Introduction

The intestinal tract is highly susceptible to hypoperfusion because of its greater level of critical oxygen delivery and countercurrent microcirculation of the villi [1]. There is increasing evidence that gastrointestinal hypoperfusion plays an important role in development of systemic inflammatory response and multiple organ failure [1,2]. Decreased splanchnic perfusion precedes the appearance of the usual indicators of hypovolemic shock, such as hypotension and lactic acidosis [3,4,5]. Gastric intramucosal acidosis and hypercapnia are observed during inadequate organ perfusion [6,7,8] and are predictive of poor clinical outcome [9,10,11]. Therefore, early detection of gastrointestinal hypoperfusion and effective treatment may improve clinical outcome. Because gastric intubation is done in most critically ill patients, gastric tonometry has traditionally been used to evaluate intramucosal pH or partial carbon dioxide tension (PCO2) indirectly during the management of critically ill patients [9,10,11,12,13,14,15]. However, gastric tonometers have some limitations. For example, air and saline tonometers may require 10–90 min for equilibration [16,17,18,19]. Reliable gastric tonome-try requires suppression of gastric acid [20], whereas gastric feedings can influence its outcome [21,22,23]. Therefore, several other sites, including esophagus, have been used for tonometric measurements [8,24,25,26,27].

Studies have demonstrated that an increase in veno–arterial PCO2 gradient could be a reliable marker of tissue hypoperfusion [28,29,30,31,32]. Knichwitz and coworkers [25] demonstrated that continuous intramucosal PCO2 measurement allows early detection of regional intestinal ischemia before the onset of changes in global hemodynamic or metabolic variables. Furthermore, measurement of tissue PCO2 in several organs has been shown to correlate with gastrointestinal perfusion [8,26,27,33]. Sato and coworkers [8] studied the relationship between gastric wall PCO2 and esophageal PCO2 (PeCO2) before, during, and after reversal of hemorrhagic shock in five spontaneously breathing rats, using an ion-sensitive field-effect transistor. They found a high correlation (r = 0.9) between the gastric wall PCO2 and PeCO2 during hemorrhagic hypotension induced reduction in splanchnic blood flow. The use of tissue PCO2 and arterial PCO2 (PaCO2) difference is a better marker of ischemia than is either gastric intra-mucosal pH or intramucosal PCO2 [34] because the gap is not influenced by alveolar ventilation [35]. Therefore, in the present study we measured intraluminal PeCO2 using a rapidly responsive fiberoptic sensor [25,35,36]. The arterial blood gases were periodically measured for determination of the PeCO2–PaCO2 gap. Our hypothesis is that the PeCO2–PaCO2 gap could be significantly increased during graded hemorrhagic hypotension and will return to baseline shortly after resuscitation.

Materials and method

Surgical procedures

The experimental protocol for the present study was approved by the Institutional Animal Care and Use Committee of Miami Children's Hospital. Ten young, albino Sprague–Dawley rats (250–350 g) were initially anesthetized with 60 mg/kg pentobarbital intraperitoneally. In a supine position, a tracheostomy was performed and an endotracheal tube (3.5 cm of a polyethylene tube, 2.4 mm diameter) was advanced to a position approximately 1 cm above the carina. Subsequently, a femoral vein and a femoral artery were exposed and cannulated. Each rat then was placed over an electric heating blanket. Rectal temperature (TH-5; Physitemp Thermalert, Clifton, NJ, USA; with a rat size thermal probe), mean arterial blood pressure (MABP), and heart rate (HR; 2001A, Datascope Corp, Paramus, NJ, USA) were continuously monitored. Normothermia (38 ± 0.5°C) was established while anesthesia (pentobarbital 15 mg/kg per hour) and fluid balance (saline 5 ml/kg per hour) were strictly maintained (Medfusion pump 2010; Medex, Duluth, CA, USA). Rats breathed room air, spontaneously, during the experiments.

Esophageal capnometry

The esophagus was intubated orally with a 22-gauge, 1.5-inch-long catheter. A fiberoptic carbon dioxide sensor (Paratrend7; Diametrics Medical Inc, Roseville, MN, USA) was introduced through the oral catheter up to 8–10 cm from the incisor teeth into lower third of the esophagus (at 2–3 cm above the gastroesophageal junction). The fiberoptic sensor consisted of two optical fibers for the measurement of PCO2 and pH, a miniature Clark electrode for determination of partial oxygen tension, and a thermocouple for measuring temperature. The sensor was automatically calibrated with precision gases under microprocessor control, as per the manufacturer's recommendations, before insertion into the esophagus.

Baseline measurements

Within 30–60 min after the insertion of the sensor, baseline values for PeCO2, core temperature, HR, and MABP were recorded. The rats then were heparinized with 200 U/kg per hour heparin and an arterial blood sample was taken for baseline (time0) gas analysis (ABL-30 Blood Gas Analyzer; Radiometer, Copenhagen, Denmark), hemoglobin, and arterial oxygen saturation (OSM3 Hemoxymeter; Radiometer). Measurements of PaCO2 and PeCO2, as well as partial arterial oxygen tension, were corrected for each animal's body temperature. Values for bicarbonate and base excess were automatically calculated by the blood gas analyzer's program.

Hemorrhagic hypotension

Three out of the 10 rats were used to test the limits of hemodynamic stability during hemorrhagic hypotension in this model. Gradual bleeding up to 15 ml/kg in these three rats led to a 30–40% reduction in MABP. Additional bleeding up to 25 ml/kg was tolerated as long as the MABP did not drop below 30 mmHg. Lower blood pressures, caused by removal of 25 ml/kg blood, created a deteriorating and irreversible systemic hypotension, accompanied by severe tachycardia. Therefore, in the actual experiments (n = 7) we considered 15 ml/kg bleeding over a 30-min period as mild hemorrhagic hypotension. Removal of 20–25 ml/kg blood, while maintaining a MABP equal to or higher than 35 mmHg, was considered severe hemorrhagic hypotension. The blood was collected in a heparinized (400 U) tube and incubated at 38°C. Thirty minutes after mild hemorrhagic hypotension, all the baseline variables were again measured. This procedure was repeated after removal of another 5–10 ml/kg blood (for generation of severe but reversible hemorrhage). All variables were recorded during severe hemorrhagic hypotension, and then the shed blood was reinfused over 20–30 min. All variables were measured again at 30 and 60 min following termination of blood reinfusion. At the end of the experiment, the animals were killed with intravenous pentobarbital and the exact position of the esophageal probe was verified.

Statistical analysis

Statistical evaluation was performed in the seven rats that completed mild and severe hemorrhage with resuscitation. All variables are presented as mean ± SD. The data were computed by repeated measures of analysis of variance followed by Dunnett multiple comparisons test, using the baseline values as controls. A linear regression analysis was also performed to evaluate association between PeCO2–PaCO2 gap and the base deficit. P < 0.05 was considered statistically significant.

Results

Hemodynamic and gas exchange variables

Mild and severe homorrhagic hypotension created average reductions of 33% and 53% in MABP, respectively. Reinfusion of the blood restored MABP to the normal range. Blood hemoglobin concentration followed a pattern similar to that of blood pressure (Table 1). The HR was significantly increased following severe hemorrhage (29%). After blood reinfusion, the HR remained significantly higher than its prehemorrhagic baseline value (Table 1). The partial arterial oxygen tension was increased significantly during both mild and severe hemorrhagic hypotension, apparently caused by hyperventilation. The latter also reduced the PaCO2 significantly (Fig. 1). Arterial saturation following blood reinfusion was not significantly different from baseline. Blood bicarbonate concentrations decreased significantly following hemorrhage, but recovery was not complete at 60 min after blood reinfusion (Table 1).

Figure 1
figure 1

Changes in partial arterial carbon dioxide tension (PaCO2), partial esophageal carbon dioxide tension (PeCO2) and esophageal–arterial PCO2 gap in seven anesthetized, spontaneously breathing rats subjected to mild and severe hemorrhagic hypotension followed by blood reinfusion. *P < 0.05, by repeated measures of analysis of variance followed by Dunnett multiple comparison test, using baseline as controls.

Full size image
Table 1 Gas exchange variables, partial esophageal carbon dioxide tension, and hemodynamic variables during mild and severe hemorrhagic hypotension and following blood reinfusion in anesthetized, spontaneously breathing rats
Full size table

Esophageal–arterial partial carbon dioxide tension gap and base deficit

The PeCO2–PaPCO2 was significantly increased after mild and severe hemorrhage, and returned to baseline values following blood reinfusion (Fig. 1). The base deficit became slightly more negative after mild hemorrhage but was significantly reduced after severe hemorrhage (-5.5 mmol/l and -14.4 mmol/l, respectively). The base deficit remained significantly high after blood reinfusion (-7.2 mmol/l after 60 min). After blood reinfusion, unlike base deficit, the PaCO2 rapidly normalized (Table 1). A significant correlation was found between base deficit and PeCO2–PaCO2 gap during hemorrhagic hypotension (Fig. 2; r2 = 0.39, P < 0.002). At the same time, there was also a significant correlation between base deficit and PeCO2 (Fig. 3; r2 = 0.24, P < 0.022).

Figure 2
figure 2

Linear regression analysis of the association between partial esophageal carbon dioxide tension (PeCO2) minus partial arterial carbon dioxide tension (PaCO2; i.e. PeCO2–PaCO2 gap) and base deficit in seven anesthetized, spontaneously breathing rats during mild and severe hemorrhagic hypotension. Broken lines represent the upper and lower limits of 95% confidence interval.

Full size image
Figure 3
figure 3

Linear regression analysis of the association between partial esophageal carbon dioxide tension (PeCO2) and base deficit in seven anesthetized, spontaneously breathing rats during mild and severe hemorrhagic hypotension. Dotted lines represent the upper and lower limits of 95% confidence interval.

Full size image

Discussion

A correlation between PeCO2 and gastric PCO2 during hemorrhagic shock was previously demonstrated in spontaneously breathing rats [8]. Our results, using a fiberoptic carbon dioxide sensor, are generally in agreement with those of Sato and coworkers [8], who used an ion-sensitive field-effect transistor sensor. In the present study, unlike that of Sato and coworkers, PeCO2 did not significantly increase during hemorrhage, whereas the PeCO2–PaCO2 gap was significantly increased. The PeCO2–PaCO2 gap returned to baseline immediately after resuscitation (Fig. 1). Our data also demonstrate a significant association between the PeCO2–PaCO2 gap and the corresponding base deficit that occurred during hemorrhagic hypotension (Fig. 2). Whereas the PeCO2–PaCO2 gap rapidly recovered after resuscitation (Fig. 1), the base deficit did not completely return to baseline after restoration of blood volume (Table 1).

The animals in our study hyperventilated because of metabolic acidosis, presumably secondary to hypoperfusion. Arterial hypocapnia can impact on the expected rise in tissue PCO2 that occurs as a result of decreased tissue perfusion. Therefore, intramucosal PCO2 as an indicator of tissue hypoperfusion is not as accurate as PeCO2–PaCO2 [34]. Moreover, the tissue PCO2 and PaCO2 gap is not influenced by alveolar ventilation [37]. However, when ventilation is controlled, the change in tissue PCO2 by itself could become a reliable indicator of tissue perfusion. In our spontaneously breathing rats the PeCO2 was lower after severe hemorrhage. We reason that the PeCO2 would have been higher if the rats were mechanically ventilated to maintain a relative arterial normo-capnia. In ventilated subjects, change in tissue PCO2 is an indicator of changes in tissue perfusion before any other global parameters of perfusion are changed [25,38]. In spontaneously breathing subjects, continuous measurements of tissue PCO2 and PaCO2 gap can be used as an early indicator of tissue hypoperfusion.

Gastric tonometry versus esophageal and sublingual capnometry

Traditionally, stomach has been used as the organ to measure intramucosal pH or PCO2 in both animal and human studies [6,7,8,9,10,11,12,13,14,15]. The low pH of stomach may interfere with tonometry, and therefore gastric acid suppression may be needed for reliable measurements [20]. Other limiting factors in gastric tonometry are related to feeding [22,23] and the large lumen of the stomach, requiring longer time for intralu-minal contents to equilibrate with intramucosal PCO2. Moreover, in the presence of low gastric pH, secretion of bicarbonate leads to intraluminal production of carbon dioxide [39]. The above factors may prevent rapid detection of changes in intramucosal PCO2. Therefore, several other sites have been used for tonometry. In animals, ileum has been used to assess splanchnic perfusion [25,36] – a clinically impractical procedure. Studies have demonstrated that sublingual capnometry, a relatively noninvasive procedure, correlates with gastric tonometry [26,40,41,42]. Practically, it may be difficult to lodge the sensor securely under the tongue in uncooperative patients, thereby preventing equilibration with tissue PCO2 [27]. Esophageal intubation, which is commonly used in critically ill patients, can be utilized to secure placement of the esophageal sensor. Similar to the gastric environment, bicarbonate is secreted in the esophagus and may enter the esophagus from salivary secretions. However, a relative alkaline pH in the esophagus, in the absence of acid reflux, may not lead to generation of a significant amount of carbon dioxide. Currently available tonometers have equilibration periods ranging between 10 and 90 min [16,17,18,19] and are therefore not efficient for rapid detection of changes in tissue perfusion on a continuous basis. Fiberoptic sensors that are used in clinical medicine for automatic and continuous measurements of blood gases [43,44] have a rapid response time [45]. Experimental evaluation of a fiberoptic PCO2 sensor, similar to that used in the present study, has shown a high degree of precision in detecting short-term changes in intramucosal PCO2 [35].

Capnometry and end-points of resuscitation

An interesting observation in the present study was the delayed recovery of base deficit after resuscitation (Table 1), whereas PeCO2, PaCO2, and the gap between them were actually recovered (Fig. 1). Porter and Ivatury [46] demonstrated that the use of base deficit, lactate, and/or gastric intramucosal pH are appropriate end-points of resuscitation for trauma patients. They also recommended that one or all of the above markers of tissue perfusion be corrected to normal range within 24 hours after injury. Povoas and coworkers [42] reported persistently elevated blood lactate level after reinfusion of blood when all other parameters of tissue perfusion, such as sublingual PCO2, gastric PCO2, and veno–arterial PCO2 gradient, were normalized. In the present study, the delay in normalization of the base deficit in the face of a rapid normalization of the PeCO2–PaCO2 gap may suggest that the PeCO2–PaCO2 gap can serve as an early indicator for resuscitation end-point rather than base deficit. Physiologically, it takes time for liver and kidneys to correct metabolic acidosis following tissue dysoxia. It is therefore anticipated that there will be a lag phase between restoration of blood volume and return of base deficit to normal.

Studies indicate that PeCO2–PaCO2 gap can continue to increase or remain abnormally high after resuscitation [25,47,48]. In those experiments [47,48], severe hemorrhage (45–47 ml/kg versus 30 ml/kg) might have contributed to ischemia/reperfusion injury, leading to persistent mucosal hypoperfusion and elevated tissue PCO2–PaCO2 gap. In the presence of ischemia/reperfusion mucosal injury, the PeCO2–PaCO2 gap may not return to normal even after restoration of circulatory volume. In such instances, base deficit (or other global parameters of tissue perfusion) may be a better index for the end-point of resuscitation.

Conclusion

The data presented here demonstrate that PeCO2–PaCO2 gap increases during hemorrhagic hypotension and reverses after resuscitation, without complete recovery of base deficit. We suggest that esophageal capnometry could be used as an alternative to gastric tonometry for assessing splanchnic hypoperfusion.

Key messages

  • Esophageal capnometry could be used as an alternative for gastric tonometry during the management of hypovolemic shock

  • PeCO2–PaCO2 gap increases during graded hemorrhagic hypotension and returns to baseline value after resuscitation, without complete reversal of the base deficit