When Aaron Kowalski was diagnosed with diabetes 24 years ago, the young woman in her twenties who lived across the street from his family had already gone blind from the disease. People with diabetes had to rely on a crude test—urine sugar levels—to gauge their control of the condition.

Today, diabetes management has improved dramatically. But for advocates such as Kowalski, 37, a senior scientist at the Juvenile Diabetes Research Foundation (JDRF) in New York, it has not fully arrived in the twenty-first century. Nor will it, he says, until diabetics can wear a cell phone–sized device that continuously monitors their blood sugar and doses them with appropriate amounts of insulin in response.

Such a system—essentially, an artificial pancreas—is moving tantalizingly closer to reality, supported by promising results from a JDRF-funded trial involving more than 300 subjects aged 8 to 72 (N. Engl. J. Med., doi:10.1056/nejmoa0805017; 2008). The treatment group wore a continuous glucose monitoring (CGM) device, which relies on a small sensor inserted just beneath the skin that sets off an alarm when blood sugar levels get too high or too low, alerting the wearer to adjust his or her insulin intake. The control group, meanwhile, used standard, intermittent testing of their blood with finger-pricking devices to guide their self-administration of insulin.

In the six-month trial a key measure of average blood sugar levels dropped half a percent in adults using CGM devices; in their control counterparts the measure remained essentially the same. Although that reduction may sound paltry, it is, in fact, a huge improvement: a single percentage point drop reduces the risk of long-term complications by about 40% (N. Engl. J. Med. 329, 977–986; 1993).

“This is a huge step forward,” says Kowalski, who himself has been wearing a CGM device for two years and heads the JDRF's Artificial Pancreas Project. “We have lots of anecdotal experience suggesting that this technology is really transformational. What this paper did is validate what we heard.”

“This paper says we should keep that development work going on the artificial pancreas,” says Richard Bergenstal, who was not involved with the study but serves as vice president for medicine and science at the American Diabetes Association.

Beyond a CGM component, an artificial pancreas would also include a subcutaneous insulin pump—a technology that has been around for decades and is used by roughly 30% of diabetics. What's now in development is the third component: a miniature computer containing algorithms that translate the continuous readings from CGMs into commands for the insulin pump, telling it how much insulin to dose out and when.

JDRF is currently funding a seven-site study of a prototype artificial pancreas employing all three elements. On the basis of the preliminary results, says Kowalski, “it's incredibly likely we're going to see another major leap forward.”