In vitro validation of the lactose 13C-ureide breath test for equine orocaecal transit time measurement

Authors

  • D. G. M. SUTTON,

    Corresponding authorSearch for more papers by this author
  • T. PRESTON,

    1. School of Veterinary Medicine, College of Medical, Veterinary and Life Sciences, University of Glasgow; and Isotope Biochemistry Laboratory, Scottish Universities Environmental Research Centre, Scottish Enterprise Technology Park, East Kilbride, UK.
    Search for more papers by this author
  • S. LOVE

    1. School of Veterinary Medicine, College of Medical, Veterinary and Life Sciences, University of Glasgow; and Isotope Biochemistry Laboratory, Scottish Universities Environmental Research Centre, Scottish Enterprise Technology Park, East Kilbride, UK.
    Search for more papers by this author

email: david.sutton@glasgow.ac.uk

Summary

Reasons for performing study: Validation of a reliable, noninvasive clinical test for quantification of equine orocaecal transit time (OCTT) is required. This would facilitate an evidence-based approach to investigation and treatment of equine small intestinal disorders.

Objectives: 1) Comparison of the lactose 13C-ureide breath test (LUBT) with the hydrogen breath test (H2BT) for OCTT measurement. 2) Identification of the characteristics of gastrointestinal microbial glycosylureide hydrolase activity in vitro. 3) Production of an optimised protocol for the LUBT for in vivo measurement of equine OCTT.

Hypothesis: Significant lactose 13C-ureide (13C-LU) hydrolase activity is restricted to the large bowel. The rate of expiratory 13CO2 production after ingestion of the isotope will provide an indirect quantifiable measure of orocaecal transit rate. Requisite bacterial activity may be enhanced by a primer dose of unlabelled substrate as shown in Man.

Methods: Combined LUBT and H2BT were performed in 8 healthy individuals. Analysis of sequential end expiratory breath samples was used to calculate OCTT and results compared. Digestion of 13C-LU was investigated in vitro using fresh faecal material or intestinal aliquots collected post mortem. Isotopic fermentation rate was measured by rate of appearance of 13CO2.

Results: Peaks in expiratory 13CO2 occurred in all individuals after ingestion of the labelled test meal, whereas H2 expiration was variable. Both faecal and intestinal microbial digestion of 13C-LU were maximised by prior exposure to 12C-LU. Induced bacterial glucoseureide hydrolase activity was significantly greater in the caecum than in the small intestine (n = 10, P<0.05).

Conclusions: Significant 13C-LU digestion is restricted to the equine large intestine under normal conditions, and is enhanced by prior exposure to 12C-LU, making 13C-LU a suitable noninvasive marker of equine OCTT. The LUBT is more reliable than the H2BT for measurement of equine OCTT.

Introduction

The development of a repeatable noninvasive test for equine orocaecal transit time (OCTT) measurement would enhance diagnostic capabilities and facilitate an evidence based approach to gastrointestinal motility research in this species. Tests reported for measurement of equine OCTT include the hydrogen breath test (H2BT) (Bracher and Steiger 1998; Murphy et al. 1998) and the sulphasalazine/sulphapyridine (SLZ/SP) test (McGreevy and Nicol 1998). The principle of the H2BT is that the arrival of undigested carbohydrate in the caecum is followed by microbial H2 production, which diffuses into mucosal capillary blood prior to producing a detectable increase in end expiratory breath H2 output (Calloway 1966). The SLZ/SP test is also dependent on bacterial activity, using the rate of appearance of SP in the blood as an indirect measure of OCTT, following bacterial cleavage of the SLZ azo-bond (Kellow et al. 1986). Due to the lack of an established standard for the detection of OCTT in equids, neither the H2BT nor the SLZ/SP test has been validated for specific assessment of this parameter, and their efficacy is dependent on restriction of the requisite microflora to the large bowel. In human medicine, the most accurate modality for detection of the arrival of ingesta in the caecum is gastroenterocolonic scintigraphy (GECS) (Camilleri et al. 1998) and this is used as the reference technique for validation of new methodology. However, scintigraphic examination of the equine abdomen is not sufficiently specific for determination of OCTT due to obscuring of the caecal base by loops of radioactive small intestine, necessitating development of an alternative technique.

Lactose 13C-ureide (13C-LU) is a glycosylureide compound, consisting of a sugar moiety (galactose-glucose) bound to a 13C-urea molecule. The glucose-urea bond itself cannot be cleaved by intestinal brush border enzymes (Ruemmele et al. 1997) but can be cleaved by colonic microflora in ruminants (Merry et al. 1982a) and humans (Heine et al. 1995), with the resultant release and hydrolysis of 13C-urea. By measuring the rate of appearance of exhaled 13CO2 after ingestion of the isotope, an indirect measure of orocaecal transit rate of ingesta is obtained. The lactose ureide breath test (LUBT) has been demonstrated to have greater sensitivity and specificity than the older lactulose-H2 breath test forOCTT measurement (Wutzke et al. 1997). Geypens et al. (1999) also reported that the LUBT was closely correlated to GECS for the measurement of OCTT (r2= 0.94, n = 22, P = 0.0001) and concluded that it was a valid alternative for determination of this parameter in Man.

Previous exposure to unlabelled glycosylureide (induction) has been shown to increase the subsequent rate of hydrolysis of the 13C-compound in both human faecal (Mohr et al. 1999; Morrison 2000) and ruminant (Merry et al. 1982a,b) microbial populations. This phenomenon results either from selected bacterial proliferation or enzymic induction, and clarifies the caecal arrival time of labelled ingesta by augmenting the expiratory 13CO2 signal. In order to maximise test accuracy, an induced test protocol is recommended for the LUBT in Man, with ingestion of unlabelled isotope on the day prior to the test (Geypens et al. 1999).

The major aim of this study was to investigate the potential value of the LUBT for equine use and to attempt its validation for OCTT measurement by comparison with existing methods. Using a preliminary protocol for the LUBT, it was aimed to compare transit time estimates with concurrent H2BT results in vivo. In order to assess validity of the LUBT for equine OCTT measurement and to optimise a test protocol for clinical use, in vitro microbial fermentation techniques were then used to evaluate the specific properties of equine microbial 13C-LU digestion.

Materials and methods

The study was approved by the Glasgow University Ethics and Welfare Committee.

Comparison of the LUBT with the H2BT for the measurement of OCTT

Subjects: Preliminary breath tracer studies were performed in 8 healthy mature individuals, comprising 4 native ponies (mean ± s.d. age 14.0 ± 4.3 years; body mass 309.5 ± 94.5 kg) and 4 Thoroughbreds (age 8.3 ± 3.3 years; body mass 541.0 ± 74.9 kg). Each subject had no known history or physical evidence of gastrointestinal disease and was in good body condition. For the purposes of the stable isotope breath test, all subjects were maintained on a ration of ryegrass seed hay to minimise fluctuations in basal 13CO2 output.

Study design: Combined H2BT and noninduced LUBTs were performed in 4 individuals, followed after 7 days by a combined tracer study using an induction protocol for the LUBT. Intermodal differences in OCTT estimation were measured. Intraindividual variation in H2 production was assessed by repeating the H2BT 3 times at weekly intervals in 4 individuals.

Hydrogen breath test: A test meal was developed to optimise subsequent H2 expiration. The final test meal consisted of 400 g high fibre pellets, 150 g crimped oats, 100 g wheat bran and 200 ml water, and was ingested voluntarily after a 14 h fast. End expiratory tidal air was collected prior to test meal ingestion and then at 30 min intervals for 24–36 h using a modified Aeromask1 and QuinTron2 collection bag. Sample H2 concentration (parts/million, ppm) was measured using an Exhaled Hydrogen Monitor3 according to the method of Corbett et al. (1981).

The lactose 13C-ureide breath test: A tracer dose of 3.0 mg/kg bwt β-lactosyl 13C-ureide dihydrate4 was added to the test meal. For the preliminary induction protocol, 5 g lactose 12C-ureide was added to the diet 12–16 h before the start of the test, as recommended by Morrison (2000). Noninduced LUBT protocols were performed before induced LUBTs in each subject, as the duration of any ‘induction’ effect was not known. Expiratory breath samples for measurement of basal 13C : 12C ratio were collected prior to test meal ingestion and then at 30 min intervals for 24–36 h, using the previously validated technique (Sutton et al. 2003). Samples were stored in Exetainers5 prior to analysis by isotope ratio mass spectrometry (ABCA)6. Storage of 13C expiratory samples even for several weeks prior to analysis does not affect sample quality (Schoeller et al. 1977).

Calculation of orocaecal transit time: For the H2BT, OCTT was determined as that time at which there first was a sustained increase of at least 5 ppm in expiratory H2 output, as recommended by Sciarretta et al. (1994) and Miller et al. (1997). For the LUBT, isotope recovery data were modelled using a delayed Maes formula (Maes et al. 1994), fitted by the least squares method using the Excel Solver function7, as given in Equation 1. OCTT was defined as that point at which 3% of the total expired cumulative percentage dose recovery (PDR) of the isotope had occurred (Equation 2).

image(Equation 1)
image(Equation 2)

where y(t) is the % 13CO2 excretion in breath/h; t is time (h); a, b and c are rate constants and d is the delay factor.

In vitro digestion of lactose 13C-ureide by equine gastrointestinal microbes

Faecal incubation: Fresh specimens were collected from mature horses and ponies maintained on a ryegrass hay diet, and which had not received medication within the previous 2 month period. A 400 g/l faecal slurry was formed using anaerobic Sorensen's phosphate buffer at pH 7.0, and homogenised for 2 min. The buffer was boiled for 10 min before use and cooled under a stream of oxygen-free nitrogen (OFN) to ensure anaerobic conditions. The slurry was then filtered through nylon meshing to remove coarse particulate matter (Edwards et al. 1996), before transferring 40 ml aliquots to 120 ml glass serum bottles. The bottles were closed with gas tight crimp lids and Teflon seals and fitted with a needle and Luer-lock tap, prior to flushing with OFN and placing in a shaking water bath (38°C, 50 revolutions/min). Headspace gas was sampled using a closed technique via the needle and Luer-lock tap, and transferred to an Exetainer5 that had previously been flushed with OFN. Isotope ratio mass spectrometric analysis (ABCA) of sequential 1 ml aliquots of headspace gas was used to measure microbial activity and 13C-substrate fermentation by measuring rate of increase in 13C : 12C ratio.

Intestinal incubations: Intestinal contents were collected from the stomach, duodenum, jejunum, ileum and caecum of horses subjected to euthanasia for conditions not relating to the gastrointestinal tract. Specimens were placed in OFN-flushed sealed containers after measurement of pH, prior to preparation of a 90% w/v slurry using preboiled Sorensen's buffer of equivalent pH. Anaerobic fermentation chambers were then prepared as above. Viable microbial activity in each chamber was assessed initially by the appearance of H2 in the chamber headspace gas. Chamber pH was remeasured prior to the addition of the 13C-substrate, and the experiment discontinued if pH had altered by >0.5 from time of preparation.

Calculation of 13C-compound fermentation rate: Recovery rate of the labelled substrate was plotted either as nmol excess 13C or % dose recovery of the isotope, against time on the x-axis. In order to calculate nmol excess 13C from ppm excess 13C, dilution in the sampling process was accounted for as shown in Equation 3:

image(Equation 3)

Since the concentration of 13CO2 was representative of the cumulative dose recovery of the isotope at any time in the closed fermentation system, its recovery was modelled using Siegel's formula, as given in Equation 4:

image(Equation 4)

where y(t) is the cumulative % 13C excretion in breath at time t; m is the total cumulative 13C recovery when time is infinite; k is the gradient of the curve and β is a rate constant. The time to hydrolysis of X% substrate as given by the modelled function was calculated as below:

image(Equation 5)

Study design

Effect of induction period on faecal glucoseureide hydrolase activity: Faecal samples (n = 6) were collected, prepared as above, and subjected to the following protocol: 12C-LU (50 mg) was added to 4 chambers at set interval (T-19 h, T-14 h, T-10 h and T0 h) before the addition of 10 mg of 13C-LU at zero time (T0 h). Three control chambers were used for each individual: these received no induction dose of 12C-LU, and were treated with either 10/20 mg 13C-LU or equimolar 13C-urea (0.11 mg) at T0 h. Basal 1 ml aliquots of headspace gas were collected from each chamber prior to addition of the 13C-LU and thereafter at 30 min intervals for 6 h to assess fermentation rate.

Intestinal distribution of 13C-LU fermenting bacteria: Intestinal luminal contents were collected from stomach, duodenum, mid-jejunum, ileum and caecal body of 3 mature horses immediately after euthanasia for orthopaedic disorders, and processed as above. Three of 4 chambers from each locus were primed with 10 mg 12C-LU at 14, 12 or 8 h prior to the addition of 2 mg 13C-LU at time 0 h; the remaining control chamber did not have a primer dose added. Samples (1 ml) of headspace gas were collected for 13C : 12C analysis at T-1 h, and thereafter hourly for 10 h.

Properties of caecal microbial induction: Anaerobic fermentation chambers were established using buffered 90% w/v fresh caecal contents from 2 mature horses collected post mortem. To each chamber, 10 mg 13C-LU was added at time 0 h, and this alone was added to the control chambers. To the remainder, one of the following substrates was added at T-24 h: 12C-LU (40 mg/9.5 × 10−5 mol); glucose (1.9 × 10−4 mol); lactose (9.5 × 10−5 mol); blood or cooked meat culture media8. Sequential headspace gas 13C:12C analysis was continued for 25 h.

Interindividual variation in locus of glucoseureide hydrolase activity: Intestinal contents were collected from 10 mature abattoir specimens as soon as possible after death, via stab incisions in the mid-jejunum, ileum (60 cm orad of the ileocaecal valve) and caecal apex. In 4/10 individuals, samples were combined from the entire small intestine due to very limited content. Samples were stored in OFN-flushed airtight containers prior to production of anaerobic fermentation chambers at appropriate pH and temperature, using 90% w/v buffered slurry. Four incubations were performed per individual: caecal controls (equimolar lactose/no induction), and duodenal/jejunal, ileal and caecal samples, all induced at T-15 h with 12C-LU at a ratio of 10:1 for primer to 13C-substrate. The rate of 13C-LU fermentation was measured over 10 h and calculated for each system (see above).

Results

Comparison of the LUBT with the H2BT for the measurement of OCTT

Expiratory H2 production was variable within and between test subjects, with one nonproducer. With the optimised test meal only 15/22 tests (68.1%) in 7 subjects resulted in significant expiratory H2 peaks, whereas peaks in expiratory 13CO2 occurred in every LUBT. The induction protocol for the LUBT produced a more discrete peak in expiratory 13CO2 in all cases, with higher cumulative dose recovery of the stable isotope and apparent shortening of OCTT. Mean ± s.d. values for OCTT (n = 4) using concurrent H2BT and induced LUBT were 2.78 ± 1.15 h and 4.43 ± 0.53 h, respectively. The mean difference of 1.65 ± 1.16 h for OCTT estimation reflected a very early H2 peak in one of 4 individuals, possibly caused by gastric or small intestinal fermentation. In the remaining individuals (n = 3), estimation of OCTT by the induced LUBT exceeded that of the H2BT by 1.07 ± 0.17 h.

Tracer recovery for one combined H2BT and induced LUBT is illustrated (Fig 1).

Figure 1.

Combined LUBT and H2BT in a mature horse after ingestion of labelled test meal at time 0 h. Unlabelled lactose ureide (5 g) was added to the diet 24 h prior to the test. Two further meals of unlabelled high fibre cubes (1.6 kg) were fed at 13 and 26 h (arrows). Three discrete peaks in H2 production occurred at regular intervals after each meal (solid circles). Cumulative expiratory 13CO2 recovery (bold line) was modelled using the multipeak formula of Maes (fine lines show individual peaks in 13CO2 recovery). OCTT estimation by LUBT and H2BT=4.49 and 4.00 h, respectively.

When the H2BT was repeated weekly (n = 3) in 4 subjects, only 50% produced a detectable increase in H2 expiration on every occasion.

In vitro digestion of lactose 13C-ureide by equine gastrointestinal microbes

Faecal glucoseureide hydrolase activity: Pretreatment with 12C-LU (‘induction’) at T-19 h resulted in a significant increase in the initial rate of 13C-LU hydrolysis, when compared with the noninduced, T-14 h and T-10 h samples. The time to reach 1% PDR was significantly shorter in the T-19 h samples in all individuals (paired t tests; n = 6; P<0.01). The cumulative dose recovery of 13C-LU was increased significantly by the earlier addition of 12C-LU to the incubation systems at T-19 h. In every control sample, 13C-urea metabolism proceeded rapidly without a lag phase (Fig 2).

Figure 2.

Rate of lactose 13C-ureide digestion by faecal anaerobe preparations from one mature horse, demonstrating effect of pretreatment with lactose 12C-ureide (50 mg) at different intervals prior to addition of lactose 13C-ureide (10 mg) at time zero (0 h). The rate of 13C-urea hydrolysis is plotted for comparison.

Intestinal distribution of 13C-LU fermenting bacteria: Caecal 13C-LU hydrolase activity was present in each case and was maximised by priming with 12C-LU. Mean ± s.d. time to 1% dose recovery for the caecal noninduced and induced samples was 4.70 ± 1.58 h and 0.55 ± 0.22 h, respectively (n = 3; P<0.001). Dose recovery at 10 h in the induced caecal samples was 39.88 ± 11.34% compared with 16.44 ± 15.72% in the noninduced fermentation chambers (P<0.1). There was no fermentation of 13C-LU in the gastric chambers with minimal digestion occurring in the duodenal/jejunal or ileal samples, and this was not enhanced significantly by priming with 12C-LU. In the induced duodenal/jejunal samples, dose recovery at 10 h was 0.51 (0.46) %, significantly less than the caecal samples (P<0.001). In the T-14 h induced ileal samples, time to 1% PDR was 3.73 (1.26) h, with a mean cumulative dose recovery of just 4.78 (1.12) %; both of these values were significantly different to the induced caecal samples (P<0.001). Summary data are given in Figure 3.

Figure 3.

Mean+s.d. rate of fermentation (n=3) of lactose 13C-ureide by equine jejunal, ileal and caecal anaerobes, as determined by appearance of 13CO2 in anaerobic chamber headspace gas. The maximum values gained for jejunal and ileal 13C-LU fermentation are shown. Lactose 13C-ureide added to each chamber at 0 h.

Caecal microbial induction: Priming with 12C-LU at T-24 h resulted in a significant reduction in time to 13C-LU 1% PDR when compared with the noninduced sample (P<0.001). This was not replicated by equimolar quantities of lactose or glucose, or by blood or meat anaerobic culture media (Fig 4). However, the cumulative dose recovery of the 13C-LU at 25 h was significantly increased by the addition of both lactose (52.20 ± 1.83%) and glucose (44.94 ± 0.23%) when compared with the 12C-LU induced (32.15 ± 0.93%) and noninduced (25.60 ± 0.42%) samples. Thus, although initial 13C-LU digestion rate was slower in these systems, it was more sustained, resulting in a higher end PDR.

Figure 4.

Effect of specific substrates and culture media on the rate of digestion of lactose 13C-ureide by equine caecal anaerobes. The mean values (n= 2) and s.d. are shown at each point. 40 mg 12C-LU or equimolar carbohydrate was added to the chambers 24 h before addition of 10 mg 13C-LU at time 0 h.

Population variation in locus of glucoseureide hydrolase activity: To quantify the differences observed in 13C-LU fermentation rate, the mean times taken to reach 1% and 3% PDR, and the cumulative isotopic dose recovery at 10 h were compared between the intestinal loci by a general linear model ANOVA, using Tukey pairwise comparisons with 95% confidence intervals (Minitab9). These values are illustrated (Fig 5) and presented numerically in Table 1. The mean rate of 13C-LU fermentation as measured by time to both 1% PDR and 3% PDR was significantly shorter in the induced caecal apex samples than in the duodenal/jejunal and ileal samples (P<0.05), and cumulative dose recovery at 10 h was also significantly higher in this group (P<0.05). In fact, in the duodenal/jejunal and ileal samples, there was negligible fermentation of 13C-LU, with mean cumulative PDRs of 1.11 ± 0.66% and 0.81 ± 0.55%, respectively, at 10 h. In the combined small intestinal samples, 2 individuals had fermentative activity, causing appreciable 13C-LU digestion. However, in each of these individuals the induced caecal samples showed a greater rate of 13C-LU fermentation.

Figure 5.

Mean+ 1 s.d. rate of lactose 13C-ureide fermentation by anaerobes from the small and large intestinal tract of 10 mature horses. Each sample was induced with a 10:1 ratio of unlabelled substrate at 15 h prior to addition of the 13C-isotope. The time to reach 3% cumulative dose recovery of the isotope in the headspace gas is shown (dotted line).

Table 1. Rate of intestinal microbial digestion of lactose 13C-ureide in vitro in 10 mature horses. Each anaerobic chamber was induced with 12C-LU at 15 h prior to addition of the 13C-isotope
Equine intestinal locusSample sizeTime to 1% PDR mean ± s.d (h)Time to 3% PDR mean ± s.d. (h)Cum PDR at 10 h mean ± s.d. (h)
  1. Matching superscript letters denote significant difference between pairs (P<0.05).

Duodenum/jejunum67.48 ± 3.95a18.58 ± 7.40e1.11 ± 0.66i
Ileum610.76 ± 1.74b,c,d15.42 ± 4.52f,g,h0.55 ± 0.81j
Combined small intestine42.59 ± 0.80c5.51 ± 1.22f8.76 ± 2.78
Caecal apex (not induced)104.53 ± 3.32d9.18 ± 3.71g5.95 ± 7.22
Caecal apex (induced)101.81 ± 1.42a,d3.65 ± 2.38e,h11.53 ± 7.44i,j

Discussion

In order to validate a new diagnostic technique for OCTT measurement, it is necessary to have an established standard method against which comparison can be drawn. The H2BT was found to be unreliable for this purpose in this study, due to the presence of nonproducers, and inter- and intraindividual variability in expiratory H2 production following test meal ingestion. Equine H2 output has been reported as variable in previous studies (Murphy 1997), potentially due to its consumption by intestinal methanogenic bacteria (Sasaki et al. 1999), and nonproducers have also been reported (Bracher and Baker 1994). A further pitfall of the H2BT is that stimulation of the gastrocolic reflex during test meal ingestion may stimulate movement and fermentation of previously consumed carbohydrate, thus increasing bacterial H2 activity and confounding interpretation (Romagnuolo et al. 2002). This phenomenon is illustrated in Figure 1. An intermodal difference in OCTT estimation was present between the 2 tracer tests, which may have resulted from different H2/13CO2 production rates by the specific caecal bacteria, or the time taken for equilibration of 13CO2 in the bicarbonate pool before expiration.

Ingestion of 13C-LU was followed in all individuals by a discrete peak in expiratory 13CO2 output and the clarity of this peak was enhanced by prior administration of 12C-LU as noted in both ruminants (Merry et al. 1982b) and man (Wutzke et al. 1997; Geypens et al. 1999). Faecal and caecal fermentations confirmed that the duration of 12C-LU exposure was critical, with a minimum period of 14 h required before a significant increase appeared in the subsequent rate of 13C-LU digestion. This effect was not attributable to changes in the rate of 13C-urea hydrolysis, as this occurred rapidly without a lag phase. Nor was it due to the provision of energy in the form of LU, as equimolar quantities of glucose and lactose were shown to have very little effect on subsequent 13C-LU digestion. It was concluded that prior exposure to LU caused either specific synthesis upregulation of a bacterial glucoseureide hydrolase, or resulted in selective multiplication of those organisms capable of its digestion.

The required 12C-LU primer : 13C-LU substrate ratio was a less important factor in fermentative activity than duration of induction. In initial further (unpublished) experiments, ratios of primer : substrate exceeding 10:1 resulted in reduced cumulative substrate digestion, possibly due to enzyme saturation. The optimum ratio using in vitro caecal preparations appeared to be between 2 and 5:1 for 12C-LU : 13C-LU ratio.

All caecal preparations (15/15) exhibited 13C-LU digestion whilst gastric and small intestinal 13C-LU fermentative activity was negligible in 10/12 (83.3%). Given the high specificity of the equine caecum for LU fermentation, it is concluded that 13C-LU is an excellent marker of caecal microbial activity. On the basis of the in vitro fermentations, the LUBT is likely to be a valid test for the measurement of equine OCTT in vivo, with arrival of ingesta in the caecum marked by an increased 13C : 12C expiratory ratio. However, the data suggest that a priming dose of 12C-LU is essential to maximise caecal 13C-LU fermentative activity and aid measurement of OCTT.

Clostridium innocuum is the only human intestinal microbe shown specifically to degrade lactose ureide and, although others may have hydrolase activity, the property is exclusive (Mohr et al. 1999). This organism does not have urease activity, but since at least 18% of equine caecal bacterial isolates are capable of urea utilisation (Maczulak et al. 1985), urea hydrolysis is unlikely to be a rate-limiting stage in the equine LUBT, as demonstrated.

In vitro models of intestinal fermentation were used in this study to avoid the requirement for surgical cannulation of multiple horses. Potential limitations of such systems include the accumulation of bacteriostatic fermentation products, and abnormal bacterial growth patterns due to the exhaustion of substrate unrelated to the one of specific interest (Merry et al. 1982b), or due to swings in chamber pH (Edwards et al. 1985). The unavoidable delay from removal of intestinal contents to re-establishment of optimum anaerobic conditions in vitro may inevitably change bacterial composition and metabolic activity. This was a particular factor when samples were being collected from multiple abattoir specimens. Given the empty small intestinal tract of 4/10 abattoir specimens, it is possible that unreported disorders, including small intestinal bacterial overgrowth, may have contributed to the small intestinal 13C-LU fermentation seen in 2 specimens.

Prior to clinical application of the LUBT, further factors must be acknowledged. Concentrate diets have been shown to promote caecal bacterial multiplication after 3–7 days (Goodson et al. 1988; Moore and Dehority 1993), which could potentially affect 13C-LU metabolism. Significant intestinal disease, such as acute grass sickness, has been shown to cause a 10-fold increase in ileal Gram-positive rods compared to controls (Garrett et al. 2002), such that reliability of OCTT estimation by LUBT might be reduced. Frequent, small episodes of caecoileal reflux also have been reported in the horse (Roger et al. 1990), causing dips in terminal ileal pH. Finally, the specific effects of antibiotic administration on LUBT have yet to be established in any species (Geypens 2000). However, based on the data produced in this study, and following the specific induction protocol, it is concluded that the LUBT is likely to provide a specific and quantitative indirect measure of OCTT in horses. As the LUBT is noninvasive and easy to perform it is anticipated that this test should prove of value in both clinical and research settings.

Authors' declaration of interests

No conflicts of interest have been declared.

Source of funding

David Sutton was funded by a Horserace Betting Levy Board Research Scholarship. The in vitro fermentations were funded by the Home of Rest for Horses grant G701.

Acknowledgements

Drs Christine Edwards and Douglas Morrison are acknowledged for assistance with development of fermentation techniques.

Manufacturers' addresses

1 Trudell Medical International, London, Ontario.

2 Quintron Instrument Company, Milwaukee, Wisconsin, USA.

3 GMI Medical Ltd, Renfrew, UK.

4 Bell College of Technology, Hamilton, South Lanarkshire, UK.

5 Labco Ltd, High Wycombe, Buckinghamshire, UK.

6 PDZ Europa, Northwich, Cheshire, UK.

7 Microsoft Corporation, Redmond, Washington, USA.

8 Oxoid, Basingstoke, Hampshire, UK.

9 Minitab Inc., State College, Pennsylvania, USA.

Ancillary