INTRODUCTION

Top Abstract Introduction Methods Results Discussion References

TONOMETRIC MEASUREMENT of gastric PCO₂, and the subsequent calculation of intracellular pH (pH_i) (6), is an attractive clinical measurement for several reasons. A greatly elevated tonometer PCO₂ measurement, and, by inference, tissue PCO₂, suggests the presence of tissue anaerobic metabolism (2, 25), resulting in proton formation (12) and titration of bicarbonate-based buffer systems to produce CO₂. A less severely elevated gastric tonometer PCO₂ suggests a relative imbalace between local tissue perfusion and metabolic demands (25). Thus, clinically, an increase in tonometer PCO₂ is useful in detecting inadequate tissue perfusion even if it has not progressed to frank anaerobic metabolism (28). Gastric tonometer PCO₂ measurements have been suggested to be particularly useful in critically ill patients because the gut may become ischemic before other organ systems in hypoperfusion states (7, 20) so that an elevated gastric tonometer PCO₂ may be an early and sensitive signal of inadequate perfusion (2, 4). These potentially important measurements have generated a great deal of enthusiasm because they are easily measured clinically by using a device no more invasive than the ubiquitous nasogastric tube in critically ill patients (8). Several important studies have shown potential for significant clinical benefit by using gastric tonometry (4, 8, 9, 15, 16, 18).

Despite the promise of clinical utility, the routine use of gastric tonometers in management of critically ill patients has not been fully embraced in very many critical care units around the world. This is due, in part, to the observation that gastric tonometer PCO₂ measurements are noisy, so, although average values in groups of patients may be predictive, they are much less valuable in guiding care in individual patients (23). A number of problems have been identified. Reflux of alkaline duodenal contents into the acidic stomach, or back diffusion of gastric mucosal bicarbonate, produces a significant amount of CO₂ unrelated to tissue hypoperfusion (11). Therefore, H₂-antagonist treatment has been used when gastric tonometric measurements are to be made (11). Gastric mucosal acid-base balance is complex so that the relationship between pH_i and tissue perfusion may not be straightforward (2, 23). In addition, the anatomic blood supply to the stomach is more extensive than that to the small bowel so that the stomach may not be particularly reflective of gut ischemia. Measurement of tonometer-solution PCO₂ is another potential source of error (27). Strategies to improve tonometric PCO₂ measurement could potentially contribute to the clinical utility of these measurements in critically ill patients.

Because the small bowel is not encumbered with the same problems as the stomach, we postulated that tonometric PCO₂ measurements in the small bowel may be superior to those in the stomach. Accordingly, in anesthetized pigs subject to progressive hemorrhage we tested the hypothesis that small bowel tonometer PCO₂ measurement is less noisy than gastric tonometer PCO₂ measurement and is superior in identifying the onset of gut ischemia.

	METHODS

Top Abstract Introduction Methods Results Discussion References

This study was approved by the Animal Care Committee of the University of British Columbia, and the animals were handled in accordance with Canadian and National Institutes of Health guidelines.

Instrumentation. Ten pigs (29 ± 6 kg) were fasted for 12 h and treated with ranitidine (150 mg iv) 2 h before experimentation. The pigs were then sedated with ketamine (500 mg im), followed by thiopental sodium (125-250 mg iv). Anesthesia was maintained during the experiment with halothane (0.5-1.5%) inhalation and ketamine infusion (0.75 µg · kg¹ · min¹ iv). The pigs were ventilated via a tracheotomy with a tidal volume of 12 ml/kg at a rate adjusted to maintain arterial PCO₂ from 37 to 42 Torr. The inspiratory O₂ fraction was set at 0.95 during instrumentation and at 0.21 after instrumentation for the duration of the experiment.

Protocol. After instrumentation and stabilization, oxygen delivery was decreased by hemorrhage of between 2 and 4 ml of blood/min by using a constant-withdrawal pump. Data were measured at the end of the stabilization period (baseline) and then at 30-min intervals during progressive hemorrhage. At each data set, measurements were made of hemodynamic parameters, gut oxygen consumption and delivery, and gastric and small bowel tonometer PCO₂.

Measurements. All expired gas from the sealed ventilator circuit was diverted through a previously validated (22) metabolic monitor (Deltatrac MBM-1000, Datex) to measure whole body oxygen consumption. Values were determined at each measurement as an average over a 3-min interval. Arterial, mixed venous, and gut venous hemoglobin and oxygen saturations were measured by using an IL 482 CO-oximeter (Instrumentation Laboratories, Lexington, MA). Corresponding blood-gas measurements and tonometer PCO₂ measurements corrected for body temperature were made by using an ABL 30 blood-gas machine (Radiometer, Copenhagen, Denmark). To measure tonometer PCO₂, the first 5 ml of saline were discarded and 2 ml saline from the tonometer catheters were then sampled anaerobically. PCO₂ was measured within 60 s of sampling in all cases.

Analysis. Blood oxygen content was calculated as hemoglobin (g/dl) × 1.39 × blood oxygen saturation (%) + 0.003 × blood oxygen tension (Torr). Cardiac output was calculated, by using the Fick principle, as whole body oxygen consumption divided by the difference between arterial and mixed venous oxygen contents. Whole body oxygen delivery was calculated as cardiac output multiplied by arterial oxygen content. Gut oxygen delivery (ml O₂ · kg¹ · min¹) was calculated as superior mesenteric vein blood flow multiplied by arterial oxygen content. Gut oxygen consumption (ml O₂ · kg¹ · min¹) was calculated as superior mesenteric vein blood flow multiplied by the difference between arterial and gut venous oxygen contents.

	RESULTS

Top Abstract Introduction Methods Results Discussion References

Progressive hemorrhage resulted in biphasic relationships between whole body oxygen delivery and consumption (Fig. 1A), with an average r² of 0.92 ± 0.14 and a critical oxygen delivery of 9.5 ± 2.3 ml O₂ · kg¹ · min¹. Similarly, biphasic gut oxygen delivery-consumption relationships (Fig. 1B) were observed with an average r² of 0.92 ± 0.08, identifying a gut critical oxygen delivery of 20.3 ± 5.7 ml O₂ · kg¹ · min¹. In 8 of 10 animals, the onset of gut anaerobic metabolism occurred before whole body anaerobic metabolism, on average for the group, at a whole body oxygen delivery of 12.5 ± 5.2 ml O₂ · kg¹ · min¹ (P < 0.05 compared with whole body critical oxygen delivery of 9.5 ± 2.2 ml O₂ · kg¹ · min¹).

View larger version (11K): [in this window] [in a new window]   Fig. 1.   O2 delivery (O2)-O2 consumption (O2) data for whole body (A) and for gut (B) during progressive hemorrhage in 1 animal. Two lines that best fit this biphasic relationship are shown together with correlation coefficient, r2, of data points to these 2 lines. The r2 value for this experiment was median for group, so that one-half of experiments had better fits to biphasic O2-O2 relationships and one-half had worse fits.

In the 10 hemorrhagic animals, the onset of anaerobic metabolism occurred in the sixth to tenth half-hourly measurement set (at the 2.5- to 4.5-h-after-baseline set at time 0). There was no significant difference in PCO₂ over time for the first four measurements in each animal. Thus measurable increases in PCO₂ due to low flow had not yet occurred. Therefore, we took the first four measurements during aerobic metabolism to define, in each animal, the mean and 95% confidence intervals for PCO₂ in that animal. In all 10 animals, small bowel PCO₂ increased outside of these 95% confidence intervals at 7.3 ± 2.5 measurement sets (on average at 3.15 h after baseline), whereas the onset of anaerobic metabolism (critical oxygen delivery point) occurred at 7.4 ± 1.6 measurement sets (on average at 3.2 h after baseline). Thus small bowel PCO₂ exceeded the 95% confidence interval in each animal at almost the same time as the onset of anerobic metabolism. Variability in the first four measurements during aerobic metabolism was much greater for gastric PCO₂ (width of 95% confidence interval 45 ± 26 Torr) than for small bowel PCO₂ (12 ± 12 Torr, P < 0.005). As a result, gastric PCO₂ in five animals did not exceed the 95% confidence intervals at any time. In the other five animals gastric PCO₂ increased outside the 95% confidence intervals at 8.2 ± 3.0 measurement sets, whereas the onset of anaerobic metabolism occurred at 7.0 ± 1.6 measurement sets.

An alternative approach is to use the 95% confidence intervals from the five control animals to define the PCO₂ threshold, recognizing that these data include additional interanimal variability. Repeated measurements in the five control experiments during aerobic metabolism demonstrated a mean gastric PCO₂ of 33 ± 39 Torr and a mean small bowel PCO₂ of 9 ± 12 Torr. Interanimal variability in gastric PCO₂ accounted for >99% of total variability in these data (interanimal mean square divided by total mean square >99%). Thus there were marked differences in PCO₂ among different control animals. In contrast, intra-animal varibility was small (intra-animal mean square divided by total mean square <1%), which indicates that PCO₂ measured repeatedly within one animal was quite stable. PCO₂ in individual animals exceeded the group threshold value (upper 95% confidence interval value was 33 Torr) in 8 of 10 animals at 7.1 ± 3.2 sets. In 2 of 10 animals, this threshold value derived from control animals was not exceeded. In contrast, gastric PCO₂ exceeded the group threshold value (upper 95% confidence interval value was 110 Torr) in only 5 of 10 animals on average at 4.6 ± 3.5 sets, essentially unrelated to the true onset of anaerobic metabolism. In the other 5 of 10 animals, this threshold value derived from control animals was not exceeded.

Biphasic relationships resulted from plots of gut oxygen delivery vs. gastric pH_i, small bowel pH_i, gastric PCO₂, and small bowel PCO₂ (Fig. 2). However, the correlation of small bowel pH_i data points with dual regression lines was significantly greater than the correlation of gastric pH_i data points (P < 0.01) (Fig. 3). Also striking were the differences in gut oxygen delivery-PCO₂ relationships, where the correlation with dual-line regression for the small bowel PCO₂ was greater than the correlation for gastric tonometer PCO₂ measurements (Figs. 2 and 4). These data indicate that there was significantly less scatter in small bowel tonometer measurements compared with gastric tonometer measurements around the biphasic relationships.

View larger version (11K): [in this window] [in a new window]   Fig. 2.   Gastric tonometer minus arterial PCO2 difference (gastric PCO2; A) or small bowel tonometer minus arterial PCO2 difference (small bowel PCO2; B) are plotted against gut oxygen delivery (gut O2). Plots are from individual experiments that had median r2 value for group. That is, one-half of gastric PCO2-gut O2 relationships were better than 1 shown and one-half were worse. Similarly, one-half of small bowel PCO2-gut O2 relationships were better than 1 shown and one-half were worse. As gut O2 decreases, small bowel PCO2 increases in a biphasic manner. In contrast, gastric PCO2 is not related to gut O2 in a readily discernable way.

View larger version (10K): [in this window] [in a new window]   Fig. 3.   r2 For data points around biphasic gut O2-intracellular pH (pHi) relationships for all experiments ().  and Error bars: means and SE, respectively. There was less scatter of data points around biphasic linear relationships (greater r2, * P < 0.05) for small bowel tonometer-derived pHi than for gastric tonometer-derived pHi.

View larger version (10K): [in this window] [in a new window]   Fig. 4.   r2 For data points around biphasic gut O2-PCO2 relationships is shown for all experiments ().  and Error bars: means and SE, respectively. There was less scatter of data points around biphasic linear relationships (greater r2, * P < 0.05) for small bowel tonometer-derived PCO2 than for gastric tonometer-derived PCO2.

The critical gastric PCO₂ was high and variable (62.9 ± 39.6 Torr), resulting in a low critical gastric pH_i (7.04 ± 0.14). In contrast, the critical small bowel PCO₂ was smaller (17.0 ± 15.1, P < 0.01) and less variable (SD 39.6 vs. 15.1, significantly different, P < 0.01) (Fig. 5). As a result, the critical small bowel pH_i (7.29 ± 0.15) was greater than the critical gastric pH_i (7.04 ± 0.14, P < 0.01).

View larger version (14K): [in this window] [in a new window]   Fig. 5.   Critical PCO2 at onset of anaerobic metabolism, determined from gut O2-O2 relationships, for gastric tonometer and the small bowel tonometer. Data from individual experiments are connected with a dashed line. Gastric tonometer critical PCO2 is high [ (mean)] and variable [SD (error bars)] compared with small bowel tonometer critical PCO2 (* P < 0.01 for both mean and SD by using t-tests).

Because tonometer and venous PCO₂ measurements should reflect tissue PCO₂, we compared both gastric tonometer PCO₂ and small bowel tonometer PCO₂ measurements with superior mesenteric vein PCO₂ from all measurement sets in all animals. We used this as an independent assessment of whether gastric or small bowel tonometer measurements were superior in assessing tissue PCO₂ in the regional circulation of the gut. Small bowel tonometer PCO₂ was closely correlated with superior mesenteric vein PCO₂ (r² = 0.81, P < 0.001), whereas gastric tonometer PCO₂ was not (r² = 0.13, P = not significant) (Fig. 6). By using a Bland-Altman analysis (1), small bowel tonometer PCO₂ was significantly more closely related to superior mesenteric vein PCO₂ both in mean (P < 0.05) and in the two SD intervals (P < 0.05) (Fig. 7).

View larger version (14K): [in this window] [in a new window]   Fig. 6.   Plot of gastric tonometer PCO2 vs. superior mesenteric vein (SMV) PCO2 (A) fails to show a significant relationship between these 2 variables. In contrast, a plot of small bowel tonometer PCO2 vs. SMV PCO2 (B) shows a significant correlation (P < 0.001).

View larger version (16K): [in this window] [in a new window]   Fig. 7.   Bland-Altman plots of difference between gastric tonometer PCO2 vs. SMV PCO2 (A) and small bowel tonometer PCO2 vs. SMV PCO2 (B) against average of tonometer PCO2. Both tonometer and SMV PCO2 are indirect clinical measurements that reflect tissue PCO2. There is poor agreement between gastric tonometer PCO2 and SMV PCO2. Small bowel tonometer PCO2 more closely reflects SMV PCO2.

To determine whether gastric or small bowel tonometry was best at identifying the onset of gut ischemia, as determined from gut oxygen delivery-consumption relationships, we compared the difference in gut critical oxygen delivery points. There was a small difference in the gut critical oxygen delivery determined from gut oxygen delivery-consumption relationships vs. gut oxygen delivery-small bowel PCO₂ relationships (mean difference in estimate of critical gut oxygen delivery 4.4 ± 3.5 ml O₂ · kg¹ · min¹). However, there was significantly greater disparity in estimates of the gut critical oxygen delivery from gut oxygen delivery-consumption relationships and gut oxygen delivery-gastric PCO₂ relationships (mean difference 9.3 ± 5.8 ml O₂ · kg¹ · min¹, P < 0.05).

	DISCUSSION

Top Abstract Introduction Methods Results Discussion References

Our key findings are that, in this large-animal model, gastric tonometer PCO₂ and pH_i measurements were noisy compared with small bowel tonometer PCO₂ and pH_i measurements. This resulted in excessively high and variable estimates of gastric PCO₂ and pH_i at the onset of gut ischemia. The 95% confidence intervals for gastric PCO₂ were wide so that values outside this range did not occur consistently, even though hemorrhagic shock progressed to death. In contrast, small bowel PCO₂ exceeded the 95% confidence intervals in all cases at approximately the onset of anaerobic metabolism. Gastric tonometer PCO₂ was not closely correlated with superior mesenteric vein PCO₂ measurements, and critical oxygen delivery points derived from these measurements were not closely related to the critical oxygen delivery points determined from biphasic gut oxygen delivery-consumption relationships. Small bowel tonometric PCO₂ measurement (and the derived pH_i calculation) was superior in all of these respects. This suggests the possibility that small bowel tonometric PCO₂ measurement may be clinically useful.

Part of the reason why measures of gut perfusion, such as tonometer PCO₂, may be useful is that the gut may be an organ that is particularly sensitive to systemic hypoperfusion (7, 20). For example, Nelson and colleagues (20) found in dogs that the onset of oxygen supply dependency occurred earlier in the gut than in the whole body. Our results confirm these previous observations, supporting the notion that evidence of gut hypoperfusion or ischemia may be an early indicator of inadequate systemic perfusion. In addition, gut ischemia may be important in the pathogenesis of a systemic inflammatory response (5, 7). Intestinal permeability increases after gut ischemia (3), resulting in leakage of bacteria and bacterial products into the portal circulation (10). This results in activation of hepatic Kupffer cells (14) and circulating leukocytes (13, 26). A number of investigators have suggested that this initiates a subsequent systemic inflammatory response, which may result in distal organ damage and dysfunction (13, 26). Thus there is a strong physiological rationale for monitoring gut perfusion in critically ill patients.

Indeed, a number of studies demonstrate that tonometry may be a useful clinical tool. Gutierrez and colleagues (8) found that, in critically ill patients with an initial gastric tonometer pH_i > 7.35, if tonometer-derived pH_i subsequently fell below 7.35, then resuscitation with fluids and dobutamine significantly improved survival. Mohsenifar and colleagues (18) have demonstrated in critically ill patients that a low pH_i, calculated from gastric juice PCO₂, predicts failure of spontaneous ventilation at the time of extubation. Studies by Marik (15) and by Maynard and colleagues (16) suggest that gastric tonometer-derived pH_i is a more robust predictor of outcome in critically ill patients than hemodynamic measurements and other common predictors. Similarly, Mohsenifar and colleagues (17) found that pH_i was a better predictor of outcome than all other presently used parameters in hemodynamically stable, mechanically ventilated patients. A pH_i < 7.25 had a sensitivity of 86% and a specificity of 83% in predicting mortality (17). Thus tonometry has the potential to play an important role in the intensive care unit, adding to other hemodynamic and physiological data in assessing adequacy of perfusion and enhancing prognostic measures, which include severity-of-illness scoring systems such as APACHE II. However, there are problems with gastric tonometer measurements that may have contributed to its limited use in many intensive care units as a clinical tool in individual critically ill patients.

Gastric tonometer PCO₂ measurement may be noisy and not accurately reflect gut intramural PCO₂ for several reasons. First, the acid-generating gastric mucosa may make the stomach a poor site for tonometry. Accurate gastric tonometry may require pretreatment with H₂ antagonists or a proton-pump inhibitor to prevent significant CO₂ generation when gastric mucosal bicarbonate diffuses into the lumen or when duodenal contents reflux through the pyloris into the acidic stomach (11). CO₂ production from titrating acid and base in this way yields erroneous information because it is unrelated to tissue perfusion. Despite our administration of a high dose of an H₂ antagonist in this study, it is conceivable that this effect may have contributed somewhat to our results. Furthermore, gastric intramural PCO₂ may not be an accurate reflection of gut intramural PCO₂ because the gastric arterial blood supply is more extensive than the blood supply of the gut, which is predominantly via the superior mesenteric artery. Thus gut ischemia does not necessarily indicate gastric ischemia. The lack of correlation between gastric PCO₂ and superior mesenteric vein PCO₂ and the close correlation between small bowel PCO₂ and superior mesenteric vein PCO₂ support the idea that it is best to consider the stomach and small bowel to be two different regional circulations. Problems common to all tonometry include that tonometer-measured PCO₂ requires a prolonged equilibration time so that a correction factor must be introduced to account for incomplete CO₂ equilibration (2). In addition, Takala and colleagues (27) have pointed out that measurement of PCO₂ in tonometer saline can introduce marked errors (27). Improved measurements with buffered solutions suggest that tonometry must be carefully done to avoid diffusion of PCO₂ from the tonometer solution into the atmosphere.

We found that small bowel tonometer PCO₂ measurements appeared to be more accurate and less variable than gastric tonometer PCO₂ measurement. Although both gastric and small bowel tonometry detect ischemia (19), compared with gastric tonometry, small bowel tonometry appears to reduce the variability of individual measurement points around biphasic oxygen delivery-consumption relationships and appears to reduce the variability of the critical PCO₂ and pH_i. Small bowel tonometry led to a closer estimate of the onset of gut ischemia, as determined from gut oxygen delivery-consumption relationships. The observation that small bowel tonometer PCO₂ was closely related to superior mesenteric vein PCO₂, whereas gastric tonometer PCO₂ was not, suggests that small bowel tonometer PCO₂ is the more accurate measurement. By improving the accuracy of tonometer measurements, small bowel tonometry may conceivably improve on the clinical utility of present gastric tonometry. Obviously, surgical placement of a small bowel tonometer, as in this animal experiment, is not clinically realistic. However, the increasing use of duodenal and jejunal feeding tubes in critically ill patients suggests a feasible clinical route.

Our data do not resolve the partly semantic issue of whether PCO₂ measurement or pH_i calculation should be used. pH_i is fundamentally more important than extracellular pH in altering cellular function, including cardiac muscle contraction, synthetic function, and work by cellular membrane ion pumps (29). Fiddian-Green (6) popularized the idea that gastric tonometer measurements of pH_i may reflect intramural pH and, by inference, pH_i. Therefore, there was theoretical reason to suspect that tonometer pH_i may be of fundamental importance in assessing tissue acidosis and would be an independent contributory parameter in evaluating hypoperfusion states in critically ill patients (6). However, the calculation of pH_i involves a number of assumptions that are generally not fulfilled (2, 23). For example, the intracellular buffer systems are not predominantly bicarbonate based, and arterial, rather than tissue, bicarbonate measurements are used. Thus these two terms (pK_a and [HCO₃]) in the three-term Henderson-Hasselbalch equation suffer significant limitations. Schlichtig and Bowles (25) have put tonometer PCO₂ measurement on a firm theoretical foundation. They show that decreased flow states will increase tissue, and thus tonometer, PCO₂ and that there is a rapid rise in tissue PCO₂ at the onset of anaerobic metabolism. There is no need to calculate pH_i because PCO₂ measurements can be used directly (2).

On the basis of this, we think that PCO₂ is the appropriate measurement to make. The counterargument could be made that pH_i is superior because biphasic gut oxygen delivery-pH_i relationships had better correlation coefficients than did gut oxygen delivery-PCO₂ relationships. We do not support this line of reasoning because pH_i measurements are less variable due to the logarithmic transformation (Henderson-Hasselbalch equation), which does not improve the value of the measurement in a fundamental way. Furthermore, the inclusion of arterial bicarbonate in pH_i calculation artificially enhances the biphasic relationship because arterial bicarbonate falls as oxygen delivery decreases. This may be considered useful in this controlled setting, starting at a normal acid-base status. However, inclusion of arterial bicarbonate is potentially a detrimental feature in the clinical use of tonometry. For example, patients with primary acid-base disturbances will have erroneously high or low calculated pH_i independent of the presence or absence of true gut ischemia, leading to inappropriate therapy for the critically ill patient (23). Furthermore, during rapid onset of gut ischemia, PCO₂ will increase, whereas arterial bicarbonate will not change significantly for a period of time. The difference in pH_i calculation between rapid and slow onset of ischemia once again could lead to clinical confusion and inappropriate management. Therefore, because the primary measurement is PCO₂ of the tonometer solution, we agree with a number of other investigators (2, 23, 25) that tonometer PCO₂, and, more importantly, the difference in PCO₂ between the tonometer solution and arterial blood (PCO₂) are the most useful clinical variables.

Surgical instumentation and measurement techniques may have influenced our results. One possibility is that, in some of the anesthetized, highly instrumented animals in this study, the gastric mucosa was much more susceptible to hemorrhage-induced ischemia than was the gut. Hence the high and variable gastric PCO₂. Although this is possible, it is our clinical experience that in patients having septic shock we reasonably frequently observe small bowel ischemia to the point of infarction, whereas we have not yet observed gastric ischemia to the point of infarction. In addition, animal studies suggesting that the gut may become oxygen supply dependent before the whole body have focused on the gut rather than on the stomach (7, 20). The variability of our gastric tonometer PCO₂ measurements in control animals was high but similar to other reported values (e.g., 57.2 ± 35.7 Torr in surviving surgical intensive care patients; Ref. 9). Other studies in healthy humans report less variability (48 ± 10 Torr; Ref. 21). It is important to note that >99% of the variability we observed in PCO₂ was interanimal variability (individual animal PCO₂ differed substantially from other animals), whereas repeated PCO₂ measurements within an individual animal were very consistent. Thus the variability in PCO₂ among different anesthetized, ventilated, and surgically instrumented pigs was greater than the variability in PCO₂ among different healthy humans (21). The tonometer measurements themselves show very little variability when repeated over substantial time intervals in individual animals, suggesting that the tonometer measurements were likely accurate. The concordance of superior mesenteric vein PCO₂ and tonometer PCO₂ provides independent corroboration of the likely accuracy of the tonometer measurements. The important comparison in our study is between gastric PCO₂, which demonstrated high variability, and small bowel PCO₂, which demonstrated lower variability. Gastric and small bowel samples were handled in exactly the same manner so that our conclusions likely are not dependent on measurement technique.

In summary, these data suggest that small bowel tonometer PCO₂ measurement is superior to gastric tonometer PCO₂ measurement in reflecting gut CO₂ accumulation, hypoperfusion, and the onset of anaerobic metabolism. This suggests the possibility of enhancing the present clinical utility of gastric tonometer PCO₂ and pH_i measurements by small bowel placement of an appropriate tonometer catheter.