Yet six years of medications and attempts at healthy living had failed to rein in her blood glucose, leaving Iaboni terrified that she was on course to have her kidneys fail "and my feet cut off" -- common consequences of uncontrolled diabetes.

Then the 45-year-old Connecticut wife, mother of two and head of human resources for a Fortune 500 company, lucked out. In 2009 she met with Dr Francisco Rubino of Weill Cornell Medical Center in New York. He had just received approval to study experimental surgery on diabetics with a relatively lean weight-to-height ratio, or body-mass index (BMI). Iaboni was among his first subjects.

Three years on, she has dropped 50 pounds to reach a healthy 145 and has normal blood pressure without medication. That isn't too surprising: Weight loss is the purpose of bariatric surgery and often reduces blood pressure. More remarkable, Iaboni no longer has diabetes.

She is not the first patient with diabetes, which can be triggered by obesity, to be cured by weight-loss surgery. But she is a rarity for having it with a BMI well below 35 and over. That's the level at which the American Diabetes Association says surgery "may be considered" and that Medicare and some private insurers cover. And Iaboni's diabetes disappeared months before she shed much weight.

Her experience has raised an intriguing possibility: that some forms of bariatric surgery treat diabetes not by making patients shed pounds. Instead, by rerouting part of the digestive system, they change what signals the gut sends to the brain and the brain sends to the liver, altering the underlying causes of diabetes.

If proven, bariatric surgery may help people with type-2 diabetes who are less obese, overweight or even of healthy weight. And it might be effective against the currently incurable type-1, or "juvenile," diabetes, too.

"Every textbook says that diabetes is chronic, irreversible, and progressive," said Rubino. "But we have thousands of patients who once had diabetes and now do not."

"INSUFFICIENT" EVIDENCE

Bariatric surgeons have long been prone to declaring victory against diabetes way too soon, before large-scale, long-term data proved their case. "The evidence for the success of bariatric surgery in patients with a BMI below 35 is not very strong," said Leonid Poretsky, director of the Friedman Diabetes Institute at Beth Israel Medical Center in New York City. "Most of the studies have been very small and not well controlled."

The American Diabetes Association rates the evidence that bariatric surgery can cure diabetes as "E," the lowest of four grades. It calls data on patients with a BMI below 35 "insufficient," and says the procedure cannot be recommended except as part of research.

The immediate risks of bariatric surgery are small -- a 0.3 percent chance of dying within 30 days of the procedure. But a small fraction of patients develop infections, leaking from the stomach into the abdominal cavity, or gallstones, and it can cause nutritional deficiencies: There is less intestine to absorb vitamins and minerals, raising the possibility of osteoporosis and anemia.

Despite these red flags, the surgical option is attracting intense interest because the quest to cure diabetes has become almost desperate. In type-1 diabetes, the pancreas does not produce enough insulin, a hormone that moves the glucose in food into cells. In type-2 diabetes, cells become resistant to insulin. In either case, glucose remains in the blood, damaging cells and blood vessels, sometimes severely enough to cause blindness, kidney failure, or gangrene requiring foot or limb amputations.

In 2010, 8.3 percent of adults worldwide had type-2 diabetes (11.3 percent did in the United States), resulting in direct medical costs of $376 billion ($116 billion in the United States). By 2030, the global incidence is projected to rise to 9.9 percent, partly because of the rising obesity rate, with costs reaching $490 billion.

The possibility that bariatric surgery could cure diabetes emerged about a decade ago. A long-term study of thousands of patients in Sweden reported in 2004 that both gastric bypass and banding improved diabetes in many subjects. A 2008 study of 55 obese patients found that 73 percent of those who underwent gastric banding saw their diabetes disappear after two years, compared to 13 percent undergoing standard medical treatment such as medication, diet and exercise.

In 2009, surgeons at the University of Minnesota analyzed 621 mostly small studies of bariatric surgery in obese, diabetic patients. Their conclusion, reported in the American Journal of Medicine: 78 percent no longer needed medication to control their blood sugar. They'd been cured. Lap banding had the worst results, worsening diabetes in some patients.

But most patients in these studies were obese, many morbidly so. (The average BMI was 48.) The improvement in glucose control could therefore be credited to the patients' weight loss, which averaged 85 pounds.

CLUES FROM THE PAST

Rubino had a hunch that something else was at work. As a research fellow in diabetes at Mount Sinai Hospital in New York in 1999, he was reviewing the medical literature one day for guidance on how to best perform bariatric surgery on a man with a BMI of 80. He found papers from the 1950s and earlier reporting that surgery for peptic ulcers had cured diabetes.

Ulcer surgery removes a portion of the stomach and reconstructs a connection to the intestine, much as gastric bypass does. Few diabetes experts had noticed the old papers; they were published in surgery journals, which endocrinologists seldom read.

His serendipitous find led Rubino to other papers describing operations on the digestive tract that cured diabetes, something that, according to medical textbooks, was unthinkable.

"Within two weeks of surgery and sometimes sooner, these patients were off their insulin, off their diabetes drugs, and with normal blood glucose levels," said Rubino. "That was too fast to explain by weight loss."

Yet that's how experts explained bariatric surgery's effect on diabetes, especially as the procedure took hold in the 1990s. Few surgeons focused on how quickly the condition disappeared, said Rubino, "or they speculated that patients weren't eating much after the surgery, and that's what cured their diabetes."

He began pursuing the idea that surgery might improve diabetes directly, rather than through weight loss. "I was ignorant of diabetes, so I wasn't burdened by too much knowledge," Rubino said. "Something that might have seemed heretical didn't seem impossible to me."

Rubino modified the popular gastric bypass surgery, called Roux-en-Y, to test his idea on diabetic lab rodents. In the classic operation, the stomach is pinched off so it can hold less food. Surgical cuts keep the rest of the stomach and the top of the small intestine, called the duodenum, from receiving any food. Instead, the stomach empties directly into the bottom of the small intestine, the jejunum. In Rubino's variation, called duodenal-jejunal bypass (DJB), the stomach is untouched, but the rest of the procedure is the same.

The rats that Rubino operated on beginning in 2000 were cured of diabetes much more quickly than their weight fell. It was the first rigorous evidence, from a well-controlled study, that gut surgery has an anti-diabetes effect.

In 2006, Rubino was ready to move from rats to people. Two patients, with BMIs of 29 and 30, underwent his procedure. Their blood sugar levels returned to normal within days, though they lost no weight. In his most recent trial, reported in March in the New England Journal of Medicine, Rubino and colleagues at Catholic University in Rome performed standard gastric bypass surgery or a procedure similar to DJB on people with type-2 diabetes. After two years, 15 of 20 bypass patients and 19 of 20 DJB patients no longer had diabetes.

Curiously, although patients shed pounds, there was no correlation between weight loss and blood glucose, the key marker of diabetes. "Bariatric surgery is more effective on diabetes than obesity," said Rubino. "Patients don't become lean, but they do not have diabetes anymore."

FROM GUT TO BRAIN

Research from the University of Toronto, reported online this month in Nature Medicine, may finally explain why. It examined the effects of bypass surgery on rats with type-1 diabetes, which is considered even harder to treat than type-2. Normally the jejunum receives only digested mush, as nutrients have already been absorbed in the duodenum, explained lead researcher Tony Lam.

Bypassing the duodenum allows the jejunum to receive an influx of nutrients for the first time, said Lam. Sensing them, the jejunum sends a "got glucose!" signal to the brain. The brain interprets that as a sign of glucose overabundance and orders the liver to decrease glucose production. Result: The rats no longer have diabetes.

"I believe that similar mechanisms are taking place in surgery for type-2 diabetes," said Lam. "It strengthens the case for the surgery treating diabetes independent of weight loss."

His rat study shows why lap banding and stomach stapling are less effective against diabetes than gastric bypass. Banding causes diabetes to go into remission in about 50 percent of patients, probably due to weight loss, said endocrinologist Dr Allison Goldfine of the Joslin Diabetes Center in Boston.

In contrast, the diabetes-remission rate after Roux-en-Y is 80 to 85 percent. "The improvements in blood glucose with Roux-en-Y appear to occur very early, by day three after surgery, so patients are being discharged with no medication," she said. Something other than weight loss "must be going on."

Goldfine has launched a study of diabetics with BMIs of 30 to 42 to compare outcomes after lap band surgery, Roux-en-Y, and intense medical management.

A year ago, Rubino began the first large study for type-2 diabetes patients with a BMI as low as 26, where "overweight" begins. The cost of the bypass surgery is covered by a grant from Covidien Plc, which makes laparoscopic instruments and surgical staplers. He aims to enroll at least 50 patients, following them for five years; he has operated on 20 so far.

(Reporting by Sharon Begley; Editing by Michele Gershberg and Prudence Crowther)

• A Permanent Economic Emergency