Materials and Methods

The study was approved by the Animal Care Protocol Review Committee of the University of Guelph, in accordance with guidelines of the Canadian Council on Animal Care. Six adult Standardbred geldings (mean ± SD weight, 437 ± 37 kg) with no known history of gastrointestinal tract disease were used in the study and were determined to be clinically normal on the basis of results of a physical examination and a CBC.

The study was conducted according to a repeated-measures-within-repeated-measures design in a randomized crossover study wherein horses were assigned to treatment (NGT placed) and control groups (no tube), and serum acetaminophen concentration was measured 0, 12, 30, 48, and 72 hours after placement of the NGT and administration of the acetaminophen. Horses were used in the crossover part of the experiment after a washout period of at least 4 weeks. All horses were weighed daily. Food and water were withheld from horses 8 hours prior to acetaminophen administration and for the duration of the experiment. During the period of feed and water withholding, fluid support was provided via IV administration of lactated Ringer's solution supplemented with potassium chloride (final potassium concentration, 20 mEq/L) at a rate of 60 mL/kg/d. Hydration and metabolic status were monitored by daily Hct, venous blood gas, and serum total protein concentration determinations. The control and treatment portions of the experiment each lasted 72 hours. For horses in the treatment group, an indwelling NGT with an outer diameter of 1.6 cm was maintained for the duration of each part of the experiment. Horses in the control group were nasogastrically intubated only to administer acetaminophen at the predetermined times.

Acetaminophen^a (dose, 20 mg/kg) was diluted in 1 L of distilled water and administered via NGT 0, 12, 30, 48, and 72 hours after nasogastric intubation to evaluate the liquid-phase GER. The entire contents of the tube were deposited in the stomach by administering a volume of air equivalent to the NGT dead space after the acetaminophen was administered. Similar to methods used in clinical practice, in the treatment horses, the exposed end of the tube was capped with a rubber sleeve to act as a 1-way valve, and the tube was secured to the horse's halter. A muzzle was placed on the treatment horses to prevent horses from dislodging the NGT and on control horses to prevent ingestion of the shavings used as bedding material. A 10-mL blood sample was collected from an IV catheter in a jugular vein immediately before acetaminophen administration for baseline measurements. A 10-mL blood sample was collected from a jugular vein 20, 40, 60, 80, 100, 120, and 180 minutes after each acetaminophen administration. Serum samples were extracted via centrifugation and stored at −80°C until analysis.

Acetaminophen concentrations in each serum sample were determined colorimetrically by use of a commercial kit^b and a microspectrophotometer^c with a filter for 620 nm calibrated in accordance with manufacturer's recommendations and with water used as the blank. Concentrations were given as the mean of 3 replicates. The inter-and intra-assay coefficients of variation were < 5% and < 2.5%, respectively.

Samples were diluted, and a calibration analysis was performed before every acetaminophen determination. For this method, the correlation of the spectrophotometric analysis with high performance liquid chromatography was 0.97. The lower limit of detection for this assay method was 6 μg/mL. Acetaminophen concentrations were plotted for treatment and control groups. A curve was fitted to the data by use of a software program.^d From that curve, C_max and T_max were estimated and used in a 1-compartment model as the metric rates for evaluation of GER.

Endoscopic examination of the nasal passages, nasopharynx, larynx, esophagus, and stomach was performed in control and treatment horses before and at the end of the 72 hours prior to reintroducing the horses to feed. The endoscopic images were reviewed with the evaluator unaware of treatment group and scored according to the number and severity of gastric and esophageal ulcers and pharyngeal tissue damage. If ulceration was observed, the location was determined and severity was scored by means of a previously described scoring system.¹⁰ Briefly, lesions in the gastric mucosa were scored as 0, intact mucosal epithelium (with possible reddening, with or without hyperkeratosis); 1, small single or multifocal lesions; 2, large single or multifocal lesions; or 3, extensive (often coalescing) lesions with areas of deep ulceration. At the end of the experiment, horses were reintroduced to feed during a period of 72 hours and turned out to pasture.

Statistical analysis—Because each horse had repeated measurements of serum acetaminophen concentration with-in each hour of interest, a summary approach was used to address the repeated measures by fitting a curve to serum acetaminophen concentration and obtaining meaningful biological estimates of gastric emptying (ie, T_max and C_max). This was repeated in succession at 0, 12, 30, 48, and 72 hours for each horse under each condition. Curve fitting was done by use of a statistical software program.^d A curve was fitted to the data according to a 1-compartment 4-parameter log normal model, and C_max and T_max were used in a 1-compartment model as the metric rates to evaluate GER. After those variables were estimated for each of the hours of interest, they were analyzed via linear mixed-model ANOVA to examine the effects of group, time, and interaction for the calculated variables of T_max and C_max. The Akaike information criterion from the ANOVA indicated that compound symmetry best fit the error structure to account for making repeated measurements over the 0-, 12-, 30-, 48-, and 72-hour time points. Results of the Shapiro-Wilk test confirmed that data for both variables were normally distributed. If the overall F-test results from the ANOVA were significant for the main effect of time or the treatment-time interaction, a Tukey test was used to make paired comparisons. No adjustment was needed for the main effect of group because there were only 2 levels. Values of P< 0.05 were considered significant. The gastric ulceration scores of the endoscopic examination were analyzed by use of the Mann-Whitney test.

Results

The fitted curve equation corresponded to a 1-compartment 4-parameter log normal model equation, from which C_max and T_max were estimated (Figure 1). As a result of technical errors related to sample handling, certain time points were missed and we were unable to fit a curve for 1 horse in the treatment group for the 12- and 72-hour time points and in 1 horse in the treatment group at the 48-hour time point. Results of the Shapiro-Wilk test indicated that the C_max and T_max data were normally distributed and met the assumptions of a repeated-measures ANOVA. Log transformation of the data improved normality on T_max and marginally on C_max (Shapiro-Wilk values, 0.176 and 0.231, respectively). Analysis was performed on untransformed data because the data were normally distributed. Examination of the residuals in the univariate analysis revealed large variations in values for C_max and T_max in both groups, although variation was larger in the treatment group, a finding that was corroborated by the SDs obtained in the means procedure. Reported means and SDs represent values obtained from ANOVA estimates, which were calculated on the basis of a pooled SE (Table 1). When the effects of horse, time, and the group-horse interaction were taken into account, C_max was significantly (P = 0.014) higher in the control group (38.11 μg/mL) than in the treatment group (29.09 μg/mL), whereas T_max was significantly (P = 0.003) shorter in the control group (22.79 minutes) than in the treatment group (35.95 minutes). This finding would support the conclusion that an indwelling NGT had a delaying effect on GER over the 72-hour period. However, there was a main effect of time for T_max, indicating that values for hour 30 were significantly (P = 0.038) different from those for hour 0, a finding that suggested that for both groups, the GER had increased by 30 hours. No group-time interactions were detected.

Concentration-versus-time curve fitted from measured serum acetaminophen concentrations at the 30-hour time point in a horse in which the NGT was removed after each administration of the drug (A) and in the same horse at the same time point when an indwelling NGT was maintained for 72 hours (B). For panel A, r² = 0.983. For panel B, r² = 0.992. *1 μmol/L = 0.151 μg/mL.

Citation: American Journal of Veterinary Research 67, 7; 10.2460/ajvr.67.7.1100

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Concentration-versus-time curve fitted from measured serum acetaminophen concentrations at the 30-hour time point in a horse in which the NGT was removed after each administration of the drug (A) and in the same horse at the same time point when an indwelling NGT was maintained for 72 hours (B). For panel A, r² = 0.983. For panel B, r² = 0.992. *1 μmol/L = 0.151 μg/mL.

Citation: American Journal of Veterinary Research 67, 7; 10.2460/ajvr.67.7.1100

• Download Figure
• Download figure as PowerPoint slide

Concentration-versus-time curve fitted from measured serum acetaminophen concentrations at the 30-hour time point in a horse in which the NGT was removed after each administration of the drug (A) and in the same horse at the same time point when an indwelling NGT was maintained for 72 hours (B). For panel A, r² = 0.983. For panel B, r² = 0.992. *1 μmol/L = 0.151 μg/mL.

Citation: American Journal of Veterinary Research 67, 7; 10.2460/ajvr.67.7.1100

• Download Figure
• Download figure as PowerPoint slide

Endoscopic evaluation of 3 horses in the control group was not performed. No differences were detected between treatment and control groups with regard to gastric ulceration (P = 0.34). The mean gastric ulcer scores for control and treatment groups were 0.3 and 1, respectively. All horses in the treatment group had abnormalities in the nasal passages and pharyngeal area, ranging from hyperemia to ulceration and bleeding. No esophageal abnormalities were seen in any horse.

Discussion

Analysis of results indicated that nasogastric intubation for 72 hours was associated with a mean delay in GER of liquid of 13 minutes (range, 14.1 to 12.2 minutes). The time effect observed in T_max at 30 hours in both groups indicated that the relationship between gastric emptying and the presence of an indwelling NGT was not uniform. It appeared that although values for horses in the control group had minimal variation, the indwelling NGT induced large variations in C_max and T_max, similar to findings in another study.⁴ It is possible that the tube elicits gastric secretion, an action that would have an unknown effect on acetaminophen absorption. It is also possible that other factors, such as contact between the tube and gastric wall, stress, or the withholding of feed, had inhibitory influences on the GER. It was considered that the crossover design of the study would minimize the effects of variation among horses.

Differences in study design prevent direct comparison of our results with those that have previously been reported.⁴ Whereas earlier investigators⁴ described a tendency toward faster GERs in an 18-hour period, the opposite effect was observed in the study reported here when mean values over a 72-hour period were considered. The 18-hour time point was not included as a measurement time point in our study, and the closest time points were 12 and 30 hours after NGT placement. In the present study, a time effect (characterized by a faster emptying rate) was only observed at 30 hours in either group. It is possible that the initial effects of intubation or placement and maintenance of an indwelling NGT were associated with abnormal patterns of gastric emptying, which could have accounted for the different results between the 2 studies and over time. Further studies are needed that include more frequent sampling and a larger number of horses so that the substantial variation that exists among individual horses can be accounted for. The power estimations for this study were 71.01% for C_max and 27.86% for T_max. The issue of insufficient power is less relevant in studies in which differences between groups are observed, as was true in the study reported here.

Errors related to the acetaminophen measurement method for determining GER may result from substantial between-subject variation in fraction of the dose that reaches the systemic circulation after first-pass metabolism through the liver, apparent volume of distribution, and rate constant of elimination.¹¹ Because between-subject variation in drug handling is usually greater than intraindividual variation, a crossover study design to offset the effects of interindividual variation was considered optimal for our study. Because of the logistics of our study design, sampling could not commence before 20 minutes preceding acetaminophen administration. A study⁹ in clinically normal horses has revealed that the range for T_max is from 27.9 to 87.3 minutes, with a median of 47.7 minutes. Therefore, sampling at 20 minutes after acetaminophen administration should have yielded reliable T_max estimations under these circumstances. Earlier and more frequent sampling would be useful to determine the initial disposition of acetaminophen under the conditions of the study. On the basis of values for T_max, the overall observed delay was approximately 13 minutes and may not be clinically important. It is possible, however, that this effect would be exacerbated in sick horses.

The study was designed to mimic, as accurately as possible, conditions that are encountered clinically. Horses are withheld from food and water and supported with fluids administered IV during episodes of ileus, which often last for 48 to 72 hours. Therefore, we elected to administer fluids IV and maintain the NGT for 72 hours. Feed withholding is known to alter gastrointestinal tract motility patterns,¹² but it was considered that the crossover study design and implementation of equal feed witholding times would control those effects.

The accuracy of the acetaminophen method of determining GER depends greatly on the absorptive index used. Despite the fact that the pharmacokinetics of acetaminophen follow a multicompartmental model, analysis by use of a 1-compartment model has been validated. In humans, the reliability of the acetaminophen method has been debated¹¹ and no consensus exists as to the optimal metric rate to use. In the study reported here, different rate metrics to evaluate GER were considered. In 1 study,⁸ values for C_max and AUC at 240 minutes were not well correlated with the criterion-referenced standard of scintigraphic measurement, whereas in another study,⁹ values for T_max were in close agreement with scintigraphically obtained values for gastric emptying in horses with experimentally induced delay in gastric emptying and in healthy horses. It has also been reported¹¹ that determination of AUC at a given time point does not offer any advantage over determination of C_max; however, it is unknown whether the time up to which AUC is integrated has a different effect on the final result. Use of T_max for calculation of GER has been validated in horses.⁹ The value for C_max is a direct measurement of the rate and extent of absorption and can be determined from data without complicated calculations.¹¹ In 1 study,⁹ T_max yielded a better estimate of GER, compared with C_max, with scintigraphic methodology serving as the criterion-referenced standard. On the basis of results of previous studies and logistic considerations, a 1-compartment model and C_max and T_max as the metric rates to evaluate the effect of indwelling NGT in GER were used in our study.^9,11,13 Because T_max is a discrete value, its reliability is dependent on the frequency of blood sampling. Mean values for T_max were within a published⁹ range. Although the optimal frequency and interval for blood sampling have not been determined, our study design was selected on the basis of published information in horses.^8,9,13 The acetaminophen method of GER determination is only valid for liquids. Differences between the emptying rates of liquid and solids are substantial,¹⁴ but because liquid accumulates in the stomach during postoperative ileus, use of a liquid marker seems appropriate.

One limitation of the study reported here was that it was performed in clinically normal horses. The pathophysiologic basis of postoperative ileus involves the effects of anesthesia, handling of intestines, electrolyte imbalances, endotoxemia, intestinal distention and ischemia, peritonitis, resection, and anastomosis.² The influences of those factors on GER were not evaluated in the our study, and the clinical application of our findings must be considered in light of that fact. It is also possible that the effects of gastric decompression during a period of ileus outweigh, to some degree, the potential delaying effects of an indwelling NGT by preventing gastric distention and improving the horse's postoperative comfort. A future study incorporating an experimental model of ileus or a properly conducted prospective clinical trial would help to answer some of these clinically relevant questions.

Size of the nasogastric tube may also have an effect, but no conclusions can be made at this time regarding the effect of a smaller NGT on GERs. In a previous report,⁷ use of a smaller NGT was recommended to minimize the adverse effects associated with an indwelling NGT. It would have been useful to measure the volume and pH of gastric contents to quantify the extent to which an indwelling NGT affects either gastric secretion or duodenal reflux.

The endoscopic examinations did not reveal a gastric ulceration effect associated with an indwelling NGT. When observed, gastric ulceration was mostly limited to the area of the margo plicatus and lesser curvature. The method used to evaluate gastric ulceration scores was a low-power test. It is possible that with a larger experimental population, an effect would have been detected. Irritation in the nasal cavity and esophagus has been reported⁶ as a possible complication of maintaining an indwelling NGT in horses, and that finding was not surprising.

Clinicians should consider the potential for delayed gastric emptying when maintaining an indwelling NGT for an extended period during the postoperative period, in addition to other reported complications. Analysis of results of this study suggest that repeated nasogastric intubation may be better for managing horses with ileus than maintenance of a long-term indwelling NGT. However, the intubation procedure has also been associated with potentially adverse effects,^6,7 and particularly in fractious horses, repeated intubation may require administration of sedatives, which may also decrease GER.¹³ The use of serum acetaminophen concentration for determination of the GER in ill horses should be a simple, inexpensive, and reliable method of obtaining information pertaining to gastric emptying.