Wednesday, February 15, 2006

Comparing Low-Carb to Low-Fat: Analyzing a review

The key to a quality meta-analysis rests in the quality of the studies included in the review. While the media continues to headline the recently published review of studies - Effects of Low-Carbohydrate vs Low-Fat Diets on Weight Loss and Cardiovascular Risk Factors - few are questioning the methods, analysis or conclusions.

As I pointed out yesterday, the researchers failed to include an important analysis in the paper - the change in cholesterol ratio (total cholesterol/HDL cholesterol) in their review. This is an important piece of information that is missing and must be part of the equation used to determine the efficacy of low-carb diets in the short and long-term.

More importantly though is the fact that the review failed to include only studies that actually met their inclusion criteria, stated in the full-text as "To be included in this meta-analysis, trials were required to use a randomized controlled design comparing the effects of a low-carbohydrate diet (defined as a diet allowing a maximum intake of 60 g of carbohydrates per day) without energy intake restriction vs a low-fat diet (defined as a diet allowing a maximum of 30% of the daily energy intake from fat) with energy intake restriction in individuals with a body mass index (calculated as weight in kilograms divided by the square of height in meters) of at least 25. Included trials had to report changes in body weight by using an intention-to-treat analysis, to have a follow-up of at least 6 months, and to include individuals 16 years and older. We excluded trials with crossover or sequential designs."

In reviewing the studies that made the final cut, it's apparent that not all met the inclusion criteria because some failed to implement the treatment as defined in the study in one or both groups.

Case in point - Dansinger ML, Gleason JA, Griffith JL, Selker HP, Schaefer EJ. Comparison of the Atkins, Ornish, Weight Watchers, and Zone diets for weight loss and heart disease risk reduction: a randomized trial. JAMA. 2005;293:43-53. FULL TEXT - set out to compare four different dietary approaches with specific macronutrient restrictions:
  • Atkins = 20g or less carbs per day, gradual increase to 50g per day
  • Ornish = 10% fat per day vegetarian
  • Weight Watchers = 24-32 "points" (1200-1600 calories) per day
  • Zone = 30% fat, 40% carbohydrate, 30% protein

So far, so good - this trial did indeed look to compare low-carb and low-fat. But, did the groups actually follow the diets?

That is an important consideration when including a trial in a meta-analysis of comparable trials. And in this instance, this trial failed to maintain the inclusion criteria for the meta-analysis since participants across all groups in the study failed to eat according to their diet group requirements - they failed in "dietary adherance" of the trial - at all timed follow-ups in the study:

  • Atkins
    1-Month = 68g carb (+18g to +48g of restriction goal)
    6-Month = 190g carb (+140g of restriction goal)
    12-Month = 190g (+140g of restriction goal)
  • Weight Watchers
    1-Month = 1477 calories (within range)
    6-Month = 1755 calories (+155 calories of restriction goal)
    12-Month = 1832 calories (+232 calories of restriction goal)
  • Zone
    1-Month = 45:32:23
    6-Month = 45:31:24
    12-Month = 40:36:24
  • Ornish
    1-Month = 17% calories from fat (+7% of restriction goal)
    6-Month = 29% calories from fat (+19% of restriction goal)
    12-Month = 32% calories from fat (+22% of restriction goal)

Now while this study wanted to follow four different groups eating four different diets, the participants were not in compliance with their respective dietary restrictions. This study simply cannot provide us with usable data about how a low-carb diet (60g or less carbs per day) effects cholesterol when compared with a low-fat diet (30% or less calories from fat per day).

Of the six studies that "made the cut" for inclusion, we find that researchers were able to achieve participant compliance in only one study used for the meta-analysis. One other did not report the diet composition at baseline or during follow-up, so we have no way of knowing if those participants actually ate the assigned low-carb diet or low-fat diet.

When attempting a meta-analysis it's important to not only find studies that meet your intial inclusion criteria, but to also assure that you're comparing apples-to-apples data. This meta-analysis is not doing that and is therefore just another that attempts to do what is currently impossible - tell us if a low-carb diet (a truly low-carb diet) is better, the same or worse than a low-fat diet.

Did I just say "impossible"...?

Yes, I did.

Unfortunately we have no really good long-term studies comparing low-carb diets to anything else - low-fat, Mediterranean, standard American - nada, zip, zilch.

We need them. We need studies that don't just set out to have subjects eat a low-carb diet and then don't - we need those participating to actually stick to it so we can really see how a low-carb diet performs not only for weight loss, but also for reduction of risks and benefits to health.

I have to tell you though, until we move past the fat-phobia that is embedded in the collective thinking, this is going to be a difficult task - studies are compromised too often by those guiding the participants when they include advice on what to eat from dietitians who are committed to the concept that dietary fat and saturated fat is harmful, who are dedicated to the idea that restricting most fruits for a period is detrimental to health or hold that one must have grains in their diet for health.

It really isn't all that hard to pick out those studies that sought to investigate the effects of a low-carb diet, yet stifled participants attempts with fat-phobic recommendations in the process. The quick way to tell if those following low-carb were given poor advice about dietary fats for a low-carb diet...ask - what amount of total fat and saturated fat did particpants eat?

If they were instucted to follow Atkins (or an Atkins-type low-carb diet) at minimum they'll eat 60% of calories from total fat and about 15-18% of calories from saturated fat each day on 1600-calories. Cholesterol intake goes way up if one is following a low-carb diet according to the "rules" established by Dr. Atkins in his books.

Yet when we look at another study included, Brehm BJ, Seeley RJ, Daniels SR, D’Alessio DA. A randomized trial comparing a very low carbohydrate diet and a calorie-restricted low fat diet on body weight and cardiovascular risk factors in healthy women. J Clin Endocrinol Metab. 2003;88:1617-1623. FULL TEXT - we find that part of the intervention included "Two registered dietitians delivered a 3-month intervention aimed at promoting dietary compliance. Group meetings with subjects on the same diet were held biweekly on the University of Cincinnati campus and addressed cooking tips, stress management, behavior modification, and relapse prevention. On alternating weeks, subjects met for individual counseling sessions during which their assigned dietitian reviewed their 3-d food records from the previous week, analyzed by Nutritionist V (First Data Bank, San Bruno, CA), and provided dietary recommendations and positive reinforcement. Subjects were advised to continue their baseline level of activity."

With this potential to corrupt the dietary recommendations to the low-carb group - to basically cause them to eat a diet with less fat - let's see how those participating in the low-carb arm ate at the follow-ups:

  • 3-Month
    1156 Calories
    57% fat (71g fat)
    15% carb (41g carb)
    28% protein (78g protein)
    Cholesterol - 461mg
  • 6-Month
    1302 Calories
    46% fat (65.5g fat)
    30% carb (97g carb)
    23% protein (74g protein)
    Cholesterol - 285mg

After years of designing menus to comply with the Atkins diet, the above ratios and gram-intakes are IMPOSSIBLE if one is just following the recommendations in Dr. Atkins' New Diet Revolution. I can speculate what happened though - at the 3-month follow-up when the dietitians saw 461mg of cholesterol in the diet of those following low-carb, they panicked and started down the road of altering the diet to reduce the intake of cholesterol.

Want to know what happened?

From baseline to the 3-month mark, those following the low-carb diet saw their total cholesterol drop from 206 to 185, their LDL drop from 124 to 113, their HDL increase from 51 to 54 and their blood pressure drop from 116/79 to 112/72. All undeniably improvements in risk markers.

After changing their macronutrient intakes - specifically reducing their fat, saturated fat and cholesterol while increasing carbohydrate intake...those following the low-carb diet watched their total cholesterol rise from 185 to 205 (back to where they started), their LDL increase to 124 (where they started) and their HDL continue to rise to 58 (thankfully since this offset their other changes in cholesterol. In addition their blood pressure went up too - to 114/74.

We find such problems in a number of studies trying to investigate low-carb diets - those who are part of the intervention are interferring with what participants are eating because they're entrenched in the belief that dietary fats and cholesterol are harmful and therefore unacceptable in the diet at any level above what is recommended.

Part of the process of discovery is to leave your assumptions behind and let those in a trial follow the study design without trying to adjust things that we "think" might be bad. Let the data come as it will - don't interfere. As we can clearly see in the above, those who were allowed to eat high levels of fat and cholesterol were having some great improvements - when they altered that and changed how they were eating, they lost those improvements rapidly.

Low-carb diets hold so much promise - yet we can't seem to grasp that study participants should just be allowed to follow the low-carb recommendations, which include eating higher amounts of cholesterol and dietary fat. Until we finally accept the idea of doing a study and allowing participants to eat freely from the foods allowed, we're not going to answer the critical questions to determine the safety or efficacy of low-carb diets!

No comments:

Post a Comment