The determination of intramucosal PCO2 values (PiCO2) by nasogastric tonometry [1,2] is a minimally invasive procedure to assess impaired gastrointestinal perfusion  and to estimate the prognosis of critical ill patients . Tonometry relies on the fact that hollow visceral tissues are highly permeable to CO2 and lead to an equilibrium between the intramucosal PCO2 and the intraluminar PCO2[5,6]. An increase in intraluminar PCO2 can therefore indicate impaired perfusion and/or metabolic changes of the mucosa.
However, methodologic problems, such as the instability of CO2 in the tonometric solution, do not permit a reliable and reproducible estimation of intraluminal CO2 and thus PiCO2[7-10]. These recently published data suggest that, in clinical practice, a determination of the severity of critical illness and outcome using this method remains nothing more than an estimation.
The direct determination of PCO2 using a fiberoptic PCO2 sensor, as in the continuous measurement of PCO2 in blood , can solve the problem of PCO2 instability in the tonometric fluid.
These considerations prompted us to develop the following hypotheses.
1. A fiberoptic PCO2 sensor determines the PCO2 of a certain fluid with better precision than conventional tonometry.
2. A fiberoptic PCO2 sensor determines PCO2 continuously, thus avoiding equilibration times and detecting PCO2 variations.
3. The PiCO2 has a good correlation to variations in arterial and mesenteric venous PCO2.
The fiberoptic PCO2 sensor (Paratrend 7 Registered Trademark, Biomedical Sensors, Highwycombe, UK) consists of two modified optical fibers for the measurements of PCO2 and pH, a miniaturized Clark electrode for the determination of PO2 and a thermocouple for the determination of temperature. All four sensing elements are combined into a device of 0.5 mm in diameter and a length of 600 mm. Since the fiberoptic PCO2 sensor was originally developed for the continuous measurement of PCO2 in blood, the first part of the present study determined the function of the fiberoptic PCO2 sensor in fluids comparable to the medium within the gastrointestinal tract.
Using the fiberoptic PCO2 sensor and for comparison a nasogastric tonometer (TRIP Registered Trademark; NGS-Katheter, Tonometrics Inc., Bethesda, MD) the PCO2 of water and humidified air with predefined PCO2 values was determined in an in vitro experiment.
In a chamber filled with warm water (37 degrees C) equilibration gas was passed through two bubblers. To facilitate optimal diffusion of the gas into the fiberoptic PCO2 sensor and the nasogastric tonometer a stirrer continuously mixed the water. Three different equilibration gas mixtures (Westfalen AG, Munster, Germany) were used: 5%, 6%, and 7% CO2 in N2. Calculation of the predefined PCO2 was performed using the Equation PCO(2) = (PB - 47 mm Hg) centered dot vol% CO2 with the actual barometric pressure (PB) and saturated vapor pressure of water vapor at 37 degrees C (PH2 O = 47 mm Hg).
After reaching steady state conditions for 10 min the fiberoptic PCO (2) sensor and three nasogastric tonometers were placed into the chamber with the CO2 equilibrated water. Saline 2.5 mL was instilled into each tonometer balloon which was aspirated and analyzed after 30, 60, and 90 min for PCO2 with the ABL 30 blood gas analyzer (Radiometer Copenhagen, Copenhagen, Denmark). Simultaneously PCO2 was continuously determined by two fiberoptic PCO2 sensors in the water and above the water in humidified air for 90 min. For every gas mixture four measurements were taken.
To estimate the methodological error of the nasogastric tonometers, a fiberoptic PCO2 sensor was placed into the saline filled balloon and one equilibration was performed with 7% CO2. Because the results of these in vitro experiments were dependent on the actual barometric pressure, we calculated the differences (mean +/- 2 SD in mm Hg) between the predefined PCO2 of the gas mixtures and the measured PCO2 by tonometry or the fiberoptic method for each gas (n = 4).
Furthermore, relative PCO2 differences (predefined PCO2- measured PCO2) were calculated for both methods (n = 12). Measured and predefined PCO (2) values were correlated for the two different methods and the coefficients of correlation (r2) were determined.
In a second part of the study, the in vivo performance of the fiberoptic PCO2 sensor was evaluated in the gastrointestinal tract of six female pigs (67 +/- 6 kg). Experiments were performed in compliance with the institutional review board for the care of animal subjects in accordance with the National Institutes of Health guidelines for ethical animal research. Animals were starved for 48 h with free access to water. Intramuscular premedication was performed using 3 mg/kg azaperone, 5 mg/kg ketamine, and 0.02 mg/kg atropine. Anesthesia was induced with 300-500 mg thiopental intravenously. Animals were placed in supine position, endotracheally intubated, and ventilated using a Servo Ventilator 900 C (Siemens-Elema, Solna, Sweden) with a fraction of inspired oxygen = 0.5, a tidal volume of 10 mL centered dot kg (-1) centered dot min-1, and the respiratory rate 10-12 breaths/min to maintain normocarbia (PaCO2 40.0 +/- 2.0 mm Hg). Anesthesia was maintained throughout the experiment with the continuous administration of ketamine (5-7 mg centered dot kg-1 centered dot h-1), flunitrazepam (0.04-0.06 mg centered dot kg-1 centered dot h-1), and additional boluses of piritramid (7.5 mg, maximum 45 mg).
After midline laparotomy and preparation, the ileum was punctured with an 18-gauge cannula and the fiberoptic PCO2 sensor was introduced 15 cm distally into the lumen. An 18-gauge catheter was placed into the superior mesenteric vein to obtain blood samples. The hemodynamic monitoring consisted of electocar-diography, direct blood pressure measurement in the abdominal aorta, and a pulmonary artery catheter introduced via the right internal jugular vein.
After preparation and 60 min of steady state conditions, the following variables were determined: cardiac index, arterial PCO2 (PaCO2), mesenteric venous PCO2 (PmvCO2), and intramucosal PCO2 (PiCO2).
The second determination of all variables was obtained after 10 min of hypoventilation (50% of minute volume). After hypoventilation steady state condition was reestablished for 30 min and the final measurement of all variables was performed after 10 min of hyperventilation (150% of minute volume).
In vivo data are presented as mean +/- SD. The correlation between PiCO (2) and PaCO2 and between PiCO2 and PmvCO2 were described by computing coefficients of correlation (r2).
(Table 1) shows absolute PCO2 differences between predefined and measured PCO2 values as mean +/- 2 SD of four measurements after 30, 60, and 90 min of equilibration for the three different gas concentrations. PCO2 obtained by tonometry showed greater differences from the predefined PCO2 than those obtained by the fiberoptic PCO2 sensor.
Relative differences from predefined PCO2 (mean of 12 measurements) after 30 min amounted to -26.2% for the tonometer (minimum value = -34.2%, maximum value = -19.6%). After 60 min relative difference was -17.9% (minimum value = -20.8%, maximum value = -16.6%) and after 90 min -12.5% (minimum value = -15.7%, maximum value = -10.6%) Figure 1. The fiberoptic PCO2 sensor was able to determine predefined PCO2 values in water and humidified air after an equilibration time of 9 min. The relative difference to predefined PCO2 values was less than 3.5% Figure 1. When the fiberoptic PCO2 sensor was placed into a saline-filled tonometer balloon, the difference between measured and predefined PCO2 was -9% at 30 min, -1% at 60 min, and +1.4% at 90 min of equilibration. The differences were always less than 2% after 54 min of equilibration time Figure 1.
After 30 min of equilibration time, coefficients of correlation between measured and predefined PCO2 were r2 = 0.91 for nasogastric tonometry, r2 = 1.0 for the fiberoptic PCO2 sensor in water, and r2 = 1.0 in humidified air Figure 2.
The values of the intraluminar PiCO2 determination measured by the fiberoptic PCO2 sensor in six pigs are shown in Figure 3. Changes in CI, PaCO (2), PmvCO2, and PiCO2 caused by ventilation are presented as mean +/- SD (n = 6) in Table 2. The coefficients of correlation for PiCO2 and PaCO (2) were r2 = 0.82, and for PiCO2 and PmvCO2, r2 = 0.94 Figure 4.
To date, nasogastric tonometry is the only available minimally invasive tool for monitoring the perfusion/metabolism of the gastrointestinal tract. In the present study it correlated well (r = 0.91) to predefined PCO2. However, the precision of the determined PCO2 is low. Even after 90 minutes of equilibration the predefined PCO2 was always underestimated by 12.5%. This was most likely caused by slow CO2 diffusion through the permeable silicon balloon of the tonometer. The manufacturer recommends correction of the measured PCO2 with a time-dependent correction factor. This factor is 1.24 after 30 minutes, 1.13 after 60 minutes, and 1.12 after 90 minutes, and is in close agreement with the finding of the present study which would result in the following correction factors: 30 minutes = 1.26; 60 minutes = 1.18; 90 minutes = 1.13. The measurement of the predefined PCO2 of 50.1 mm Hg by the fiberoptic PCO2 sensor introduced into the tonometer balloon confirms the hypothesis of a delayed CO2 diffusion through the silicon balloon. An exact determination of intraluminar CO2 by tonometry should be possible after 54 minutes. The remaining inaccuracy is probably caused by the inhomogeneous deadspace mixing in the tonometer duct when aspirating saline.
However, with tonometry more methodological faults have to be considered. One problem is caused by the difficulty of injecting and aspirating airless saline. Of major importance is the faulty PCO2 determination in saline with standard blood gas analyzers [7-10]. Depending on the blood gas analyzers used, there were CO2 losses resulting in different instrumental biases in PCO2 up to 57.5% when saline was used . A phosphate-buffered solution instead of saline can eliminate the extensive inaccuracy [7,8] of PCO2 determination by the blood gas analyzers. However, the methodological problems (i.e., handling, inhomogeneous deadspace mixing, and dependency of correction factors) would still exist. Finally the determination of PCO2 by tonometry is inaccurate if measurements are performed before 60 minutes of equilibration.
Many studies have been published to demonstrate the importance of measuring intramucosal PCO2 for gastrointestinal monitoring [2,4]. However, in our opinion the inaccuracy of the method necessitates a critical review of these studies. As a basis for a therapeutical approach, a definition of the pathological range needs to be defined. This stands in contrast to the lack of precision and reproducibility of the method, which calls into question the usefulness in clinical settings.
A possible solution is the direct determination of intraluminar PCO2 by a fiberoptic PCO2 sensor. This is the first study that uses a fiberoptic PCO2 sensor for the continuous determination of intestinal CO2.
The validation of the CO2 determination using a fiberoptic PCO2 sensor was performed in an in vitro experiment using predefined media consisting of water and humidified air. This was thought to closely resemble the inhomogeneous medium present in the gastrointestinal tract. In both media, the sensor had a good correlation with predefined PCO2 after nine minutes of equilibration (maximal error +/- 3.5%), with a maximal correlation of r2 = 1.0 in humidified air and water.
The applicability of the fiberoptic PCO2 sensor was tested in an in vivo experiment that confirmed the in vitro results in the inhomogeneous intraluminar milieu of the ileum. The increase in arterial PCO2 caused by hypoventilation led to an intraluminar increase of CO2 according to the hypothesis of CO2 balance between capillaries of the intestinal mucosa and the intestinal lumen. The increase in CO2 caused by hypoventilation and the decrease in CO2 caused by hyperventilation was determined within 10 minutes with a high correlation between PiCO2 and PaCO2 of r2 = 0.82 and between PiCO2 and PmvCO2 of r2 = 0.94.
In every case the determined PiCO2 was greater than the corresponding PmvCO2. The reason for this observation cannot be determined from the present data. However, the PmvCO2 values suggest that the PiCO2 might be closer to the capillary venous PCO2 than to PaCO2. This hypothesis is supported by the higher correlation to PmvCO2. The present data was not sufficient to establish "normal range" for PiCO2.
In conclusion, the present study demonstrates that a fiberoptic PCO2 sensor can exactly determine the intraluminal PCO2 and its changes. Thus it resembles the only technique to reliably monitor PiCO2. In contrast to tonometry, short-term changes of PiCO2 are determined with high precision and reproducibility. The fiberoptic construction, length, and diameter, allow nasogastric placement through a common gastric tube.
The clinical use of this fiberoptic PCO2 sensor can significantly improve the measurement of PiCO2. For the first time, it would be possible to make intraindividual comparisons and to determine valid pathological values of PiCO (2) to treat a gastrointestinal perfusion mismatch.
1. Fiddian-Green RG, Pittenger BS, Whitehouse WM. Back diffusion of CO2 and its influence on the intramural pH in gastric mucosa. J Surg Res 1982;33:39-48.
2. Fiddian-Green RG, Baker S. Predictive value of the stomach wall pH for complications after cardiac operations: comparison with other monitoring. Crit Care Med 1987;15:153-6.
3. Antonsson JB, Boyle CC, Kruithoff KL, et al. Validity of tonometric measures of gut intramural pH during endotoxemia and mesenteric occlusion in pigs. Am J Physiol 1990;259:G519-23.
4. Gutierrez G, Palizas F, Doglio G, et al. Gastric intramucosal pH as a therapeutic index of tissue oxygenation in critically ill patients. Lancet 1992;339:195-9.
5. Bergofsky EH. Determination of tissue O2
tensions by hollow visceral tonometers: effect of breathing enriched O2
mixtures. J Clin Invest 1964;43:192-200.
6. Dawson AM, Trenchard D, Guz A. Small bowel tonometry: assessment of small gut mucosal oxygen tension in dog and man. Nature 1965;206:943-4.
7. Knichwitz G, Mertes N, Kuhmann M. Improved PCO2-measurement
of six standard blood gas analysers with a phosphate-buffered solution for gastric tonometry. Anaesthesia 1995;50:532-4.
8. Knichwitz G, Kuhmann M, Brodner G, et al. Gastric tonometry: precision and reliability are improved by a phosphate buffered solution. Crit Care Med 1996;24:512-6.
9. Riddington D, Venkatesh B, Clutton-Brock TH, et al. Potential hazards in estimation of gastric intramucosal pH. Lancet 1992;340:547.
10. Takala J, Parviainen I, Siloaho M, et al. Saline PCO2
is an important source of error in the assessment of gastric intramucosal pH. Crit Care Med 1994;22:1877-9.
11. Venkatesh B, Clutton-Brock TH, Hendry SP. A multiparameter sensor for continuous intra-arterial blood gas monitoring: A prospective evaluation. Crit Care Med 1994;22:588-94.