In our Special Topics analysis on diabetes research
over the past decade, the work of Dr. David Nathan ranks at
#7 by total cites, and #1 by cites/paper, based on 72
papers cited a total of 6,040 times. According to
Essential Science IndicatorsSM from
Reuters, his record includes 102 papers, the majority
of which are classified in Clinical Medicine, cited a total
of 8,333 times between January 1, 1999 and February 28,
Dr. Nathan is Professor of Medicine at Harvard Medical School, as well
as the Director of the General Clinical Research Center of the Diabetes
Center at Massachusetts General Hospital. He is the chairman of the
Diabetes Prevention Program, and co-chairman of the Epidemiology of
Diabetes Interventions and Complications Study, both NIH-sponsored trials
in diabetes research.
In the interview below,
ScienceWatch.com correspondent Gary Taubes talks with
Dr. Nathan about his highly cited work.
Your 2002 New England Journal of
Medicine article on the Diabetes Prevention Program (DPP) is cited
more times, by a factor of 10, than any other paper you've published in
the last decade. Do you consider this your major contribution to
Actually, what I'm probably best known for is the Diabetes Control and
Complications Trial (DCCT), which was a multi-center, NIH-funded study in
type 1 diabetic patients. But the major results from that were published in
1993, before the Special Topics analysis. We eventually published about 150
papers out of that study and many of them are among my most-cited papers.
The original 1993 paper, of which I was the first author, may be the most
frequently cited paper in all of medicine—virtually every clinical
and clinical research diabetes paper cites it. Of course, these multicenter
studies represent the work of hundreds of investigators and the invaluable
contributions of thousands of volunteers.
In the past decade, the two papers that are most cited are that DPP study,
again one of those wonderful NIH-supported opportunities to study important
clinical questions across the US, and a paper from the DCCT published in
the New England Journal of Medicine in 2005 that looked at the
effectiveness of glucose control on heart disease. (Nathan DM, et
al., "Intensive diabetes treatment and cardiovascular disease in
patients with type 1 diabetes," 353:2643-53, 22 December 2005).
"The DPP study was important because
it showed how to combat the epidemic of type
2 diabetes. The DCCT study was the first to
show that controlling blood sugar can
beneficially affect heart disease, which is a
major killer of persons with
The DPP study was important because it showed how to combat the epidemic of
type 2 diabetes. The DCCT study was the first to show that controlling
blood sugar can beneficially affect heart disease, which is a major killer
of persons with diabetes. From the science point of view, both are
similarly important; from the clinical point of view, they are as well.
How did the Diabetes Prevention Program get
started, and what question did you set out to ask?
The institute at the NIH that supports my work, and to which I am forever
indebted, is the NIDDK. Historically, it was not as active as the two other
large NIH institutes—NHLBI and NCI—with regard to clinical
trials. The DCCT was really the first great clinical trial hit for NIDDK in
the 1980-90s. With it came tremendous enthusiasm that we can successfully
perform these multicenter trials and that it was important that we do more.
After the DCCT was published, the NIDDK started thinking about what
directions should be taken next.
I was part of a small working group, along with leadership people at the
NIDDK, that looked at a number of different studies that might be done with
money put aside to do large clinical trials. Two large ones came out of the
planning. One was the Diabetes Prevention Trial 1 (DPT1) to study whether
type 1 diabetes could be prevented from occurring in high-risk individuals.
The results of that study were essentially negative—the selected
interventions didn't work.
What was the intervention used in the study to try
to prevent type 1 diabetes?
The intervention selected was insulin, not to treat diabetes, but to
prevent it. More than 100,000 young relatives—mostly siblings and
children—of patients with type 1 diabetes, who were at high risk were
identified while their blood glucose levels were still normal, but with
decreased insulin secretion. They were on the way to developing diabetes.
They were treated with insulin, injected or oral. It was a very
experimental approach to see whether it could prevent diabetes, and it
turned out that neither of the approaches worked.
And the DPP?
That was the second of the two large multicenter clinical trials that were
proposed and funded by NIDDK in the mid-1990s. It was originally called
DPT2, like type 2 diabetes, but we didn't like that name and changed the
name to the Diabetes Prevention Program or DPP. The NIDDK put out a request
for applications, and numerous universities, academic hospitals, research
centers applied. The most meritorious applications and centers were
selected with a peer-review process, and the selected Principal
Investigators subsequently designed and conducted the study. Again, these
types of studies depend on a large group of talented investigators
including, in the case of the DPP, research coordinators, dietitians and
experts in behavioral medicine, and, of course a committed and loyal group
of research volunteers. We showed that we were able to prevent
What interventions were used and how did they work
We had a control group, of course, which was given a diabetes education
that everyone received in the study. The volunteers were randomly assigned
either to lifestyle interventions—the goal was to achieve 7% weight
loss and increase activity to 150 minutes per week—to metformin, the
most commonly used diabetes drug in the world, here used to prevent
diabetes, to the thiazolidinedione troglitazone, or to placebo. We stopped
the troglitazone part of the study early, because of concern regarding the
liver problems associated with it.
Did you see in the other groups what you expected
We had a hypothesis we were testing, which was that we would see a decrease
in the development of diabetes in these very high-risk people we selected.
But the magnitude of the effect was surprising. The 58% reduction in the
development of diabetes in the lifestyle group was much more dramatic than
we expected; in fact, we ended up stopping that study ahead of time because
we got such a large effect. That was also what happened with the
DCCT—we stopped that trial ahead of time since we were seeing such a
large effect. The 31% reduction of diabetes development with metformin was
Did you continue to follow your subjects after the
When the study ended, we informed the patients of the main results. We
actually taught them all, or offered to teach them all, how to do the
lifestyle changes that had been so effective. Then we proposed a follow-up
of the study and the vast majority of the participants enrolled in the DPP
You started planning DPP well before the research
community became aware of the existence of the obesity epidemic, so
you must have been quite prescient to be anticipating a diabetes
"The current projection is that
there will be 225 million persons with
diabetes worldwide in the next 15
I like to think we were, but it was clear that an epidemic was coming. In
1984-85, the worldwide population of type 2 diabetics was said to be about
35 million. By 1995, it was clearly on the upswing, increasing along with
obesity. While we knew type 2 diabetes was increasing, we'd never have
guessed by how much. The current projection is that there will be 225
million persons with diabetes worldwide in the next 15 years.
Do you have other large diabetes studies in the
works that we should know about?
I have recently written an editorial in JAMA about clinical
research and what's coming up ("Progress in diabetes research—what's
next," 301: 1599-1601, 15 April 2009). I suggested that we now have a
wealth of information about how to treat type 1 and type 2 diabetes better
and how to prevent type 2 diabetes. The challenge is to make sure that the
interventions are being used effectively and cost effectively.
I have become increasingly focused on the public health implications of
diabetes research. We have to make certain, as best we can afford, that as
many people as possible are offered the interventions that have been proven
to work. Prevention must be given a high priority.
It requires real research to determine how to translate the results of
clinical research into clinical care most effectively and cost effectively,
for the largest number of people.
When you're looking at prevention, were you able to
show in the DPP that increasing physical activity prevented diabetes,
since physical activity was a part of your lifestyle
There's only been one study, in China, that looked at exercise
independently. That study randomized people to diet, exercise, and diet
plus exercise. The design of the study made it a little difficult to
interpret, but the investigators couldn't show any independent or added
effect of exercise. In the DPP, we could only look at the effects of
exercise alone in a secondary analysis, because we didn't have a separate
When we did the secondary analysis it looked like the major effect on
diabetes prevention was through weight loss. We did notice, though, that
those who exercised had the best weight loss. For example, the participants
who reached their exercise and weight-loss goals, they had the
best results. The people who didn't reach their weight-loss goals, but did
reach their exercise goal also benefited. Exercise apparently plays an
important, supportive role.
So these were all overweight individuals to begin
Their average body mass index was 34, which qualifies as obesity. These
were big Americans, selected specifically to be at least overweight.
Seventy percent or so had a family history of diabetes. We also recruited
from ethnic and racial groups that were known to be at particularly high
risk to develop diabetes. Our study population was a very high-risk group.
What do you consider the biggest challenge to doing
the type of diabetes research you've been doing?
The biggest challenge from the investigators point of view? I think the
biggest challenge is to create a study and a study environment in which the
research volunteers are respected and feel respected. In the really
successful studies in which I've been fortunate to participate, the
adherence and retention of our study volunteers has been extraordinarily
high—often exceeding 95% over many years. This is in contrast to many
– often less demanding studies—in which as many as 20% of the
volunteers "drop out" in the first six months.
Keeping study volunteers actively involved in studies requires planning and
effort. It is important to make certain that the subjects feel respected,
that they're involved in a critical way. If participants are just told what
to do and don’t understand why they are doing it, they may be made to
feel like lab rats. Volunteers should feel part of the research
David M. Nathan, M.D.
Harvard Medical School
Massachusetts General Hospital
Boston, MA, USA