Glycaemic and Satiating Properties of Potato Products

ORIGINAL ARTICLE Glycaemic and satiating properties of potato products

M Leeman, E O¨ stman and I Bjo¨rck

Applied Nutrition and Food Chemistry, Department of Food Technology, Engineering and Nutrition, Lund University, Lund, Sweden

Objective: To investigate glycaemic and satiating properties of potato products in healthy subjects using energy-equivalent or carbohydrate-equivalent test meals, respectively. Subjects and setting: Thirteen healthy subjects volunteered for the first study, and 14 for the second. The tests were performed at Applied Nutrition and Food Chemistry, Lund University, Sweden. Experimental design and test meals: All meals were served as breakfast in random order after an overnight fast. Study 1 included four energy-equivalent (1000 kJ) meals of boiled potatoes, french fries, or mashed potatoes; the latter varying in portion size by use of different amounts of water. The available carbohydrate content varied between 32.5 and 50.3 g/portion. Capillary blood samples were collected during 240 min for analysis of glucose, and satiety was measured with a subjective rating scale. Study 2 included four carbohydrate-equivalent meals (50 g available carbohydrates) of french fries, boiled potatoes served with and without addition of oil, and white wheat bread (reference). The energy content varied between 963 and 1534 kJ/portion. Capillary blood samples were collected during 180 min for analysis of glucose, and satiety was measured using a subjective rating scale. Results: Study 1: boiled potatoes induced higher subjective satiety than french fries when compared on an energy-equivalent basis. The french fries elicited the lowest early glycaemic response and was less satiating in the early postprandial phase (area under the curve (AUC) 0–45 min). No differences were found in glycaemic or satiety response between boiled or mashed potatoes. Study 2: french fries resulted in a significantly lower glycaemic response (glycaemic index (GI) ¼ 77) than boiled potatoes either with or without addition of oil (GI ¼ 131 and 111, respectively). No differences were found in subjective satiety response between the products served on carbohydrate equivalence. Conclusions: Boiled potatoes were more satiating than french fries on an energy-equivalent basis, the effect being most prominent in the early postprandial phase, whereas no difference in satiety could be seen on a carbohydrate-equivalent basis. The lowered GI for french fries, showing a typical prolonged low-GI profile, could not be explained by the fat content per se. Sponsorships: The Swedish Agency for Innovation Systems (Project No P11900-3 A) and O¨ resund Starch Profiles (O¨ SP). European Journal of Clinical Nutrition (2008) 62, 87–95; doi:10.1038/sj.ejcn.1602677; published online 28 February 2007

Keywords: potatoes; blood glucose; satiety; french fries; glycaemic index

Introduction studies as reviewed by for example, Ludwig (2000), Pawlak et al. (2002) and Raben (2002). However, the conclusions The role of glycaemic index (GI) on satiety sensation and regarding a potential relationship differ. In a study with weight maintenance has been the subject of a number of isovolumetric beverages, Anderson et al. (2002) found that a high glycaemic response was associated with increased Correspondence: Dr E O¨ stman, Applied Nutrition and Food Chemistry, satiety in the early postprandial period (1 h). In elderly Department of Food Technology, Engineering and Nutrition, Lund University, subjects, mashed potatoes (high-GI food) induced signifi- PO Box 124, SE-221 00 Lund, Sweden. cantly higher glycaemia and satiety compared with a barley E-mail: [email protected] Guarantor:IBjo¨rck. meal (low-GI food) containing the same amount of available Contributors: ML and IB made the design of the studies. ML was responsible for carbohydrates as measured during 120 min (Kaplan and ¨ collection of data and statistical analysis of the collected data. ML, EO and IB Greenwood, 2002). In contrast, a lower satiety was observed contributed to the writing of the manuscript. Received 27 April 2006; revised 23 November 2006; accepted 4 January 2007; in obese children after consumption of a rapidly digested published online 28 February 2007 carbohydrate meal (including potatoes) compared with a Glycaemic and satiating properties of potato products M Leeman et al 88 lente meal (including spaghetti; Alvina and Araya, 2004). In Materials and methods the latter study, no difference was found in normal weight children. More recent data suggest that consumption of Study 1 low-GI foods at breakfast may result in a lowered voluntary Subjects and test meals. Thirteen healthy subjects, nine men food intake at the subsequent lunch as studied in non- and four women, aged 19–27 years, with normal BMIs overweight and obese children (Warren et al., 2003). Low-GI (BMI7s.d.; 21.873.1 kg/m2) and without drug therapy, diets have also been examined in the treatment of paediatric participated in the study. Four meals were included in the obesity; and as opposed to an energy-restricted reduced study, each meal containing 1000 kJ. French fries (Ica fat diet, consumption of a low-GI/regular fat diet during Handlarna, Solna, Sweden) were bought locally. The french 4 months resulted in a lowered body mass index (BMI) fries, containing 8% fat (sunflower oil), were heated in an and body weight in obese children (Spieth et al., 2000). oven at 2501C for 9 min (according to the manufacturer’s Some studies have been investigating the satiating effect instructions), before being served with an addition of 0.5 g of different foods on an energy-equivalent basis. Holt et al. NaCl. Mashed potatoes were prepared from an instant potato (2001) found that energy-equivalent portions of bread powder (Felix, Eslo¨v, Sweden); 69 g of powder was recon- differed both in satiating and glycaemic response, stituted with either 200 or 330 g water before being served although no correlation between satiety and glycaemia (mashed small and mashed large, respectively). The potato was seen. When comparing energy-equivalent (1000 kJ) powder contains 4 g NaCl/100 g powder. Potatoes of the portions of carbohydrate foods, high-GI potatoes variety King Edward (mealy winter variety) were bought (Foster-Powell et al., 2002) promoted a considerably higher locally. Before being served, the potatoes were peeled and satiety than many other foods, irrespective of their GI boiled for 18–20 min (until soft). At serving, 0.5 g NaCl was character (Holt et al., 1995). In the Holt study, there sprinkled on the potatoes. The potatoes were freshly cooked was a large range in portion size and thus, for example meal before each test meal to avoid starch retrogradation. The volume varied. There are studies in the literature indicating a meals were served either with water or with a combination of positive influence of volume on satiety (Rolls et al., 1999, milk and water. Commercial milk (1.5% fat) produced by 2000). In the study by Holt et al. (1995), large differences in Ska˚nemejerier (Malmo¨, Sweden) was bought locally. Milk subjective rating of satiety were found between different (60 g) was served with boiled potatoes and the french fries to potato products. In the literature there are also large compensate for the milk component in the mashed potatoes. differences in glycaemia after different potato products, In total, 250 ml of liquid was served with each meal. Also, depending on, for example, processing conditions; 150 ml of coffee or tea was included in each meal. The test canned potatoes and instant mashed potatoes ranging in subjects were allowed to choose between these drinks at the GI between 87 and 139 (Foster-Powell et al., 2002). In one first occasion and then retained the same drink through study, potatoes, regardless of variety, cooking method or all the test meals. The subjects were served the meals for maturity had exceptionally high-GI values, or from 93 to breakfast in random order after an overnight fast. The tests 144 (Soh and Brand-Miller, 1999). Recently, GI values were were performed approximately 1 week apart and commenced published for potato products available in Great Britain atthesametimeinthemorning.Allmealswereconsumed (Henry et al., 2005) and North America, respectively over 12 min. The nutritional composition of each food per (Fernandes et al., 2005); different potato products displaying 1000 kJ is listed in Table 1. Data on french fries are based on GIs between 78 and 132. Still, the GI data for well- manufacturer’s data, whereas the remaining values are from characterized products are scarce, and there are for example food tables (Livsmedelsverket, 2001). The starch content in the only two values found for french fries (GI ¼ 90 and 107) potatoes was calculated by use of a formula stated by Leeman (Wolever et al., 1994; Fernandes et al., 2005). et al. (2005), where a strong correlation between dry matter The aim of the present study was to evaluate postprandial content and available starch content was found in boiled satiety of potato products, and its relation to glycaemic potatoes. In Table 1, there are also data on serving weight, meal response in healthy subjects using well-defined potato volume, energy density and dry matter content. Volume was products. Both energy-equivalent (1000 kJ as used in the measured with a rapeseed equipment. Milk added to french study by Holt et al., 1995) and carbohydrate-equivalent fries and boiled potatoes is included in the nutrient composi- portions (50 g available carbohydrates) were used as basis tion, but not in volume, weight or dry matter measurements. for comparison. White wheat bread (WWB) was included in the equi-carbohydrate study to allow for calculation of GI. Mashed potatoes prepared from an instant potato powder Study 2 were tested at two different portion sizes by using different Subjects and test meals. Fourteen healthy subjects, six men amounts of water for reconstitution. Two of the potato test and eight women, aged 20–28 years, with normal BMIs meals (boiled potatoes with oil and french fries) contained a (21.972.0 kg/m2) and without drug therapy, participated in similar amount of sunflower oil to make it possible to study the study. Four meals were included in the study and all test the potential effects of fat co-ingestion on satiety and meals contained 50 g available carbohydrates. The french glycaemia, respectively. fries were of the same brand and were prepared in the same

European Journal of Clinical Nutrition Glycaemic and satiating properties of potato products M Leeman et al 89 manner as in Study 1. Potatoes of the cultivar Asterix (firm graded between ‘painfully hungry’ to ‘full to nausea’ winter variety) were bought from a local farmer. They were (Liljeberg et al., 1995). The test subjects filled out the forms peeled and freshly cooked for 30 min before each test to themselves. The study was approved by the Ethics Commit- avoid starch retrogradation. At serving, 0.5 g NaCl was added tee of the Faculty of Medicine at Lund University. to the potatoes. The boiled potatoes were served either with or without the addition of 15.4 g sunflower oil (Ica Handlarna, Solna, Sweden) to match the type of oil and lipid Calculations and statistical methods content of the french fries portion. WWB was used as a For each subject and test meal, the incremental blood reference meal and was prepared according to Liljeberg and glucose- and satiety areas under the curves (AUCs) were Bjo¨rck (1994). NaCl content of WWB is approximately calculated (GraphPad Prism, version 3.0; GraphPad Software, 1.4 g/portion. The composition of the meals is shown in San Diego, CA, USA). GI was calculated from the 120 min Table 2. The starch content of WWB, potatoes and french fries incremental blood glucose area by using WWB as a reference was analysed according to Holm et al. (1986), and remaining (GI ¼ 100) (FAO/WHO, 1998). Values are presented as data were obtained from manufacturer’s data or food tables means7s.e.m. and significant differences among the time (Livsmedelsverket, 2001). The added oil is not included in the points and AUCs were assessed with a general linear model dry matter content of the meal ‘boiled potatoes þ oil’. The (analysis of variance) followed by Tukey’s multiple compar- subjects were served the meals for breakfast in random order isons test (MINITAB Statistical Software, release 13 for after an overnight fast. The tests were performed approxi- Windows; Minitab Inc, State College, PA, USA). Spearman’s mately 1 week apart and commenced at the same time in rank correlation was used to study the relations between the morning. All test meals were served with 150 ml water blood glucose, satiety and meal composition parameters. and consumed over 12 min. Also, 150 ml coffee or tea was Correlations were performed for AUCs 0–45 and 70–180 min, included in each meal. The test subjects were allowed to respectively. A correlation for each subject was calculated choose between these drinks at the first occasion and then and from these values the mean value of Spearman’s retained the same drink through all the test meals. correlation coefficient was obtained. To determine the P-value, a permutation test was performed by using MATLAB Sampling and analysis. Finger-prick capillary blood samples with the null hypothesis that no correlation existed (the were taken using Haemolancet (HaeMedic AB, Munka alternative hypothesis was that the data were correlated). Ljungby, Sweden) before the meal (0) and at 20, 45, 70, 180 Values of Po0.05 were considered significant. and 240 min after the meal in Study 1 and before the meal (0) and at 20, 45, 70, 120 and 180 min after the meal in Study 2 for determination of blood glucose (HemoCue B-glucose Results analyser; HemoCue AB, A¨ngelholm, Sweden). Assessments of feelings of hunger/satiety were performed immediately after Energy-equivalent meals (Study 1) each blood sampling. The rating scale used at each time The blood glucose response at 20 and 45 min after the test point was bipolar and contained nine different wordings meal with french fries was significantly lower than after all

Table 1 Composition of the meals in study 1, each meal containing 1000 kJ

Meal Protein (g) Carbohydrate (g) Fat (g) Volume (ml) Weight (g) Energy density (kJ/g) Dry matter (%)

French fries 5.1 32.5 10.3 125 118 8.5 45 Potatoes 6.7 50.3 0.9 280 305 3.3 21 Mashed small 5.9 46.2 3.5 300 250 4.0 24 Mashed large 5.9 46.2 3.5 375 380 2.6 17

Values for french fries are based on manufacturer’s data. Remaining values from food tables (Livsmedelsverket, 2001). The starch content in the boiled potatoes was calculated by use of a formula stated by Leeman et al. (2005).

Table 2 Composition of the carbohydrate-equivalent meals in study 2

Meal Protein (g) Carbohydrate (g) Fat (g) Volume (ml) Weight (g) Energy (kJ) Energy density (kJ/g) Dry matter (%)

WWB 10.3 50 4.2 310 120 1134 9.5 52 French fries 4.8 50 15.4 260 193 1448 7.5 45 Potatoes 1.6 50 0.1 290 272 963 3.5 21 Potatoes þ oil 1.6 50 15.4 305 287 1534 5.3 21

Abbreviation: WWB, white wheat bread. Values for french fries are based on manufacturer’s data. Remaining values from food tables (Livsmedelsverket, 2001).

European Journal of Clinical Nutrition Glycaemic and satiating properties of potato products M Leeman et al 90 other meals (Figure 1; Po0.05). In addition, at 70 min the Boiled potatoes produced the highest satiety scores and blood glucose response was lower after the french fries deviated significantly from the french fries at 20–70 min compared with the small portion of mashed potatoes and postprandially (Figure 2; Po0.05). Also, the smaller portion the boiled potatoes (Po0.01). The blood glucose AUC was with mashed potatoes was more satiating at 20–45 min significantly lower after french fries compared with the other after the meal than were the french fries. Satiety AUC was test meals (Table 3; Po0.05). No significant differences were significantly lower for french fries compared with boiled seen in blood glucose increment between the meals with potatoes at all time intervals and compared with the small boiled potatoes and mashed potatoes; neither at specific portion of mashed potatoes at 0–45 and 0–70 min (Table 4; time points nor in AUCs. Po0.05).

Figure 1 Mean blood glucose responses in 13 healthy subjects at a Figure 2 Mean satiety scores in 13 healthy subjects at a breakfast breakfast with energy-equivalent portions of boiled potatoes (m), with energy-equivalent portions of boiled potatoes (m), french fries french fries (’), mashed potatoes in a small portion (.) and a large (’), mashed potatoes in a small portion (.) and a large portion portion (E). Values are presented as means (n ¼ 13) with bars (E). Values are presented as means (n ¼ 13) with bars indicating indicating s.e.m. Significant differences between blood glucose s.e.m. Significant differences between satiety scores at certain time levels at certain time points are indicated with letters (Po0.05). points are indicated with letters (Po0.05).

Table 3 Acute postprandial blood glucose areas in 13 healthy subjects after energy-equivalent meals (study 1)

Meal AUC (mmol/l min)

0–45 0–70 0–180 0–240 70–180

French fries 34.574.0a 51.676.4a 69.0712.4a 73.9714.4a 17.577.5a Potatoes 75.278.0b 129715.1b 192727.7b 195727.3b 63.0714.2b Mashed small 80.977.6b 141715.8b 211730.3b 218730.6b 70.4716.2b Mashed large 84.175.9b 129712.4b 167718.9b 171719.4b 38.378.8ab

Abbreviation: AUC, area under the curve. Values are means7s.e.m., n ¼ 13. Values within a column not sharing the same letter are significantly different (Po0.05).

Table 4 Area under curve for satiety scores in 13 healthy subjects after energy-equivalent meals (study 1)

Meal AUC

0–45 0–70 0–180 0–240 70–180

French fries 68.3717.8a 101727.0a 152743.4a 153743.9a 50.8719.2a Potatoes 201740.5b 320766.9b 5947137b 6177146b 274774.8b Mashed small 169736.7b 261755.6b 4247111ab 4447121ab 163764.6ab Mashed large 148739.5ab 235762.1ab 4397121ab 4567127ab 204763.5ab

Abbreviation: AUC, area under the curve. Values are means7s.e.m., n ¼ 13. Values within a column not sharing the same letter are significantly different (Po0.05).

European Journal of Clinical Nutrition Glycaemic and satiating properties of potato products M Leeman et al 91 No correlations were found between blood glucose and and at 20–70 min compared with boiled potatoes with oil satiety or between blood glucose or satiety and nutritive or (Figure 3; Po0.05). At 120 and 180 min, both meals of boiled physical characteristics of the different meals. potatoes gave significantly lower blood glucose responses than french fries and at 120 min, both boiled potato meals also gave significantly lower glucose responses Carbohydrate-equivalent meals (Study 2) than WWB. Blood glucose AUCs, as calculated for 0–45, When comparing carbohydrate-equivalent meals, the blood 0–70, 0–120 and 0–180 min, were significantly lower after glucose response was significantly lower after french fries at french fries compared with potatoes with oil (Table 5; 20–45 min compared with boiled potatoes without oil, Po0.05). Furthermore, the AUCs after french fries were significantly lower than with potatoes without oil as calculated for 0–45, 0–70 and 0–120 min (Table 6; Po0.05). GI was significantly lower for french fries (77) compared with boiled potatoes with oil (131) and without oil (111). No significant differences in satiety sensation were found between the meals; neither for AUCs nor at specific time points (Table 6, Figure 4; Po0.05). No correlations were found between blood glucose and satiety or between blood glucose or satiety and the nutritive or physical characteristics of the different meals.

Discussion

Energy-equivalent meals (Study 1) The main finding of the first study was that boiled potatoes were significantly more satiating than french fries in the Figure 3 Mean blood glucose responses in 14 healthy subjects at a early postprandial phase (70 min) when served as energy- breakfast with carbohydrate-equivalent portions of white wheat equivalent meals. This finding supports the view that an bread (K), french fries (’) and boiled potatoes with (*) and without (m) addition of oil. Values are presented as means (n ¼ 14) with bars elevated early glycaemia per se may be advantageous in indicating s.e.m. Significant differences between blood glucose relation to satiety. The result is also in line with the study by levels at certain time points are indicated with letters (Po0.05). Holt et al. (1995), where potatoes were found to be more

Table 5 Acute postprandial blood glucose areas and GI values at 120 min in healthy subjects after carbohydrate-equivalent meals (study 2)

Meal AUC (mmol /l min) GI

0–45 0–70 0–120 0–180 70–180 120 min

WWB 53.475.0b 10778.1bc 155712.3ab 170713.9ab 62.278.8ab 100ab French fries 44.174.3b 82.078.4b 116714.9b 139719.3b 56.5711.4ab 77710.0b Potatoes 70.075.0a 126710.8ac 160717.9a 161718.1ab 35.379.0a 111713.7a Potatoes þ oil 68.375.4a 136712.8a 194721.4a 202721.9a 65.6711.6b 131714.6a

Abbreviations: AUC, area under the curve; WWB, white wheat bread. Values are means7s.e.m., n ¼ 14. Values within a column not sharing the same letter are significantly different (Po0.05).

Table 6 AUCs for satiety in healthy subjects after carbohydrate-equivalent meals (study 2)

Meal AUC

0–45 0–70 0–120 0–180 70–180

WWB 196734.8 321756.1 490787.5 5987117 277773.0 French fries 217738.8 335762.0 5057106 6137145 277784.9 Potatoes 215728.6 350742.1 520766.6 596783.8 245750.7 Potatoes þ oil 214727.7 330742.0 467770.8 5637109 233774.8

Abbreviations: AUC, area under the curve; WWB, white wheat bread. Values are means7s.e.m., n ¼ 14.

European Journal of Clinical Nutrition Glycaemic and satiating properties of potato products M Leeman et al 92 than mashed potatoes. In fact, some of the participants perceived the meal with boiled potatoes as difficult to consume, mainly because the portion was experienced as large but also owing to the fact that subjects were not accustomed to have potatoes for breakfast. However, no measurements on subjective palatability and/or physical comfort/discomfort were performed. In the present study, no correlations were found between satiety and nutritive or physical characteristics of the meals. One study emphasized portion size and energy density as being the strongest determinants of the satiating effect of energy-equivalent meals (Holt et al., 1996). Energy density (kJ/g) has also been proposed as one of the key determinants of energy intake (Rolls and Bell, 1999; Kral et al., 2004). Furthermore, Kral et al. (2004) conclude that energy density Figure 4 Mean satiety scores in 14 healthy subjects at a breakfast and portion size act independently to affect energy intake. It with carbohydrate-equivalent portions of white wheat bread (K), has also been proposed that we tend to eat the same amount/ ’ m french fries ( ) and boiled potatoes with (*) and without ( ) volume that we are accustomed to, independently of energy addition of oil. Values are presented as means (n ¼ 14) with bars indicating s.e.m. Significant differences between satiety scores at content or composition of the meal (Rolls and Bell, 1999). In certain time points are indicated with letters (Po0.05). the present study, the portion with boiled potatoes in particular was perceived as larger than most subjects would normally consume, whereas the portion of french fries was satiating than many other foods, including french fries, perceived as quite small. when served as 1000 kJ portions. According to a review by Anderson and Woodend (2003), high but not low glycaemic responses are associated with increased satiety and/or Carbohydrate-equivalent meals (Study 2) reduced food intake in the early postprandial phase (1 h) In the study comparing carbohydrate-equivalent meals, (Holt et al., 1995; Anderson et al., 2002), whereas the reverse french fries resulted in a significantly lower GI (77) than occurs in the later phase after a meal (up to 6 h) (van boiled potatoes; both with (GI ¼ 131) and without (GI ¼ 111) Amelsvoort and Weststrate, 1992). An inverse relationship addition of 15 g of sunflower oil to match the nutrient between glycaemic response and satiety sensation has also composition and fat quality of the french fries. The meal of been reported, suggesting that low-GI foods are more french fries exhibits a curve with a low glucose peak followed satiating (Ludwig, 2000; Roberts, 2000), although it has by a sustained net increment. Few GI data of french fries are been pointed out that in some of these studies the control of available in the literature. The GI of frozen/oven-heated confounding variables was not adequate (Roberts et al., french fries in the present study is lower than that reported 2002). Thus, the overall glycaemic profile is important, by Wolever et al. (1994) for frozen/microwave-heated french and high-GI foods may show benefits in the early phase after fries (107), or for frozen/oven-heated french fries (90) as food intake, whereas maintained glycaemia may result in reported by Fernandes et al. (2005). A number of studies have preserved satiety also in the later postprandial phase. If the been dealing with the effects of co-ingestion of fat on the experimental design had been based on ad libitum food glycaemic response, some showing a lowering effect (Collier intake rather than energy-equivalent loads, it cannot be and O’Dea, 1983; Lunetta et al., 1995) whereas no effect was excluded that the voluntary food intake instead should have observed in a study with non-insulin-dependent diabetics been reduced in the case of boiled potatoes. It should (Gannon et al., 1993). According to Wolever and Bolognesi be noted that the lowered glycaemia obtained with french (1996), variations in protein and fat content of mixed meals fries vs boiled potatoes using energy balanced meals does not had negligible effects on post-meal glycaemia. Furthermore, result only from a more lente delivery of glucose to the in a study by Owen and Wolever (2003), variation of fat blood, but also from a lower carbohydrate load. intake across the normal range (17–44% energy) did not Holt et al. (1995) suggested that the higher bulk in the case affect the glycaemic response to white bread. In most of the of the boiled potatoes (368 g) possibly explain the increased studies referred to above, the fat incorporated in the meals satiety response compared with french fries (93 g). However, was butter or margarine, and it cannot be excluded that the in our study, the meal with boiled potatoes (300 g) tended to degree of saturation of the fat might have affected the produce a higher satiety than the large portion of mashed outcome. However, in a study where fat with different potatoes (380 g). A normal portion of boiled potatoes is degrees of saturation was added to meals of mashed potatoes, generally considered to be within the range of 150–200 g no differences appeared neither in glycaemia, insulinaemia (Enghardt Barbieri and Lindvall, 2003). Thus, it could be that nor in satiating response (MacIntosh et al., 2003). Thus, it boiled potato tubers were perceived as more difficult to eat seems as if addition of sunflower oil in the amounts tested

European Journal of Clinical Nutrition Glycaemic and satiating properties of potato products M Leeman et al 93 here should not be expected to lower the postprandial (2005) found that floury potatoes exhibited a somewhat blood glucose response. No correlation was found between higher GI than waxy varieties although the differences did blood glucose response and fat content. not reach statistical difference. Neither in a study by Soh and It is known that the frying process increases the amount of Brand-Miller (1999) were any differences noted between resistant starch (RS) in potato products as measured in vitro three different varieties. Unfortunately, the texture of the (Garcia-Alonso and Goni, 2000). When studying the effect of different tubers was not specified in this study. different food preparation methods on the in vitro digest- The meal with french fries was about 1.5 times larger in the ibility of starch in potatoes, frying of potato slices increased study with carbohydrate-equivalent meals compared with the amount of RS (Kingman and Englyst, 1994). However, the french fries served on energy equivalence. Surprisingly, the type of fat used for frying (lard, vegetable oil or butter) the satiety response was four times higher with the did not affect the digestibility of starch. Although not a larger (AUC 0–180 min ¼ 613) compared with the smaller prerequisite, formation of RS during processing is frequently portion of french fries (AUC 0–180 min ¼ 152). This raises associated with a lowered rate of starch hydrolysis of the the question whether a threshold value in terms of meal size starch bulk, thus causing a lowered glycaemic response. is required before the satiety sensation is affected. According Another possible explanation to the lower glycaemic to most manufacturers, a normal french fries portion is response observed after french fries compared with boiled around 150 g, that is in between the two sizes tested, potatoes served with oil in the present study, may be and it would have been interesting to study satiety response formation of amylose–lipid complexes during frying and/or after a normal sized portion. It should be kept in mind reheating. These complexes are more slowly digested by that in the present studies, subjective satiety rather a-amylase in vitro (Cui and Oates, 1999; Tufvesson et al., than actual food intake was determined and it cannot be 2001) and in vivo (Holm et al., 1983). The lowered glycaemic disregarded that such a measurement may have added response after french fries compared with boiled potatoes valuable information. served with oil is noteworthy, and it would have been Overall, no direct correlation was observed between interesting to study whether this effect can be achieved postprandial blood glucose levels and satiety ranking. with less oil, other types of oils or following other methods However, when comparing energy-equivalent meals, of potato processing (including different surface-area/ the french fries which elicited the lowest glycaemic volume ratios). response (AUC 0–45 min ¼ 35 vs ¼ 75–84 for boiled or In the present study, no differences were observed in mashed potatoes) was less satiating in the early postprandial satiating capacity between the different meals and no phase (AUC 0–45 min ¼ 68 vs ¼ 148–201 for boiled or correlation was found either between blood glucose or mashed potatoes). The connection between blood glucose satiety rating and the composition of the meals. and satiety seem to vary in different studies, some showing a correlation (van Amelsvoort and Weststrate, 1992; Ludwig et al., 1999) whereas others do not (Holt General discussion et al., 1996, 2001). The design of the present studies has the advantage of In the present study, no effect on satiety was seen owing to being more similar to a natural setting in which equivalent GI features as judged from results obtained using equi- amounts of energy or carbohydrates are consumed, but the carbohydrate test meals. A suggested mechanism for the present design does not allow identification of the parameter higher satiating capacity of low-GI foods in general are that that had the highest impact on the satiety sensation. An they are less refined than their high-GI counterparts, and interesting aspect regarding the possible connection between that they owing to this are more difficult to consume, glycaemic response and satiety is the mechanism by which mainly because of a lower energy density (owing to a higher the low-GI features have been achieved; and it could be fibre content and/or water content) and lower palatability hypothesized that certain mechanisms are more effective (Holt et al., 1996). Interestingly, postprandial blood glucose as modulators of satiety than others. Studies with bread AUCs for boiled potatoes turned out to be within the same products added with Na-propionate (Liljeberg et al., 1995; range in both studies as measured in the early postprandial Liljeberg and Bjo¨rck, 1996) or acetic acid (O¨ stman et al., phase; for example AUC 0–70 min ¼ 129 in the energy- 2005) as a means of lowering GI have resulted in higher equivalent study vs AUC 0–70 min ¼ 126 in the study with satiety. In these cases, the lowered glycaemia is mediated carbohydrate-equivalent meals. The carbohydrate contents through a delay in gastric emptying rate. In contrast, in a were equal (50 g) and it thus seems as if it is possible to study with pasta and bread made from exactly the same receive reproducible results when measuring glycaemia after ingredients, no differences could be found neither in meals with boiled potatoes, despite the use of different satiety response nor in ad libitum food intake after 3 h subjects and potato varieties. To our knowledge, there is at (Barkeling et al., 1995). Furthermore, many other preabsorp- present no data in the literature indicating that the firmness tive and postabsorptive signals for satiety exist, such as of the potato tubers should have an influence on satiety. gastrointestinal hormones, and these may also contribute to When it comes to blood glucose responses, Henry et al. the satiety sensation.

European Journal of Clinical Nutrition Glycaemic and satiating properties of potato products M Leeman et al 94 Conclusion Holt S, Miller JB, Petocz P (1996). Interrelationships among post prandial satiety, glucose and insulin responses and changes in Boiled potatoes displayed significantly higher blood glucose subsequent food intake. Eur J Clin Nutr 50, 788–797. Holt SH, Brand-Miller JC, Stitt PA (2001). The effects of equal-energy responses than french fries in healthy subjects whether portions of different breads on blood glucose levels, feelings of compared on energy or carbohydrate basis. The GI obtained fullness and subsequent food intake. J Am Diet Assoc 101, 767–773. for french fries (GI ¼ 77) was lower than previously reported Holt SH, Miller JC, Petocz P, Farmakalidis E (1995). A satiety index of in the literature (GI ¼ 90 and 107; Wolever et al., 1994; common foods. Eur J Clin Nutr 49, 675–690. Kaplan RJ, Greenwood CE (2002). Influence of dietary carbohydrates Fernandes et al., 2005) and was not explained by the fat and glycaemic response on subjective appetite and food intake in content per se. Boiled and mashed potatoes displayed a high healthy elderly persons. Int J Food Sci Nutr 53, 305–316. post-meal subjective satiety when using energy-equivalent Kingman SM, Englyst HN (1994). The influence of food preparation test meals. Meal sizes within the range studied did not methods on the in-vitro digestibility of starch in potatoes. Food Chem 49, 181–186. affect satiety. Kral TV, Roe LS, Rolls BJ (2004). Combined effects of energy density and portion size on energy intake in women. Am J Clin Nutr 79, 962–968. Leeman AM, Ba˚rstro¨m LM, Bjo¨rck IME (2005). In vitro availability of Acknowledgements starch in heat-treated potatoes as related to genotype, weight and storage time. J Sci Food Agr 85, 751–756. The authors thank Giorgia Guizzetti, Kristina Andersson and Liljeberg H, Bjo¨rck I (1994). Bioavailability of starch in bread Yvonne Granfeldt for invaluable assistance. products. Postprandial glucose and insulin responses in healthy subjects and in vitro resistant starch content. Eur J Clin Nutr 48, 151–163. Liljeberg HG, Bjo¨rck IM (1996). Delayed gastric emptying rate as a References potential mechanism for lowered glycemia after eating sourdough bread: studies in humans and rats using test products with added Alvina M, Araya H (2004). Rapid carbohydrate digestion rate organic acids or an organic salt. Am J Clin Nutr 64, 886–893. produced lesser short-term satiety in obese preschool children. Liljeberg HG, Lo¨nner CH, Bjo¨rck IM (1995). Sourdough fermentation Eur J Clin Nutr 58, 637–642. or addition of organic acids or corresponding salts to bread Anderson GH, Catherine NL, Woodend DM, Wolever TM (2002). improves nutritional properties of starch in healthy humans. Inverse association between the effect of carbohydrates on blood JNutr125, 1503–1511. glucose and subsequent short-term food intake in young men. Livsmedelsverket (2001). Livsmedelstabell Energi och na¨ringsa¨mnen Am J Clin Nutr 76, 1023–1030. 2002. Livsmedelsverket: Uppsala. Anderson GH, Woodend D (2003). Effect of glycemic carbohydrates Ludwig DS (2000). Dietary glycemic index and obesity. JNutr130, on short-term satiety and food intake. Nutr Rev 61, S17–S26. 280S–283S. Barkeling B, Granfelt Y, Bjo¨rck I, Ro¨ssner S (1995). Effects of Ludwig DS, Majzoub JA, Al-Zahrani A, Dallal GE, Blanco I, Roberts SB carbohydrates in the form of pasta and bread on food-intake and (1999). High glycemic index foods, overeating, and obesity. satiety in man. Nutr Res 15, 467–476. Pediatrics 103, E26. Collier G, O’Dea K (1983). The effect of coingestion of fat on the Lunetta M, Di Mauro M, Crimi S, Mughini L (1995). Influence of glucose, insulin, and gastric inhibitory polypeptide responses to different cooking processes on the glycaemic response to carbohydrate and protein. Am J Clin Nutr 37, 941–944. potatoes in non-insulin dependent diabetic patients. Diab Nutr Cui R, Oates CG (1999). The effect of amylose-lipid complex Metab 8, 49–53. formation on enzyme susceptibility of sage starch. Food Chem MacIntosh CG, Holt SH, Brand-Miller JC (2003). The degree of 65, 417–425. fat saturation does not alter glycemic, insulinemic or Enghardt Barbieri H, Lindvall C (2003). De svenska na¨ringsrekommen- satiety responses to a starchy staple in healthy men. J Nutr 133, dationerna o¨versatta till livsmedel. Livsmedelsverket: Uppsala. 2577–2580. FAO/WHO (1998). Carbohydrates in human nutrition. Report of a Owen B, Wolever TM (2003). Effect of fat on glycaemic responses in joint FAO/WHO expert consultation. FAO Food and Nutrition Paper normal subjects: a dose-response study. Nutr Res 23, 1341–1347. 66, 1–140. Pawlak DB, Ebbeling CB, Ludwig DS (2002). Should obese patients be Fernandes G, Velangi A, Wolever TM (2005). Glycemic index of counselled to follow a low-glycaemic index diet? Yes Obes Rev 3, potatoes commonly consumed in North America. J Am Diet Assoc 235–243. 105, 557–562. Raben A (2002). Should obese patients be counselled to follow a low- Foster-Powell K, Holt SH, Brand-Miller JC (2002). International table glycaemic index diet? No Obes Rev 3, 245–256. of glycemic index and glycemic load values: 2002. Am J Clin Nutr Roberts SB (2000). High-glycemic index foods, hunger, and obesity: is 76, 5–56. there a connection? Nutr Rev 58, 163–169. Gannon MC, Ercan N, Westphal SA, Nuttall FQ (1993). Effect of Roberts SB, McCrory MA, Saltzman E (2002). The influence of dietary added fat on plasma glucose and insulin response to ingested composition on energy intake and body weight. J Am Coll Nutr 21, potato in individuals with NIDDM. Diabet Care 16, 874–880. 140S–145S. Garcia-Alonso A, Goni I (2000). Effect of processing on potato starch: Rolls BJ, Bell EA (1999). Intake of fat and carbohydrate: role of energy in vitro availability and glycaemic index. Nahrung 44, 19–22. density. Eur J Clin Nutr 53 (Suppl 1), S166–S173. Henry CJ, Lightowler HJ, Strik CM, Storey M (2005). Glycaemic index Rolls BJ, Bell EA, Thorwart ML (1999). Water incorporated into a values for commercially available potatoes in Great Britain. Br J food but not served with a food decreases energy intake in lean Nutr 94, 917–921. women. Am J Clin Nutr 70, 448–455. Holm J, Bjo¨rck I, Drews A, Asp N-G (1986). A rapid method for the Rolls BJ, Bell EA, Waugh BA (2000). Increasing the volume of a food by analysis of starch. Starch/Sta¨rke 38, 224–226. incorporating air affects satiety in men. Am J Clin Nutr 72, 361–368. Holm J, Bjo¨rck I, Ostrowska S, Eliasson AC, Asp NG, Larsson K et al. Soh NL, Brand-Miller J (1999). The glycaemic index of potatoes: the (1983). Digestibility of amylose-lipid complexes in vitro and in vivo. effect of variety, cooking method and maturity. Eur J Clin Nutr 53, Starch/Sta¨rke 35, 294–297. 249–254.

European Journal of Clinical Nutrition Glycaemic and satiating properties of potato products M Leeman et al 95 Spieth LE, Harnish JD, Lenders CM, Raezer LB, Pereira MA, Hangen SJ Wolever TM, Bolognesi C (1996). Prediction of glucose and insulin et al. (2000). A low-glycemic index diet in the treatment of responses of normal subjects after consuming mixed meals pediatric obesity. Arch Pediatr Adolesc Med 154, 947–951. varying in energy, protein, fat, carbohydrate and glycemic index. Tufvesson F, Skrabanja V, Bjo¨rck I, Elmsta˚hl HL, Eliasson AC (2001). J Nutr 126, 2807–2812. Digestibility of starch systems containing amylose-glycerol mono- Wolever TMS, Katzmanrelle L, Jenkins AL, Vuksan V, Josse palmitin complexes. Lebensm-Wiss Technol 34, 131–139. RG, Jenkins DJA (1994). Glycemic index of 102 complex van Amelsvoort JM, Weststrate JA (1992). Amylose-amylopectin ratio carbohydrate foods in patients with diabetes. Nutr Res 14, in a meal affects postprandial variables in male volunteers. Am J 651–669. Clin Nutr 55, 712–718. O¨ stman E, Granfeldt Y, Persson L, Bjo¨rck I (2005). Vinegar Warren JM, Henry CJK, Simonite V (2003). Low glycemic index supplementation lowers glucose and insulin responses and breakfasts and reduced food intake in preadolescent children. increases satiety after a bread meal in healthy subjects. Eur J Clin Pediatrics 112, E414–E419. Nutr 59, 983–988.

European Journal of Clinical Nutrition