Follow this blog:
RSS

Atkins does not extend your life — more to come

By | September 7, 2010, 7:57 AM PDT

A close friend of mine recently lost over 100 pounds on the Atkins diet.

That’s the after picture, taken in March. He’s kept the weight off.

I know because I visited his home last month. He took me to Whole Foods, bought some ground buffalo and hamburger, then proceeded to make us some dinner. He had two burgers, with no buns.

He’s a professional baker, so I was impressed with his self-discipline. He’s really a great one. (Try the scones.) He’s missing some wondrous eating by not eating what he makes.

He may also miss some years of retirement, according to a new study in the Annals of Internal Medicine.

The study group was led by Teresa Fung of Simmons University in Boston. She combined two studies, of doctors and nurses, going back up to 30 years, and matched diets to health outcomes over time.

Here is her conclusion:

A low-carbohydrate diet based on animal sources was associated with higher all-cause mortality in both men and women, whereas a vegetable-based low-carbohydrate diet was associated with lower all-cause and cardiovascular disease mortality rates.

In other words, Double Down and you’re going down hard.

Dean Ornish felt vindicated, saying he had recommended for years that if you’re going low-carb, you need to do it with veggies, not meat. CBS emphasized the higher risk from cancer (although heart-related deaths were also higher in the meat-eating group). Meat is the issue, wrote HealthDay.

Maybe. I wish my friend well. He’s a great guy. Would he have been better off fat? I don’t know — the Fung study doesn’t tell us.

And that’s the real news here. Because future Fung studies may indeed tell us.

Dr. Fung, who holds a Doctor of Science degree, specializes in analyzing firehoses of data for health patterns. A study she had published in February of last year showed a link between regular consumption of sweetened sodas and heart disease.

Until now her work has mainly involved looking at data stores like the Nurses Health Study. These are good studies, but the numbers are limited, especially when you look at narrow risk factors, and the cohort you’re studying may be fairly uniform — the nurses’ study studied nurses.

As the use of Electronic Medical Records (EMRs) expands, and EMRs carry more detail, researchers like Dr. Fung will have a much broader field of data on which to work. Studies like this will have bigger numbers, from broader populations, and they will become more reliable.

We will soon know, pretty conclusively, what factors extend life and what factors reduce it. There remains an enormous amount of noise, and some strange conclusions, in this field, like the recent study showing drinkers live longer than teetotalers. That particular study covered 1,800 people. What happens when you look at 180,000, or 18 million?

What happens is trends become clearer, conclusions become firmer. And people like Dr. Teresa Fung become more important.

Start your week smarter with our weekly e-mail newsletter. It's your cheat sheet for good ideas. Get it.

Dana Blankenhorn

About Dana Blankenhorn

Dana Blankenhorn was a contributing editor for SmartPlanet from 2009 to 2010.

Dana Blankenhorn

Dana Blankenhorn

Contributing Editor, Healthcare

Dana Blankenhorn has written for the Chicago Tribune, Advertising Age's "NetMarketing" supplement and founded the Interactive Age Daily for CMP Media. He holds degrees from Rice and Northwestern universities. He is based in Atlanta.

Follow him on Twitter.

Dana Blankenhorn

Dana Blankenhorn

Dana Blankenhorn has been a technology reporter since 1982, a business reporter since 1978, and a writer for as long as he can remember. His Schwab IRA has a few tech stocks in it, most notably some Intel and Applied Materials bought over 10 years ago. But the vast majority of his tiny fortune (emphasis on the word tiny) is invested in mutual funds. He presently writes for no one else but ZDNet, SmartPlanet and himself. But if you've got an opportunity let him know. If he takes the gig he"ll first add it to this disclosure page.

He writes for SmartPlanet and is not an employee of CBS.

15
Comments

Join the conversation!

Follow via:
RSS
0 Votes
+ -
RE: Atkins does not extend your life - more to come
A meta analysis provides direction for doing clinical trials, for more precise information. Meta analysis is put together from different studies that didn't necessarily control for the factors the analyst is trying to associate with a specific outcome. The result can not be used as a hard and fast guideline.
Posted by Stephen-Engard
7th Sep 2010
0 Votes
+ -
Stephen-Engard
I don't think this is a study meant to lead to clinical trials. How would
you design one? How could you be certain people were eating as
prescribed? And how long would the trial have to last before you
declared a result?

This is a completely different kind of thing. It's a population study.
Bigger populations with data mean more data and more reliable
studies.
Posted by DanaBlankenhorn
7th Sep 2010
0 Votes
+ -
RE: Atkins does not extend your life - more to come
You're joking, right?

More data does not in itself ensure "more reliable studies" (meta-analyses are often the most biased studies published).

But even if these results are valid (I have serious doubts), the paper itself says there is at most "a modest increase in overall mortality" from low-carb diets.

When you compare that to the mortality increases from being 100 pounds overweight, it's clear your friend made the right decision. But, then, like I said--you were joking, right?
Posted by DrPamB
7th Sep 2010
0 Votes
+ -
DrPamB
The difference in mortality between low-carb and low-fat diets lies in
those low-carb diets which depend on animal protein. That was the
study's conclusion, looking at 84,000 records.

If they look at 800,000 or 8 million records, the conclusion can be
nailed down with more certainty.
Posted by DanaBlankenhorn
7th Sep 2010
0 Votes
+ -
What about privacy?
What you did not say is that the 230,000 nurses participating in the Nurses Health Study are all volunteers. They gave permission for their medical data to be used.

Are you saying researchers should have access to EMRs without the participants' permission? If not, how would you entice people to participate? One could argue that altruists who volunteer for studies such as these may be more inclined to take better care of themselves, thus biasing the results.

And what about using the vast store of data collected from grocery store loyalty cards? For most people (it must be at least 100 million people) their grocery stores have a fairly complete record of what they've eaten for at least 10 years. Combine this with the data from EMRs and you could probably find some unsuspected correlations between diet, disease, and overall health. But it's unlikely many people will volunteer the use of their loyalty card data. How do you balance privacy against the huge potential benefits?
Posted by zackers
7th Sep 2010
0 Votes
+ -
zackers
Gross data is easy to anonymize. Any number below a certain
threshold isn't reported. That's already the case.

As the number of records in the pot grows, this becomes less of an
issue. You can raise the threshold without losing valuable data.
Posted by DanaBlankenhorn
8th Sep 2010
0 Votes
+ -
...what factors extend life...
The author asserts "...We will soon know, pretty conclusively, what factors extend life and what factors reduce it..."

I doubt it. For over 60 years, the popular research has been concluding that high carbs, low fat, low cholesterol are the recipe for health. During that same time, the nation's population has ballooned (about 1/3 are over-fat or obese), diabetes has sky-rocketed from 0.5% to ~30%, and heart disease, cancers, and a host of other ailments and diseases, have multiplied.

Today's popular beliefs also include the "I can't change because it's genetic" mantra. For those who are receptive to information that shows how we got into this mess, and why we believe fat is bad, cholesterol is bad, salt is bad, etc., get Gary Taubes "Good Calories, Bad Calories."

Do what you believe is "right" for you, and you will get the logical results.
Posted by SteveMak
8th Sep 2010
0 Votes
+ -
SteveMak
That's not what I meant at all. I meant that the firehose of data which
will emerge from EMRs, when matched to systematic research as
found in this study, is going to detect patterns that are definitive on
many questions we have only prejudices to answer for today.
Posted by DanaBlankenhorn
8th Sep 2010
0 Votes
+ -
You cannot anonymize records enough
@DanaBlankenhorn: Sorry, but it's become well-known that only a few items of information are enough to identify someone, no matter how big the pot or how careful the attempts at anonymization.

For example, the cookies on your web browser (which can be read by any website) uniquely identify that browser and by extension the person doing the browsing. This is true even if you exclude cookies that contain explicit identification of the person. See http://conferences.sigcomm.org/sigcomm/2009/workshops/wosn/papers/p7.pdf .

You may remember that Netflix ran a famous challenge where they released an anonymized sample of about 500,000 users' movie preferences so people participating in the challenge could refine Netflix's algorithm for making movie suggestions. Netflix was going to do a second challenge, but withdrew it when researchers from UT Austin showed they could "reidentify" people from the data of the first challenge ( http://userweb.cs.utexas.edu/~shmat/shmat_oak08netflix.pdf ).

Given that there is no real anonymity in large sets of "stripped" data, do you still support use of these sets without the users permission?
Posted by zackers
8th Sep 2010
0 Votes
+ -
...what factors extend life...
@DanaBlankenhorn replied "That's not what I meant at all. I meant that the firehose of data which will emerge from EMRs, when matched to systematic research as found in this study, is going to detect patterns that are definitive on many questions we have only prejudices to answer for today."

It is exactly those patterns, or "correlation", that have given rise to may of today's beliefs. Humans are outstanding at detecting patterns. So good, that they will perceive patterns in noise. And so, with this ocean of data, we will detect even more correlations, and come to still more false conclusions.

The issue in all this, is that most of us are not practicing Good Science. Our beliefs, even in the scientific and research circles, leads us to poor practices and incorrect conclusions. The "patterns" you speak of, will establish correlations, and still not prove *causality*. Shake. Repeat. More of the same.

The more firmly entrenched a belief, the more resistant we are to receive information to the contrary. This also exists in the scientific and research circles. That's why today, most North Americans still believe that eating fat makes us fat, eating cholesterol increased serum (blood) cholesterol, salt is bad and increases hypertension, and so on.

To explain this in a different way: Data + Process = Conclusion. Our process is severely flawed, due to well-intentioned clinging to factually incorrect beliefs. Add more data, using the same process, and we still get garbage conclusions.
Posted by SteveMak
9th Sep 2010
0 Votes
+ -
SteveMak
There are problems with scientists and the scientific method. But
there are opportunities in what you say, for other scientists to use
the method and the growing body of data to draw conclusions you
say may be more correct. Assuming they can be found. And you can
defend your methodology. Before scientists.
Posted by DanaBlankenhorn
9th Sep 2010
0 Votes
+ -
RE: Atkins does not extend your life - more to come
Thanks to all in the discussion above- One of the most enjoyable & discerning serial presentations re the logic & ethics of research methodology I've had the pleasure to read online.
Posted by kmarin@...
9th Sep 2010
0 Votes
+ -
RE: Atkins does not extend your life - more to come
This study was based on the notoriously unreliable "food questionnaire" approach. Is a firehose full of bad data preferable to a trickle?

I also think that your headline badly misrepresents the thrust of the study. Fung herself is on record as stating that the main result of her study is that a heavily meat-laden low-carb diet is less healthy than a more plant-oriented low-carb diet. Atkins simply advocated low-carb diets in general, and even had sections in his books for low-carb vegetarians.

On the other hand, you seem to have a bit of an axe to grind. I note that all of your articles on the topic on this site appear to be anti-low-carb (which you characterize as "Atkins"). You never seem to mention truly paradigm-shifting research, like the recent studies of Volek, et al, who found all the markers of cardiovascular disease are lowered more on low-carb diets than on low-fat diets.
Posted by DrPamB
11th Sep 2010
0 Votes
+ -
RE: Atkins does not extend your life - more to come
You also don't mention that Fung's study found no significant health problems resulting from low-carb dieting relative to high-carb dieting.

Readers also might like to note the editorial that was published alongside Fung's study. In this study, even the upper limits of the confidence intervals are in the "lower range of clinical importance," they wrote, and it didn't show a "clear dose-response relationship in that there was not a clear progression of risk moving up or down the diet deciles."

They said the study "addresses a critical, unresolved public health question of diet but cannot satisfy us with a definitive answer," and called for a large-scale clinical trial "to provide a more definitive answer to the largest public health crisis in the U.S.: the effect of diet on obesity, chronic disease, and mortality."

"The current state of the evidence," they wrote, "is such that no one can legitimately claim that a low-carbohydrate diet is either harmful or safe with any degree of certainty until a large-scale, randomized study with meaningful clinical endpoints is done."
Posted by DrPamB
11th Sep 2010
0 Votes
+ -
DrPamB
Thanks for your posts. You can learn a lot from links. That's why I
use as many as possible.
Posted by DanaBlankenhorn
13th Sep 2010
Join the conversation
Formatting +
BB Codes - Note: HTML is not supported in forums
  • [b] Bold [/b]
  • [i] Italic [/i]
  • [u] Underline [/u]
  • [s] Strikethrough [/s]
  • [q] "Quote" [/q]
  • [ol][*] 1. Ordered List [/ol]
  • [ul][*] · Unordered List [/ul]
  • [pre] Preformat [/pre]
  • [quote] "Blockquote" [/quote]

Join the SmartPlanet Community and join the conversation! Signing-up is free and quick, Do it now, we want to hear your opinion.