Automatic Blood Gas Research Articles

Combined inhibition of nitric oxide and vasodilating prostaglandins abolishes forearm vasodilatation to systemic hypoxia in healthy humans

We tested the hypothesis that nitric oxide (NO) and vasodilating prostaglandins (PGs) contribute independently to hypoxic vasodilatation, and that combined inhibition would reveal a synergistic role for these two pathways in the regulation of peripheral vascular tone. In 20 healthy adults, we measured forearm blood flow (Doppler ultrasound) and calculated forearm vascular conductance (FVC) responses to steady-state (SS) isocapnic hypoxia (O₂ saturation ~85%). All trials were performed during local α- and β-adrenoceptor blockade (via a brachial artery catheter) to eliminate sympathoadrenal influences on vascular tone and thus isolate local vasodilatory mechanisms. The individual and combined effects of NO synthase (NOS) and cyclooxygenase (COX) inhibition were determined by quantifying the vasodilatation from rest to SS hypoxia, as well as by quantifying how each inhibitor reduced vascular tone during hypoxia. Three hypoxia trials were performed in each subject. In group 1 (n = 10), trial 1, 5 min of SS hypoxia increased FVC from baseline (21 ± 3%; P < 0.05). Infusion of N(G)-nitro-L-arginine methyl ester (L-NAME) for 5 min to inhibit NOS during continuous SS hypoxia reduced FVC by -33 ± 3% (P < 0.05). In Trial 2 with continuous NOS inhibition, the increase in FVC from baseline to SS hypoxia was similar to control conditions (20 ± 3%), and infusion of ketorolac for 5 min to inhibit COX during continuous SS hypoxia reduced FVC by -15 ± 3% (P < 0.05). In Trial 3 with combined NOS and COX inhibition, the increase in FVC from baseline to SS hypoxia was abolished (~3%; NS vs. zero). In group 2 (n = 10), the order of NOS and COX inhibition was reversed. In trial 1, five minutes of SS hypoxia increased FVC from baseline (by 24 ± 5%; P < 0.05), and infusion of ketorolac during SS hypoxia had minimal impact on FVC (-4 ± 3%; NS). In Trial 2 with continuous COX inhibition, the increase in FVC from baseline to SS hypoxia was similar to control conditions (27 ± 4%), and infusion of L-NAME during continuous SS hypoxia reduced FVC by -36 ± 7% (P < 0.05). In Trial 3 with combined NOS and COX inhibition, the increase in FVC from baseline to SS hypoxia was abolished (~3%; NS vs. zero). Our collective findings indicate that (1) neither NO nor PGs are obligatory to observe the normal local vasodilatory response from rest to SS hypoxia; (2) NO regulates vascular tone during hypoxia independent of the COX pathway, whereas PGs only regulate vascular tone during hypoxia when NOS is inhibited; and (3) combined inhibition of NO and PGs abolishes local hypoxic vasodilatation (from rest to SS hypoxia) in the forearm circulation of healthy humans during systemic hypoxia.

Glucose Measurement in the Operating Room

Abnormalities of blood glucose are common in patients undergoing surgery, and in recent years there has been considerable interest in tight control of glucose in the perioperative period. Implementation of any regime of close glycemic control requires more frequent measurement of blood glucose, a function for which small, inexpensive, and rapidly responding point-of-care devices might seem highly suitable. However, what is not well understood by many anesthesiologists and other staff caring for patients in the perioperative period is the lack of accuracy of home glucose meters that were designed for self-monitoring of blood glucose by patients. These devices have been remarketed to hospitals without appropriate additional testing and without an appropriate regulatory framework. Clinicians who are accustomed to the high level of accuracy of glucose measurement by a central laboratory device or by an automated blood gas analyzer may be unaware of the potential for harmful clinical errors that are caused by the inaccuracy exhibited by many self-monitoring of blood glucose devices, especially in the hypoglycemic range. Knowledge of the limitations of these meters is essential for the perioperative physician to minimize the possibility of a harmful measurement error. In this article, we will highlight these areas of interest and review the indications, technology, accuracy, and regulation of glucose measurement devices used in the perioperative setting.

Evaluation of transcutaneous CO2 responses following acute changes in PaCO2 in healthy subjects

Transcutaneous blood gas-monitoring systems with miniaturized SpO(2) (peripheral blood oxygen saturation)/PCO(2) combined sensors (TOSCATM) have been widely used. There are no reports of the inter- and intra-individual variability in transcutaneous measurements of PaCO(2) (PtcCO(2)) in response to acute progressive changes in PaCO(2). This study examined inter- and intra-individual variability of PtcCO(2) measurements under semi-steady-state conditions, and characterized the behaviour of PtcCO(2) in response to acute progressive changes in PaCO(2). Subjects breathed mixed gases through a mouthpiece connected to an automatic arterial blood gas controller. Using end-tidal PCO(2) and PO(2) as guides, PaCO(2) was controlled to increase and/or decrease between baseline and > or = 60 mm Hg, in a stepwise (n = 9) or progressive fashion (n = 6). Arterial blood was sampled when needed. Intra-individual correlation coefficients between PtcCO(2) and PaCO(2) were excellent in all subjects (0.971-0.989); however, the slope of the regression line varied among subjects (1.040-1.335). Bias and limits of agreement (+/- 2 SD from bias) between PtcCO(2) and PaCO(2) were -1.8 mm Hg and -7.7 to 4.1 mm Hg. Changes in PtcCO(2) in response to acute progressive changes in PaCO(2) also varied among subjects. The PtcCO(2) measurement system allows reliable estimation of PaCO(2) in a given subject. However, caution is needed when comparing absolute values between subjects or when acute changes in PaCO(2) occur.

Validation study of a new transcutaneous PCO2 measurement system using an automatic blood gas controller in healthy subjects and clinical usefulness of a new transcutaneous PCO2 measurement system for patients

Validation study of a new transcutaneous PCO₂ measurement system using an automatic blood gas controller in healthy subjects and clinical usefulness of a new transcutaneous PCO₂ measurement system for patients

24-norUrsodeoxycholic Acid Is Superior to Ursodeoxycholic Acid in the Treatment of Sclerosing Cholangitis in Mdr2 (Abcb4) Knockout Mice

Current therapy for primary sclerosing cholangitis is of limited efficacy. Multidrug resistance gene 2 knockout mice (Mdr2(-/-)) represent a well-characterized model for sclerosing cholangitis. Experiments were performed to test in such mice the therapeutic effects of 24-norUrsodeoxycholic acid, a C(23) homologue of ursodeoxycholic acid with 1 fewer methylene group in its side chain. Mdr2(-/-) mice were fed a diet containing 24-norUrsodeoxycholic acid (0.5% wt/wt) or ursodeoxycholic acid (0.5% wt/wt) as a clinical comparator for 4 weeks; controls received standard chow. Effects on serum liver tests, liver histology, markers of inflammation and fibrosis, and bile acid transport and metabolism were compared. 24-norUrsodeoxycholic acid metabolism was studied in serum, liver, bile, and urine. 24-norUrsodeoxycholic acid markedly improved liver tests and liver histology and significantly reduced hydroxyproline content and the number of infiltrating neutrophils and proliferating hepatocytes and cholangiocytes. 24-norUrsodeoxycholic acid underwent extensive phase I/II metabolism (hydroxylation, sulfation, and glucuronidation), thereby increasing the hydrophilicity of biliary bile acid secretion. There was a coordinated induction of bile acid detoxifying enzymes (Cyp2b10, Cyp3a11, and Sult2a1) and efflux pumps (Mrp3 and Mrp4). Ursodeoxycholic acid, in contrast, increased alanine transaminase and alkaline phosphatase levels, had no significant effects on hydroxyproline content, and induced biliary transporters and detoxification enzymes to a much smaller extent than 24-norUrsodeoxycholic acid. 24-norUrsodeoxycholic acid ameliorates sclerosing cholangitis in Mdr2(-/-) mice. Its therapeutic mechanisms involve (1) increasing the hydrophilicity of biliary bile acids, (2) stimulating bile flow with flushing of injured bile ducts, and (3) inducing detoxification and elimination routes for bile acids.

Messgrößen in der Perinatalmedizin - der Basenexzess

There are convincing data from neonatal studies that the base excess (BE, mmol/L) measured in capillary or cord blood offers special diagnostic and prognostic power in the new-born compromised by hypoxia. For computation of BE the hemoglobin concentration (Hb, %), the oxygen saturation (%) and the distribution of the fetal fluid compartments are necessary. Until now these three factors, which differ in the fetus when compared with adults, have not been taken into account using automatic blood gas equipment. This study therefore attempts to analyse and verify which BE will offer the best fit with the outcome conditions in the newborn. Using a cohort of 7701 singletons delivered at term by the vaginal route in whom a blood gas analysis in cord blood together with Hb measurements were available and plausible, three BE values were computed: the ordinary BE (actual BE, according to Radiometer determinations), the BE corrected to the real oxygen saturation (BE,oxy) and the BE in the extracellular fluid compartment of the fetus (BE,ecff) using the algorithms of O. Siggaard-Andersen. Moreover, three different oxygen saturation curves for fetal Hb (HbF) were applied, namely those of Zander et al., Hellegers and Schruefer and Ruiz et al. (Severinghaus). The prognostic potential of the different BE values was analysed using a correlation with the APGAR-score after 1 min, the sum of the APGAR indices after 1 and 5 min, and additionally with the fetal heart frequency (FHF) of the last 30 min ante-partum quantified by a score electronically in a group of 342 fetuses. The median of the actual BE in umbilical artery blood amounted to -4.9 mmol/L. Correction to the real oxygen saturation of the blood sample leads to a median of -7.7 mmol/L. Using three different oxygen saturation curves and thus taking into account HbF reveals little influence on the numerical values of BE. However, application of the true distribution of fetal fluid compartments in the evaluation of BE (BE,ecff) leads to a significant reduction of the numerical value of BE (median = -3.2 mmol/L). Computational correction of BE using the real oxygen saturation (%) in each fetus leads to a important and highly significant increase of the correlation (r and rho) with the APGAR indices. The correlation with the FHF score also increased but failed to reach significance. Our data show that, in perinatal medicine, a computational correction of BE according to the real oxygen saturation of the blood sample is mandatory. An algorithm for HbF should be used. The BE in the extracellular fluid compartment does not fit to reality. The determination of the Hb concentration in the umbilical blood of each fetus is essential since its variability is higher than expected. In legal controversies the computational origin of BE should be documented and carefully regarded.

Postoperative Pulmonary Changes after Laparoscopic Cholecystectomy

Postoperative Pulmonary Changes after Laparoscopic Cholecystectomy

Quantification of dose-dependent respiratory depression in isoflurane-anesthetized horses

The magnitude of respiratory depression in equidae anesthetized with isoflurane has not been thoroughly described. The aim of this study was to quantify changes in ventilation as a function of isoflurane (ISO) dose in the horse. Anesthesia was induced and maintained in five healthy adult horses weighing 564 ± 52 kg (mean ± SD) in left lateral recumbency using only ISO in O2. During spontaneous ventilation, four doses of constant end-tidal ISO measured by a calibrated infra-red analyzer were studied in randomized order: 1.31, 1.57, 1.83, and 2.10%; corresponding to the minimum alveolar concentration (MAC), 1.2 MAC, 1.4 MAC, and 1.6 MAC, respectively. The PaO2 and PaCO2 were measured using an automated blood gas analyzer and corrected to tonometrically derived standard curves. Respiratory rate (f) and tidal volume (VT) were measured via calibrated signal integration from a Silverman-style pneumotachometer; volumes were corrected to BTPS. Minute ventilation (VE) was the calculated product of VT and f. Data were analyzed using a repeated measures anova; Tukey's test was used for pairwise comparisons (p < 0.05). The following results are listed in order of ascending MAC multiples: PaCO2 (mm Hg) = 68 ± 14, 68 ± 9, 72 ± 9, 78 ± 5; f (min−1) = 4.3 ± 2.4, 3.5 ± 1.6, 2.5 ± 1.0, 2.0 ± 0.7; VT (mL kg−1) = 23 ± 11, 22 ± 11, 22 ± 12, 21 ± 7; (mL minute−1 kg−1) = 84 ± 41, 67 ± 27, 50 ± 23, 39 ± 11. Significant differences were observed between the highest and lowest MAC multiples for PaCO2, f, and VE, but not for VT. The PaO2 was always >80 mm Hg. The equations given below express median minute ventilation and median PaCO2 as a function of ISOMAC. Median values at each dose were used instead of mean values owing to the departure from sample variance uniformity. The y-intercepts coincide with published reports of and PaCO2 for awake, unmedicated horses (0 MAC). The x-intercept of Equation 1 predicts respiratory arrest at 2.4 MAC, near the apneic threshold for ISO. Future work directed towards increasing sample size in this study may remedy heterogeneous treatment variances present within the data and permit further refinement of dose–ventilation relationships in the horse. (1) PaCO2 = 22 × MAC + 39(2)

The distribution of major lymphocyte subsets in cord blood is associated with its pH

Objectives: To assess in venous cord blood the distribution of major lymphocyte subsets according to pH and medications used during labor. Design and methods: Venous cord blood was sampled immediately after labor from 70 newborns (35 males and 35 females) delivered vaginally. Lymphocytes were immunophenotyped by flow cytometry and pH was measured using the AVL 900 automated blood gas analysis system. Data on birth weight, gestational age at delivery, length of labor, presence of stained amniotic fluid, medications used during labor, maternal risk factors, age and parity were collected. Results: The percentage of T lymphocytes decreased while the percentage of NK lymphocytes increased with decreasing pH over the whole range of pH values. The proportions of T and NK lymphocytes were associated with the administration of neuroplegics, spasmolytics or dihydroergotoxin in the first stage of labor. Conclusions: Cord blood pH and labor-associated variables should be taken into account to adequately interpret the profile of major lymphocyte subsets as a marker of the effect of different prenatal factors on the immune system of neonates.

Gestion à distance de l'analyse délocalisée contrôle de qualité automatique

In the first part the authors present within the framework of delocalized biology the possibility of the remote control of several automatic blood gas analyzers from a central laboratory. This enables biologists to monitor laboratory exams which would otherwise be beyond their control and guarantee a quality of the results for the benefit of the patient. In the second part, they present an automatic control module enabling this control to be standardized thereby avoiding likely errors arising form human handling.

Computer analysis system of blood oxygen saturation in an animal hypoxia model.

An experimental animal hypoxia model has been developed. It consists of two sensors (an in vitro and in vivo model), an experimental device and a computer signal processing system. This method can easily be applied to determine and analyse blood oxygen saturation at various hypoxia levels. It can also be used to evaluate the accuracy of pulse oximetry over a wide range of oxyhemoglobin desaturation levels. The DC and AC components of recorded red and infra-red signals, the dual-wavelength ratio R12 and the reading of a pulse oximeter (SpO2) can be automatically calculated and displayed on a computer screen. Preliminary results of the animal hypoxia test indicate that the measurements made by the instrument correlate well with the oxygen saturation readings of the automatic blood gas analyser AVL945. The computer analysis system is suitable for repeated estimations in the animal model.

Effects of a combination of detomidine and butorphanol on respiratory function in horses with or without chronic obstructive pulmonary disease

Abstract Objective To evaluate the effects of detomidine and butorphanol in combination on respiratory function in horses and to determine whether these effects are more severe in horses with pre-existing respiratory dysfunction. Design Pulmonary function testing and arterial blood gas analyses were performed before and after administration of a combination of detomidine (10 μg/kg of body weight, IV) and butorphanol (20 μg/kg, IV). Animals 5 horses with chronic obstructive pulmonary disease and 5 horses free of respiratory disease (controls). Procedures Flow rates were obtained from a pneumotachograph attached to a face mask, and esophageal pressure was measured with a catheter placed in the distal third of the esophagus. Respiratory rate, tidal volume (VT), minute ventilation (VE), maximal change in transpulmonary pressure, pulmonary resistance, and dynamic compliance were calculated by use of a pulmonary function computer. Arterial blood was collected from the transverse facial artery, and blood gas partial pressures were measured with an automated blood gas analyzer. Results The combination of detomidine and butorphanol caused significant changes with time in respiratory rate, VT, VE, Paco2, and Paco2. Changes with time in VT, VE, and maximal change in transpulmonary pressure were significantly different between groups. Conclusion The combination of detomidine and butorphanol affects respiratory function in horses, and the observed changes are affected by the presence of preexisting respiratory dysfunction. Clinical Relevance Changes in respiratory function after administration of a combination of detomidine and butorphanol are not necessarily more severe in horses with pre-existing respiratory dysfunction. (Am J Vet Res 1996; 57:705–709)

Energy Metabolism of Thoracic Surgical Patients in the Early Postoperative Period: Effect of Posture

To determine the effect of elective thoracic surgery on energy metabolism and gas exchange and to evaluate whether the 30-degree sitting position would affect these variables. Prospective, unblinded, controlled study. Surgical ICU in a university hospital. Twenty-two adult patients undergoing elective pulmonary resection. Posture change from supine to 30-degree sitting position. Oxygen consumption (VO2), carbon dioxide production (VCO2), respiratory quotient (RQ), and energy expenditure (EE) were measured by means of computerized indirect calorimetry before and after surgery. Heart rate and respiratory frequency were measured continuously during gas exchange measurement. Blood gases were analyzed with an automated blood gas analyzer. Preoperatively, altering position did not affect energy metabolism, gas exchange, and cardiopulmonary variables. Postoperatively, the measured EE was 116% of the expected value. Mean EE and VO2 values for each position were higher than the preoperative values for the corresponding postures (p<0.05 for each position), while VCO2 increased only in the supine position (p<0.05). Mean percent increases in EE, VO2, and VCO2 were significantly lower in the 30-degree sitting position than in the supine position (EE: 7.9+/-2.7% vs 14.4+/-2.3%; p<0.001; VO2: 9.0+/-3.0% vs 16.4+/- 2.6%; p<0.001; VCO2: 3.2+/-2.1% vs 6.5+/-1.4%: p<0.05). Arterial oxygen tension and all the physiologic indexes of gas exchange for each position were worse than the preoperative values for the corresponding postures (p<0.05 for each position). Mean arterial pressure, heart rate, and respiratory frequency for each position were higher than the preoperative values for the corresponding postures (p<0.05 for each position). No changes in mean values of these variables occurred between the two positions postoperatively. The early postoperative period of patients undergoing elective thoracic surgery is characterized by a condition of impaired gas exchange and by a hypermetabolic state. Hypermetabolism can be partly mitigated by assuming the 30-degree sitting position.

Normobaric Measurement of Arterial Oxygen Tension in Subjects Exposed to Hyperbaric Oxygen

This study demonstrates the ability of an automated blood gas analyzer (Radiometer ABL 330) operated at atmospheric pressure to measure the arterial oxygen tension (PaO2) of ten healthy volunteers exposed to hyperbaric oxygen (HBO2) up to 3.0 atmospheres absolute. Arterial blood samples were aspirated from subjects compressed in a single-person hyperbaric chamber and were analyzed immediately in the blood gas analyzer. The subjects' values for PaO2 correlated with the calculated alveolar O2 tension (PAO2) (measured PaO2 = 0.827 x PAO2-15.1) (r2 = 0.97). Tonometric experiments indicated a difference between saline and blood PO2 measurements. We therefore derived a correction factor for blood measurements (corrected PaO2 = 0.908 x PAO2-52.4) (r2 = 0.98). These results compared favorably with PaO2 measurements made with blood gas analyzers calibrated inside walk-in hyperbaric chambers. We conclude that the PaO2 of normal subjects exposed to HBO2 can be measured accurately at atmospheric pressure with this automated blood gas analyzer. Prior to this study, hyperbaric PaO2 measurements could only be performed within walk-in chambers. Our observations generalize the normobaric measurement of hyperbaric PaO2 to patients treated in single-person and walk-in chambers.

Acidification of epididymal fluid in the boar

The present study describes the measurement of pH made in vivo in the rete testis fluid and in different regions of the boar epididymis. Furthermore, samples of whole ejaculates, semen fractions, testicular (ductuli efferents/rete testis), epididymal and deferential fluids collected from the same fertile boars, were analysed for their acid/base status with an automatic blood gas analyser. A pH gradient of activity was found between the fluid entering the ductus epididymis (pH 7.2) and the region of sperm storage at the cauda (pH 6.5). A significantly lower concentration of bicarbonate ion was found in the cauda epididymidis (3-4 mM) compared to rete testis fluid (30 mM), which might be related to the quiescence of the spermatozoa. A significant increase in extracellular pH and bicarbonate concentration occurred at ejaculation, the bicarbonate levels being 9-10-fold higher in the semen fraction rich in seminal vesicle fluid, where sperm showed higher motility, than in the cauda epididymis.

Dose calcium balanced heparin affect blood gas and electrolyte analysis?

The effect of calcium-balanced heparin (471896, CIBA CORNING) on blood gas and electrolyte analysis was evaluated, by comparing with that of sodium heparin (Na heparin). One ml of whole blood was collected into a syringe, which contained calcium-balance heparin (Ca balanced heparin) or Na heparin. 122 pairs of blood samples obtained from 15 patients were analyzed for Na, K, ionized calcium (Ca(++)), total hemoglobin, pH, P(CO)(2), and P(O)(2) by an automatic blood gas and electrolyte analyzer, CIBA CORNING model 288. There was a significant difference ( P << 0.05) in pH, P(CO)(2), Na, and Ca(++) between the two different groups. Ca(++) concentration was significantly less in Na heparin group than in Ca balanced heparin group, probably due to more chelation of Ca(++) by Na heparin than Ca balanced heparin. The present study suggests that the Ca balanced heparin has minimal effect on the blood gas and electrolyte analysis, and is a suitable anticoagulant for the Ca(++) measurement.

The clinical value of the haemoglobin concentration analysis (g/L) and the calculation of oxyhaemoglobin saturation (%) in an automatic blood gas analyser (ABL2, Radiometer A/S, Denmark)

This study was aimed to evaluate the accuracy and thereby the clinical value of the haemoglobin concentration (g/L) analysis and of the calculation of oxyhaemoglobin saturation (%) in a commonly used automatic blood gas analyser (ABL2, Radiometer A/S, Denmark). The measured and calculated values (ABL2) were compared to the readings obtained in a last generation haemoximeter (OSM3, Radiometer A/S, Denmark), which served as reference. The haemoglobin concentration measurements performed in the ABL2 apparatus showed an increasing underestimation of the true haemoglobin value at larger haemoglobin concentration values. This together with a poor correlation makes the analysis unsuitable for clinical use. The oxyhaemoglobin saturation values (%) obtained using the ABL2, however, had a good correlation to the corresponding values measured in the OSM3. It is suggested that the oxyhaemoglobin saturation value obtained in the ABL2 can be used clinically, e.g., for invasive control of pulsoximeters.

Precision of arterial blood gases and cutaneous oxygen saturation in healthy non-smokers

Arterial blood from 20 healthy non-smokers (mean age 36 years) was sampled with a pre-heparinized syringe (Radiometer B129), and analysed with an automatic blood gas analyser (AVL945). Cutaneous oxygen saturation (sO2) was recorded on an ear oxymeter (Biox IIA). Mean pH +/- 2 SD was 7.392 +/- 0.072, mean arterial pCO2 +/- 2 SD was 5.15 +/- 0.90 kPa, and mean arterial pO2 +/- 2 SD was 13.68 +/- 1.90 kPa. Standard deviation for differences in pairs of analyses (SD) was, for pH 0.008, 0.20 kPa for pCO2, and 0.33 kPa for pO2. In arterial blood from 20 hypoxic patients, SD for pO2 was 0.10 kPa. Mean cutaneous sO2 +/- 2 SD was 96.7 +/- 2.2%, and mean calculated arterial sO2 +/- 2 SD was 97.6 +/- 1.0%. Mean arterial pO2 was higher than that expected from the literature.

Knowledge-based instrument systems: Integrating knowledge software with microprocessor-based instruments

Integrating knowledge software with existing microprocessor-based instruments leads to a new generation of instrument systems. Our basic idea is to build an instrument system supported by a microcomputer workstation running knowledge software. The initial phase of implementation has been realised on a microprocessor-controlled automatic blood gas analyzer, interfaced to a Great-Wall 0520C-H, an IBM-XT upwardly-compatible workstation. The implementation of a knowledge system module which is programmed in LISP with lists expressed in Chinese characters and transformed into C which provides more compact code is described. It obtains a virtual link between instrument nodes according to the author's model. Further progress leads to a bedside instrument system of a life-support system for paediatric cardiac surgical patients. It serves as a testbed for exploring principles and guidelines for designing and implementing knowledge-based instrument systems as defined in the present paper.

Inter-Instrument Comparison of Blood Gas Analyzers and Assessment of Tonometry Using Fresh Heparinized Whole Human Blood

Because of limitations associated with commercial blood gas quality controls and tonometry of stored blood, fresh heparinized blood was used to compare PO2 and PCO2 performance of ten blood gas analyzers. Function of nine gas mixer/tonometer systems was evaluated. These were used to create blood samples with target values for PO2 and PCO2. All ten analyzers had high precision; this magnified small differences between observed results and target values. Grand mean results from all ten analyzers were within 0.8 mm Hg of the target for PO2 of 40.0 and 100.0 and for PCO2 of 20.0 and 70.0 mm Hg. Eight automated blood gas analyzers gave clinically indistinguishable results for PO2 and PCO2 that were sufficiently accurate for clinical purposes. Also reported is the practicality of using tonometry of fresh heparinized blood in a per shift quality control program for PO2 and PCO2.