| JOURNAL HOME | CME HOME | THIS MONTH | PAST ISSUES | ETOC | COLLECTIONS |
| AUTHORS | REVIEWERS | EDITORIAL BOARD | FEEDBACK | RSS | HELP |
 
|
|
|
||
|
|
© 2002 International Anesthesia Research Society
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
Abstract |
|---|
 Top Abstract IntroductionMethodsResultsDiscussionReferences |
|---|
We compared estimates for base excess of extracellular fluid (BEecf; mmol/L) obtained in five clinically used blood gas analyzers: AVL Compact 2 (Roche Diagnostics, Mannheim, Germany), Ciba-Corning 860 (Bayer Diagnostics, Fernwald, Germany), IL 1620 (Instrumentation Laboratories, Lexington, MA), Stat Profile Ultra (Nova Biomedical, Waltham, MA), and ABL 510 (Radiometer, Copenhagen, Denmark). A total of 134 measurements per analyzer were obtained in arterial and venous blood samples from 10 patients undergoing cardiac surgery and 65 measurements per analyzer in venous blood samples from 2 healthy volunteers. The blood samples were equilibrated in a tonometer with gases of known composition (37°C). Additional theoretical studies were performed to evaluate the relationship between pH and calculated BEecf value (with varied PCO2) using the formulas of the various analyzers. The standard deviations of repeated measurements were 0.24 mmol/L for ABL 510 and approximately 0.45 mmol/L for the other 4 analyzers. The maximal systematic difference between the average of all measurements of each analyzer was 3.7 mmol/L; this was primarily attributable to differences in measuring pH, and, to a lesser extent, to differences in calculation and determination of PCO2. Comparison of the results from samples with different oxygen saturation showed that the relative alkalinity of deoxygenated hemoglobin (Haldane effect) can also influence the determinations of BEecf.
IMPLICATIONS: A clinically useful way to quantify nonrespiratory disturbances of the acid-base balance is calculation of the base excess of extracellular fluid by using blood gas analyzers. In this study, we found significant variability in estimates of base excess of extracellular fluid obtained with five analyzers from different manufacturers. This variability is attributable to multiple factors, including lack of correction for deoxygenated hemoglobin (Haldane effect).
|
Introduction |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Values for base excess are used widely to evaluate the nonrespiratory component of acid-base balance in patients. The determinations of base excess in the extracellular fluid (BEecf) are usually performed by using instruments, so-called "blood gas analyzers," that simultaneously measure the partial pressures of oxygen (PO2) and carbon dioxide (PCO2), as well as of pH in a blood sample. Calculated estimates for BEecf are derived by applying these measured values to an algorithm that is unique for the particular analyzer. Thus, these estimates for BEecf may vary depending on which analyzer is used. This variability is compounded by the inherent error involved in measurements of PO2, PCO2, and pH.
Errors in the estimates for BEecf can lead to misdiagnosis and potentially harmful treatment strategies, or to an underestimation of the severity of the metabolic derangement, which can result in inadequate or delayed treatment. This can be potentially life threatening because of impairments on the biochemical, cellular, and tissue levels, leading to failure of vital organs. Moreover, both acidosis and alkalosis can diminish the effectiveness of drug therapy and other treatment maneuvers, including controlled ventilation. Thus, it is important that the clinician appreciate the limitations of the clinically available estimates for BEecf, and to what extent they can be relied on as a basis for treatment.
The current study was conducted to compare, in a systematic manner, the ability of five commercially available, clinically used blood gas analyzers to provide estimates of BEecf from measurements of PCO2, PO2, and pH in samples of blood. Arterial, venous, and tonometered blood samples from patients and healthy volunteers were used. Mathematical correction for the effect of oxygen saturation (SO2) on estimates of base excess of whole blood (BEbl), the so-called "Haldane effect," has been suggested (1,2). We assessed the advantage of this correction in the estimates of BEecf.
|
Methods |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
The study was comprised of five modern automatic blood gas analyzers: the AVL Compact 2 (Roche Diagnostics, Mannheim, Germany), the Ciba-Corning 860 (Bayer Diagnostics, Fernwald, Germany), the IL 1620 (Instrumentation Laboratories, Lexington, MA), the Stat Profile Ultra (Nova Biomedical, Waltham, MA), and the ABL 510 (Radiometer, Copenhagen, Denmark). The instruments were standard models for both point-of-care and laboratory use and they were provided, put into operation, and maintained by the respective manufacturers or their authorized contractors. The same standards for test performance are used for the point-of-care and laboratory applications of these instruments, as indicated by the manufacturer. Two-point calibrations were performed just before the test series each day (manually started on each of the analyzers simultaneously) followed by intervals that were programmed in the devices.
The Radiometer ABL 510 uses a relatively complex algebraic simulation of the acid-base nomogram introduced by Siggaard-Andersen (3) to calculate BEecf. On the basis of the recommendation by the International Federation of Clinical Chemistry (IFCC) (4), the other analyzers have simplified these formulas:
|
|
with
|
|
The respective constants are presented in Table 1.
|
A correction was made for the Haldane effect, i.e., incomplete oxygen saturation, in all estimates of BEecf by using the following formulas (2,5,6):
|
|
with
|
|
with
|
|
Only the Radiometer ABL 510 (measurement) and Nova Stat Profile Ultra (determination via hematocrit) analyzers provide the concentration of hemoglobin (cHb, [g/dL]), which is a value required in Equation 3.
Theoretical studies were performed to assess differences in the calculated BEecf values obtained using the equations used by the various blood gas analyzers. A plot of the relationship between the pH values and the calculated values for BEecf was generated at PCO2 values approximating hypocapnia, normocapnia, and moderate and severe hypercapnia.
Experimental studies were performed in which a value for BEecf was obtained for the same blood sample using each of the five blood gas analyzers in sequence. Arterial and central venous blood samples were obtained from 10 patients undergoing cardiac surgery (63 ± 12 yr of age, elective operations under clinically stable conditions). Seven determinations of BEecf were obtained for each of the 20 blood samples in each analyzer. This number of determinations was chosen on the basis of statistical, cost, and time considerations.
In addition, we obtained venous blood samples from two healthy volunteers and analyzed them for gases and pH with and without equilibration with specific gas mixtures using a Tonometer (IL 237; Instrumentation Laboratories). The samples of venous blood were analyzed for BEecf eight times before, and three times after, tonometry. A comparison of the pre- and post-tonometry values for BEecf provided insight into the effect of time, e.g., storage duration, on the values for BEecf. The samples were stored in ice water for 6.5 h between the BEecf determinations. The venous blood samples were equilibrated in the tonometer with 4 gas mixtures: 5% CO2, 95% O2, 0% N2; 5% CO2, 20% O2, 75% N2; 10% CO2, 0% O2, 90% N2; 8% CO2, 92% O2, 0% N2. These gas mixtures were standard mixtures provided by commercial suppliers of medical gases (AGA Gas, Neunkirchen, Germany), which encompassed a wide range of values for PCO2 and PO2.
We also analyzed in each instrument an arterial blood sample from one of the volunteers five times before and five times after the measurements on tonometered blood to evaluate the differences between arterial and venous samples drawn simultaneously (50 measurements). Thus, the total number of measurements was: 700 (during operations) + 2 · 55 (healthy, venous) + 2 · 120 (healthy, tonometered) + 50 (healthy, arterial) = 1100.
Written informed consent was obtained from all patients and volunteers in accordance with the regulations of the local ethics committees. In the surgical patients, the blood samples were obtained from arterial and central venous catheters that had been inserted for routine monitoring purposes. Before withdrawing the samples for analysis, 5 mL of blood was withdrawn into a separate syringe to avoid contamination by stale blood or drug solutions that may have been in the catheters. Direct punctures of the cubital vein and femoral artery were used to obtain blood samples from the volunteers. Ten-milliliter polypropylene syringes were used for blood sampling in the study (B. Braun Melsungen AG, Melsungen, Ger- many). The untreated arterial and venous samples were introduced into the blood gas analyzers directly from these syringes. A special procedure, using 2-mL syringes of the same type, was used to transfer blood from the tonometer to the analyzers. The walls of all syringes used in this study were moistened and the dead space was filled with sodium heparin solution (5000 IU/mL, Liquemin N 25,000; Hoffmann-La Roche AG, Grenzach-Wyhlen, Germany) to prevent clotting of blood and the introduction of air bubbles into the syringe. All blood samples were put in ice water immediately and transported to the laboratory within 10 min. The storage time on ice ranged between 1 and 4 h for the blood samples from the patients and between 4 and 7 h for the blood samples from the healthy volunteers. The syringes were removed from the ice water immediately before analysis, rolled between the palms for 5 s, and turned three times upside-down to ensure adequate mixing. Then, the content of the cone of the syringe was expelled into a compress. The order in which each blood sample was analyzed by the different instruments was varied randomly to exclude experimental bias.
Data acquisition was performed automatically by using two standard personal computer systems directly connected to the blood gas analyzers. Statistical analysis included calculation of means and standard deviations and the use of the Student’s t-test, Mann-Whitney U-test, and Wilcoxon’s matched-pairs, signed rank test to uncover statistically significant differences. A P value < 0.05 was considered statistically significant.
|
Results |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Differences in Calculated Values for BEecf
A comparison of the BEecf values calculated with the various formulas is presented in Figure 1. Under physiologic conditions (pH = 7.4, PCO2 = 40 mm Hg), the maximal difference in the calculated BEecf values was 0.64 mmol/L (Radiometer versus Ciba-Corning). The differences between the BEecf values were enhanced during metabolic alkalosis and acidosis. The greatest difference during metabolic alkalosis in the observed range of values for BEecf (±20 mmol/L) was 2.3 mmol/L; the formula used by IL 1620 and Nova STAT Profile Ultra yielded a BEecf value of 20.0 mmol/L, whereas that used by the ABL 510 yielded a BEecf value of 17.7 mmol/L (pH = 7.619 and PCO2 = 40 mm Hg). The formulas differed even more in the case of extreme metabolic alkalosis; the calculated value for BEecf was 34.6 mmol/L with the IL/Nova formulas and 30.0 mmol/L with the ABL 510 formula (pH = 7.65 and PCO2 = 50 mm Hg). In the case of metabolic acidosis, the maximal difference in the calculated BEecf value was 1.8 mmol/L, i.e., the formula used by Ciba-Corning 860 yielded a BEecf of -20.0 mmol/L, whereas that used by the ABL 510 yielded a value of -18.2 mmol/L at pH = 7.041 and PCO2 = 40 mm Hg.
|
Values for pH and PCO2 within the usual reference range resulted in widely varying calculated BEecf values: -6.7 mmol/L using the formula in Ciba-Corning 860 (pH = 7.35, PCO2 = 35 mm Hg) and +7.4 mmol/L using the formulas in IL and Nova instruments (pH = 7.45, PCO2 = 45 mm Hg).
Reliability of Repeated Measurements
Of the 220 planned measurements of BEecf, 199 were obtained in the 5 blood gas analyzers. Eight of the measurements were unsuccessful because the Nova Stat Profile Ultra reported error messages without results when analyzing blood samples tonometered with oxygen-free gas (contrary to the manual of this analyzer which specifies a range from 0.0 to 800.0 mm Hg for PO2). The other 13 missing values were caused by various error messages, which were distributed among the analyzers.
On average, each sample was analyzed 5.9 times in each of the blood gas analyzers. The differences between the individual measurements and the mean of the repetitions are displayed in Figure 2. The standard deviation of the differences for all analyzers was 0.41 mmol/L. The maximal difference between an individual value for BEecf and the mean value of all results measured by the same analyzer on the same sample was 3.19 mmol/L, which is nearly 8 times the standard deviation. This indicates that the errors do not follow a normal distribution (the Kolmogorov-Smirnov Goodness of Fit test confirms this with a nominal probability of P = 0.0000 for the normal distribution). Thirty-three measurements showed differences of >1 mmol/L, 7 of >1.5 mmol/L, and 3 of >2 mmol/L. Assuming a normal distribution, this would be expected in 1.47% (15 of the 995 cases), 0.03% (0 cases), or 0.00006% (0 cases).
|
|
The pH results showed a standard deviation (means calculated separately for each analyzer and sample) of 0.0073 for the measurements on patients’ blood and of 0.0051 for the other ones (significant with P < 0.0001 in Mann-Whitney test); the standard deviations of PCO2 and BEecf did not show significant differences between the patients’ and healthy volunteers’ blood.
Differences Between the Analyzers
Differences as large as 3.7 mmol/L were detected among the BEecf values obtained using the various analyzers (mean results of 199 parallel measurements by Nova Stat Profile Ultra minus Ciba-Corning 860; Table 2). These differences were mainly attributed to systematic differences in the measured pH values (2.67 mmol/L [BEecf]) between these two analyzers, and to a lesser extent, to differences in calculation of BEecf and determination of PCO2 values. By using a unified algorithm for calculation [IFCC recommendation (4)], the maximal systematic differences of BEecf could be reduced by 16%.
Differences Depending on the Type of Sample
The mean difference in BEecf for the venous and arterial samples taken during surgery was 1.98 mmol/L (95% confidence interval of t-test for paired samples [-0.853; -0.506]; means: venous +0.65, arterial -1.33 mmol/L). Correction for the Haldane effect (Eqs. 3–5) together with the IFCC-recommended formula (Eqs. 1 and 2 with the constants of the last line in Table 2) reduced the difference to 0.18 mmol/L ([-0.35; -0.01]; means: -1.03, -1.21 mmol/L, respectively; also significantly different with P = 0.001 in Wilcoxon’s matched-pairs signed rank test). For the non-tonometered blood samples from the healthy volunteers, the difference of the reported values was 2.52 mmol/L (95% confidence interval of t-test for independent samples [1.91; 3.13]; means: +3.18, +0.66 mmol/L, respectively); calculation with Equations 1–5 caused a significant overcompensation for the Haldane effect (meaning that the venous values were then less than the arterial ones) of 0.54 mmol/L ([0.20; 1.40]; means: -0.03, 0.57 mmol/L, respectively; in Mann-Whitney test different with P = 0.011). The difference between the samples tonometered with the oxygen-free gas and those tonometered with a gas containing enough oxygen to reach nearly full saturation of the hemoglobin was 4.65 mmol/L ([3.90; 5.41]; means: 2.74, -1.91 mmol/L, respectively; n = 56, 184, respectively). When Equations 1–5 were used, a mean difference of 0.64 mmol/L resulted ([-0.13; +1.42]; means: -1.10, -1.74 mmol/L, respectively; according to Mann-Whitney test, however, significant overcompensation with P = 0.02; a marked skewness of -0.96 toward negative values limited the reliability of the t-test in this case).
Tonometry reduced the BEecf values by an average of 2.8 mmol/L after correction for the Haldane effect (95% confidence interval of t-test for independent samples [-3.3; -2.3]; means of the two types of samples +1.6 and -1.2 mmol/L). There was no statistically significant effect of the various gas mixtures on the decrease in BEecf; however, differences between the test gases of up to 5.6 mmol/L were found.
Storage of the venous samples in ice water for 6.5 h did not affect the BEecf values (mean 0.05 mmol/L, P = 0.473 in Mann-Whitney U-Test).
|
Discussion |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
BEecf values are calculated from measurements of pH and PCO2 obtained in a blood gas analyzer. These values are used clinically to identify metabolic acidosis and alkalosis, and as a basis for corrective therapy. This study showed significant variability in calculated BEecf values using various analyzers, thus raising questions about the clinical usefulness of this index.
There is no consensus in the literature regarding the acceptable clinical range for BEecf values, and which values should be of concern to the clinician. The reference intervals for BEecf provided in the instrument manuals are typically in the range of ±3 mmol/L, but the exact values vary considerably, and two of the manufacturers (IL, Nova Biomedical) provide no reference interval in their instrument manuals. The reference intervals offered for BEecf are similar (or identical) to those offered for BEbl, which is considered redundant for clinical purposes (7) and no longer supported by the IFCC (4). All of the instruments tested in this study offered values for both BEecf and BEbl.
BEecf is an estimate of how blood and the interstitial fluid that is in direct diffusional contact with the blood, respond to acid-base changes in vivo, whereas BEbl is a measure of base excess in an isolated blood sample. In the estimate for BEecf, the volume of the interstitial fluid is considered to be approximately 2 times that of the blood (4). The sum of the blood and the interstitial fluid is an estimate of the "extracellular fluid." Within the physiologic range for pH, PCO2, and cHb, the calculated values for BEbl and BEecf are nearly identical. However, in the case of hypercapnia, BEbl is considerably lower than BEecf. For example, at PCO2 = 80 mm Hg, pH = 7.2 and cHb = 20 g/dL, the IL (IL 1620) and Nova Biomedical (Stat Profile Ultra) devices calculate a BEbl of +2.6 mmol/L and a BEecf of +6.1 mmol/L. At PCO2 = 80 mm Hg, pH = 7.2 and cHb = 15 g/dL, the Roche device (AVL Compact 2) calculates a BEbl of -0.7 mmol/L and a BEecf of +2.7 mmol/L. Larger hemoglobin concentrations (only relevant for the calculation of BEbl) further magnify this difference (BEbl = -1.9 mmol/L at PCO2 = 80 mm Hg, pH = 7.2, and cHb = 20 g/dL).
Clinical decisions are often based on changes in BEecf over time rather than on a single measurement. Studies on trauma patients have revealed that a decrease in BEecf of approximately 3 mmol/L shows a good correlation to intraabdominal injury (8), and is a predictor of mortality (9–11). In the multivariate logistic regression model developed by Rixen et al. (11), a decrease of base excess by 3 mmol/L corresponds to a mortality odds ratio of 1.29, i.e., an approximately 30% increased risk to die. Smith et al. (12) demonstrated that a base excess value of less than -4 mmol/L predicted the risk of mortality with both sensitivity and specificity of approximately 70% for patients admitted to intensive care units. Unfortunately, these investigators did not provide the method of determination of base excess, e.g., whether BEecf or BEbl was estimated, or the number and types of analyzers and formulas used.
Our findings indicating as much as a 3.7 mmol/L systematic difference for BEecf values have obvious clinical relevance, because this difference corresponds to a 36% variation in mortality risk (11). A patient with a nearly normal BEecf of -0.4 mmol/L on one analyzer would be in the high-risk group according to the results of another analyzer. In the study in trauma patients alluded to above (8), if the analyzer with a tendency toward the highest values was used on admission and that with a tendency toward the lowest values for the control measurements, most patients would be presumed to have a high risk for intraabdominal injury. If the sequence of the analyzers were reversed, a true decrease in BEecf would likely never be discovered.
We found standard deviations between 0.24 and 0.48 mmol/L (see Table 2) for repeated measurements on the same sample in a given instrument. These findings imply that deviations of >2 mmol/L (approximately 4 standard deviations) must be expected less than once every 10,000 measurements, which would appear acceptable for clinical purposes. However, the observed deviations did not follow a "normal distribution," and large differences for parallel measurements were more frequent. This suggests that a repeat measurement may be in order if there is doubt about the accuracy of the first.
Our findings demonstrated that the variability of measurements within a given instrument was much less than that between instruments. Although the absence of a reference method ("gold standard") does not permit a determination of which of the tested instruments was more accurate, our findings suggest that at least some of them have errors in calibration.
To provide insight into the influence of time-related factors on analyzer reliability, we compared the present findings with those obtained three months earlier. This comparison demonstrated that the maximal difference in the value for pH increased with time (0.026 versus 0.037, respectively), resulting in a larger range for the values for BEecf (1.0 versus 3.7 mmol/L). This may be related to deterioration of the buffer solutions used for calibration. The time-related alteration in the pH measurements was not detected despite the regular use of quality-control, buffer solutions, in accordance with the guidelines of the manufacturers. This is because the accepted intervals for pH and PCO2 are wide, and capable themselves of producing differences in base excess of up to 8 mmol/L (13).
The larger variability in the pH values of blood from patients versus that from volunteers is noteworthy, because it underscores the importance of not limiting reference measurements to healthy subjects; measurements from patients in various critical conditions would seem essential. This variability in pH values was most marked for cHb. The standard deviations (means calculated separately for each analyzer and sample) were 0.16 mmol/L = 0.26 g/dL (ABL 510) and 0.21 mmol/L = 0.34 g/dL for the samples from healthy volunteers; however, they were 0.86 mmol/L = 1.37 g/dL, respectively, 0.64 mmol/L = 1.03 g/dL for the samples from patients.
Another source of variation in the BEecf values was the use of different equations for calculation. This is the result of two commissions (IFCC and National Committee for Clinical Laboratory Standards) suggesting different algorithms and of most manufacturers using others. The buffering properties of blood are very complex and affected by pathologic conditions, which makes them difficult to quantify (10,13). This limits the reliability of all the algorithms now in use. It is not possible to choose one algorithm over the others.
None of the currently used algorithms corrects for incomplete saturation of the hemoglobin with oxygen (Eq. 3); Roche has announced that it will add this feature to the AVL analyzers (2), but as yet has not done so. Thus, at this time, the values for BEecf in fully saturated samples must be considered the most reliable. The results of our study suggest that, for both healthy persons and patients in stable condition undergoing surgery, venous and arterial samples yield nearly equivalent values for BEecf, as long as the Haldane effect is quantified using the correction factor 0.03 mmol H+ per gram of hemoglobin (Eq. 3). This implies that the correction factor introduced by Zander (2) (0.02 mmol/g) for BEbl may not be sufficient for BEecf. Correction for the Haldane effect is more important for estimates of BEecf than for those of BEbl, because BEecf is more affected by increases in PCO2, which occur concomitantly with decreases in oxygen saturation within the capillaries. Consequently, values for venous BEecf are generally higher than the corresponding values for BEbl. During resuscitation and other critical conditions, venous BEecf can be expected to decrease more than arterial BEecf, particularly if sodium bicarbonate is administered (14–17). This arteriovenous difference in BEecf values can serve as a valuable diagnostic tool in clinically critical situations. Thus, it is important that it not be obscured by lack of compensation for the Haldane effect.
Previous studies have suggested that the commercially available solutions may not have sufficient reliability for quality control of determinations of base excess (13,18,19). Our findings suggest that untreated venous blood may provide a readily available and more reliable substitute in patients with normal acid-base balance, because the BEbl does not change (by definition) if PCO2 changes during handling of the sample, unlike the base excess of these quality-control solutions, which have very different buffering properties. However, this approach would require compensation for the Haldane effect, because, otherwise, the rendered BEecf values would be approximately +2 mmol/L, and they would vary inversely with the SO2 of the venous blood sample.
The number of patients used in the present study was based on data from previous studies (2,13), and our clinical experience suggesting that this number would be necessary to encompass a variety of clinical conditions and a wide range of pH values. Variation in the acid-base status for healthy individuals was assumed to be small (as borne out in the present results), and thus only two normal volunteers were studied.
In conclusion, although values for base excess can be useful in detecting critical conditions in emergency and critical care medicine, they are derived values with significant sources of error and variability. It is important that the clinician appreciate these limitations of values for base excess, and do not rely on them exclusively to evaluate the metabolic status of the patient.
|
Acknowledgments |
|---|
We are grateful to our patients and to the volunteers for their cooperation, and to the manufacturers of the analyzers for supplying and maintaining them.
|
References |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
- Thews G. Ein Nomogramm für die O2-Abhängigkeit des Säure-Basen-Status im menschlichen Blut. Pflügers Arch Gesamte Physiol Menschen Tiere 1967;296:212–21.
- Zander R. Die korrekte Bestimmung des Base Excess (BE, mmol/L) im Blut. Anaesthesiol Intensivmed Notfallmed Schmerzther 1995; 30 (Suppl 1): 36–8.
- Christiansen TF. An algorithm for calculating the concentration of the base excess of blood. In: Siggaard-Andersen O, ed. Proceedings of the IFCC Expert Panel on pH and blood gases held at Herlev Hospital 1980. Copenhagen: Radiometer A/S, 1981; 77–81.
- Burnett RW, Covington AK, Fogh-Andersen N, et al. International Federation of Clinical Chemistry (IFCC), Committee on pH, Blood Gases and Electrolytes: approved IFCC recommendation on definitions of quantities and conventions related to blood gases and pH. Eur J Clin Chem Clin Biochem 1995; 33: 399–404.[Web of Science][Medline]
- Siggaard-Andersen O. The acid-base status of the blood. Copenhagen: Munksgaard, 1963.
-
Dill DB, Edwards HT, Consolazio WT. Blood as a physiochemical system. XI. Man at rest. J Biol Chem 1937; 118: 635–48.
[Free Full Text] -
Siggaard-Andersen O, G
thgen IH. Oxygen and acid-base parameters of arterial and mixed-venous blood, relevant versus redundant. Acta Anaesthesiol Scand Suppl 1995; 107: 21–7.[Medline]
-
Mackersie RC, Tiwary AD, Shackford SR, Hoyt DB. Intraabdominal injury following blunt trauma: identifying the high-risk patient using objective risk factors. Arch Surg 1989; 124: 809–13.
[Abstract/Free Full Text] - Rutherford EJ, Morris JA Jr, Reed GW, Hall KS. Base deficit stratifies mortality and determines therapy. J Trauma 1992; 33: 417–23.[Web of Science][Medline]
-
Siegel JH, Rivkind AI, Dalal S, Goodarzi S. Early physiologic predictors of injury severity and death in blunt multiple trauma. Arch Surg 1990; 125: 498–508.
[Abstract/Free Full Text] - Rixen D, Raum M, Bouillon B, et al. Prognoseabschätzung des Schwerverletzten: Eine Analyse von 2069 Patienten des Traumaregisters der DGU. [Predicting the outcome in severe injuries: an analysis of 2069 patients from the trauma register of the German Society of Traumatology.] Unfallchirurg 2001; 104: 230–9.[Web of Science][Medline]
- Smith I, Kumar P, Molloy S, et al. Base excess and lactate as prognostic indicators for patients admitted to intensive care. Intensive Care Med 2001; 27: 74–83.[Web of Science][Medline]
- Zander R. Qualitätskontrolle des Base Excess (BE, mmol/L) im Blutgasanalysator. Anaesthesiol Intensivmed Notfallmed Schmerzther 1995; 30 (Suppl 1): 39–40.
- Adrogué HJ, Rashad MN, Gorin AB, et al. Assessing acid-base status in circulatory failure: differences between arterial and central venous blood. N Engl J Med 1989; 320: 1312–6.[Abstract]
- Mark JB, FitzGerald D, Fenton T, et al. Continuous arterial and venous blood gas monitoring during cardiopulmonary bypass. J Thorac Cardiovasc Surg 1991; 102: 431–9.[Abstract]
- Trancik VJ, Barta E, Siska K, et al. Vergleich der Werte des Säure-Basen-Status und der Blutgase im zentralvenösen, kapillären und arteriellen Blut bei kardiochirurgischen Patienten. Zentralbl Chir 1974; 99: 1509–20.[Medline]
- Weil MH, Rackow EC, Trevino R, et al. Difference in acid-base state between venous and arterial blood during cardiopulmonary resuscitation. N Engl J Med 1986; 315: 153–6.[Abstract]
-
Hansen JE, Casaburi R. Patterns of dissimilarities among instrument models in measuring PO2, PCO2, and pH in blood gas laboratories. Chest 1998; 113: 780–7.
[Abstract/Free Full Text] - Witek F. Standardisierte Qualitätskontrolle der Blutgasanalyse: Bericht über zwei Ringtests in Österreich. Wien Med Wochenschr 1978; 128: 272–6.[Medline]
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
