Nutrition

July 26, 2014

Low-Carb vs Balanced-Diets: The Debate Rages On

Few nutritional topics spark more debate than the efficacy of low-carb diets. It’s therefore no wonder that a recent meta-analysis titled Low Carbohydrate versus Isoenergetic Balanced Diets for Reducing Weight and Cardiovascular Risk: A Systematic Review and Meta-Analysis is currently stirring up quite a bit of controversy in internet forums and social media outlets.

In case you’re not aware, a meta-analysis involves pooling the results of studies on a given topic to achieve clarity on the body of evidence. Here is an overview of the study with my analysis of methods and findings as well as commentary on its implications:

Inclusion Criteria
An important component of a meta-analysis is the inclusion/exclusion criteria. Simply stated, this refers to what conditions must be met for studies to be included in the analysis. For this meta-analysis, researchers required that studies had the following constraints:
• Compared a “low-carb” diet (less than 45% of calories from carbs) to a “balanced” diet (45-65% calories from carbs) that were isoenergetic (same number of calories between groups) in randomized, controlled fashion
• Comprised overweight or obese subjects
• Included macronutrient breakdowns
• Spanned at least 12 weeks in length
• Contained at least 10 subjects in each group

For statistical analysis, the low-carb diets were stratified into “high fat” (containing greater than 35% lipid) or high protein (containing greater than 20% protein). The diets were then further stratified into those where subjects were type 2 diabetics or non-diabetic. This stratification allowed for sub-analysis in a manner that helped reduce the potential confounding.

At first glance it would be fair to question the fact that “low carb” was categorized as any diet containing less than 45% of total calories. However, percentages can be misleading. The only truly relevant number here is the total grams consumed from carbohydrate. Let’s take a look at how this factored in to the included trials.

For the high-fat, non-diabetic studies there was one true ketogenic diet (4% of total calories from carbs) and the others ranged between 26-38% total calories from carbs. Given that energy intakes varied from about 1500 – 1700 calories per day, this puts total carb intake at about 97 to 161 grams/day (discounting the one true keto study). On the lower end this would put most in ketosis while on the higher end it would not. Carb intake in the high-fat diabetic studies averaged 20% of total calories, which would almost certainly translate into a ketogenic state in these subjects. The “high-protein” studies were basically all “Zone” type diets using the 40-30-30 approach. In these studies the total carb intake would have been greater than 150 grams and hence not induce ketosis. Bottom line: The “high fat” groups could fairly be classified as low-carb for the most part (at least if you pool the means of these studies) while the “high protein” groups would be more appropriately placed in a balanced category.

Overall the inclusion criteria allowed for the ability to examine an important issue on the topic, namely the effects of carb intake when total calories are kept constant. As will always happen in such situations, a number of studies that can have relevance are ultimately excluded from analysis. The extent to which this impacts results cannot be determined and the entire body of literature should always be taken into consideration when drawing evidence-based conclusions for practical application to nutritional approaches.

Analytic Specifics
A total of 19 RCTs met inclusion criteria encompassing 3209 participants. The duration of the studies spanned from 3 months to 2 years. In scrutinizing the methodology, the researchers appeared to have done a nice job collecting and analyzing data. Two different researchers were involved in the search and coding process. This serves as a double-check to help minimize the prospect of errors in data entry. They screened for various types of bias (i.e. selection, performance, detection, attrition, and reporting) and did report instances where these issues could have impacted results. The one thing I did not see mentioned was an attempt to re-code a random number of the studies to check for “coder drift” (a change in the interpretation of coding items over time). It’s unlikely that this significantly impacted results, but the possibility cannot be ruled out based on what is presented in the methods section.

Results
There were no significant differences in any of the outcomes at any of the time-points measured; weight loss was similar between all of the diets as were health-related outcomes (blood pressure, blood lipids, fasting glucose). The forest plots highlight the disparity between studies, with no trend whatsoever for superiority of one diet over another. The evidence presented suggests that when calories are equated, there is no difference in weight loss or health-related markers regardless of carbohydrate intake.

Limitations
The primary limitation of the analysis is the fact that participants did not fully adhere to prescribed macronutrient goals in a majority of trials, and adherence declined with longer time periods. Compounding matters further, self-reporting of food intake is historically inaccurate, particularly in those who are overweight and obese. Thus, the strength of evidence is compromised here, making it is difficult to formulate clear conclusions from the analysis.

On the other hand, what is reinforced from this data is just how difficult it is to stick with a diet – any diet – over the long-term. As the authors of the study point out, this is especially true in diets that exclude entire food groups such as low-carb diets (although it should be noted that adherence in the balanced diet was equally poor in the studies analyzed). With respect to weight loss, nothing is more important than dietary adherence; you can’t achieve results if you don’t follow the diet.

The other point to keep in mind is that the subjects were all overweight or obese. Thus, results cannot necessarily be generalized to a healthy, non-overweight population. Now I’d point out that those who are lean tend to be more insulin-sensitive compared to the overweight/obese, and therefore low-carb diet would seemingly have less utility for these individuals. This would be particularly true of those who are serious exercisers, as both aerobic exercise and resistance training enhance insulin sensitivity. Still, the relevance of findings to lean or athletic populations remains questionable.

Perspectives
This meta-analysis provides evidence that energy balance – not macronutrient composition – is what dictates weight loss, although findings must be interpreted with caution due to poor dietary adherence across protocols. Despite this inherent limitation, results seem to be consistent with current theory on weight loss. While ketogenic diets can be a viable approach for some, I’m aware of no evidence showing that they have a universal metabolic superiority over balanced diets provided calories and protein are equated between dietary strategies. In fact, the few studies that have investigated the topic under controlled conditions failed to show any such metabolic advantage:

• Johnston et al compared a ketogenic diet (33 g carbs) to a balanced diet (157 g carbs) in a sample of 20 sedentary overweight/obese men and women. Total protein and calorie consumption were held constant so the only thing that differed between diets was intake of carbohydrate. No differences were found in fat loss or markers of cardiovascular health. The big strength of this study was that all meals were individually prepared giving a high degree of confidence in the results. The study was limited by a duration of only 6 weeks and a small sample size.
• Soenen et al. conducted an elegant study that included four isoenergetic groups of varying protein and carb content, including groups where protein was matched but carb intake varied. The study was carried out over a 12 month period with an initial 3-month phase where subjects consumed 33% of their maintenance calories followed by a 9-month phase where subjects consumed calories at 66% of maintenance. During the initial 3-month phase the low-carb group consumed 5% of calories from carbohydrate; during the second phase carb intake increased to 25% of total calories. The average total caloric intake was not disclosed, but given the percent carb values and the fairly substantial energy restriction, it would certainly appear that the subjects were in ketosis throughout the study duration. Results? Here is a direct quote from the authors: “The study showed irrefutably, that, despite the success all-over with all four diets, the answer is that it is the relatively high-protein content per se, that supports the even greater success, and not the relatively lower carbohydrate content.”

The primary take-home message here is that there is no universal “best” diet. There is compelling evidence that higher protein intakes (at least 1.5 g/kg and generally higher in those who are lifting weights) are beneficial for optimizing body composition and enhancing satiety. A certain amount of dietary lipid is also essential for proper health, particularly with respect to polyunsaturated fats. Otherwise your approach to nutrition is largely an individual choice that, within fairly wide limits, should be based on preference, goals and lifestyle. Most importantly, calories do count!

References
Naude CE, Schoonees A, Senekal M, Young T, Garner P, Volmink J. Low Carbohydrate versus Isoenergetic Balanced Diets for Reducing Weight and Cardiovascular Risk: A Systematic Review and Meta-Analysis. PLoS One. 2014 Jul 9;9(7):e100652

Johnston CS, Tjonn SL, Swan PD, White A, Hutchins H, Sears B. Ketogenic low-carbohydrate diets have no metabolic advantage over nonketogenic low-carbohydrate diets. Am J Clin Nutr. 2006 May;83(5):1055-61

Soenen S, Bonomi AG, Lemmens SG, Scholte J, Thijssen MA, van Berkum F, Westerterp-Plantenga MS. Relatively high-protein or ‘low-carb’ energy-restricted diets for body weight loss and body weight maintenance? Physiol Behav. 2012 Oct 10;107(3):374-80