|
|
||||||||
Guest Access | Sign In via User Name/Password |
|||||||||
* From Critical Care Medicine, The Mercy Hospital of Pittsburgh, Pittsburgh, PA.
Correspondence to: Paul Marik, MD, FCCP, Critical Care Medicine, Mercy Hospital of Pittsburgh, 1400 Locust St, Pittsburgh, PA 15219-5166; e-mail: pmarik{at}zbzoom.net
| Abstract |
|---|
|
|
|---|
Design: A prospective, validation study.
Setting: The medical and coronary ICUs of a community teaching hospital.
Patients: Consecutive patients with severe sepsis, septic shock, or cardiogenic shock requiring pulmonary artery catheterization for hemodynamic management.
Interventions: During the first 24 h of ICU admission, the PslCO2, PimCO2, and blood lactate concentrations as well conventional hemodynamic and oxygenation parameters were recorded every 4 to 6 h. The PslCO2-PaCO2 and PimCO2-PaCO2 differences were used as indexes of tissue dysoxia. These variables were correlated with each other as well as with the traditional markers of tissue oxygenation.
Results: Seventy-six data sets were obtained on 22 patients. Fifteen patients had severe sepsis/septic shock, and 7 patients did not have sepsis. A patient with ischemic bowel who had a large PimCO2-PslCO2 difference (60.2 mm Hg) was excluded. The initial PslCO2 and PimCO2 measurements were 43.5 ± 10.4 mm Hg and 42.8 ± 10.9 mm Hg, respectively (correlation coefficient [r] of 0.86; p < 0.001). The mean PslCO2 and PimCO2 for the entire data set were 48.0 ± 13.4 mm Hg and 46.1 ± 12.3 mm Hg, respectively (r = 0.78; p < 0.001). Ten patients died. The initial PslCO2-PaCO2 difference was 9.2 ± 5.0 mm Hg in the survivors and 17.8 ± 11.5 mm Hg in the nonsurvivors (p = 0.04). The initial PimCO2-PaCO2 difference was 8.4 ± 4.8 mm Hg in the survivors and 16.1 ± 13.7 mm Hg in the nonsurvivors (p = 0.08, not significant). The initial PslCO2-PaCO2 difference correlated with the initial mixed venous-arterial CO2 gradient (r = 0.66; p = 0.001), but correlated poorly with the initial blood lactate concentration (r = 0.38), mixed venous PO2 (r = 0.05), and systemic oxygen delivery (r = - 0.39).
Conclusion: In this study, sublingual capnometry yielded measurements that correlated well with those of gastric tonometry. PslCO2 may serve as a technically simple and noninvasive clinical measurement of tissue dysoxia in critically ill and injured patients.
Key Words: gastric tonometry intramucosal hypercarbia lactate pulmonary artery catheter sublingual capnography tissue oxygenation
| Introduction |
|---|
|
|
|---|
Tissue dysoxia, however, is exceedingly difficult to detect at the bedside; there are no specific clinical signs and no simple laboratory tests. Global indexes of oxygen delivery (DO2) and oxygen consumption provide no useful information as to the adequacy of tissue oxygenation.4 5 7 Blood lactate concentration and mixed venous oxygen tension (PmvO2) are crude and insensitive indexes of global oxygen balance.4 The limitation of traditional ICU monitoring devices has driven the development of organ-specific monitors.
Gastric tonometry is the only organ-specific monitor of tissue dysoxia currently available. Gastric tonometry is based on the principle that tissue production of CO2 rises sharply with tissue dysoxia.8 9 10 11 Present evidence indicates that the source of increased tissue CO2 is intracellular buffering of excesses of hydrogen ions by bicarbonate. The excesses of hydrogen ions are, in turn, traced to anaerobic generation of excesses of lactic acid and degeneration of high-energy phosphate compounds during tissue dysoxia. Tissue PCO2 tension is therefore an index of tissue energy status; tissue dysoxia whether due to decreased oxygen availability (hypoxic, anemic, or stagnant hypoxia) or diminished ability to utilize oxygen (cytopathic hypoxia) results in an increase in tissue PCO2.8 9 10
Gastric tonometry is a monitor of gastric mucosal hypoxia. Measurement of the adequacy of the splanchnic circulation may be particularly important in the critically ill or injured patient. Splanchnic hypoperfusion occurs early in shock, and may occur before the usual indicators of shock such as hypotension or lactic acidosis are present.11 12 13 Furthermore, the gut is highly susceptible to diminished tissue perfusion and oxygenation as it has a higher critical DO2 than the whole body and other vital organs, and the mucosal counter-current microcirculation renders the villi particularly vulnerable to ischemia.11 14 Researchers15 16 17 18 have demonstrated that the very proximal GI tract, namely, the tongue and/or sublingual mucosa, may serve as appropriate sites for measurement of tissue PCO2. These authors15 16 17 18 have demonstrated an increase in sublingual PCO2 (PslCO2) that was closely related to decreases in arterial pressure and cardiac index during circulatory shock produced by hemorrhage and sepsis.
The aim of this study was to compare PslCO2 with gastric intramucosal PCO2 (PimCO2) as well as with the "traditional" indexes of tissue oxygenation in hemodynamically unstable ICU patients.
| Materials and Methods |
|---|
|
|
|---|
Interventions and Measurements
After insertion of a pulmonary artery catheter (Abbott 7F
Thermodilution Catheter; Abbott Critical Care Systems; North Chicago,
IL) and after informed consent was obtained, a nasogastric tonometer
(16F Trip NGS-TC Catheter; Instrumentation Corporation; Helsinki
Finland) was inserted and its position confirmed radiologically. After
the tonometer was in place and the monitor (Tonocap; Datex
Instrumentation; Helsinki, Finland) had cycled at least three times, a
baseline data set was obtained. The data set was repeated every 4 to
6 h within the first 24 h of the ICU admission. All patients
were pretreated with IV histamine type-2 blockers, and tube feeds were
withheld for the duration of the study period.20
Each data set included hemoglobin and arterial lactate concentration, core body temperature, arterial and mixed venous blood gas values (IL 1620 Blood Gas Analyzer and IL 682 Co-Oximeter; Instrumentation Laboratory; Lexington, MA) as well as the measurement of PimCO2 and PslCO2. Hemoglobin concentration was measured using a Coulter analyzer (Coulter Electronics; Hialeah, FL). Arterial blood specimens for lactate determination were analyzed using an enzymatic method (2300 Stat Analyzer; YSI; Yellow-Spring, OH). The normal range of plasma lactate for our laboratory is 1.2 to 2.2 mmol/L.
Tonometric Measurements: The PimCO2 was measured with an automated tonometer system (Tonocap; Datex Instrumentation). The Silastic balloon of the tonometric catheter is automatically inflated in a closed circuit, and the intraluminal PCO2 is measured by infrared spectroscopy every 15 min. The PimCO2-PaCO2 gradient was determined by subtracting the PaCO2 from the simultaneously measured PimCO2.
PslCO2 Measurements: The PslCO2 was measured using a sublingual measurement device (CapnoProbe; Optical Sensors; Minneapolis, MN), which consists of the following major components: (1) a disposable PCO2 sensor, (2) a fiberoptic cable that connects the disposable sensor to the instrument, and (3) a blood gas monitoring instrument with software modifications. The disposable PCO2 sensor is a fiberoptic technology-based sensor. An optical fiber is terminated with a silicone membrane that contains a fluorescent dye that is sensitive to CO2 concentration. The silicone membrane is permeable to CO2 gas. CO2 passes through the silicone membrane and comes in contact with the fluorescent dye. When the dye is presented with a light from the instrument, the dye fluoresces and emits light in direct correlation to the amount of CO2 present. Light signals are converted to numeric PCO2 values that are displayed on the instrument. The PslCO2 measurements are performed by placing the disposable sensor under the tongue (much like taking an oral temperature). The equilibrated PslCO2 measurement was recorded (2- to 4-min equilibration time).
Data Collection and Analysis
The patients demographic, clinical, and laboratory data were
recorded in an electronic database (Access 2000; Microsoft; Redmond,
WA). At the end of the data collection, summary statistics were
compiled to allow a description of the patient population. Statistical
analysis was done using software (NCSS 2000; NCSS Statistical Software;
Kaysville, UT). The bias and precision between the
PslCO2 and
PimCO2 measurements were analyzed by
the method described by Bland and Altman.21
This method
describes the limits of agreement by defining the mean difference and
SDs of the difference between the two techniques. The limits of
agreement are defined as the mean difference ± 2 SD of the
difference. Correlations between the variables of interest were
performed using Pearsons product moment correlation. Continuous data
were compared using Students t test. Unless otherwise
stated, all data are expressed as mean (with SDs) with statistical
significance declared for probability values of
0.05.
| Results |
|---|
|
|
|---|
In the patient with ischemic bowel, there was a large discrepancy between the mean PimCO2 and mean PslCO2 (101 mm Hg vs 42 mm Hg); this patients data were excluded from the data set. The initial PslCO2 and PimCO2 measurement was 43.5 ± 10.4 mm Hg and 42.8 ± 10.9 mm Hg, respectively (correlation coefficient [r] of 0.86; p < 0.001). The mean PslCO2 and PimCO2 for the entire data set was 48.0 ± 13.4 mm Hg and 46.1 ± 12.3 mm Hg, respectively (r = 0.78; p < 0.001) The scatter plot of the simultaneously measured PslCO2 and PimCO2 is shown in Figure 1 . The bias between the two methods was 2.1 mm Hg, and the limits of agreement were 16.1 to - 11.9 mm Hg (Fig 2 ). The initial lactate concentration was 3.7 ± 3.1 mmol/L, the PmvO2 was 37.3 ± 9.4 mm Hg, and the mixed-venous arterial CO2 difference was 6.7 ± 6.1 mm Hg. Ten patients died. The initial PslCO2 was 42.4 ± 7.6 mm Hg in the survivors and 45.0 ± 13.7 mm Hg in the nonsurvivors (not significant [NS]). Similarly, the initial PimCO2 was 41.5 ± 5.9 mm Hg in the survivors and 43.3 ± 15.1 mm Hg in the nonsurvivors (NS). The initial PslCO2-PaCO2 gradient was 9.2 ± 5.0 mm Hg in the survivors and 17.8 ± 11.5 mm Hg in the nonsurvivors (p = 0.04). The initial PimCO2-PaCO2 gradient was 8.4 ± 4.8 mm Hg in the survivors compared to 16.1 ± 13.7 mm Hg in the nonsurvivors (p = 0.08, NS). The initial PslCO2-PaCO2 difference correlated with the initial mixed venous-arterial CO2 gradient (r = 0.66; p = 0.001); however, it correlated poorly with the initial lactate concentration (r = 0.38, NS), mixed venous PO2 (r = 0.05, NS), and systemic DO2 (r = -0.39, NS).
|
|
| Discussion |
|---|
|
|
|---|
Decreased tissue oxygen tension (hypoxic hypoxia), decreased microcapillary blood flow (stagnant hypoxia), or an inability to utilize oxygen (cytopathic hypoxia) will result in increased tissue PCO2 tension. However, since tissue PCO2 concentration is dependent on the PCO2 of the perfusing arterial blood, as well as tissue CO2 production and removal, tissue PCO2 must be interpreted in conjunction with an arterial blood gas analysis. This is important, as an anion-gap metabolic acidosis usually accompanies tissue dysoxia and this will increase minute ventilation and CO2 elimination in an attempt to correct arterial pH. Furthermore, hyperventilation is an early and characteristic feature of sepsis.31 In an experimental model, Pernat et al15 investigated the effects of hyperventilation and hypoventilation on PimCO2 and PslCO2 before, during, and after reversal of hemorrhagic shock. These authors15 demonstrated that the PimCO2 and PslCO2 varied directly with changes in arterial PCO2. These findings are supported by our study: the PslCO2-PaCO2 gradient and not the PslCO2 predicted outcome.
An interesting observation in this study was the poor correlation between the initial PslCO2-PaCO2 gradient and the conventional markers of tissue dysoxia, namely, arterial lactate concentration and PmvO2. Hyperlactemia may occur in situations in which tissues are well perfused and in the absence of tissue dysoxia. Hyperlactemia commonly occurs in critical illnesses associated with hypermetabolic states, such as sepsis, burns, and trauma. The hyperlactemia in these conditions usually occurs in association with a normal lactate:pyruvate ratio and little evidence of a tissue oxygen debt.32 33 34 35 36 37 38 Furthermore, the interpretation of blood lactate concentration is complicated by additional factors: (1) the blood lactate concentration depends on the balance between tissue lactate production and hepatic removal. The liver has a large capacity for lactate removal, and therefore lactate production has to be substantially increased before the metabolic threshold of the liver is exceeded, and increased blood concentrations occur. Therefore, tissue dysoxia may be present despite a normal lactate concentration, (2) due to delayed hepatic clearance of lactate, tissue hypoxia may have resolved, yet the blood lactate concentration may remain elevated; and (3) other conditions interfere with lactate production, so that despite tissue hypoxia, blood lactate concentration will be normal. In severe malnutrition, for example, glucose stores are insufficient to sustain glycolysis. Similarly, the interpretation of PmvO2 is fraught with difficulties. PmvO2 monitoring is based on the principle that inadequate tissue DO2 will be associated with a low PmvO2. However, PmvO2/saturation monitoring has a number of major limitations, including (1) it neither measures nor reflects tissue PO2, (2) there appears to be no clear-cut threshold that defines inadequate DO2 and tissue hypoxia, (3) patients with cytopathic hypoxia may have an elevated PmvO2, (4) PmvO2 correlates poorly with cardiac output and DO2, and (5) PmvO2 is insensitive to regional desaturation in organs whose venous blood comprises a small percentage of total mixed venous blood (eg, brain and intestine). It is noteworthy that the initial PslCO2-PaCO2 gradient correlated with the initial mixed venous-arterial CO2 gradient. Researchers39 40 41 have previously demonstrated the mixed venous-arterial CO2 gradient to be increased in various forms of low-flow states. While the mixed venous-arterial CO2 gradient is a global measure of low flow, the PslCO2-PaCO2 gradient is a specific end-organ marker of tissue dysoxia.
The difficulty in assessing the performance of sublingual capnography is that there is no "gold standard" with which to compare. Gastric tonometry is currently the only end-organ metabolic monitoring system currently available. In this study, the performance of sublingual capnography was at least as good as that of tonometry. Blood lactate and PmvO2 are poor markers of tissue dysoxia, particularly in patients with sepsis. Sublingual capnography appears to have promise as a simple and noninvasive method of monitoring tissue dysoxia. This study is limited by the small sample size; additional studies are needed to determine the "normal" and "abnormal" ranges of the PslCO2-PaCO2 gradient in critically ill patients, the prognostic value of this marker, and the effect on patient outcome of PslCO2 titrated treatment.
| Footnotes |
|---|
Received for publication July 28, 2000. Accepted for publication February 6, 2001.
| References |
|---|
|
|
|---|
This article has been cited by other articles:
![]() |
G. A. A. M. Cammarata, M. H. Weil, M. Fries, W. Tang, S. Sun, and C. J. Castillo Buccal capnometry to guide management of massive blood loss J Appl Physiol, January 1, 2006; 100(1): 304 - 306. [Abstract] [Full Text] [PDF] |
||||
![]() |
E. J. Bridges and S. Dukes Cardiovascular Aspects of Septic Shock: Pathophysiology, Monitoring, and Treatment Crit. Care Nurse, April 1, 2005; 25(2): 14 - 40. [Full Text] [PDF] |
||||
![]() |
J. A. Guzman, M. S. Dikin, and J. A. Kruse Lingual, splanchnic, and systemic hemodynamic and carbon dioxide tension changes during endotoxic shock and resuscitation J Appl Physiol, January 1, 2005; 98(1): 108 - 113. [Abstract] [Full Text] [PDF] |
||||
![]() |
S. O. Heard Gastric Tonometry: The Hemodynamic Monitor of Choice (Pro) Chest, May 1, 2003; 123(5_suppl): 469S - 474S. [Abstract] [Full Text] [PDF] |
||||
![]() |
L. J. Caruso, A. J. Layon, and A. Gabrielli What Is the Best Way To Measure Cardiac Output? : Who Cares, Anyway? Chest, September 1, 2002; 122(3): 771 - 774. [Full Text] [PDF] |
||||
![]() |
C. Ince The Microcirculation Unveiled Am. J. Respir. Crit. Care Med., July 1, 2002; 166(1): 1 - 2. [Full Text] [PDF] |
||||
![]() |
D. De Backer, J. Creteur, J.-C. Preiser, M.-J. Dubois, and J.-L. Vincent Microvascular Blood Flow Is Altered in Patients with Sepsis Am. J. Respir. Crit. Care Med., July 1, 2002; 166(1): 98 - 104. [Abstract] [Full Text] [PDF] |
||||
![]() |
P. E. Marik, J. Varon, F. Abroug, L. Besbes, S. Nouira, S. Sarkar, Y. Kupfer, S. Tessler, E. P. Rivers, and H. B. Nguyen Goal-Directed Therapy for Severe Sepsis N. Engl. J. Med., March 28, 2002; 346(13): 1025 - 1026. [Full Text] [PDF] |
||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |