Is a Calorie a Calorie?
In addition to the endless Protein Controversies and Carbohydrate and Fat Controversies in the area of nutrition, another long-standing argument has to do with the question of: Is a calorie a calorie?
Simply put, the debate comes down to this: is all that matters is caloric balance (calories in versus calories out) or do the source of those calories matter? The short and simple answer, of course, is ‘No’.
The longer answer is what this article is about.
As usual, both sides of the argument can bring lots of data to the table in support of their contentions. Frequently, as you’ll see below, they end up arguing slightly different issues. In looking the topic, I want to look at three distinct data sets, each of which generates slightly different results and answers to the question. Basically, this is where a big part of the confusion comes from: people are trying to comparing data from dissimilar sets of studies and are reaching bogus conclusions because of it.
Studies Varying Protein Intake
Most commonly, when folks want to argue that ‘a calorie is not a calorie’, they will use studies comparing higher and lower protein intakes. With very few exceptions, dietes providing adequate protein intake (for dieters 1.5 g/kg lean body mass or higher would be a minimum) to lower intakes find better results than diets with lower protein intakes. This is especially apparent under dieting conditions with any number of studies support the need for higher protein intake to support muscle growth.
That is, given an identical caloric intake, the group that gets sufficient protein will generally show better muscle mass maintenance than the lower-protein group. As well, since weight losses are typically similar, that means that slightly more fat is lost. Other studies show that protein blunts hunger better (meaning it’s easier to reduce calories) than carbs or fats and a recent study showed better blood glucose maintenance in the diet containing higher protein. Aha, folks say, the source of the calories do matter!
Tangentially, I suspect that folks reporting better results from low-carb diets compared to higher-carb diets is related to this. Because of the reliance on meat, it’s nearly impossible NOT to get sufficient protein intake on a low-carbohydrate diet; folks on high-carbohydrate diets frequently overemphasize carbs to the extent that protein intake gets shorted.
The above is, honestly no surprise. Anyone who has read my books, my other articles or my forum postings knows that after setting calories properly, ‘Getting adequate protein’ is by far and away the single most important factor in setting up a proper diet. There’s just really no argument about this.
But for that reason, I tend to consider studies out of this set to be meaningless. I consider sufficient protein intake (which may be 0.8-1.5 g/lb depending on the specifics) to be assumed in any diet I’m interested in looking at. Arguments about studies comparing the inadequate RDA to what I consider appropriate amounts of protein simply don’t matter to me.
Given that built-in assumption, the question then changes slightly: given adequate protein intake to begin with, does the source of the other calories (carbohydrates versus fat) affect anything or is it simply a calorie in versus calorie out issue. In adressing this, I want to describe two other data sets.
Studies Where Calories are Rigorously Controlled
The first set of studies, which tend to be in the minority are those studies where subject’s caloric intakes are strictly controlled. These are usually the studies that the ‘a calorie is a calorie’ folks use to support their argument.
These studies are typically done by locking subjects in a hospital type of situation and measuring their food intake or by giving them pre-made food packets to use at home. Sometimes, studies are done in hospital patients being fed through a feeding tube. As you might imagine, these studies are hellishly expensive (especially if they are done over more than a few days) and, for that reason, aren’t being done as often anymore.
There is also the question of whether or not they have relevance to the real-world but that’s a separate issue. I should also mention that frequently very short-term studies (looking at a single meal or a day or two of intake) sometimes find differences for different diets but these have no bearing in the real-world where you’re looking at intakes over weeks or months.
However, in those studies, you generally see minimal (if any) differences in terms of the amount or composition of the weight lost when you vary the different nutrients. Studies have compared high to low-carbohydrate diets and even varying low-carbohydrate diets. With minor variation (maybe a pound or two here or there), any differences in the total amount of weight loss or the composition of the weight lost (again this assumes adequate protein intake in the first place) are very minor. Rather, the majority (easiliy 90% or more) of the change can be attributed directly to the caloric intake of the diet. Macronutrient composition makes a tiny, approaching negligble difference.
I’d note that recent research is suggesting that the interaction of diet with genetics may play a role here. In the article Insulin Sensitivity and Fat Loss, I examine recent research showing an interaction between carbohydrate intake and insulin sensitivity. However, this data doesn’t support that any single diet is de facto superior; only that a given diet might or might not be better for a given individual (depending on their individual genetics and such).
I should mention that studies comparing high to low-carbohydrate diets typically show greater weight losses in the low-carbohydrate group but this can generally be attributed to greater water losses. One or two studies have shown a slight trend towards greater fat loss in the low-carbohydrate group but it’s rarely huge and is invariably confounded by the issue of hugely varying protein intake. Without exception, the ‘low-carbohydrate’ group ends up eating more protein; this raises the question of whether the benefit is due to the diet being lower in carbohydrate, or higher in protein.
Of course, athletes and bodybuilders will retort that few studies are done in very lean individuals and this is very true. It’s possible that an athlete trying to get to single digit bodyfat levels might find a given diet to produce superior results but it’s poorly studied.
On that note, I have had the benefit of receiving endless feedback from athletes and bodybuilders who have compared various diets at the same calorie level. In general, differences in terms of fat loss (or muscle mass maintenance) tend to be small and highly variable.
Occasionally, you’ll find someone who loses 2-3 lbs more fat (and thus keeps 2-3 more pounds of muscle) on a cyclical ketogenic diet compared so something like the Isocaloric diet (moderate carb/moderate fat) but you can just as readily find folks who report the opposite: more muscle loss and less fat loss on the ketogenic compared to the carb-based diet. It could be genetic difference or something else causing the difference. As you’ll learn in the chapters on partitioning, factors unrelated to diet or training control the majority of what you lose on a diet in the first place.
I want to mention that relatively fewer studies have been done comparing different sources of carbohdyrates or fat. There are studies looking at the impact of sucrose (table sugar) vs. starch within the context of strictly controlled caloric intakes and they usually show no difference in fat loss or anything else. That is, given an identical caloric intake, the source of the carbohydrates shows minimal differences. Similar studies have been done with dietary fat, typically showing similarly small differences. This is especially true when calories are restricted.
Unfortunately, overfeeding hasn’t been examined in as great a detail in humans. There are studies comparing overfeeding of fat to carbohydrates (in the form of glucose, sucrose, or fructose) and, over the long-term gain in bodyfat are pretty much identical. The mechanism of the fat gain is different but, when the same number of calories are overfed, the same amount of fat is gained. I’d note that these studies rarely if ever include activity and that has the potential to impact on where calories ‘go’ quite significantly. Calorie Partitioning is its own complicated topic.
Studies looking at overfeeding of medium chain triglycerides (MCTs) or some newly developed dietary fats (diglycerols) also show some benefits in terms of decreased fat gain but the difference is typically small as well.
It’s also conceivable that at the extremes of obesity, where all manners of metabolic problems are occurring, a difference might be seen for different macronutrient composition diets. Even there, studies where calories are rigorously controlled generally show little to no difference for varying macronutrient composition in terms of weight loss or body composition. I’d note again that recent research is suggesting an interaction with diet and insulin sensitivity; given that insulin sensitivity often accompanies obesity, differential results might be seen.
I should probably mention that, in studies of weight changes, there is quite frequently a large degree of variance in weight loss or weight gain given an identical number of calories. As it turns out, and I discuss in Metabolic Rate: Overview, this ends up being more an issue of individual metabolism and how it adapts than the diet itself. That is, some people’s metabolic rate goes up (or down) more in response to over (or under) feeding. There is no evidence that the composition of the diet affects this to any significant degree; rather it’s a genetically based metabolic effect.
Summing up this mini-section: for the most part, studies where protein is adequate (or at least close to it), varying carbs and fats within the context of an identical caloric intake tends to have a minimal overall effect. What effect is occasionally seen tends to be small and highly variable (some subjects do better with one diet than another but there’s no consistent advantage). With the possible exception of extreme conditions (folks looking for super-leanness or folks who are super-obese), caloric intake is the greater determinant of results than the macronutrient composition.
Studies Where Calories are not Rigorously Controlled
As you might have guessed, these are generally the studies that the ‘a calorie is NOT a calorie’ folks refer to. In actuality, there are two different sets of studies in this group. The first is studies which are looking at nutrient intake on various diets. In such studies, subjects are simply given dietary guidelines (such as reduce fat to below 30% or reduce carbohydrates to 50 g/day or less) and intakes are examined.
Another data set of relevance to this discussion is studies comparing different diets (for example, recent studies have compared low-carbohydrate diets to the American Heart Association diet) under more real-world free-living conditions. Generally, in those studies, the subjects are given recommendations for the diet and let go. They typically report back to the researchers at some interval and frequently food intake is determined by means of self-reporting (which I’ve mentioned can be notoriously misleading). I want to look at each since both are illuminating to this discussion, as well as to how to choose a given diet.
The studies that look at average intakes given various recommendations are important because they often point to the real reason that a given diet works. For example, in studies where folks are told to reduce fat intake below 30% (or some other value), there is frequently an initial reduction in total caloric intake. That is, when they reduce dietary fat, their total energy intake generally goes down (at least initially). This is accompanied by weight loss. But this is not because of some magical effect of dietary fat, it’s simply because they are eating less calories. Of course, longer term studies show that most people end up compensating, eating more of other foods, so the result is pretty short lived.
Studies of low-carbohydrate diets tend to show similar results. Tell folks to reduce (or even remove) all of the carbohydrates from their diet and they tend to eat less automatically without thinking about it. Usually a lot less. What typically happens in such studies is that folks keep their protein and fat intakes roughly the same. So, by removing a food category that might make up 50% or more of total calories, they can’t help but eat less. Of course, this causes weight loss. But it’s not because of the carbs (or lack thereof) per se; rather it’s because they are eating less. There are other reasons, of course, such as decreased hunger (which not everyone experiences) and stable blood glucose that contribute to the reduction in calories but the removal of an entire food group is the main effect.
As I discuss in detail in Introduction to Dieting, many, many, many diet books rely on the rather simple prescription of ‘reduce or remove food X’ to lose weight. With X being something that contributes a lot of calories to the body, such as fat, sugars or highly refined carbohdyrates. But while such diet books typically use all kinds of pseudo-physiology to explain the effect, it’s really quite simple: if food X contributes a lot of calories to your diet and you remove food X, you’ll eat less total calories and lose weight. No magic, simple caloric restriction.
So let’s look at the second set of studies within the context. As I mentioned above, typically such studies look at the effect of different diets under free-living conditions. Similar to the results above, such studies frequently find that a given diet approach generates greater weight (or fat loss) but the effect is almost always due to differences in caloric intake.
For example, a study comparing a low-fat (but calorie uncontrolled) diet to a higher fat (calorie uncontrolled) diet will frequently see more weight/fat loss in the low-fat trial becuase the subjects ate less calories. The same goes for other comparisons. And while a few studies have shown drastically differential effects (such as greater weight loss at higher caloric intakes for a given type of diet), the methodology leaves a good bit to be desired. As I mentioned above, most use self-reporting of food intakes which tend to be notoriously inaccurate.
I should mention that, very frequently, the variance in weight loss tends to be humongous, as do reported caloric intakes. What this would tend to suggest is that, sometimes a certain diet type will reduce (or increase) food intake and sometimes it won’t. Individuals variance and food preferences can play a role as much as anything else.
Explaining the Discrepancy
So now, perhaps, we have a little bit better handle on why two totally different arguments about whether or not ‘a calorie is a calorie’ can come out of the resarch. The problem is that, most commonly, folks are referring to different data sets in making their argument.
As mentioned in the first section, there’s no doubt that studies comparing varying protein intakes almost always find better results with the higher protein intake. As you’ll see next chapter, all of the diets described in this book are based on adequate protein intake so these studies, have no relevance here. From the standpoint of anything I’ll ever write about or talk about, the real debate comes out of studies which keep protein constant and vary carbohydrates and fat and there are two data sets in this regards.
On the one hand are the studies were calories are rigorously controlled, where the subjects are provided their daily food intake. In those studies, differences in weight loss or body composition changes tend to be small and highly variable (some people do slightly better on one diet versus another but there’s no consistent pattern). This is the pattern I’ve observed in the real-world as well: some people do report slightly better results on one diet versus another but there’s no consistent superiority of a given approach.
Simply put, different diets are better under circumstances as discussed in Comparing the Diets.
On the other hand are studies examining spontaneous food intakes on various diets, typically examining a single diet such as low-fat or low-carbohydrate. Such studies frequently find that spontaneous food intake goes down or up given certain macronutrient intakes. For example, when fat intake is reduced below a certain point, claoric intake frequently goes down. The same occurs when carbohydrate intake goes below a certain point. Diets high in both fat (40% of total calories) and carbs frequently show higher spontaneous caloric intakes.
There is also a set of studies looking at changes comparing different diets to one another, using self-reported intakes to estimate caloric intake. While such studies frequently show differences in terms of weight loss, it’s generally related to caloric intake: if a given diet causes people to reduce calories more than another (through whatever mechanism), those people lose weight.
Is A Calorie a Calorie?
So is a calorie a calorie? Yes and no. Based on the data, my general feeling is this:
- A sufficient protein intake will always beat out an insufficient protein intake, no matter what you do. Since all of the diets described in this book are based around sufficient protein, this is a non-issue.
- Assuming caloric intake can be controlled (and protein is adequate of course), shuffling of carbs and fats tends to have a minor, approaching negligble effect.
- There might be exceptions at the extremes (folks going to single digit bodyfat or extreme obesity) but that doesn’t apply to the majority of folks.
In this respect, given adequate protein, it seems to matter very little what diet is chosen. From a weight or bodyfat standpoint, high carb should be as good as low-carb. Right? Well, no. The problem is that there’s a HUGE assumption built into statement #2 above: that calories can be controlled under a given set of conditions. As has been found repeatedly in the real-world, this simply isn’t a safe assumption.
Put a little bit differently, it might very well be possible to lose all the weight/fat you wanted on a calorie controlled junk-food diet with some high quality protein source. The problem that would probably arise is that most people wouldn’t be able to control their hunger or appetite on such a diet and they’d probably end up eating more in the long run. In eating more, they’d either lose less weight/fat or even gain it. Even if a given dietary approach appears optimal for some reason, if you can’t control your caloric intake, and end up eating more because of it, it won’t produce results.
Meaning this: you’ll frequently see folks make comparisons along the lines of ‘well, it’s easier to eat 300 calories from food X than from food Y, therefore a calorie isn’t a calorie’. They may be generally correct but this criticism is tangential to the main issue. This is why I divided the data sets into studies where calories are controlled (usually in a highly artificial fashion) and where they are not (having more real world application).
It’s obviously easier to overconsume calories from jelly beans or candy than from vegetable just as it’s easier to eat 3000 calories from butter than from celery (no human alive could eat enough celery to get 3000 digestible calories). That matters hugely under conditions where folks are allowed to eat whatever they want. Quite in fact, many many diets are based around this simple fact: make people eat less of the foods that are easy to overconsume and/or make them eat lots of those foods that are tough to overeat and they will lose weight because they automatically reduce their caloric intake. I’ll discuss that topic more in the next chapter.
But that only applies to the situation where calories aren’t being monitored. When calories are being controlled rigidly, the source of calories (whether you’re comparing carbs to fat, or even different sources of carbs and fat) matters to a much smaller degree.
Once again, my point is that if calories are being strictly controlled, the source doesn’t appear to make a humongous difference in terms of body composition changes. As well, once you get protein intake to proper levels, fooling around with carbohydate and fat ratios (within the context of identical caloric intakes) don’t seem to make a huge amount of difference either. The bottom line still comes down to calories in versus calories out; it’s simply that it may be easier to affect calories in (food intake) or calories out (through activity) with different macronutrient breakdowns.
As well, the source of calories can affect other aspects of physiology beyond body composition. Health, energy levels, hunger/appetite and all the rest interact here. So while a calorie controlled diet of jelly beans, butter and protein powder might very well work to lose weight/fat, it probably wouldn’t be as healthy compared to a diet of low GI carbohydrates, healthier oils and lean protein sources.
Understand me here? Issues such as hunger control, long-term adherence, individual variance, athletic performance, and a few others all go into the determination of what food might or might not be a better choice under a given set of circumstances. So while a calorie might be more or less a calorie under somewhat artificial conditions (where calories are or can be strictly controlled), it’s a little more complex than that in the real world. Other issues interact. The next few chapters will adress those other issues.