Continuous emergency department monitoring of arterial saturation in adult patients with respiratory distress

ORIGINAL CONTRIBUTION arterial saturation, monitoring

Continuous Emergency Department Monitoring of Arterial Saturation in Adult Patients With Respiratory Distress

Continuous measurement of arterial oxygen saturation (Sa02) using pulse oximetry has become a common monitoring and management technique in critically ill hospitalized patients. To determine the impact of SaO 2 monitoring on emergency patient management, we conducted a prospective un- controlled clinical trial on 40 adult patients presenting to the emergency department with acute respiratory illness, such as emphysema, asthma, or pulmonary edema. Recorded data included hemograms, arterial blood gases, subsequent therapy, and response to treatment. Additionally, the "early warning" capability of SaO 2 monitoring was analyzed by recording the severity and outcome of hypoxemic events during treatment. Mean duration of usage for the 40 oximeters in the ED was 1.8 hours; all probes functioned feb'ably over a wide range of systolic pressures (80 to 206 m m Hg), heart rates (40 to 180 beats per minute), and hematocrits (20% to 58%). There was good correlation between simultaneous pulse oximeter values and both directly measured SaO 2 (r = 0.95) and saturations derived from measured arterial PaO 2 (r = O. 94). The device detected several otherwise unrecognized drops in arterial saturation that were confirmed by laboratory analysis. Other clinical situations in which the pulse oximeter was found useful in the ED are reviewed. We conclude that continuous measurement of SaO 2 can improve the monitoring of ED patients, increase the precision of therapy, detect hypoxemia during intubation, suctioning, and other treatments, and detect clinically unsuspected changes in arterial oxygenation. ]Jones J, Heiselman D, Cannon L, Oradisek R: Continuous emergency department monitoring of arterial saturation m adult patients with respiratory distress. Ann Emerg Med May 1988;17:463-468.]

INTRODUCTION Assuring adequate oxygenation of tissues is a primary goal in the acute

management of patients with respiratory failure. Conventional analysis of arterial blood samples has many shortcomings, including the need for arterial puncture and the risk of errors due to sampling technique, sample transport, and analysis. Furthermore, because of the intermittent availability of data, acute alterations in tissue oxygen levels may not be detected until after clinical deterioration occurs.

Noninvasive methods of measuring oxygenation parameters are now available. Transcutaneous oxygen sensors measure gas partial pressures at the skin surface and are structurally similar to the electrodes used in conventional blood gas machines3 This allows for continuous measurements, but their response time varies, frequent calibration is necessary, and change of skin site is required to prevent skin damage. There may also be a poor correlation with PaO 2 depending on the age of the patient, site of electrode, oxygen levels, cardiac output, peripheral perfusion, and condition of the skin.2,3

Recently, a pulse oximeter has been developed that measures arterial oxygen saturation (SaO2) beat by beat. 4

Clinical experience with surgical patients in the operating room and intensive care unit has demonstrated that pulse oximetry accurately reflects oxygen delivery and is a clinically useful, noninvasive monitoring technique in these settings, us However, there is little direct evidence that routine use of SaO 2 monitoring in emergency medicine has a beneficial role. The purpose

Jeffrey Jones, MD Darell Heiselman, DO, FACC, FCCP Louis Cannon, MD Richard Gradisek, MD Akron, Ohio

From the Department of Emergency Medicine and Department of Internal Medicine, Division of Critical Care Medicine, Akron General Medical Center, Northeastern Ohio Universities College of Medicine, Akron, Ohio.

Received for publication September 11, 1987. Revision received January 15, 1988. Accepted for publication February 12, 1988.

Presented at the University Association for Emergency Medicine Annual Meeting in Philadelphia, May 1987.

This study was supported in part by a grant from Ohmeda, Boulder, Colorado.

Address for reprints: Jeffrey Jones, MD, Department of Emergency Medicine, Akron General Medical Center, 400 Wabash Avenue, Akron, Ohio 44307.

17:5 May 1988 Annals of Emergency Medicine 463/53

ARTERIAL SATURATION Jones et al

FIGURE 1. The Biox 3700 ® pulse oximeter. Front panel LEDs on the microprocessor continuously display SaO 2 heart rate status and alarm lim- its. Signal strength indicator and plethysmograph waveform indicate blood perfusion changes (Courtesy of Ohmeda, Boulder, Colorado).

of our study was to evaluate the accuracy of this device and its clinical usefulness in patients with acute respiratory illness.

MATERIALS AND METHODS The Pulse Oximeter System

The noninvasive, microprocessor- based pulse ox imeter (Biox 3700 ®, Ohmeda, Boulder, Colorado) measures heart rate and SaO 2 cont inuous ly with a single sensor located on an ex- tremity. The ins t rument probe con- sists of two light-emitting diodes and a photosemiconductor attached to the monitor by fiberoptic cable. No calibration of the microprocessor is necessary; saturation is displayed digitally and audibly with each heart beat {Fig- tire 1). Blood volume changes are displayed by con t inuous p l e thysmo- graphic waveform. A linear array of lights alongside the graph provides qualitative indication of perfusion or signal strength.

Trend memory is available to store SaO 2 and heart rate data from the previous eight hours. The analogue output was connected to an external recorder for data collection.

Clinical Series We studied 40 consecutive adult pa-

tients who presented to the emergency d e p a r t m e n t at Akron Genera l Medical Center with varying degrees of respiratory distress. The mean age of the patients was 52 years (range, 19 to 81). These patients were suffering from various respiratory disorders, including chronic obstructive pulmonary disease (13 patients), a s thma (nine), cardiogenic pulmonary edema (nine), pneumonia (six), adult respiratory distress syndrome (one), pulmonary edema secondary to near drowning (one), and thoracic gunshot wound (one). M1 were hemodynamically sta- ble with systolic pressures of more than 80 m m Hg. Patients were re- suscitated by the emergency physi- cians using standard clinical criteria. Five patients required endotracheal intubation and mechanical ventilation.

Exclusion criteria included signs of compromised peripheral circulation (ie, pallor, cool skin, slow nail bed fill- ing) or hemodynamic instability (systolic blood pressure less than 80 m m Hg}. These criteria may reduce vascular pulsat ions and affect the oximeter 's ability to calculate saturation. Because carboxyhemoglobin may erroneously increase readings, patients who had smoked during the 24-hour period prior to admission were ex- cluded from the study.

Study Protocol M1 conditions and procedures of our

experiment were approved by the In- s t i tu t iona l Review Board for Bio- medical Research at our institution. Informed consent was obtained from each pa t ien t prior to entering the study.

The pulse ox ime te r i n s t r u m e n t probe was placed on the finger as soon as possible after admission of the patient to the ED. Any limb with compromised peripheral circulation was not used. The beat- to-beat display mode was used, and the value was accepted only if the pulse oximeter- monitored heart rate did not differ by more than five beats from that on the independent bedside ECG monitor.

Initial data were collected on each patient prior to the administration of oxygen or drug therapy and included the fol lowing: t ime, t empera tu re , heart rate, hematocrit, blood gas analysis, blood pressure, calculated SaO2, direct SaO 2 measurement, and pulse ox imete r SaO2. Arterial blood gas

Annals of Emergency Medicine

samples were obtained anaerobically in 5-mL heparinized syringes from a radial or femoral a r t e~ and stored on ice until analysis with a Coming 175 ® oximeter (Coming Medical and Scien- tific, Coming, New York).

SaO 2 was measured in the laboratory using an IL 282 Co-oximeter ® (In- strumental Laboratory Incorporated, Lexington, Massachusetts) . Percent saturation was defined as the ratio of hemoglobin bound to oxygen divided by the to ta l amount of hemoglobin available for reversibly binding to oxygen. SaO 2 levels below 91% were re- garded as unacceptable, corresponding to PaO 2 values of less than 60 m m Hg, and levels of 85% and below were re- garded as severe hypoxemla. Carboxy- h e m o g l o b i n and m e t h e m o g l o b i n , when present, were not included in the calculation of SaO 2.

In vivo SaO 2 values were recorded continuously on a strip chart recorder connected to the oximeter. Medica- tions (eg, furosemide, oxygen, bronchodilators) and therapeutic treatment (eg, aerosol t reatments , intubation) were noted on the record strip. Pro- cedures and blood gas determinations were done at the discretion of the housestaff. Serial measurements were made at any time when SaO 2 values decreased by 5% or more f rom the baseline value. Data collection was continued throughout pulmonary resuscitation until patients were trans- ferred from the ED.

Statistical Analysis Clinical and h e m o d y n a m i c data

54/464 17:5 May 1988

100-

9 0 .

80.

7 0 ¸ .

60.

50 50

r = 0.95

' ' = I w I 6b 7b do 9o 1oo

Pulse Oximeter S a O 2, °/o

Measured SaO2, % (IL 282 Co-oxirneter)

Pulse Oximeter Sa02, O/o

80-

. :

60-

40-

20"

0 o 2'5 do 7's 40 1;s

PaO 2, mm Hg 3

were analyzed for means, standard de- viations, and standard errors of the mean. Unless otherwise specified, the data are presented as mean -+ standard deviation. Linear regression analysis was used to compare SaO2, deter- mined by the pulse oximeter, wi th measured and calculated SaO2 and

PaO 2. Differences were considered significant at the .05 level.

RESULTS Accuracy

The correlation of in vivo and in vitro measured oxygen saturations is shown (Figure 2). The range of arterial

FIGURE 2. Regression analysis of 52 s imultaneous measurements of SaO 2.

FIGURE 3. In vivo oxygen dissociation curve comparing PaO 2 and pulse oximeter SaO 2. Regression l ine is nor- m a l oxygen d i s soc ia t ion curve for adult pat ient w~th pH, 7.4; p C O 2, 40 m m Hg; and temperature, 37 C.

oxygen saturation was 51% to 99%. At the t ime the compar isons were made, the hematocr i t values ranged f rom 20% to 58.1%. Systolic blood pressures ranged from 80 to 206 m m Fig, and pulse rates from 40 to 180 beats per minute. The correlation co- efficient (r) of 0.95 (P < .001) compares well with previous observations using a different pulse oximeter 2 and with o ther s tudies us ing this system.6, 7 The mean difference between measured SaO~ was + 1 . 8 6 + 3.11% (range, - 8 . 0 to +9.3). There was no significant correla t ion between age, blood pressure, heart rate, hematocrit, or PaO2, and the difference between measured and pulse oximeter saturations.

Regress ion analysis of pulse oximeter SaO 2 versus calculated SaO2 showed an r of 0.94 (P < .001) with an equation of y = 0.96x + 1.12. The mean difference between calculated and oximetric 8aOz was +0.96% -+ 3.81% (range, - 7 . 8 to +9.0). The in vJvo oxygen dissociation curves constructed with PaO 2 versus measured SaO2, pulse oximeter saturation (Fig- ure 3), and calculated SaO 2 all showed similar curves. The oxygen dissociation curve constructed with the calculated SaO 2 represented, in fact, a reliable m e a n d i s soc i a t i on of all ou r patients.

Technical Performance Mean duration of usage for the ox-

imeter in the ED was 1.8 hours per patient. The mon i to r func t ioned well throughout the study. No error was in- troduced nor was the machine limited by skin pigmentation. The readout of sa tu ra t ion occur red w i th in several heart beats, as soon as pnlsatile flow was picked up by the monitor. We obtained a stronger signal more consis- tently from the finger probe compared with the ear probe, and so we chose to use that sensor. Malposition of the finger probe was an occasional problem, especially in the restless patient. If the sensor lost finger contact, a panel indicator light and alarm were activated.

17:5 May 1988 Annals of Emergency Medicine 465/55


4

SaO 2 (%)

90

80-

70" l A

60-

5O o 6 do 3~ 4'o do

Time (rain)

SaO 2 (O/o)

5

100-

90-

80-

70-

60-

5O o 6 2~ 35 ;o do

Time (rain)

FIGURE 4. Record o f c o n t i n u o u s SaO 2 monitoring during repeated at- t e m p t s at e n d o t r a c h e a l i n t u b a t i o n (A), m e c h a n i c a l ven t i la t ion w i th a PEEP of 5 cm (B), 10 cm (C), and 20 cm of water (D).

FIGURE 5. Record o f c o n t i n u o u s SaO 2 monitoring showing changes in SaO 2 associated wi th administration of oxygen (A) and tracheal suctioning (B). Diuresis wi th IV furosemide (C) led to improvement in SaO 2 as pulmonary edema was reduced.

In addition, the continuous plethys- mographic waveform was lost.

Clinical Usefulness To assess the utility of pulse oxime-

try as a trend monitor of arterial gases, strip chart records were analyzed to determine the severity of desaturation episodes. We compared oximetric versus measured changes for SaO 2. Thir- ty-six simultaneous data points were obtained from 29 patients. There was a highly significant positive correlation between increases or decreases of 5% or more in oximetric SaO 2 and corresponding changes in arterial SaO 2 measured in the laboratory (r = 0.965, P < .001).

Continuous measurement and display of SaO 2 allows real time changes to be observed and therapeutic inter- vent ion commenced if necessary. Nineteen patients had an initial SaO 2 of less than 90% (mean PaO~ 54 +- 7.8 mm Hg). Interventions directed at

improving SaO a were successful in all but eight cases and included oxygenation (38 patients), bronchodilators (14), fluid management (13), diuretics lten), mechanical venti lat ion (five), digi- talization (four), and tube thorocostomy (one).

In eight patients, the SaO 2 remained below 90% despite initial resuscitation measures. All eight patients were subsequently hospitalized, five requiring intubation and mechanical ventilation in the ED. The use of oximetric SaO 2 to identify hypoxia associated with prolonged intubation at- tempts is shown (Figure 4). The data are from a 32-year-old man wit h pulmonary edema secondary to near drowning. The patient required ven- tilatory support and the use of positive end-expiratory pressure (PEEP). We were able to titrate changes in FIO 2 and PEEP by noting their effect on SaO 2.

The value of bedside monitoring Of SaO 2 in alerting ED personnel to the effects of various nursing maneuvers leg, administration of oxygen or suctioning) is shown (Figure 5). The trac- ing is from a 72-year-old man with congestive cardiomyopathy and acute pulmonary edema. In an attempt to avoid in tubat ion, hypoxemia was treated with high-flow oxygen (FIO2, 0.8). The oxygen mask was removed during tracheal suctioning with a subsequent decrease in SaO 2. We noted similar changes, often unexpected, in oxygen saturat ion wi th vomit ing, coughing, and repositioning of pa-

tients (Table). The oximetric monitor also serves

as a warning device for impending respiratory failure. During our study, two patients with a normal SaO 2 on admission had a gradual decrease in arterial saturation despite emergency therapy. These drops in 8aO 2 were confirmed by laboratory analysis. One patient, with acute exacerbation of asthma, responded slowly to IV the- ophylline and aerosol treatment. The second patient had a gunshot wound to the chest and subsequently developed a left hemothorax requiring tube thorocostomy,

DISCUSSION The diagnosis of respiratory illness

and the determination of its severity in a given patient are difficult at the bedside. There is a poor correlation between the extent of disease and the' symptoms related by the patient. Tra- ditionally, detection of acute hypoxia has d e p e n d e d - o n r e c o g n i t i o n of clinical signs, such as bradycardia or cyanosis, or arterial blood gas analysis. Clinical observation of cyanosis is only useful when blood oxygen ten- sions are below the range of 48 to 56 mm Hg. 8 This equates with a hemoglobin saturation of approximately 80%, the beginning of the steepest part of the oxyhemoglobin dissociation curve (Figure 3}. Once overt signs of hypoxemia are apparent, any fur- ther fall in the arterial oxygen tension will lead to a dramatic decline in oxygen delivery and reduction in the ter-

56/466 Annals of Emergency Medicine 17:5 May 1988

TABLE. Transient changes in SaO 2

Observations (No.)

Suctioning (13)

Coughing (9)

Intubation (5)

Vomiting (4)

Lying supine (6)

*Values are mean _+ standard deviation

SaO 2 Decrease* 5.3 + 3.2

3.1 _+ 0.9

5.2 _+ 3.9

3.8 _+ 1.4

2.6 _+ 1.7

minal blood/tissue gradient that deter- mines oxygen entry into the cells. 8

A monitoring system that detects covert hypoxemia before clinical deterioration of the patient has great po- tential for improving the quality of ED resuscitation. There are presently two noninvasive methods of objec- tively assessing a patient's oxygenation at the bedside: transcutaneous measurement of PaO 2 and pulse oximetry to estimate SaO 2. Until recently, PaO a measurement had been the method most commonly used, but the transcutaneous electrode has varying response t ime and m u s t be changed at regular intervals to avoid second-degree bums. 4

As a consequence of the shape of the oxyhemoglobin dissociation curve, the oxygen content of the blood of hy- poxic patients is more sensitively ex- pressed in terms of hemoglobin saturation than arterial oxygen tension, s On the steep portion of the curve, a large change in saturation is accom- panied by only a small change in oxygen tension (Figure 3).

Furthermore, defining oxygenation in terms of hemoglobin saturation avOids factors that may displace the oxyhemoglobin-affinity curve (eg, hypothermia, acidosis, blood transfusion) and thus influence the oxygen content of the blood at a given tension.a, 5

Pulse oximetry is a new method that measures SaO 2 directly by com- bining plethysmography with oximetry. It functions by positioning any pulsating arterial vascular bed between a two-wavelength light source and a photodetector. The pulsatile flow creates a transient change in the light path, modifying the amount of light received by the photodetector. Factors that reduce cutaneous blood flow, such as hypothermia, hypoten- sion, and vasopressor infusions, decrease the accuracy of transcutaneous oxygen monitors, but do not appear to

affect the accuracy of pulse oximetry as long as vascular pulsations can be detected.9,1o Thus, while pulse oximetry is valuable in assessing respiratory function, it may be less effective in assessing tissue oxygenation under low- flow conditions.

The use of only two wavelengths in- dicates that pulse oximeters do not differentiate between oxyhemoglobin (HbO2) , carboxyhemoglobin (HbCO), and methemoglobin (MetHb). The oximeter overestimates oxygen saturation in cases of increased HbCO or MetHb levels. The degree of error de- pends on the quantity of dyshemoglo- bin present, as well as the relative spectral extinction of the dyshemoglo- bin at the specific wavelength used. 4

Also, it should be remembered that pulse oximeters do not measure hemoglobin levels. In cases of hemodilu- tion, the monitor does show a satis- factory oxygen saturation to Hb level of 5 g/dL. 9 The presence of fetal hemoglobin does not appear to affect the accuracy of the oximeter, and there is a high degree of correlation between pulse oximetry values and blood gas determinations in newborn infants. 2

Despite these limitations our data show that oximetry can be an accurate, reliable tool in the real-time as- sessment of a patient's oxygenation status. Readings remained accurate over a wide range of SaO2, PaO2, ages, heart rates, hematocrits, and systolic pressures. Our study also demonstrated a highly significant positive correlation between increases or decreases of 5% or more in oximetric SaO a and corresponding changes in measured saturation (r = 0.965, P < .001). Thus, the probability of making erroneous clinical decisions based on oximetric data is small.

In the present series, those patients who did not have a progressive increase in SaO 2 during the initial ED resuscitation had ongoing hypoxia and

Annals of Emergency Medicine

poor clinical outcomes (eg, mechanical ventilation, prolonged hospital stays). Measurement of SaO 2 may therefore identify those patients with a poor prognosis who warrant more aggressive therapy, such as early intubation or tube thorocostomy. In these cases, any inaccurate reading is readily apparent from the pulse display.2, s

Oximetric SaO 2 levels below 91% indicate poor oxygen delivery in patients without longstanding pulmonary disease. 11 A rapid or prolonged decrease from this level can provide an early warning of a substant ia l deterioration in oxygenation. In two patients, the pulse oximeter detected an otherwise unrecognized drop in arterial saturation. As a result of this and other experiences, we routinely set the low limit alarm on the oximeter to approximately 91%.

Numerous reports during the last five years,2,4, 6-8 have shown that the response time of the oximetric system is short and the system is able to re- spond to rapidly changing conditions. This accurate and dynamic measurement facilitates optimal patient management by identifying acute changes in oxygen delivery and provides an al- mos t immed ia t e feedback of the efficacy of various therapeutic interventions. For example, the hypoxia following such routine nursing procedures as suctioning or repositioning in bed (Table) has led us to modify our approach to these procedures. Mem- bers of our nursing staff are now more sensitive to the need for hyperoxygen- ation before suctioning. Its use also has improved the safety of invasive procedures such as endotracheal or fiberoptic bronchoscopy.12

Finally, the titration of respiratory therapy (PEEP, inspired oxygen con- centration) by continuous measurement of SaO 2 decreased the need for repeti t ive measurements of blood gases, la Flick et a114 monitored the SaO 2 in ten patients wi th severe obstructive lung disease. They found the oximeter invaluable for establishing trends in arterial oxygenation, in- dicating appropriate times to draw blood for analysis and establishing the time course of desaturation episodes.

C O N C L U S I O N We have found that oximetric moni-

toring in selected patients in the ED is extremely valuable during pulmonary resuscitation. Such continuous real- time information may allow better as-

17:5 May 1988 467/57


sessment of patient status, facilitate titration of respiratory therapies, and serve as a useful warning system for acute deterioration in oxygenation,

T h e a u t h o r s t h a n k the ED n u r s e s for the i r pa t i ence , u n d e r s t a n d i n g , and he lp ; Dr E Logue for p r o v i d i n g s t a t i s t i c a l ana l y s i s ; a n d Dr J a m e s D o u g h e r t y for r e v i e w i n g t h i s m a n u s c r i p t . T h e y a l s o t h a n k Joe M e s s n e r and S u z a n n e Worces te r for prep- a ra t ion and typing.

REFERENCES 1. Tremper KK, Shoemaker We: Cont inuous CPt~ monitoring with t ranscutaneous oxygen and carbon dioxide sensors. Crit Care Med 1981;9:417-418.

2. Fanconi S, Soherty P, . . . . . . . ~umunu~ jP, ct ~.~" ~"!~o~ oximetry in pediatric in tensive care: Com- parison with measured saturations and transcutaneous oxygen tension. J Pediatr 1985;107: 362-366.

3. Tremper KK, Shoemaker We: Transcutane-

ous oxygen monitor ing of critically ill adults with and wi thout low flow shock. Crit Care Med 1981;9:706-709.

4. Yelderman M, New W: Evaluation of pulse oximetry. Anesthesiology 1983;59:349-352.

5. Deckhardt R, Steward DJ: Noninvasive arterial hemoglobin oxygen sa tura t ion versus t ranscutaneous oxygen tension monitoring in the pre term infant . Crit Care Med 1984;12: 935-939.

6. Miyasaka K: Use of non-invasive oximetry during the induction of anaesthesia in children, in Payne J~ Severinghaus ,..w¢ ,ed~l'-~ Pulse O~:- imetry. New York, Springer-Verlag, 1986, p 95-100.

7. Plenderleith JL, Dougall J, Asbury AJ: Use of the Ohmeda Biox III oximeter in an intensive care unit, in Payne JE Severinghaus JW (eds]: Pulse Oximetry. New York, Springer-Verlag, !986, p 56-6!.

8. Abbott MA: Monitoring oxygen saturation levels in the early recovery phase of general anaesthesia, in Payne JP, Severinghans JW (eds): Pulse Oximetry. New York, Springer-Verlag, 1986, p 165-171.

9. Mertzlufft F, Zander R: Non-invasive oximetry using the Biox III oximeter: Clinical evaluation and physiological aspects, in Payne ]P, Se veringhaus JW teds]: Pulse Oximetry. New York, Springer-Verlag, 1986, p 71-77.

10. Brooks TD, Gravenstein N: Pulse oximetry for early detection of hypoxemia in anesthe- tized infants. J Clin Monit 1985;1:135-137.

11. Elling A, Hanning CD: Oxygenation during postoperative transportation, in Payne JP, Sever- inghaus JW {eds): Pulse Oximetr~ New York, Springer-Verlag, 1986, p 161-164.

12. Rebuck AS, Chapman KR, D'Urzo A: The accuracy and response characteristics ot a s lm plified ear oximeter. Chest 1983;83:860-864.

13. Dautzenberg B, Gallinari C, Moreau A, et al: The advantages of real-time oximetry over in termit tent arterial blood gas analyses in a chest department, in Payne JR Severinghaus JW [eds}: Pulse Oximetrg New York, Springer-Vcr- lag, 1986, p 63-65.

14. Flick MR, Block J: C o n t i n u o u s in-vivo monitoring of arterial oxygenation in chronic obstruction of lung disease, Ann Intern Med 1977;86:725-730.

58/468 Annals of Emergency Medicine 17:5 May 1988

Documents

Continuous emergency department monitoring of arterial saturation in adult patients with respiratory distress