R. C. Dennis1 and C. R. Valeri2. Materials and Methods

Transcription

1 CLN.CHEM (1980) MeasuringPercentOxygenSaturationof HemoglobinPercent Carboxyhemoglobinand Methemoglobinand Concentrationsof Total Hemoglobinand Oxygenin Bloodof Man Dog and Baboon R. C. Dennis1 and C. R. Valeri2 We used an automated four-wavelength spectrometer to measure the concentration of total hemoglobin percent oxyhemogiobin carboxyhemogiobin and methemoglobin and concentration of oxygen bound to hemoglobin in the blood of humans dogs and baboons under clinical and various experimental conditions. Measurements of total hemoglobin and methemoglobin with this simple method were comparable to those with standard spectrometric procedures. Carboxyhemoglobin measurements were comparable to those made with gas chromatography and measurements of oxygen content were comparable to those made with the galvanic cell method. The new instrument is as accurate as the comparison methods used to evaluate itinallparameters is reliable and measurements take only 63 s per sample. nadditionitrequires minimal operator training infrequent need for calibration and no sample preparation. Addtional Keyphrases: blood gases. spectrophotometry L 282 CO-Oximeter performance The classic method for measuring oxygen content in blood has been the van Slyke vacuum extraction and manometric procedure which although accurate requires great skill and considerable time for each measurement (1). The Lex-02- CON oxygen galvanic cell has been shown to give results that agree well with those by the van Slyke method but it also requires operator skill and an independent control standard (2 3). This paper reports the clinical and laboratory evaluation of a new simple method for measuring several hemoglobin species. The L 282 CO-Oximeter (nstrumentation Laboratory Lexington MA 02173) is an automated four-wavelength spectrometer designed for research the intensive-care unit and the clinical laboratory. n an anaerobic 300-FL sample of whole anticoagulated blood the instrument measures the concentration of total hemoglobin (T Hb) percent oxyhemoglobin (% 02 Hb) percent carboxyhemoglobin (% CO Hb) and percent methemoglobin (% Met Hb) and calculates the concentration of total hemoglobin-bound oxygen (Vol. % O2). A single digital thumb-wheel on the rear panel enables selection of spectral absorption coefficients for human dog or baboon hemoglobin. n addition no special sample prepara- 1 Department of Surgery Boston University School of Medicine 80 East Concord St. Boston MA Naval Blood Research Laboratory Boston University School of Medicine 615 Albany St. Boston MA Nonstandard abbreviations used: T Hb total hemoglobin concentration; % 02 Hb percent oxyhemoglobin; % CO Hb percent carboxyhemoglobin; % Met Hb percent methemoglobin; Vol. % 02 concentration of total hemoglobin-bound oxygen. The units and abbreviations used in this report are those that are digitally displayed by the instrument. Received Oct ; accepted May tion daily instrument preparation calibration or warm-up time is required. The instrument is self-cleaning and a total of 63 s is required for each blood sample. Only the T Hb channel requires periodic calibration and this is accomplished by sampling blood of known hemoglobin concentration by any reference method and turning a set screw until the L 282 CO-Oximeter digital readout is correct. The instrument utilizes a thalliumneon hollow-cathode lamp which emits light at the desired wavelengths. With activation of the sample button a 300-1zL sample of blood is aspirated into the instrument mixed with diluent at a ratio of nine parts sample to one part diluent hemolyzed and entered into a 125-tim pathlength cuvette thermostated to 37 #{176}C. Four spectral lines ( and nm) are isolated by means of interference filters mounted on a motor-driven filter wheel. Zeroing-solution blank absorbances are automatically measured and subtracted from the sample absorbance at each wavelength for each measurement. The built-in computer then solves four simultaneous equations using the measured absorbances and the species-specific coefficient matrix stored in permanent memory. The results are digitally displayed without need for manual calculations. An error-detection system alerts the operator of faults including the cuvette temperature not being within 37 ± 0.3 #{176}C the absorbance values being outside of expected ranges or a clot in the cuvette. Materials and Methods The L 282 CO-Oximeter was evaluated in a clinical and research setting by technicians with no formal training and only a few minutes instruction on its operation to determine if its ease of operation and minimal calibration and maintenance requirements could be matched by accurate results. Precision of Measurements Precision or reproducibility of the measurements was assessed by analysis in triplicate of 20 human blood samples collected anaerobically in heparinized 3- or 10-mL plastic syringes and stored at 4 #{176}C. Before each measurement the sample was mixed within the syringe by simultaneously rolling the syringe between the hands and turning it end-on-end for 5 mm. Total-Hemoglobin Measurement in Human Blood The instrument was calibrated for total hemoglobin concentrations according to the instruction manual with a Coleman Jr. Spectrometer (4) and the cyanmethemoglobmn method as the reference and used in a clinical study of patients undergoing coronary artery bypass surgery. Heparmnized samples of arterial mixed venous and coronary sinus blood were obtained anaerobically before during and after bypass from 14 patients over a four-month period. The instrument s total-hemoglobin channel was calibrated once during this period and confirmed six times CLNCALCHEMSTRYVol. 26No.91980

2 Table 1. Summary of Linear Regression Data for Comparison Methods (y) and the L 282 Co-Oxlmeter (x) Comparison Comparison method. y-ntercept Slope (SD) (SD) n r Hb human Coleman Jr.(cyanmethemoglobin) (0.020) 1.094(0.222) T Hb human CoulterHemoglobinometer(cyanmethemoglobin) 1.010(0.0003) (0.003) Vol.% human Lex 02 CON TL (0.007) (0.091) Vol.% dog (1.34 LexO2CONTL 1.036(0.010) (0.103) factor) Vol.% dog (1.39 Lex 02 CON TL 0.999(0.009) (0.100) factor) Vol. % baboon Lex 02 CON TL 1.023(0.007) (0.090) (1.39factor) Vol.% baboon Lex 02 CON TL (0.008) (0.090) (1.41 factor) % 02 Hb dog Lex 02 CON TL 1.006(0.005) (0.304) % Met Hb human Spectrometric 0-10% 0.943(0.024) 0.040(0.090) % Met Hb human Spectrometric0-80% (0.0002) (0.006) % CO Hb human Gas chromatograph0-17% (0.007) 0.809(0.058) % CO Hb human Gas chromatograph0-100% (0.013) (0.851) We compared the instrument with the Coulter Hemoglobinometer (Coulter Electronics Hialeah FL 33010) cyanmethemoglobmn method. Fresh heparmnized human blood was centrifuged and various combinations of cells and plasma were combined and run in random order on the L 282 CO- Oximeter (single measurements) and on the Coulter Hemoglobinometer (mean of duplicate or triplicate measurements). To determine same-day imprecision a sample of heparinized blood was run 20 times yielding a T Hb of (SD 0.08) gdl. Same-day imprecision of the other channels and between-day imprecision were not measured because the vanables being measured also vary with time. Total Oxygen Concentration n patients. Total oxygen concentration in blood from the patients undergoing coronary artery bypass surgery was measured with the L 282 CO-Oximeter and the Lex-02-CON TL oxygen galvanic cell (Lexington nstruments Waltham MA 02154). Arterial mixed venous and coronary sinus blood samples were obtained in heparinized plastic syringes before and after hemodilution providing a wide range of percent saturation (35-100%) and hemoglobin concentrations ( gdl). Blood samples were anaerobically analyzed in the L 282 CO-Oximeter the L 813 Blood Gas Analyzer and the Lex-02-CON TL. The dissolved oxygen concentration was calculated from the oxygen tension measured in the Blood Gas Analyzer by multiplication with the solubility coefficient of oxygen in plasma at 37 #{176}C (p2 in mmhg X ) and was added to the Vol. % 02 from the L 282 CO-Oximeter. Arterial oxygen tension on bypass often exceeded 300 mmhg (40 kpa). n dog and baboon blood. Fresh heparinized blood from two mongrel dogs was used to calibrate the T Hb with the Coulter Hemoglobinometer as a comparison method and with use of hemoglobin spectral absorption coefficients for the dog in the L 282 CO-Oximeter. The blood was centrifuged and various combinations of cells and plasma were recombined for each dog to vary hemoglobin. Samples were tonometered with use of the L 237 Tonometer and L 208 Gas Mixing Module against gas mixtures with oxygen concentrations ranging from 0 to 95%. Samples were simultaneously aspirated directly into the Blood Gas Analyzer and the L 282 CO-Oximeter through gas-impermeable fine-bore tubing immediately after withdrawing 0.5 to 0.6 ml into a 1-mL syringe for analysis in the Lex-02-CON TL. Dissolved oxygen was calculated and added to the L 282 CO-Oximeter value for hemoglobin-bound oxygen (Vol. %). A procedure similar to that described above was used for baboon blood after turning the species dial on the instrument to baboon coefficients. Percent Oxygen Saturation in Dog Blood The percent oxygen saturation of hemoglobin in dog blood was calculated from the measured Lex-02-CON TL total oxygen concentration minus the concentration of dissolved oxygen divided by the oxygen capacity of dog hemoglobin as calculated from the hemoglobin concentration multiplied by the oxygen-carrying capacity of dog hemoglobin. The calculated Lex-O2-CON TL percent oxygen saturation was compared to the percent saturation measured directly in the L 282 CO-Oximeter. Percent Carboxyhemoglobin in Human Blood Heparinized fresh and 24-h-old and outdated blood-bank blood samples were mixed with various amounts of carbon monoxide and analyzed in the L 282 CO-Oximeter and in a Hewlett-Packard Model 7620A Research Gas Chromatograph by a modification of the technique of Collison Rodkey and O Neal (6). Percent Methemoglobin in Human Blood Fresh heparmnized human blood was analyzed for methemoglobin by the RodkeyO Neal spectrometric method (7) and run in the L 282 CO-Oximeter. Various methemoglobin percentages from 1 to 70% were made by adding appropriate amounts of K3Fe(CN)s to aliquots of blood. Data Analysis Data comparing values obtained by the comparison methods to values obtained with the L 282 CO-Oximeter over a wide range of values were analyzed by use of a Hewlett- Packard System 45 Computer Non-Linear Regression Package. Slopes and intercepts plus or minus one standard deviation are given. Results Precision of Measurements f samples were well mixed in the syringe it was most Un- CLNCAL CHEMSTRY Vol. 26 No

3 . aus O ;.1 -a 6 Scc.. O.S.02O #{149} S..OOJ 20X71 2.r L 282 TOTAL HEMOGLOBN (dl) Fig. 1. Comparson of total hemoglobin concentration as measured n the Coleman Jr. Spectrometer by the cyanmethemoglobin method and inthe L282 CO-Oximeter Blood samplings were obtained from patients undergoing open-heart surgery over a four-month period usual to have triplicate T Hb readings in the L 282 CO-Oximeter vary by more than 0.1 gdl. Thus the L 282 CO-Oximeter was more precise than our comparison methods the cyanmethemoglobin method used in the Coleman Jr. Spectrophotometer and the Coulter Hemoglobinometer. The % 02 Hb % CO Hb and % Met Hb varied by less than 0.4% on triplicate determinations. Correlations between the reference values and the L 282 CO-Oximeter values for the various parameters are summarized in Table 1. The comparison of total hemoglobin measured in the Coleman Jr. Spectrophotometer and in the L 282 CO-Oximeter during a patient study over a four-month period is also shown in Figure 1. The L 282 CO-Oximeter T Hb channel was checked six times during this period but required no adjustment. A comparison of the Coulter Hemoglobinometer and the L 282 CO-Oximeter was done on one day L 282 CO-DCMETER O CONCENTRATON + DSSOLVED O(.%) Fig.3. Comparisonoftotal oxygen concentration as measured by thelex-02-con T and thel282 CO-Oxlmeter in patients undergoingopen-heartsurgery Samples of arterial mixed venous and coronary sinus blood were obtained before and after hemodilution from each of 14 patients. The y-intercept of is significantly different from zero (p <0.005) indicating that the L 282 CO-Oximeter. like the van Slyke apparatus gives a slightly higher total oxygen concentration than does the Lex-02-CON L with a range of hemoglobins from 2.3 to 24 gdl (Figure 2 and Table 1). Duplicate measurements were made in the Coulter Hemoglobinometer and single ones in the L 282 CO-Oximeter. We compared total oxygen concentrations in 14 patients undergoing open-heart surgery during a four-month period as measured by the Lex-02-CON TL galvanic cell and the L 282 CO-Oximeter over a range of total oxygen concentrations from 4 to 21 Vol. % hemoglobin concentrations from 6.2 to 15.2 gdl ( gl) and per cent saturations from 35 to 100% (Figure 3). The difference between pairs was significantly different from zero (p <0.005). Table 1 summarizes our total oxygen concentration mea r y1m $ N B 2 #{149} l P Sac t n.35 l V.4 4. V LEU OO(METER O GONCEJ(TRATOW + DSSOLVED 0 (VoLt) Fig.4. Comparison oftotaloxygen concentration in dog blood. obtained anaerobically and measured directly with the Lox- 02-CON and oxygen concentration measured in the L 282 CO Oxlmeterfromthe total hemoglobin concentration the percent 1L282 TOTAL. HEMOGLOeN(odL) oxygen saturation of hemoglobin the assumed oxygen-carrying capacity ofdog hemoglobinof 1.39mL of oxygen per gram of Fig.2. Comparisonoftotalhemoglobin concentration of human hemoglobin plus the dissolved oxygen bloodas measured inthe CoulterHemoglobinometerby the The y-intercept of s significantly less than zero (p <0.005) while the cyanmethemoglobin method and in the L 282 CO-Oximeter slope of is not significantly different from unity indicating that the Lex- Human blood samples were prepared with hemoglobin concentrations ranging 02-CON TL consistently gives slightly lower t ta oxygen concentrations than from 2.3 to 24 gdl the L 282 CO-Oximeter 1306 CLNCALCHEMSTRY Vol.26No.91980

4 L 282 O CONCENTRATON + DSSOLVED01(Vo.S) Fig.5. Comparisonofthetotal oxygen concentration n baboon bloodobtainedanaerobically and measured directly with the Lex-02-CON and measured inthel282co-oximeterfromthe total hemoglobin concentration the oxygen saturation of hemoglobin and the assumed oxygen carrying capacity of 1.39 ml of oxygen per gram of hemoglobin plus the dissolved oxygen When an oxygen carrying capacity value of 1.41 was used the slope was 1.00 Sc 2O ±0.99 Scc day old Blood 4 #{149} F.sfl m *admol. Blood #{149} 24Aold L282 %COHb SATURATON Fig. 7. Comparisonof percent carboxyhemoglobin up to 100% in human blood as measured by gas chromatography and with the L 282 CO-Oxlmeter The y-intercept is not significantly different from zero and the slope is not significantly different from unity surements in dog blood. The L 282 CO-Oximeter total oxygen concentration was calculated by the following formula: (% 0 Hb x [T Hb] X 1.34 ml 02g Hb) + ( X po2). The 1.34-mL factor yielded an r of a slope of and a y-intercept of n Figure 4 the same data were recalculated using the factor 1.39 ml 02g Hb and yielded an r of a slope of and a y-intercept of The factor takes into account the fraction of T Hb that can combine reversibly with 02. The y-intercept of is significantly different from zero (p <0.005). Figure 5 shows an oxygen-concentration analysis for baboon blood with use of the human factor of 1.39 ml O2g Hb in the calculations and yielded an r of a slope of and a y-intercept of When an oxygen carrying capacity of 1.41 ml 02g Hb was used a slope of 1.00 was observed with a y-intercept significantly less than zero (p <0.01). Methemoglobin was produced and was analyzed by a biochemical assay and in the L 282 CO-Oximer yielding an r C 6(- #{149} L 282 #{149} %M.thsmOgObn 6 r yntcrc#{149}pt ±006 j Sops n 36 Fig. 6. Comparison of percent methemogiobin up to 70% in human bloodmeasuredbythe Rodkey and O Neaspectrometric assay and with the L 282 CO-Oximeter. 7 value of with a slope of for levels to 70% (Figure 6). A comparison of % CO Hb in the gas chromatograph and the L 282 CO-Oximeter for levels of % CO Hb up to 100% gave an r value of and a slope of (Figure 7).4 Discussion The van Slyke apparatus has been the traditional method of measuring total oxygen concentration (both dissolved and hemoglobin-bound) but it is bulky contains mercury and its use requires considerable time and skill. The Lex-02-CON TL is simpler to operate but still requires skill up to 6 mm between samples and 30 to 45 mm to calibrate. n addition an independent reference method is required because the cell deteriorates with time (3). Nevertheless data from this laboratory and others have shown excellent agreement between results by the van Slyke and the Lex-O2-CON instruments although values obtained with the Lex-02-CON were consistently slightly lower (2 3). n the study reported here we compared the Lex-02-CON TL method of measuring total oxygen with the L 282 CO- Oximeter and found the Lex-02-CON values also to be lower (p <0.05) with a y-intercept of and a slope of (Figure 3). The L 282 CO-Oximeter displays the calculated Vol. % 02 from the measured [T Hbj X (% 02 Hb) X 1.39 when the unit for T Hb is gdl and 1.39 mlg is the oxygen carrying capacity of human hemoglobin programmed in the machine. The Vol. % 0 displayed does not include dissolved 02. Therefore to determine total oxygen concentration dissolved oxygen must be calculated and added to the hemoglobin oxygen content read from the L 282 CO-Oximeter. The factor 1.39 mlg can be simply changed by a service representative if desired. Figure 4 and Table 1 show that the assumed oxygen-carrying capacity of 1.34 ml 02g Hb for dog blood produced too large a slope (1.036) while 1.39 mlg gave a slope of near unity (0.999) indicating that the oxygen carrying capacity of dog blood is 1.39 ml 02g Hb. For baboon blood use of the 1.39 ml 02g Hb value gave a slope of (Figure 5) whereas a Figures shcwing details of those relationships not illustrated here are available from the Editorial Office of this journal or from the authors. CLNCAL CHEMSTRY Vol. 26 No

5 factor of 1.41 ml 02g Hb gave a slope of 1.00 suggesting that the oxygen carrying capacity of baboon hemoglobin is 141 ml of oxygen per gram. Good sample mixing was mandatory to obtain accurate T Hb requiring vigorous twisting-rotation motion of the syringe for a full 5 mm; lack of a gas-blood interphase or a washer in the syringe hindered rapid mixing of the anaerobic blood sample. nadequate mixing improper T Hb calibration or a change of pump windings without recalibrating T Hb gave inaccurate T Hb readings but because % 02 Hb % CO Hb and % Met Hb are ratios they remain accurate independent of T Hb calibration or poor sample mixing. During the four-month patient study the L 282 CO-Oximeter remained calibrated for T Hb without any adjustments or changes of pump windings and there were no mechanical failures. There are no adjustments for % 02 Hb % CO Hb or % Met Hb. Unlike the earlier L 182 CO-Oximeter the L 282 CO-Oximeter has a thalliumneon hollow-cathode light source which emits light of discrete wavelengths allowing the use of much more stable light filters. Also unlike the older Model 182 little interference between channels was noted. As carbon monoxide was varied between 0 and 80% % Met Hb remained stable and % CO Hb was stable as methemoglobin was altered from 0 to 70%. As % 02 Hb was varied between 0 and 98% % CO Hb increased from 1.5 to 2.2%. Total hemoglobin values measured in the Coleman Jr. Spectrometer had more scatter than total hemoglobin values measured in the L 282 CO-Oximeter and the Coulter Hemoglobinometer. The human hemoglobins studied were normal. We have had no experience with abnormal hemoglobins in the L 282 CO- Oximeter. Measurement of plasma hemoglobin concentrations ranging from 0.7 to 7.3 gdl (7-73 gl) and of plasma methemoglobin concentrations ranging from 7 to 53% of total plasma hemoglobin with the L 282 CO-Oximeter has been reported by Sehgal et al. (8) with excellent correlation with comparison methods. Currently the L 282 CO-Oximeter costs $11900 and about 2.5 per measurement for reagents. n addition tubing now costing $35.00 must be replaced about every three or four months. Our data indicate that the L 282 CO-Oximeter is more precise and as accurate as our comparison methods for all variables measured is reliable yet requires no great skill to operate only infrequent calibration no manual calculations and no sample preparation. This work was supported by the U. S. Navy (Naval Medical Research and Development Command Research Task No N- MR and Office of Naval Research Contract No. N C-0168). The opinions or assertions contained herein are those of the authors and are not to be construed as official or reflecting the views of the Navy Department or the Naval Service at large. We acknowledge the technical assistance of J. Joens and J. Mostacci. References 1. van Slyke D. D. and Neill J. M. The determination of gases in blood and other solutions by vacuum extraction and manometric measurements. J. Biol. Chem (1924). 2. Valeri C. R. Zaroulis C. G. Marchionni L. and Patti K. J. A simple method for measuring oxygen content in blood. J. Lab. Clin. Med (1972). 3. Clerbaux T. H. Frans A. Detry J. M. and Rousseau M. Drift in the oxygen content measured with the Lex 02 CON: Long-term assessment. J. Lab. Clin. Med (1976). 4. Crosby W. H. Munn J.. and Furth F. W. Standardizing a method for clinical hemoglobinometry. U.S. Armed Forces Med. J (1954). 5. Sendroy J. Jr. Dillon R. T. and van Slyke D. D. Studies of gas and electrolyte equilibria in blood. XX. The solubility and physical state of uncombined oxygen in blood. J. Biol. Chem (1934). 6. Collison H. A. Rodkey F. L. and O Neal J. D. Determination of carbon monoxide in blood by gas chromatography. Clin. Chem (1968). 7. Rodkey F. L. and O Neal J. D. Effects of carboxyhemoglobin on the determination of methemoglobin in blood. Biochem. Med (1974). 8. Sehgal H. L. Sehgal L. R. Rosen A. L. et al. Sensitivity of the L 282 C0-Oximeter to low hemoglobin concentration and high proportions of methemoglobin. Clin. Chem (1980) CLNCAL CHEMSTRY Vol. 26 No