Best Practice & Research Clinical AnaesthesiologyVol. 19, No. 1, pp. 111–123, 2005

8

Does perioperative pulse oximetry improve

outcome? Seeking the best available evidence

to answer the clinical question

Tom Pedersen* MD, PhD

Director of the Head & Orthopaedic Centre, and Coordinating Editor,

The Cochrane Anaesthesia Review Group

Copenhagen University, Rigshospitalet, Blegdamsvej 9, 2100 Copenhagen, Denmark

The aim of this chapter is to clarify the effect of perioperative monitoring with pulse oximetry andto identify the adverse outcomes that might be prevented or improved by its use. Trials wereidentified by computerized searches of The Cochrane Library, MEDLINE and EMBASE, and bychecking the reference lists of trials and review articles. All controlled trials that randomizedpatients to either pulse oximetry or no pulse oximetry during the perioperative period, includingthe operating and recovery room, were included in the search. Pulse oximetry monitoring wasshown to substantially reduce the extent of perioperative hypoxaemia, enabling the detection andtreatment of hypoxaemia and related respiratory events, and promoting several changes inpatient care. The implementation of perioperative pulse oximetry monitoring does notsignificantly reduce the number of postoperative complications, but the question remainswhether pulse oximetry can improve outcomes in other situations. Pulse oximetry has beenadopted all over the world in clinical practice as a tool that guides anaesthesiologists in the dailymanagement of patients: in teaching situations, in emergencies, and especially in the care ofchildren. Given the relatively small number of patients studied in these trials and the rare eventsbeing sought, the studies of perioperative monitoring with pulse oximetry were not able to showan improvement in the outcomes studied.

Key words: monitoring; pulse oximetry; perioperative; hypoxaemia; cognitive dysfunction;complications; arterial oxygen saturation; monitoring; respiratory monitoring.

Hypoxaemia and perioperative respiratory complications have long been recognized asa risk to patients in the operating room and postanaesthesia care unit. Monitoring ofhaemoglobin oxygen saturation with pulse oximetry has become standard care in theseareas.

doi:10.1016/j.bpa.2004.07.002available online at http://www.sciencedirect.com

1521-6896/$ - see front matter Q 2005 Elsevier Ltd. All rights reserved.

* Tel.: C45-4535-3050.

E-mail address: doctp@yahoo.com

112 T. Pedersen

There is a need to clarify the real effect of perioperative monitoring with pulseoximetry and to identify the adverse outcomes that might be either prevented orimproved by the use of this technique. There is growing evidence to suggest that laterpostoperative hypoxaemia may also play a role in organ dysfunction leading tomorbidity and mortality. Economic pressures to move patients earlier from expensivepostoperative recovery and intensive care areas to the general care floor (where nurse-to-patient ratios are lower) may lead to inadequate surveillance of patients at risk.

This chapter discusses whether the use of pulse oximetry is associated with animprovement in the detection and treatment of hypoxaemia, and whether the use ofpulse oximetry per se reduces morbidity as well as mortality in the perioperativeperiod.

MONITORING SATURATION WITH PULSE OXIMETRY

Arterial hypoxaemia, particularly that associated with hypoventilation, is a major risk topatients during surgical anaesthesia, especially for abdominal and thoracic procedures,and it continues to be a risk through the early postoperative period in thepostanaesthesia care unit (PACU). While the risk of hypoxaemia has generally beenthought to end by the time of discharge from the PACU, a growing body of thoughtsuggests that hypoxaemia can persist in some patients on the general care floor duringthe first postoperative week.1

The risk of hypoxaemia and the need for pulse oximetry monitoring of haemoglobinsaturation in both the operating room and the postanaesthesia care unit are wellrecognized and have been formalized in treatment standards adopted by the AmericanSociety of Anesthesiologists (ASA).2 These standards were instituted to help maximizesuccessful tolerance of—and recovery from—ventilatory and respiratory pertur-bations associated with anaesthesia and surgery, and they have received widespreadsupport and compliance within the anaesthesiology community. Since its introductioninto clinical practice in the mid-1980s, pulse oximetry monitoring has enabledcontinuous measurement of blood haemoglobin saturation in a variety of clinicalsettings and in virtually all of the studies cited above. Its clinical and scientific utilitystems from the fact that it provides a non-invasive and continuous estimation of anotherwise difficult-to-measure clinical parameter. The instrumentation is relativelyinexpensive, the technology is stable, and it is amenable to computer networking forcentral monitoring, as described in recent comprehensive reviews.3–5

As with all clinical instrumentation, there are limitations that must be recognized andmitigated for effective use of pulse oximetry in clinical settings.6,7 Pulse oximetrymonitoring provides a spectroscopic estimate of haemoglobin oxygen saturation ratherthan a direct measure of the total oxygen content of the blood, yet haemoglobinsaturation fraction is generally taken to reflect changes of ventilatory and respiratorystatus rather than changes of haemoglobin concentration that would affect total oxygencontent. It is important to remember, also, that registered oximetry values can beconfounded by methaemaglobinaemias, certain vital dyes, excessive ambient light, andcircumstances where tissue perfusion is limited. In addition, false alarms may occur dueto artefacts of patient motion, transducer displacement, or when there is venouscongestion or pulsation.8,9 Recently developed artefact rejection algorithms can betterdistinguish arterial pulse waveforms (i.e. rapid onset, slow decay) from movement-induced pulse (generally slow onset and decay) and can substantially reduce

Perioperative pulse oximetry 113

the incidence of false alarms and movement artefacts.9,10 Although these technicallimitations can be controlled11,12, the clinical effectiveness of pulse oximetry dependson establishing clear guidelines for using pulse oximetry monitoring and proper trainingfor medical and nursing staff in its theory and practice.13–15

EXPERIENCE WITH PULSE OXIMETRY IN THE PERIOPERATIVEPERIOD

The introduction of the pulse oximeter, a clinical monitor of oxygen saturation andpulsation, has made it possible to monitor perioperative hypoxaemia with a non-invasive continuous measuring technique.16 The main value of pulse oximetry is that itcan provide an early warning of hypoxaemia. For the individual patient, it isunpredictable at what level of hypoxaemia the brain, heart and other organs willsuffer, and to what extent irreversible damage may occur. Many factors—such ascardiac output, haemoglobin concentration and oxygen demand—can affectthe smallest tolerable value of oxyhaemoglobin saturation.17 The occurrence andpossible pathogenesis of perioperative hypoxaemia were first described many yearsago.18,19

Monitoring with pulse oximetry might improve patient outcome by enabling an earlydiagnosis and, consequently, correction of perioperative events that might causepostoperative complications or even death.20 An operational definition of an event is anundesirable incident that required intervention and did, or possibly could, causecomplications or death. Such events may be attributed to pathophysiological processes,to malfunction of gas supply or equipment, or to human error—e.g. oesophagealintubation or anaesthetic mismanagement.21 For many of these events, hypoxaemia isprobably the most common mechanism responsible for the eventual adverseoutcomes.22 Therefore, the monitoring of perioperative oxygenation may improve apatient’s outcome following anaesthesia.

Recent studies have suggested that hypoxaemia in the operating theatre andrecovery room is common, and that monitoring with pulse oximetry permits its earlydiagnosis and treatment of hypoxaemia.23,24 Thus pulse oximetry can reduce theincidence and severity of hypoxaemia. Only a few randomized clinical trials of pulseoximetry have been performed during anaesthesia and in the recovery room whichdescribe the perioperative hypoxaemic events, postoperative cardiopulmonarycomplications and cognitive dysfunction.24–27

SEEKING THE BEST AVAILABLE EVIDENCE TO ANSWERTHE CLINICAL QUESTION: DOES PERIOPERATIVE PULSEOXIMETRY IMPROVE OUTCOME?

Randomized controlled trials were identified by searching MEDLINE (1966–2003),EMBASE (1980–2003), The Cochrane Anaesthesia Review Group’s specializedregister, and The Cochrane Library 2003, issue 128,29, using the Cochrane searchstrategy.28 The bibliography of each article was scanned for relevant references. Studiespublished in any language were considered. Studies included in the systematic reviewhad to fulfil the following criteria: randomized or quasi-randomized controlled trialsdealing with the use of pulse oximetry or no pulse oximetry during the perioperative

114 T. Pedersen

period, including the operating and recovery room. Trials were included irrespectiveof their blinding, number of patients randomized, or the language of the article. Theparticipants were patients 18 years of age or older undergoing surgery withanaesthesia.

The use of pulse oximetry according to outcomes

In the systematic Cochrane review28, the following outcomes were studied:

1.

Events detectable by pulse oximetry: hypoxaemia (SpO2!90% corresponding to anarterial oxygen tension !60 mmHg).

2.

Causes of events: patient respiratory causes of hypoxaemia (pneumothorax,bronchospasm, air embolus, respiratory depression, apnoea, airway obstruction,pneumonia, ventilatory failure and pulmonary embolus); patient mechanical causesof hypoxaemia (oesophageal or main stem intubation, mucus plug, kinkedendotracheal tube), delivery system causes of hypoxaemia (anaesthesia machineand gas supply problems).

3.

Interventions that may prevent, attenuate or shorten these events: airway support,endotracheal intubation, manual or mechanical ventilation, oxygen treatment,pressors and inotropes, fluid treatment.

4.

Any serious complications occurring during anaesthesia or in the postoperativeperiod. These comprised postoperative intensive care admission due to respiratoryinsufficiency, circulatory insufficiency, or infection. These complications weredefined as follows. Respiratory insufficiency: pneumonia (fever, chest radiographor positive culture), atelectasis (chest radiograph), pneumothorax (diagnosed onchest radiograph), and respiratory insufficiency (requiring intervention); andcardiovascular insufficiency: cardiac arrest, cardiac failure or myocardial infarction.Other complications were renal and hepatic insufficiency, neurological and cognitivedysfunction, and serious infection requiring antibiotics.

5.

Intra- or postoperative mortality. 6. Duration of recovery or intensive care stay.

Types of outcome measures

†

hypoxaemia † respiratory events † interventions that may prevent, attenuate or shorten these events † any serious complications that occurs in the perioperative period † intra- or postoperative mortality † duration of recovery or intensive care stay

WHAT WERE THE RESULTS OF THE RANDOMIZED CONTROLLEDSTUDIES IN RELATION TO OUTCOMES?

The systematic Cochrane review28 searching yielded six reports. Of those six reports,two were judged ineligible for analysis because they either had no control group or theylacked information on the relevant postoperative outcomes.21,30 All outcome measures

Table 1. Details of the randomized controlled trials of pulse oximeter monitoring versus no pulse

oximeter monitoring in the perioperative period.

Authors Methods

Number of

patients Outcomes

Bierman et al, 199231 Blinded,

control group

20/15 Clinically unapparent

desaturationa

Moller et al, 199232 Blinded,

control group

100/100 Hypoxaemia in the

recovery roomb

Moller et al, 199334 Quasi-randomized

Blinded comparison

358/378 Postoperative cognitive

dysfunctionc

Moller et al, 199335 Quasi-randomized 10 312/10 490 Postoperative complications:

respiratory,

cardiovascular, and infectious

a Clinically undetected episodes of desaturation for more than 1 minute.b Saturation !85% for more than 30 seconds during the recovery period.c Problems with recent memory (telephone numbers, appointment, etc.).

Perioperative pulse oximetry 115

in the included studies were extracted and can be seen in detail in Table 1. The types ofoutcome measures were separated into events detectable by pulse oximetry that couldresult in complications, and perioperative complications detected in other ways. Theincluded studies31,32,34,35 used a number of different ways of assessing postoperativeoutcome:

1.

Events with hypoxaemia measured either with blood gas analyses or pulse oximetry. 2. Tests of cognitive function: Wechsler memory scale, continuous reaction time and

subjective perception of cognitive dysfunction (test of memory).

3. Clinical outcome: respiratory, cardiovascular and neurological complications

following anaesthesia.

Studies using blood gas analysis and pulse oximetry to assess hypoxaemia

In the Bierman study31, the staff were instructed to use the pulse oximetry data in lieuof arterial blood gas analysis whenever possible. The desaturation alarm on theoximeters was set to sound at values %93%. For patients in group 2, blood gasanalyses were obtained every hour or more frequently as was clinically indicated.These patients were monitored with a modified pulse oximeter from which SpO2 wasrecorded continuously at a distant site, but not displayed at the patients’ bedside. Thealarms for desaturation and pulse rate were deactivated in the bedside unit. The useof pulse oximetry allowed a significant reduction in arterial blood gases without anyadverse events. Clinically unapparent desaturations were detected in seven of 15patients in the group without pulse oximetry, in comparison with none detected inthe pulse oximetry group. There was no difference between the two groups in thenumber of changes in ventilatory support per postoperative intensive care unit (ICU)stay, whereas the dose of supplemental oxygen was adjusted more frequently in thegroup without pulse oximetry. There was no evidence of a significant differencebetween the groups regarding duration of postoperative mechanical ventilation orICU stay.

Table 2. Effect of pulse oximeter perioperative monitoring on postoperative hypoxaemia and

complications.

Authors Outcomes

Pulse

oximeter

group (%)

Control

group (%)

Odds ratio

(CI 95%)

Bierman et al, 199231 Clinically undetected

desaturation in

intensive care unit

0.0 46.7 –

Moller et al, 199232 Hypoxaemia in the

recovery room

7.0 31.0 0.17a (0.07–0.40)

Moller et al, 199334 Late cognitive

dysfunction

7.0 11.0 0.61 (0.36–1.03)

Moller et al, 199335 Respiratory 3.8 3.2 1.21a (1.04–1.40)

Cardiovascular 2.9 2.7 1.07 (0.91–1.27)

Total complications 10.0 9.4 1.07 (0.98–1.17)

a P!0.05.

116 T. Pedersen

Moller et al32 found that hypoxaemia was reduced in the pulse oximetry group bothin the operating theatre and in the recovery room (Table 2). During observation in therecovery room, the incidence of hypoxaemia in the pulse oximetry group was 1.5–3times less than in the group without pulse oximetry. No patient in the pulse oximetrygroup experienced extreme or severe hypoxaemia.

As a consequence of the pulse oximetry monitoring, changes in the recovery roomresulted in several interventions. For example, in the recovery room, on average, thepatients in the oximetry group received an increased FiO2; more patients in this groupreceived naloxone; and the patients had a longer stay.33 The number of patientsdischarged from the recovery room with an order for supplemental oxygen was 13.3%in the oximeter group and 3.5% in the control group.

Study using tests of cognitive dysfunction

Moller et al34 demonstrated that postoperative cognitive function as measured by theWechsler memory scale and continuous reaction time was independent ofperioperative monitoring with pulse oximetry. The postoperative subjective reports(by questionnaire) of cognitive deficits revealed no difference: 7% of patients in thepulse oximetry group, and 11% in the group without pulse oximetry believed that theircognitive abilities had decreased (Table 2). There was no significant difference betweenthe two groups in ability to concentrate (10 versus 9%). The study showed no evidenceof less postoperative cognitive impairment after perioperative monitoring with pulseoximetry.

Study using clinical measures of complications to discharge

The study of Moller et al35—which included 20 802 surgical patients assigned to bemonitored with pulse oximetry or not—found that one or more postoperativecomplications occurred in 10% of the patients in the oximetry group and in 9.4% in thecontrol group. The two groups did not differ in cardiovascular, neurological, orinfectious complications; however, the incidence of respiratory complications was 3.8%

Perioperative pulse oximetry 117

in the pulsoximeter group compared with 3.2% in the control group (odds ratioZ1.21)(Table 2). The duration of hospital stay was a median of 5 days in both groups. Equalnumbers of in-hospital deaths occurred in the two groups: 1.1% in the oximeter groupand 1.0% in the control group. Seven deaths were classified as possibly anaesthesia-related: three deaths were in the oximetry group and four in the control group. Theseven deaths did not display any specific pattern. A questionnaire, completed by theanaesthesiologists, revealed that 18% of the anaesthesiologists had experienced asituation in which a pulse oximeter monitor helped to avoid either a serious event orcomplication, and that 80% of the anaesthesiologists felt more secure when they used apulse oximeter. Although monitoring with pulse oximetry prompted a number ofchanges in patient care, no evidence was found to indicate a significant reduction in theoverall rate of postoperative complications using perioperative pulse oximetry. Thestudies confirmed that pulse oximetry can detect hypoxaemia and related events.However, in our searching we were unable to find reliable evidence that pulse oximetryaffects the outcome of anaesthesia.

PERIOPERATIVE HYPOXAEMIA AND POSTOPERATIVECOMPLICATIONS

It is difficult to make a comparison between the overall rates of perioperativehypoxaemia and postoperative complications in the present randomized studiesbecause of the few study populations and differences in type of outcomes. It wouldappear that in the present studies the general rates of hypoxaemia and complicationare at the same level as in other studies.36–39

The positive effects of pulse oximetry monitoring—reduced hypoxaemia, the abilityto detect and correct potential harmful events, and the finding of several changes inpatient care—must be compared with the fact that no reduction in the number ofpostoperative complications was found.35 It is also worth emphasizing that patientsmonitored with pulse oximetry have no change in their outcome, despite the fact thatthey tend to be given more oxygen and naloxone than patients monitored withoutpulse oximetry.35 This would suggest that merely increasing a patient’s saturation levelsfrom marginal to satisfactory is probably not going to make any difference to theiroutcome. In other words, the use of pulse oximetry as an early warning of moderatehypoxaemia does not appear to be beneficial, even if the appropriate responses areinstituted earlier than they would have been without pulse oximetry. This resultconflicts with most anaesthesiologists’ beliefs. In the closed claims analyses of adverserespiratory events in anaesthesia, reviewers judged that better monitoring would haveprevented adverse outcome in 72% of the claims.40,41 In the general analysis of the roleof monitoring devices in the prevention of anaesthetic mishaps, nearly 60% of theinstances of death and brain damage were considered preventable by application ofadditional monitors. These studies exhibit a number of limitations, including absence ofa control group, a probable bias toward adverse outcomes, and reliance on data fromdirect participants rather than objective observers. The study of Moller et al35 haslimitations. If he had involved a substantially greater number of patients, perhaps hewould have shown significant outcome differences (a type II error). The fact that theconclusions of the closed claims study were different from those of Moller’s study mightsuggest power differences rather than lack of efficacy of the pulse oximeter. The studyof rare events requires unusual methodology, and randomized controlled trials may not

118 T. Pedersen

be an appropriate method of looking at outcome measures in a study of anaestheticcomplications.

POSTOPERATIVE COGNITIVE DYSFUNCTION

The relationship between perioperative hypoxaemia and impaired postoperativecognitive function has been debated.42 Moller et al found that 9% of the surgical patientsthought that their mental function had deteriorated.34 In their recent study they foundthat postoperative cognitive dysfunction in the elderly, as identified with neuropsycho-logical tests, was present in 25.8% of patients 1 week after surgery, and in 9.9%3 months after surgery.27 However, hypoxaemia was not a significant risk factor forcognitive dysfunction at any time. Perioperative monitoring with pulse oximetry did notappear to affect the patients’ postoperative cognitive function.

Methods of assessing cognitive function

Cognitive function tests are valuable when studying anaesthetic drug effects; however, anumber of patients have unexplained complaints of impaired cognitive function that arenot verified by objective tests.27 One may speculate that application of a broader rangeof neuropsychological assessments that were used in the study27 could have detectedvarying deficits of an enduring nature. Using a broad range of tests, several investigatorshave described moderate to severe cognitive dysfunction lasting for several monthsafter coronary bypass surgery.43

WHY ARE RANDOMIZED CONTROLLED TRIALS PREFERABLE?

The real value of randomization is that, if it is done properly, it reduces the risk ofserious imbalance in unknown but important factors that could influence the clinicalcourse of the participants. No other study design allows investigators to balance theseunknown factors. Proper randomization is the only means by which one can berelatively confident that there are no fundamental differences between groups(concerning known and unknown prognostic factors). When the method ofrandomization is not described, or is open to manipulation (through allocation bybirth date, for example), treatment effect can be overestimated by 30–50%.44–46 Thereis also a clear tendency to overestimate effect in the control groups of historical studiesand in studies that are unblinded.44,46 This overestimation can be 30–40%47 or evengreater if there are very few patients in the study.48 When observational cohort studiesare subjected to strict epidemiological principles, this discrepancy can be partiallyoffset. The results can also undervalue the treatment effect, which means that thenewer techniques do not make the observational studies more reliable.48,49

METHODOLOGICAL QUALITY OF RANDOMIZED CONTROLLEDSTUDIES

The quality of blinding and allocation concealment of included studies are variable.Power analyses were seldom conducted to determine adequate sample sizes.

Perioperative pulse oximetry 119

Consequently, even the studies with high-quality blinding and allocation concealmentmay still not provide reliable results. Furthermore, the authors of the Cochranesystematic review28 did not perform a meta-analysis because of the few studies, thevery different outcomes measured in the studies, and the heterogeneity in the studies.

PULSE OXIMETRY MONITORING AND THE SAFETYOF ANAESTHESIA

The proliferation of monitors in anaesthesia is obvious. The goal of monitoring as anadjunct to clinical decision-making is to directly or indirectly reduce the incidence ofcomplications. This is based on the premise that unambiguous and accurateinformation, readily interpretable and available, will assist the anaesthesiologist indeciding upon and initiating correct therapeutic interventions. The constantdevelopment of new anaesthesia monitoring instruments means that the anaesthesiol-ogist will be faced with an overabundance of data about the patient and anaestheticdelivery system. The unanswered question is whether the individual anaesthesiologist’sperformance—the human factor—is of far more importance than the implementationof new monitoring equipment or other new safety initiatives in reducing the rate ofpostoperative complications.

Another factor which has to be considered is that the overwhelming majority ofpatients in the studies come from a region where standards of anaesthesia and nursingcare are good. Almost all the data were collected by a single group of people. This doesreduce the possibility of generalizing the results in terms of what might be found inother geographical areas where standards of care and assessment methods may bedifferent. Since detected hypoxic events were treated, we do not really know what thedifferences in outcome would have been if hypoxic events were neither detected nortreated. The studies were relatively controlled and did not reproduce situations wherethere was a high likelihood of disaster.

RESOURCE-EFFECTIVENESS ISSUES

Although additional and more definitive data would be desirable, pulse oximetrymonitoring on the general care floor has been advocated for at-risk patients as a cost-effective approach to maximizing quality of care while enhancing risk management. Ifadequate guidelines for use and suitable staff training are in place then the relatively lowexpense of pulse oximetry systems may be offset by operational cost savings.50 Becausepulse oximetry can easily indicate whether a patient needs oxygen therapy or not,prophylactic and therapeutic oxygen use can be rationalized and minimized. In addition,use of oximetry can reduce the number of diagnostic blood gas analyses that need to beperformed.31 Further, telemetric pulse oximetry systems can aid in the efficientutilization of nursing resources. By identifying the contribution of hypoxaemia to thedisruptive and distressing effects of postoperative delirium, oximetry monitoring canalso improve patient care and even patient quality of life. Perhaps more significantly,oximetry use in late postoperative patients could potentially decrease the risk ofcatastrophic late sequellae of anaesthesia and surgery, as well as the specific dangers ofventilatory depression from pain therapy opiates—whether administered intrave-nously, epidurally, or intrathecally—with virtually no added risk of its own.

120 T. Pedersen

Furthermore, this potential benefit appears to be available at a favourable cost. Oneearly cost-effectiveness analysis estimated that one fatal hypoxaemic episode in 40 000moderately hypoxaemic patients would justify the cost of the oximetry.50

IMPLICATIONS FOR PRACTICE

Pulse oximetry monitoring substantially reduced the extent of perioperativehypoxaemia, enabled the detection and treatment of hypoxaemia and relatedrespiratory events, and promoted several changes in patient care. The implementationof perioperative pulse oximetry monitoring, however, was not the breakthrough whichcould reduce the number of postoperative complications, and the question remainswhether pulse oximetry improves outcomes in other situations. Pulse oximetry hasalready been adopted all over the world in clinical practice, and may be a tool thatguides the anaesthesiologist in the daily management of patients, in teaching situations,in emergencies, and especially in care of the children. The results of the studies, despitean intense, methodical collection of data from a large population, indicate thatperioperative monitoring with pulse oximetry does not improve reliable outcomes,effectiveness and efficiency.

IMPLICATIONS FOR RESEARCH

It is time to focus on which technologies are needed. The science of human factorsincludes the psychological and mental influences affecting performance in theworkplace. Monitoring systems should fit in naturally with the way the anaesthetistworks, thinks, and interacts with the patient, equipment, and operating roomenvironment in order to bring significant benefits.

Future work in this area would benefit from greater attention to methods ofrandomization, as only a few appropriate randomized studies were found. The impactof new monitoring equipment on outcomes should be evaluated and controlled, as iscustomary when introducing new drugs. As the present studies illustrate, the problemsare multitudinous. The worst problem is clearly the huge number of patients needed. Bylimiting the inclusion criteria to specific subgroups of patients (e.g. elderly patients agedO65 years, patients with cardiac risk factors, ASA physical status III and IV, acuteabdominal surgery), isolating more rigorous outcome variables, and establishing wideco-operation between departments and countries, new monitoring and anaesthesiasafety studies could be launched in the future.

SUMMARY

Monitoring with pulse oximetry might improve patient outcome by enabling an earlydiagnosis and, consequently, correction of perioperative events that might otherwisecause postoperative complications or even death. Trials were identified bycomputerized searches of The Cochrane Library, MEDLINE and EMBASE, and bychecking the reference lists of trials and review articles. All controlled trials thatrandomized patients to either pulse oximetry or no pulse oximetry during theperioperative period, including the operating and recovery room, were included in

Perioperative pulse oximetry 121

the search. The search identified six reports. Of these six reports, four studies withdata from a total of 21 773 patients were considered eligible for analysis. Two studiesspecifically addressed the outcomes in question; neither found any effect on the rate ofpostoperative complications using perioperative pulse oximetry. Hypoxaemia wasfound to be reduced in the pulse oximetry group both in the operating theatre and inthe recovery room. During observation in the recovery room, the incidence ofhypoxaemia in the pulse oximetry group was 1.5–3 times less. Postoperativecomplications occurred in 10% of the patients in the oximetry group and in 9.4% inthe control group. The duration of hospital stay was similar in both groups, and equalnumbers of in-hospital deaths were registered in both groups. The studies confirmedthat pulse oximetry can detect hypoxaemia and related events. However, therandomized controlled trials were not able to show an improvement in variousoutcomes.

Practice points

† randomized controlled trials confirmed that pulse oximetry can detecthypoxaemia and related events

† no evidence was found that pulse oximetry affects the outcome of anaesthesiain the perioperative period

† the conflicting subjective and objective results of the trials, despite an intense,methodical collection of data from a relatively large population, indicate thatthe value of perioperative monitoring with pulse oximetry is questionable inrelation to improved reliable outcomes, effectiveness and efficiency

REFERENCES

1. Eichhorn JH. Pulse oximetry monitoring and late postoperative hypoxemia on the general care floor.

Journal of Clinical Monitoring and Computing 1998; 14: 49–55.

2. Eichhorn JH. Pulse oximetry as a standard of practice in anesthesia. Anesthesiology 1993; 78: 423–426.

3. Hanning CD & Alexander-Williams JM. Pulse oximetry: a practical review. BMJ 1995; 311: 367–370.

4. Lindberg LG, Lennmarken C & Vegfors M. Pulse oximetry—clinical implications and recent technical

developments. Acta Anaesthesiologica Scandinavica 1995; 39: 279–282.

5. Severinghaus JW. History and recent developments in pulse oximetry. Scandinavian Journal of Clinical and

Laboratory Investigation. Supplementum 1993; 214(supplement): 105–111. 648–50.

6. Inman KJ, Sibbald WJ, Rutledge FS et al. Does implementing pulse oximetry in a critical care unit result in

substantial arterial blood gas savings? Chest 1993; 104: 542–546.

7. Mardirossian G & Schneider RE. Limitations of pulse oximetry. Anesthesia Progress 1993; 39: 194–196.

8. Pan PH & Gravenstein N. Intraoperative pulse oximetry: frequency and distribution of discrepant data.

Journal of Clinical Anesthesia 1994; 6: 491–495.

9. Wiklund L, Hok B, Stahl K & Jordeby-Jonsson A. Post anaesthesia monitoring revisited: frequency of true

and false alarms from different monitoring devices. Journal of Clinical Anesthesia 1994; 6: 182–188.

10. Lie C, Kehlet H & Rosenberg J. Comparison of the Nellcor N-200 and N-3000 pulse oximeters during

simulated post-operative activities. Anaesthesia 1997; 52: 450–452.

11. Plummer JL, Ilsley AH, Fronsko RRL & Owen H. Identification of movement artefact by the Nellcor N-200

and N-3000 pulse oximeters. Journal of Clinical Monitoring 1997; 13(2): 109–113.

12. Porter R, Merlin M & Heller M. The fifth vital sign. Emergency 1990; 22: 37–41.

122 T. Pedersen

13. Stoneham MD, Saville GM & Wilson IH. Knowledge about pulse oximetry among medical and nursing

staff. Lancet 1994; 344: 1339–1342.

14. Bowton DL, Scuderi PE, Harris L & Haponik EF. Pulse oximetry monitoring outside the intensive care

unit: progress or problem? Annals of Internal Medicine 1991; 115: 450–454.

15. McGaffigan PA. Hazards of hypoxaemia: how to protect your patient from low oxygen levels. Nursing

1996; 26: 41–46.

16. Severinghaus JW & Kelleher JF. Recent developments in pulse oximetry. Anesthesiology 1992; 76:

1018–1038.

17. Bendixen HH, Hedley-Whyte J & Laver MB. Impaired oxygenation in surgical patients during

general anaesthesia with controlled ventilation. The New England Journal of Medicine 1963; 269: 991–

996.

18. Laver MB, Morgan J, Bendixen HH & Radford Jr. EP. Lung volume, compliance, and arterial oxygen

tensions during controlled ventilation. Journal of Applied Physiology 1964; 19: 725–733.

19. Nunn JF, Bergman NA & Coleman AJ. Factors influencing the arterial oxygen tension during anaesthesia

with artificial ventilation. British Journal of Anaesthesia 1965; 37: 898–914.

20. Cooper JB, Newbower RS & Kitz RJ. An analysis of major errors and equipment failures in anesthesia

management: considerations for prevention and detection. Anesthesiology 1984; 60: 34–42.

21. Mateer JR, Olson DW, Stueven HA & Aufderheide TP. Continuous pulse oximetry during emergency

endotracheal intubation. Annals of Emergency Medicine 1993; 22: 675–679.

22. Cooper JB, Cullen DJ, Nemeskal R et al. Effects of information feedback and pulse oximetry on the

incidence of anesthesia complications. Anaesthesiolgy 1987; 67: 686–694.

23. Moller JT, Wittrup M & Johansen SH. Hypoxaemia in the post-anaesthesia care unit: an observer study.

Anesthesiology 1990; 73: 890–895.

24. Cote CJ, Rolf N, Liu LM et al. A single-blind study of combined pulse oximetry and capnography in

children. Anesthesiology 1991; 74: 980–987.

25. Cote CJ, Goldstein EA, Cote MA et al. A single-blind study of pulse oximetry in children. Anesthesiology

1988; 68: 184–188.

26. Moller JT. Anaesthesia Related Hypoxaemia. The Effect of Pulse Oximetry Monitoring on Perioperative Events

and Postoperative Complications (thesis). Copenhagen: Laegeforeningens Forlag; 1994.

27. Moller JT, Cluitmans P, Rasmussen LS et al. Long-term postoperative cognitive dysfunction in the elderly

ISPOCD1 study. ISPOCD investigators. International study of post-operative cognitive dysfunction.

Lancet 1998; 351: 857–861.

28. Pedersen T, Pedersen BD & Moller AM. Pulse oximetry for perioperative monitoring. Cochrane Database

of Systematic Reviews 2003; 3: CD002013. Review.

29. Pedersen T, Moller AM & Pedersen BD. Pulse oximetry for perioperative monitoring: systematic review

of randomized, controlled trials. Anesthesia and Analgesia 2003; 96(2): 426–431.

30. Cullen DJ, Nemeskal AR, Cooper JB et al. Effect of pulse oximetry, age, and ASA physical status on the

frequency of patients admitted unexpectedly to a postoperative intensive care unit and the severity of

their anesthesia-related complications. Anesthesia and Analgesia 1992; 74: 181–188.

31. Bierman MI, Stein KL & Snyder JV. Pulse oximetry in the postoperative care of cardiac surgical patients.

A randomized controlled trail. Chest 1992; 102: 1367–1370.

32. Moller JT, Jensen PF, Johannessen NW et al. Hypoxaemia is reduced by pulse oximetry monitoring in the

operating theatre and in the recovery room. British Journal of Anaesthesia 1992; 68: 146–150.

33. Moller JT, Pedersen T, Rasmussen LS et al. Randomized evaluation of pulse oximetry in 20,802 patients: I.

Design, demography, pulse oximetry failure rate, and overall complication rate. Anesthesiology 1993; 78:

436–444.

34. Moller JT, Svennild I, Johannessen NW et al. Perioperative monitoring with pulse oximetry and late

postoperative cognitive dysfunction. British Journal of Anaesthesia 1993; 71: 340–347.

35. Moller JT, Johannessen NW, Espersen K et al. Randomized evaluation of pulse oximetry in

20,802 patients: II. Perioperative events and postoperative complications. Anesthesiology 1993; 78:

445–453.

36. Mlinaric J, Nincevic N, Kostov D & Gnjatovic D. Pulse oximetry and capnometry in the prevention of

perioperative morbidity and mortality. Lijecnicki Vjesnik 1997; 119: 113–116.

37. Pedersen T. Complications and Death Following Anaesthesia (thesis). Copenhagen: Laegeforeningens Forlag;

1994.

Perioperative pulse oximetry 123

38. Rheineck-Leyssius AT, Kalkman CJ & Trouwborst A. Influence of motivation of care providers on

the incidence of postoperative hypoxaemia in the recovery room. British Journal of Anaesthesia 1996; 77:

453–457.

39. Stausholm K, Rosenberg-Adamsen S, Edvardsen L et al. Validation of pulse oximetry for monitoring of

hypoxaemic episodes in the late postoperative period. British Journal of Anaesthesia 1997; 78: 86–87.

40. Tinker JH, Dull DL, Caplan RA et al. Role of monitoring devices in prevention of anesthetic mishaps:

a closed claims analysis. Anesthesiology 1989; 71: 541–546.

41. Caplan RA, Posner KL, Ward RJ & Cheney FW. Adverse respiratory events in anesthesia: a closed claims

analysis. Anesthesiology 1990; 72: 828–833.

42. Krasheninnikoff M, Ellitsgaard N, Rude C et al. Hypoxaemia after osteosynthesis of hip fractures.

International Orthopaedics 1993; 17: 27–29.

43. Townes BD, Bashein G, Hornbein TF et al. Neurobehavioral outcomes in cardiac operations.

A prospective controlled study. The Journal of Thoracic and Cardiovascular Surgery 1989; 98: 774–782.

44. Schulz KF, Chalmers I, Hayes RJ & Altman DG. Empiral evidence of bias. Dimensions of methodological

quality associated with estimates of treatment effects in controlled trials. JAMA 1995; 273: 408–412.

45. Moher D, Pham B, Jones A et al. Does quality of reports of randomised trials affect estimates of

intervention efficacy reported in meta-analyses? Lancet 1998; 352: 609–613.

46. Kjærgaard LL, Villumsen J & Gluud C. Quality of randomised clinical trials affects estimates of intervention

efficacy. VII Cochrane Colloquium, Rome 1999; B10: 57.

47. Sacks H, Chalmers TC & Smith Jr. H. Randomized versus historical controls for clinical trials. The American

Journal of Medicine 1982; 72: 233–240.

48. Berlin JA, Begg CB & Louis TA. An assessment of publication bias using a sample of published clinical trials.

The Journal of the American Statistical Association 1989; 84: 381–392.

49. McKee M, Britton A, Black N & McPherson K. Choose in between randomised and non-randomised

studies. In Black N, Brazier J, Fitzpatrick R & Reeves B (eds.) Health Services Research Methods. London:

BMJ books, 1998, pp. 61–72.

50. Stausholm K, Kehlet H & Rosenberg J. Oxygen therapy reduces postoperative tachycardia. Anaesthesia

1995; 50: 737–739.