Effect of dietary fat on satiation within and between meals

Blundell et al performed a series of four experiments to determine the effect of fat content of a meal on satiety, both during and after meals. [1 full text ]

The first  experiment involved giving 16 lean, healthy subjects a standard breakfast of 440 kcal, or the same breakfast supplemented with either fat or carbohydrate calculated to provide ~360 kcal.  The standard breakfast consisted of orange juice, scones, and fruit yogurt.  The supplements consisted of either polyunsaturated margarine and cream, or a combination of sucrose, maltodextrin, and glucose.  The following table provides the data on the three types of breakfasts:

Each individual tested each breakfast with a one week interval between tests.  The subjects rated the palatability of the meals, and rated hunger, desire to eat, fullness, and prospective consumption before the breakfasts and periodically during the rest of the day.

After the breakfasts, the subjects ate measured meals, provided by the experimenters, for lunch and dinner on the same day, and kept weighed food records for the period between dinner that day and breakfast the next.

In this experiment, neither the high fat nor the high carbohydrate breakfasts appeared to exert any significant effect on intakes of macronutrients at the subsequent lunch or dinner.  However, subjective reports of hunger did differ between the high carbohydrate and high fat meals.   Specifically, the subjects reported less hunger with the carbohydrate-supplemented meal compared to the baseline or fat-supplemented breakfasts.  The following figure depicts the effects:

When the subjects ate the carbohydrate-enriched breakfast, they experienced less post-meal hunger than when given the fat-enriched breakfast, indicating that they found fat less satisfying than carbohydrate during the post-ingestive phase. 

In the second experiment, 12 lean healthy individuals consumed the same breakfasts given in experiment one, followed by a snack provided 90 minutes after the breakfasts.  Subjects rated themselves as less hungry and more full after the carbohydrate-enriched breakfasts, compared to the fat-enriched breakfasts.   They also ate smaller snacks at the 90 minute mark when they had the carbohydrate-enriched breakfast, compared to when they ate the fat-enriched breakfast.  The following figure depicts the effects:

In the third experiment, 16 lean healthy subjects consumed the same breakfasts as in experiment one, followed by a snack 90 minutes after the breakfast, or a meal 270 minutes after the breakfast.   The following table depicts the results:

When given the carbohydrate-enriched breakfast, the subjects  voluntarily ate a smaller snack at 90 minutes after, compared to what they ate after the fat-enriched breakfast.  Of interest, although the fat-enriched breakfast supplied ~800 kcal, 90 minutes after that breakfast they voluntarily consumed a snack the same size as they had 90 minutes after the 440 kcal baseline breakfast.  This means that at 90 minutes after the breakfast, the 800 kcal breakfast providing 57% energy from fat was no more satisfying than the 440 kcal breakfast providing only 10% energy from fat.  

The greater satiating effect of the high carbohydrate breakfast disappeared at 270 minutes.  Blundell et al attributed the greater satiating effect of the carbohydrate-supplemented breakfast at 90 minutes to the greater elevation of blood glucose achieved by the carbohydrate-rich breakfast.  As the glucose was oxidized or stored over the next 180 minutes, this satiating power declined.

In the fourth experiment, 12 obese women ate either one of two lunches, each on two different occasions.  One lunch supplied 527 kcal, the other supplied 985 kcal.  Between lunch and dinner, the subjects rated their hunger at one hour intervals.  At dinner, each woman was offered either foods supplying 50% of energy as fat, or 50% of energy as carbohydrate and allowed to eat as much as desired.  So, each woman had four different procedures:

1.  A 527 kcal lunch, followed by an ad libitum high-fat, low-carbohydrate (50% energy as fat) meal for dinner.

2.  A 527 kcal lunch, followed by an ad libitum low-fat, high-carbohydrate (50% energy as carbohydrate) meal for dinner.

3.  A 985 kcal lunch, followed by an ad libitum high-fat, low-carbohydrate (50% energy as fat) meal for dinner.

4. A 985 kcal lunch, followed by an ad libitum low-fat, high-carbohydrate (50% energy as carbohydrate) meal for dinner.

Not surprisingly, the size of the mid-day meal determined the course of subjective hunger during the afternoon, i.e. the smaller meal was followed by earlier return and greater intensity of hunger as depicted in the following figure:

However, the size of the midday meal did not affect the energy content of the ad libitum dinner meal as much as the relative fat and carbohydrate contents of the offered dinners.  The following table displays the impact of midday meal size and dinner composition on satiation during the dinner meal:

Regardless of whether given the high-energy or low-energy midday meal, the subjects consumed an average of 5.6 MJ/1336 kcal for dinner when given high-fat foods, but only 2.8MJ/677 kcal when given the high-carbohydrate/low-fat foods.  This demonstrated that within a meal, high-carbohydrate foods appear to have a greater satiating effect, i.e. subjects voluntarily consumed less food energy when given a high-carbohydrate selection compared to when given a high-fat selection.

Since the high-fat meal was at least 50% energy from fat, and the average intake was 1336 kcal, this means the average intake of non-fat nutrients at the high-fat meals was at most 668 kcal, approximately the same as the 677 kcal the women consumed when eating the low-fat high-carbohydrate meals. It appears as if the women were eating to achieve a certain intake of non-fat nutrients (carbohydrate or protein), regardless of the fat content of the food.  Since the high-carbohydrate meals supplied more carbohydrate and protein per ingested gram of food, the apparent carbohydrate or protein drive was satisfied with less total fat and energy intake.

The researchers followed the food intake of these women after dinner and throughout the next day as well.  Some would predict that the high fat intake of the high-fat meal would reduce post-meal and next-day food intake.  That did not happen.

 The women consumed after-dinner snacks averaging 310 kcal after the high-fat meal, and 391 kcal after the low-fat, high-carbohydrate meal.   The 81 kcal lower energy intake is insignificant compared to the nearly 700 kcal greater kcal intake at the preceding high- meal. 

The day after having the high-fat dinner, the women consumed an average of 1800 kcal, whereas the day after having the low-fat dinner, the women consumed an average of 1556 kcal.  This ~250 kcal difference did not reach statistical significance, but the direction was opposite of the prediction that they would compensate for the high-energy intake of the previous day by reducing energy intake in subsequent days.  It may even suggest that the high-carbohydrate dinner had a satiating effect that lasted into the next day. 

Now, looking at this from an evolutionary perspective:  If humans find carbohydrate more satisfying than fat, this suggests that evolutionary diets were high in carbohydrate and low in fat.  An organism geared to consuming fat would get the most satisfaction from fat.  An organism geared to consuming carbohydrate would continue eating until it either satisfied its carbohydrate requirement, or reached the limit of  its ability to convert protein or glycerol to carbohydrate, whichever comes first, regardless of “energy” intake--exactly as seen in these women. 

Since the brain regulates eating behavior and it primarily relies on glucose as its main fuel, we can reasonably expect that the brain has a carbohydrate drive, meaning that it drives people to eat until they ingest adequate glucose, or enough protein to provide the brain with adequate glucose.   

In other words, I would predict that, barring interference from the conscious mind (i.e. so-called "discipline") attempting to control macronutrient ingestion,  hungry people will keep eating until they at least minimally satisfy their carbohydrate requirements either directly from dietary carbohydrate, or indirectly from dietary protein, or until in the latter case they reach the limit the body imposes on protein ingestion, whichever comes first, regardless of total fat (or energy) intake.

Several studies appear to indicate that primitive Eskimo diets aligned with this prediction.  Several studies have indicated that Eskimos consume  very high percentage of energy as protein [Table from 2 full text ]:

On average these studies suggest that free living Eskimos derive 48% of energy from protein.  Assuming a 3000 kcal diet for an active male, this would be 1440 kcal/360 g of protein daily.  Since the human protein (70 kg reference man) requirement ranges from 50 to 75 g per day under most circumstances, these data indicate that the Eskimo may consume ~300 g excess protein daily.  According to Jungas et al, about 58% of catabolized protein will appear in the blood stream as glucose [3 ] .  Therefore, an Eskimo consuming about 300 g excess protein daily will generate from this about 174 g glucose, approximately the minimal amount required by the central nervous system. 

Some have criticized the data in the table above claiming that Eskimos eat ~80% of kcalories as fat based on claims made by Stefansson.  I find it extremely unlikely that four separate investigations produced incorrect data on Eskimo macronutrient consumption, and since Stefansson did not directly measure the macronutrient intake of Eskimos, I see no reason we should accept his estimate as more accurate.

The idea that Eskimos couldn’t have eaten a diet providing 48% of energy as protein is based on the claim that a protein intake of this magnitude will lead to so-called “rabbit starvation” from excess protein intake.  About “rabbit starvation” Cordain et al [4 ] have written:

“Excess consumption of dietary protein from the lean meats of wild animals leads to a condition referred to by early American explorers as “rabbit starvation,” which initially results in nausea, then diarrhea, and then death (39). Clinical documentation of this syndrome is virtually nonexistent, except for a single case study (42). Despite the paucity of clinical data, it is quite likely that the symptoms of rabbit starvation result primarily from the finite ability of the liver to up-regulate enzymes necessary for urea synthesis in the face of increasing dietary protein intake.” [Emphasis added]

Aside from the fact that clinical documentation of “rabbit starvation” is “virtually nonexistent,” Shaefer reported that primitive Eskimos on native diets had enlarged livers in comparison to Caucasians, and when they reduced protein intake, substituting carbohydrate, their livers reduced in size [5, full reference below].  This may suggest that Eskimos on native diets had livers adapted to chronically very high protein intakes via hypertrophy, whereas explorers (including Stefansson) may have experienced acute “rabbit starvation” when forced to eat very lean meat because unlike Eskimos they did not have previous lifelong exposure and hepatic adaptation to very high protein intakes.  

A Word About Protein And Satiety

A number of studies have shown that when given energy-restricted diets, people find higher protein intakes more satiating (within meals) and satisfying  (between meals) than lower protein diets.  For a while I felt impressed by this, thinking that protein is more satiating than any other nutrient. 

However, I now think this finding simply reflects the long-known fact that when when we restrict total food energy intake, and therefore carbohydrate intake, protein requirements increase, because carbohydrate restriction increases the use of lean mass to produce glucose.  Thus, under hypocaloric conditions, a drive to meet increased protein requirements--what we might call "protein hunger"-- may surface. 

Again, it has been known for a long time that protein requirements increase under hypocaloric conditions, so these recent studies showing higher protein diets to be more satiating under hypocaloric conditions appear to me to just be late application of something we have known for decades.  These findings do not mean that protein is the most satiating nutrient under all conditions.  I performed a quick PubMed search for studies of the satiating effects of protein under ad libitum conditions, and found only one study  [6 abstract ] which reported both "higher protein led to greater daily fullness" and "Protein quantity did not influence daily hunger, glucose, or insulin concentrations," i.e. inconsistent effects.


Given what we know about the satiating power of carbohydrate and protein, and fat balance versus energy balance, Astrup suggests that the optimal diet for reducing body fat might be very low in fat, high in carbohydrate, and moderately high in protein, for example 60-65/20-25/15 carbohydrate/protein/fat [7 abstract, 8 full text].

Take Home

1)  This series of clinical studies found that both lean and obese people not invested in consciously controlling their macronutrient intakes experienced most satisfaction and less hunger from high-carbohydrate than from high-fat meals.

2) The biological basis for this probably lies in the brain's demand for glucose.  The brain drives eating behavior, and since it prefers glucose to other fuels, it has a drive to satisfy its own requirement for glucose.

3)  Barring conscious control of macronutrient intake/ratios, a majority of people probably will eat to achieve an adequate intake of carbohydrate, either by consuming carbohydrate directly, or by consuming enough protein to produce adequate carbohydrate, continuing to eat until either they satisfy their carbohydrate drive, or they meet the limit of the body's ability to convert protein to carbohydrate, whichever comes first, and regardless of total energy intake. 

4) If people choose high-fat foods to satisfy carbohydrate requirements, they will very likely consume more fat (grams) than they can burn in a day, leading to progressive gain of fat weight.

Notes:

5. Schaefer O. Eskimos (Inuit). In: Burkitt DP, Trowell HC, eds. Western Diseases: Their Emergence and Prevention. Cambridge, MA:  Harvard University Press, 1981:114.