INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Gastric tonometry measuring regional PCO₂ (Pr_CO2) of the gastric lumen has been considered to be a safe and noninvasive, organ-specific, metabolically oriented monitor of the adequacy of gut perfusion, oxygenation, and cellular energy balance (1, 2). Nevertheless, tonometry has yet not been widely accepted in clinical practice. This may relate to potential methodological pitfalls, serious doubts on the validity of intramucosal pH calculation, and finally substantial measurement errors, in particular with saline tonometry (2).

A semicontinuous automated air tonometry technique eliminates many of the technical problems inherent in the manual procedure, including the protracted equilibration time (2). Buffering of gastric acid by bicarbonate secreted by the stomach or entering from the duodenum or esophagus may extensively elevate Pr_CO2 (3, 4), thus still representing a most important source of error even with automated air tonometry. Consequently, the role of gastric juice pH and the value of histamine-2 (H₂)-receptor antagonists for reliable gastric tonometry is still a point of debate (5, 6). It has yet not been considered whether proton pump inhibitors might improve the validity of gastric tonometry determination.

Therefore we investigated the influence of changes in gastric juice pH due to pentagastrin and omeprazole administration on Pr_CO2 in healthy human volunteers, as determined by semicontinuous automated air tonometry.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Study Protocol

The institutional review board for human studies at the University Ulm approved the study protocol, and written informed consent was obtained from 17 healthy volunteers (male n = 11, female n = 6) after completing a medical history and a physical examination.

After an overnight fasting period of at least 10 h the volunteers received maintenance fluid infusion with 1 ml/kg/h lactated Ringers solution (Delta-Pharma, Pfullingen, Germany). All monitoring devices were placed under local anesthesia at least 30 min before baseline measurements, which were obtained between 8:00 A.M. and 9:00 A.M. Every 30 min we registered mean arterial pressure (MAP), heart rate (HR), arterial blood gases, and gastric juice pH and gastric Pr_CO2. To amplify gastric acid secretion we continuously administered 0.6 µg/ kg/h pentagastrin (Peptavlon; Cambridge Laboratory, New Castle, UK) intravenously for 1 h (9:00 A.M. to 10:00 A.M.). Gastric secretion usually begins within 10 min, peaks after 20-30 min, and lasts up to 1 h (7). After cessation of pentagastrin infusion another observation hour without pharmacological intervention followed (10:00 A.M. to 11:00 A.M.). To subsequentlly inhibit gastric acid secretion we employed a proton pump blockade with omeprazole. At 11:00 A.M. the volunteers received 40 mg omeprazole (Antra pro infusione, Astra Zeneca, Bad Homburg, Germany), which was administered as a short infusion over 30 min. The observation period finished at 3:00 P.M.

Gastric pH Measurement and Tonometry

Gastric juice pH was measured continuously using a combined glass electrode with an integral Ag/AgCl reference electrode near the tip (LOT 440; Medical Instruments Corporation, Solothum, Switzerland). The electrode was connected to a receiver (Gastrograph Mark II; Fresenius, Bad Homburg, Germany) for continuous long-term pH-metry. The glass electrode was introduced through the nose and advanced under direct pH measurement until the tip was 10 cm below the gastroesophageal junction, identified by an abrupt pH decrease as the probe entered the stomach.

A nasogastric tonometer tube (Trip NSG; Tonometrics, Helsinki, Finland) equipped with a semipermeable silastic balloon was also placed nasogastrically at the same distance from the nose as the pH electrode. Injecting air and auscultating a gurgling sound over the stomach as well confirmed the satisfactory placement of the tonometer. Subsequently the gastric tonometer was connected to an automated gas analyzer (Tonocap TC-200; Datex, Helsinki, Finland).

Hemodynamics and Arterial Blood Gas Analysis

The hemodynamic monitoring was carried out with an electrocardiographic system (Sirecust 404-1; Siemens AG, Erlangen, Germany) and a 20-G arterial line, which was inserted into the radial artery under local anesthesia. MAP was measured with disposable transducers (MX 8024-AS; Medex Medical, St. Crispin, UK). Moreover, arterial blood gas values for PO₂, PCO₂, base excess, bicarbonate, and pH were determined by using an automated analyzer (Stat 5; Nova Biomedical, Waltham, MA).

Statistics

The data were analyzed using a personal computer and commercially available software (Sat View 5.0; Abacus concept Inc., Berkeley, CA). Data are given as mean ± standard deviation (SD). We tested one principal null hypotheses: There are no differences in gastric juice pH, Pr_CO2, and all other measured variables over the time. For all data conflicting with normal distribution we employed log (x) transformation before statistical analysis. Differences over the time were examined with a paired Student's t test with Bonferroni correction (' < 0.0071, '' < 0.0045), respectively. Significant differences were assumed with an  error of < 0.05.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

All studied volunteers had stable hemodynamics (MAP, HR), and arterial blood gases (pHa, Pa_O2, Pa_CO2, and bicarbonate) throughout the whole observation period between 8:00 A.M. and 3:00 P.M.

Gastric Juice pH

With mean gastric juice pH values of 1.2 ± 0.4 the volunteers demonstrated a normal acid secretion behavior during baseline measurements. The pentagastrin stimulation of acid secretion resulted in an immediate decrease of gastric juice pH to a mean peak value of 0.6 ± 0.3 at 10:00 A.M., which returned to baseline values 1 h after cessation of the pentagastrin infusion. Omeprazole administration subsequently produced a marked increase in gastric juice pH to a mean peak value of 4.4 ± 1.7. Toward the end of the observation period the gastric juice pH again decreased to almost baseline values (Figure 1).

View larger version (18K): [in this window] [in a new window]   Figure 1.   Gastric juice pH (lower panel ) and regional to arterial PCO2 gradient (Pr-aCO2, upper panel ) during baseline (8:00-9:00) as well as following pentagastrin (9:00-11:00) and omeprazole (12:00- 15:00) administration in 17 healthy volunteers. Data are given mean ± SD. *p < 0.007 versus baseline 8:00; p < 0.002 versus pentagastrin 11:00.

Regional PCO₂

The regional to arterial gradient of PCO₂ (Pr-a_CO2) (Figure 1) and the gastric lumen Pr_CO2 demonstrated corresponding changes during the whole observation period. Already under baseline conditions we observed a great interindividual variability of Pr_CO2 with values ranging from 33 to 74 mm Hg in the healthy human volunteers. This variability of Pr_CO2 even significantly increased after pentagastrin administration. We found mean peak values 61 ± 17 mm Hg 1 h after cessation of the pentagastrin infusion. Following omeprazole administration we confirmed a substantial decrease in Pr_CO2. Two hours after omeprazole we detected the lowest Pr_CO2 (44 ± 5 mm Hg) along with a marked reduction of individual deviation. Toward the end of the observation period the Pr_CO2 tended to increase again.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

To the best of our knowledge this is the first study showing the subsequent effects of pentagastrin and omeprazole administration on gastric Pr_CO2 measurements in healthy human volunteers. The principal findings are as follows:

1. Changes in gastric juice pH are in fact a potential source of error.

2. Omeprazole significantly decreases gastric Pr_CO2 along with a marked reduction of the interindividual variability in healthy volunteers.

In young healthy humans the PCO₂ of the gastric mucosa should more or less equal blood PCO₂ (4), but we already measured a considerable interindividual variability of tonometer to blood PCO₂ gradients from 8 to +31 mm Hg under baseline conditions. Normal Pr-a_CO2 values amount to about 9 mm Hg (8, 9). In contrast to theoretical considerations (1), three volunteers demonstrated a constant negative Pr-a_CO2 gradient, which most likely represents a measurement artifact due to swallowing environmental air (2). Stimulating acid secretion with pentagastrin resulted in a significant increase in gastric PrCO₂, which occurred despite a well-known dose-dependent rise in gastric mucosal blood flow following pentagastrin administration (10). Augmentation of regional blood flow normally decreases rather than increases Pr_CO2 (1). Therefore both the high Pr_CO2 values at baseline and the significant increase following pentagastrin most likely indicate generation of intraluminal PCO₂. This is probably due to buffering of normal or stimulated acid load by salivary, duodenal, or gastric bicarbonate, independently from mucosal blood flow and transmucosal CO₂ diffusion (3, 4, 11). In addition, the diffusion of CO₂ down its concentration gradient from luminal fluid into tissue is rather slow, thus allowing generation of a relatively large PCO₂ gradient across the mucosa. (11). Finally, buffering of gastric acid by bicarbonate in the first part of the duodenum, followed by reflux of carbon dioxide through the pylorus, must as well be considered (4), as PCO₂ values of up to 500 mm Hg in the duodenal bulb have been recorded in humans (12).

Intraluminal CO₂ generation is further supported by the fact that the transient blockade of acid secretion following omeprazole produced a significant decrease of Pr_CO2 along with a marked reduction of the individual deviation. While abolishing acid and inhibiting bicarbonate secretion, the proton pump inhibitor omeprazole maintains basal gastric mucosal blood flow (10), in contrast to H₂ blockers that decrease both acid secretion as well as mucosal perfusion (13). The antisecretory effect of omeprazole was merely transient (2-3 h) in our study. In fact recent data support the finding that inhibition of acid secretion by a continuous infusion is superior to single injections of omeprazole and as well seemingly more effective as a continuous infusion of ranitidine (14). Our data confirm previous findings with ranitidine in healthy humans (3, 4, 8) and in perioperative patients (15) that demonstrated that inhibition of gastric acid secretion prevented the formation of CO₂ from bicarbonate and hydrogen ions thereby improving accuracy and precision of the measurement of gastric Pr_CO2. The present study extends former results by analyzing the effects of both pentagastrin stimulation and subsequent proton pump inhibition as well as by using semicontinuous automated air tonometry for gastric Pr_CO2 determination.

It can be questioned whether the data presented in healthy subjects apply to critically ill patients, as ranitidine had no effect on Pr_CO2 in seriously ill patients (6). This study, however, has specific limitations, including methodological drawbacks due to the use of semiquantitative gastric luminal pH measurement (16) and saline tonometry (1, 2). Moreover, baseline gastric pH in the former study appears to be rather high3.9 in the ranitidine group and 3.7 in the placebo group. Gastric juice pH is definitely unpredictable in perioperative (15) or in the critically ill (17), inasmuch as some patients may have an impaired ability to acidify the gastric lumen, whereas others do in fact respond to physiological stimuli such as pentagastrin (5). In patients with severe sepsis, polytrauma, and severe head injury Wiedeck and coworkers (17) recently demonstrated gastric pH between 1.8 and 6.2 in untreated patients as compared with 5.5-7.2 in omeprazole (2 × 40 mg/24 h)-treated patients.

Costs of care together with the concept that inhibition of acid secretion predisposes to nosocomial infections is probably the reason that most authors do not use H₂ or proton pump blockers before tonometry in critically ill patients (1, 2). An increase in Pr_CO2 without variation in gastric juice pH (pH > 4-5) likely indicates compromised mucosal perfusion, whereas alterations in Pr_CO2 along with an acid or changing gastric juice pH could as well indicate buffering of gastric acid by bicarbonate and thus luminal CO₂ generation (2). Hence gastric juice pH measurements appear to be mandatory for reliably interpreting gastric tonometry data. The administration of ranitidine or omeprazole, which both certainly do improve the bias and precision of gastric tonometry, may not render gastric pH measurements superfluous (2), because patients may vary in their response and secretion inhibition may only be transient (14).

In conclusion, we demonstrated that changes in gastric acid secretion due to pentagastrin and omeprazole might significantly affect gastric PCO₂ tonometry thereby remarkably confounding adequate interpretation. Administration of omeprazole reduces mean gastric Pr_CO2 and the individual variability in healthy human volunteers.