Skip to main content

That diabetes study that ended early ...

Did you hear?  From the New York Times last week:  Diabetes Study Ends Early With a Surprising Result.  Thus far there's been some of the usual responses from the low carb community, and I expect more to come.  For starters, the writer is Gina Kolata, a woman reviled by Gary Taubes and thus his adoring fans, so we'll get the usual round up of Gina doesn't get it that low carb is the bestest and only way to treat diabetes ... duh!!   Let's look a bit closer, shall we?  

First, keep in mind this study has not been published up, so we're dealing with press releases and quotes here.  Secondly, ending the study early makes it sound quite a bit more dire than what we know to have happened.  The study was apparently to last 13 years (odd duration, that), but was ended at 11 years because no differences were seen between the study and control groups over that time.  Sounds prudent, because likely they looked at the numbers and probably concluded that there would need to be an unfathomable number of incidents/deaths in the remaining two years to alter the results.  Save the money.  Good.

Let's look at the study:  

  • 5,145 overweight or obese people with Type 2 diabetes, randomly assigned to intervention or general health counseling sessions.
  • Intervention:  Diet (12-1500 cal/day for those under 250 lbs, 15-1800 cal/day for those over 250 lbs) and Exercise (~ 3 hrs/week, moderate -- presumably aerobic -- exercise).  
  • Control:  General health counseling
After 11 years, from the article:
... the intensive program did not help.
... the two groups had nearly identical rates of heart attacks, strokes and cardiovascular deaths.

“I was surprised,” said Rena Wing, the study’s chairwoman and a professor of psychiatry and human behavior at Brown University’s medical school.
Like many, she had assumed diet and exercise would help, in part because short-term studies had found that those strategies lowered blood sugar levels, blood pressure and cholesterol levels.
OK, so my first question is this:  How old were the participants? We're all aware of the statistics we're reminded of in the article, that diabetes roughly doubles (or a bit more) risk of CVD. But if they weren't following a population in which we see significant amounts of such events to begin with, it's no surprise to me when no difference is seen regardless of the intervention.
But assuming this was in a population in which we would expect to see enough events over a decade or so to see meaningful differences, let's look a bit further.
The study participants assigned to the intensive exercise and diet program did lose about 5 percent of their weight and managed to keep it off during the study.  That was enough to reduce cardiovascular risk factors.
Since the diet prescriptions included a weight demarcation of 250 lbs, let's use that number as an "average" participant.  Five percent of 250 is 12.5 lbs.  So the average person assigned to the "rigorous intervention" went from obese to less obese at 237.5 lbs.  Those who were less obese to begin with, say someone weighing 180 lbs to start lost a whopping average of 9 lbs.  Perhaps this study can put an end to the "lowering the bar of expectations" practice of hyping the trade-offs from nominal changes in body weight.   So these people are being congratulated for accomplishing very little.
I can hear it now, this just goes to show how useless ELMM is ... gotta go low carb!!  Does anyone out there truly believe that if your average 250 lb person complied with a diet of 1500 cal/day for a decade would still weigh nearly as much?  It sounds like this study included a diet phase early on (6 months perhaps?) during which the participants lost weight, and then maintained the losses.  Did they continue the exercise?  These are details I'd like to see when this study is eventually published up.  How many are even following the "rigorous program" to any degree?  
We know from countless short term studies that compliance varies greatly. Drop-out rates can be as high as 50% with 30% not being unusual.   In the rather more rigorous Shai study,  the six year followup offers further insights.  In that study, the diets were partly administered for two years in a workplace environment.  As I recall, the cafeteria had foods labeled as appropriate for the particular diet and/or with information readily available so participants could make easily informed choices.  Presumably even if this labeling stopped, these participants could still recognize their foods improving the likelihood of compliance.  We learn that under those conditions ~95% completed the 2 year study, but the adherence rate was only 85%.  So was that 85% of the completers which would be roughly 80% of the initial group.   They were able to follow up with 95% of those who completed at two years which translated to roughly 80% of the original participants, even if we presume 85% compliance amongst these, we're now down under 70% of the original participants under these rather more ideal conditions.  Since all groups regained some weight, it's fair to say that compliance was likely nowhere near 85% at the six year point.  So imagine what the compliance might be in this study we're talking about, a decade plus out?
I'd be surprised to learn that the current study was worth the investment at all.  I'm thinking not.  Here's another thought.  Imagine you're an overweight/obese diabetic and have been recruited for a study and are assigned to the general health counseling group.  No doubt you are further apprised of all the negative impacts out-of-control diabetes can have on your health.  This "control" group didn't do nothing, they may well have undertaken lifestyle changes on their own.  Perhaps they lost an average of 2% of their original weights.  A control group instructed not to make lifestyle changes would be unethical, so this is not really a true control group after all.
In any case, and again we need to wait for the study publication to get some exacts on this, later in the article we learn:
... the results meant that people with diabetes might have a choice. The group assigned to diet and exercise ended up with about the same levels of cholesterol, blood pressure and blood sugar as those in the control group, but the dieters used fewer medications.
I'd say that's a good thing, no?  So the dieters were able to achieve the same result with fewer medications.  Isn't that what all the low carbers are after?  Isn't that what most of those requiring medications are after?  I'd think so.  
You know what would be interesting?  An analysis of this study data looking at change-in-weight vs. outcomes.  If the average weight loss amongst dieters was 5%, no doubt some lost a lot more and others lost none or even gained weight.  Also no doubt some of the control group undertook their own lifestyle changes (maybe even low carb!) during the ensuing decade.  What I would hate to see is folks getting the take home message some of the low carbers (usual suspects folks, no need to link) are piling on along with the mainstream -- that diet and exercise are ineffective treatments for diabetes.  I'm going to call bull and shit on that one.  
A final note:  Is anyone surprised by these (sketchy) results?  I'm not.  Lifestyle change is difficult ... most of us here oughta know.  Short term studies are always promising for those who complete them and comply.  While weight is not an ultimate determinant of health, excessive fat is simply not health improving for anyone.  It's not necessarily detrimental to all, or equally detrimental for that matter, but it's unlikely to improve health (and note the word excessive, not a little extra).  To me the take-away message (as of now) from this study is that lifestyle change with some medications can be equally effective as more medications.  The low carbers who believe their lifestyle change is superior should seize on that and continue to promote low carb, but keep those Shai results in mind.  On average ....


Unknown said…

Clinical Trial

"5.4.5 Diet

Restriction of caloric intake is the primary method of achieving weight loss. In order to aim for a weight loss of 10% of initial weight, the calorie goals are 1200-1500 kcal/day for individuals weighing 250 lbs (114 kg) or less at baseline and 1500-1800 kcal/day for individuals who weigh more than 250 lbs. These goals can be reduced to 1000-1200 kcal/day and 1200-1500 kcal/day, respectively, if participants do not lose weight satisfactorily. These calorie levels should promote a weight loss of approximately one to two lbs/week.

The composition of the diet is structured to enhance glycemic control and to minimize cardiovascular risk factors. The recommended diet is based on guidelines of the ADA and National Cholesterol Education program96,97 and includes a maximum of 30% of total calories from total fat, a maximum of 10% of total calories from saturated fat, and a minimum of 15% of total calories from protein."

"During the first four weeks of the intervention, participants are encouraged to follow a portion-controlled diet, given findings that this approach produces significantly larger weight losses than having participants consume a self-selected diet of conventional foods."

"Participants choose from two prototype diets. The first includes the use of a commercially available liquid meal replacement that will replace two meals and snacks each day. This regimen is combined with an evening meal of either a frozen entrée or conventional table foods to provide a total of 1200-1800 kcal/day depending on the individual’s baseline weight. The second option, for those who do not accept or tolerate the liquid/prepared meal prototypes, involves the consumption of a very structured meal plan, with the same calorie range, using foods that participants prepare themselves."

Unknown said…

Curriculum Vitae - Rena R. Wing, PH.D

Trained in Psychology with Postgraduate work in Psychiatry
Paul Jaminet said…
Good point about the change in weight vs outcomes, that would be interesting. Hard to say anything about this trial based on the press releases and news stories. I haven't seen any useful data yet.
Anonymous said…

Doesn't this mean that the participants were followed for up to 11.5 years, but the intensive support was only for 4 years?

Here's the four-year report:
Paul N said…

Peter at Hyperlipid also wrote about this study.

more information here;

which included the age of the participants - 45 to 76 at the start of the study.

One of his commenters provided a link to an interview with the study leader;

From that website:
"In addition, early data showed that treadmill fitness levels, hemoglobin A1c levels, systolic and diastolic blood pressure, HDL-cholesterol levels, and triglyceride levels were all significantly improved among patients in the lifestyle-intervention arm when compared with the control group."

So, we have a trial that showed a modest weight loss, an "improvement" in a bunch of markers, yet all of this, after 11 years, made zero difference to the number of CV events.

But the original goal of the program was, *specifically* to improve CV health, and it didn't.

You can download the entire study protocol (118pages!) from the lookahead site.. In it is this statement (p9);

"The primary hypothesis is that the incidence rate of the first post-randomization occurrence of a composite outcome, which includes cardiovascular death (including fatal myocardial infarction and stroke),
non-fatal myocardial infarction, hospitalized angina, and non-fatal stroke,
over a planned follow-up period of up to 13.5 years will be reduced among participants assigned to the Lifestyle Intervention compared to those assigned to the control condition, Diabetes Support and Education."

And the diet? From p30;
"...The composition of the diet is structured to enhance glycemic control and to minimize cardiovascular risk factors. The recommended diet is based on guidelines of the ADA and National Cholesterol Education program and includes a maximum of 30% of total calories from total fat, a maximum of 10% of total calories from saturated fat, and a minimum of 15% of total calories from protein."

So lets be clear, this study, carefully designed to minimise CV risk factors, *didn't* make any difference to CV risk!

Will the study authors admit then that their primary hypothesis is disproven?

Will they also admit that the diet recommended by the ADA and the NCE program doesn;t improve CV health?

And that neither did their secondary markers (improving fitness, HDL cholesterol, lowering blood pressure, eating less saturated fats, etc etc)

And what are the implications of this for addressing CV health?
Basically, the current methods, officially endorsed by the various health bodies, don;t work...

So, if this trial shows that the current accepted methods for managing CV health don't work, then I think that is exceptional value for money. Now they can change their thinking and start looking for methods that DO work.

No wonder they are ending the trial early.

Be interesting to see if the publish all their data...

Unknown said…
Another round in the Study Wars, all participants will step ten paces, turn around at the count of three and reveal their studies.

I think improving fitness is a pretty good thing in and of itself, whether it has an effect on CV, who the hell knows? It think it is a difficult thing to study due to the myriad factors you have to deal with. Also it is probably illegal to force-feed people with Big Macs and chicken-fried steak.

But there's a lot to be said for being able to function better at the physical level. I think there is a valid argument to be made with the "Eat Less" part of the equation but I don't I think there are any problems with the "Move More" part, I doubt anyone on their deathbed ever thought "I really wish I had moved less."
Unknown said…


"The basic flaw in the Look AHEAD study was that it was designed to bring about weight loss, and hoped that weight loss would improve health.

A better intervention would seek to improve health through a more PHD-like diet and through circadian rhythm therapies. Successful health improvement would, more than likely, lead to weight loss.

For the overweight and for diabetics, the focus should not be on weight, but on health. Improve health, and weight loss will follow. Focus on weight with a simple-minded “eat less, move more” intervention without tending to the quality of your diet and lifestyle, and you might be doing yourself more harm than good."

Unknown said…
"Participants choose from two prototype diets. The first includes the use of a commercially available liquid meal replacement that will replace two meals and snacks each day. This regimen is combined with an evening meal of either a frozen entrée or conventional table foods to provide a total of 1200-1800 kcal/day depending on the individual’s baseline weight. The second option, for those who do not accept or tolerate the liquid/prepared meal prototypes, involves the consumption of a very structured meal plan, with the same calorie range, using foods that participants prepare themselves."
Anonymous said…
This is what comes to mind when I read R. Wing's name:

I am not sure I am reading this correctly. It looks like you may have the ethical associations switched.

Unless a food has been proven to be dangerous to health, there is nothing unethical about feeding a person Big Macs or chicken fried steaks.

In contrast, an exertion heart attack is a known phenomenon. The annual wave of heart attacks set off by people shoveling snow in winter or mowing the lawn in the summer is a known phenomenon.

I would think it is far more unethical to send someone out to "move more" than to simply "eat less".
If I remember correctly, the Minnesota starvation experiments of Ancel Keys had the men eating 1800 calories a day for 6 months with sometimes horrific consequences. But these people were supposed to stick to 1800 cal/day for 13 years?

Did it ever occur to anyone that an overweight person may be even MORE sensitive to the stressful effects of starvation than a normal weight person? I mean, that could be a possible reason they got to be overweight in the first place.

And where did they find people over 250 lbs. with so little dieting experience that they actually thought they could stick to eating 1800 cal/day for 13 years? Maybe there are more of those people than I realize.

The longest I ever lasted on a diet was eating 1800 cal/day. I lasted for 5 months. I lost about 50 lbs. I also ended up having my gall bladder removed. I subsequently found out that needing a cholecystectomy after prolonged calorie restriction is not that uncommon. I was not aware of that and I would have been PISSED if that had happened during a medical study.
Mir Writes said…
Every time I see a weight study where, when you look at the small print, the people lost maybe 4 to 10 pounds, I'm hardly impressed. I'd like to see the studies helping folks lose significant weight and then monitoring that longer-term to see effects on health markers, sure, but also to see WHAT factors are most key in achieving and maintaining the necessary lifestyle changes.

I've lost 42% of my highest weight (300) and stopped being obese last summer. I maintained the loss for a bit over a year now...but I am still overweight, though not obese. I had significant improvements in various medical and quality of life markers (though LDL seriously sucks, still). However, the fearful cloud over my head is the dread odds re regain (though statistics seem to be unclear, the consensus from experts "out there" is that MOST regain).

Still, 5 pounds weight loss trials don't even make me bat an eye. Puhlease. On what weight-challenged person would that even be significant for health makeover purposes?

Mir aka Princess Dieter
Mir Writes said…
I regain on 1800 calories. That may be starvation for some, but that's maintenance or REGAIN level calories for others. There's one blogger monitoring intake, and she regains over 1400....

I have to stay around 1600-1700 to maintain, and I'm overweight. Maybe it's the Hashimoto's, maybe it's middle-age, maybe it's just normal for women not to be able to eat a whole lot unless quite athletic. But 1800 is not starvation level for moi. It's the beginning ground of "fatten up" levels.

I ate as 1200 for several months, mostly at 1400 in my losing phases. I'm maintaining at 1600 or so. And that's since 2 years ago plus. If I don't keep it up, I'll be obese again in the blink of an eye.

I was aware about the gallbladder issues on extended lowfat diets, hence i made sure to have olive oil, butter, and coconut oil in my diet. It meant less actual bulk food, sure, cause of the concentrated calories in fat, but I didn't want to lose my gallbladder, and I'd seen the theory that it was due to lowfat (not causing gallbladder to secrete, etc). Maybe that helped. My mom had hers out, so I worried.

Most of the successful weight-losers that I've seen online (ie lost and are still keeping it off) eat not a whole lot of calories. I've seen as low as 1200 for maintenance in a few to 1800 in others, but a lot seem to be around 1500-1600. I know that one famous The Biggest Loser winner maintains at around 1400 calories, and she exercises.

It's grim.
Susanne said…
Yes, I know very well what 1500 calories is from my own tracking, and that is what struck me most from the NYT article. It's very hard to imagine that with strict compliance, over 11 years, that the average weight loss was so low. There are always going to be outliers, but most 250lb bodies are going to shed fairly notable amounts on that -- more than 5% -- and settle at some lower level. And if they are exercising and eating sufficient protein, some of that at least will be fat.

And yet none of the authors' comments included in the article mentioned potential compliance issues: ? I wonder too if that will appear in the final article, or if the authors are just that clueless. But for me right now the takeaway message from this is more, "It's really hard to lose weight and maintain the loss" rather than "Diet and exercise doesn't work for CVD."
Susanne said…
LOL. Because "A Trial in Reducing CVD Risk Factors by Attempting to Get Participants to Comply With Either a Slim-Fast-Based(TM) or a Lean-Cuisine-Based(TM) Diet for Thirteen Fricking Years Fails to Produce Significant Results," is not an article title which looks good for any of the parties involved. :)
That is shocking!

What in the world could be causing the metabolism to slow down so much that a person has to limit themselves to 1400 calories/day AND exercise just to be a normal weight? That is crazy!

The idea that one would have to count calories just to have a normal body size is absolutely an indication that something is wrong. To further have to reduce calories to such a low amount is even MORE indication that something is wrong. That is so low that it would be very difficult to get proper nutrition without careful planning. That is unnatural and clearly pathological.

This is freaking insane!
CarbSane said…
You made me spit my coffee Susanne! Yeah, euler, it seems rather vague what the duration of the actual intervention was the diets listed were for four weeks, and then they transitioned to real foods? From the 4 year report (, we see that they gained back almost half the 1 year losses by the 4 year mark and there's no reason to believe this trend didn't continue. The averages obscure the fact that 25% of the control group lost over 5% of their weight, just over half as many as the diet group.
CarbSane said…
Paul, I don't think this study will tell us anything unless there's some attention paid to compliance rates. Only half of the diet group had maintained a piddly 5% reduction in weight at the 4 year mark, while about a quarter of the "control" group also lost that much weight.
CarbSane said…
I think you're getting a little sarcastic here, eh? My point was that it would be unethical to encourage a true control group to just change nothing knowing the risks.
CarbSane said…
I tend to disagree with Paul on this. I do not believe that switching to a PHD-style diet would improve health and weight loss follows health improvements. I'm not saying one cannot improve health without weight loss, actually it might be more important to eat a good diet than lose a nominal amount of weight, but this is presuming all individuals were eating a crappy diet to begin with.

I cringe to see Paul citing Peter's nonsense speculation and expanding on it to this notion that ELMM might do more harm than good. Yes, ridiculous crash dieting and overexercising, but not what's being suggested here.
CarbSane said…
The Minnesota experiment is irrelevant. That study involved young men who were lean to begin with, and cut their intake in half to starve them to 25% below normal body weight. Their diet was almost entirely carbs with very little protein, but likely sufficient fat and they were made to exercise a lot. Speaking as a formerly 250 lb person, the "exercise" in this study would be otherwise known as normal activity for the thinner me. According to NHANES, the average woman in the 1970's ate around 15-1600 cal/day. Even if under-estimated, that's what it was. When you add up normal amounts of food prepared at home, that does sound about right.
CarbSane said…
Hey Mir ... Delving just a little deeper we learn that 55% of the diet group did NOT maintain even this piddly 5% loss. I hate that fearful cloud. I think I'll write a blog with some further thoughts on these studies in general.
Unknown said…
The guy who has a heart attack shoveling snow is a perfect example of why the "move more" part is non-debatable.

What is the benefit of "move more"? The benefit is that you maintain the ability to move without cause undue stress or injury. This benefit accrues regardless of whether you lose weight, gain weight or remain at a constant weight.

It is particularly important if you a overweight, because the extra weight makes physical activity more stressful. You have to work to retain your mobility.
CarbSane said…
Interesting that the chairwoman of this study is better known for the National Weight Control Registry. That recent JAMA study (you remember the one where VLC burned more calories but they defied scientific laws by maintaining their weight) the participants lost almost 15% of their body weight in 12 weeks eating roughly these caloric levels (perhaps even a bit more) with no exercise. So, I'm calling foul on compliance with anywhere near the prescribed intervention.
Anonymous said…
'The primary objective of the Look AHEAD clinical trial is to assess the long-term effects (up to 11.5 years) of an intensive weight loss program delivered over 4 years in overweight and obese individuals with type 2 diabetes.'

That's the primary objective.

So, if you lose weight and then regain it back, what's that going to do to your CVD risk? If you don't regain it, what would be the impact on your CVD risk? Would that be the point of the study? Or is this another study about the difficulty of maintaining weight loss?

The Weight Control Registry (R Wing was a founder) tracks the success of people who lost weight and maintained: the results have been that more physical activity and keeping calorie intake in check have been the winning factors there! The one flaw in taking any of this to heart is that the registry is not a randomized control trial. The people who participate have their doctors' cooperation (confirming the weight loss and maintenance), and the participants are also a pretty motivated bunch!

The Look AHEAD study did randomize type 2 diabetics into two groups. After one year:

'At 1 year, ILI resulted in clinically significant weight loss in people with type 2 diabetes. This was associated with improved diabetes control and CVD risk factors and reduced medicine use in ILI versus DSE. Continued intervention and follow-up will determine whether these changes are maintained and will reduce CVD risk.'

(ILI is lifestyle intervention. DSE is diabetes education.)

'The primary study outcome is time to incidence of a major CVD event. The study is designed to provide a 0.90 probability of detecting an 18% difference in major CVD event rates between the two groups. Other outcomes include components of CVD risk, cost and cost-effectiveness, diabetes control and complications, hospitalizations, intervention processes, and quality of life.'

At the very least, there will be a lot of data about maintenance of weight loss. Doesn't sound like a waste of time to me. One year results don't count when you consider the primary objective. However, there's still a lot to learn from the data they have obtained.

ProudDaddy said…
I often comment here that we can sometimes learn as much from the outliers as the averages. In this case, it would appear that we might be able to learn more. The correlation that Evelyn suggests might prove to be the only new knowledge this study can produce. When are the guidelines for clinical studies going to insist that the raw individual data be made available for everyone to analyze?

On a personal note, I might not be commenting much in the future - Mrs. ProudDaddy is pregnant! At 71, I'm not planning on gaining 40 pounds as I did during "our" last pregnancy, but...
CarbSane said…

I think studies like this just scream for sharing of raw data, at least after some period of time.

@euler: I would agree that there's lots of great data here depending on the reporting of counseling sessions as well. It's just pretty useless data, IMO, to assess the primary objective. If they're going to look at 7-10% weight loss, they really have to look at that and only that.
Blogger said…
EasyHits4U - Your Free Traffic Exchange - 1:1 Exchange Ratio, Business networking. Get FREE Advertising!