International Journal of Clinical Monitoring and Computing9: 45-52, 1992, 9 1992KluwerAcademic Publishers. Printedin the Netherlands.
Clinical evaluation - continuous real-time intra-arterial blood gas monitoring during anesthesia and surgery by fiber optic sensor Bradley E. Smith, Paul H. King & Les Schlain Department of Anesthesiology, Vanderbilt University Medical Center, Nashville, Tennessee and Optex Biomedical, Inc., The Woodlands, Texas Accepted 13 January 1992
Key words: clinical monitoring, anesthesia, analysis of blood gases Summary
A clinical evaluation of the clinical utility, techniques of use, durability, accuracy, and potential complications of a newly available system for the continuous real-time, intra-arterial monitoring of arterial blood gas and acid base status (ABG) has been studied (Optex BioSentry | System, Optex Biomedical, Incorporated, The Woodlands, Texas, U.S.A.). The system consists of three separate fiber optic channels with contained fluorescent and absorbant chemical analytes imbedded in a single probe and capable of insertion inside of a twenty gauge indwelling arterial catheter (The Optex Optode | Sensor), along with an external monitor. The system was utilized during anesthesia and surgery in the care of five informed patients. Constant trend monitoring of all three variables was deemed satisfactory in four of the patients. The fifth sensor was damaged by a surgical assistant while in place and ceased to function. Comparison of Optode | sensor readings with standard clinical laboratory ABG analysis was excellent in three uncomplicated patients (pH: bias - 0.0183 pH units, precision: 0.0237 pH units; PCO2: bias 3.22 mmHg, precision 2.04 mmHg; PO2: bias - 5.94 mmHg, precision 11.74 mmHg). Postoperative study suggested that discrepancies in a fourth patient may have been due to an incorrect 'offset' applied to the Optode | sensor yielding a constant error.
Introduction
Although frequent analysis of arterial PO2, PCO2, and pH measurements and the immediate application of this knowledge to real-time management of patients undergoing anesthesia and surgery has become routine during the past two decades, the standard procedure frequently entails lag time in the presentation of data and has become extremely expensive in actual hospital applications [12]. The ability to continuously monitor moment to moment changes in A B G has long been desired for use in the operating room and in critical care units, and would provide obvious advantages in the care of a wide range of acutely and critically ill patients.
Toward this objective, several proposed devices have appeared in the past two decades, but have thus far presented technical limitations or have been withdrawn from the market for economic reasons. These include miniaturized electrochemical biosensors and various fiber optic techniques, Electrochemical sensors often represent a combination of two transducers, one biochemical, tile other physical, in intimate contact with each other. These relate the concentration of an anatyte to a measurable electrical signal [11]. In clinical use, various models of these types have proven susceptible to damage and particularly to calibration drift. Attempts at clinical application of such devices have included polarographic oxygen electrodes [6],
46 electrodes based upon the Severinghaus model [7], and intravascular placement of fiber optic reflectance devices estimating oxygen tension from oxygen saturation [9]. Except for the latter, most have, at this time, been abandoned [2]. Great interest is currently centered on promising newer technologies based on various designs incorporating transmission of light through fiber optic fibers and manipulation of the transmission of this light by chemicals, usually absorbant or fluorescent, whose characteristics change in proportion to changes of pH, PCO~, and/or PO2 [14, 4]. The instrument utilized in this evaluation is of the latter type. Preliminary studies of the utilization of somewhat similar instruments for the constant measuring of intra-arterial blood acid base and physiologic gas status have been carried out in dogs [3], in human volunteers [8], in a small group of critical care patients [13] and in fourteen patients undergoing anesthesia and surgery [2]. Satisfactory human trials of the device utilized in this study extending up to ninety-six hours of use have been carried out in a limited number of patients in the critical care setting (Optex Biomedical, The Woodlands, Texas, Personal Communication), however, this is the first evaluation of the use of this particular device during anesthesia and surgery.
Methods
Evaluation of the clinical utility, techniques of use, durability, accuracy, and potential complications associated with the use of the Optex BioSentry | Blood Gas Monitoring System an Optode | sensor placed in the radial artery were carried out in five consenting surgical patients.
Patient number one
A forty year old male weighing 72kg, height 165 cm, with an ASA Classification of II [1] underwent a craniotomy for cranioplasty. The duration of anesthesia was 239 min. Anesthesia consisted of sufentanyl 50 tzg, thiopental sodium 500 mg, vecuronium 13 mg, midazolam 2 mg, and isoflurane in-
haled from 0.5 to 0.75% via oral-tracheal tube. There were no complications of anesthesia or surgery, however, the blood pressure rose from 120 systolic to 160 systolic during the final fifteen minutes of emergence from anesthesia.
Patient number two
A sixty-one year old female weighing 70 kg, height 170 cm, with an ASA Classification of Ill also experienced a craniotomy procedure for cranioplasty. The duration of anesthesia was 131min and anesthesia consisted of sufentanyl 25/xg, thiopental sodium 250mg, vecuronium 10mg, labetolol 20 mg, esmolol 15 mg, ephedrine 5 mg and isoflurane from 0.5 to 1.38% inhaled via an oral-tracheal tube. Intraoperative complications consisted of intermittent hypertension as high as 190/110 early in the case, treated by blockers but recurring briefly during emergence.
Patient number three
A seventy-nine year old female weighing 65 kg, height 165 cm, with an ASA Classification of III underwent laparotomy for total abdominal hysterectomy and bilateral salpingo-oophorectomy. The duration of anesthesia was 157 rain and anesthesia consisted mainly of a continuous epidural block consisting 100mg of 0.5% bupivacaine. Other drugs included thiopental sodium 250 mg, vecuronium 5mg, midazolam 3mg, phenylephrine 300/xg, and approximately 60% nitrous oxide inhaled via an oral-tracheal tube. Intraoperative complications consisted only of brief hypotension to approximately 77 mmHg systolic over 30 diastolic on three accasions. During emergence from anesthesia, the blood pressure arose to 150/70.
Patient number four
A forty-two year old female weighing 50 kg, height 170 cm, with an ASA Classification of II also experienced an exploratory laparotomy, total abdom-
47
/
Clear L u e r
pH and P C O 2 During Anesthesia
Lock
/
Patient # 1 Connector
7.6
755
Sensors
...............
. . . . . . . . . . . . .
f ~
~
/ ~ - ~
5O
C = Change in ventilator 55
0 = O[fset
,~_ C
~
C
l 50
L/
I Arrow
20
Io_
f4
o
g35
~)
73
725-
45
40 o h35
i
~i
' 3O
g. Catheter
25
Fig. 1. The Optode | Sensor (Optex BioSentry| Monitor, The Woodlands, Texas, United States of America).
T i m e (hours) O p t o a e TM pH L a b o r a t o r y pH
inal hysterectomy, and bilateral salpingo oophorectomy. The duration of anesthesia was 129 min and anesthesia consisted of thiopental sodium 125 mg, succinylcholine 100 rag, nitrous oxide up to 65% mixed with up to 1.34% isoflurane inhaled via an oral-tracheal tube. Intraoperative complications consisted of very brief hypotension as low as 65/55 on two occasions with no apparent clinical impairment.
Patient number five A sixty-three year old male weighing 58 kg, height 160 cm, with an ASA Classification of III experienced right carotid endarterectomy. The duration of anesthesia was 110 min and anesthesia consisted of propofol 140 rag, fentanyl 300 txg, nitrous oxide up to 50% mixed with isoflurane up to 0.5% inhaled via an oral-tracheal tube. In addition, phenylephrine 200/xg, ephedrine 200 mg, and sodium nitroprusside 200 ~g, were utilized in the patient's care. Intraoperative complications consisted only of hypertension early in the procedure which was well controlled during the latter part of the procedure. In all patients, the radial arterial puncture was carried out in the usual sterile manner under local anesthesia before induction of general anesthesia. Before establishment of the intra-arterial sensor, the Optode | sensor (Fig. 1) was calibrated in a sealed container with a sterile chemical solution and was equilibrated with appropriate reference gases. After warm up and cessation of drift while in the reference solution, the sensor was ready for
+
O p t o d e rv PCO~
- -
L a b o r a t o r y PCO 2
Fig. 2. pH and PCO2 During Anesthesia.
placement (Fig. 2). Attention was given to placement of the catheter in such a manner that the sensor experienced a straight course of insertion and was not obliged to bend. This was carried out easily in all five patients. After establishment of the arterial catheter and placement of the sensor, a formed plastic splint was added external to the forehand in order to protect the sensor. After establishment of general anesthesia and when the patient had achieved a stable anesthetic level, an ABG sample was obtained from each patient via the port provided at the base of the Optode | sensor device. Withdrawal of the blood sample was unimpeded and was carried out in a routine fashion. Constant radial artery pressure monitoring was easily accomplished through the same port in all five patients. The time of withdrawal of the sample was noted and when the report of values from the clinical laboratory value became available, an offsetting current was added to the BioSentry| monitor for the purpose of bringing the values on the BioSentry| monitor (pH, PCO2, and PO2) into agreement with the clinical laboratory readings. Thereafter, additional offset adjustments could be made, but have not been required due to the demonstrated stability of the Optex BioSentry| system.
Results The Optex BioSentry| system appears to be reia-
48 tively simple in clinical application. No problem was encountered in any of the five evaluations with initial warm up, calibration, or stabilization of the sensor in the standard solution. Insertion of the sterile, calibrated, and stabilized sensor into the radial artery presented no problems in any of the five patients. After insertion into the patient's radial artery, stabilization appeared to take place in approximately twenty to thirty minutes. Figure 1 demonstrates the plotted Optode | sensor readings of pH and PCO2 in Patient Number One. The readings are updated every thirty seconds and can be obtained as digital data and in a trended presentation simultaneously. Figure 1 also illustrates comparisons of pH and PCO~ from arterial samples analyzed by the Hospital Clinical Laboratory. Similar trend graphs were prepared for pH, PCO2, and PO2 in all five patients and in all of the first four patients trending closely approximated and followed clinical laboratory values of the extracted samples. Table 1 displays all data for all five patients,
including clinical laboratory ABG values, Optode | sensor readings at the same time and the difference between the two systems for each variable. The course of monitoring appeared to proceed without incident in Patients Number One and Two except that it was noted that there was a far greater variation between clinical laboratory and Optode | sensor values in the ABG samples taken during emergence from anesthesia. The monitoring course in Patient Number Three was similarly uncomplicated, but no arterial blood gas sample was collected during emergence from anesthesia. Patient Number Four, displayed greater discrepancy between the ABG and the Optode | sensor readings. However, sensor trending seemed to be stable and representative. On post-operative study, it was noted that the ABG sample for the original 'offset' reading was inadvertantly drawn during a period of clinical hypotension. A postulation was offered that the ABG drawn at this time represented a transient due to the physiologic disturbance and that the sensor had not yet equili-
Table 1. Arterial Blood Gases (ABG) During Anesthesia and Surgery. Patient No.
1
2
3
Time
Clinical Laboratory Optode | Sensor
Difference
pH
PO,
pH
PCO,
PO~
pH
PCO~
Comments PCO,
PO,
9:20 a m
7.46
3l
131
7.45
31.9
134.1
-0.1105
0.948
9:40 a m
7.46
29
154
7.47
30.9
1411.3
0.01/5
1.845
3.10
10:05 a m
7.47
25
171
7.48
29.6
149.3
0.008
4.639
- 21.712
[0:31 a m
7.47
25
172
7.49
29.6
145.0
1/.015
4.631
-26.984
10:54 a m
7.49
29
156
7.49
29.6
143.7
0.111t3
0.563
- 12.285
ll:17am
7.33
411
166
7.40
4 l. 1
181.9
0.070
1.049
1:33 p m
7.39
31
167
7.37
37.9
160.9
- 1/./120
6.91
- 6.08
l:57pm
7.37
35
154
7.33
41.4
156.6
- 11.042
6.42
2.62
2:19pm
7.38
40
157
7.32
41.7
145.5
- 0.060
1.74
- 11.55
2:41pm
7.50
26
371
7.41
29.l
227.1
- 0.086
3.05
- 143.91)
9:13 a m
7.44
29
109
7.41
31.0
112.2
- 0.030
2.03
3.21
9:43 a m
7.45
28
107
7.41
31.1
ll8.11
-0.036
3.12
11.05
Uncomplicated
- 13.650
15.875
10:14am
7.46
27
114
7.42
29.6
121.0
- 0.039
2.59
6.96
4
12:44 p m
7.35
39
168
7.38
38.8
125.9
0.034
- 0.19
- 42.07
l:18pm
7.35
39
148
7.45
35.4
1211.9
0.102
- 3.59
- 27.15
5
8:32am
7.3l
37
263
7.30
36.6
286.0
- 0.010
- 0.40
23.00
8:43am
7.31
411
256
7.29
31.9
272.0
- 0.018
- 8.12
16.113
291.9 290.7
- 0.029 - 1t.1151
- 4.27 - 2.38
72.90 57.70
9:15 a m
7.32
42
9:47 am
7.33
40
219 233
7.29 7.28
37.7 37.6
10:15 a m
7.31
41
243
7.26
40.2
291.8
- 11.048
- 0.84
48.80
10:43 a m
7.36
40
226
7.31
39.6
284.8
- 0.045
- 0.45
58.84
(emergence A BG) Uncomplicated
(emergence ABG) Uncomplicated
(no emergence A BG) Offset Problem Sensor Damage
49 brated with this change, resulting in an incorrect offset. This may have given rise to a constant error in absolute comparison of A B G and sensor values. In Patient N u m b e r Five, insertion of the Optode | sensor was uneventful. After 'offset' (8:32 am), the next comparison of A B G and sensor values (8:43 am) appeared to be reasonably accurate. Soon thereafter, a surgical assistant was observed to lean on the site of the sensor with significant weight. Thereafter, the sensor values were greatly different from the A B G values. After the end of anesthesia and surgery the sensor was withdrawn and found to be grossly intact, but on magnified inspection the fiber optic path was indeed damaged. Statistical analysis of the comparison of corresponding A B G laboratory readings with Optode | sensor readings is presented in Table 2 and Table 3, 'Bias' is defined as the mean of differences between sensor and arterial values and 'Precision' is one standard deviation (1SD) of the differences [8]. Table 2 presents comparison which exclude the A B G gas samples taken during anesthetic emergence and eliminates patient four and five who experienced known problems. Table 3 presents statistical analysis of the same data but also includes A B G samples drawn during anesthesia emergence. The senior author physician examined all patients from the second to the fourth day post-anesthesia and surgery and found no evidence of complications of the use of the sensor either systemically or at the local site of insertion.
Table 2. Optode | Sensor - Clinical Laboratory Agreement Uncomplicated Patients*.
pH
PCO,
OPTICAL SENSOR SYSTEMS
\
i n/ ............ ~aem~ra~e t , ct ~ e
i
Fig. 3. Optical Sensor Systems.
Discussion
The Optex BioSentry | system operates on principle similar to other fiber optic sensor systems. An optical sensor indicates variation in the surrounding blood by changes in either the color or intensity of the light passing through the sensor. The sensor is in contact with the surrounding medium through a semipermeable m e m b r a n e and the light pathway is lead through the chemical c h a m b e r passing then to a detector, then to a signal processor and finMly to a user interface. A diagram of such a system is provided in Fig. 3. Designs vary between manufacturers, and in some cases parallel light channels or reflection techniques through monofiber channels are utilized. Some designs [10, 4] depend on reflectance of the light from the indicator complex which is located at the tip of a catheter sensor, Fiber optic systems are inherentiy immune to electromagnetic noise, can be extremely small, and provide electrical isolation of the patient.
Table 3. Optode | Sensor - Clinica~ Laboratory Agrecmem Uncomplicated Patients*
PO2 pH
Bias 1SD Precision # of ABGs
- 0.0183 0.0237 11
3.22 2.04 11
* Excludes emergence period. Eliminates Patient 4 with erroneous offset. Eliminates Patient 5 with damaged sensor,
s ire
- 5.94 11.74 11
Bias ISD Precision # of ABGs
- 0.0167 0.0377 13
PCO~
PO~
3.04 [.97 13
- 14.87 39,2! 13
* Eliminates Patient 4 with erroneous offset. Eliminates Patient 5 with damaged sensor.
50 FIBER
OPTIC
- pH
SPECTROSCOPY
-
PRINCIPLES OF OPERATION OF OPTODE T M SENSOR
CHEMICAL Anaiyte
Chemistry of amphoteric pH indicators light ~
1
indicator
Analyte
light
)
pH
= -Log(K a ) + Log[(A)/(HA)] intrinsic physical constant of indicator concentration of ionized indicator concentration of neutral indicator
a
(,4-) = I - INDICATOR
(HA) =
Fig. 6. pH - Chemical Principles of Operation of Optode|
Sensor.
Dual Reflection
Monofiber
Transmission
Fig. 4. Fiber Optic Spectroscopy.
Figure 4 diagrams several alternative system configurations. The Optode | sensor incorporates a 180 degree bend at the tip of the optical fiber, which allows transmission of light through the analyte chamber and also provides for placement of the analyte chamber away from the tip of the sen-
- pH -
sor, thus reducing reflection interference and other problems of tip mounted indicator chambers. The Optex Optode | sensor utilizes principles of classical UV-visible spectrophotometry as illustrated by the statement of the Beer-Lambert Equation in Fig. 5. The Optex system uses light absorbance by indicators for the detection and measurement of pH and PCO~. A pH indicator react by absorbing, or blocking, light of a particular wavelength in proportion to the pH of the sensor environment. The ratio of the intensity of light projected into the sensor to the intensity of light emitted from the sensor is used to predict the pH. As CO2 enters the sensor through a gas permeable membrane it equilibrates with a carbonic acid/bicarbonate buffer system. The pH of this system, as predicted from the absorbance of the indicator, is used to predict PCO~. The chemistry of amphoteric pH indicators is further illustrated in the equation shown in Fig. 6
CHEMICAL PRINCIPLES OF OPERATION OF OPTODE TM SENSOR
- PCO Beer-Lambert Equation (U.V.- visible spectrophotometry)
PRINCIPLES OF OPERATION OF OPTODE T M SENSOR
CHEMICAL
-Log(l o/li) = el(c) 1o 1~ e l (c~
= = = = =
wavelength of light entering sample intensity of light transmitted through the sample intrinsic physical constant of chemical interest path length of light through sample concentration of sample
2 -
Henderson-Hasselbach
equation
p H = -Log(K a ) + Log[(HC03 )/0.03PC021
Fig. 5. pH - Chemical Principles of Operation of Optode |
Fig. 7. PCO, - Chemical Principles of Operation of Optode |
Sensor.
Sensor.
51
-
PO
z -
CHEMICAL PRINCIPLES OF OPERATION OF OFFODE TM SENSOR Stern-Vollmer equation (fluorescence quenching) - L o g ( I o IIs ) - I
= K,,(PO, )
Io
=
intensity o f light emitted from a compound in the
1i
=
intensity of fluorescent light emitted at an unknown P02
K,v
=
an intrinsic physical constant of the fluorescent indicator compound
absence of oxygen
Fig. 8. PO, - Chemical Principles of Operation of Optode@
Sensor. and the familiar chemistry of the Henderson-Hasselbach equation utilized in detection of CO 2 is illustrated in Fig. 7. The measurement of oxygen in the Optex system is dependent on the quenching of fluorescent dyes by the presence of oxygen. When stimulated with ultraviolet light, certain dye molecules may obtain enough energy to emit visible light. Oxygen molecules can prevent emission by stealing some of this energy during collisions with the dye. The fluorescence emission of quenchable dye in a sensor is related to the amount of stimulus light projected into the sensor and the amount of oxygen present in the sensor. The principles of 'fluorescence quenching' utilized in oxygen detection which are illustrated in the Stern-Vollmer equation are represented in Fig. 8. Exaggerated s e n s o r - ABG differences noted during emergence might be due to a lag in equili-
bration by the Optode | sensor, however, the variance is in opposite direction in Patients One and Two. Others have postulated a difference in actual ABG at the sampling site in the radial artery from blood in the ventricle [8] and this phenomenon might be exaggerated during the excitement of emergence. However, interaction between the sensor and the arterial wall might also be exaggerated during the vascular reactivity and hypertension noted in both patients. Table 5 lists expected advantages of the Optex Optode | sensor system, which include support of the analyte on all sides, and improved strength and integrity of the analyte chamber. The side placement of the membrane provides more surface area for interaction than on tip mounted sensors. Side versus tip placement may reduce thrombogenicity and inaccuracies introduced by localized metabolic processes in the thrombus noted by previous investigators [8] and the design may reduce sensorarterial wall interaction which has also been encountered by others [8]. Complete separation of the analyte containing sites in three separate fiber optic paths reduces the risk of light leak and reflectance interference between analyte chambers. An increased selectivity and sensitivity of the transmission technique versus the light reflective technique is expected. Internal reflective interference is expected to be reduced due to separate transmission cells. Furthermore, the sensor body is of high light transmission capability, therefore allowing more light to be received distal to the cell at the detector. Side versus tip placement of the membrane may reduce thrombogenicity.
Table 4. Advantages of Optex | Light Transmission System.
1 Mechanical 2 Chemical 3
Optical
4
Clinical
material suppor! separation of sites reduces cross reactivity increased sensitivity reduction of reflective interference high transmission of sensor body probe mechanically stable side window is less thrombogenic improved accuracy, less chemical or physical interference
Table 5. I)ritt of Optode | Sensor*.
Analyte Average Drift Per Hour
Deviation After 12 Hours
pH PCO, PO,
0.025 pHu 10 mmHg 10 mmHg
0.002 pHu 0.8 mmHg 0.8 mmHg
* Stable controlled pH, PCO> PO2, temperature, etc.
52 Conclusions A preliminary evaluation of the clinical usefulness, reliability and potential complications in application of a new continuous real-time intra-arterial blood gas monitoring system (Optex BioSentry | Optex Biomedical, Inc., The Woodlands, Texas, U.S.A.) has been carried out in five consenting anesthetized patients during surgery. Trend monitoring of pH, PCO2, and PO2, was deemed satisfactory in four of the five patients (In the fifth patient, the sensor was damaged by the weight of a surgical assistant). Comparison of Optode | sensor readings with clinical laboratory ABG readings was excellent in three uncomplicated patients. However, in a fourth patient, insertion of offset to the sensor based on A B G sampled during clinical hypotension may have inserted a small constant error in later Optode | sensor readings. Further evaluation is indicated to better understand the relationship of equilibration time of the sensor to physiologic changes, and to the variables which may affect delivery of blood representative of the current physiologic state to the sensor sampling site. Further evaluations should also reveal whether the one damaged sensor indicates fragility of the system in clinical use.
Acknowledgements
Supported in part by a contract from Optex BioMedical, Inc., The Woodlands, Texas, United States of America.
References 1. American Society of Anesthesiologists. New classification of physical status. Anesthesiology 1963; 24: 111.
2. Barker SJ, Hyatt J. Continuous measurement of intraarterial pHa, PaCO,, and PaO~ in the operating room. Anesth Analg 1991; 73: 43-8. 3. Barker SJ, Tremper KK, Hyatt J, Zaccari J, Heitzmann HA, Holman BM, et al. Continuous fiberoptic arterial oxygen tension measurements in dogs. J Clin Monit 1987; 3: 48-52. 4. Barnard SPM, Walt DR. A fiber-optic chemical sensorwith discrete sensing sites. Nature 1991; 353: 338-40. 5. Ouatro Pro 3.0. Scotts Valley (Ca, USA): Borland International, Incorporated. 1991 Mar 27. 6. Conway M, Durbin GM, Ingram D, Mclntosh N, Parker D, Reynolds EO, et al. Continuous monitoring of arterial oxygen tension using a catheter tip polarographic electrode in infants. Pediatrics 1976; 57: 244-50. 7. Coon RL, Lai NCJ, Kampine JP. Evaluation of a dual function pH and PCO_, in vivo sensor. J Appl Physiol 1976; 40: 625-9. 8. Mahutte CK, Sassoon CSH, Muro JR, Hansmann DR, Maxwell TP, Miller WW, et al. Progress in the development of a fluorescent intravascular blood gas system in man. J Clin Monit 1990; 6: 147-57. 9. Martin WE, Cheung PW, Johnson CC, Wong KC. Continuous monitoring of mixed venous oxygen saturation in man. Anesth Analg 1973; 52: 784-93. 10. Peterson JI, Goldstein SR, Fitzgerald RV, Buckhold DK. Fiber optic pH probe for physiologic use. Anal Chem 1980; 52: 864-9. 11. Reynolds ER, Yacynych AM. Miniaturized electrochemical biosensors. American Laboratories 1991 Mar 19-27. 12. Shapiro BA, Cane RD. Blood Gas Monitoring: Yesterday, today, and tomorrow. Crit Care Med 1989; 17: 573-81. 13. Shapiro BA, Cane RD, Chonka CM, Bandala LE, Peruzzi WT. Preliminary evaluation of an intra-arterial blood gas system in dogs and humans. Crit Care Med 1989; 17: 45560. 14. Stipp D. Sensor designed for insertion into arteries. Wall Street Journal 1991 Sep 26; CC XVIII (218) 62: B-I, B-4.
Address for offprints: B.E. Smith, Profesor and Chair, Department of Anesthesiology, Vanderbilt University Medical Center, The Vanderbilt Clinic, Room 2301, Nashville, Tennessee 37232-2125, USA
Clinical evaluation - continuous real-time intra-arterial blood gas monitoring during anesthesia and surgery by fiber optic sensor Bradley E. Smith, Paul H. King & Les Schlain Department of Anesthesiology, Vanderbilt University Medical Center, Nashville, Tennessee and Optex Biomedical, Inc., The Woodlands, Texas Accepted 13 January 1992
Key words: clinical monitoring, anesthesia, analysis of blood gases Summary
A clinical evaluation of the clinical utility, techniques of use, durability, accuracy, and potential complications of a newly available system for the continuous real-time, intra-arterial monitoring of arterial blood gas and acid base status (ABG) has been studied (Optex BioSentry | System, Optex Biomedical, Incorporated, The Woodlands, Texas, U.S.A.). The system consists of three separate fiber optic channels with contained fluorescent and absorbant chemical analytes imbedded in a single probe and capable of insertion inside of a twenty gauge indwelling arterial catheter (The Optex Optode | Sensor), along with an external monitor. The system was utilized during anesthesia and surgery in the care of five informed patients. Constant trend monitoring of all three variables was deemed satisfactory in four of the patients. The fifth sensor was damaged by a surgical assistant while in place and ceased to function. Comparison of Optode | sensor readings with standard clinical laboratory ABG analysis was excellent in three uncomplicated patients (pH: bias - 0.0183 pH units, precision: 0.0237 pH units; PCO2: bias 3.22 mmHg, precision 2.04 mmHg; PO2: bias - 5.94 mmHg, precision 11.74 mmHg). Postoperative study suggested that discrepancies in a fourth patient may have been due to an incorrect 'offset' applied to the Optode | sensor yielding a constant error.
Introduction
Although frequent analysis of arterial PO2, PCO2, and pH measurements and the immediate application of this knowledge to real-time management of patients undergoing anesthesia and surgery has become routine during the past two decades, the standard procedure frequently entails lag time in the presentation of data and has become extremely expensive in actual hospital applications [12]. The ability to continuously monitor moment to moment changes in A B G has long been desired for use in the operating room and in critical care units, and would provide obvious advantages in the care of a wide range of acutely and critically ill patients.
Toward this objective, several proposed devices have appeared in the past two decades, but have thus far presented technical limitations or have been withdrawn from the market for economic reasons. These include miniaturized electrochemical biosensors and various fiber optic techniques, Electrochemical sensors often represent a combination of two transducers, one biochemical, tile other physical, in intimate contact with each other. These relate the concentration of an anatyte to a measurable electrical signal [11]. In clinical use, various models of these types have proven susceptible to damage and particularly to calibration drift. Attempts at clinical application of such devices have included polarographic oxygen electrodes [6],
46 electrodes based upon the Severinghaus model [7], and intravascular placement of fiber optic reflectance devices estimating oxygen tension from oxygen saturation [9]. Except for the latter, most have, at this time, been abandoned [2]. Great interest is currently centered on promising newer technologies based on various designs incorporating transmission of light through fiber optic fibers and manipulation of the transmission of this light by chemicals, usually absorbant or fluorescent, whose characteristics change in proportion to changes of pH, PCO~, and/or PO2 [14, 4]. The instrument utilized in this evaluation is of the latter type. Preliminary studies of the utilization of somewhat similar instruments for the constant measuring of intra-arterial blood acid base and physiologic gas status have been carried out in dogs [3], in human volunteers [8], in a small group of critical care patients [13] and in fourteen patients undergoing anesthesia and surgery [2]. Satisfactory human trials of the device utilized in this study extending up to ninety-six hours of use have been carried out in a limited number of patients in the critical care setting (Optex Biomedical, The Woodlands, Texas, Personal Communication), however, this is the first evaluation of the use of this particular device during anesthesia and surgery.
Methods
Evaluation of the clinical utility, techniques of use, durability, accuracy, and potential complications associated with the use of the Optex BioSentry | Blood Gas Monitoring System an Optode | sensor placed in the radial artery were carried out in five consenting surgical patients.
Patient number one
A forty year old male weighing 72kg, height 165 cm, with an ASA Classification of II [1] underwent a craniotomy for cranioplasty. The duration of anesthesia was 239 min. Anesthesia consisted of sufentanyl 50 tzg, thiopental sodium 500 mg, vecuronium 13 mg, midazolam 2 mg, and isoflurane in-
haled from 0.5 to 0.75% via oral-tracheal tube. There were no complications of anesthesia or surgery, however, the blood pressure rose from 120 systolic to 160 systolic during the final fifteen minutes of emergence from anesthesia.
Patient number two
A sixty-one year old female weighing 70 kg, height 170 cm, with an ASA Classification of Ill also experienced a craniotomy procedure for cranioplasty. The duration of anesthesia was 131min and anesthesia consisted of sufentanyl 25/xg, thiopental sodium 250mg, vecuronium 10mg, labetolol 20 mg, esmolol 15 mg, ephedrine 5 mg and isoflurane from 0.5 to 1.38% inhaled via an oral-tracheal tube. Intraoperative complications consisted of intermittent hypertension as high as 190/110 early in the case, treated by blockers but recurring briefly during emergence.
Patient number three
A seventy-nine year old female weighing 65 kg, height 165 cm, with an ASA Classification of III underwent laparotomy for total abdominal hysterectomy and bilateral salpingo-oophorectomy. The duration of anesthesia was 157 rain and anesthesia consisted mainly of a continuous epidural block consisting 100mg of 0.5% bupivacaine. Other drugs included thiopental sodium 250 mg, vecuronium 5mg, midazolam 3mg, phenylephrine 300/xg, and approximately 60% nitrous oxide inhaled via an oral-tracheal tube. Intraoperative complications consisted only of brief hypotension to approximately 77 mmHg systolic over 30 diastolic on three accasions. During emergence from anesthesia, the blood pressure arose to 150/70.
Patient number four
A forty-two year old female weighing 50 kg, height 170 cm, with an ASA Classification of II also experienced an exploratory laparotomy, total abdom-
47
/
Clear L u e r
pH and P C O 2 During Anesthesia
Lock
/
Patient # 1 Connector
7.6
755
Sensors
...............
. . . . . . . . . . . . .
f ~
~
/ ~ - ~
5O
C = Change in ventilator 55
0 = O[fset
,~_ C
~
C
l 50
L/
I Arrow
20
Io_
f4
o
g35
~)
73
725-
45
40 o h35
i
~i
' 3O
g. Catheter
25
Fig. 1. The Optode | Sensor (Optex BioSentry| Monitor, The Woodlands, Texas, United States of America).
T i m e (hours) O p t o a e TM pH L a b o r a t o r y pH
inal hysterectomy, and bilateral salpingo oophorectomy. The duration of anesthesia was 129 min and anesthesia consisted of thiopental sodium 125 mg, succinylcholine 100 rag, nitrous oxide up to 65% mixed with up to 1.34% isoflurane inhaled via an oral-tracheal tube. Intraoperative complications consisted of very brief hypotension as low as 65/55 on two occasions with no apparent clinical impairment.
Patient number five A sixty-three year old male weighing 58 kg, height 160 cm, with an ASA Classification of III experienced right carotid endarterectomy. The duration of anesthesia was 110 min and anesthesia consisted of propofol 140 rag, fentanyl 300 txg, nitrous oxide up to 50% mixed with isoflurane up to 0.5% inhaled via an oral-tracheal tube. In addition, phenylephrine 200/xg, ephedrine 200 mg, and sodium nitroprusside 200 ~g, were utilized in the patient's care. Intraoperative complications consisted only of hypertension early in the procedure which was well controlled during the latter part of the procedure. In all patients, the radial arterial puncture was carried out in the usual sterile manner under local anesthesia before induction of general anesthesia. Before establishment of the intra-arterial sensor, the Optode | sensor (Fig. 1) was calibrated in a sealed container with a sterile chemical solution and was equilibrated with appropriate reference gases. After warm up and cessation of drift while in the reference solution, the sensor was ready for
+
O p t o d e rv PCO~
- -
L a b o r a t o r y PCO 2
Fig. 2. pH and PCO2 During Anesthesia.
placement (Fig. 2). Attention was given to placement of the catheter in such a manner that the sensor experienced a straight course of insertion and was not obliged to bend. This was carried out easily in all five patients. After establishment of the arterial catheter and placement of the sensor, a formed plastic splint was added external to the forehand in order to protect the sensor. After establishment of general anesthesia and when the patient had achieved a stable anesthetic level, an ABG sample was obtained from each patient via the port provided at the base of the Optode | sensor device. Withdrawal of the blood sample was unimpeded and was carried out in a routine fashion. Constant radial artery pressure monitoring was easily accomplished through the same port in all five patients. The time of withdrawal of the sample was noted and when the report of values from the clinical laboratory value became available, an offsetting current was added to the BioSentry| monitor for the purpose of bringing the values on the BioSentry| monitor (pH, PCO2, and PO2) into agreement with the clinical laboratory readings. Thereafter, additional offset adjustments could be made, but have not been required due to the demonstrated stability of the Optex BioSentry| system.
Results The Optex BioSentry| system appears to be reia-
48 tively simple in clinical application. No problem was encountered in any of the five evaluations with initial warm up, calibration, or stabilization of the sensor in the standard solution. Insertion of the sterile, calibrated, and stabilized sensor into the radial artery presented no problems in any of the five patients. After insertion into the patient's radial artery, stabilization appeared to take place in approximately twenty to thirty minutes. Figure 1 demonstrates the plotted Optode | sensor readings of pH and PCO2 in Patient Number One. The readings are updated every thirty seconds and can be obtained as digital data and in a trended presentation simultaneously. Figure 1 also illustrates comparisons of pH and PCO~ from arterial samples analyzed by the Hospital Clinical Laboratory. Similar trend graphs were prepared for pH, PCO2, and PO2 in all five patients and in all of the first four patients trending closely approximated and followed clinical laboratory values of the extracted samples. Table 1 displays all data for all five patients,
including clinical laboratory ABG values, Optode | sensor readings at the same time and the difference between the two systems for each variable. The course of monitoring appeared to proceed without incident in Patients Number One and Two except that it was noted that there was a far greater variation between clinical laboratory and Optode | sensor values in the ABG samples taken during emergence from anesthesia. The monitoring course in Patient Number Three was similarly uncomplicated, but no arterial blood gas sample was collected during emergence from anesthesia. Patient Number Four, displayed greater discrepancy between the ABG and the Optode | sensor readings. However, sensor trending seemed to be stable and representative. On post-operative study, it was noted that the ABG sample for the original 'offset' reading was inadvertantly drawn during a period of clinical hypotension. A postulation was offered that the ABG drawn at this time represented a transient due to the physiologic disturbance and that the sensor had not yet equili-
Table 1. Arterial Blood Gases (ABG) During Anesthesia and Surgery. Patient No.
1
2
3
Time
Clinical Laboratory Optode | Sensor
Difference
pH
PO,
pH
PCO,
PO~
pH
PCO~
Comments PCO,
PO,
9:20 a m
7.46
3l
131
7.45
31.9
134.1
-0.1105
0.948
9:40 a m
7.46
29
154
7.47
30.9
1411.3
0.01/5
1.845
3.10
10:05 a m
7.47
25
171
7.48
29.6
149.3
0.008
4.639
- 21.712
[0:31 a m
7.47
25
172
7.49
29.6
145.0
1/.015
4.631
-26.984
10:54 a m
7.49
29
156
7.49
29.6
143.7
0.111t3
0.563
- 12.285
ll:17am
7.33
411
166
7.40
4 l. 1
181.9
0.070
1.049
1:33 p m
7.39
31
167
7.37
37.9
160.9
- 1/./120
6.91
- 6.08
l:57pm
7.37
35
154
7.33
41.4
156.6
- 11.042
6.42
2.62
2:19pm
7.38
40
157
7.32
41.7
145.5
- 0.060
1.74
- 11.55
2:41pm
7.50
26
371
7.41
29.l
227.1
- 0.086
3.05
- 143.91)
9:13 a m
7.44
29
109
7.41
31.0
112.2
- 0.030
2.03
3.21
9:43 a m
7.45
28
107
7.41
31.1
ll8.11
-0.036
3.12
11.05
Uncomplicated
- 13.650
15.875
10:14am
7.46
27
114
7.42
29.6
121.0
- 0.039
2.59
6.96
4
12:44 p m
7.35
39
168
7.38
38.8
125.9
0.034
- 0.19
- 42.07
l:18pm
7.35
39
148
7.45
35.4
1211.9
0.102
- 3.59
- 27.15
5
8:32am
7.3l
37
263
7.30
36.6
286.0
- 0.010
- 0.40
23.00
8:43am
7.31
411
256
7.29
31.9
272.0
- 0.018
- 8.12
16.113
291.9 290.7
- 0.029 - 1t.1151
- 4.27 - 2.38
72.90 57.70
9:15 a m
7.32
42
9:47 am
7.33
40
219 233
7.29 7.28
37.7 37.6
10:15 a m
7.31
41
243
7.26
40.2
291.8
- 11.048
- 0.84
48.80
10:43 a m
7.36
40
226
7.31
39.6
284.8
- 0.045
- 0.45
58.84
(emergence A BG) Uncomplicated
(emergence ABG) Uncomplicated
(no emergence A BG) Offset Problem Sensor Damage
49 brated with this change, resulting in an incorrect offset. This may have given rise to a constant error in absolute comparison of A B G and sensor values. In Patient N u m b e r Five, insertion of the Optode | sensor was uneventful. After 'offset' (8:32 am), the next comparison of A B G and sensor values (8:43 am) appeared to be reasonably accurate. Soon thereafter, a surgical assistant was observed to lean on the site of the sensor with significant weight. Thereafter, the sensor values were greatly different from the A B G values. After the end of anesthesia and surgery the sensor was withdrawn and found to be grossly intact, but on magnified inspection the fiber optic path was indeed damaged. Statistical analysis of the comparison of corresponding A B G laboratory readings with Optode | sensor readings is presented in Table 2 and Table 3, 'Bias' is defined as the mean of differences between sensor and arterial values and 'Precision' is one standard deviation (1SD) of the differences [8]. Table 2 presents comparison which exclude the A B G gas samples taken during anesthetic emergence and eliminates patient four and five who experienced known problems. Table 3 presents statistical analysis of the same data but also includes A B G samples drawn during anesthesia emergence. The senior author physician examined all patients from the second to the fourth day post-anesthesia and surgery and found no evidence of complications of the use of the sensor either systemically or at the local site of insertion.
Table 2. Optode | Sensor - Clinical Laboratory Agreement Uncomplicated Patients*.
pH
PCO,
OPTICAL SENSOR SYSTEMS
\
i n/ ............ ~aem~ra~e t , ct ~ e
i
Fig. 3. Optical Sensor Systems.
Discussion
The Optex BioSentry | system operates on principle similar to other fiber optic sensor systems. An optical sensor indicates variation in the surrounding blood by changes in either the color or intensity of the light passing through the sensor. The sensor is in contact with the surrounding medium through a semipermeable m e m b r a n e and the light pathway is lead through the chemical c h a m b e r passing then to a detector, then to a signal processor and finMly to a user interface. A diagram of such a system is provided in Fig. 3. Designs vary between manufacturers, and in some cases parallel light channels or reflection techniques through monofiber channels are utilized. Some designs [10, 4] depend on reflectance of the light from the indicator complex which is located at the tip of a catheter sensor, Fiber optic systems are inherentiy immune to electromagnetic noise, can be extremely small, and provide electrical isolation of the patient.
Table 3. Optode | Sensor - Clinica~ Laboratory Agrecmem Uncomplicated Patients*
PO2 pH
Bias 1SD Precision # of ABGs
- 0.0183 0.0237 11
3.22 2.04 11
* Excludes emergence period. Eliminates Patient 4 with erroneous offset. Eliminates Patient 5 with damaged sensor,
s ire
- 5.94 11.74 11
Bias ISD Precision # of ABGs
- 0.0167 0.0377 13
PCO~
PO~
3.04 [.97 13
- 14.87 39,2! 13
* Eliminates Patient 4 with erroneous offset. Eliminates Patient 5 with damaged sensor.
50 FIBER
OPTIC
- pH
SPECTROSCOPY
-
PRINCIPLES OF OPERATION OF OPTODE T M SENSOR
CHEMICAL Anaiyte
Chemistry of amphoteric pH indicators light ~
1
indicator
Analyte
light
)
pH
= -Log(K a ) + Log[(A)/(HA)] intrinsic physical constant of indicator concentration of ionized indicator concentration of neutral indicator
a
(,4-) = I - INDICATOR
(HA) =
Fig. 6. pH - Chemical Principles of Operation of Optode|
Sensor.
Dual Reflection
Monofiber
Transmission
Fig. 4. Fiber Optic Spectroscopy.
Figure 4 diagrams several alternative system configurations. The Optode | sensor incorporates a 180 degree bend at the tip of the optical fiber, which allows transmission of light through the analyte chamber and also provides for placement of the analyte chamber away from the tip of the sen-
- pH -
sor, thus reducing reflection interference and other problems of tip mounted indicator chambers. The Optex Optode | sensor utilizes principles of classical UV-visible spectrophotometry as illustrated by the statement of the Beer-Lambert Equation in Fig. 5. The Optex system uses light absorbance by indicators for the detection and measurement of pH and PCO~. A pH indicator react by absorbing, or blocking, light of a particular wavelength in proportion to the pH of the sensor environment. The ratio of the intensity of light projected into the sensor to the intensity of light emitted from the sensor is used to predict the pH. As CO2 enters the sensor through a gas permeable membrane it equilibrates with a carbonic acid/bicarbonate buffer system. The pH of this system, as predicted from the absorbance of the indicator, is used to predict PCO~. The chemistry of amphoteric pH indicators is further illustrated in the equation shown in Fig. 6
CHEMICAL PRINCIPLES OF OPERATION OF OPTODE TM SENSOR
- PCO Beer-Lambert Equation (U.V.- visible spectrophotometry)
PRINCIPLES OF OPERATION OF OPTODE T M SENSOR
CHEMICAL
-Log(l o/li) = el(c) 1o 1~ e l (c~
= = = = =
wavelength of light entering sample intensity of light transmitted through the sample intrinsic physical constant of chemical interest path length of light through sample concentration of sample
2 -
Henderson-Hasselbach
equation
p H = -Log(K a ) + Log[(HC03 )/0.03PC021
Fig. 5. pH - Chemical Principles of Operation of Optode |
Fig. 7. PCO, - Chemical Principles of Operation of Optode |
Sensor.
Sensor.
51
-
PO
z -
CHEMICAL PRINCIPLES OF OPERATION OF OFFODE TM SENSOR Stern-Vollmer equation (fluorescence quenching) - L o g ( I o IIs ) - I
= K,,(PO, )
Io
=
intensity o f light emitted from a compound in the
1i
=
intensity of fluorescent light emitted at an unknown P02
K,v
=
an intrinsic physical constant of the fluorescent indicator compound
absence of oxygen
Fig. 8. PO, - Chemical Principles of Operation of Optode@
Sensor. and the familiar chemistry of the Henderson-Hasselbach equation utilized in detection of CO 2 is illustrated in Fig. 7. The measurement of oxygen in the Optex system is dependent on the quenching of fluorescent dyes by the presence of oxygen. When stimulated with ultraviolet light, certain dye molecules may obtain enough energy to emit visible light. Oxygen molecules can prevent emission by stealing some of this energy during collisions with the dye. The fluorescence emission of quenchable dye in a sensor is related to the amount of stimulus light projected into the sensor and the amount of oxygen present in the sensor. The principles of 'fluorescence quenching' utilized in oxygen detection which are illustrated in the Stern-Vollmer equation are represented in Fig. 8. Exaggerated s e n s o r - ABG differences noted during emergence might be due to a lag in equili-
bration by the Optode | sensor, however, the variance is in opposite direction in Patients One and Two. Others have postulated a difference in actual ABG at the sampling site in the radial artery from blood in the ventricle [8] and this phenomenon might be exaggerated during the excitement of emergence. However, interaction between the sensor and the arterial wall might also be exaggerated during the vascular reactivity and hypertension noted in both patients. Table 5 lists expected advantages of the Optex Optode | sensor system, which include support of the analyte on all sides, and improved strength and integrity of the analyte chamber. The side placement of the membrane provides more surface area for interaction than on tip mounted sensors. Side versus tip placement may reduce thrombogenicity and inaccuracies introduced by localized metabolic processes in the thrombus noted by previous investigators [8] and the design may reduce sensorarterial wall interaction which has also been encountered by others [8]. Complete separation of the analyte containing sites in three separate fiber optic paths reduces the risk of light leak and reflectance interference between analyte chambers. An increased selectivity and sensitivity of the transmission technique versus the light reflective technique is expected. Internal reflective interference is expected to be reduced due to separate transmission cells. Furthermore, the sensor body is of high light transmission capability, therefore allowing more light to be received distal to the cell at the detector. Side versus tip placement of the membrane may reduce thrombogenicity.
Table 4. Advantages of Optex | Light Transmission System.
1 Mechanical 2 Chemical 3
Optical
4
Clinical
material suppor! separation of sites reduces cross reactivity increased sensitivity reduction of reflective interference high transmission of sensor body probe mechanically stable side window is less thrombogenic improved accuracy, less chemical or physical interference
Table 5. I)ritt of Optode | Sensor*.
Analyte Average Drift Per Hour
Deviation After 12 Hours
pH PCO, PO,
0.025 pHu 10 mmHg 10 mmHg
0.002 pHu 0.8 mmHg 0.8 mmHg
* Stable controlled pH, PCO> PO2, temperature, etc.
52 Conclusions A preliminary evaluation of the clinical usefulness, reliability and potential complications in application of a new continuous real-time intra-arterial blood gas monitoring system (Optex BioSentry | Optex Biomedical, Inc., The Woodlands, Texas, U.S.A.) has been carried out in five consenting anesthetized patients during surgery. Trend monitoring of pH, PCO2, and PO2, was deemed satisfactory in four of the five patients (In the fifth patient, the sensor was damaged by the weight of a surgical assistant). Comparison of Optode | sensor readings with clinical laboratory ABG readings was excellent in three uncomplicated patients. However, in a fourth patient, insertion of offset to the sensor based on A B G sampled during clinical hypotension may have inserted a small constant error in later Optode | sensor readings. Further evaluation is indicated to better understand the relationship of equilibration time of the sensor to physiologic changes, and to the variables which may affect delivery of blood representative of the current physiologic state to the sensor sampling site. Further evaluations should also reveal whether the one damaged sensor indicates fragility of the system in clinical use.
Acknowledgements
Supported in part by a contract from Optex BioMedical, Inc., The Woodlands, Texas, United States of America.
References 1. American Society of Anesthesiologists. New classification of physical status. Anesthesiology 1963; 24: 111.
2. Barker SJ, Hyatt J. Continuous measurement of intraarterial pHa, PaCO,, and PaO~ in the operating room. Anesth Analg 1991; 73: 43-8. 3. Barker SJ, Tremper KK, Hyatt J, Zaccari J, Heitzmann HA, Holman BM, et al. Continuous fiberoptic arterial oxygen tension measurements in dogs. J Clin Monit 1987; 3: 48-52. 4. Barnard SPM, Walt DR. A fiber-optic chemical sensorwith discrete sensing sites. Nature 1991; 353: 338-40. 5. Ouatro Pro 3.0. Scotts Valley (Ca, USA): Borland International, Incorporated. 1991 Mar 27. 6. Conway M, Durbin GM, Ingram D, Mclntosh N, Parker D, Reynolds EO, et al. Continuous monitoring of arterial oxygen tension using a catheter tip polarographic electrode in infants. Pediatrics 1976; 57: 244-50. 7. Coon RL, Lai NCJ, Kampine JP. Evaluation of a dual function pH and PCO_, in vivo sensor. J Appl Physiol 1976; 40: 625-9. 8. Mahutte CK, Sassoon CSH, Muro JR, Hansmann DR, Maxwell TP, Miller WW, et al. Progress in the development of a fluorescent intravascular blood gas system in man. J Clin Monit 1990; 6: 147-57. 9. Martin WE, Cheung PW, Johnson CC, Wong KC. Continuous monitoring of mixed venous oxygen saturation in man. Anesth Analg 1973; 52: 784-93. 10. Peterson JI, Goldstein SR, Fitzgerald RV, Buckhold DK. Fiber optic pH probe for physiologic use. Anal Chem 1980; 52: 864-9. 11. Reynolds ER, Yacynych AM. Miniaturized electrochemical biosensors. American Laboratories 1991 Mar 19-27. 12. Shapiro BA, Cane RD. Blood Gas Monitoring: Yesterday, today, and tomorrow. Crit Care Med 1989; 17: 573-81. 13. Shapiro BA, Cane RD, Chonka CM, Bandala LE, Peruzzi WT. Preliminary evaluation of an intra-arterial blood gas system in dogs and humans. Crit Care Med 1989; 17: 45560. 14. Stipp D. Sensor designed for insertion into arteries. Wall Street Journal 1991 Sep 26; CC XVIII (218) 62: B-I, B-4.
Address for offprints: B.E. Smith, Profesor and Chair, Department of Anesthesiology, Vanderbilt University Medical Center, The Vanderbilt Clinic, Room 2301, Nashville, Tennessee 37232-2125, USA
