetic patients would trade off blood sugar control for cholesterol, assuming this would prevent heart disease and lengthen their lives. While the guidelines explicitly acknowledged that the ADA authorities had no idea if this was the right thing to do, the advice would be given anyway.

Read More: Why You're Not Losing Weight

By 1986, the ADA was recommending diabetic patients get “ideally up to 55-60% of total calories” from carbohydrates, while researchers led by the Stanford endocrinologist Gerald Reaven had established that such a diet was almost assuredly doing more harm than good. That same year, the NIH held a “consensus conference” on diet and exercise in Type 2 diabetes. The assembled authorities concluded that, at best, the nature of a healthy diet for diabetes remained unknown. The conference chairman, Robert Silverman of the NIH, summed the state of affairs up this way: “High protein levels can be bad for the kidneys. High fat is bad for your heart. Now Reaven is saying not to eat high carbohydrates. We have to eat something.” And then he added, “Sometimes we wish it would go away, because nobody knows how to deal with it.”

The modern era of the diabetes-diet relationship began 25 years ago, with the awareness that the nation was in the midst of an obesity epidemic. Physicians, confronted with ever more obese and diabetic patients and the apparent failure of conventional advice—eat less, exercise more—suggested instead the only obvious options, the approaches suggested by popular diet books. Many of these—Dr. Atkins’ Diet Revolution, Protein Power, Sugar Busters—were touting modern incarnations of Rollo’s animal diet.

The Diet Trials

The result was a series of small, independent clinical trials, comparing, for the first time, the conflicting dietary philosophies of a century before. Is it better for patients with Type 2 diabetes, specifically, to avoid dietary fat and, if they’re gaining weight, restrict total calories (both carbohydrates and fat), or will they do better by avoiding carbohydrate-rich foods alone and perhaps entirely? The earliest trials focused on treating obesity, but many of the participants also struggled with Type 2 diabetes. In 2003, physicians at the Philadelphia VA Medical Center published the results from the first of such trials in the New England Journal of Medicine: patients with both obesity and diabetes counseled to eat as much food as they desired but to avoid carbohydrates, became both leaner and healthier than patients counseled to eat the low-fat, carbohydrate-rich, calorie-restricted diet prescribed by both the American Heart Association and ADA. The numerous trials since then have concluded much the same.

Among the profound assumptions about Type 2 diabetes that these trials have now challenged is that it is, indeed, a progressive, degenerative disorder. This may only be true in the context of the carbohydrate-rich diets that the ADA has recommended. In 2019, researchers led by the late Sarah Hallberg of the University of Indiana, working with a healthcare start-up called Virta Health, reported that more than half of the participants in their clinical trial were able to reverse their type 2 diabetes by eating what amounts to a 21^st century version of Rollo’s animal diet or the Newburgh and Marsh approach. They were able to discontinue their insulin therapy and all but the most benign of their diabetes medications (known as metformin) while achieving healthy blood sugar control. A third of these patients remained in remission, with no sign of their disease, for the five years, so far, that their progress has been tracked.

As for Type 1 diabetes, in 2018, a collaboration led by the Harvard endocrinologists Belinda Lennerz and David Ludwig reported on a survey of members of a Facebook Group called TypeOneGrit dedicated to using the dietary therapy promoted by Dr. Richard Bernstein in his book Dr. Bernstein’s Diabetes Solution. Bernstein’s approach requires patients to self-experiment until they find the diet that provides stable healthy levels of blood sugar with the smallest doses of insulin. Such a diet, invariably, is very low in carbohydrates with more fat than either the ADA or AHA would deem healthy. Both youth and adults in the Harvard survey maintained near-normal blood sugar with surprisingly few signs of the kind of complications—including very low blood sugar, known as hypoglycemia—that make the life of a patient with Type 1 diabetes so burdensome. The TypeOneGrit survey, Lennerz said, revealed “a finding that was thought to not exist. No one thought it possible that people with type one diabetes could have [blood sugar levels] in the healthy range.” This does not mean that such diets are benign. They may still have the potential to cause significant harm, as Lennerz and Ludwig and their colleagues made clear. That, again, has never been tested.

One consequence of the diabetes associations embracing and prescribing a dietary philosophy in 1971 that has only recently been tested is that we’re back to the kind of situation that led to the evidence based medicine movement to begin with: enormous variation in therapeutic options from physician to physician and clinic to clinic with potentially enormous variations in benefits, harms and costs.

Even the ADA advice itself varies from document to document and expert panel to expert panel. In 2019, for instance, the ADA published two consensus reports on lifestyle therapy for diabetes. The first was the association’s consensus report on the standard of care for patients with diabetes. The authors were physicians; their report repeated the conventional dietary wisdom about eating “vegetables, fruits, legumes, whole grains….” It emphasized “healthful eating patterns”, with “less focus on specific nutrients,” and singled out Mediterranean diets, Dietary Approaches to Stop Hypertension (known as the DASH diet) and plant-based diets as examples that could be offered to patients. This ADA report still argued for the benefits of low-fat and so carbohydrate-rich diets, while suggesting that the “challenges with long-term sustainability” of carbohydrate-restricted eating plans made them of limited use.

Three months later, the ADA released a five-year update on nutrition therapy. This was authored by a 14-member committee of physicians, dietitians and nutritionists. Among the conclusions was that the diets recommended as examples of healthful eating patterns in the lifestyle management report—low-fat diets, Mediterranean diets, plant-based diets and the DASH diet—were supported by surprisingly little evidence. In the few short-term clinical trials that had been done, the results had been inconsistent. As for carbohydrate-restricted high fat eating patterns, they were now “among the most studied eating patterns for Type 2 diabetes,” and the only diets for which the results had been consistent. “Reducing overall carbohydrate intake for individuals with diabetes,” this ADA report stated, “has demonstrated the most evidence for improving glycemia [high blood sugar] and may be applied in a variety of eating patterns that meet individual needs and preferences.”

Physician awarenessof the potential benefits of carbohydrate-restriction for Type 2 diabetes, meanwhile, still often comes from their patients, not their professional organizations. In the United Kingdom, for instance, David Unwin, a senior partner in a medium-sized practice began suggesting carbohydrate-restricted high fat diets to his patients in 2011, after seeing the results in one such patient who chose to do it on her own and lost 50 pounds. When results of her blood tests came back, says Unwin, they both realized that she was no longer suffering from diabetes. Both the weight loss and the reversal of diabetes were unique in Unwin’s experience. After reading up on the burgeoning literature on carbohydrate restriction, Unwin began counseling his diabetic patients to follow a very-low-carbohydrate, high-fat eating pattern. In 2017, the UK’s National Health Service awarded Unwin its “innovator of the year” award for applying a 200-year-old approach to diabetes therapy, as Unwin says, that “was routine until 1923.” Unwin has now published two papers documenting the experience in his medical practice. As of last year, 20 percent of the clinic’s diabetic patients—94 in total—had chosen to follow this restricted dietary approach and put their Type 2 diabetes into remission.

If the diabetes community is to solve the formidable problems confronting it, even as drug therapies get ever more sophisticated, it will have to accept that some of its fundamental preconceptions about diabetes and diet may indeed be wrong. As it does so, it will have to provide support for those living with diabetes who decide that what theyhave been doing is not working. Some patients, when confronted with the choice between following a restricted eating pattern that seemingly maximizes their health and wellbeing or eating whatever they want and treating the symptoms and complications with drug therapy, will prefer the former. For those who do, the informed guidance of their physicians and diabetes educators will be  invaluable.

When I interviewed individuals living with Type 1 diabetes, among the most poignant comments I heard was from a nutrition consultant diagnosed in 1977 when she was eight years old. She told me that she finally had faith she could manage her blood sugar and live with her disease when she met a physician who said to her “What can I do to help you?” That’s what changed her life, as much as any technology or medical intervention. In the context of the dietary therapies we’re discussing, that requires practitioners who are themselves open-minded and willing to spend the necessary time and effort to truly understand an approach to controlling diabetes that is, by definition, unconventional and, in Type 1 diabetes, still lacking clinical trials that test (or testify to) its safety and efficacy. Easy as it is for physicians to continue believing that what they should be doing is what they have been doing, they do not serve their patients best by doing so.

Adapted from Gary Taubes' new book Rethinking Diabetes: What Science Reveals About Diet, Insulin and Successful Treatments