Effects of Physical Properties of Feed on Microbial Ecology and Survival of Salmonella enterica Serovar Typhimurium in the Pig Gastrointestinal Tract

A two-by-two factorial experiment with pigs was conducted tostudy the effect of feed grinding (fine and coarse) and feedprocessing (pelleted and nonpelleted) on physicochemical properties,microbial populations, and survival of Salmonella enterica serovarTyphimurium DT12 in the gastrointestinal tracts of pigs. Resultsdemonstrated a strong effect of diet on parameters measuredin the stomachs of the pigs, whereas the effect was less inthe other parts of the gastrointestinal tract. Pigs fed thecoarse nonpelleted (C-NP) diet showed more solid gastric contentwith higher dry matter content than pigs fed the fine nonpelleted(F-NP), coarse pelleted (C-P), or fine pelleted (F-P) diet.Pigs fed the C-NP diet also showed significantly increased numberof anaerobic bacteria (P < 0.05), increased concentrationsof organic acids, and reduced pH in the stomach. In addition,pigs fed the C-NP diet showed increased in vitro death rateof S. enterica serovar Typhimurium DT12 in content from thestomach (P < 0.001). Pigs fed the C-NP diet had a significantlyhigher concentration of undissociated lactic acid in gastriccontent than pigs fed the other diets (P < 0.001). A strongcorrelation between the concentration of undissociated lacticacid and the death rate of S. enterica serovar Typhimurium DT12was found. In the distal small intestine, cecum, and midcolon,significantly lower numbers of coliform bacteria were observedin pigs fed the coarse diets than in pigs fed the fine diets(P < 0.01). Pigs fed the C-NP diet showed the lowest numberof coliform bacteria in these segments of the gastrointestinaltract. Pigs fed the coarse diets showed increased concentrationof butyric acid in the cecum (P < 0.05) and colon (P <0.10) compared with pigs fed the fine diets. It was concludedthat feeding a coarsely ground meal feed to pigs changes thephysicochemical and microbial properties of content in the stomach,which decreases the survival of Salmonella during passage throughthe stomach. In this way the stomach acts as a barrier preventingharmful bacteria from entering and proliferating in the lowerpart of the gastrointestinal tract.

INTRODUCTION

Top

Introduction

Human food-borne disease outbreaks are a public health concernand have economic importance to the food-producing industry.Subclinical Salmonella enterica infections in pig herds arerecognized as an important source of human salmonellosis inDenmark (29). A Danish Salmonella surveillance and control programin pig herds and slaughterhouses has been applied nationallysince 1995, and the goal is to reduce the Salmonella prevalencein pork meat to 0.5% (23). In year 2002, the incidence of Salmonella-positivecarcass samples in Danish slaughterhouses was 1.4% and the correspondingSalmonella prevalence at herd level monitored by serologicaltesting was 3.2% (1).

There is a growing awareness that it is only possible to reducethe prevalence of Salmonella in pork, and consequently the numberof infections in people consuming pork and pork products, throughan integration and cooperation of all stages of the pork productionchain. This is indicated by several studies. Berends et al.(3) found a strong correlation between the number of live animalsthat carry Salmonella spp. in their feces and the number ofcontaminated carcasses in the end of the slaughter line (3).Hurd et al. (12) reported that transport to the slaughterhouseand especially holding in pens at the slaughterhouse increasedthe prevalence of S. enterica-positive pigs by sevenfold comparedwith the prevalence among pigs from the same farms but necropsiedon the farm (12). These studies also indicate that, if salmonella-freepigs are produced, consumers can be provided with virtuallysalmonella-free pork products.

There is strong evidence that the physical characteristics ofthe feed influence the susceptibility of pigs to Salmonella.Feeding a ready-mixed pelleted dry feed increases the risk ofSalmonella infection compared with home-mixed, nonpelleted wetfeed (17, 28). Coarse grinding of feed also reduces the incidenceof Salmonella in pigs compared to fine grinding (15, 30). However,feeding a coarse and/or nonpelleted feed to pigs results inpoor growth performance (6, 15, 19, 31, 32) and consequentlyis an economic disadvantage for the pig producer. To designalternative feeds with similar Salmonella-reducing propertiesbut with improved growth performance, it would be of great importanceto determine how coarse feed influences the survival of Salmonellain the gastrointestinal tracts of pigs.

The aim of the present investigation was to study the factorsbehind the lower incidences of Salmonella in pigs fed a coarsefeed. A two-by-two factorial experiment in which pigs were feda fine or coarse feed given to the animals either as nonpelletedor pelleted feed was carried out. The effect of the experimentaldiets on physicochemical properties and microbial populationsand activity in the gastrointestinal tracts of slaughtered pigswas investigated. An in vitro method was performed to studythe survival of S. enterica serovar Typhimurium DT12 in stomachand cecal contents from the pigs fed the different experimentaldiets. This Salmonella strain is a frequent cause of human salmonellosisand is also common in pig herds in Denmark.

MATERIALS AND METHODS

Top

Materials and Methods

Animals and feed.
A total of 96 crossbreed Danish Landrace x Yorkshire pigs (averageweight, 33 ± 7 kg) were obtained from the herd at theResearch Center Foulum, Foulum, Denmark, and assigned to oneof four treatment groups in a two-by-two factorial design experiment.The pigs were allocated to the treatment groups with an equaldistribution for litter and sex. The experiment was carriedout in six replicates of four pigs per group housed togetherin a pen with no physical or visual contact between animalsin different pens. The pigs were fed ad libitum and had freeaccess to water. The feed contained barley (35.9%), wheat (35.9%),soybean meal (12.3%), rapeseed meal (5%), sunflower meal (5%),animal fat (2.3%), dicalcium phosphate (1.2%), calcium carbonate(0.8%), sodium chloride (0.4%), lysine (0.3%), methionine (0.12%),threonine (0.08%), vitamin-mineral mixture (0.2%), and chromicoxide (0.4%). The feed was ground at two levels to obtain afine or coarse feed and given to the animals as either nonpelletedor pelleted feed.

The four types of experimental feed, fine nonpelleted (F-NP),fine pelleted (F-P), coarse nonpelleted (C-NP), and coarse pelleted(C-P), were analyzed for particle size distribution by a wet-sievemethod. Four hundred milliliters of deionized water at 20°Cwas added to 100 g of feed. The solution was kept for 1 h atroom temperature without stirring. The suspension was then sievedfor 2 min, at a 3-mm amplitude, with an AS200 control "g" sieveshaker (RETSCH GmbH & Co., Haan, Germany). The sieves (interiordimensions, 200 by 50 mm; DIN ISO 3310/1) were made of stainlesssteel, and the following mesh sizes were used 3,150, 2,000,1,000, 500, and 250 mm. The different particle fractions werethen dried to constant weight at 100°C and weighed.

Weights of the pigs and feed consumption per pen were registeredevery week during the experiment. Performance of the pigs (30to 100 kg) per pen was calculated on the basis of four pigsper pen from day 0 to 28 of the experiment and three pigs perpen for the remaining period until the pigs weighed approximately100 kg. The pigs were checked daily for signs of diarrhea andgiven a score on a scale from 0 to 3, with 3 corresponding toaqueous feces, 2 corresponding to thin feces, 1 correspondingto porridge-like feces, and 0 corresponding to normal consistencyof feces. The experiment complied with the guidelines of theDanish Ministry of Justice with respect to animal experimentationand care of animals under study.

Collection of intestinal samples.
After 28 days on the experimental diets, one pig per replicatefrom each experimental group (n = 6) was killed with a lethalinjection of pentobarbital sodium (200 g/liter). The gastrointestinaltract was immediately removed and divided into eight segments:stomach, three equal parts of the small intestine, cecum, andthree equal parts of the colon including the rectum. The totalcontent from each segment was collected and weighed. Immediately,pH was measured by direct insertion of a pH electrode (RadiometerS.A., Villeurbanne, France) into the matrices, and samples fordetermination of microbial activity were processed as describedbelow. The intestinal samples were then taken to the laboratoryfor further analysis.

Examination for gastric ulcer.
At slaughter, the gastric epithelium of the pars oesophageawas visually examined for possible alterations. The scale proposedby Nielsen and Ingvartsen (24) was used, with 0 correspondingto normal epithelium and 10 corresponding to severe constrictiondue to gastric ulcer.

Physicochemical analyses.
Intestinal samples from all segments were analyzed for dry mattercontent, determined by freeze-drying, and for the concentrationof organic acids (formic acid, short-chain fatty acids [SCFA],lactic acid, and succinic acid) by the method described by Jensenet al. (14). A rheological characterization was done on samplesfrom the stomach. In brief, the samples were centrifuged at12,000 x g for 20 min at 4°C in capped 50-ml centrifugetubes. After the tubes were weighed, the supernatant fractionwas removed by suction and the centrifuge tubes were turnedupside down to drain off residual supernatant fraction for atleast 1 h before reweighing. The sediment was subsequently driedat 105°C to constant weight. The water-binding capacitywas calculated as the amount of water that was left after centrifugationand removed during the drying process. The sedimentation abilitywas measured by removing 100 ml of stomach content, placingit in a cylinder tube at 5°C, and subsequently measuringthe distribution in fluid and sediment after 24 h.

Microbial determinations.
Feed samples (10.0 g) were mixed with 90 ml of peptone watercontaining 1% Bacto Peptone (1.07213; Merck, Darmstadt, Germany)and 0.1% Tween 80 and homogenized in a stomacher laboratoryblender (Interscience, St. Nom, France) for 2 min. Tenfold dilutionseries in peptone water were prepared. Total aerobic bacteriawere enumerated on plate count agar (Merck; 5463) and coliformbacteria and lactose-negative enterobacteria were enumeratedon MacConkey agar (Merck; 5465) after incubation at 38°Cfor 24 h. Lactic acid bacteria were enumerated on de Man, Rogosa,and Sharpe agar (MRS; Merck; 10660) after anaerobic incubationat 20°C for 48 h. Yeast was enumerated on malt chloramphenicolagar (MCA) containing 10 g of glucose, 5 g of malt extract,10 g of peptone, 15 g of agar, and 0.050 g of chloramphenicol/literof distilled water after incubation at 20°C for 72 h.

Gastrointestinal samples (10.0 g) from the stomach, distal thirdof the small intestine, cecum, and midthird of the colon weretransferred under a flow of CO₂ into sterile serum bottles containing90 ml of a prereduced salt medium (11). The suspension was pouredinto a CO₂-flushed plastic bag and homogenized in a stomacherlaboratory blender for 2 min. Serial 10-fold dilutions wereperformed in prereduced salt medium by the technique of Millerand Wolin (21). The number of total culturable anaerobic bacteriawas determined with anaerobic roll tubes containing rumen fluid-glucose-cellobiose-agar(11) after incubation at 38°C for 7 days. Coliform bacteriaand lactose-negative enterobacteria were enumerated on MacConkeyagar after incubation at 38°C for 24 h. Lactic acid bacteriaand yeast were enumerated on MRS and MCA, respectively, afteranaerobic incubation of MRS plates and aerobic incubation ofMCA plates at 38°C for 48 h.

Determination of microbial activity.
Content (5.0 g) from each of the eight segments was extractedwith 10.0 ml of perchloric acid (PCA)-EDTA (2 M cold PCA containing10 mM EDTA) and stored at –80°C. The concentrationof ATP in digesta was determined by the luciferin-luciferasemethod as described by Jensen and Jørgensen (13).

In vitro survival of Salmonella in stomach and cecal content.
Contents from the stomach and cecum (stored at –20°Cuntil use) were thawed, and 10.0-g samples were added to smallstomacher plastic bags and inoculated with a 1.0-ml suspensionof S. enterica serovar Typhimurium DT12 grown in brain heartinfusion broth (Merck; 1.10493). The inoculum was prepared asdescribed by Naughton et al. (22). The stomacher bags were flushedwith CO₂, sealed airtight, and incubated at 38°C in a shakingwater bath. For each sample, four identical subsamples in stomacherbags were prepared, one for each sampling time, at 0, 2, 4,and 6 h of incubation. At sampling, the stomacher bag was cutopen and pH was measured by direct insertion of a pH electrode(Radiometer S.A.) into the contents. Then, 90 ml of peptonewater was added and the whole solution was poured into a CO₂-flushedstomacher bag and homogenized for 2 min. A sample (1 ml) wastaken and analyzed for the concentration of organic acids (formicacid, SCFA, lactic acid, and succinic acid) as described byJensen et al. (14). Then, serial 10-fold dilutions were performedin peptone water. The enumeration of salmonellae was performedon MacConkey agar and brilliant phenol-lysine-sucrose agar (Merck;1.10747) after aerobic incubation at 38°C for 16 h.

Calculations.
The death rate (µ) of S. enterica serovar TyphimuriumDT12 in contents from the stomach and cecum was calculated fromsemilogarithmic plots of CFU during the 6-h incubation periodfor stomach contents and 4-h incubation period for cecal contents.The bacterial half-life (T_1/2) was calculated from the equationT_1/2 = ln 2/µ as described by Knarreborg et al. (16).

Concentrations of undissociated SCFA and lactic acid were calculatedfrom the Henderson-Hasselbach equation: pH = pK_a + log₁₀ [A^–]/[HA],where [A^–] and [HA] represent the concentrations of dissociatedand undissociated acids, respectively. The average pH and concentrationsof SCFA and lactic acid during the 6- and 4-h incubation periodsfor stomach and cecal contents, respectively, were used forthe calculations. The pK_a values are 4.75 for acetic acid, 4.81for butyric acid, 4.87 for propionic acid, and 3.84 for lacticacid (5).

Statistics.
The effect of grinding (fine versus coarse), form of feed (pelletedversus nonpelleted), and diet (interaction between grindingand form of feed) was analyzed statistically by a two-factoriallinear model using the PROC MIXED procedure of SAS (18). Litterwas included in the statistical model as a random effect. Ifthe interaction between grinding and form of feed was not significant(P < 0.05), it was omitted from the model. Data obtainedfrom the gastrointestinal tract were analyzed by segment. Statisticalanalysis of bacterial counts was performed after logarithmicconversion of the data. Results are expressed as least-squaremeans and standard errors of the means (SEM). If the overalleffect of a main factor was significant, differences betweenmeans were compared by the Fisher's least significant differenceand considered significant at P values of <0.05.

RESULTS

Top

Results

Feed.
The distributions of particle sizes in the experimental dietsare given in Table 1. The highest percentage of larger particles(>1,000 µm) was found in the C-NP diet, followed bythe C-P diet, with an intermediate percentage, and low percentagesof particles larger than 1,000 µm in the F-NP and F-Pdiets. The distribution of bacteria in the feed is also givenin Table 1. The numbers of total aerobic bacteria, coliformbacteria, and yeast in the nonpelleted feed were significantlyhigher than those in the pelleted feed (P < 0.01). Therewere no differences in the number of lactic acid bacteria amongthe diets (Table 1).

View this table:
[in this window]
[in a new window]
TABLE 1. Particle size distribution and microbiological counts of the experimental diets

Animal performance.
There were no significant differences in daily weight gain orfeed intake between pigs fed the different experimental diets.The feed conversion ratio was significantly better (2.53 versus2.69 kg of feed/kg of weight gain; P < 0.01) for pigs fedthe pelleted feed than for those fed the nonpelleted feed (resultsnot shown). Pigs fed the pelleted feed showed a significantlyhigher diarrhea score than those fed the nonpelleted feed (0.33versus 0.02; P < 0.01). There was no effect of grinding (fineversus coarse) on the diarrhea scores (results not shown). Thegastric alteration scores are given in Table 2. All pigs fedthe C-NP diet had a gastric alteration score of 0, which correspondsto an undamaged epithelium. Both fine versus coarse feed andpelleted versus nonpelleted feed afforded significantly higherscores for altered gastric epithelia of the pigs (Table 2).

View this table:
[in this window]
[in a new window]
TABLE 2. Gastric alterations and physicochemical properties of stomach content

Rheological characteristics of stomach content.
The sedimentation ability and water-binding capacity of stomachcontent are shown in Table 2. A consistent high sedimentationpercentage (100%), corresponding to no separation into fluidand sediment, was observed for stomach contents of pigs fedthe C-NP diet. Conversely, the lowest sedimentation percentage,corresponding to the most pronounced separation of stomach contentinto a watery phase and a precipitate, was observed for pigsfed the F-P diet (Table 2). The coarse versus fine feed andthe nonpelleted versus pelleted feed both resulted in significantlyhigher sedimentation percentages of the stomach content (P <0.0001 and P < 0.01, respectively). Significantly higherwater-binding capacity was observed for stomach contents frompigs fed the C-NP diet (P < 0.01) than for those from pigsfed the other diets (Table 2).

Gastrointestinal digesta, pH, and concentration of organic acids.
The amounts of digesta from the various segments of the gastrointestinaltract were not affected by the experimental diets (results notshown). The dry matter percentage of content from the stomachwas significantly higher for pigs fed the coarse feed than forpig fed the fine feed (P < 0.01) and was also higher forpigs fed the nonpelleted feed than for pigs fed the pelletedfeed (P < 0.05; Table 2). No differences in dry matter percentagesof digesta were observed in the small intestines, ceca, andcolons of the pigs (results not shown).

The pH and organic acid concentrations in stomach contents ofpigs fed the experimental diets are also given in Table 2. ThepH in stomach content was lowest for pigs fed the C-NP dietand differed significantly from the pH of stomach contents foundin pigs fed the C-P diet (Table 2). In general, no differencesin pH were observed in the small intestines, ceca, and colonsof pigs fed the different diets (results not shown).

The concentration of organic acids in the stomach of the pigswas strongly affected by the experimental diet (Table 2). Thehighest concentrations of the various organic acids in stomachcontent were observed for pigs fed the C-NP diet, although theeffect of diet was not significant for all the individual organicacids (Table 2). The concentration of acetic acid was significantlyhigher in pigs fed the coarse feed than in pigs fed the finefeed (P < 0.001) and was significantly higher in pigs fedthe nonpelleted feed than in pigs fed the pelleted feed (P <0.05). The concentration of propionic acid was significantlyhigher in pigs fed the C-NP diet than in pigs fed the otherdiets (P < 0.05). The concentrations of butyric and lacticacid were significantly higher in stomach contents of pigs fedthe coarse feed than in those of pigs fed the fine feed (P <0.05 and P < 0.001, respectively). Valeric acid was observedin minor concentrations only in pigs fed the C-NP diet (Table2).

In the other compartments of the intestinal tract, generallylittle or no effect on the concentrations of organic acids wasobserved (results not shown). However, the distal small intestinesof pigs fed the nonpelleted feed had a significantly higherconcentration of formic acid than those of pigs fed the pelletedfeed (17.5 versus 12.4 mmol/kg; P < 0.05). The concentrationof butyric acid was significantly higher in the ceca (14.8 versus11.6 mmol/kg; P < 0.05) and almost significantly higher inthe midcolons (19.9 versus 18.0 mmol/kg; P < 0.10) of pigsfed the coarse feed than in those of pigs fed the fine feed(results not shown).

Microbial populations and activity.
The numbers of total anaerobic bacteria, lactic acid bacteria,coliform bacteria, and yeast and the concentration of ATP inthe gastrointestinal tracts of pigs fed the experimental dietsare given in Tables 3 and 4. The number of total anaerobic bacteriain the stomachs of pigs fed the C-NP diet was significantlyhigher than in those of pigs fed the other diets (P < 0.05).Besides this, the coarse feed resulted in significantly highernumbers of lactic acid bacteria (P < 0.05) in the stomachthan fine feed and nonpelleted feed gave significantly highernumbers of yeast (P < 0.05) in the stomach than pelletedfeed (Table 3). There were no differences in the numbers ofcoliform bacteria in the stomachs of pigs fed the differentdiets. However, in the distal small intestine, cecum, and midcolon,significantly higher numbers of coliform bacteria were observedin pigs fed the fine feed than in pigs fed the coarse feed (Table3). The midcolons of pigs fed the F-NP diet had a significantlyhigher (P < 0.01) number of coliform bacteria than thoseof pigs fed the other diets (Table 3). The number of lacticacid bacteria was significantly higher (P < 0.05) in thececa of pigs fed the F-P diet than in ceca of pigs fed the F-NPdiet. Likewise, the number of total anaerobic bacteria in themidcolons of pigs fed the F-P diet was significantly higher(P < 0.05) than in those of pigs fed the F-NP diet (Table3).

View this table:
[in this window]
[in a new window]
TABLE 3. Microbiological counts in gastrointestinal content

View this table:
[in this window]
[in a new window]
TABLE 4. Concentration of ATP in gastrointestinal content^a

A significantly higher concentration of ATP was found in thestomachs of pigs fed the C-NP diet (P < 0.05) than in thoseof pigs fed the other diets (Table 4). Higher ATP concentrationswere found the distal small intestines, ceca, and colons ofpigs fed the nonpelleted than in those of pigs fed the pelletedfeed, and this difference was significant in the distal smallintestine and cecum (P < 0.05).

In vitro death rate of Salmonella serovar Typhimurium DT12 in contents from the stomach and cecum.
Figure 1 shows an example of typical time courses for the densityof the inoculated S. enterica serovar Typhimurium DT12, pH,and the concentrations of SCFA and lactic acid in content fromthe stomach and cecum during the in vitro incubation. In contentfrom the stomach, the density of S. enterica serovar TyphimuriumDT12 declined exponentially during the incubation and the specificdeath rates of S. enterica serovar Typhimurium DT12 were calculatedfrom the slope of the linear regression for the 6-h incubationperiod (Fig. 1). The pH and concentrations of SCFA and lacticacid were generally constant during incubation with stomachcontent, and the results for these parameters given in Table5 are means of values measured during the 6-h incubation period(Table 5).

View larger version (23K):
[in this window]
[in a new window]
FIG. 1. Typical time courses for the density of S. enterica serovar Typhimurium DT12, pH, and the concentration of organic acids in stomach and cecal contents during the in vitro incubation. ${blacklozenge}$ , S. enterica serovar Typhimurium DT12; ${blacksquare}$ , pH; ${triangleup}$ , lactic acid; ${circ}$ , acetic acid; ${lozenge}$ , propionic acid; ${square}$ , butyric acid.

View this table:
[in this window]
[in a new window]
TABLE 5. Growth rates of S. enterica serovar Typhimurium DT12 and physicochemical properties measured in vitro in content from the stomachs and ceca of pigs fed the experimental diets

The results showed a significantly higher specific death rateof S. enterica serovar Typhimurium DT12 (P < 0.001) frompigs fed the C-NP diet than from pigs fed the other diets (Table5). The specific death rate of 1.9/h observed in stomach contentfrom pigs fed the C-NP diet corresponds to a T_1/2 of 22 min.This was almost 10 times faster than the T_1/2 of 219 min observedfor pigs fed the C-P diet (results not shown). The T_1/2s were139 and 95 min for pigs fed the F-NP and F-P diets, respectively(results not shown). The pH was significantly lower for theC-NP diet than for the C-P diet (P < 0.05), and the concentrationof propionic acid was significantly higher for the C-NP dietthan for the other diets (P < 0.001). The numerically highestconcentrations of acetic acid, butyric acid, and lactic acidwere observed in stomach contents from pigs fed the C-NP diet(Table 5). The concentrations of SCFA and lactic acid were significantlyhigher for pigs fed the coarse feed than for pigs fed the finefeed (Table 5). The concentration of undissociated lactic acidwas significantly higher for the C-NP diet than for the otherdiets (P < 0.001).

Figure 1 also shows a typical time course for the density ofS. enterica serovar Typhimurium DT12, pH, and SCFA and lacticacid concentrations in cecal content. In cecal content, a drasticincrease in the concentration of lactic acid, a decrease inpH, and a decrease in the density of S. enterica serovar TyphimuriumDT12 were observed at the end of the incubation period (Fig.1). For that reason, the specific death rates of S. entericaserovar Typhimurium DT12 were calculated from the slope of thelinear regression for the 4-h incubation period (Fig. 1) andthe results on pH and concentrations of SCFA and lactic acidare means of values measured during the 4-h incubation period(Table 5). The specific death rates of S. enterica serovar TyphimuriumDT12, pH, and concentrations of SCFA in cecal content were notsignificantly affected by the experimental diets (Table 5).However, the concentration of butyric acid during in vitro incubationwas significantly higher (P < 0.01) in cecal content frompigs fed the coarse feed than in that from pigs fed the finefeed (Table 5). Lactic acid accumulated at the end of the invitro incubations, and the highest lactic acid concentrationwas observed for the C-NP diet. No significant differences inthe concentration of undissociated lactic acid or total organicacids in cecal content were observed (Table 5).

The specific death rates for all observations of the in vitroincubations with S. enterica serovar Typhimurium DT12 in stomachcontent (regardless of diet) were plotted against the correspondingconcentrations of undissociated lactic acid, and the correlation(R²) was calculated by linear regression (y = 0.15x + 0.19)as 0.71 (Fig. 2).

View larger version (14K):
[in this window]
[in a new window]
FIG. 2. Correlation between the death rate of S. enterica serovar Typhimurium DT12 and the concentration of undissociated lactic acid HA (millimolar) in stomach content measured in vitro.

DISCUSSION

Top

Discussion

The present study shows that pigs fed a nonpelleted feed haveless-efficient feed conversion than pigs fed a pelleted feed.This supports data from previous studies (6, 15, 31, 32). Theresults also show that pigs fed the nonpelleted feed were lesslikely to experience diarrhea. This effect of feed processingon diarrhea score has not, to our knowledge, been reported before.Furthermore, pigs fed the coarsely ground meal feed all hadundamaged tissue of the nonglandular esophageal region in thestomach, whereas pigs fed the other diets showed some degreeof lesions and keratinization of this tissue. These resultsare in accordance with other studies (7, 24, 25, 32). The reasonfor this prevention of gastric ulcer may be related to the consistencyof the stomach content. Stomach content from pigs fed the coarselyground meal feed had the highest dry matter content and hadsignificantly higher water binding capacity and a more porridge-likeconsistency (that did not divide into a precipitate and a wateryphase on standing) than stomach contents from pigs fed the otherdiets. These marked effects of the coarsely ground meal feedoccurred in spite of only minor differences in the particlesize distribution between the coarsely ground meal feed andthe coarsely ground pelleted feed. The observed rheologicalcharacteristics of stomach content are in accordance with resultsfrom other studies (20, 24, 25).

A strong effect of the experimental diets on several microbialparameters was observed in the present study. Pigs fed the coarselyground meal feed showed significantly increased number of totalanaerobic bacteria, significantly higher concentration of ATP,increased concentrations of various organic acids, and lowerpH in the stomach than pigs fed the other diets. These resultsdemonstrate that pigs fed the coarsely ground meal feed hadmuch higher microbial fermentation in their stomachs than pigsfed the other diets. Others have reported similar results (15).The coarsely ground meal feed also led to a different profileof organic acids produced, suggesting that the microbiota inthe stomach of these pigs had a different composition and/oractivity. A slower gastric passage rate when pigs are fed acoarsely versus a finely ground feed has been reported (20,25). A slower passage rate may be the explanation for the highermicrobial activity found in the stomachs of pigs fed the coarselyground meal feed due to an increase in the time for the microbiotato proliferate in the stomach. In addition, the consistencyof the stomach content may influence the growth of the microbiotain the stomach. In this regard, it has been reported that thematrix microstructure in a model porous medium exerts a majorinfluence on the survival of food-borne pathogens (10) and thatbacterial survival depends on the formation of microscopic water-richdomains (9). It is possible that the higher water binding capacityand more porridge-like consistency of stomach content assistedmicrobial growth in the stomachs of pigs fed the coarsely groundmeal feed in the present study.

S. enterica serovar Typhimurium DT12 is one of the most frequentlyisolated strains in human salmonellosis in Denmark and is alsocommon in pig herds (1, 2). The very strong killing effect onS. enterica serovar Typhimurium DT12 in stomach contents frompigs fed the coarsely ground meal feed observed in the presentstudy is in agreement with results showing that the growth ofenterobacteria is inhibited at high levels of lactic acid atlow pH (16). It is proposed that the undissociated form of lacticacid and other organic acids exerts the antibacterial effect.The undissociated form of the acids can freely cross the bacterialmembrane and enter the bacterial cell, where the acids dissociateand cause intracellular pH to drop. As a result, enzymatic processesstop and the proton motive force collapses, causing cell death(26). This is supported by the present results showing a highcorrelation between the death rate of S. enterica serovar TyphimuriumDT12 and the concentration of undissociated lactic acid in thestomach content. The retention time for stomach content in pigsis approximately 3 to 4 h (8). From the present results, itcan be estimated that a potential population of Salmonella wouldbe halved about 10 times, resulting in a 1,000-fold reductionin population size, in stomach contents from pigs fed the coarselyground meal feed. In comparison, the Salmonella population wouldbe halved about one to three times (eightfold reduction) inthe stomachs of pigs fed the other diets. Thus, the presentresults strongly suggest that a combination of low pH and highconcentrations of lactic acid in the stomach is important forreducing the number of Salmonella cells entering the small andlarge intestines of pigs.

A study by Jørgensen et al. (15) has demonstrated a clearpositive correlation between the incidence of Salmonella-positivepigs and the density of coliform bacteria in the gastrointestinaltracts of pigs (15). These results indicate that the densityof coliform bacteria is a reliable indicator of the populationof Salmonella in pigs. In the present study, a significant reductionin the number of coliform bacteria was observed in the distalsmall intestines, ceca, and colons of pigs fed the coarse feedcompared with pigs fed the fine feed. The lowest number of coliformbacteria was observed in pigs fed the coarsely ground meal feed.These results show that the grinding and processing of feedalso have an effect on the number of Enterobacteria, such ascoliforms and salmonellae, in the lower parts of the intestinaltract.

The most marked effect of grinding and processing of feed wason the microbial activity in the stomachs of pigs in the presentstudy. Less-marked effects in the lower compartments of thegastrointestinal tract were observed. The observed in vitrosurvival of S. enterica serovar Typhimurium DT12 in cecal contentindicates that there is neither an increase nor decrease inSalmonella numbers in the ceca of pigs. The results also demonstratethat the present in vitro method must be applied with some cautionwhen used with cecal content. The calculations of specific deathrates, pH, and SCFA concentrations should be based on the 4-hincubation period rather than a 6-h incubation period, due toan accumulation of lactic acid and a decrease in pH in the second2 h of the incubation period.

Higher concentrations of butyric acid were observed in the cecaand colons of pigs fed the coarsely ground feed than in thoseof pigs fed the finely ground feed. Stimulation of butyric acidin the hindgut is considered beneficial because it promotesthe growth of epithelial cells (27). Thus, the present resultsare in accordance with a study reported by Brunsgaard, showingsignificantly larger crypts and epithelial cell proliferationin the colons of pigs fed a coarse diet than in those of pigsfed a fine diet (4).

In conclusion, the present study has demonstrated numerous beneficialeffects of feeding pigs a coarsely ground meal feed insteadof a finely ground pelleted feed. The coarsely ground meal feedpromotes animal welfare by reducing gastric ulcer and diarrheaamong the pigs and promotes gastrointestinal health by reducingthe number of Salmonella cells and potentially harmful coliformsin the stomach and in the lower parts of the intestinal tract.The present study has put forward an explanation for the Salmonella-reducingproperties of coarsely ground meal feed. The coarsely groundmeal feed induces physical conditions in the stomach that promotethe growth of anaerobic bacteria, accompanied by the productionof organic acids, especially lactic acid, and a lowering ofthe pH. The high levels of lactic acid and low pH in the stomachthen kill Salmonella and Escherichia coli and consequently preventthese bacteria from entering, or reduce the numbers bacteriathat enter, the parts of the gastrointestinal tract in whichthey normally proliferate. In this way, the stomach acts asa barrier and breaks the cycle where animals in problematicherds infect each other and are reinfected through the consumptionof feces. In addition, the coarsely ground meal feed induceshigher concentrations of butyric acid in the hindgut, whichmay have beneficial effect on gut health by promoting growthof epithelial cells. A drawback of feeding a coarse meal feedto pigs, however, is the poor growth performance and consequentlythe financial loss to the pig producer. Nevertheless, the presentresults have contributed to our knowledge regarding the beneficialattributes of the coarse meal feed. This information may beused to design alternative feeds that can reduce the incidencesof Salmonella in pig herds without a negative effect on growthperformance.

ACKNOWLEDGMENTS

This work was financially supported by The Danish Bacon andMeat Council Basic Research Funding Program.

We thank Thomas Rebsdorf, Trine Poulsen, Karin Durup, and MonaDinsen for technical assistance.

FOOTNOTES

* Corresponding author. Mailing address: Danish Institute of Agricultural Sciences, Research Centre Foulum, P.O. Box 50, DK-8830 Tjele, Denmark. Phone: 45 8999 1133. Fax: 45 8999 1370. E-mail: LeneL.Mikkelsen{at}agrsci.dk.

REFERENCES

Top