Apparent nutrient digestibility and gastrointestinal evacuation time in European seabass ( Dicentrarchus labrax) fed diets containing different levels of legumes

Aquaculture 289 (2009) 106–112

Contents lists available at ScienceDirect

Aquaculture

j ourna l homepage: www.e lsev ie r.com/ locate /aqua-on l ine

Apparent nutrient digestibility and gastrointestinal evacuation time in Europeanseabass (Dicentrarchus labrax) fed diets containing different levels of legumes

Styliani Adamidou a,⁎, Ioannis Nengas b, Maria Alexis b, Eleni Foundoulaki b, Dimitra Nikolopoulou b,Patrick Campbell c, Ioannis Karacostas d, George Rigos b, Gordon J. Bell a, Kim Jauncey a

a Institute of Aquaculture, University of Stirling, Stirling, Scotland FK9 4LA, UKb Hellenic Centre for Marine Research, Institute of Aquaculture, Agios Kosmas, 16777 Athens, Hellas, Greecec BioMar U.K. Grangemouth, Scotland, UKd BioMar Hellenic S.A. Block No. 6 2nd I.Z. of Volos, GR-37500, Velestino, Greece

⁎ Corresponding author. Tel.: +44 1786467892; fax: +E-mail address: [email protected] (S. Ada

0044-8486/$ – see front matter © 2009 Elsevier B.V. Adoi:10.1016/j.aquaculture.2009.01.015

a b s t r a c t

Article history:

Keywords:Wheat substitutionLegumesGastrointestinal evacuationGlucose loadPhysical properties

astrointestinal evacuation time for diets containing two levels of legumes weredetermined in European seabass (initial wt. 150 g). Seven isonitrogenous (44 g crude protein 100 g−1) andisoenergetic (20 kJ g−1) extruded diets were tested. Experimental diets contained 15 g or 30 g 100 g−1 of eachlegume including faba bean meal, (FB15, FB30), chickpea meal (CP15, CP30), field pea meal (FP15, FP30) and acontrol, wheat meal containing, diet. Inclusion of faba beans in diets significantly enhanced apparentdigestibility coefficients (ADCs) of dietary protein and energy at both inclusion levels. Starch digestibility wassignificantly lower for the control and FP30 diet and significantly higher for the FB15 diet. Fat digestibilitycoefficients were elevated significantly for FB15, FB30, CP15, FP15 and FP30 diets, compared to the CP30 andthe control diet. The incorporation of faba bean and chickpea greatly affected gastrointestinal evacuationtime compared to fish fed the control diet. Longer passage times were observed when these legumes wereincluded in the diet. Glucose serum peak value was delayed for FB30 and a slower decreasing rate wasobserved for the legume diets. The addition of legumes had a positive effect on physical properties of thepellets by increasing the hardness while water activity remained within the safety limits. The findings of thisstudy suggest that the legumes tested are potential candidates for carbohydrate replacement (wheat) and, toa lesser extent, for protein substitution in diets for European seabass.

© 2009 Elsevier B.V. All rights reserved.

1. Introduction

Wheat is an important ingredient and the main source ofcarbohydrate in aquafeeds. Wheat starch is a good binder and is alsoconsidered a good source of energy (Hertrampf and Piedad-Pascal,2000). However, recent expansion of global biofuel industries has ledto increased demands for starch and sugar-rich feedstuffs with conse-quent price increases for such crops (Runge and Senauer, 2007).Indeed, wheat is expected to undergo further price increases in thefuture (Runge and Senauer, 2007).

Furthermore, limited European production of soybeans (Grainlegumes, 2006) and the general position of EU countries with respectto genetically modified soybeans reinforce the importance of plantsources other than soybeans in aquafeeds (Euroactiv, 2006).

Legumes represent a significant constituent of global grain pro-duction (Grain legumes, 2007) and are highly valued feed componentsfor both ruminants and monogastric livestock due to their relativelyhigh protein and energy values compared to cereal grains (Petterson

44 1786472133.midou).

ll rights reserved.

and Mackintosh, 1994). Several legumes, including field peas, chick-peas and faba beans, may be able to replace wheat as a source ofcarbohydrate since they are characterized by high carbohydratecontent (Knudsen and Bach, 1997). Moreover, the protein content oflegumes may enable partial replacement of fish meal (Gouveia andDavies, 1998), which is rapidly increasing in price (Josupeit, 2008).

A drawback to the use of legumes in fish diets is the presence of avariety of endogenous antinutritional factors including proteolyticenzyme inhibitors, amylase inhibitor, phytic acid, tannins and lectinsthat adversely affect enzyme action or the absorption of minerals andother nutrients (Tacon, 1993). Antinutritional factors in legume seedscan be partially or totally inactivated by heat processing such as roast-ing, autoclaving, extruding or cooking or by other processing methodssuch as soaking, germination or enzyme addition, prior to inclusion infish feeds (Francis et al., 2001). Digestibility coefficients either for eachlegume or for diets containing the legumes have been investigated.Field peas have been evaluated as potential feed ingredients, wholeor dehulled, raw or processed, for several aquatic species includingEuropean seabass (Gouveia and Davies, 1998, 2000), Australian silverperch (Allan et al., 2000), Atlantic salmon (Carter and Hauler, 2000),rainbow trout (Gomes et al., 1995a) and turbot (Burel et al., 2000).

Table 2Diet composition

Ingredient (g 100 g−1) Control B15 CP15 P15 B30 CP30 P30

Fish meal 70%, Norway 61.4 57.1 57.9 57.7 54.1 55.0 55.3Wheat 24.2 13.1 12.9 12.5 1.3Faba beans 15.0 30.0Chickpeas 15.0 30.0Field peas 15.0 30.0Fish oil SA, Peru 14.1 14.5 13.9 14.5 15.6 13.4 14.4Premixa 0.3 0.3 0.3 0.3 0.3 0.3 0.3

Proximate composition (%)Ash 9.5 10.1 9.6 10.0 10.0 9.8 9.9Crude fat 17.3 18.1 17.6 20.1 20.8 19.4 18.9Crude protein 42.3 43.8 42.9 42.9 44.3 44.0 43.1Total starch 15.8 14.8 14.2 14.6 11.7 12.4 13.3Wheat starchb 15.8 8.9 8.8 7.8 0.0 0.9 0.0Legume starchb 0.0 5.9 5.3 6.8 11.7 11.5 13.3Total NSPc 5.0 5.4 4.8 5.2 6.9 5.1 6.1S-NSPd 1.6 2.5 1.5 1.8 2.4 1.5 2.3I-NSPe 3.4 2.6 3.4 3.6 3.7 3.6 4.6Total carbohydrate 20.8 20.2 19.0 19.7 18.7 17.5 19.4Water content 6.7 6.3 7.4 6.6 8.3 6.8 7.3Energy (kJg−1) 20.1 19.5 20.6 20.7 20.2 19.9 20.1

a Vitamins (perkgofdry feed): vitaminA3750 IU, cholocarciferol (D3) 750 IU,α-tocopherol(E) 131.3 mg, thiamine 7.5 mg, riboflavin 15 mg, pyridoxine 7.5 mg, Vitamin B12 2.25μg, andvitamin K3 7.5 mg. Minerals (per kg of dry feed): zinc 75 mg, iodine 0.9 mg, copper 3.75 mg,manganese 22.5 mg, cobalt 0.75 mg, and selenium mineral 0.19 mg. Mineral premix alsoincludes Y2O3 to give final concentration of 0.1% in the diet.

b Calculated from total starch values and the concentration of each ingredient in thediet.

c,d,e Total, soluble and insoluble non starch polysaccharides.

107S. Adamidou et al. / Aquaculture 289 (2009) 106–112

However, there is less published research for faba beans and chick-peas, specifically for Australian silver perch (Allan et al., 2000; Boothet al., 2001) and rainbow trout fed diets including faba beans (Gouveiaet al., 1993; Gomes et al., 1995b).

Processing is also known to gelatinize starch. Wheat when gelatin-ized is known to be highly digestible by European sea bass (Peres andOliva-Teles, 2002), while legume starches are known to be digested ata slower rate in humans (Tharanathan and Mahadevamma, 2003).

Gastrointestinal evacuation reflects the time needed for an animalto digest a meal and is usually affected by a number of parameterssuch as the fish species and its physiology as well as the physical andchemical composition of the diet (Jobling, 1987). Gastrointestinalevacuation rate has proven to be an important parameter for modell-ing daily feed intake (Jobling, 1981) as is partly responsible for thecontrol of fish appetite (Riche et al., 2004).

Different ingredients including whole grains or starch sources ingeneral can undoubtedly affect the physical characteristics of thepellets (Thomas et al., 1998; Briggs et al., 1999) and these in turn candifferentiate the digestion procedure. For example hydration of thepellets in the stomach is depended on the texture of the diet and isclosely connected to the gastric emptying (Grove et al., 2001).

The objectives of the present study were to evaluate the effect ofpartial or total replacement of wheat by whole grain flours of fieldpeas, chickpeas and faba beans, on diet digestibility, gastrointestinalevacuation time and glucose loading for European sea bass. Thephysical characteristics of the pellets were also determined for thetested diets.

2. Materials and methods

2.1. Feed ingredients and feed formulation

Three different legumes, including field peas (Pisum sativum L.),chickpeas (Cicer arietinum L.) (cultivated in Greece, in the area ofThessaly) and faba beans (Vicia faba L.) (cultivated in Denmark), wereincluded in the experimental diets. Proximate composition of the rawmaterials is given in Table 1. Six experimental diets were formulatedto be isoenergetic and isonitrogenous containing 15 g 100 g−1 (CP15,FP15 and FB15) and 30 g 100 g−1 (CP30, FP30 and FB30) of each legumeplus a control diet (Table 2). All diets weremanufactured at the BioMarTec Centre in Brande, Denmark, using a twin-screw extruder (ClextralBC 45, screw speed 340 rpm, outlet die temperature 70 °C). The dietmixture was pre-conditioned before entering the extruder, dried ina six-level drier, coated with fat under vacuum and cooled. Thelegumes were ground to 1.5 mm and the pellet diameter intended tobe 4.5 mm. All diets contained 0.1% yttrium oxide (Y2O3) as an indi-gestible marker.

Table 1Proximate composition of the feedstuffs

g 100 g−1 DM Chickpeaextruded

Faba beanextruded

Field peaextruded

Wheatextruded

Fishmeal

Ash 3.6 4.5 3.7 1.9 15.3Crude protein 27.5 32.1 25.6 14.8 76.3Crude fat 5.1 2.0 1.9 2.0 10.8Starch 47.4 38.1 43.2 70.6 –aS-NSP 3.8 4.9 4.6 3.0 –bI-NSP 9.8 15.6 10.9 7.3 –cTotal NSP 13.6 20.5 15.5 10.3 –dRSO 4.68 4.40 5.52 1.40 –

Total 101.9 101.6 95.4 101.0 102.4Tannins (%) 0.59 1.14 0.67 0.67 –eTI(mg TI/g) 1.63 1.24 1.26 1.26 –

a S-NSP: Soluble non starch polysaccharides.b I-NSP: Insoluble non starch polysaccharides.c Total NSP: Total non starch polysaccharides.d RSO: Raffinose-series oligosaccharides.e TI: Trypsin Inhibitors.

2.2. Fish and experimental design

Four hundred twenty European sea bass (Dicentrarchus labrax)of initial weight 152.3±12.1 g were transferred from a commercialfarm (located in Epidavros, Greece) to the facilities of the laboratory ofFish Nutrition and Pathology, Institute of Aquaculutre, Hellenic Centrefor Marine Research in Athens. A total of 20 fish were stocked ran-domly into 21×250 L cylindroconical tanks and acclimatized fortwo months. Digestibility trials were carried out in triplicate. Faeceswere collected in a trap using a modification of the Guelph method(Cho et al., 1982). The faecal trap was surrounded with ice duringfaecal settlement to minimize bacterial degradation. Sea bass werefed daily at 1% of their biomass, twice a day (09.00 and 16.00 h). Faecalsamples were removed each morning for 8 days, prior to feeding.Faeces were centrifuged and kept at −20 °C until they were freeze-dried.

2.3. Water quality

Seawater was supplied continuously at 150L h−1 from the sea andmechanically filtered (5 μm) before entering the tanks. Water temper-ature was controlled at 18±1°C during the experimental period. Indi-vidual stone diffusers were located in each tank to provide sufficientoxygen. Oxygen concentration was approximately 8±1 mg L−1, pHranged from 7.8 to 8.0, total ammonia levels from 0.2 to 0.3 mg L−1 andsalinity 38 ‰.

2.4. Calculation of digestibility coefficients

Apparent digestibility coefficients (ADC) for the control and testdiets were calculated according to the formulae:

ADCk = 100× 1− F×Dyð Þ= D×Fyð Þ½ �;

where F=nutrient concentration or energy in faeces, D=nutrientconcentration or energy in diet, Dy=Yttrium in concentration diet andFy=Yttrium concentration in faeces.

Table 3Apparent digestibility coefficients (ADC) for protein, fat, starch and energy

Diet % ADC protein % ADC fat % ADC starch % ADC energy

Control 92.9±0.3a 96.7±0.6ab 94.0±1.9ab 94.3±0.5a

CP15 93.2±0.4ab 97.2±0.4bc 96.4±0.6cd 95.0±0.3b

FP15 93.9±0.2bc 97.7±0.2c 96.4±0.5cd 95.6±0.2bc

FB15 94.8±0.2d 97.7±0.1c 97.0±0.7d 96.0±0.03c

CP30 93.0±0.7a 96.2±0.5a 94.5±0.2bc 94.2±0.4a

FP30 92.9±0.5a 97.0±0.2bc 92.3±1.1a 94.2±0.3a

FB30 94.2±0.2cd 97.7±0.1c 95.7±0.2bcd 95.6±0.1bc

Different superscripts in each column show significant differences among the valuesaccording to Tukey test (P≤0.05). Values are given as means±SD (n=3).

108 S. Adamidou et al. / Aquaculture 289 (2009) 106–112

Energy in diets and faeces was determined according to theformula of Blaxter (1989):

Energy kJg−1� �

= 23:6×P + 17:3×S + 39:5×F;

where P=% protein in faeces or in diet, S=% starch in faeces or in diet,F=% fat in faeces or in diet.

2.5. Gastrointestinal evacuation and blood sampling

Following the digestibility trials, 4 diets were selected (control,CP30, FB30 and P30) for gastrointestinal evacuation time studies. Fishwere fasted for 72 h before being fed to make sure that the gastro-intestinal tract was empty. Six fish per treatment were sacrificed at 0,8, 16, 24, 32 and 48 h after feeding a single meal equal to 1% biomass,following phenoxyethanol/ethanol (1/1, 0.4 ml L−1) induced anaesthe-sia. The time intervals were selected based on test sampling and takinginto account the low temperature as proposed by Finstad (2005).

The method used was the serial slaughter technique. This methodinvolves sacrificing fish at regular time intervals after been fed. Bloodsamples were taken, fish were killed with a blow to the head and thedigestive tract was carefully removed and separated into three parts:stomach, foregut and hindgut. Foregut was defined as the section fromthe pyloric sphincter to the ileocaecal valve and hindgut from theileocaecal valve to the anus. The gut contents were removed andplaced in pre-weighed tubes. Samples were freeze-dried and dryweights were used to estimate the gastric evacuation time in com-parison to the body weight of each fish.

Geometric means of stomach and intestinal contents wereregressed against time in order to examine a possible fit to a modelfor calculating time and rate of gastric and intestinal evacuation.

2.5.1. StomachData from all treatments in the case of stomach and gastric evacua-

tion rates were calculated using the formula described by Elliot (1972)

Wt = A−rt

Where Wt is the geometric mean weight of stomach dry matterdigesta at time t, A is a parameter calculated from the formula of theregression and r is the rate of gastric evacuation. Data were plotted in alogarithmic form as

LnWt = 1nA−rt;

The slope of r corresponding to GERwas calculated. Gastric evacua-tion time (GET) was calculated from the above equation as the time (t)necessary to empty 50, 75 and 90% of the stomach contents.

2.5.2. ForegutThe time needed for the digesta to completely pass the foregut was

estimated from the intercept on the x axis and the rate of evacuationby the slope of each equation.

2.5.3. HindgutPoints from each curve were estimated to determine the time of

maximum hindgut content and evacuation time.

2.6. Chemical analysis

Proximate composition of the extruded ingredients, feeds andfreeze-dried faeces was determined as follows. Dry matter and ashwere determined according to AOAC (1990) and crude protein accord-ing to the Kjeldahl method. Starch was measured by an enzymaticmethod (Megazyme Total Starch Assay kit, Megazyme International,Ireland) using thermostable α-amylase and amyloglucosidase. Totalfat in extruded feeds and ingredients was determined by ether extrac-tion after being hydrolysed in acid and total lipid in faeces by the

phosphovanillinmethod (Nengas et al., 1995). Total and insoluble non-starch polysaccharides (NSP) were determined spectrophotometri-cally as described by Englyst et al. (1994) and soluble NSP calculatedby subtracting insoluble from total NSP. Oligosaccharides were alsomeasured spectrophotometrically (Magazyme Raffinose–series oligo-saccharides/D-Glucose assay kit, Megazyme, International, Ireland).Blood samples were centrifuged and plasma was analyzed for glucoseby the glucose oxidase method (GOD-PAP, cat. No 10260, HUMAN).Trypsin inhibitor activity was determined according to Smith et al.(1980). Measurement of total tannins was based on the Prussian Bluemethod as described by Budini et al. (1980). Yttrium oxide was deter-mined by ICP-MS according to Refstie et al. (1997).

2.7. Physical characteristics of the pellets

Determination of the texture was performed on each pellet of theexperimental feeds, using a Stable Micro Systems TA-XT2 Texture Anal-yser. For sample, a blade was placed on the texture analyzer and cutthrough the sample at a speed of 1 mm s−1. Twelve pellets were exam-ined for each diet and the shear force was measured (Newton) as themaximum force required for cutting through the samples thatwas equalto the peak height. For the determination of water activity (aw), Hygro-meter A3, Rotronic was used. Pellet density d, measured by the formulad=m/V, where,m is themass as determined in an analytical balance andV the volume of the pellet. Dimensions of 12 pellets, chosen to have theshape of a well formulated cylinder, were determined with a calliper.

Water absorption of the pellets was measured by immersing 15pellets in seawater in triplicate groups for 5 s and then the moisturecontentwasdetermined. Totalmoisture (%)was calculated including theinitial moisture of the pellets.

In avolumetric cylinderof 1 L and34cmheight, seawater (of 20 ˚Candsalinity 38)wasaddedandthepellets allowed to fall through thewater tothe bottom of the cylinder. The time (t) needed for each pellet to gothrough 34 cm of the cylinder was measured and the velocity (v) of thepellets of each diet determined according to the formula v=34 t−1 cm s−1.

2.8. Statistical analysis

Digestibility coefficients and glucose values were analysed by one-way ANOVA and significant differences were determined by the Tukeypost hoc test using SPSS 13.0 at a significance level of P=0.05.

Regressions were performed using the statistical program Stat-graphics Plus 2.1. The linear fits and comparison of slopeswere tested ata significance level of P=0.05. Different parameters were correlated by aPearson 2-tailed significance correlation.

3. Results

3.1. Digestibility

ADCs for the nutrients tested are summarized in Table 3. Digest-ibility coefficients were high for all diets, but significant differenceswere revealed among diets after applying Tukey’s post hoc test.

Fig. 3. Foregut linear relationship curves and equations for dry matter % of body weight(DM %BW) over the 48 h sampling period in fish fed the four experimental diets.

Fig. 2. Stomach curves plotted linearly in a logarithmic form show the gastric evacua-tion time for dry matter % of body weight over the 32 h sampling period in fish fed thefour experimental diets.

Fig. 1. Stomach exponential curves and equations show gastric evacuation time of drymatter % of body weight (DM %BW) over the 32 h sampling period in fish fed the fourexperimental diets.

109S. Adamidou et al. / Aquaculture 289 (2009) 106–112

Protein ADC was significantly lower for the Control and higher forFB15 diet. Starch digestibility was significantly lower for the FP30(92.3%) and Control diets and significantly improved for all three 15%diets. Finally, energy ADCs showed similar trends to starch ADCsdemonstrating best values for FB15 (96%).

3.2. Gastrointestinal evacuation

Gastric evacuation time (GET) was evaluated as the geometricmean of dry matter of feed per g of body weight of each fish using anexponential model (Fig. 1). The values observed immediately afterfeeding (0 h) were lower than 1% of body weight partially because ofthe calculation of dry matter and partially because the sampled fishhad eaten slightly less than 1%.

Table 4 presents GET in relation to the percentage of feed re-maining in the stomach. The Control diet required 5.3 h to evacuatehalf of the initial feed content, followed by FP30, 5.6 h, FB30 with 8.0 hand still longer for CP30, 9.8 h. Comparison of the slopes for GastricEvacuation (Fig. 2) indicated significantly longer GET for CP30 com-pared to the Control and FP30 diets, while FB30 did not differ from allother diet emptying times.

3.3. Foregut evacuation

The data extracted for the foregut fitted a linear model. Accordingto the equations (Fig. 3), foregut was evacuated in 31.7 h for fish fedthe Control diet, 38.2 h for CP30, 39.5 h for FP30 and 46.5 h for FB30.

3.4. Hindgut evacuation

Hindgut data best fitted a quadratic model. According to the equa-tions (Fig. 4), evacuation was estimated to be 34.2 h for fish fed theControl diet, 41.4 h for FP30, 51.9 h for CP30 and 51.7 for FB30.

3.5. Serum glucose

In the fist sampling (0 h after feeding) serum glucose values rangedfrom 4.56 to 5.30 mmol dl−1 (Table 5). After 8 h fish fed the wheat-based Control diet showed the most rapid increase that became sig-

Table 4The gastric evacuation time (GET, h), foregut evacuation time (FGET), hindgut evacua-tion time (HGET) for European sea bass fed the diets and correlation coefficient (r) of theregression lines based on exponential model of gastric evacuation

Gastric evacuation GET, 50% GET, 75% GET, 90% r

Control 5.3 10.7 17.7 −0.9892FP30 5.6 11.2 18.6 −0.9696FB30 8.0 15.9 26.4 −0.9849CP30 9.8 19.6 32.6 −0.9753

nificant at 16 h. Diets CP30, FP30 and FB30 showed significant increaseafter 24 h, while FB30 showed a delay with glucose starting to increaseafter 8 h. After 24 h the Control, CP30 and FP30 diets exhibited a peak,while diet FB30 showed the highest value after 32 h. Fish fed the

Fig. 4. Hindgut quadratic curves and equations describe the evacuation of dry matter %of body weight (DM %BW) over the 48 h sampling period in fish fed the four experi-mental diets.

Table 7Correlation coefficients among physical characteristics of the pellets

Hardness (Newton) Density (g cm3) Sinking rate Water absorption

Density 0.168Settling v. 0.101 0.362⁎⁎Water abs −0.295 −0.444⁎ −0.447⁎Moisture −0.062 −0.240 0.219 −0.133

⁎, ⁎⁎Significant at Pb0.05 and Pb0.01, respectively.

Table 5Plasma glucose values (mmol/dl) measured over 48 h, in 8 h intervals for diets control,FP30, FB30 and CP30

Plasma glucose (mmol/dl)

Hours⁎ Control FP30 CP30 FB30

0 4.87±0.53ab 5.04±1.83a 5.30±1.13ab 4.56±0.97a

8 5.67±1.62b 5.39±1.31ab 5.30±0.95a 4.45±0.72a

16 7.14±1.49c 5.80±0.02ab 6.18±1.40abc 5.05±0.51ab

24 7.24±1.45c 6.69±1.18b 7.27±1.31c 6.02±0.87b

32 4.10±0.54a 5.47±1.52ab 7.15±1.01bc 6.17±1.23b

48 4.94±0.53ab 5.18±1.31a 5.10±1.94a 4.67±0.31a

⁎Hours of sampling after feeding.Different superscripts in each column show significant differences among the valuesaccording to Tukey test (P≤0.05).Values are given as means±SD (n=6).

110 S. Adamidou et al. / Aquaculture 289 (2009) 106–112

Control diet showed a significant decrease of serum glucose at 32 hin contrast to the rest of the fish fed the legume diets revealing a‘smoother’ glucose decrease. Statistical analysis also showed that fishfed the legume diets showed no significant differences between 16and 32 h at glucose levels.

3.6. Physical characteristics of the pellets

Physical characteristics of the pellets are presented in Table 6.Pellets of the FB30 diet were significantly harder than both chickpeadiets while no other significant difference on this parameter wasdetected. Diet FB30 had significantly less total water after immersingthan the FB15, FP15 and CP30 diets, while the Control diet showedno differences compared to any of the diets. The quicker sinking ratemeasured for the Control and FP30 diets was significantly differentfrom the CP15, CP30 and FB30 diets. Significantly higher density wasmeasured for the Control diet, followed by the FP, FB and CP diets. Astrong positive correlation (Pb0.01) between pellet density and sink-ing ratewas found aswas expected and a negative correlation betweensinking rate–water absorption and density–water absorption occurred(Table 7).

4. Discussion

In the present study all three legumes tested proved promising ascarbohydrate sources at both inclusion levels tested in European seabass as indicated by their ADCs with best values measured for the fababean diets. Similar studies evaluating legumes for carbohydrate re-placement in fish diets are very limited in the literature. Gouveiaand Davies (1998) tested diets containing 20 and 40% field peas inEuropean sea bass and showed no differences for protein ADCs, butcarbohydrate digestibility was significantly lower for a 40% substi-

Table 6Physical characteristics of the pellets

1Hardness(N)

2Density(g cm−3)

3Sinking rate(cm s−1)

4aw at23 ˚C

5Waterabsorption(%)

Dietmoisture(%)

Control 12.4±0.6ab 1.27±0.02d 11.2±0.9c 0.61 11.8±0.6abcd 6.7±0.5a

CP 15 11.9±0.2a 1.04±0.03a 6.7±0.6a 0.53 11.4±0.5abc 6.6±0.5a

CP 30 11.1±0.6a 0.99±0.03a 7.4±0.6a 0.52 12.5±0.5cd 7.3±0.5ab

FP 15 12.3±0.4ab 1.16±0.01c 9.5±0.9bc 0.57 12.0±0.1bcd 6.3±0.5a

FP 30 12.3±0.7ab 1.13±0.01c 10.7±0.6c 0.54 11.0±0.2ab 8.3±0.7b

FB 15 12.4±0.7ab 1.06±0.03ab 10.3±0.2bc 0.59 13.0±0.3d 7.4±0.6b

FB 30 13.8±0.8b 1.13±0.03bc 8.4±0.7ab 0.53 10.7±0.6a 6.8±0.4ab

Different superscripts in each column show significant differences among the valuesaccording to Tukey test (P≤0.05). Values expressed as mean±SD.an=12 treatments per diet (Newton).bn=12 treatments per diet (g cm−3), n=3 treatments per diet.cSinking rate n=12 treatment per diet (cm s−1).dWater activity n=1 per diet.eWater absorption (total moisture %).

tution level. The differences in energy and carbohydrate digestibilityfor field pea diets between the above and the present study (64.8 vs.96.4% for low levels and 56.7 vs. 92.3% for high levels) can be attrib-uted to different processing of the diets used but also to the determi-nation of carbohydrate instead of starch evaluated in the present study.Extrusion, as opposed to compressed pellets, was used in the presentstudy, possibly resulting in enhanced carbohydrate digestibility andthus improving energy utilization. The same authors in a more recentstudy in European sea bass (Gouveia andDavies, 2000) tested extrudedpea seedmeal in diets up to 30% and all nutrient ADCs did not differ forthe different inclusion levels, while both protein and carbohydrateADCs were higher compared to their previous research (Gouveia andDavies, 1998). When pea seed meal was included at 18% and substi-tuted for wheat in Atlantic salmon extruded diets (Aslaksen et al.,2007), all nutrient digestibilities were lower compared to those foundin the present study, while the starch ADC was lower compared to thepresent findings. The same authors also evaluated 22% faba beansubstitution for wheat and found similar values to the Control ADC,contrary to lower starchADC,which disagreeswith the current results.Gouveia et al. (1993) partially replaced fish meal and dextrinwith 26%of faba beans in rainbow trout diets, but no differences were found forprotein or fat ADCs in contrast to the improvement of the sameADCs inthe present study. No similar studies appear to be available for chick-peas, with the exception of Booth et al. (2001) who investigated thedigestibility in Australian silver perch and found values of 88.1 and72.1% for protein and energy ADCs, respectively.

There is little published information regarding the ability of dif-ferent starch sources to influence utilization of other nutrients inthe diet. It seems as though more easily digestible starch (gelatinized)does not influence protein digestibility but does improve energy andstarch digestibility (Peres and Oliva-Teles, 2002). In the present study,higher starch ADC was observed for diets combining wheat andlegume starch (with higher values being observed for FB15) ratherthan the diets including only wheat or any of the tested legumes. Apositive correlationwas found among the ADCs for each diet, with thediets showing overall lower or higher ADCs.

Non starch polysaccharides in monogastric animals can delayintestinal absorption of glucose, possibly through a reduced rate ofgastric emptying, leading to delayed absorption (Knudsen and Bach,2001). Furthermore, soluble dietary NSP can increase the viscosity ofthe stomach contents (Storebakken, 1985). In the current study, total,soluble and insoluble NSP did not play any significant role in starchdigestibility, however, the diet with the higher inclusion level of totalNSP (FB30) showed the longest evacuation time for the gut and one ofthe highest values for starch digestibility.

The majority of gastric evacuation studies have focused on inves-tigating the nutritional habits of wild fish by creating specific models(Andersen and Beyer, 2005), comparing pelleted and prey feed(Andrade et al., 1996) or investigating the effects of prey size (DosSantos and Jobling, 1991) and temperature (Finstad, 2005). Most ofthese studies have shown that the relationship between food remain-ing in the gastrointestinal tract and time can be described by a linear,square root or exponential equation (Bromley, 1994). In the presentstudy, the mathematical models giving the best fit were found to bedifferent for each part of the tract for all tested diets being exponentialfor the stomach, linear for the foregut and quadratic for the hindgut.Stomach curves followed the formula described by Elliot (1972) and

111S. Adamidou et al. / Aquaculture 289 (2009) 106–112

supported by Persson (1986) as a model, fitting better for voluntaryfeeding. An exponential model was also used by Venou et al. (2003)and Sveier et al. (1999) to express gastric evacuation time of giltheadseabream, and Atlantic salmon, respectively, however no comparabledata are available for foregut and hindgut models.

High standard deviations were observed for the hindgut dryweight values in the present study, due to the small length of thehindgut and the low quantity of the sample obtained as well as anincreased mucus content present in this part of the intestine. Hindgutquadratic curves showed that this part of the intestine accumulatesthe indigestible contents of the foregut which reaches a maximumlevel and defecation starts or continues at a faster rate. Sveier et al.(1999) found a similar trend for Atlantic salmon hindgut evacuationtime, although the curves were plotted against an external marker andno model was applied for the data.

It has been assumed that in evacuation rate studies, it is preferableto measure the meal size of each individual fish (Bromley, 1994).During communal feeding, it is difficult to control the amount of feedeaten by individual fish but the results can be less biased than whenfish are fed individually or forced fed (Bromley, 1994). For this reason,the variance in initial stomach content is expected to be high (Jensenand Berg,1993; Bromley,1994). These initial variations inmeal size arealso reflected in the variations measured in the amount of feed re-trieved from the different parts of the digestive tract. Such substantialvariations were also observed in the current study, but in order toavoid extreme values when a fish was overfed or under-fed, thesefish were rejected and immediately replaced by another during thesampling procedure.

A single meal experimental protocol was followed in the presenttrial, like most similar published GET experiments (Rouhonen et al.,1997; Sveier et al., 1999; Sweka et al., 2004). The results presented inthe present study give a comparative evaluation for the different diets,although absolute values would be different in multiple meal, realproduction processes.

In practical terms, inclusion of legumes in diets for European seabass clearly increased the GET of the feed with faba beans having thestrongest effect. NSP inclusion level may have caused this delay(Knudsen and Bach, 2001), also in the hindgut. Different extrudedstarch sources of wheat and corn result in different gastric evacuationrates for gilthead sea bream (Venou et al., 2003). It is possible that thestructure of the extruded starchmolecules of the tested legumes couldresult in this delay. Along with this observation, digestibility coef-ficients were not affected by the gastrointestinal evacuation time andneither was it for Atlantic salmon (Sveier et al., 1999).

The rapidness of starch digestion can be reflected in the glucoseloading in blood after each meal (Hemre and Hansen, 1998). Theperiod needed for the fish to recover from glucose loading depends onthe starch level and the water temperature (Hemre et al., 1995),whereas serum glucose levels in European sea bass has only beeninvestigated after glucose injection (Peres et al., 1999), but not afterfeeding different starch sources. Legume starches promoted slow andmoderate postprandial glucose and insulin responses and results inlower glycaemic indexes in humans (Tharanathan and Mahadevamma,2003), while fish seem to have a longer postprandial hyperglycemiaafter carbohydrate intake (Brauge et al., 1994). Peres et al. (1999), in aglucose tolerance test of European sea bass at 22 °C, found a largerrange between the lower and higher values and peaked 3 to 6 h after aglucose injection, while in a similar test with Atlantic salmon at 2–3 °C(Hemre and Hansen, 1998), the peak was found 3 h after glucoseinjection. The present study was performed to evaluate the effect ofdifferent starch sources on postprandial glucose levels and thus theresults are not comparable to the above mentioned studies. The in-crease of evacuation time in the foregut could also have affected theglucose absorption rate. Fish fed the Control, CP30 and FP30 dietsevacuated 90% of the digesta after 32, 38 and 39 h, respectively, whileserum glucose rapidly decreased for fish fed the wheat-starch diet,

however a delay in serum glucose reduction was observed in fish fedthe pea and chickpea starch diets. Fish fed diet FB30 evacuated thedigesta after 46 h, showed better serum glucose regulation, with adelay in serum glucose load and a lower peak value, indicating a pos-sible strong correlation of these two findings.

Experimental diet pellets were not affected negatively by theaddition of low or high levels of legumes with respect to hardness,settling velocity, water activity or pellet density. High inclusion of fababeans in the experimental diets led to an increase in hardness as alsofound for soybean meal when included in fish feed diets (Sorensenet al., 2009), This result can be considered a positive effect on pelletquality as harder pellets are possibly more durable (Sorensen et al.,2009) and could be used in automatic feeders with less losses. Wateractivity values were within the safety limits (Schwertner et al., 2003)for all diets indicating that long storage even at room temperature ispossible without microbial development and spoilage of the pellets.

In conclusion, all three legumes tested proved to have potentialas feed ingredients for European sea bass, mainly as carbohydratereplacers for wheat and to a lesser extent for protein substitution.Digestibility coefficients showed satisfactory values for all diets eitherincluding low or high levels of legumes. In addition, starch digest-ibility was slightly improved for the diets that combined wheat anda legume starch source. Gastric evacuation was significantly delayedby the inclusion of chickpeas, while foregut evacuation rates werereduced for all legume diets with faba beans showing the strongesteffect. Glucose levels in European sea bass serumwas also affected bythe type of carbohydrate ingested with wheat starch showing themost rapid increase and decrease in serum glucose compared to fishfed pea and chickpea diets, while faba bean starch had a delay in theserum glucose peak and a lower range of glucose values. The additionof legumes in European sea bass feed pellets had a positive effect byincreasing the hardness of the pellets.

Acknowledgements

This study was supported by Greek Scholarship Foundation (IKY),BioMar SA, Hellenic Centre for Marine Research in Athens and theUniversity of Stirling. Special thanks to Dr. Petros Taoukis and Dr. FaniTsironi for their assistance and guidance in the lab of NationalTechnical University of Athens, School of Chemical Engineering.

References

Allan, G.L., Parkinson, S., Booth, M.A., Stone, D.A.J., Rowland, S.J., Frances, J., Warner-Smith, R., 2000. Replacement of fish meal in diets for Australian silver perch,Bidyanus bidyanus: I. Digestibility of alternative ingredients. Aquaculture 186,293–310.

Andersen, N.G., Beyer, J.E., 2005. Gastric evacuation of mixed stomach contents inpredatory gadoids: an expanded application of the square root model to estimatefood rations. J. Fish Biol. 67, 1413–1433.

Andrade, P.J., Erzini, K., Palma, J., 1996. Gastric evacuation and feeding in the gilthead seabream reared under semi intensive conditions. Aquacult. Int. 4, 129–141.

AOAC, 1990. Official Methods of Analysis of the Association of the Official AnalyticalChemists, 15th edn. AOAC, Arlington, VA, USA.

Aslaksen, M.A., Kraugerud, O.F., Penn, M., Svihus, B., Denstadli, V., Jørgensen, H.Y.,Hillestad, M., Krogdahl, Å., Storebakken, T., 2007. Screening of nutrient digest-ibilities and intestinal pathologies in Atlantic salmon, Salmo salar, fed diets withlegumes, oilseeds, or cereals. Aquaculture 272, 541–555.

Blaxter, K.L., 1989. Energy Metabolism in Animal and Man. Cambridge University Press,Cambridge, UK.

Booth, M.A., Allan, G.L., Frances, J., Parkinson, S., 2001. Replacement of fish meal in dietsfor Australian silver perch, Bidyanus bidyanus: IV. Effects of dehulling and proteinconcentration on digestibility of grain legumes. Aquaculture 196, 67–85.

Brauge, C., Medale, F., Corraze, G., 1994. Effect of dietary carbohydrate levels on growth,body composition and glycaemia in rainbow trout, Oncorhynchus mykiss, reared inseawater. Aquaculture 123, 109–120.

Briggs, J., Maier, D., Watkins, B., Behnke, K., 1999. Effect of ingredients and processingparameters on pellet quality. Poultry Sci. 78, 1464–1471.

Bromley, P.J., 1994. The role of gastric evacuation experiments in quantifying the feedingrates of predatory fish. Rev. Fish Biol. Fish 4, 36–66.

Budini, R., Tonelli, D., Girotti, S., 1980. Analysis of total phenols using the Prussian Bluemethod. J. Agric. Food. Chem. 28, 1236–1238.

112 S. Adamidou et al. / Aquaculture 289 (2009) 106–112

Burel, C., Boujard, T., Tulli, F., Kaushik, S.J., 2000. Digestibility of extruded peas, extrudedlupin and rapeseed meal in rainbow trout (Oncorhynchus mykiss) and turbot (Psettamaxima). Aquaculture 188, 285–298.

Carter, C.G., Hauler, R.C., 2000. Fish meal replacement by plant meals in extruded feedsfor Atlantic salmon, Salmo salar L. Aquaculture 185, 265–275.

Cho, C.Y., Slinger, S.J., Bayley, H.S., 1982. Bioenergetic of salmonid fish: energy intake,expenditure and productivity. Comp. Biochem. Physiol. 72B, 25–41.

Dos Santos, J., Jobling, M., 1991. Factors affecting gastric evacuation in cod, Gadusmorhua L., fed single-meals of natural prey. J. Fish Biol. 38, 697–713.

Elliot, J.M., 1972. Rates of gastric evacuation in brown trout, Salmo trutta. FreshwaterBiol. 2, 1–18.

Englyst, H.N., Quigley, M.E., Hudson, G.J., 1994. Determination of dietary fibre as non-starch polysaccharides with gas–liquid chromatographic, high-performance liquidchromatographic or spectophotometric measurement of constituent sugars.Analyst 119, 1497–1509.

Euroactiv, 2006. http://www.euractiv.com/en/biotech/genetically-modified-organisms/article-117498.

Finstad, A.G., 2005. Effects of sampling interval and temperature on the accuracy of foodconsumption estimates from stomach contents. J. Fish Biol. 66, 33–44.

Francis, G., Makkar, P.S.H., Becker, K., 2001. Antinutritional factors present in plant-derived alternate fish feed ingredients and their effects in fish. Aquaculture 199,197–227.

Gomes, E.F., Rema, P., Kaushik, S.J., 1995a. Replacement of fish meal by plant proteins inthe diet of rainbow trout (Oncorhynchus mykiss): digestibility and growth perfor-mance. Aquaculture 130, 177–186.

Gomes, E.F., Rema, P., Gouveia, A., Oliva-Teles, A., 1995b. Replacement of fish meal byplant proteins in diets for rainbow trout (Oncorhynchus mykiss) effect of the qualityof the fishmeal based control diets on digestibility and nutrient balances. Water Sci.Technol. 31, 205–211.

Gouveia, A., Davies, S.J., 1998. Preliminary nutritional evaluation of pea seed meal(Pisum sativum) for juvenile European seabass (Dicentrarchus labrax). Aquaculture166, 311–320.

Gouveia, A., Davies, S.J., 2000. Inclusion of an extruded dehulled pea seed meal in dietsfor juvenile European seabass (Dicentrarchus labrax). Aquaculture 182, 183–193.

Gouveia, A., Oliva-Teles, A., Gomes, E., Rema, P., 1993. Effect of cooking/expansion ofthree legume seeds on growth and feed utilization by rainbow trout. In: Kaushik, S.J.,Luquet, P. (Eds.), Fish Nutrition in Practice. 24–27 June 1991. Special Publication, LesColloques, vol. 6. INRA, Biarritz, France, pp. 933–938.

Grain legumes, 2006. Grain Legumes in EU Agriculture. http://www.grainlegumes.com/aep/production/trends_in_production/european_production/grain_legumes_in_eu_agriculture.

Grain legumes, 2007. Main Grain Legume Producing Countries. http://www.grainlegumes.com/aep/production/trends_in_production/world_scene/main_grain_legume_producing_countries.

Grove, D., Genna, R., Pralika, V., Boraston, J., Hornyold, M.G., Siemens, R., 2001. Effects ofdietary water content on meal size, daily food intake, digestion and growth inturbot, Scophthalmus maximus (L.). Aquacult. Res. 32, 433–442.

Hemre, G.-I., Hansen, T., 1998. Utilisation of different dietary starch sources andtolerance to glucose loading in Atlantic salmon (Salmo salar), during parr-smolttransformation. Aquaculture 161, 145–157.

Hemre, G.-I., Torrissen, O., Krogdahl, A., Lie, O., 1995. Glucose tolerance in Atlanticsalmon, Salmo salar L., dependence on adaption to dietary starch and watertemperature. Aquacult. Nutr. 1, 69–75.

Hertrampf, J.W., Piedad-Pascal, F., 2000. Pulses. Handbook on Ingredients for AquacultureFeeds. Kluwer Academic Publishers, Dordrecht, The Netherlands, pp. 531–542.

Jensen, J.W., Berg, T., 1993. Food rations and rate of gastric evacuation in brown trout fedpellets. Prog. Fish-Cult. 55, 244–249.

Jobling, M., 1981. Dietary digestibility and the influence of food components on gastricevacuation in plaice, Pleuronectes platessa L. J. Fish Biol. 19, 29–36.

Jobling, M., 1987. Influence of food particle size and dietary energy content on patternsof gastric evacuation in fish: test of physiological model of gastric emptying. J. FishBiol. 30, 299–314.

Josupeit,H., 2008. FishmealMarketReport for FAOGlobefish. Fishmeal PricesUp,ButMightDecline Soon. http://www.thefishsite.com/articles/452/fismeal-market-report.

Knudsen Bach, K.E., 1997. Carbohydrate and lignin contents of plant materials used inanimal feeding. Anim. Feed Sci. Technol. 67, 319–338.

Knudsen Bach, K.E., 2001. The nutritional significance of “dietary fibre” analysis. Anim.Feed Sci. Technol. 90, 3–20.

Nengas, I., Alexis, M.N., Davies, S.J., Petichakis, G., 1995. An investigation to determinethe digestibility coefficients of various raw materials in diets for the gilthead seabream (Sparus aurata L.). Aquacult. Res. 26, 185–194.

Peres, H., Oliva-Teles, A., 2002. Utilization of raw and gelatinized starch by Europeanseabass (Dicentrarchus labrax) juveniles. Aquaculture 205, 287–299.

Peres, H., Goncalves, P., Oliva-Teles, A., 1999. Glucose tolerance in gilthead seabream(Sparus aurata) and European seabass (Dicentrarchus labrax). Aquaculture 179,415–423.

Persson, L., 1986. Patterns of food evacuation in fishes: a critical review. Environ. Biol.Fishes 16, 51–58.

Petterson, D.S., Mackintosh, J.B., 1994. The Chemical Composition and Nutritive Value ofAustralian Grain Legumes. Grains Research and Development, Canberra, Australia.

Refstie, S., Helland, S.J., Storebakken, T., 1997. Adaptation to soybean meal in diets forrainbow trout, Oncorhynchus mykiss. Aquaculture 153, 263–272.

Riche, M., Haley, D.I., Oetker, M., Garbrecht, S., Garling, D.L., 2004. Effect of feedingfrequency on gastric evacuation and return of appetite in tilapia Oreochromisniloticus (L.). Aquculture 234, 657–673.

Rouhonen, K., Grove, D.J., McIlroy, J.T., 1997. The amount of food ingested in a singlemeal by rainbow trout offered chopped herring, dry and wet diets. J. Fish Biol. 51,93–105.

Runge, C.F., Senauer, B., 2007. How Biofuels Could Starve the Poor. Foreign Affairs, May/June 2007. http://www.foreignaffairs.org.

Schwertner, M.A., Liu, K.K.M., Barrows, F.T., Hardy, R.W., Dong, F.M., 2003. Performanceof post-juvenile rainbow trout Onchorynchus mykiss fed diets manufactured bydifferent processing techniques. J. World Aquacult. Soc. 34, 162–174.

Smith, C., Megen, V.W., Twaalfhoven, L., Hitchcock, C., 1980. The determination oftrypsin inhibitor levels in foodstuffs. J Sci. Food Agric. 31, 341–350.

Sorensen, M., Stjepanovic, N., Romarheim, O.H., Krekling, T., Storebakken, T., 2009.Soybean Meal Improves the Physical Quality of Extruded Fish Feed. Anim. Feed Sci.Technol. 149, 149–161.

Storebakken, T., 1985. Binders in fish feeds: I. Effect of alginate and guar gum on growth,digestibility, feed intake and passage through the gastrointestinal tract of rainbowtrout. Aquaculture 47, 11–26.

Sveier, H., Wathne, E., Lied, E., 1999. Growth, feed utilisation and gastrointestinalevacuation time in Atlantic salmon (Salmo salar L.): the effect of dietary fish mealparticle size and protein concentration. Aquaculture 180, 265–282.

Sweka, J.A., Cox, M.K., Hartman, K.J., 2004. Gastric evacuation rates of Brook trout. Trans.Am. Fish. Soc. 133, 204–210.

Tacon, A.G.J., 1993. Feed Ingredients forWarmwater Fish: FishMeal and Other ProcessedFeedstuffs, FAO Fish. Circ. No. 856. FAO, Rome, Italy, p. 64.

Tharanathan, R.N., Mahadevamma, S., 2003. Grain legumes—a boon to human nutrition.Trends Food Sci. Technol. 14, 507–518.

Thomas, M., van Vliet, T., van der Poel, A.F.B., 1998. Physical quality of pelleted animalfeed. 3. Contribution of feedstuff components. Anim. Feed Sci. Techn. 70, 59–78.

Venou, B., Alexis, M.N., Fountoulaki, E., Nengas, I., Apostolopoulou, M., Castritsi-Cathariou, I., 2003. Effect of extrusion of wheat and corn on gilthead sea bream(Sparus aurata) growth, nutrient utilization efficiency, rates of gastric evacuationand digestive enzyme activities. Aquaculture 225, 207–223.