Please note, this is recovered content from the former website of the New York Obesity Research Center website.


Producing Positive Satiety Effects With Protein Intake Concerning Anorexia Nervosa, Bulimia & Binge Eating Disorder

Date: Thursday, September 12th, 2002
Title: “Macronutrient Effects on Satiety & Binge Eating in Bulimia Nervosa & Binge Eating Disorder”
Speaker: Janet Latner, Rutgers University, New Brunswick, NJ

bulimiaWomen with bulimia nervosa (BN) and those with binge eating disorder (BED) have marked disturbances in satiety. For example, they consume greater food intake than controls throughout the day (Rossiter et al., 1992; Weltzin et al., 1991) and during binge eating episodes compared to episodes of controls asked to binge eat (Hadigan et al., 1992; Yanovski et al., 1992).

In addition, women with BN report lower satiety after meals than controls (Geracioti & Liddle, 1989). These findings are consistent with disturbances in physiological indices of satiety in these women, such as the blunted release of the satiety agent cholecystokinin following food intake (Geracioti & Liddle, 1989) and increased gastric capacity (Geliebter et al., 1992) found in women with BN.

Another difference between women with BN or BED and controls is in their dietary selection. Women with these disorders have been found to consume lower proportions of protein during their binge episodes compared to the episodes of controls asked to binge eat (Walsh et al., 1989; Yanovski et al., 1992).

Women with BN also ingest lower proportions of protein during binges compared to their own non-binge episodes (Van der Ster Wallin et al., 1994), and in the overall diet compared to controls (Hetherington et al., 1994). Women with BN begin binge episodes with dessert and snack foods, whereas controls begin binges with fish and meat (Hadigan et al., 1989).

These abnormalities in dietary selection may have particular significance in light of the different satiating effects of the macronutrients in non-eatingdisordered individuals. Protein leads to greater suppression of food intake and greater reported fullness than other macronutrients (Latner & Schwartz, 1999).

For example, 12 female college students without eating disorders were given three 450-kcal liquid lunches on non-consecutive days, high-protein (71%), high-carbohydrate (99%), and an equal mixture of these two liquid lunches, in counterbalanced order. Participants were asked to eat as much as they felt like at a buffet-style dinner 4 hours later. They consumed 24% less at dinner after protein, and 17% less after the mixture, than after carbohydrate. Greater hunger and excitement about eating before dinner were reported in the carbohydrate lunch condition compared to the protein lunch condition.

Greater enjoyment of dinner was reported in the carbohydrate condition compared to mixture and protein conditions (Latner & Schwartz, 1999). Such shorter-term findings are consistent with reports that over the long-term, only dietary protein (as a proportion of total energy intake) is inversely correlated with later intake, unlike carbohydrate or fat (DeCastro, 1994).

The satiety deficits in BN and BED might in part be addressed by inducing dietary patterns designed to increase satiety. It would be valuable to discover whether the different satiating effects of the three macronutrients also apply to individuals with eating disorders. The aim of the study was to compare the satiating effects of protein and carbohydrate in women with BN or BED.

To examine this, 18 women (mean age = 34.8) diagnosed with either BED (n = 11, mean BMI = 31.1) or BN (n = 7, mean BMI = 22.3) were given two weeks of liquid supplements, high-protein or high-carbohydrate (280 kcal), in a repeated-measures, counterbalanced design. Supplements were to be consumed three times each day at one hour prior to typical meal times.

They were balanced across conditions for appearance, flavor, volume, fiber and sweetness. Participants recorded their food intake and binge eating during an initial baseline week, throughout each of the two weeks of high-protein or high-carbohydrate supplementation, and during a seven day wash-out period between the phases.

On the morning of the final day of each phase, participants were given a 420 kcal high-protein or high-carbohydrate supplement, corresponding with the current phase. Three hours later, participants were served a buffet-style test meal in a private room.

Ample portions of foods typical of both meals and binges, varying in macronutrient content and energy density, were offered. Participants were instructed to let themselves eat as much as they felt like eating. Before and after this meal, participants completed visual analogue scales, anchored with the words “most imaginable” and “not at all,” assessing their hunger and fullness, and satisfaction.

At the end of both phases, participants were asked to compare their levels of hunger, satisfaction, interest in food, temptation to binge, and feelings of sickness across the two phases, on 3-point scales labeled “no difference,” “slightly more,” and “much more” during either phase.

Binge eating episodes, or objective bulimic episodes (OBEs), the consumption of unequivocally large amount of food accompanied by a loss of control over eating (Fairburn & Cooper, 1993), were identified from participants’ food records. Binge frequency, food intake at test meals, and hunger and satiety levels before and after these meals, were compared between phases using paired-sample t-tests.

Binge frequency during each phase was also compared with baseline frequency. Chi-square tests were used to compare participants’ reports of hunger, satiation, interest in food, desire to binge, and feelings of sickness across phases.

Binge episodes occurred less frequently during the protein phase (1.19 OBE/week) than during the carbohydrate phase (2.94 OBE/week; t(17) = 4.58, p < .001). During supplementation of protein, but not during carbohydrate supplementation, average daily OBE frequency fell below baseline level (3.01 OBE/week; t(17) = 2.98, p < .01).

The frequency of subjective bulimic episodes (SBEs), the occurrence of a loss of control over eating in the absence of the consumption of a large amount of food (Fairburn & Cooper, 1993) did not differ significantly during either baseline, protein, or carbohydrate supplementation. A greater weight change occurred during the carbohydrate phase than during the protein phase (1.06 vs. 0.27 kg gained; t(17) = 2.92, p < .01).

Unlike the weight change during protein, the gain during the carbohydrate phase was significantly greater than zero (t(17) = 3.78, p < .005).

Three hours following supplements on the final day of each phase, hunger was lower (t(17) = 2.32, p < .05) and fullness was higher (t(17) = 2.25, p < .05) after protein supplements than after carbohydrate supplements. Satisfaction was marginally higher after protein intake (t(17) = 2.06, p = .055).

At the ad libitum test meal, participants consumed less food 3 hours after protein intake (673 kcal) than after carbohydrate intake (856 kcal; t(17) = 2.75, p < .02). The total weight of the food consumed after carbohydrate intake was greater than after protein intake (t(17) = 3.66, p < .005).

There were no differences in macronutrient composition or energy density between the meals. Following these meals, there were no differences in participants’ reported hunger, fullness, or satisfaction, suggesting that participants may have needed to consume less food after protein than after carbohydrate in order to feel as satiated.

At the end of both phases, 61% of participants reported having felt more hungry during the carbohydrate phase, whereas 17% were hungrier during protein (c2(2) = 6.33, p < .05). 61% of participants reported feeling more satiated during the protein phase, compared with 11% during carbohydrate (c2(2) = 7.00 p < .05). 39% vs. 11% reported greater temptation to binge during carbohydrate and protein, respectively (c2(2) = 6.33, p < .05). There were no differences in the proportion of participants in each response category for interest in food or for feeling sick.

These results indicate that in women with BN or BED, dietary supplementation with protein substantially decreased binge-eating, compared to supplementation with carbohydrate and compared to their baseline binge frequency. Protein intake also led to greater fullness, lower hunger, and less ad libitum food intake than carbohydrate intake.

One direction for future research is examining the physiological mechanisms that mediate the effect of protein on satiety. Another research objective might be identifying other aspects of foods that increase satiety in patients with eating disorders. Findings from such research could ultimately inform a longer-term treatment intervention directed at restoring satiety.

It would be useful to examine the clinical implications for these findings. Increasing the level of protein in the diet of patients in treatment for BN and BED might help to correct or to compensate for the satiety deficits and abnormalities in macronutrient selection present in these disorders. Such an intervention might also enhance the ability of patients to remain on a schedule of regular, controlled meals and snacks over the long-term.


Q. How did you assess binge episodes during your study?
A. We had patients keep a diary…

Follow-up: Has that method of monitoring been validated in any way? The schema for assessment of binge eating is still ‘up for debate…’

A. All of the current techniques used to track the frequency of binge episodes are open to criticism, but there is no evidence that other tools are more precise.

Q. What is the difference between Binge Eating Disorder (BED) and BulimiaNervosa (BN)?
A. From a diagnostic perspective, the key distinction is that patients with BN engage in inappropriate compensatory behaviors, such as vomiting. From a clinical perspective, patients with BED are almost always obese, while those with BN tend to be of normal weight. Furthermore, while patients with BN almost always report restricting dieting prior to the development of their disorder, patients with BED are no more likely to have recently dieted than weight-matched subjects without BED.

Q. What’s the difference between BN and BED with regard to insulin release?
A. Studies have shown that post-prandial insulin release is excessively elevated in BN, while among patients with BED (and the obese in general), insulin levels are chronically elevated.

Q. You have referred to your subjects as ‘patients’; is that because they are receiving treatment?
A. No, they are not yet in treatment but rather they are ‘treatment-seeking’ and will begin treatment once the study concludes.

Q. Since ‘Very Low Calorie Diets’ (VLCDs) are high in volume- relative to their energy content- are their effects on food intake related to their filling properties, or to their high protein content?
A. That is an interesting perspective, but I’m not sure anyone has directly studied the mechanism responsible for the efficacy of VLCDs.

Q. What amounts of protein produced lower total intake?
A. The studies I referred to were correlational in design; they looked at long-term spontaneous feeding behavior in humans and found that higher proportions of protein in the diet correlated with lower total intake. Another long-term study with monkeys found that a 36% protein diet suppressed food intake more than a 14% protein diet.

Q. Is there any difference between VLCDs comprised of carbohydrate versus protein, or between the types of proteins used in solid versus liquid forms of VLCDs?
A. I’m not sure about the forms of protein used in liquid versus solid VLCDs.

Q. In your first study were the conditions counterbalanced?
A. Yes they were.

Q. Can you describe the scale to which you refer to as the ‘Excitement Scale’?
A. This scale asked subjects to respond to the following: “How excited are you about eating right now?”

Q. Since the subjects ate different amounts under the different conditions, were there differences in how full they felt?
A. This sensation was not measured in the first study I presented. In the larger study of women with eating disorders, fullness and hunger after the test meals were no different across conditions.

Q. Patients with BN rate protein-containing foods as less palatable than do normal subjects, and they tend to avoid it. So how can you encourage them to increase protein intake if they don’t like it?
A. Certainly it will be challenging to encourage patients with BN to increase their protein intake if they don’t like it, but that is part of what we work on during Cognitive Behavioral Therapy, or CBT.

Q. How many of the subjects typically skipped breakfast?
A. We’re not sure.

Q. How were the diaries validated?
A. We did not validate the food diaries in our study, but others have compared dietary recall obtained from a food diary, with actual laboratory measurement of food intake (1). These studies found that, among patients with BN, information obtained from a food diary accurately reflects the amounts actually consumed by the patients.

Q. Was palatability measured for the foods offered in the test meals?
A. Yes; we tailored each subject’s meal to her own preferences.

Q. Was there a ‘run-in’ time?
A. Yes; the run-in time was one week.

Q. Was the intent of the run-in period to lower binge-eating or to familiarize patients?
A. The intent was to assess baseline binge eating frequency during a run-in period of self-monitoring of food intake. Significant reactivity to self-monitoring did occur, where binge eating frequency decreased during self-monitoring as compared to the frequency reported during the initial interviews.

Q. Can you clarify your categories of binge episodes?
A. An ‘Objective Binge Episode’ (OBE) involved both a loss of control over eating behavior and the consumption of an unequivocally large amount of food. In contrast, a ‘Subjective Binge Episode’ (SBE) involved a sense of loss of control over one’s eating, but the eating episode was not deemed to be unequivocally large.

Q. In your paradigm, did you give the test meal at the end of each test-phase? Also, did you give one at the beginning so you could measure the change in response to your diet?
A. Test meals were given at the end of each phase of the study (protein and carbohydrate). We wanted to include test meals at the start of each phase as well, but this was not feasible.

Q. Did you ask the subjects how much they thought they had eaten?
A. We asked them this question at the end of the actual test meals, but not at the end of the two study-phases.

Q. Were your subjects weight-stable?
A. No, but women with binge eating disorder who are seeking treatment are often not weight-stable.

Q. Were there any racial differences?
A. No, we did not see any differences.

Q. Was there any evidence that these patients had satiety deficits?
A. None at all.

Follow-up: So you found no differences between the ‘binge-eating’ subjects and the normal individuals?

A. The binge-eating subjects may be avoiding protein, or at least they seem to be taking in less protein. Therefore, by providing them with protein in the form of a supplement, we may be able to augment their sensations of satiety.


  1. DeCastro JM (1987). Macronutrient relationships with meal patterns and mood in the spontaneous feeding behavior of humans. Physiology and Behavior, 39, 561-9.
  2. Fairburn CG & Cooper Z. (1993) The Eating Disorder Examination (12th Ed). In CG Fairburn & GT Wilson (Eds), Binge eating: Nature, assessment, and treatment, pp. 317-32. New York: Guilford Press.
  3. Geliebter A, Melton PM, McCray RS, Gallagher DR, Gage D & Hashim SA (1992). Gastric capacity, gastric emptying, and test-meal intake in normal and bulimic women. American Journal of Clinical Nutrition, 56, 656-61.
  4. Geracioti TD & Liddle RA (1989). Impaired cholecystokinin secretion in bulimia nervosa. New England Journal of Medicine, 319, 683-88.
  5. Hadigan CM, Walsh BT, Devlin MJ, LaChaussee JL & Kissileff HR (1992). Behavioral assessment of satiety in bulimia nervosa. Appetite, 18, 233-41.
  6. Hadigan CM, Kissileff HR & Walsh BT (1989). Patterns of food selection during meals in women with bulimia. American Journal of Clinical Nutrition, 50, 759-66.
  7. Hetherington MM, Altemus M, Nelson M, Bernat AA & Gold PW (1994). Eating behavior in bulimia nervosa: Multiple meal analyses. American Journal of Clinical Nutrition, 60, 864-73.
  8. Latner JD & Schwartz M (1999). The effects of a high-carbohydrate, high-protein, or balanced lunch upon later food intake and hunger ratings. Appetite, 33, 119-28.
  9. Rossiter EM, Agras WS, Telch CF, Bruce B (1992). The eating patterns of non-purging bulimic subjects. International Journal of Eating Disorders, 11, 111-20.
  10. Van der Ster Wallin G, Noring C & Holmgren S (1994). Binge eating versus nonpurged eating in bulimics: Is there a carbohydrate craving after all? Acta Psychiatrica Scandinavica, 89, 376-81.
  11. Walsh BT, Kissileff HR, Cassidy SM & Dantzic S (1989). Eating behavior of women with bulimia. Archives of General Psychiatry, 46, 54-8.
  12. Weltzin LK, Hsu LKG, Pollice C & Kaye WH (1991). Feeding patterns in bulimia nervosa. Biological Psychiatry, 30, 1093-110.
  13. Yanovski SZ, Leet M, Yanovski JA, Flood M, Gold PW, Kissileff HR & Walsh BT (1992). Food selection and intake of obese women with and without binge-eating disorder. American Journal of Clinical Nutrition, 56, 975-80.
  14. Hadigan CM, LaChaussee JL, Walsh BT, Kissileff HR & Devlin MJ (1992). The accuracy of 24-hour dietary recall in patients with bulimia nervosa. International Journal of Eating Disorders, 12, 107-11.