Effect of sodium bicarbonate supplementation on feed intake, digestibility, digesta kinetics, nitrogen balance and ruminal fermentation in young fattening lambs

Twenty-two Merino lambs (average weight=15.3 kg) were used to study the effects of inclusion of sodium bicarbonate in the concentrate on feed intake, digestibility, rate of passage, nitrogen balance and ruminal fermentation in vivo and in vitro . Lambs were allocated to two experimental groups receiving concentrate and 20 g kg -1 sodium bicarbonate (group Bic ) or concentrate alone (group Control ). Both groups received barley straw ad libitum . Faeces and urine were collected for 5 days to estimate digestibility, nitrogen balance and rate of passage. After slaughter (at 25 kg live weight), samples of rumen fluid were collected from each lamb to determine parameters of ruminal fermentation and to be used as inocula for batch cultures of rumen microorganisms. There were no significant differences between treatments (P>0.10) in concentrate intake, dry matter digestibility, nitrogen balance and digesta kinetics. However, straw intake was greater (P<0.05) and neu-tral-detergent fibre digestibility showed a tendency to be higher in the group Bic (P<0.10). No differences, due to the supplementation with sodium bicarbonate, were observed for in vivo pH, ammonia-N and volatile fatty acids concentrations (P>0.10). Results for rumen fermentation parameters determined in in vitro batch cultures and for fermentation kinetics estimated with the gas production technique followed a similar trend to results observed in vivo . Most parameters showed no significant differences between groups. Nevertheless, the extent of degradation of barley grain in vitro tended to be stimulated (P<0.10) by the use of sodium bicarbonate.


Introduction
Fattening lambs are generally fed rations containing over 800 g kg -1 concentrate in order to achieve high levels of energy intake and daily weight gains (Normand et al., 2001). This results in reduction of the molar proportion of acetate and increase of the molar proportion of propionate (Enemark et al., 2002), which reduces methane production and enhances energy retention (Russell, 1998). However, diets containing high levels of concentrates may be also associated with digestive disorders as a result of decreases of the buffering capacity of the rumen and subsequent increases in rumen acidity (McKinnon et al., 1990).
Buffers can enhance ruminal environmental conditions by modulating acidity of the ruminal contents, preventing severe drops in pH (Le Ruyet and Tucker, 1992). Some salts, such as sodium bicarbonate, are routinely added to ruminant diets to buffer rumen pH, and have been widely used for fattening lambs.
The utilization of sodium bicarbonate has been reported to result in increases in digestibility and rate of passage, and in changes in the proportions of volatile fatty acids (Hart and Doyle, 1985;James and Wohlt, 1985). It has also been suggested that it may improve the amount and efficiency of ruminal microbial protein synthesis, which occurs independently of changes in ruminal fluid dilution rates (Mees et al., 1985), and enhance bacterial uptake of ammonia (Newbold et al., 1988), these effects being eventually associated to a higher feed intake and a subsequent increased daily gain (Tripathi et al., 2004).
Nevertheless, available information on the effects of addition of sodium bicarbonate to lamb diets is not consistent. Thus, in addition to the positive effects mentioned above, some reports have indicated no effect or even negative effects have been observed on ruminal pH (Hadjipanayiotou, 1982;Hart and Doyle, 1985), volatile fatty acid production (James and Wohlt, 1985;Kawas et al., 2007), dry matter intake (Hart and Doyle, 1985;James and Wohlt, 1985) or feed to gain ratio (Hart and Doyle, 1985). Possible explanations for this inconsisten-cy in responses to buffer supplementation may include differences in the mode of action associated to or in addition to a change of pH (Haaland and Tyrrell, 1982) as well as the persistence of the effects. Thus, buffers could be beneficial when added to diets that cause a severe decline in pH during fermentation in the rumen, because these salts increase or stabilize the pH, improving the environmental conditions for microbial and enzymatic activity (Paggi et al., 1999). Buffers may also enhance passage of starch from the rumen (Russell and Chow, 1993).
There is no much information about the use of sodium bicarbonate as buffer in the most common fattening system in Mediterranean countries, in which feedlot lambs are finished under intensive conditions over a short period of time (5-7 weeks after weaning) and slaughtered at a live body weight of 25-30 kg (Sañudo et al., 1998). Therefore, recommendations on the use of dietary sodium bicarbonate in this lamb fattening system cannot be derived from information available on other fattening systems that utilize older and heavier animals.
Therefore, this experiment was conducted to study the effects of the inclusion of 20 g kg -1 sodium bicarbonate in the concentrate of young lambs, under a typical Mediterranean fattening system, on feed intake, digestibility, rate of passage, nitrogen balance and ruminal fermentation.

Animals
Twenty-two male Merino lambs (average initial age of 8-9 weeks) were divided according to their live weight (LW, 15.3 ± 0.14 kg) into two experimental groups of eleven animals each: one group was used as the control (Control) and the other group received concentrate supplemented with sodium bicarbonate (Bic). The lambs were housed in individual floor pens.
Co-EDTA was used to estimate the liquid rate of passage, and was prepared according to the method of Udén et al. (1980). On day 22 of experimental trial, the animals were dosed orally with the marker (1.125 g Co-EDTA dissolved in 30 mL of water). Faeces were collected at 0,4,8,12,16,22,28,34,40,48,60,72,96,120 and 144 h after marker administration, weighed and a subsample (10%) was taken and stored at -30ºC. Samples were dried to constant weight before analysis.

Slaughter and rumen environment
When lambs reached 25 kg of LW, they were slaughtered following standard procedures. The day before slaughter, feed was withdrawn at 20.00 h. Then, two hours prior to slaughter, feed was offered again to all lambs for 1 hour so that each had a full rumen and active fermentation at the time of slaughter. Each lamb was euthanized with an intravenous injection of barbiturate (Euta-lender ® , Normon, Spain), slaughtered by exsanguination from the jugular vein and eviscerated. Total rumen contents from each slaughtered lamb were collected, mixed thoroughly and sampled. About 400 g of this mixture of rumen contents were strained through two layers of cheesecloth and the pH was measured immediately. After centrifugation (600 g at 4ºC for 10 min), a sample of 5 mL of the supernatant was acidified with 5 mL 0.2 N HCl for ammonia determination. Another 0.8 mL of the supernatant was added to 0.5 mL of a deproteinising solution (2% metaphosphoric and 0.4% crotonic acids, w/v, in 0.5 N HCl) for determination of volatile fatty acids (VFA). The samples for NH 3 and VFA were stored at -30ºC until analysed. Remaining filtered rumen fluid from twelve lambs (six from the Control group and six from the Bic group) was used as rumen inoculum for the batch cultures.

Batch cultures of rumen microorganisms
In vitro fermentation kinetics and substrate disappearance (as indicators of the degradative activity of rumen contents) of two feedstuffs, namely barley straw and barley grain, were studied by following the in vitro gas production technique as described by Theodorou et al. (1994). Chemical composition of the substrates is shown in Table 1.
Previously, the lambs had remained stalled with their mothers, with free access to a commercial starter concentrate and alfalfa hay until the commencement of the trial. Immediately after birth, lambs had been treated with Vitasel (Lab. Ovejero, Spain) to prevent white muscle disease, and later on with Miloxan (Merial Lab., Spain) to prevent enterotoxaemia and with albendazol 2.5% Ganadexil ® (Industrial Veterinaria, Spain) to control internal parasites.
The experiment was carried out in accordance with the Royal Decree 1201/2005 for the protection of animals used for experimental and other scientific purposes (BOE, 2005).

Diets
Lambs were fed barley straw and a concentrate containing barley grain, maize grain, soya bean meal, cane molasses and a mineral/vitamin mix, supplemented (Bic) or not (Control) with sodium bicarbonate (20 g kg -1 ). Ingredients and chemical composition of the concentrates and the barley straw are presented in Table 1. All animals had free access to fresh water.
Barley straw and concentrates were offered ad libitum, once a day (approx. at 09.00 h). The amount of feed offered was adjusted daily on the basis of the previous day intake, allowing refusals of 20%. Both concentrate and straw refusals were removed daily, pooled weekly for each animal and weighed and dried to determine dry matter (DM) intake.

Digestibility trial, urine collection and rate of passage
On day 20, lambs (22.5 ± 0.26 kg LW) were confined in metabolism cages fitted with specific devices to collect faeces and urine separately. After 4 days of adaptation, faeces of each animal were collected daily, weighed, mixed thoroughly and subsampled (10%), for 5 days. Aliquots from each sheep were pooled and stored at -30ºC. Pooled samples were dried to constant weight before analysis.
Urine was collected in a solution of H 2 SO 4 (10%; v/v) to maintain the pH<3. Daily urine was weighed, its density was measured and a subsample (20%) was taken for each lamb for 5 days. Daily samples were pooled to form composite samples and stored at -30ºC until analysed for total nitrogen and purine derivatives.
Rumen fluid inocula were obtained separately from 12 lambs (six from the Control group and six from the Bic group), selected randomly among the lambs of each experimental group. Rumen contents were immediately transferred to the laboratory in pre-warmed thermos flasks, strained again through a double layer of muslin and kept under a flushing of CO 2 .
Each feedstuff used as fermentation substrate (barley straw and barley grain) was incubated in duplicate in each of the 12 inocula. Incubations were carried out in 120 mL serum bottles in which 500 mg DM of the appropriate feedstuff was weighed out, and 10 mL strained rumen fluid and 40 mL medium were dispensed anaerobically. The medium contained buffer, macroand micro-mineral, resazurin and reducing solutions according to Menke and Steingass (1988). Then the bottles were sealed and placed in an incubator at 39ºC. Accumulated head-space pressures and gas volumes were measured, using a pressure transducer (Bailey & Mackey, UK) and a graduated syringe, throughout the incubation period (at 2, 4, 6, 8, 10, 12, 15, 19, 24, 30, 36, 48, 72, 96, 120 and 144 h post-inoculation). The values were corrected for the quantity of substrate organic matter (OM) incubated and gas released from blank cultures (i.e., rumen fluid plus buffered medium, without substrate, with and without sodium bicarbonate).
After 144 h of incubation, fermentations were stopped by swirling the flasks on ice. In vitro dry matter disappearance (DMD; g kg -1 ) at this end-point was estimated by filtering residues using pre-weighed sintered glass crucibles (100-160 µm pore size; Pyrex, UK). Residues were then analysed for neutral-detergent fibre (NDF) content to determine in vitro NDF degradation (NDFD).

Ammonia, VFA, and DM and NDF degradation
Twenty-four more cultures per substrate (2 treatments × 6 inocula [replicates] per treatment × 2 flasks per inoculum) were incubated to compare fermentation patterns when rumen fluid from either Bic or Control lambs was used as inoculum. In this case, end-point in vitro incubations at either 8 h (barley grain) or 24 h (barley straw) were carried out. Immediately after termination of the incubation, headspace pressure and gas volume were measured, and bottle contents were centrifuged (600 g, 4ºC, 15 min). Samples of the supernatant were taken and stored at -30ºC for ammonia and VFA determination as described above. Remaining contents of the centrifuge tube were filtered through sintered glass crucibles for subsequent determination of in vitro DM and NDF disappearance, as described above.

Analytical procedures
Procedures described by AOAC (2003) were used to determine DM (AOAC official method 934.01), ash (AOAC official method 942.05), and Kjeldahl nitrogen (N; AOAC official method 976.06). Neutral-detergent fibre (expressed inclusive of residual ash) was determined by the method of Van Soest et al. (1991), adding sodium sulphite to the solution. Only samples of concentrates and barley grain were assayed with alpha amylase.
Ammonia concentration was determined as described by Weatherburn (1967) and VFA by gas chromatography, using crotonic acid as internal standard (Ottenstein and Bartley, 1971), both in centrifuged samples. Concentration of purine derivatives (allantoin, uric acid, xanthine and hypoxanthine) in urine was determined by HPLC according to Balcells et al. (1992).
Samples of faeces collected in the passage kinetics trial were digested with HNO 3 (65% w/v) in a microwave (Microwave Sample Preparation System MDS-2000, CEM Co., USA) and filtered (Whatman, UK). Concentration of Co was determined by inductively coupled atomic emission spectroscopy (ICP-AES, Perkin-Elmer Optima 2000DV, USA).

Calculations and statistical analysis
The generalized Mitscherlich model proposed by France et al. (2000) was fitted to the gas production profiles where G is the cumulative gas production (mL g -1 OM) at time t (h) of incubation, A (mL g -1 OM) represents the asymptotic gas production, c (h -1 ) and d (h -1/2 ) are parameters defining the fractional fermentation rate, and Lag (h) is the initial delay in the onset of gas production. The parameters A, c, d and Lag were estimated by an iterative least squares procedure using the non-linear regression procedure (NLIN) of the Statistical Analysis System package (SAS, 1999). For barley grain, it was shown that parameter d was not significantly different from zero, resulting in a simple exponential equation as a special case of the generalized model.
Other interesting measures were derived from the model parameters, such as time to half-asymptote (t ½ , h): Then, fractional rate of fermentation at t ½ (µ, h -1 ) was calculated as: and average fermentation rate (AFR; mL gas h -1 ), defined as the average gas production rate between the start of the incubation and the time at which the cumulative gas production was half of its asymptotic value, was calculated as: Extent of degradation (ED; g kg -1 ) was estimated, assuming a rumen particulate outflow (k p ) of 0.0625 h -1 , according to the equation proposed by France et al. (2000): Methane production (CH 4 ; mmol g -1 OM) was calculated stoichiometrically according to the amounts of acetate (Ac), propionate (P) and butyrate (B) produced, following the equation proposed by Blümmel et al. (1997): Microbial N flow to the duodenum (MNDF) was estimated from daily urinary excretion of purine derivatives as described by Chen et al. (1992). According to this method, the amount of microbial purines absorbed (MPA, mmol day -1 ) corresponding to the purine derivatives excreted (PDE, mmol day -1 ) was calculated from the following relationship: This model corrects for the contribution of endogenous purines, which is represented by the component within brackets (taking into account the body weight, BW), and which decreases as exogenous purines become available for utilization by the animal. With the assumption that the purine:protein ratio of mixed ruminal microbes remains constant, the amount of MNDF(g day -1 ) was calculated as follows: where 70 is the N content of purines (g mol -1 ), 0.83 is the digestibility coefficient for microbial purines and 0.116 is the ratio of purine N to total N in mixed microbial biomass.
A multi-compartmental model (Dhanoa et al., 1985) was fitted to faecal marker excretion curves: FMC = S · e -k 1 ·t · e -(N-2)·e (k 1 -k 2 )t , where FMC (µg L -1 ) is the faecal marker concentration at time t after dosing (h), k 1 and k 2 are estimates of the   slow and fast fractional outflow rates of digesta (h -1 ), which are usually associated with the rates in the reticulum-rumen and in the caecum and proximal colon, respectively, N is the number of compartments and S is a scale parameter.
Transit time (TT, h) was calculated as: and total mean retention time (TMRT, h) as: All data were analysed as a one-way analysis of variance, with bicarbonate addition as the only source of variation (Control vs. Bic), using the general lineal model (GLM) procedure of the Statistical Analysis System programme (SAS, 1999).

Results
There were no significant differences between groups in the dry matter intake of concentrate (P>0.10) ( Table 2). On the other hand, lambs supplemented with sodium bicarbonate consumed a higher amount of dry matter of barley straw (P=0.015).
No treatment effects (P>0.10) were observed on the digestibility of DM and crude protein. However, NDF digestibility tended to be significantly higher (15%) in group Bic than in Control lambs (0.533 vs 0.462; P=0.080).
There were no differences between treatments for any of the parameters related to the rate of passage (k 1 , k 2 , TT and TMRT; P>0.10) or nitrogen balance.
Mean values of in vivo pH, ammonia and VFA concentrations are given in Table 3. No differences, due to the supplementation with sodium bicarbonate, were found for any of these parameters with the exception of isobutyrate concentration, which was significantly higher in the control group (1.29 vs 0.66 mmol L -1 , P<0.022). Molar proportions of VFA referred to as "others" (sum of valerate + isobutyrate + isovalerate + + caproate) were also significantly higher in the Control than in the Bic lambs (0.0612 vs 0.0412, respectively; P=0.027).
Parameters describing in vitro gas production, substrate disappearance, extent of degradation, ammonia concentration and VFA production when barley grain and barley straw were incubated, are shown in Table 4.
Extent of degradation tended to be greater (P=0.079) when barley grain was incubated with rumen inocula derived from lambs receiving bicarbonate. However, all the other variables (gas production parameters, DM and NDF disappearance, ammonia, VFA, etc.) were observed to be similar in the two experimental groups (P>0.10).
When barley straw was incubated, only production of propionate or butyrate showed a tendency (P=0.075 and 0.097, respectively) to be positively affected by the treatment. No significant effect of supplementation with sodium bicarbonate was found for any other parameter describing the fermentation of barley straw (P>0.10).

Discussion
Sodium bicarbonate is supposed to benefit ruminants eating large amounts of readily fermentable carbohydrates and it has therefore been commonly used in commercial fattening lamb systems. It is considered innocuous to the animal and safe for the consumer without any adverse effect on the meat attributes.
The addition of sodium bicarbonate to the diet has been reported to affect, among others, ruminal pH and osmolality, volatile fatty acid production, rate of passage and voluntary feed intake. In the present experiment, dry matter intake was not significantly affected with the inclusion of bicarbonate in the diet, in agreement with the results reported by Mandebvu and Galbraith (1999) using a lower dose of sodium bicarbonate (15 g kg -1 ). Other authors have observed a higher concentrate intake when adding this salt at similar (Phy and Provenza, 1998) or higher doses (Corcuera et al., 1977).
On the other hand, despite the low proportion that barley straw represents in the diet consumed by the animals, its intake was on average 40% higher in those lambs receiving sodium bicarbonate, which could be related to the tendency to an increased NDF digestibility. Hadjipanayiotou (1982) added sodium bicarbonate to a diet of sheep fed mainly concentrates and obtained an increase in NDF digestibility (22% vs 15% in the present study). Contrary to our result, Hadjipanayiotou (1982) also reported improvements in the DM digestibility. However, Cooper et al. (1996), which included up to 40 g kg -1 sodium bicarbonate in their diet, did not obtain differences in NDF or DM digestibility. Other authors, however, using dehydrated alfalfa with sodium bicarbonate reported greater increases in NDF and DM digestibility (Stroud et al., 1985), which may be due to the high efficiency of the forage to elevate ruminal pH (Hadjipanayiotou, 1982). It is known that the inclusion of sodium bicarbonate prevents severe decline in rumen pH, alleviating the depressive effect of a low pH on cellulolysis, and so enhancing fibre degradation (Mould et al., 1983;Wedekind et al., 1986).
Previous studies have shown that the use of sodium bicarbonate may result in increased rates of passage of the liquid phase of the digesta. Stokes (1983) reported a quadratic increase in the dilution rate from the rumen of sheep fed diets supplemented with increasing levels of sodium bicarbonate. In this study, although total mean retention time (TMRT) of digesta in the gastrointestinal tract was on average 11% shorter in lambs eating the sodium bicarbonate diet than in those fed the control diet, the differences were not statistically significant (P>0.10). Whereas some authors have attributed the changes in rate of passage to a rise in acetic acid production , others have suggested a mode of action related to a higher osmolality and water intake (Russell and Chow, 1993;Cooper et al., 1996). Acetic acid concentration in rumen fluid or its production in vitro, when incubating rumen fluid with either barley grain or straw, did not change in lambs supplemented with sodium bicarbonate (see Tables 3, 4 and 5) but water consumption was not measured. In agreement with our results, Mees et al. (1985), adding 35 g kg -1 sodium bicarbonate to the diet, did not observe changes in liquid flow rates. The lack of response to buffer treatment in this and other studies was probably due to supplementary buffers having little effect on turnover rate in animals that already show fast digesta passage rates (Harrison et al., 1975). It must be noted that the analyses of faecal Co-EDTA concentration curves are only an estimate of dilution rate (Dhanoa et al., 1985). However, this method is widely accepted and leads to accurate estimate of mean retention time in the rumen, and is consistent with the current trend in animal research to use non-invasive techniques for assessing digestive parameters (Bernard et al., 1998).
No differences in ruminal pH in vivo due to buffer addition were observed (see Table 3), in agreement with other results reported in the literature (Hadjipanayiotou, 1982;Hart and Doyle, 1985;Russell and Chow, 1993;Khorasani and Kennelly, 2001;Kawas et al., 2007). Nevertheless, it should be mentioned that, in the current experiment, values of pH were only measured immediately after slaughter, and diurnal variations in ruminal pH were not recorded.
Published results about changes in VFA production in response to sodium bicarbonate administration are   (Thomas and Hall, 1984;Hart and Doyle, 1985;Khorasani and Kennelly, 2001) and others decreases (James and Wohlt, 1985) or no changes (Mees et al., 1985;Kawas et al., 2007). No significant effects were detected in the present experiment, with the exception of the molar proportion of the sum of valerate + isobutyrate + isovalerate + caproate. Some authors have observed higher acetate and lower propionate molar proportions related, respectively, to greater fibre degradation (Van Soest et al., 1991), and to an increased ruminal dilution rate and resulting washout of soluble carbohydrates from the rumen (Russell and Chow, 1993).
The use of sodium bicarbonate as a dietary additive did not affect the acetate to propionate ratio in the rumen, showing a low value characteristic of high intakes of concentrate (starchy) feeds, as suggested by Khorasani and Kennelly (2001). It is noteworthy that these low ratios (<1.5) fell into the range (from 0.9 to 4) previously reported by other authors (Woods and Luther, 1962). A negative effect of sodium bicarbonate on propionate production would have been very advantageous in fattening lambs because when in excess, propionate might be a precursor for the synthesis of oddnumbered and methyl branched chain fatty acids, whose proportions are high in soft adipose tissues (Normand et al., 2001).
Although some studies have suggested that sodium bicarbonate can enhance the rate of ammonia utilization by the rumen bacteria (Newbold et al., 1988), other researchers working with sheep (Stokes, 1983) and cattle (Khorasani and Kennelly, 2001) have shown no effect on ruminal ammonia levels, which is in line with the results observed in this experiment. Some studies also suggested that sodium bicarbonate supplementation may enhance microbial protein synthesis (Harrison et al., 1975) as result of changes in the dilution rate (Newbold et al., 1988). In the present experiment no differences in nitrogen balance or flow of microbial protein from the rumen were estimated from purine derivatives excretion, which agrees with the lack of significant variations found in the digesta kinetics.
Characterization of in vitro ruminal fermentation when sodium bicarbonate is included in the diet has been reported for cattle and sheep (Le Ruyet and Tuc ker, 1992;Cobos-Peralta et al., 2005), but available information for young fattening lambs is particularly scarce. Results for rumen fermentation characteristics obtained by the in vitro batch cultures and the gas production technique followed a similar pattern to results obtained in vivo. Sodium bicarbonate had no significant effect on parameters such as ammonia or VFA production, although in incubations there was a tendency for higher production of propionic and butyric acids with barley straw.
The extent of degradation of barley grain tended to be stimulated slightly (P<0.10) in animals receiving the buffer. On the other hand, no significant differences were detected when barley straw was incubated. This lack of effect of sodium bicarbonate on in vitro NDF degradability in contrast to the improvement observed in in vivo NDF digestibility could be accounted for by the highly buffered incubation medium ("artificial saliva", McDougall, 1948) used in the in vitro batch cultures. Nevertheless, it is noteworthy that extent of degradation of straw estimated from the fermentation kinetics was very low (<0.12) when ruminal fluid from young lambs fed a diet rich in concentrates was used as inoculum for the mixed ruminal microorganisms cultures. In contrast, extent of degradation of barley grain can be considered within a normal range. Thus, whereas amylolytic activity is normal in rumen fluid of these  . In vitro gas production parameters (A, Lag, t ½ and µ), average fermentation rate (AFR), extent of degradation (ED), ammonia concentration, volatile fatty acids (VFA) and methane production, and dry matter (DM-D) and neutral-detergent fibre (NDF-D) disappearance after 8 and 144 h incubation for barley grain and barley straw From the results shown here, it can be concluded that the addition of 20 g kg -1 sodium bicarbonate to concentrate fed to young fattening lambs can improve the intake of straw and NDF digestibility. However, the mode of action of this buffer additive remains unclear. Further studies will be necessary to test the hypothesis that the lack of a stronger effect in this kind of young animals may be probably related to the short duration of the finishing period under the typical Mediterranean lamb fattening system.