CLIN. CHEM. 36/8(B), 1562-1 566 (1990)

Current Analytical Approaches to Measuring Blood AnalytesMary F. Burrltt

In recent years, laboratory testing in the critical-care settinghas increased, a trend due, in part, to the evolution ofelectrochemical sensors. Various innovations have extendedsensor lifetimes, reduced sensor maintenance, and led to thedevelopment of single-use and unit-use disposable sensors.These sensor technologies allow the accurate and precisedetermination, either at or near the bedside, of severalanalytes including Po2’ Pco2’ pH, Na, K, Cl, ionized calcium,hematocrit, total hemoglobin, and glucose. Use of these newsystems, however, has raised new issues regarding sensorcalibration and sample handling and collection. The numberof direct-reading analyzers for electrolyte determinations hasalso increased dramatically. Issues regarding calibration ofion-selective electrodes (ISE5) for NaJK have also beenraised after demonstrations of between-instrument variation.Recently, collaborative efforts between eight ISE instrumentmanufacturers and the National Institute of Standards andTechnology resulted in the development of a StandardReference Material, SRM 956, for the purpose of standard-izing direct-reading Na/K ISEs to the flame photometer.Other widely used technologies that provide noninvasive,continuous monitoring include pulse oximetry and trans-cutaneous gas electrodes. These trends are expected tocontinue and to produce a new generation of electrochemicaland optical sensors.

Additional Keyphrases: ion-selective electrodes referencematerials calibration electrochemical sensors elec-trolytes . pulse oximetry . transcutaneous gas electrodes

In recent years, laboratory testing in the critical-caresetting has increased, both in the numbers and variety oftests offered, and also in the types of individual performingthe tests. The acute-care or “stat” laboratory has evolvedfrom individual blood gas analyzer stations to an inte-grated, multidisciplinary laboratory offering testing forsurgical procedures and for emergency room and intensive-care-unit patients. The development of these laboratorieshas paralleled the technological advances in both labora-tory equipment and medical care. In a study reported in1984, Hall and Shapiro (1) surveyed 227 acute-care/bloodgas laboratories with regard to operational characteristics,personnel, analyses performed, and quality-assurance pro-cedures. All of the laboratories performed blood gas/pHanalyses and many also determined associated analyses,such as oxyhemoglobin, oxygen saturation, and hemato-crit/hemoglobin concentration; 18% also provided one ormore tests that are not directly associated with respiration,e.g., determinations of sodium/potassium, calcium, osmom-etry, and glucose. Of the 227 laboratories, only 130 (67%)used appropriate quality-control procedures for blood gas!pH. A greater percentage (84-100%) were performing ac-

Divi8ion of Laboratory Medicine, Mayo Clinic/Foundation,Rochester, MN 55905.

Received March 5, 1990; accepted May 16, 1990.

1562 CLINICAL CHEMISTRY, Vol. 36, No. 8(B), 1990

ceptable quality-control procedures for the other analytesmeasured. Burrin and Fyffe (2) surveyed 14 clinical bio-chemistry laboratories in the Northwest Thames region inEngland with regard to instrumentation, locations in thehospital, users, and quality assurance. Of the 22 blood gasanalyzers they found in use outside the clinical laborato-ries, most were located in intensive-care units or special-care pediatric units and were mainly used by the anesthe-sia staff or medical personnel. Little laboratory traininghad been given to any of the users of these instruments,despite the fact that the clinical laboratory was responsiblefor maintenance of 12 of the analyzers. For 21 of theinstruments, acceptable internal quality-control proce-dures were used, but none was enrolled in any externalquality-assessment scheme.

Blood Gas/Electrolytes

This trend in increased testing outside the traditionallaboratory and the emergence of the acute-care laboratoryis due, in part, to the evolution of electrochemical sensorsover the past three decades. These innovations have led toextended sensor lifetimes, reduced sensor maintenance, thedevelopment of single-use disposable sensors, and the useof small sample volumes. Today’s sensor technologies allowthe accurate and precise determination of several analytes,either at or near the bedside. The introduction of thecombination of blood gas/electrolyte analyzers allows rapid,direct analyses of various analytes simultaneously inwhole blood, reducing response time to 1-2 miii. Themultichannel systems are available from several compa-nies, including Nova Biomedical (Waltham, MA), CibaCorning (Medfield, MA), Instrumentation Laboratory (Lax-ington, MA), and Mallinckrodt Sensors (Ann Arbor, MI),and perform various combinations of tests, e.g., for pH,Pco’ Po2’ sOdimTi (Na), potassium (K), chloride (Cl), ionizedcalcium (Ca2), hematocrit, total hemoglobin, and glucose.The first three manufacturers design equipment that isbest suited for a laboratory setting, whereas the last oneproduces an instrument (Gem-Stat) suitable for the ward orpossibly the bedside. In Gem-Stat, exemplifying a unit-useformat, the calibrators, sensors, and waste container areprepackaged in a cartridge that can be used for 48 hanalyze as many as 50 individual specimens.

The appearance of these combination systems has raia new issue regarding calibration of the ion-selective electrodes (ISEs) used in these analyzers.1 The phosphabuffers developed at the National Institute of Stanand Technology (NIST) and used in most stand-alone bIgas systems are not acceptable for calibration in the multichannel analyzers in which pH and electrolytes are calibrated simultaneously (3). Binding of Na, K, and Ca24 bthe phosphate buffers reduces the Na, K, and Ca24 io

1 Nonstandard abbreviations: ISE, ion-selective electrodNIST, National Institute of Standards and Technology; NCCNationalCommittee forClinicalLaboratory Standards; and SStandard Reference Material.

activity coefficients. Use of the zwitterionic buffers, 3-(N-morpholino)propanesulfonic acid and 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (MOPSand HEPES, respec-tively) to eliminate this problem is necessary for simulta-neous calibration of pH and electrolytes. Moreover, theinterchange of calibrating solutions between blood gasanalyzers and multichannel analyzers even from the samemanufacturer may not be possible, and may cause errone-ous results.Use ofthesesystemsalsonecessitatesa reviewofsample

storage criteria and choice of anticoagulant. Misiano andMoran (4, 5) recently addressed the issue of sample trans-port and storage in relation to potassium values. Althoughstorage in an ice-water slurry is considered the ideal forblood gas determinations, it has adverse effects on potas-sium determinations.In theirstudy,clinicalspecimenswere sent un-iced to the laboratory and assayed. The excessspecimens were divided into two equal aliquots: one wasstored at room temperature, the other in crushed ice. Thespecimens were then analyzed at timed intervals, from 25to 80 mm later. As has been reported previously(6,7),storing whole-blood specimens on ice before measurementincreases potassium values by amounts that are bothclinically (0.14-0.40 mmolJL) and statistically significant(P <0.001 at all times measured). Individual specimensyielded differences in excess of 1.5 mmol/L. Other factors,such as instrumentation and interindividual variations,also contributed. The presence of cold agglutinins in asample that is iced before analysis may increase potassiumby one -or twofold, depending on how long the sample is inthe ice.

Although the effect of excess lithium and sodium heparinon ionized calcium values has been reported previously(8-10), its impact has become more widespread as ionizedcalcium sensors are added to combination blood gas/elec-trolyte systems. The commonly used 3-mL pre-packagedsyringes for blood gas determination contain between 75and 200 USP units of heparin; filled, the syringes containabout 25-66 USP units of heparmn per milliliter of blood.Urban et al. (9) suggest that sodium heparin concentrationin excess of 10-15 USP units/mL will cause significantbinding of ionized calcium. Calcium-titrated heparin orelectrolyte-balanced heparin (available from Radiometer

erica, Inc., Westlake, OH, and from Ciba Corning) willeduce this effect and can be used in concentrations up to 50t. units/mL (11). Studies of paired samples in our labora-ry compared ionized calcium values in 3-mL syringesntaiing dry sodium heparin at 25 USP units/mL vs

iquid calcium-titrated heparin at about 30 mt. units/mL.se of sodium heparin gave a consistent decrease in ionized

cium of 0.1 mmollL (Table 1).The number of direct-reading sensors for determination

f electrolytes, including lithium, has also increased dra-atically. These include traditional electrochemical sen-rs, as well as single-use disposable sensors. Among the

atter are the Kodak DT6O and E700 (Eastman Kodak,hester, NY) and the ChemPro (Johnson & Johnsonfessional Diagnostics, Roseville, MN). The Kodak

uipment can assay serum or plasma; the ChemPro ispable of assaying whole blood, serum, or plasma. Both

ystems use a multilayer approach, in which the tradi-ional electrochemical sensors are placed on planar sub-trates. As this and other similar technologies develop, we

ticipate several improvements, including multi-sensorays on single slides or cards, improved uniformity and

Table 1. Effect of Heparln Ofl 10Blood

nlzed Calcium In Whole

Heparin typeCalcIum2,

mmol/L pHcalcIum2, mmol/L

@pH 7.40

Calcium-titrated 1.33 7.37 1.32Sodium 1.22 7.37 1.20Calcium-titrated 1.25 7.38 1.24Sodium 1.14 7.37 1.13Calcium-titrated 1.30 7.36 1.28Sodium 1.20 7.36 1.18Calcium-titrated 1.25 7.36 1.24Sodium 1.15 7.36 1.13Calcium-titrated 1.25 7.42 1.25Sodium 1.16 7.40 1.16Calcium-titrated 1.29 7.36 1.26Sodium 1.23 7.34 1.19Calcium-titrated 1.21 7.42 1.22

Sodium 1.12 7.42 1.14a Calcium-titrated hepann was in liquid form (Radiometer, Copenhagen,

Denmark); its approximate concentration in blood was 25-30 i. units/mL.Sodium heparin was in dry form in prepackaged syringes (Sherwood Medical,St. Louis. MO); its approximate concentration in a filled syringe was 25 USPunits/mL.

reliability,and low cost.Calibration of the ISEs for electrolyte measurements has

been a controversial topic for many years. Most manufac-turers recommend using their own calibration standardsand control material, materials that range from simpleaqueous solutions to complex mixtures containing bovineserum albumin. Given the differences in calibration ap-proaches, it is not surprising that these instruments givedifferent electrolyte values for the same sample of humanserum. Gunaratna et al. (12) have shown that ISE-basedinstruments from different manufacturers give results forNa/K determinations that differ by 2-5%. In 1985, a re-search project sponsored by the National Committee forClinical Laboratory Standards (NCCLS) and funded byeight instrument manufacturers (Amdev, AVL, Ciba Corn-ing, Eastman Kodak, Instrumentation Laboratory, Kone,Nova, and Radiometer) was undertaken at NIST. The goalof the project was to develop standard reference materialsto be used by manufacturers and laboratories to bringconformity to direct potentiometric measurements of Na/K(12, 13). A series of four interlaboratory round-robin testswas performed by the eight manufacturers on their equip-ment and by four hospital laboratories. In each round-robinprotocol, between three and 11 candidate reference mate-rials, including aqueous standards, standards containingbovine serum albumin, and human-serum-based materials,including variations of NIST Standard Reference Material909, were assayed. The results indicated that use of anultraliltered human serum pool (at three concentrations)asa standard material for calibrating direct potentiometricISEs would bring conformity. among instruments fromvarious manufacturers. After post-calibrationwith thismaterial, interlaboratoryprecision was improved by two-to threefold for both Na and K. The new Standard Refer-ence Material, designated SRM 956, should be available in1990. Sodium and potassium values will be determined byboth Definitive and Reference Methods at NIST. It is hopedthat future efforts in this area will lead to development of

CLINICAL CHEMISTRY, Vol. 36, No. 8(B), 1990 1563

similar Standard Reference Materials for ionized calciumanalyzers.A recently publishedNCCLS ProposedStandard (C29-P)

recommends standardization ofdirect ISEs (those that donot requirepredilutionof the specimen) for Na/K to theflame photometricreferencemethod (14). It further recom-mends use of SRM 956 as the standard material, thuseliminating any variation due to differences in standardi-zation of flame photometers by the manufacturers. Indivi-dual laboratoriescan utilizeSRM 956 to verify calibrationby manufacturers and also to ensure comparability ofresults between direct ISEs, indirect ISEs, and flame pho-tometers in use in their institutions.

Continuous Noninvasive Monitors

Several technologies have been used clinically to providenoninvasive, continuous monitoring of Pa02 and PaCO2#{149}TWO

of the most prominent are pulse oximetry and transcuta-neous gas electrodes (15), both of which make their mea-surement through the patient’s skin and monitor the mea-surements continuously without requiring withdrawal of aspecimen. In addition, miniaturized noninvasive gas sen-sors have been developed for placement at the inner liningof the eyelid (conjunctival sensors) (16).

The development and clinical utility of transcutaneousmeasurements (termed p2 and Po) have been welldocumented (17-20). The sensors for these determinationsare the same as their counterparts used in conventionalblood gas analyzers, i.e., the Clark-style oxygen and theStow-Severinghaus-stylecarbon dioxideelectrodes. Bothindividual and combination electrodes are available (19).

In all cases, the sensors are heated at 43-45 #{176}Cto producea stable hyperemic flow of arterialized blood in the dermisand to reduce the diffusion barrier of the epidermis. The netresult is oxygen diffusion through the skin in concentra-tions resembling the Pa02 (15).

Although the transcutaneous electrodes are useful astrend monitors, one must recognize that transcutaneousblood gases are not the same as arterial blood gases.Several important factors affect the relationship betweenthe two. First, oxygen is consumed by the skin between thevascular bed and the transcutaneous electrode, which tendsto decrease the p2 measurement and increase theSecond, the electrode heats the underlying skin, whichtends to increase both p2 and Pteco#{149}For p these twofactors nearly cancel one another so tltat p and Pa02

similar, particularly in neonates. Forp, die two factorsare additive, making Ptco2 greater than PaCO, (21).

The use of transcutaneous gas electrodes has found wideacceptance for neonatal monitoring (20,22), and more re-cently in adults (23). In most patients with normal cardiacoutput, Pto2 follows the trend inpaoz and decreases relativetoPa as the patients’ ages increase (24). However, in thepresence of severely reduced cardiac output and peripheralperfusion, the p2 values will deviate from po2 values andbecome flow dependent, thus providing quantitative infor-mation regarding cardiac output (22). This has been testedexperimentally in dogs (25) and subsequently confirmed inclinical series (23). Similar diversions between Po2 andPaCO2 have been demonstrated (26). Therefore, a decreasein p2 may be due to a decrease in Pa02 or a decrease incardiac output. Thus, the p2 is an indicator of tissueoxygen delivery: arterial oxygenation and cardiac output(21). Clinically, this divergence is not an error to beavoided, but rather a highly significant noninvasive mdi-

1564 CLINICALCHEMISTRY,Vol.36, No.8(B),1990

cator of circulatory compromise (22).Several technical considerationsmust be observed whes

these sensors are being used: proper maintenance, membrane preparation and placement, calibration, and propeipreparation of the skin. Because of the potential for causingskin burns with the heated electrodes, some guidelineehave been developed. A temperature of 43.0 to 43.5 #{176}Cit

commonly used for premature infants, and the location olthe sensor on the skin is changed every 2 h. The electrodetemperature and length of time it can be left on the skiscan be increased as the patient matures and skin thicknestincreases. For newborns, a temperature of 43.5 #{176}Cfor 4 h itsafe; for children and adults, 44#{176}Cand 4-S h is safe (24),Several reports have evaluated the effect of halothane onthe transcutaneous oxygen electrode. Although some reported significant upward drift in the p reading (27),another report showed no clinically significant differencee(28). Clearly, electrode design and membrane compositionare important factors, and each sensor must be evaluatedindividually.

Pulse OximetryAnother of the continuous, noninvasive technologies in

wide use is pulse oximetry, which provides a measurementof arterial oxygen saturation (Sa02). The pulse oximetercombines the principles of spectrophotometric oximetryand plethysmography and functions by placing a livingtissue (usually the ear, foot, toe, hand, or finger) between atwo-wavelength light source and photodiode detector. Lightis then transmitted through the tissue at the two wave-lengths, about 660 and 940 nm, and monitored during eachpulse. The degree of change in transmitted light is propor-tional to the arterial pulse change, wavelength of lightused, and oxyhemoglobin saturation. Assuming that thepulsatile waveform is entirely due to the passage of arterialblood, one can calculate aO2 continuously (15). Most pulseoximeters also display pulse rate.

Although widely accepted in monitoring neonates, chil-dren, and adults, there are limitations to this technology. Iperfusion is poor, the pulse oximeter may not be ableadequately differentiate between arterial pulsation anbackground noise. In addition, the oximeters are subjectmotion artifact (21) and to other factors such as hypothermia, vasopressor drugs, and stray ambient lighting (29)

Because pulse oximeters use two-wavelength spectrophtometry, they measure only oxyhemoglobmn and deoxyhemoglobin and are unable to detect dyshemoglobins. Foexample, pulse oximetry is blind to the effects of carboxyand methemoglobin and will overestimate the true oxygesaturation in the presence of these hemoglobin derivativeFetal hemoglobin, however, apparently has little or nadverse effect (21). Severinghaus and Naifeh (30) assethe ability of six different pulse oximeters to accuratelmeasure low oxygen saturation (a2 <70%). They recordSaO, during rapidly induced plateaus of profound hypoxi(40-70% saturation) in normal volunteers. They not onifound between-instrument variation, but also determinthat some instruments displayed near normal saturatiowhen the in vitro measurement gave values in the range40-70%. Thus, the performance of pulse oximeters belo70% saturation remains marginal at best. Finally, bototal hemoglobin and the presence of dyshemoglobins mbe assessed before calculating oxygen content from a mesurement of by pulse oximetry (s02).

When used in the proper context, pulse oximetry offe

CLINICAL CHEMISTRY, Vol. 36, No. 8(B), 1990 1565

many advantages in the continuous monitoring of criticallyill patients: rapid and simple setup, no warm-up time, nocalibration, no sensor heating, and no hourly maintenance.Excellentcorrelationswith resultsfrom multiwavelengthco-oximetershave been reportedin very low-birth-weightinfants (31), neonates (32-34), children, and adults (35-37).In most cases, the study populations had SaO values be-tween 75% and 100%. Several studies have pound pulseoximetry to be an appropriate alternative top (31-33),particularly in situations of poor peripheral per?usion. Inaddition, saturation may be a more sensitive indicator ofoxygenation than is Pa02 in infants with relative hypox-emia, owing to the shape of the oxyhemoglobin dissociationcurve. On the steep portion of the curve, small changes in

Pa02 are associated with marked saturation changes (38).At the other extreme (the flat portion of the curve), largeincreases in p()2 may be associated with only smallchanges in Sa02, so that saturation may be a lesssensitiveindex ofoxygenation.Therefore,in acutelyillinfantswhomay be hyperoxemic, great caution must be exercisedinpredictingPa02 from S,,02. Walsh et al.(38) assessedtheabilityofpulse oximetry to predict PaO, in two groups ofinfants, one with chroniclung diseaseand the otherwithacutecardiorespiratorydisease. In theinfantswith chroniclung disease, Po2 derived from pulse oximetry was within10 mmllg of measured Pa02 in 73% of comparisons; in theacute group, calculated Po2 was within 10 mmHg of mea-sured Pa02 in only 50% of comparisons. Correcting theoxygen dissociation curves for the concentration of hemo-globin F in the acute infants did not change the results. Theauthors concluded that s and derived PaO, are useful ininfants with chronic lung disease, but emphasized that fullawareness of the limitations is important. In infants withacute cardiorespiratory problems, pulse oximetry unreli-ably reflects Pa02, but may be useful in detecting clinicaldeterioration.

Pulse oximetry has proven very useful for the continuousmonitoring of oxygenation. However, no data are availableto support the use of pulse oximetry as a replacement forarterial blood gases. When clinical conditions do not agreewith pulse oximetry readings, arterial analysis is war-ranted (15).

Additional Testing Methods

Reflectance meters for bedside glucose testing have beenused extensively for both emergency and routine monitor-ng as well as in the home. Several published reports have

ocumented their successful use (39-41), but a few haveported poor correlation between bedside determination of

lucose and the quantitative laboratory procedure (42,43).learly, these techniques are operator-dependent and re-uire careful training and monitoring to be successful.ost glucose meters utilize strips that incorporate theell-established glucose oxidase/peroxidase chemistry. In a

ently developed glucose meter system, The ExacTechMediSense, Cambridge, MA), the strip incorporates anlectron transfer mediator, ferrocene, which takes the placef oxygen in the reaction of glucose oxidase with glucose

d is detected amperometrically. This system is similar tohe single-use disposable electrodes previously discussed.

Lactate concentration in whole blood can also be deter-med by several analyzers available from YSI Inc. (Yellow

prings, OH). These systems utilize an enzyme electrodehnology in which the enzyme L-lactate oxidase is immo-

ilized in a thin membrane and placed over the sensor. The

enzyme catalyzes the conversion of L-lactate to pyruvateand hydrogen peroxide,the latter then being determinedamperometrically. The longest response time of these sys-tems, from sampling aspiration to result, is 1 mm.

In summary, it is clear that rapid development of newsensors and improvement of existing sensors will continue.The evaluation of single-use and unit-use disposable sen-sors has brought critical-care testing closer to the patient’sbedside. The combination multichannel blood gas/electro-lyte systems have allowed us to offer an array of criticalanalyses simultaneously, rapidly, and with enhanced pre-cision. In the future, we will undoubtedly see a continua-tion of these trends and the evolution of a new generationof electrochemical and optical sensors.

References1. Hall JR, Shapiro BA. Acute care/blood gas laboratories: profileof current operations. Crit Care Med 1984;12:530-3.2.Bun-in JM, Fyffe JA. Diagnostic equipment outside the labora-tory. J Clin Pathol 1988;41:925-8.3. D’Orazio P. The effects of HEPES and MOPS Zwitterionic bufferson the activities of sodium and potassium in multi-ISE calibratingsolutions and their contributions to the residual liquid junctionpotential at open KCIjunctions. In: Burritt MF, Cormier AD, MassAHJ, Moran RF, O’Connell KM, eds. Methodology and clinicalapplications of ion-selective electrodes. Vol. 9. Washington, DC:AACC, Electrolyte/Blood Gas Division, 1988.4. Misiano DR, Moran RF. Effects of storage temperature onwhole blood potassium measurements: implications in combinedanalyses ofblood gasesand electrolytes. In: Burritt MF, Moran RF,eds. Methodology and clinical applications of electrochemical andfiber optic sensors. Vol. 11. Washington, DC: AACC, Electrolyte/Blood Gas Division, 1990 (inpress).5. Moran RF, Misiano DR. Multichannel analysis of blood gas,electrolytes, and related analytes: implications of pre-analyticaltreatment of samples. Ibid.6. Tietz NW, Pruden EL, Siggaard-Andersen 0. Electrolytes,blood gases and acid-base balance. In: Tietz NW, ed. Texthook ofclinical chemistry. Philadelphia: WB Saunders & Co., 1986:110-29.7. Goodman JT, Vincent J, Rosen I. Serum potassium changes inblood clots. Am J Clin Pathol 1954;24:111-3.8. Ladenson JH, Bowers GN. Free calcium in serum. I. Determi-nation with ion-specifIc electrode, and factors affecting the results.Clin Chem 1973;19:565-74.9. Urban P, Buchxnann B, Scheidegger D. Facilitated determina-tion of ionized calcium. Clin Chem 1985;31:264-6.10. Biswas CK, Rarnos JM, Kerr DNS. Heparin effect on ionizedcalcium concentration. Clin Chim Acta 1981;116:343-7.11. Burns J, Law RWG, Paterson CR. Measurement of plasmaionised calcium: comparison of calcium-titrated sodium heparinand lithium heparin as anticoagulants. Med Lab Sci 1987;44:187-9.12. Gunaratna PC, Paule RC,Koch WF. Development of referencematerials for potentiometric sodium and potassium measurementsinserum. Op.cit. (ref. 3).13. Koch WF, Paule RC. Development of a standard referencematerial for ISE measurements of sodium and potassium. In: MaasAILJ, Buckley B, Manzoni A, Moran BY, Siggaard-Andersen 0,Sprokholt R, eds. Methodology and clinical applications of ion-selective electrodes. Vol. 10. IFCC Secretariat, 1989.14. Burritt MF, Fallon KD, Graham G, et al. Standardization ofsodium and potassium ion-selective electrode systems to the flamephotometric reference method. NCCLS ProposedStandard C29-P.Villanova, PA: NCCLS, 1989.15. Shapiro BA, Cane RD. Blood gas monitoring: yesterday, todayand tomorrow. Crit Care Med 1989;17:573-81.16. Abraham E. Noninvasive monitoring of cardiorespiratory pa-rameters. Emerg Med Clin North Am 1986;4:791-804.17. Lubbers DW. History of transcutaneous blood gas measure-ments. Crit Care Med 1981;9:693.18. Lubbers D. Theoretical basis of the transcutaneous blood gas

1566 CLINICAL CHEMISTRY, Vol. 36, No. 8(B), 1990

measurements. Crit Care Med 1981;9:721-33.19. Severinghaus JW. Transcutaneous blood gas analysis. RespirCare 1982;27:152-9.20. CassadyG. Transcutaneous monitoring in the newborn infant.J Pediatr 1983;103:837-48.21. Moran BY, Hess DR. Blood gases: the measurement of pH,P#{176}2,pCO2and related quantitives. In: Branson R, Hess D, Chat-burn R, eds. Respiratory care equipment. Philadelphia-Toronto:JB Lippincott Co., 1990 (in press).22. Shoemaker WC, Vidyasagar D. Physiological and clinicalsignificance of PtcO2 and PtcCO2 measurements. Crit Care Med1981;9:689-90.23. Tremper KK, Shoemaker WC. Transcutaneous oxygen moni-toring of critically ill adults, with and without low flow shock. CritCare Med 1981;9:706-.9.24. Tremper KK, Barker SJ. Transcutaneous oxygen measure-ment: experimental studies and adult applications. Int AnesthesiolClin 1987;25:67-96.25. Tremper KK, Waxman K, Shoemaker WC. Effects of hypoxiaand shock on transcutaneous P02 values in dogs. Crit Care Med1979;7:526-31.26. Tremper KK, Mentelos RA, Shoemaker WC. Effects of hyper-carbia and shock on transcutaneous carbon dioxide at differentelectrode temperatures. Crit Care Med 1980;8:608-12.27. Muravchich S.Teflon membranes do not eliminate halothaneinterference with transcutaneous oxygen electrodes [Abstract].Anesthesiology 1982;57:A168.28. Tremper KK, Barker SJ, Blatt DH, Wendei- RH. Effects ofanesthetic agents on the drift of a transcutaneous P02 sensor. JClin Monit 1986;2:234-6.29. Barrington KJ, Finer NN, Ryan CA. Evaluation of pulseoximetry as a continuous monitoring technique in the neonatalintensive care unit. Crit Care Med 1988;16:1147-53.30. Severinghaus JW, Naifeh KH. Accuracy of response of sixpulse oximeters to profound hypoxia. Anesthesiology 1987;67:551-8.

31. Ramanathan R, Durand M, Larrazabal C. Pulse oximetry invery low birth weight infants with acute andchronic lung disease.Pediatrics 1987;79:612-7.32. Durand M, Ramanathan R. Pulse oxinietry for continuousoxygen monitoring in sick newborn infants. J Pediatr1986;109:1052-6.33. Jennis MS, Peabody JL Pulse oximetry: an alternativemethod for the assessment of oxygenation in newborn infants.Pediatrics 1987;79:524-8.34. Sendak MJ, Harris AP, Donham RT. Use of pulse oximetry toassessarterial oxygen saturation during newborn resuscitation.Crit Care Med 1986;14:739-40.35. Fait CD, Wetzel RC, Dean JM, Schleien CL, Gioia FR. Pulseoximetry in critically ill children. J Clin Monit 1985;1:232-5.36. Fanconi S, Doherty P, Edmonds JF, Barker GA, Bohn DJ.Pulse oximetry in pediatric intensive care: comparison with mea-sured saturations and transcutaneous oxygen tension. J Pediatr1985;107:362-6.37. Tyler IL, Tantisira B, Winter PM, Motoyana EK. Continuousmonitoring of arterial oxygen saturation with pulse oximetry dur-ing transfer to the recovery room. Anesth Analg 1985;64:1108-12.38. Walsh MC, NobleLM, CarloWA, Martin RJ. Relationship ofpulse oximetry to arterialoxygen tension in infants. Crit Care Med1987;15:1102-6.39. Leroux ML, Desjardins PRE. Establishment and maintenanceof a hospital glucose meter program. Lab Med 1989;20:97-9.40. Belsey R, Morrison JI, Whitlow KJ, et al. Managing bedsideglucose testing in the hospital. J Am Med Assoc 1987;258:1634-8.41. North DS, SleinerJF, Woodhouse KM, et al. Home monitors ofblood glucose:comparison of precision and accuracy. Diabetes Care1987;10:360-6.42. Chu SY, Edney-Parker H. Evaluation of the reliability ofbedside glucose testing. Lab Med 1989;20:93-6.43. Cohen FE, Sater B, Feingold KR. Potential danger of extend-ing SMBG techniques to hospital wards. Diabetes Care1986;9:320-2.