We have shown that 20 g of casein or pea protein has a stronger effect on lowering food intake 30 min later compared to whey protein, egg albumin and maltodextrin. This was further supported through higher feelings of satiety after the casein and pea protein preload. However, this effect on food intake was attenuated when the preload was consumed immediately before the ad libitum meal.
Ad libitum energy intake was lower after the pea protein and the casein preloads in the first experiment and showed a trend in the second experiment. These findings demonstrate pea protein and casein as candidate proteins for satiety. In the literature, there are inconsistent findings related to protein source and satiety when a 45-50 g dose is used. Previous studies have shown similar  or lower effect on food intake  when whey was compared to casein. These inconsistencies can be attributed to different reasons including dose, study design, subject sample, as well as different physical properties of the proteins used. Even within the same source of protein, attributes can differ with regards to degree of hydrolyzation of the peptides, aggregation of the peptides (micelles) and purity of the isolates used.
Unlike casein protein, pea protein has not been extensively investigated. In the present study, we show for the first time that pea protein is effective in lowering short-term food intake. A few studies have shown either similar or lower food intake after pea protein  and pea protein hydrolyzate  respectively compared to other proteins. Mechanisms were not explored but gastric emptying might play a role. Casein has been shown to exhibit slower gastric emptying compared to whey protein . Other potential mechanisms can be related to the action of satiety hormones.
When the preload was consumed 30 min before the ad libitum meal, total food intake was similar after casein and pea protein compared to after water control, with a caloric compensation of around 100%. This means that subjects were able to compensate only for the calories in the casein or pea protein preload and not more. Except for a limited number of fiber studies [28, 29], it is rare to find an effect of a preload on total food intake that is superior to the calories of the preload. Accordingly, there is limited basis for recommending the ingestion of a preload or a snack to reduce total food intake when simply drinking water or not ingesting the preload or snack results in the same effect. We therefore suggested conducting a second experiment to measure satiation where the preload is given right before the ad libitum meal in order to maximize its effect on the meal or caloric compensation. Contrary to our hypothesis, we found that administering the pea protein or casein preload as a starter before the meal did not lower ad libitum food intake. Furthermore, total intake was higher compared to the water control. In the literature, there are no reported studies on protein source and satiation. The observed lack of effect on satiation was perhaps because the drinks were consumed fast and therefore the volume effect might have overridden any potential functional effect of the protein preload. In both experiments, all treatments were iso-volumetric and subjects were instructed to drink the preloads quickly, as a shot. Previous studies have shown that by shortening the delay between the preload and the ad libitum meal, energy compensation is affected by the volume of the undigested preload and not the type of preload served [30–32]. A slow eating rate has been associated with lower food intake and higher satiety [33–35] perhaps due to longer oro-sensory exposure as well as interaction with the gastrointestinal tract to release of satiety signals. Ratings of satiety in the first experiment were higher after the pea protein and casein preloads compared to other preloads. This might explain the observed lower energy intake. In the second experiment, the ratings were measured only after the ad libitum meal and as expected, they did not differ amongst the preloads.
Postprandial glycaemia was measured as a secondary outcome to investigate if the reported second meal effect of whey protein [16, 36] persists when the preload is administered 30 min before the meal. Indeed, whey protein blunted the blood glucose response to the ad libitum meal compared to the other protein preloads and water control. Unlike the literature where a fixed meal was administered, which used a fixed meal, our results were reported after an ad libitum meal and therefore should be considered with caution. We did not investigate the mechanisms responsible for this decrease in blood glucose, but others have shown an increase in plasma insulin concentrations after whey protein which could explain the blunted glucose response [16, 36]. Furthermore, the time of ingestion of the ad libitum meal (30 min after the preload) corresponds to the peak insulin response after protein and carbohydrate preloads [7, 37]
One limitation in the first study is the palatability of the pea protein preload which was lower than the other preloads. It was not possible to control for palatability in the statistical analysis due to its correlation with energy intake. It is not clear how much of the observed effect on energy intake is driven by palatability. For instance, the casein preload, in the first experiment, had a similar palatability to albumin and whey preloads but still resulted in significantly lower energy intake. Studies by Spitzer et al.  and Poppitt et al.  found little or no evidence for a reduced food intake after low palatable foods. Furthermore, De graaf et al have shown that palatability has an effect on the intake of the preload itself and not on subsequent food intake and satiety .
Other limitations include including only male subjects which raises the question of whether these findings can be applied to women as well, and the single blinded design of the studies which could result in a bias on the results. However for the latter, the person administering the treatments was not the same person preparing them, and given that the drinks were served in opaque and covered cups, the bias is quite limited.