Individual animal variability in ruminal bacterial communities and ruminal acidosis in primiparous Holstein cows during the periparturient period.

The purpose of this study was to investigate variability among individual cows in their severity of ruminal acidosis (RA) pre- and postpartum, and determine whether this variability was related to differences in their ruminal bacterial community composition (BCC). Variability in the severity of RA among individual cows was characterized based on ruminal fermentation variables. Effects of prepartum dietary treatment on the severity of RA were also examined. Fourteen Holstein heifers paired by expected calving date and BCS were allotted to 1 of 2 prepartum dietary treatments: low-concentrate or high-concentrate diets. All cows received the same lactation diet postpartum. Microbial DNA extracted from 58 ruminal digesta samples in total collected prepartum (d -50, -31, and -14; 27 samples) and postpartum (d +14 and +52; 31 samples) and amplified by PCR were subjected to automated ribosomal intergenic spacer analysis. Changes in ruminal variables over time [pH, volatile fatty acids (VFA), and acidosis indicators, including duration and area under the rumen pH curve below 5.8, 5.5, and 5.2, measured on d -54, -35, -14, -3, +3, +17, +37, and +58] were analyzed using principal components analysis. Based on the shift (defined as the distance of the mean loadings) between the prepartum and postpartum period for each cow, the 14 cows were classified into 3 groups: least acidotic (n=5), most acidotic (n=5), and intermediate (n=4). Cows in the most acidotic group had greater severity of RA (measured as duration of total RA, mild RA, moderate RA, and acute RA; area under the pH curve for total RA, mild RA, and moderate RA) postpartum than prepartum, and this difference between periods was greater than for the least acidotic cows. Similarly, the RA index (total area of pH <5.8 normalized to intake) showed an interaction between severity of RA and period. The variation in the severity of RA was independent of intake, total VFA concentration, and individual VFA proportions. Production variables (milk yield, fat percentage, fat yield, fat-corrected milk, and efficiency of milk production) were not influenced by the severity of RA. Ruminal BCC was not influenced by dietary treatment or period. However, some cows experienced greater shift in BCC than other cows across the periods. Based on the magnitude of the shift in BCC (distance between mean ordination values across the periods for each cow), cows were grouped into 3 BCC profile categories: stable (5 cows with lesser shift), unstable (5 cows with greater shift), and intermediate (4 cows with average shift). Cows demonstrating a greater shift in BCC were not necessarily those in the most acidotic group and vice versa. The shift in ruminal fermentation variables (principal components analysis rankings) and the shift in BCC (automated ribosomal intergenic spacer analysis rankings) between pre- and postpartum were not related (n=14; R(2)=0.00). It was concluded that not all cows are equally susceptible to RA and postpartum shifts in BCC appear to be independent of the differences in the severity of RA postpartum.

Changes in ruminal bacterial community composition following feeding of alfalfa ensiled with a lactic acid bacterial inoculant.

Some silage inoculants help to improve silage quality and promote an increase in milk production, possibly through altering the rumen microflora. We hypothesized that rumen bacterial community composition (BCC) would be different in cows fed alfalfa ensiled with the inoculant Lactobacillus plantarum MTD/1 (LP) compared with those fed alfalfa ensiled without the inoculant (Ctrl). Eight ruminally cannulated Holstein cows were allotted to 2 diets (Ctrl or LP) in a double crossover design with four 28-d periods. Diets were formulated to contain (% dry matter basis) 28.0% neutral detergent fiber and 16.2% crude protein, and contained alfalfa silage, 50.9; corn silage, 20.6; high-moisture shelled corn, 21.4; soy hulls, 4.7; plus minerals and vitamins, 2.4. Ruminal digesta were collected just before feeding on 3 consecutive days near the end of each period, and were separated into solid and liquid phases. Microbial DNA was extracted from each phase, amplified by PCR using domain-level bacterial primers, and subjected to automated ribosomal intergenic spacer analysis. The pH was 4.56 and 4.86 and the lactate-to-acetate ratio 9.8 and 4.4, respectively, for the treated and untreated alfalfa silages. Dry matter intakes and milk production data were not influenced by diets but showed a cow effect. Total volatile fatty acids (mM) tended to be greater for LP compared with Ctrl. Individual volatile fatty acids were not influenced by diets but showed a significant cow effect. Ruminal acetate (mol/100 mol) and acetate-to-propionate ratio were lower and propionate (mol/100 mol) greater for the 2 milk fat-depressed (MFD; <3.2% fat content) cows compared with the other 6 cows. Correspondence analysis of the 265 peaks in the automated ribosomal intergenic spacer analysis profile across the 188 samples revealed that the first 2 components contributed 7.1 and 3.8% to the total variation in the profile. The ordination points representing the liquid and solid phases clustered separately, indicating that these phases differed in BCC. The analysis of similarity data showed differences between Ctrl and LP. The lactic acid bacterial counts (log(10) cfu/g of wet silage) were 3.94 and 4.53 for the untreated and treated silage, respectively, at ensiling. The relative population size (RPS) of L. plantarum, determined by real-time PCR of 16S rRNA gene copies, was greater in LP compared with Ctrl. The ordination points corresponding to certain individual cows clustered separately, and the most distinctive bacterial communities were those associated with MFD cows. The RPS of Megasphaera elsdenii was greater in 1 of the 2 MFD cows, although mean RPS of M. elsdenii did not differ between the treatments. In addition to the differences in rumen BCC between LP and Ctrl, MFD cows also displayed differences in BCC compared with cows with normal milk fat yield.

Host specificity of the ruminal bacterial community in the dairy cow following near-total exchange of ruminal contents.

The purpose of this study was to examine the stability and host specificity of a cow's ruminal bacterial community following massive challenge with ruminal microflora from another cow. In each of 2 experiments, 1 pair of cows was selected on the basis of differences in ruminal bacterial community composition (BCC), determined by automated ribosomal intergenic spacer analysis (ARISA), a culture-independent "community fingerprinting" technique. Each pair of cows was then subjected to a 1-time exchange of >95% of ruminal contents without changing the composition of a corn silage/alfalfa haylage-based TMR. In experiment 1, the 2 cows differed (P<0.01) in prefeed ruminal pH (mean = 6.88 vs. 6.14) and prefeed total VFA concentration (mean = 57 vs. 77 mM), averaged over 3 d. Following exchange of ruminal contents, ruminal pH and total VFA concentration in both cows returned to their preexchange values within 24h. Ruminal BCC also returned to near its original profile, but this change required 14 d for 1 cow and 61 d for the other cow. In experiment 2, the 2 other cows differed in prefeed ruminal pH (mean = 6.69 vs. 6.20) and total VFA concentration (mean = 101 vs. 136 mM). Following exchange of ruminal contents, the first cow returned to its preexchange pH and VFA values within 24h; the second cow's rumen rapidly stabilized to a higher prefeed pH (mean = 6.47) and lower prefeed VFA concentration (mean = 120 mM) that was retained over the 62-d test period. Both cows reached somewhat different BCC than before the exchange. However, the BCC of both cows remained distinct and were ultimately more similar to that of the preexchange BCC than of the donor animal BCC. The data indicate that the host animal can quickly reestablish its characteristic ruminal pH and VFA concentration despite dramatic perturbation of its ruminal microbial community. The data also suggest that ruminal BCC displays substantial host specificity that can reestablish itself with varying success when challenged with a microbial community optimally adapted to ruminal conditions of a different host animal.

pH dynamics and bacterial community composition in the rumen of lactating dairy cows.

The influence of pH dynamics on ruminal bacterial community composition was studied in 8 ruminally cannulated Holstein cows fitted with indwelling electrodes that recorded pH at 10-min intervals over a 54-h period. Cows were fed a silage-based total mixed ration supplemented with monensin. Ruminal samples were collected each day just before feeding and at 3 and 6h after feeding. Solid and liquid phases were separated at collection, and extracted DNA was subjected to PCR amplification followed by automated ribosomal intergenic spacer analysis (ARISA). Although cows displayed widely different pH profiles (mean pH=6.11 to 6.51, diurnal pH range=0.45 to 1.39), correspondence analysis of the ARISA profiles revealed that 6 of the 8 cows showed very similar bacterial community compositions. The 2 cows having substantially different community compositions had intermediate mean pH values (6.30 and 6.33) and intermediate diurnal pH ranges (averaging 0.89 and 0.81 pH units). Fortuitously, these 2 cows alone also displayed milk fat depression, along with markedly higher ruminal populations of 1 bacterial operational taxonomic unit (OTU) and reduced populations of another ARISA amplicon. Cloning and sequencing of the elevated OTU revealed phylogenetic similarity to Megasphaera elsdenii, a species reportedly associated with milk fat depression. The higher populations of both M. elsdenii and OTU246 in these 2 cows were confirmed using quantitative real-time PCR (qPCR) with species-specific primers, and the fraction of total bacterial rDNA copies contributed by these 2 taxa were very highly correlated within individual cows. By contrast, the fraction of total bacterial rDNA copies contributed by Streptococcus bovis and genus Ruminococcus, 2 taxa expected to respond to ruminal pH, did not differ among cows (mean= <0.01 and 10.6%, respectively, of rRNA gene copies, determined by qPCR). The data indicate that cows with widely differing pH profiles can have similar ruminal bacterial community compositions, and that milk fat depression can occur at intermediate ruminal pH. The results support recent reports that milk fat depression is associated with shifts in bacterial community composition in rumine and is specifically related to the relative abundance of Megasphaera elsdenii.

Shifts in bacterial community composition in the rumen of lactating dairy cows under milk fat-depressing conditions.

Eighteen ruminally cannulated dairy cattle were fed a series of diets (in 28-d periods) designed to elicit different degrees of milk fat depression (MFD) for the purpose of relating MFD to ruminal bacterial populations. Cows were fed a TMR containing 25% starch (DM basis) supplied as corn silage, a slowly fermented starch (SFS treatment, period 1), then switched to a TMR containing 27% starch, much of it supplied as ground high-moisture corn, a rapidly fermented starch (RFS treatment, period 2). In period 3, the RFS diet was amended with 13.6 mg of monensin/kg of DM (RFS/Mon treatment), and in period 4, the cows were returned to the RFS diet without monensin (RFS/Post treatment). Effect of both starch source and monensin on milk fat percentage varied by cow, and cluster analysis identified 4 pairs of cows having distinct milk fat patterns. Archived ruminal liquors and solids from the 4 pairs were processed to isolate bacterial DNA, which was subjected to automated ribosomal intergenic spacer analysis followed by correspondence analysis to visualize bacterial community composition (BCC). One pair of cows (S-responsive) showed MFD on RFS feeding, but displayed no additional MFD upon monensin feeding and a fat rebound upon monensin withdrawal. The second pair of cows (M-responsive) showed no MFD upon switch from the SFS diet to the RFS diet, but displayed strong MFD upon monensin feeding and no recovery after monensin withdrawal. Both groups displayed major shifts in BCC upon dietary shifts, including dietary shifts that both did and did not change milk fat production. The third pair of cows (SM-responsive) displayed reduction of milk fat on both RFS and RFS/Mon diets, and fat returned to the levels on the RFS diet upon monensin withdrawal; these cows showed a more gradual shift in BCC in response to both starch source and monensin. The fourth pair of cows (nonresponsive) did not display changes in milk fat percentage with dietary treatment and showed only minor shifts in BCC with dietary treatment. Regardless of milk fat response, BCC did not reassemble its original state upon monensin withdrawal, though the difference was strongest in M-responsive cows. One amplicon length (representing a single bacterial species) was elevated in most, but not all, MFD-susceptible (S-, M-, or SM-responsive) cows relative to milk fat-nonresponsive cows, whereas 2 amplicon lengths displayed reduced abundance under MFD conditions. Overall, this study demonstrates an association between MFD and wholesale shifts of microbial communities in the rumen.

Isolation and characterization of a Trichodermastrain capable of fermenting cellulose to ethanol.

The direct fermentation of cellulosic biomass to ethanol has long been a desired goal. To this end, we screened the environment for fungal strains capable of this conversion when grown on minimal medium. One strain, identified as a member of the genus Trichoderma and designated strain A10, was isolated from cow dung and initially produced about 0.4 g ethanol l(-1). This strain cannot grow on any substrate under anaerobic conditions, but can ferment microcrystalline cellulose or several sugars to ethanol. Ethanol accumulation was eventually increased, by selection and the use of a vented fermentation flask, to 2 g l(-1) when the fermentation was carried out in submerged culture in minimal medium. The highest levels of ethanol, >5.0 g l(-1), were obtained by the fermentation of glucose. Little ethanol was produced by the fermentation of xylose, although other fermentation products such as succinate and acetate were observed. Strain A10 was also found to utilize (aerobically) a wide range of carbon sources. In addition, auxotrophic mutants were generated and used to demonstrate parasexuality by complementation between auxotrophs and between morphological mutants. The ability of this strain to use a wide variety of carbohydrates (including crystalline cellulose) combined with its minimal nutrient requirements and the availability of a genetic system suggests that the strain merits further investigation of its ability to convert biomass to ethanol.

Analysis of the replicon region and identification of an rRNA operon on pBM400 of Bacillus megaterium QM B1551.

An 18 633 bp region containing the replicon from the approximately 53 kb pBM400 plasmid of Bacillus megaterium QM B1551 has been sequenced and characterized. This region contained a complete rRNA operon plus 10 other potential open reading frames (ORFs). The replicon consisted of an upstream promoter and three contiguous genes (repM400, orfB and orfC) that could encode putative proteins of 428, 251 and 289 amino acids respectively. A 1.6 kb minimal replicon was defined and contained most of repM400. OrfB was shown to be required for stability. Three 12 bp identical tandem repeats were located within the coding region of repM400, and their presence on another plasmid caused incompatibility with their own cognate replicon. Nonsense, frameshift and deletion mutations in repM400 prevented replication, but each mutation could be complemented in trans. RepM400 had no significant similarity to sequences in the GenBank database, whereas five other ORFs had some similarity to gene products from other plasmids and the Bacillus genome. An rRNA operon was located upstream of the replication region and is the first rRNA operon to be sequenced from B. megaterium. Its unusual location on non-essential plasmid DNA has implications for systematics and evolutionary biology.

Characterization of 13 newly isolated strains of anaerobic, cellulolytic, thermophilic bacteria.

Characteristics of 13 newly isolated thermophilic, anaerobic, and cellulolytic strains were compared with previously described strains of Clostridium thermocellum: ATCC 27405 and JW20 (ATCC 31549). Colony morphology, antibiotic sensitivity, fermentation end-products, and cellulose degradation were documented. All 13 strains were sensitive to erythromycin (5 microg/ml) and chloramphenicol (25 microg/ml), and all strains but one were sensitive to kanamycin (20 microg/ml). Polymerase chain reaction (PCR) amplification using primers based on gene sequences from C. thermocellum ATCC 27405 was successful for all 13 strains in the case of the hydrogenase gene and 11 strains in the case of phosphotransacetylase/acetate kinase genes. Ten strains amplified a product of the expected size with primers developed to be specific for C. thermocellum 16SrRNA primers. Two of the 13 strains did not amplify any product with the PCR primers designed for the phosphotransacetylase/acetate kinase and 16SrRNA primers. A MboI-like GATC- recognizing restriction activity was present in all of the five strains examined. The results of this study have several positive implications with respect to future development of a transformation system for cellulolytic thermophiles.

Characterization of a theta plasmid replicon with homology to all four large plasmids of Bacillus megaterium QM B1551.

A replicon from one of an array of seven indigenous compatible plasmids of Bacillus megaterium QM B1551 has been cloned and sequenced. The replicon hybridized with all four of the large plasmids (165, 108, 71, and 47 kb) of strain QM B1551. The cloned 2374-bp HindIII fragment was sequenced and contained two upstream palindromes and a large (>419-amino-acid) open reading frame (ORF) truncated at the 3' end. Unlike most plasmid origins, a region of four tandem 12-bp direct repeats was located within the ORF. The direct repeats alone were incompatible with the replicon, suggesting that they are iterons and that the plasmid probably replicates by theta replication. The ORF product was shown to act in trans. A small region with similarity to the B. subtilis chromosomal origin membrane binding region was detected as were possible binding sites for DnaA and IHF proteins. Deletion analysis showed the minimal replicon to be a 1675-bp fragment containing the incomplete ORF plus 536 bp upstream. The predicted ORF protein of >48 kDa was basic and rich in glutamate + glutamine (16%). There was no significant amino acid similarity to any gene, nor were there any obvious motifs present in the ORF. The data suggest that this is a theta replicon with an expressed rep gene required for replication. The replicon contains its iterons within the gene and has no homology to reported replicons. It is the first characterization of a B. megaterium replicon.

Flow characteristics of bioprosthetic heart valves.

A review of the in vivo and in vitro fluid dynamic performance of three bioprosthetic heart valves is presented. Data on Hancock porcine valves (standard models 242 aortic and 342 mitral and modified orifice model 250 aortic), Carpentier-Edwards porcine valves (model 2625 aortic and 6625 mitral), and the Ionescu-Shiley pericardial valve are reviewed. These valves were chosen because of their past or present popularity in clinical use and because of the variation in fluid dynamic performance reported by different investigators. The flow parameters that are reported include in vivo and in vitro mean pressure drop, cardiac output or cardiac index, regurgitant volume, effective orifice area, and performance index. These data provide a framework for differentiation of normal and abnormal bioprosthetic valve function.

Flow characteristics of four commonly used mechanical heart valves.

The in vivo and in vitro fluid dynamic performance of 4 mechanical heart valves was reviewed: Starr-Edwards silicon-rubber ball valves (models 1200/1260 aortic and 6120 mitral valves), Björk-Shiley tilting disc valves (standard spherical model, modified and unmodified convexo-concave [60 degrees and 70 degrees C-C] models), the Medtronic-Hall (Hall-Kaster) tilting disc valve and the St. Jude Medical bileaflet valve. These valves were chosen because of their past or present popularity in clinical use and because they encompass most of the basic mechanical valve designs used during the past 2 decades. The flow measurements reported include in vivo and in vitro mean pressure drop, cardiac output or cardiac index, regurgitant volume, effective orifice area and performance index.

Numerical simulation of steady turbulent flow through trileaflet aortic heart valves--I. Computational scheme and methodology.

A numerical simulation model and technique are described to simulate steady turbulent blood flow through trileaflet tissue valves of varying degrees of stenosis. The aortic trileaflet tissue valve design was chosen as the subject of this study, since it is the only popular valve in current clinical use which is approximately axisymmetric. An axisymmetric geometry is computationally more convenient since it involves only two dimensional equations. The geometry and dimensions of the aorta were designed from angiographic studies and measurements made from cadavers. The valve dimensions were obtained from tissue leaflet photography studies conducted on tissue bioprostheses of varying degrees of stenosis. The non-rectangular nature of this valve necessitated the use of a body conforming grid. Thompson's method coupled with a Chimera grid system was chosen for this purpose. The Chimera grid was used to avoid a grid with highly skewed cells. Turbulence was simulated by using the kappa-epsilon model with the wall function method. This decision was made after comparing the kappa-epsilon model's performance with that of lower order models, and after considering the increased computer time requirements and decreased stability of more complex models, such as the Reynolds stress model. The results of the study which are very encouraging and compare favorably with in vitro experimental data, are described in Part II of the paper.

Numerical simulation of steady turbulent flow through trileaflet aortic heart valves--II. Results on five models.

Turbulent flow simulations are run for five aortic trileaflet valve geometries, ranging from a valve leaflet orifice area of 1.1 cm2 (Model A1--very stenotic) to 5.0 cm2 (Model A5--natural valve). The simulated data compares well with experimental measurements made downstream of various aortic trileaflet valves by Woo (PhD Thesis, 1984). The location and approximate width and length of recirculation regions are correctly predicted. The less stenotic valve models reattach at the end of the aortic sinus region, 1.1 diameters downstream of the valve. The central jet exiting the less stenotic valve models is not significantly different from fully developed flow, and therefore recovers very quickly downstream of the reattachment point. The more stenotic valves disturb the flow to a greater degree, generating recirculation regions large enough to escape the sinuses and reattach further downstream. Peak turbulent shear stress values downstream of the aortic valve models which approximated prosthetic valves are 125 and 300 Nm-2, very near experimental observations of 150 to 350 Nm-2. The predicted Reynolds stress profiles also present the correct shape, a double peak profile, with the location of the peak occurring at the location of maximum velocity gradient, which occurs near the recirculation region. The pressure drop across model A2 (leaflet orifice area 1.6 cm2) is 20 mmHg at 1.6 diameters downstream. This compares well with values ranging from 19.5 to 26.2 mmHg for valves of similar orifice areas. The pressure drop decreases with decreasing valve stenosis, to a negligible value across the least stenotic valve model. Based on the good agreement between experimental measurements of velocity, shear stress and pressure drop, compared to the simulated data, the model has the potential to be a valuable tool in the analysis of heart valve designs.

An on-line method for evaluation of the in vitro pulsatile pressure drop and regurgitative characteristics of prosthetic heart valves.

An on-line digital method is described for analyzing the pressure drop and regurgitative characteristics of prosthetic heart valves in a pulse duplicator system. The method is based on the Apple II microcomputer system but could be modified to be used with most microcomputers currently available. The digital data collection system is synchronized with the pulse duplicator system, and is programmed to collect the relevant pressure (upstream and downstream) and volumetric flow data for 10 cardiac cycles at a given time. The system is relatively easy to use and gives the scientist a fair amount of flexibility in terms of data collection, storage, and analysis. The on-line method has been used with the pulse duplicator of the Georgia Institute of Technology for the past 3 years.

In vitro fluid dynamic characteristics of Ionescu-Shiley and Carpentier-Edwards tissue bioprostheses.

In the study reported here, the in vitro fluid dynamic characteristics of the Ionescu-Shiley (calf pericardial) and Carpentier-Edwards (porcine) aortic tissue valves were studied. The experiments conducted were pressure drop measurements, leaflet photography, flow visualization, and velocity measurements. The pressure drop studies indicated that both types of tissue valves created relatively large pressure drops. These pressure drops were larger than those observed with the corresponding sizes of Bjork-Shiley, Hall-Kaster, and St. Jude aortic prostheses. The photographs of the opening of the valve leaflets indicated that the tissue valves do not open as ideally as do the natural valves. It was also observed that the Ionescu-Shiley aortic valves opened more symmetrically and with reproducibility than the corresponding Carpentier-Edwards aortic valves. Velocity and shear stress measurements made with a laser-Doppler anemometer indicated that the flow that emerged from the leaflets for both types of tissue valves was like a jet and could lead to turbulent shear stress on the order of 1,000-3,000 dynes/cm2. Such turbulent shear stresses could be harmful to blood components. The jet-type flow could also damage the endothelial lining of the wall of the ascending aorta. The velocity measurements also indicated an annular region of stagnant fluid between the outflow surfaces of the leaflets and the flow channel wall. Such a region could lead to the build-up of thrombotic, fibrotic, and/or calcific material on the outflow surfaces of the leaflets. Both types of valve designs, however, created relatively low wall shear stresses and regurgitant volumes.

In vitro fluid dynamic characteristics of the Medtronic-Hall pivoting disc heart valve prosthesis.

The need for better low-profile mechanical valves led to the design and development of the Medtronic-Hall (formerly known as the Hall-Kaster valve) pivoting disc heart valve prosthesis in 1976. In vitro flow studies indicate that it has improved pressure drop characteristics compared to the Lillehei-Kaster and convexoconcave Björk-Shiley (60 degrees model) tilting disc valves. It does, however, have a somewhat larger regurgitant volume compared to the Björk-Shiley valve design. Velocity measurements with a laser-Doppler anemometer in the immediate downstream vicinity of the Medtronic-Hall valve indicate no region of stagnation near the outflow face of the disc. Regions of stagnation were, however, observed adjacent to the two titanium "pivot stops" situated on either side of the disc in the major orifice and along the pivot post in the minor orifice, together with a region of flow separation adjacent to the sewing ring of the minor outflow region. The results of the present in vitro study indicate a small but significant improvement in the overall fluid dynamic performance of the Medtronic-Hall valve, compared to the convexo-concave Björk-Shiley (60 degrees model) and Lillehei-Kaster tilting disc prostheses in current clinical use.

The Björk-Shiley heart valve prosthesis. Flow characteristics of the new 70 degrees model.

Thrombus formation on the outflow face and tissue overgrowth along the sewing ring adjacent to the minor outflow region are major clinical pathologic problems observed with the Björk-Shiley tilting disc valve. In the hope of reducing these pathologic conditions the convexo-concave Björk-Shiley valve was designed. A new modification to the convexo-concave model has further improved most of the fluid dynamic characteristics of the valve. In vitro flow studies indicate an average improvement in pressure drop characteristics of about 20 to 30%. Velocity measurements made with a laser-Doppler anemometer in the immediate downstream vicinity of the modified convexo-concave valve indicate that the design changes have further reduced the size of the stagnation zone near the outflow face of disc. It was observed, however, that the regurgitation across the valve increased by about one to two percentage points.