This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Email this article to a colleague Similar articles in this journal Similar articles in Web of Science Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Permissions Citing Articles Citing Articles via Web of Science (6) Citing Articles via Google Scholar Google Scholar Articles by Lebuffe, G. Articles by Webb, A. Search for Related Content PubMed PubMed Citation Articles by Lebuffe, G. Articles by Webb, A. Related Collections Economics and Health Care Research Postanesthetic Care Unit Monitoring (Non-cardiac)

---

ECONOMICS, EDUCATION, AND HEALTH SYSTEMS RESEARCH

A European, Multicenter, Observational Study to Assess the Value of Gastric-to-End Tidal PCO₂ Difference in Predicting Postoperative Complications

Gilles Lebuffe, MD PhD^*, Benoît Vallet, MD PhD^*, Jukka Takala, MD PhD, Gary Hartstein, MD DSc, Maurice Lamy, MD, Monty Mythen, MD FRCA, Jan Bakker, MD PhD^||, David Bennett, MD FRCP^**, Owen Boyd, MD FRCA^**, and Andrew Webb, MD FRCP

*Department of Anesthesiology 2, CHU de Lille, Lille, France; Department of Intensive Care, Kuopio University Hospital, Kuopio, Finland; Department of Anesthesiology and Intensive Care, CHU de Liege, Liege, Belgium; Department of Intensive Care, UCL Hospitals Mortimer St, London; ||Department of Intensive Care, University Hospital, Apeldoorn, Netherlands; **Department of Intensive Care, St George’s Hospital, London, UK

Address correspondence and reprint requests to A. R. Webb, MD, FRCP, Medical Director, UCL Hospitals NHS Trust, John Astor House, Foley Street, London W1W 6DN, United Kingdom. Address email to a.webb{at}uclh.org

	Abstract

Top Abstract Introduction Methods Results Discussion References

 
Automated online tonometry displays a rapid, semi-continuous measurement of gastric-to-endtidal carbon dioxide (Pr-etCO₂) as an index of gastrointestinal perfusion during surgery. Its use to predict postoperative outcome has not been studied in general surgery patients. We, therefore, studied ASA physical status III–IV patients operated on for elective surgery under general anesthesia and a planned duration of >2 h in a European, multicenter study. As each center was equipped with only 1 tonometric monitor, a randomization was performed if more than one patient was eligible the same day. Patients not monitored with tonometry were assessed only for follow-up. The main outcome measure was the assessment of postoperative functional recovery delay (FRD) on day 8. Among the 290 patients studied, 34% had FRD associated with a longer hospital stay. The most common FRDs were gastrointestinal (45%), infection (39%), and respiratory (35%). In those monitored with tonometry (n = 179), maximum Pr-etCO₂ proved to be the best predictor increasing the probability of FRD from 34% for all patients to 65% at a cut-off of 21 mm Hg (2.8kPa) (sensitivity 0.27, specificity 0.92, positive predictive value 64%, negative predictive value 70%). We conclude that intraoperative Pr-etCO₂ measurement may be a useful prognostic index of postoperative morbidity.

IMPLICATIONS: Gastric-to-end tidal partial pressure difference of carbon dioxide (Pr-etCO₂) is recognized as an index of gastrointestinal perfusion during surgery. In a high-risk surgical population with an expected duration of surgery of more than 2 h, this European, multicenter observational study suggests that automated semi-continuous monitoring of Pr-etCO₂ can be used as an intraoperative predictor of poor outcome.

	Introduction

Top Abstract Introduction Methods Results Discussion References

 
Multiple organ failure (MOF) is the most common cause of death in surgical intensive care units (ICUs). It is increasingly recognized that milder forms of MOF are associated with a less severe form of organ dysfunction that also increases hospital length of stay and cost. This problem is common; data from two studies revealed an average of 8.5 to 20 additional days in hospital (1,2) that occurred in approximately 15% of patients undergoing major elective surgery (2,3).

The etiology of postoperative morbidity is multifactorial, including both severe life-threatening complications such as organ failure but also less severe complications such as wound infection. These overall complications that contribute to prolonged hospitalization and health care cost can be easily and reproducibly identified by the postoperative morbidity survey (POMS) (4).

Compromised physiologic reserve (5) and occult hypovolemia (6,7) associated with extensive surgery are thought to represent the most common cause of organ dysfunction and death. In this context, the body reflex defense mechanisms reduce flow to certain end organs (e.g., the gut) to preserve flow to the heart, lung, and brain (6,7). Because the gut is particularly susceptible to hypoperfusion, it has been advocated to provoke distant organ injury by exacerbating the magnitude of systemic inflammation in response to an ischemia/reperfusion injury as well as to release bacteria-derived products (8,9). Consequently, appropriate intervention at an early stage of hypoperfusion might limit the development of MOF and its associated increased morbidity and mortality in high-risk surgical patients.

Gastric tonometry is the sole clinically available device allowing monitoring of gastrointestinal perfusion. The initial methodological drawbacks with saline tonometry have been solved by the development of an automated method using air tonometry to measure gastric carbon dioxide (PrCO₂) semi-continuously. By referencing PrCO₂ to arterial CO₂ (Pr-aCO₂), this measurement has proved to be an early indicator of shock and a good predictor of postoperative complications in cardiac surgery (10). Its clinical use is limited by the need for arterial sampling, leading to the proposal of gastric-to-end tidal carbon dioxide difference (Pr-etCO₂) as a noninvasive marker of regional perfusion. Its usefulness as a reliable method for continuous monitoring of gastric mucosal perfusion has been reported in critically ill (11) and surgical (12) patients. In addition, Lebuffe et al. (13) found that a high Pr-etCO₂ on admission to the ICU as well as high Pr-aCO₂ were associated with a poorer outcome postcardiac surgery. Hence, semi-continuous monitoring of Pr-etCO₂ might allow earlier intraoperative prediction of poor outcome at a time when intervention may be helpful.

Therefore, this study was conducted to investigate the relationship between intraoperative increased Pr-etCO₂ and the incidence of postoperative complications assessed by POMS.

	Methods

Top Abstract Introduction Methods Results Discussion References

 
This prospective, observational study was conducted in 6 major European hospitals. Ethical approval was gained by each participating center before the onset of the study. After written consent was obtained from patients who met the eligibility criteria during a preoperative visit, ASA physical status III–IV patients scheduled for elective, noncardiac and non-neurological surgery with general anesthesia and a planned duration of more than 2 h were included in the study.

The following exclusion criteria were observed: participation in another study, age <18 yr, pregnancy, contraindication to passage of a nasogastric tube, esophageal pathology, and gastric or esophageal surgery.

The study was conducted over a period of 6 mo in each center during 1998. Each center was equipped to study only one patient with tonometry at any one time. If more than one patient was eligible in any center on any one day, then one patient was randomly selected to be monitored with automated online tonometry (Tonocap®, Datex-Ohmeda Division, Instrumentarium Corp., Helsinki, Finland). Other patients were assessed for functional recovery delay (FRD) by follow-up only. Randomization was performed by computer (custom software, Datex-Ohmeda Division, Instrumentarium Corp., Helsinki, Finland) before the day’s surgery starting. If only one patient was eligible for study he or she was monitored with automated online tonometry.

Anesthetic Management
Patients were given ranitidine 150 mg PO on the night before surgery and on the morning of surgery to reduce gastric luminal CO₂ production as a result of reflux buffering of gastric acid. Anesthesia was induced by an IV technique and maintained with a balanced inhaled technique. Choice of anesthetic was not standardized. Supplemental epidural analgesia was allowed. Ventilation was adjusted to maintain PaCO₂ at 35–45 mm Hg (4.6–6.0 kPa) throughout.

Monitoring
After the induction of anesthesia an arterial cannula was inserted in routine fashion and a gastrointestinal tonometer (Datex-Ohmeda Division, Instrumentarium Corp.) was introduced into the patient’s stomach. Correct placement was confirmed by the injection of air down the nasogastric lumen while auscultating over the epigastrium. The tonometer and the airway sampling line attached to the patient’s airway adapter were connected to the Tonocap®. After the tonometer balloon was filled with air and allowed to equilibrate for 10 min, the gas was automatically sampled and measured by infrared spectroscopy. The device interrupts the monitoring of end-tidal CO₂ (PETCO₂) once every 10 min for the determination of gastric mucosal CO₂ (PrCO₂). The Tonocap® screen was covered to blind operating department staff to the PrCO₂ data. Tonocap® data were stored to a personal computer continuously.

Measurements
Patients monitored with the automated online tonometry were risk assessed according to Shoemaker et al.’s risk categories (14) on study entry and with the physiological and operative severity score for the enumeration of mortality and morbidity (POSSUM) score (15). Arterial blood gases were sampled on insertion of the cannula and hourly throughout surgery allowing calculation of gastric-to-arterial PCO₂ difference (Pr-aCO₂), end-surgical base deficit, and end-surgical pHi. Because the Tonocap® measures PrCO₂ at body temperature end-surgical pHi and base deficit were corrected for body temperature according to the following formula:

where T is end surgical core temperature.

End surgical measurements were performed before discharging patients from the operating room. At that time, the following intraoperative data were collected: blood loss >1000 mL, the number of units of blood transfused, the volume of crystalloid and colloid infusion, the end surgical core temperature, and the duration of surgery.

The postoperative follow-up was assessed by recording the length of ICU stay and overall postoperative length of stay and hospital outcome. All hospitalized patients were evaluated on postoperative days 2 and 8 using the POMS (4) listed in Table 1. Each assessment was done by speaking with the patient, examining the patient, and consulting the medical record and/or the nurses. FRD was the primary outcome of this study. Given previous work reporting that postoperative length of stay more than 7 days was associated with persistent dysfunction of at least one organ system in noncardiac and non-neurological surgery (4), the presence of at least one criterion of the POMS on postoperative day 8 defined FRD.

A number of potential predictors of FRD were analyzed. These included Shoemaker risk categories, POSSUM score, maximum Pr-aCO₂ difference, maximum Pr-etCO₂ difference, end-surgical pHi, and end-surgical base deficit. The primary outcome variable was the ability of Pr-etCO₂ to predict FRD. Sensitivity and specificity for the prediction of FRD was calculated for multiple cut-off points according to standard formulae. A routine was written in Microsoft FoxPro to automate the calculations. Data were imported to Microsoft Excel 7.0 for Windows for construction of receiver operating characteristic (ROC) curves. Using the sensitivities and specificities for each cut-off point and the prevalence of FRD among all patients studied, Bayes’ theorem was used to calculate the probability of FRD in a patient with an abnormal test positive predictive value (PPV) and the probability of FRD in a patient with a normal test (equivalent to 1-negative predictive value [NPV]) (16).

The information content (uncertainty before the prediction minus the uncertainty after the prediction) of various ROC cut-off points was assessed according to information theory, taking into account sensitivity, false positive ratio, and the prevalence of FRD in the study patients (17). Information content (I) was calculated for pairs of conditional true positive and false positive probabilities to assess the maximum amount of information available in a series of ROC curves. The test cut-off with the maximum information (I_max) was used as the definition of abnormal for the test. The highest I_max reduces the uncertainty of a decision based on the cut-off to a minimum.

	Results

Top Abstract Introduction Methods Results Discussion References

 
Of 290 patients enrolled in the study over the 6-mo period, 179 patients were monitored with the Tonocap®. For the 290 patients the median age was 69 (21–93) years, the proportion of males was 59%, and the median length of stay in ICU and in hospital was 0 (0–99) and 11 (1–99) days, respectively (Table 2). On postoperative day 8, 100 patients had FRD (34%) and 15 (5%) died. Because death was not selected as a primary outcome for this study, patients who died were assessed for morbidity on the relevant days if they were alive at that time.

Information theory was used to assess prediction of FRD for the scoring systems, arterial base deficit, and tonometric indices. The abnormal threshold values for predicting FRD (values with the highest I_max) were 2 for Shoemaker risk category, 31 for POSSUM score, 3 mmol/L for end-surgical base deficit, 7.15 for end-surgical pHi, 11 mm Hg (1.5 kPa) for maximum Pr-aCO₂, and 21 mm Hg (2.8 kPa) for maximum Pr-etCO₂. At these cut-off values, the best predictive power was observed for maximum Pr-etCO₂, as I_max value (0.045) was higher than for maximum Pr-aCO₂ (0.028), POSSUM score (0.019), end-surgical pHi (0.018), end-surgical base deficit (0.010), or Shoemaker risk category (0.006). At this cut-off value, maximum Pr-etCO₂ increased probability of FRD from 35% to 65% (Fig. 1). Predictive value for each test is given in Table 5.

View larger version (14K): [in this window] [in a new window]   Figure 1. The left panel shows the probability of functional recovery delay (FRD) at various cut-off thresholds for maximum Pr-etCO2. The dotted plot represents the probability of FRD for all patients studied. The probability of FRD if the test is positive at that cut-off point is shown in black. The probability of FRD if the test is negative at that cut-off point is shown in gray. Imax occurred at a cut-off threshold of 21 mm Hg (2.8 kPa). The right panel shows the corresponding ROC curve.

	Discussion

Top Abstract Introduction Methods Results Discussion References

 
The use of Pr-etCO₂ offers the possibility for online measurement of gastrointestinal CO₂ referenced to systemic values without the need for arterial blood sampling. A high Pr-etCO₂ is regarded as indicative of an imbalance between gastric perfusion and alveolar ventilation and is not affected by systemic acid-base status. The principal observation of the present study is that maximum Pr-etCO₂ measured during the operative course could predict FRD.

This is the first study to evaluate the capacity of intraoperative Pr-etCO₂ to predict postoperative complications in high-risk surgical patients after major surgery. After cardiac surgery, Bennett-Guerrero et al. (10), using gas tonometry, reported that Pr-aCO₂ was a predictor of prolonged hospitalization. Similarly, Lebuffe et al. (13) observed that hospital length of stay was significantly increased (5 days) in postcardiac surgery patients with a Pr-aCO₂ between 12 mm Hg (1.6 kPa) and 20 mm Hg (2.7 kPa). Interestingly, in these patients the same trend was observed with Pr-etCO₂ values ranging between 18 mm Hg (2.4 kPa) and 28 mm Hg (3.7 kPa). In the present study, the cut-off that maximized the certainty of prediction was 21 mm Hg (2.8 kPa) for maximum Pr-etCO₂. Because most previous studies were performed in critically ill patients using pHi, there are currently no data concerning the threshold value of Pr-etCO₂ in surgical patients. In the ICU, Pr-aCO₂ has recently been shown to predict death for values more than 20 mm Hg (2.6 kPa) (18). However, Pr-etCO₂ depends both on Pr-aCO₂ and on the difference between arterial and end tidal PCO₂ (Pa-etCO₂). Previous data in patients without preoperative cardiac dysfunction suggest that a high Pr-etCO₂ could be mainly attributed to Pr-aCO₂ increase whereas Pa-etCO₂ would not exceed 3.8 to 7.6 mm Hg (0.5 to 1.0 kPa) (12,13). This suggests that the cut-off we found for Pr-etCO₂ corrected for Pa-etCO₂ is near the value of 20 mm Hg (2.6 kPa) for Pr-aCO₂ reported in an experimental study as the cut-off point for anaerobic metabolism (19).

Gastric tonometry can detect hypovolemia within minutes (20). With its brief period of equilibration, the automated-air gastric tonometry provides Pr-etCO₂ values every 10 minutes, allowing calculation of a cumulative sum of Pr-etCO₂ above the "normal" baseline gap. There is currently no evidence that brief transient episodes of gastric hypoperfusion are clinically relevant. In contrast, several authors have demonstrated in critically-ill patients that a persistently abnormal pHi (21) or Pr-aCO₂ (18) value during the first 24 hours after ICU admission was associated with increased mortality. In our study, the maximum certainty for prediction was observed at a cumulative Pr-etCO₂ of 150 mm Hg (20 kPa) above a "normal" level of 8.25 mm Hg (1.1 kPa). There was no gain in predictive power of this calculation compared to assessment of maximum Pr-etCO₂ gap. It is interesting to note that 12 measurements of a Pr-etCO₂ value just above 21 mm Hg (2.8 kPa) would be required to achieve the cut-off of 150 mm Hg (20 kPa). Because automated gastric tonometry can display up to 6 measurements per hour, a patient considered to be abnormal by the maximum Pr-etCO₂ would remain normal by the cumulative Pr-etCO₂ for 2 hours. This observation may explain why severe and long-lasting episodes of gastric hypoperfusion, as previously reported by low pHi or high Pr-aCO₂ at the end of surgery and postoperatively, were as or more robust predictors of postoperative complications than transient episodes with higher sensitivity and specificity. This issue requires further study in which a therapeutic change aimed at decreasing a high intraoperative cumulative Pr-etCO₂ value is assessed against outcome in a selected high-risk surgical population.

Our choice of a primary outcome variable warrants comment. Although mortality is the most widely used measure in assessing patient outcome in the ICU, it is an insensitive measure of outcome for elective surgical patients because death is relatively rare. In the present work, the overall mortality was 5%, a value that is consistent with the mortality observed in cardiovascular surgery (10). More important is the assessment of postoperative morbidity as a measure of surgical outcome. However, there are no accepted methods for assessing postoperative morbidity in patients undergoing surgical procedures. The duration of hospital stay cannot be used as the sole criterion to evaluate postoperative morbidity because it is affected by non-medical factors. We decided to measure FRD on postoperative day 8 based on the observation that a postoperative length of stay more than 7 days was associated with persistent dysfunction of at least one organ system (4). The presence of evidence of delayed recovery on postoperative day 8 was considered as a clinically relevant estimate of outcome in our patients. The clinical relevance of this end-point was further supported by the significant longer median hospital length of stay (20 days) in patients with a FRD on day 8.

Some limitations of this study should be noted. Inclusion criteria were based on the ASA physical status and the expected duration of surgery. Intraoperative management was not standardized; therefore, some patients may have benefited from more aggressive therapeutic intervention (Hawthorne phenomenon) influencing outcome. The operative categories between centers were not homogeneous and specific operations within categories varied between centers. A preponderance of a particular operation in one center might indicate a particular expertise for that surgery with a different rate of FRD than in other centers. We noted that patients undergoing surgery for peripheral vascular disease were at increased risk of postoperative complications compared with patients operated on for plastic or orthopedic back surgery. Therefore, we cannot exclude the possibility that surgical subgroups altered the prognostic value of Pr-etCO₂. Although the expected length of surgery was more than 2 hours, approximately 20% of patients had an operative time ranging between 25 and 120 min. Shorter surgical time in these patients may also have reduced the prognostic value of Pr-etCO₂. Finally, as with any study, there were some incomplete data sets (6 patients had some missing data) and some protocol violations, particularly with respect to timings of blood gas analyses (only 36% of blood gases taken within 30 min of the start of the surgery and 47% within 30 min of the end of surgery). The protocol violations with respect to blood gas analysis affect the interpretation of end-surgical pHi, end-surgical base deficit, and Pr-aCO₂ as predictors.

Despite these limitations, our study demonstrates maximum Pr-etCO₂ may be a useful prognostic index of postoperative morbidity in a high-risk surgical population with an expected duration of surgery more than 2 hours. These results need validation in another high-risk surgical population. Randomized clinical trials are needed to assess whether intervention to improve the intraoperative Pr-etCO₂ measured by an automated online tonometer can improve end organ perfusion and postoperative outcome in high-risk surgical patients.

	Footnotes

 
Tonometers and Tonocap monitors were provided for this study by Datex-Ohmeda Division, Instrumentarium Corp., Helsinki, Finland.

	References

Top Abstract Introduction Methods Results Discussion References

 

1. Hanley B, Truesdale A, King A et al. Involving consumers in designing, conducting, and interpreting randomised controlled trials: questionnaire survey. BMJ 2001; 322: 519–23.[Abstract/Free Full Text]
2. Bellomo R, Goldsmith D, Russell S, Uchino S. Postoperative serious adverse events in a teaching hospital: a prospective study. Med J Aust 2002; 176: 216–8.[Web of Science][Medline]
3. Vincent C, Neale G, Woloshynowych M. Adverse events in British hospitals: preliminary retrospective record review. BMJ 2001; 322: 517–9.[Abstract/Free Full Text]
4. Bennett-Guerrero E, Welsby I, Dunn TJ, et al. The use of a postoperative morbidity survey to evaluate patients with prolonged hospitalization after routine, moderate-risk, elective surgery. Anesth Analg 1999; 89: 514–9.[Abstract/Free Full Text]
5. Older P, Smith R, Courtney P, Hone R. Preoperative evaluation of cardiac failure and ischemia in elderly patients by cardiopulmonary exercise testing. Chest 1993; 104: 701–4.[Abstract/Free Full Text]
6. Mythen MG, Webb AR. Intraoperative gut mucosal hypoperfusion is associated with increased postoperative complications and cost. Intensive Care Med 1994; 20: 99–104.[Web of Science][Medline]
7. Mythen MG, Webb AR. The role of gut mucosal hypoperfusion in the pathogenesis of postoperative organ dysfunction. Intensive Care Med 1994; 20: 203–9.[Web of Science][Medline]
8. Baue AE. The role of the gut in the development of multiple organ dysfunction in cardiothoracic patients. Ann Thorac Surg 1993; 55: 822–9.[Abstract]
9. Soong CV, Halliday MI, Barclay GR, et al. Intramucosal acidosis and systemic host responses in abdominal aortic aneurysm surgery. Crit Care Med 1997; 25: 1472–9.[Web of Science][Medline]
10. Bennett-Guerrero E, Panah MH, Bodian CA, et al. Automated detection of gastric luminal partial pressure of carbon dioxide during cardiovascular surgery using the Tonocap. Anesthesiology 2000; 92: 38–45.[Web of Science][Medline]
11. Uusaro A, Lahtinen P, Parviainen I, Takala J. Gastric mucosal end-tidal PCO₂ difference as a continuous indicator of splanchnic perfusion. Br J Anaesth 2000; 85: 563–9.[Abstract/Free Full Text]
12. Lebuffe G, Onimus T, Vallet B. Gastric mucosal-to-end-tidal PCO₂ difference during major abdominal surgery: influence of the arterial-to-end-tidal PCO₂ difference? Eur J Anaesthesiol 2003; 20: 147–52.[Web of Science][Medline]
13. Lebuffe G, Decoene C, Pol A, et al. Regional capnometry with air-automated tonometry detects circulatory failure earlier than conventional hemodynamics after cardiac surgery. Anesth Analg 1999; 89: 1084–90.[Abstract/Free Full Text]
14. Shoemaker WC, Appel PL, Kram HB, et al. Prospective trial of supranormal values of survivors as therapeutic goals in high-risk surgical patients. Chest 1988; 94: 1176–86.[Abstract/Free Full Text]
15. Copeland GP, Jones D, Walters M. POSSUM: a scoring system for surgical audit. Br J Surg 1991; 78: 355–60.[Medline]
16. McNeil BJ, Keller E, Adelstein SJ. Primer on certain elements of medical decision making. N Engl J Med 1975; 293: 211–5.[Abstract]
17. Metz CE, Goodenough DJ, Rossmann K. Evaluation of receiver operating characteristic curve data in terms of information theory, with applications in radiography. Radiology 1973; 109: 297–303.[Web of Science][Medline]
18. Levy B, Gawalkiewicz P, Vallet B, et al. Gastric capnometry with air-automated tonometry predicts outcome in critically ill patients. Crit Care Med 2003; 31: 474–80.[Web of Science][Medline]
19. Schlichtig R, Bowles SA. Distinguishing between aerobic and anaerobic appearance of dissolved CO₂ in intestine during low flow. J Appl Physiol 1994; 76: 2443–51.[Abstract/Free Full Text]
20. Hamilton-Davies C, Mythen MG, Salmon JB, et al. Comparison of commonly used clinical indicators of hypovolaemia with gastrointestinal tonometry. Intensive Care Med 1997; 23: 276–81.[Web of Science][Medline]
21. Chang MC, Cheatham ML, Nelson LD, et al. Gastric tonometry supplements information provided by systemic indicators of oxygen transport. J Trauma 1994; 37: 488–94.[Web of Science][Medline]

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Email this article to a colleague Similar articles in this journal Similar articles in Web of Science Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Permissions Citing Articles Citing Articles via Web of Science (6) Citing Articles via Google Scholar Google Scholar Articles by Lebuffe, G. Articles by Webb, A. Search for Related Content PubMed PubMed Citation Articles by Lebuffe, G. Articles by Webb, A. Related Collections Economics and Health Care Research Postanesthetic Care Unit Monitoring (Non-cardiac)