Comparison of oesophageal and gastric air tonometry in patients with circulatory failure

U. Janssens*, H. Groesdonk, J. Graf, P. W. Radke, W. Lepper and P. Hanrath

FESC, Medical Clinic I, University of Aachen, Pauwelsstraße 30, D-52057 Aachen, Germany*Corresponding author

Accepted for publication: March 7, 2002


    Abstract
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Conclusions
 References
 
Background. Gastric PCO2 measured by balloon tonometry can estimate the adequacy of splanchnic perfusion. However, enteral feeding and gastric content can interfere with gastric PCO2 assessment. Tonometry in other sites of the body could avoid these problems. We therefore tested the hypothesis that oesophageal air tonometry would give results similar to gastric tonometry.

Methods. We studied 20 consecutive patients (mean age 68 (SD 9) [range 49–81] yr, 18 males, SAPS II score 55 (SD 18), ICU mortality 55%) with circulatory disorders during mechanical ventilation in the intensive care unit. Tonometer probes were placed via the nose, one into the stomach and the other in the oesophagus. PCO2 was measured with two automated gas analysers, at admission and 30 min, 1, 2, 3, 32, 40, and 48 h thereafter.

Results. One hundred and forty-eight paired measurements were obtained. Gastric PCO2 was greater than oesophageal PCO2 on admission (7.19 (1.43) vs 5.89 (0.73) kPa, P<0.01) and subsequently. Differences between the measures correlated (r=0.67) with the mean absolute value, indicating that overestimation increased as gastric PCO2 increased.

Conclusions. Oesophageal PCO2 is less than gastric PCO2, and the difference is greater when gastric PCO2 levels are greater. Air tonometry may not measure regional PCO2 levels in the oesophagus satisfactorily. Other methods and sites for carbon dioxide tonometry should be examined.

Br J Anaesth 2002; 89: 237–41

Keywords: carbon dioxide, measurement; carbon dioxide, partial pressure; ventilation, mechanical


    Introduction
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Conclusions
 References
 
Measurement of gastric PCO2 by balloon tonometry can quantify the adequacy of splanchnic perfusion. However, enteral feeding1 2 and acid gastric content3 4 can increase gastric PCO2 values. In the normally perfused upper gastrointestinal tract, luminal PCO2 may indicate the amount of acid secreted and buffered in the stomach.57 PCO2 increases can occur from the buffering of gastric acid by bicarbonate secreted from the gastric mucosa6 or by reflux of alkaline duodenal contents. Suppression of gastric acid secretion by H2 blockers has been suggested3 4 8 9 before gastric tonometry. However, inhibition of gastric acid secretion in critically ill patients may allow bacterial colonization of the upper gastrointestinal tract which can predispose to pneumonia.10 11

These problems may be avoided by capnometry in other parts of the body. Weil and co-workers found that oesophageal tonometry and more recently sublingual capnometry are practical alternatives to gastric tonometry.1217 Oesophageal luminal PCO215 and sublingual PCO212 could be correlated with gastric PCO2 in animals with haemorrhagic shock and in severely ill patients.17

We therefore tested the hypothesis that regional oesophageal air tonometry oesophageal PCO2 provides results similar to gastric tonometry, in patients with circulatory disturbances.


    Materials and methods
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Conclusions
 References
 
After approval by the local Institutional Review Board, we studied patients with acute circulatory disorders (Table 1). Patients with a history of acute or chronic disease of the upper gastrointestinal tract such as erosive oesophagitis or gastritis, gastric, or duodenal ulcer, or patients with signs of upper gastrointestinal bleeding were excluded. All patients were sedated and receiving mechanical ventilation. Assent was obtained from the closest relatives. Measurements were made at the time of admission and 30 min, 1, 2, 3, 32, 40, and 48 h thereafter. The study ended if gastric tonometers were removed for clinical reasons (e.g. extubation), if the patient died, or 48 h after the start of the study.


View this table:
[in this window]
[in a new window]
 
Table 1 Patient details on admission. CPR=cardiopulmonary resuscitation; AMI=acute myocardial infarction; CS=cardiogenic shock; MAP=mean arterial pressure; HR=heart rate; CI=cardiac index; PaCO2=arterial PCO2; HCO3=arterial bicarbonate; BE=base excess. P-value indicates a significant difference between survivors and non-survivors. Data are displayed as mean (SD). *Data are not normally distributed and displayed with median and interquartile range
 
None of the patients received enteral nutrition. Famotidine (Pepdul®, MSD Chibropharm, München, Germany), 20 mg was given i.v. twice daily. Sedation was maintained with a continuous infusion of fentanyl (Fentanyl®-Janssen, Janssen-Cilag GmbH, Neuss, Germany) and midazolam (Dormicum®, Hoffman-La Roche AG, Grenzach-Wyhlen, Germany). Catecholamines were given depending on clinical requirements. The treatment of the patients was not influenced by the study. All measurements were obtained by a single investigator.

A tonometer probe (Trip catheter 14 F, Datex-Ohmeda, Finland) was placed via the nose into the stomach. The position was verified by aspiration of gastric juice followed by air insufflation and auscultation. A second probe was inserted into the stomach and correct placement confirmed clinically too. Then this probe was drawn back to lie with the tip 30–35 cm from the nasal opening. Air was insufflated and the upper abdomen auscultated to check that the probe was not in the stomach. Finally, correct placement of both probes was verified radiologically. A gas analyser (Tonocap TC 200, Datex-Ohmeda) was connected to each tonometer catheter. With each measurement cycle, 6 ml of air is pumped into the balloon. A sample is then automatically drawn. The first 1.2 ml are from the dead space of the system and are discarded. The PCO2 is measured in the remainder with an infrared sensor.

Arterial blood gas analysis was done using a standard blood gas analyser (ABL 505 blood gas analyser, Radiometer, Copenhagen, Denmark). Samples were analysed by the blood gas analyser at 37°C and corrected for the patient’s actual body temperature using internal software according to the manufacturer’s handbook

(PCO2[temp]=PCO2[37°C]x100.021x[temp–37°C])

The gastric or oesophageal PCO2 – arterial PCO2 (PaCO2) difference (CO2 Gap[G,O]) was determined according to the formula:

CO2 Gap[G,O]=PCO2 [G,O]–PaCO2

where the carbon dioxide Gap[G,O] is in kPa.

All patients had a radial arterial cannula (Leadercath 11509, 20 gauge 8 cm, Vygon, Aachen, Germany) and a pulmonary artery catheter (Continuous Cardiac Output Combo Catheter®, Baxter Edwards Critical Care, Irvine, CA), which was connected to a computer system (VigilanceTM, Baxter Edwards Critical-Care). Vascular pressures (central venous pressure (CVP); mean pulmonary artery pressure (MPAP); pulmonary artery occlusion pressure (PAOP); systemic mean arterial pressure (MAP), mm Hg) and cardiac output were measured at the given time points. Cardiac index (CI) in litre min–1 m–2, oxygen delivery index (DO2I) in ml O2 min–1 m–2 and oxygen consumption index (VO2I) in ml O2 min–1 m–2 were calculated from standard formulae.

Statistical analysis on a microcomputer was with the Statistical Package for the Social Sciences (SPSS 10.0, SPSS Inc., Chicago, IL). All continuous data were tested for normal distribution by the Kolomogorov–Smirnov test. Normally distributed data are presented as mean and standard deviation (SD), non-normally distributed data as median and interquartile range. Survivors and non-survivors were compared at study entry using an unpaired Student’s t-test if data were normally distributed and the Mann– Whitney U-test if non-normally distributed. A chi-squared test was used for nominal data. To compare changes with time, repeated measures analysis of variance was used. Values at any time point were compared between survivors and non-survivors by unpaired Student’s t-test or Mann– Whitney U-test, respectively. Furthermore, mean gastric PCO2 and oesophageal PCO2 values at each of the eight time points were correlated with each other and the corresponding mean values of haemodynamic variables (CI, MAP, MPAP, PAOP, SVRI, PVRI, DO2I, VO2I) and arterial blood gases. Simple linear regression analysis was performed on the differences of both methods against the means. Moreover, we calculated the mean bias and the average mean of both methods for any patient and performed simple linear regression analysis. Correlations were calculated using Pearson’s correlation coefficient. Significance was assumed at P<0.05 unless otherwise stated.


    Results
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Conclusions
 References
 
Patient details are shown in Table 1. One patient died at 6 h, one patient at 13 h, one patient at 37 h, and one patient at 44 h after the start of the study. One patient was extubated 6 h after admission. Thus, 15 patients completed the study. Complications associated with the tonometer probes, such as nasopharyngeal injury, oesophageal or gastric bleeding, were not seen.

Table 2 gives haemodynamic and blood gas values throughout the study. One hundred and forty-eight paired measurements of gastric PCO2 and oesophageal PCO2 were obtained. Gastric PCO2 was significantly greater than oesophageal PCO2 at admission (7.19 (1.43) vs 5.89 (0.73) kPa, P<0.01), and gastric carbon dioxide gap was greater than oesophageal carbon dioxide gap (1.73 (1.09) vs 0.45 (0.87) kPa, P<0.01). The differences persisted for the duration of the study (Fig. 1).


View this table:
[in this window]
[in a new window]
 
Table 2 Mean values of haemodynamic and blood gas values. CI=cardiac index; HR=heart rate; DO2I=oxygen delivery index; VO2I=oxygen consumption index; MAD=mean arterial pressure; MPAP=mean pulmonary artery pressure; PAOP=pulmonary artery occlusion pressure; apH=arterial pH; PaCO2=arterial PCO2; BE = base excess. *Data are not normally distributed and displayed as median with interquartile range
 


View larger version (24K):
[in this window]
[in a new window]
 
Fig 1 (A) Gastric PCO2 (PCO2[G]) and oesophageal PCO2 (PCO2[O]). (B) Gastric carbon dioxide gap (CO2 Gap[G]) and oesophageal carbon dioxide gap (CO2 Gap[O]) during the study period.

 
There was a good correlation between the differences of both methods and the means (r=0.67, P<0.001) indicating that overestimation increased with greater values of gastric PCO2. Analysing the mean bias of any given patient showed that the discrepancy between the two methods was significantly pronounced in patients with a high regional carbon dioxide level (Fig. 2). In addition, mean bias at any time correlated with the average means of both methods at any time point (r=0.72, P=0.04).



View larger version (16K):
[in this window]
[in a new window]
 
Fig 2 Scatter plot showing the components of simple linear regression analysis. The explanatory variable is the average of the means of both methods for each patient, the response variable the mean of the bias for any given patient. PCO2[G]=gastric PCO2; PCO2[O]=oesophageal PCO2.

 
There were no differences between survivors and non-survivors in tonometry, haemodynamic, and blood gas values at any given time point and no correlation between haemodynamic measures (CI, DO2I, VO2I) and regional gastric and oesophageal PCO2 values. Arterial pH (r=–0.7, P<0.001) and PaCO2 (r=0.75, P<0.001) correlated well with gastric PCO2 but not with oesophageal PCO2 (r=–0.31, P=0.188 and r=0.39, P=0.08, respectively).


    Discussion
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Conclusions
 References
 
Although lumen tonometry has been widely used to predict outcome,1820 assess the effects of treatment,21 22 or as an early indicator of colon ischaemia23 24 several technical and methodological drawbacks have hindered its widespread use in intensive care. In particular, carbon dioxide formed in the stomach by gastric acid reacting with bicarbonate, or bacterial fermentation in the gut may reduce the ability of luminal PCO2 to indicate the adequacy of mucosal perfusion in the stomach and large bowel, respectively.5 Some workers suggested measurement of the visceral PCO2 in places such as in the oesophagus or under the tongue.12 13 15 17 25 Using ion-sensitive field-effect transistor (ISFET) measurements15 or capnometric recirculating gas tonometry (CRGT),25 animal studies found a reliable relationship between oesophageal PCO2 and gastric PCO2. However, no human data are currently available.

We found that oesophageal PCO2 was different from gastric PCO2 in critically ill patients with circulatory disturbances. The difference increased as gastric PCO2 increased. Differences of arterial blood supply in the oesophagus and other regions of the gastrointestinal tract, compared with the gastric mucosa, may limit the sensitivity of tonometry for early detection of general hypoperfusion,5 as selective vasoconstriction may occur in the stomach or small bowel.26 In contrast, Weil and co-workers found that increases in gastric PCO2 in rats with a haemorrhagic shock were highly correlated with increases in oesophageal PCO2,15 and that oesophageal blood flow and gastric blood flow decreased proportionally. Likewise, in dogs with septic shock, oesophageal PCO2 and gastric PCO2 measured with CRGT showed a similar trend.25 However, after endotoxin treatment, gastric PCO2 increased more rapidly than oesophageal PCO2. Sublingual tissue PCO2 measured with a miniature carbon dioxide electrode in rats with haemorrhage showed a highly significant linear correlation with gastric PCO2.13 Increased sublingual PCO2 correlated well (r2=0.84, P<0.001) with increased lactate concentrations in 46 patients with life-threatening illnesses.17 Unfortunately, gastric PCO2 levels were not assessed in this convincing study.

Our findings do not support these results. The methods for regional PCO2 measurement in the above experiments were quite different from our study except for the CRGT used by Guzman and co-workers,25 which is similar to the automated gas analyser of our study. In contrast to the air tonometry the ISFET method measures regional PCO2 directly in the gastric wall.15 The ISFET technology and CRGT allow continuous measurement and display, giving rapid easy measurements.15 25 In addition, the speed of onset and severity of haemorrhagic shock were profound in the animal study of Sato and co-workers. The close relationship they found between gastric PCO2 and oesophageal PCO2 may not apply under conditions of lesser severity in other settings such as septic shock.15

Secondly, we studied patients with acute circulatory disorders. Although overall mortality was high, patients were not profoundly shocked. Such patients may differ considerably from the haemorrhagic or septic shock studied in the animal experiments.

Finally, poor correlation of oesophageal PCO2 and arterial PCO2 compared with the close correlation of gastric PCO2 and arterial PCO2 suggests a methodological bias. The oesophagus is not a closed system and carbon dioxide may partly diffuse to the rhinopharynx or be diluted with ambient air. Gas diffusion is proportional to the difference in partial pressure between two phases and the area of the exchange surface, so that diffusion will increase with an increased gradient. This may have contributed to the observed disparity between oesophageal PCO2 and gastric PCO2 and explain the greater bias with greater regional PCO2 levels. In addition, the equilibration period will cause carbon dioxide loss near the balloon. This effect could be circumvented by shorter equilibration times or measurements using systems such as the COMOCADOF (fibre-optical catheter equipped with a carbon dioxide sensor capable of continuous carbon dioxide assessment).27

There are differences between the stomach and the oesophagus. Buffering of acid reflux into the oesophageal lumen from the stomach is by bicarbonate ions entering the oesophageal lumen, from swallowed saliva, and from the oesophageal wall.28 29 The epithelium of the oesophagus contains carbonic anhydrases (CA) II,30 which catalyse the interconversion of carbon dioxide and H2CO3.29 Oeso phageal luminal carbon dioxide may not come just from oesophageal tissues but also from fluid, which contains H2CO3, CA, and gastric acid. These possibilities do not explain our findings but suggest that oesophageal PCO2 measurements have the same uncertainties as gastric PCO2.

Although our patients received i.v. histamine receptor block we did not control the intragastric or intraoesophageal pH, and we cannot rule out possible interference of H+ with HCO3 in the oesophageal lumen. Routine use of H2 blockers to reduce the error of gastric PCO2 measurements has been questioned.31 The recommendations are based mainly on studies in human volunteers. H2 blockers in the critically ill patients may not affect the assessment of gastric PCO2.32 Discrepancies between healthy volunteers and critically ill patients may be related to reduced gastric acid secretion in the latter who have reduced visceral perfusion.33

Despite extensive study there are no accepted normal values for gastric PCO2, so critical appraisal of oesophageal PCO2 alone will be difficult.


    Conclusions
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Conclusions
 References
 
Oesophageal air tonometry severely underestimates gastric PCO2 and difference increases at greater PCO2 values. Air tonometry may not be the appropriate technique to assess regional PCO2 levels in the oesophagus. Future research should target alternative methods and sites for continuous carbon dioxide measurement.


    References
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Conclusions
 References
 
1 Levy B, Perrigault PF, Gawalkiewicz P, et al. Gastric versus duodenal feeding and gastric tonometric measurements. Crit Care Med 1998; 26: 1991–4[ISI][Medline]

2 Marik PE, Lorenzana A. Effect of tube feedings on the measurement of gastric intramucosal pH. Crit Care Med 1996; 24: 1498–500[ISI][Medline]

3 Kolkman JJ, Groeneveld AB, Meuwissen SG. Effect of ranitidine on basal and bicarbonate enhanced intragastric PCO2: a tonometric study. Gut 1994; 35: 737–41[Abstract]

4 Heard SO, Helsmoortel CM, Kent JC, Shahnarian A, Fink MP. Gastric tonometry in healthy volunteers: effect of ranitidine on calculated intramural pH. Crit Care Med 1991; 19: 271–4[ISI][Medline]

5 Kolkman JJ, Otte JA, Groeneveld AB. Gastrointestinal luminal PCO2 tonometry: an update on physiology, methodology and clinical applications. Br J Anaesth 2000; 84: 74–86[Free Full Text]

6 Odes HS, Hogan DL, Steinbach JH, Ballesteros MA, Koss MA, Isenberg JI. Measurement of gastric bicarbonate secretion in the human stomach: different methods produce discordant results. Scand J Gastroenterol 1992; 27: 829–36[ISI][Medline]

7 Rune SJ. Acid–base parameters of duodenal contents in man. Gastroenterology 1972; 62: 533–9[ISI][Medline]

8 Bams JL, Kolkman JJ, Roukens MP, et al. Reliable gastric tonometry after coronary artery surgery: need for acid secretion suppression despite transient failure of acid secretion. Intensive Care Med 1998; 24: 1139–43[ISI][Medline]

9 Heard SO. Suppression of gastric acid secretion: preventing excess gas. Intensive Care Med 1998; 24: 1123–5[ISI][Medline]

10 Cook DJ, Laine LA, Guyatt GH, Raffin TA. Nosocomial pneumonia and the role of gastric pH. A meta-analysis. Chest 1991; 100: 7–13[Abstract]

11 Cook DJ. Stress ulcer prophylaxis: gastrointestinal bleeding and nosocomial pneumonia. Best evidence synthesis. Scand J Gastroenterol Suppl 1995; 210: 48–52

12 Jin X, Weil MH, Sun S, Tang W, Bisera J, Mason EJ. Decreases in organ blood flows associated with increases in sublingual PCO2 during hemorrhagic shock. J Appl Physiol 1998; 85: 2360–4[Abstract/Free Full Text]

13 Nakagawa Y, Weil MH, Tang W, et al. Sublingual capnometry for diagnosis and quantitation of circulatory shock. Am J Respir Crit Care Med 1998; 157: 1838–43[Abstract/Free Full Text]

14 Noc M, Weil MH, Sun S, Gazmuri RJ, Tang W, Pakula JL. Comparison of gastric luminal and gastric wall PCO2 during hemorrhagic shock. Circ Shock 1993; 40: 194–9[ISI][Medline]

15 Sato Y, Weil MH, Tang W, et al. Esophageal PCO2 as a monitor of perfusion failure during hemorrhagic shock. J Appl Physiol 1997; 82: 558–62[Abstract/Free Full Text]

16 Pernat A, Weil MH, Tang W, et al. Effects of hyper- and hypoventilation on gastric and sublingual PCO2. J Appl Physiol 1999; 87: 933–7[Abstract/Free Full Text]

17 Weil MH, Nakagawa Y, Tang W, et al. Sublingual capnometry: a new noninvasive measurement for diagnosis and quantitation of severity of circulatory shock. Crit Care Med 1999; 27: 1225–9[ISI][Medline]

18 Doglio G, Pusajo J, Egurrola M, Bonfigli G, Gutierrez G. Gastric mucosal pH as a prognostic index of mortality in critically ill patients. Crit Care Med 1991; 19: 1037–40[ISI][Medline]

19 Maynard N, Bihari D, Beale R, et al. Assessment of splanchnic oxygenation by gastric tonometry in patients with acute circulatory failure. JAMA 1993; 270: 1203–10[Abstract]

20 Marik PE. Gastric intramucosal pH. A better predictor of multiorgan dysfunction syndrome and death than oxygen-derived variables in patients with sepsis. Chest 1993; 104: 225–9[Abstract]

21 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[ISI][Medline]

22 Mythen MG, Webb AR. Perioperative plasma volume expansion reduces the incidence of gut mucosal hypoperfusion during cardiac surgery. Arch Surg 1995; 130: 423–9[Abstract]

23 Klok T, Moll FL, Leusink JA, Theunissen DJ, Gerrits CM, Keijer C. The relationship between sigmoidal intramucosal pH and intestinal arterial occlusion during aortic reconstructive surgery. Eur J Vasc Endovasc Surg 1996; 11: 304–7[ISI][Medline]

24 Soong CV, Halliday IM, Barclay GR, Hood JM, Rowlands BJ, Barros D’SA AA. Intramucosal acidosis and systemic host responses in abdominal aortic aneurysm surgery. Crit Care Med 1997; 25: 1472–9[ISI][Medline]

25 Guzman JA, Lacoma FJ, Kruse JA. Gastric and esophageal intramucosal PCO2 (PiCO2) during endotoxemia: assessment of raw PiCO2 and PCO2 gradients as indicators of hypoperfusion in a canine model of septic shock. Chest 1998; 113: 1078–83[Abstract/Free Full Text]

26 Edouard AR, Degremont AC, Duranteau J, Pussard E, Berdeaux A, Samii K. Heterogeneous regional vascular responses to simulated transient hypovolemia in man. Intensive Care Med 1994; 20: 414–20[ISI][Medline]

27 Lueger A, Scherr D, Brunner G, Trettnak W, Mencaglia A, Baldini F. Continuous monitoring of gastric carbon dioxide with optical fibers: European project COMOCADOF, in vitro evaluation. Crit Care 2000; 4 (Suppl 1): S85–S6

28 Meyers RL, Orlando RC. In vivo bicarbonate secretion by human esophagus. Gastroenterology 1992; 103: 1174–8[ISI][Medline]

29 Parkkila S, Parkkila AK. Carbonic anhydrase in the alimentary tract. Roles of the different isozymes and salivary factors in the maintenance of optimal conditions in the gastrointestinal canal. Scand J Gastroenterol 1996; 31: 305–17[ISI][Medline]

30 Parkkila S, Parkkila AK, Juvonen T, Rajaniemi H. Distribution of the carbonic anhydrase isoenzymes I, II, and VI in the human alimentary tract. Gut 1994; 35: 646–50[Abstract]

31 Baigorri F, Calvet X, Joseph D. Gastric intramucosal pH measurement. Crit Care 1997; 1: 61–4[Medline]

32 Calvet X, Baigorri F, Duarte M, et al. Effect of ranitidine on gastric intramucosal pH in critically ill patients. Intensive Care Med 1998; 24: 12–7[ISI][Medline]

33 Higgins D, Mythen MG, Webb AR. Low intramucosal pH is associated with failure to acidify the gastric lumen in response to pentagastrin. Intensive Care Med 1994; 20: 105–8[ISI][Medline]





This Article
Abstract
Full Text (PDF)
E-Letters: Submit a response to the article
Alert me when this article is cited
Alert me when E-letters are posted
Alert me if a correction is posted
Services
Email this article to a friend
Similar articles in this journal
Similar articles in ISI Web of Science
Similar articles in PubMed
Alert me to new issues of the journal
Add to My Personal Archive
Download to citation manager
Disclaimer
Request Permissions
Google Scholar
Articles by Janssens, U.
Articles by Hanrath, P.
PubMed
PubMed Citation
Articles by Janssens, U.
Articles by Hanrath, P.