Original Article

European Journal of Clinical Nutrition (2006) 60, 897–902. doi:10.1038/sj.ejcn.1602397; published online 15 February 2006

A comparison of effects of fish and beef protein on satiety in normal weight men

Guarantor: S Rössner.

Contributors: BB and SR designed the study. SB designed and cooked the test meals. BB and SB recruited subjects and collected data. ATP and MN conducted the statistical analyses. BB wrote the initial manuscript draft, assisted by SR. None of the authors had any personal or financial conflicts of interest with regard to the study.

S Borzoei1, M Neovius1, B Barkeling1, A Teixeira-Pinto2,3 and S Rössner1

  1. 1Obesity Unit M73, Department of Internal Medicine, Karolinska University Hospital, Huddinge, Stockholm, Sweden
  2. 2Department of Biostatistics and Medical Informatics, Faculty of Medicine, University of Porto, Porto, Portugal
  3. 3Department of Biostatistics, Harvard School of Public Health, Boston, MA, USA

Correspondence: Professor S Rössner, Obesity Unit M73, Department of Internal Medicine, Karolinska University Hospital, Huddinge, Stockholm SE-141 86, Sweden. E-mail: stephan.rossner@medhs.ki.se

Received 4 May 2005; Revised 21 November 2005; Accepted 12 December 2005; Published online 15 February 2006.





Previous studies have indicated that fish protein may have a greater effect on satiety compared to other protein sources of animal origin.



To compare the effects of fish protein and beef protein meals on hunger and satiety.



Twenty-three normal non-smoking, healthy males aged 20–32 years, body mass index 22.5plusminus1.8 (s.d.) kg/m2 participated in a study, with within-subjects design and 1 week between test days. In the morning of the test days, subjects received a standardized breakfast. Four hours after breakfast, subjects were served an iso-energetic protein-rich (40 energy % protein) lunch meal, consisting of either a fish protein dish or a beef protein dish. Four hours after the start of the lunch meals, an ad libitum standardized evening meal was served and the intake of food was measured. Appetite was rated by visual analogue scales (VAS) immediately before and after the meals, as well as every hour between the meals. After the evening meal until bedtime, subjects were asked to record in detail foods and drinks consumed.



The repeated VAS-ratings of hunger, satiety and prospective consumption were modelled in a random effects model, taking pre-lunch VAS-ratings into account. After the fish meal, the point estimates were lower for hunger (-2plusminus4.8), higher for satiety (8.7plusminus6.0) and lower for prospective consumption (-4.9plusminus4.7), but they did not reach statistical significance (P satiety=0.88; P hunger=0.15; P prospective=0.30). However, the energy intake at the evening meal displayed significant differences with subjects eating less after the fish protein lunch (2765 vs 3080 KJ, P<0.01) without feeling less satiated. No later energy compensation after the evening meal was found on the test day.



Although no significant differences in VAS-ratings of satiety or hunger were detected, subjects displayed an 11% reduction in energy intake at the subsequent evening meal.


appetite, beef, energy intake, fish, protein, satiety



Most previous studies, comparing the short-term effect on satiety of protein in comparison with carbohydrates and fat in iso-energetic meals, have found protein to be the most satiating macronutrient (Eisenstein et al., 2002; Westerterp-Plantenga, 2003). Whether different sources of protein, for example, fish protein vs meat protein, exert different effects on satiety has been studied only to a limited extent. In a study by Uhe et al. (1992), beef, chicken and fish protein were compared. Subjects were served grilled whole chunks of 50 g of protein from each food type and were then asked to rate repeatedly how hungry or full they felt during 180 min following commencement of the meals. The results showed that satiety was greater after the fish protein meal compared to the other protein sources. In a study by Holt et al. (1995), subjects were served 1000 KJ (240 kcal) of 38 different food items and were then asked to rate subjective satiety every 15 min during 120 min. A satiety index (SI) was calculated for each test food by dividing the area under the satiety response curve with the area under the satiety response curve for the reference food item, which consisted of white bread. The results from this study indicated that fish protein had a higher SI than all other protein-rich food items tested (e.g. beef steak, eggs) and actually had the second highest SI of all food items tested.

Although both these studies indicate that fish protein has a significant effect on satiety compared to other protein sources of animal origin, these studies do not reveal whether the higher subjective satiety ratings when eating fish protein actually affect subsequent food intake. Furthermore, there are some methodological concerns regarding these previous studies: in the study by Uhe et al. (1992), the texture of whole fillets of beef and fish differed to a large extent, which may have had an impact on the results. In the study by Holt et al. (1995), the protein-rich food items tested differed largely in protein content, with fish having the highest protein content of all items tested (56 g of protein for fish compared to 42 g of protein for beef, in the test portion).

The aim of the present study was to investigate whether a lunch meal with fish protein had a different short-term effect on satiety, compared to an iso-energetic lunch meal with beef protein, by measuring the intake of a subsequent ad libitum evening meal intake and subjective ratings of appetite. The aim was also to make this comparison with meals made of natural, commonly used foods instead of synthetic meals and to keep the content of macronutrients and fibre constant and the texture, appearance and taste as similar as possible.


Subjects and methods


Non-smoking healthy (i.e. without any known diseases) males aged 20–50 years of age were invited by advertising at university sites close to the Karolinska University Hospital.

Twenty-five male subjects were initially included in the study. Two subjects were withdrawn from the study because they failed to consume the entire test meals at lunch, which was stated as inclusion criterion. Thus, 23 non-smoking, healthy, young normal weight male subjects aged 25plusminus4 (meanplusminuss.d.) years with body mass index 22.5plusminus1.8 kg/m2 completed the study. All participants had low scores (less than or equal to50% of the maximum score) on the cognitive restraint scale measured by the short revised 18-item Three-Factor Eating Questionnaire (TFEQ) (Karlsson et al., 2000).

Measurements of appetite and food intake

Ratings of subjective feelings of appetite, that is, desire to eat, hunger, satiety and how much they could eat (prospective consumption) were made by visual analogue scales (VAS) (Rogers and Blundell, 1990; Barkeling et al., 1995). For example, the question 'How strong is your desire to eat now?' was rated along a 100 mm scale anchored with 'Not strong at all' on the left and 'Very, very strong' on the right. Subjects were asked to make a vertical mark across the line corresponding to their feeling at the present time. A further VAS was used at the end of the lunch meals where subjects rated how pleasant they had found the meal. Quantification of the VAS-ratings was made by measuring in mm the distance from the left of the line to the mark.

Since the meal at the ad libitum dinner was of a homogeneous type, the energy intake of the food ingested was simply calculated by measuring the weight of the food eaten. The energy intake from after dinner until bedtime was calculated from a food diary, where participants recorded in detail everything they consumed, giving detailed information on type of food and amounts, using household measures.


The standardized, typical Swedish breakfast, which had to be entirely consumed by the participants, consisted of three slices of white bread (90 g), low-fat margarine (10 g), cheese 17% fat (45 g), a glass of orange juice (178 g), cucumber slices (15 g), slices of bell pepper (15 g) and a cup of coffee or tea. The breakfast energy content was 2093 KJ (500 kcal) (protein 23.5 g (19 energy % (E%)), fat 15.5 g (27 E%), carbohydrate 65.8 g (53 E%), fibre 2.2 g).

The two lunch meals were of the same energy, macronutrient and fibre content (Tables 1 and 2) and differed only in type of protein (beef vs fish). They consisted of boiled rice with a sauce of minced meat or minced fish. In order to make the texture of the meals as similar as possible, fillet of beef and of cod were minced in mincing-machine with a hole size of 3 mm, before pan frying. Appearance and taste of the meals were made similar by colouring the minced cod and beef fillet with canned tomatoes and tomato paste, and by spicing with a generous amount of dried basil. The appearance of the prepared meals was almost identical, except that the fish protein meal was slightly more pale in colour than the beef protein meal. This meal would be considered a normal dish by the volunteers.

The ad libitum evening meal served was an excess portion (1000 g) of a homogeneous, industrially produced typical Swedish hash dish with a standard energy content of 700 KJ (170 kcal)/100 g (protein 6 g (15 E%), fat 8 g (45 E%), carbohydrate 16 g (40 E%)) consisting of diced beef meat, onions and potatoes mixed and fried (Oxpytt, Findus, Bjuv, Sweden), which is a common Swedish dish.


Each subject participated twice, eating each type of test meal (fish/beef) in a counterbalanced order with 1 week in between test days. The day prior to the first test day, subjects were instructed to write a food and physical activity diary and were then asked to maintain these registered meal and activity patterns including bedtime at the day before the next test day, with emphasis on eating the last meal of the day at the same hour. This was checked in detail upon arrival on the second test day. Subjects were also instructed not to drink alcohol the day prior to test days and also to refrain from eating and drinking, except for water after 2200 hours. Subjects arrived fasting to the clinic in the morning of the test days (at the same hour each test day) and were served a standardized breakfast (for content see above). Before breakfast, on the first test day, body weight and height were measured. Four hours after the start of the breakfast, participants were served the lunch meal, consisting of either a fish protein dish or a beef protein dish (see Table 1 for content) with a glass of water throughout the meal, and were asked to consume the entire meal. Four hours after the start of the lunch meals, an ad libitum standardized evening meal (for content see above) was served and participants were asked to eat until satisfied.

VAS scales to monitor subjective feelings of appetite were applied immediately before and after the meals, as well as every hour between the meals. During the whole test days until after the evening meal, all volunteers performed only sedentary activities such as reading and studying.

After the evening meal, participants were allowed to leave the laboratory but were asked to record in detail in a food diary all foods and drinks, including caffeine-containing drinks, consumed for the rest of the day until bedtime for further calculations of the energy intake. Subjects were also asked to record physical activities. These food and activity diaries were mailed back and if necessary a dietitian called the subjects by phone to get additional information.

The participants were informed that they were taking part in a study to test the palatability and the experience of appetite of different meals. The content of the meals served was not described and no mention that food intake was measured at dinner meals was made. In order to distract subjects from the purpose of the study, they were given after each lunch meal an open-ended questionnaire on the taste, appearance, texture, smell and size of the meal.


Statistical analyses were conducted using SAS (version 9, SAS Institute Inc., Cary, NC, USA). All values are expressed as meansplusminuss.d.'s, unless otherwise specified. The level of significance was set at 0.05. Paired t-tests were performed for comparisons of subsequent energy intake and VAS-ratings measuring appetite at single time points. The post-prandial time period from immediately after lunch to before dinner was of particular interest for evaluating the effect of the protein meals. In order to take into account the repeated measures for each subject, a separate linear model with random intercept was fitted for hunger, satiety and prospective consumption using PROC MIXED. Although there were no significant differences in the baseline VAS-ratings (measurement at 11.55), these were subtracted from each measurement after lunch as an adjustment for baseline. Time of the measurement and type of meal were the covariates used in the models. Interaction terms (type of meal times time) were tested but were not significant (all P>0.05) and therefore excluded from the final models.



Energy intake

At the ad libitum evening meal, 4 h after the consumption of the protein meals, 17 of the total 23 subjects ate less food after the fish protein lunch meal compared to after the beef protein lunch meal. The mean energy intake at the evening meal was 2805plusminus1095 after the fish protein lunch compared to 3133plusminus904 KJ after the beef protein lunch (P=0.007; Figure 1). The meal duration was similar on both occasions. The energy intake from after dinner until bedtime calculated from the subjects' food diaries revealed no energy compensation later in the day: when including all subjects in the analyses, the energy intakes after dinner until bedtime were 2901plusminus1924 and 2891plusminus1571 KJ on the fish protein and beef protein test days, respectively. Five of the 23 subjects consumed alcohol during the evening, after they left the laboratory. Two of the subjects consumed alcohol only in the evening after the fish protein day. Two of the subjects consumed alcohol after both test days, although in larger amounts after the fish protein day, and one subject consumed alcohol only after the beef protein day. Since consumption of alcohol adds to the energy intake from food and may also have an appetite-stimulating effect (Yeomans et al., 2003), a further calculation of the evening energy intake was performed, where the five subjects which had consumed alcohol in the evening of the test days were excluded from the analyses. The energy intakes after dinner until bedtime for the remaining 18 non-alcohol consumers were 2278plusminus1651 and 2622plusminus1492 KJ (P=0.44) at the fish protein and beef protein test days, respectively. Caffeine intake did not differ between study occasions.

Figure 1.
Figure 1 - Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact help@nature.com or the author

Mean energy intake (KJ) at the ad libitum meal after the meat and fish protein meal (n=23).

Full figure and legend (37K)

Subjective ratings of appetite

Subjective ratings of appetite were collected during the whole test days. As an example, the temporal profile for hunger is shown in Figure 2. At the time point immediately after the lunch meals, subjective ratings revealed that subjects experienced less hunger (6plusminus8 vs 14plusminus18, P<0.05), were more satiated (84plusminus15 vs 74plusminus20, P<0.05) and thought they could eat less ((prospective consumption) 12plusminus12 vs 23plusminus21, P<0.01) after the fish protein lunch meal compared with after the beef protein lunch meal.

Figure 2.
Figure 2 - Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact help@nature.com or the author

Ratings of hunger during test days are shown. The study procedure until lunch is identical on both occasions. After lunch, the effects on ratings of eating the fish protein vs the beef protein lunch meals are shown. Mean and s.e.m. are given.

Full figure and legend (17K)

In Figure 3, subjective ratings of satiety are shown as changes in ratings from pre-lunch meal levels (delta values). In the random effects models, taking all the ratings between the meals into account, adjusting for baseline score and type of meal, hunger increased by 9.1plusminus0.6 units/h, satiety decreased by 11.6plusminus0.7 units/h and prospective consumption increased by 8.6plusminus0.5 units/h, on average. Although the point estimates (Table 3) for the regression coefficients for fish indicated less hunger (-2plusminus4.8), greater satiety (8.7plusminus6.0) and reduced prospective consumption (-4.9plusminus4.7), none of them reached statistical significance (Psatiety=0.88; Phunger=0.15; Pprospective=0.29). When not adjusting for the nonsignificant difference in VAS scores before lunch (at 1155 hours), the signs of the coefficients were the same and they reached statistical significance.

Figure 3.
Figure 3 - Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact help@nature.com or the author

Changes in ratings of satiety after the intake of lunch meals with fish protein and beef protein are shown. Mean and s.e.m. are given.

Full figure and legend (14K)

Immediately after the ad libitum dinner, there were no differences in ratings of desire to eat, hunger, fullness and prospective consumption, although subjects ate less after the fish protein lunch meal.

Impressions of the type of meal

The beef protein meal was rated as more pleasant than the fish protein meal (65plusminus16 vs 55plusminus23, P<0.05). The open-ended questionnaires and the verbal comments from the subjects revealed that they could not distinguish the two equally spiced meals from each other. In a few cases, a 'fishy' smell was reported from either meal. The records from the physical activity diaries did not reveal any significant differences between test days (P>0.05).



This study demonstrates that a protein-rich lunch meal with fish protein reduced subsequent energy intake compared to an iso-energetic lunch meal of beef protein, in normal weight young men. The subjective ratings of appetite did, however, not display significant differences, although the point estimates indicated effects in the same direction. These results indicate that type of protein, that is, fish vs beef, may be of importance for satiety, at least in a short-term perspective. Certainly the satiating effects of fish may depend on other factors than the protein content, such as fat quality. Cod however is a very lean fish (0.7 g fat/100 g).

The rated pleasantness of eating the fish protein meal was lower compared to the beef protein meal. A number of studies have shown that palatability affects satiation, that is, affects the amount of food eaten within a meal, with an increased intake as palatability increases (Sorensen et al., 2003). Whether palatability also affects satiety, that is, subsequent hunger and satiety and subsequent ingestion of food, has also been studied. The effect on post-prandial appetite ratings after fixed pre-loads of different palatability, but with the same nutritional content, where palatability has been manipulated by adding flavours (not sweeteners), has been examined in three studies (Rogers and Blundell, 1990; Warwick et al., 1993; De Graaf et al., 1999). In one of the studies (Warwick et al., 1993), hunger ratings were lower and fullness greater after the more palatable meal. In the two other studies (Rogers and Blundell, 1990; De Graaf et al., 1999), there was no effect of palatability on post-prandial appetite ratings. In these three studies (Rogers and Blundell, 1990; Warwick et al., 1993; De Graaf et al., 1999), the effect on subsequent food intake was also measured 30 min, 90 min or 3 h after the pre-loads, and in none of them the subsequent food intake was affected by the palatability.

In the present study, we wanted to use cooked lunch meals made of commonly used food items (in contrast to synthetic formulas), where the protein to be tested was incorporated. The protein content of the lunch meals (47 E% of protein) in this study was high. If lower protein content had been used, possible differences in satiety effects between fish and beef protein may not have been detectable.

The mechanisms for the reduction in subsequent energy intake after the fish protein meal compared with the beef protein meal were not tested in this study. However, in the study by Uhe et al. (1992), comparing the satiety effect of chunks of fillets of fish, beef and chicken, the plasma amino-acid profiles, dietary amino-acid profiles, plasma glucose and insulin were measured. A study by Hall et al. (2003) have similarly described how casein and whey protein appear to exert differential effects on satiety, possibly mediated by post-absorptive increases in amino acids, cholecystokinin and glucagon-like peptide 1 (CCK and GLP-1). The results obtained after consuming fish differed significantly from after consumption of beef and chicken in a number of factors, which may be related to satiety. The tryptophan: large neutral amino-acid (LNAA) ratio decline after the meal was slower, the amino-acid concentration took longer to reach peak levels and the dietary and plasma concentrations of taurine and plasma concentration of methionine were significantly higher after fish consumption compared to after beef and chicken. As also pointed out by Uhe et al. (1992), serontoninergic factors as well as slower digestibility of fish proteins could help to explain our extended observations of fish protein enhanced satiety.

Further long-term trials may help to reveal whether a fish protein-rich diet may help to control energy balance.



  1. Barkeling B, Rossner S, Sjoberg A 1995. Methodological studies on single meal food intake characteristics in normal weight and obese men and women. Int J Obes Relat Metab Disord 19, 284–290. | PubMed | ChemPort |
  2. De Graaf C, De Jong LS, Lambers AC 1999. Palatability affects satiation but not satiety. Physiol Behav 66, 681–688. | Article | PubMed | ISI | ChemPort |
  3. Eisenstein J, Roberts SB, Dallal G, Saltzman E 2002. High-protein weight-loss diets: are they safe and do they work? A review of the experimental and epidemiologic data. Nutr Rev 60, 189–200. | Article | PubMed | ISI |
  4. Hall WL, Millward DJ, Long SJ, Morgan LM 2003. Casein and whey exert different effects on plasma amino acid profiles, gastrointestinal hormone secretion and appetite. Br J Nutr 89, 239–248. | Article | PubMed | ISI | ChemPort |
  5. Holt SH, Miller JC, Petocz P, Farmakalidis E 1995. A satiety index of common foods. Eur J Clin Nutr 49, 675–690. | PubMed | ISI | ChemPort |
  6. Karlsson J, Persson LO, Sjostrom L, Sullivan M 2000. Psychometric properties and factor structure of the Three-Factor Eating Questionnaire (TFEQ) in obese men and women. Results from the Swedish Obese Subjects (SOS) study. Int J Obes Relat Metab Disord 24, 1715–1725. | Article | PubMed | ChemPort |
  7. Rogers PJ, Blundell JE 1990. Umami and appetite: effects of monosodium glutamate on hunger and food intake in human subjects. Physiol Behav 48, 801–804. | Article | PubMed | ISI | ChemPort |
  8. Sorensen LB, Moller P, Flint A, Martens M, Raben A 2003. Effect of sensory perception of foods on appetite and food intake: a review of studies on humans. Int J Obes Relat Metab Disord 27, 1152–1166. | Article | PubMed | ChemPort |
  9. Uhe AM, Collier GR, O'Dea K 1992. A comparison of the effects of beef, chicken and fish protein on satiety and amino acid profiles in lean male subjects. J Nutr 122, 467–472. | PubMed | ISI | ChemPort |
  10. Warwick ZS, Hall WG, Pappas TN, Schiffman SS 1993. Taste and smell sensations enhance the satiating effect of both a high-carbohydrate and a high-fat meal in humans. Physiol Behav 53, 553–563. | Article | PubMed | ISI | ChemPort |
  11. Westerterp-Plantenga MS 2003. The significance of protein in food intake and body weight regulation. Curr Opin Clin Nutr Metab Care 6, 635–638. | PubMed | ISI | ChemPort |
  12. Yeomans MR, Caton S, Hetherington MM 2003. Alcohol and food intake. Curr Opin Clin Nutr Metab Care 6, 639–644. | PubMed | ISI |


These links to content published by NPG are automatically generated

Extra navigation