PERIOPERATIVE MONITORING WITH PULSE OXIMETRY AND LATE POSTOPERATIVE COGNITIVE DYSFUNCTION

British Journal of Anaesthesia 1993; 71: 340-347

PERIOPERATIVE MONITORING WITH PULSE OXIMETRY AND LATE POSTOPERATIVE COGNITIVE DYSFUNCTION J. T. MOLLER, I. SVENNILD, N. W. JOHANNESSEN, P. F. JENSEN, K. ESPERSEN, J. S. GRAVENSTEIN, J. B. COOPER, M. DJERNES AND S. H. JOHANSEN SUMMARY

Anaesth. 1993; 71 : 340-347) KEY WORDS Complications: hypoxaemia. Memory.

Postoperative cognitive dysfunction has been reported even after non-cardiac and non-neurosurgical operations and anaesthesia [1-5]. Perioperative hypoxaemia has been suggested as a major risk factor leading to postoperative cognitive dysfunction [3, 6—10]. Impaired cognitive functions range from momentary postoperative confusion and delirium [8—10] to permanent brain damage [3, 6, 7]. In a few studies, postoperative patients were examined with neuropsychological tests [9, 10] and the role of

PATIENTS AND METHODS

Because- pulse oximetry was not widely used in Denmark at the time of the study, a controlled, randomized trial seemed to be an appropriate step in assessing the effect of this monitoring technique on anaesthesia care. The study was approved by the Regional Ethics Committee for Copenhagen and Esbjerg, and patients gave informed consent. The study was carried out during 15 consecutive months at two major Danish hospitals, one of which was a University hospital. The patients were 18 yr or older, admitted for elective surgery requiring general, spinal or extradural anaesthesia expected to last more than 20 min. We excluded outpatients and patients undergoing J. T . MOLLER*, M.D.; K. ESPERSEN, M.D.; S. H . JOHANSEN, M.D.;

Department of Anaesthesia, Herlev Hospital, University of Copenhagen, Denmark. I. SVENNILD, PH.D. (Department of Psychology); N. W. JOHANNESSEN, M.D., P. F. JENSEN, M.D.,

M. DjERNES, M.D. (Department of Anaesthesia); Esbjerg Central Hospital, Denmark. J. S. GRAVENSTEIN, M.D., DR.MED.H.C, De-

partment of Anesthesiology, Gainesville, University of Florida, U.S.A. J. B. COOPER, PH.D., Department of Anesthesiology, Massachusetts General Hospital, U.S.A. Aceepted for Publication: March 1, 1993. *Address for correspondence: Department of Anaesthesia, Rigshospitalet, University of Copenhagen, DK-2100 Copenhagen, Denmark.

Downloaded from http://bja.oxfordjournals.org/ at Georgetown University on September 6, 2014

In a randomized, blinded clinical study, we have used objective and subjective measures to determine if perioperative monitoring with pulse oximetry—by virtue of its potential to lessen hypoxaemia—would decrease late postoperative cognitive dysfunction. We investigated 736 adult patients undergoing elective procedures (other than cardiac, neurosurgical or for cancer) under regional or general anaesthesia, allocated randomly to undergo (group I) or not to undergo (group II) pulse oximetry monitoring in the operating theatre and recovery room. Cognitive function was evaluated using the Wechsler memory scale (WMS) and continuous reaction time (RT) test the day before surgery, and on the 7th day after operation or at discharge if that occurred before postoperative day 7. A questionnaire sent 6 weeks after surgery elicited patients' subjective perceptions regarding cognitive abilities. There were no significant differences between the two groups in either the total WMS score, the score for each WMS subtests or RT test. The questionnaire revealed that 7% in group I and 11% in group II believed cognitive abilities had decreased (ns). For the 40 patients whose WMS scores were 10 points less after than before operation, a follow-up study was undertaken 3 months after surgery. At that time, the median WMS score had returned to the preoperative value. We conclude that, for these 736 patients, subjective and objective measures did not indicate less postoperative cognitive impairment after perioperative monitoring with pulse oximetry. (Br. J.

hypoxaemia was considered, as mild to moderate hypoxaemia is associated with cognitive dysfunction [11-14]. Recent studies have suggested that hypoxaemia in the operating theatre and recovery room is common [15-17], and that monitoring with pulse oximetry would make possible early diagnosis and treatment and thus reduce markedly the incidence and severity of this condition [18-20]. We postulated that perioperative monitoring with pulse oximetry should reduce the incidence or severity of perioperative cognitive dysfunction. In our recent, randomized evaluation of pulse oximetry in 20802 patients, we found no association between the rate of postoperative neurological complications and the use of pulse oximetry [21, 22]. However, because we searched only for protracted confusion, stroke and coma, more subtle and diffuse cognitive effects would have been missed. In this randomized study using objective and subjective measures, we have assessed any possible relationship between the use of pulse oximetry and late postoperative cognitive dysfunction.

PULSE OXIMETRY AND COGNITIVE DYSFUNCTION

Cognitive testing

Cognitive testing consisted of administration of the Wechsler memory scale (WMS) [23] and continuous reaction time (RT) test the day before operation (WMS I and RT I) and, after operation, on the 7th postoperative day or on the day of discharge from the hospital if that occurred before postoperative day 7 (WMS II and RT II). The tests were administered by specially trained nurses who were unaware of the random allocation of pulse oximetry monitoring. The WMS requires approximately 45 min to administer, and consists of a series of eight subtests of attention and memory, including personal and current information, orientation, mental control, logical memory, memory span, visual reproduction and associated learning [23]. It is available in two versions; both are standardized and contain the same number and kinds of subtests. As a check on reliability of the standardization, we used the two versions in random order for each patient. The WMS is perhaps the most widely used and acknowledged test of memory function. The RT test requires approximately 8 min to complete and was administered with a computer (Olivetti M24) with specially designed software (developed by B. Fries, medical technician, T. Kann, medical engineer, I. Svennild, Ph.D., psychologist,

and J. T. Moller, M.D.). Patients were asked to respond to a series of 100 light flashes by pushing a button as quickly as possible with the dominant hand. The flashes occurred without warning every 1-8 s. A learning session using 25 light flashes familiarized the patient with the procedure before actual testing. Patients were asked also to evaluate subjectively if their cognitive abilities had changed since surgery: a short questionnaire sent by mail 6 weeks after surgery elicited details of mood, memory and the ability to concentrate. To study the influence of time on cognitive dysfunction following anaesthesia, we also studied a subsample of patients. All patients whose WMS score decreased by 10 points or more (total WMS score 93; median approximately 63) after operation were asked to participate in a new test 3 months later ("cases")- For each of these patients, a control patient was selected randomly from all study patients (monitored and not-monitored), without regard for WMS scores. For each patient, the version of WMS used before operation was used for the 3-month follow-up. Monitoring and anaesthesia

In group I, pulse oximetry was begun just before induction of anaesthesia and continued until discharge of the patient from the recovery room. Two different pulse oximeters were used: an Ohmeda 3700 at one hospital and a Radiometer OXI at the other. A finger probe was used in all instances. All personnel were instructed to maintain oxyhaemoglobin saturation (Sp02) equal or greater than 93 % in group I. Any reading less than this required intervention. Exceptions to this policy could be sanctioned only by the anaesthetist in charge of the individual patient. In group II, patients were not monitored by pulse oximetry at any time. Other operating theatre monitoring consisted of continuous electrocardiography, measurement of arterial pressure at least every 5 min and, when warranted by the type of surgery or the patient's physical condition, measurement of central venous pressure. In the recovery room, arterial pressure, heart rate and ventilatory frequency were measured at least every 15 min; when indicated, electrocardiogram and central venous pressure were monitored continuously. The preoperative visit, premedication, anaesthetic regimen, postoperative observation and other anaesthesia- and recovery-related care followed the routine guidelines of each hospital. Each patient's physician, unaware of the patient's possible allocation to oximetry monitoring, planned the

TABLE I. Patient data and duration of surgery. Group I : Monitored with pulse oximeter; group II: no pulse oximeter monitoring. *Significantly uneven distribution (P < 0.05)

Group II (n = 378)

Group I (n = 358)

Age (yr) Weight (kg) Height (cm) Duration of surgery (min)

Mean

Median

52.2 74.2 169.6 88.0

56 73* 168 90

Range

Mean

Median

18-81 40-136 147-193 20-320

52.3 71.6 169.4 89.0

55 71* 169 88

Range 18-82 45-115 135-192 20-^30

Downloaded from http://bja.oxfordjournals.org/ at Georgetown University on September 6, 2014

neurological, cardiac or cancer operations. Patients with suspected or confirmed cancer were excluded because a pilot study of 100 subjects indicated a confounding psychological difference between patients with cancer and those without cancer [Moller JT, unpublished data, 1988]. The need for reoperation was another exclusion criterion. Patients were allocated to an operating theatre the day before surgery according to the department's routine, after which time pulse oximeters were allocated randomly to 50% of those theatres. A sealed envelope contained the daily random assignment and was not opened until the patients had been allocated to an operating theatre. The random assignment of an individual patient to receive monitoring with pulse oximetry (group I) or not (group II) could not be changed by moving the patient to another OT. Every day, as many as four patients from each surgical department were chosen randomly to undergo cognitive testing. The patients selected for study alternated between the surgical specialties according to a predetermined rotation schedule known only by the investigators. The physicians allocating patients to the operating theatres did not know which patients would undergo cognitive testing.

341

BRITISH JOURNAL OF ANAESTHESIA

342 TABLE II. Major patient-related risk factors, type of anaesthesia and surgical procedures. Group I: Monitored with pulse oximeter; group II: no pulse oximeter monitoring. No significant difference between groups

Group I (» = 358) No.

%

Group II (» = 378) No.

% 0

Risk factors 62.3 31.0 6.4 0.3 9.5 9.9

249 110 19 0 24 28

65.9 29.1 5.0 0.0 6.5 7.5

O

27.3 23.1 45.5

26 15 129 172 15

7.3 4.2

3.9

36.0 48.0 4.2

103 86 172 17

27.3 22.8 45.5

33 16 128 186 16

8.7 4.2

I

—• O

O

I O

I

I

— ' i n

1 d o

1

14 o

98 83 163 14

I

1

0? in o o o

:o8) to 8) -21 .5 to 16.51

223 111 23 1 33 35

I

in oo —

1 1 1 1 11

o o — o o —• o I -

4.5

11

33.9 49.2 4.2

anaesthetic regimen as usual. For premedication, patients received either a benzodiazepine orally or an opioid i.m. before induction of anaesthesia. General anaesthesia was classified as either inhalation or i.v., based on the primary maintenance agent used. The lungs of all patients were ventilated with an inspired oxygen concentration of at least 30%. When applicable, residual neuromuscular block was antagonized with neostigmine 2.5 mg and atropine 1 mg i.v. Regional anaesthesia was either extradural or spinal block. Supplementary oxygen at a minimum flow of 3 litre min"1 was administered to all patients during regional anaesthesia and to all patients after regional and general anaesthesia upon admission to the recovery room, where analgesia consisted of i.v., i.m. or extradural administration of morphine.

i ^

^

•gco

P XJ.122 m r- in o ^^ in in

o ~ ^ -^ o

•S.8

^4

in

^/>,7

"^ ^ 1

4

—

IA

VO

op

9^"

^ -• M 1 • I I I r*-

en CJ o M -J tn in

in "t o

Data collection and analysis

Because information is limited on cognitive dysfunction beyond 4 days after routine anaesthesia, it was difficult to determine the precise number of patients needed for study. A sample size of 750 patients was estimated to give a minimum power of 90% (type II error < 0.10) to detect a difference of total WMS score of 3 points (5 %) between the two groups. We used specially designed forms to collect data. One form was used to record descriptive information copied from medical and preanaesthetic evaluation records, such as a history of heart disease, hypertension or lung disease. We also noted details of the anaesthetic technique, the patient's recovery and postoperative complications. All personnel were instructed to register events with hypoxaemia and changes in patient care as a consequence of the oximeter monitoring. Hypoxaemia was denned as 5po2 sg 90 % or cyanosis, PaOs sj 7.8 kPa or SaO2 ^ 90%. WMS results were recorded on a special WMS form; RT test results were stored on com-

c^i ~-> r n

>o

op

i vo (n 4 co int~-oocscoo

C 2 M u

.a •3

X) 3

6 6 c c u o c 8 B S 2 P 2 ffl'3

§1 8 a2 1 co

- 1 1 1>

«

73 y

<

h

Downloaded from http://bja.oxfordjournals.org/ at Georgetown University on September 6, 2014

Heart disease Lung disease Anaesthesia General anaesthesia Potent inhalation I.v. Regional GA + regional Surgical procedure Eye and ENT Upper abdominal Lower abdominal Procedures on extremities Other procedures

" —

I

:o 1)

ASA I II III IV

S 2 S S S " —• • * in in ON —

PULSE OXIMETRY AND COGNITIVE DYSFUNCTION

343

TABLE IV. Scores for the Continuous Reaction Time (RT) test for 358 patients who had (group I), and 378 patients who did not have (group II), perioperative monitoring with pulse oximetry. *Difference calculated for each patient. None of the differences between RT I and RT II or between the two groups was significant

Preoperative score (ms) (RT I)

Postoperative score (ms) (RT II)

Difference (ms) ( R T I I - R T I)*

Group I

Group II

Group I

Group II

Group I

Group II

232 270 363 296

228 273 356 292

239 283 376 307

234 277 365 301

6 8 12 12

6 9 11 8

10% fractile Median 90 % fractile Mean

lOO-i 90 80 -

6050 403020-

/ /

100-25

-20

-15

-10

-5 0 5 WMS difference

10

15

20

25

FIG. 1. Distribution of differences between preoperative and postoperative total scores on the Wechsler Memory Scale (WMS) for 358 patients who were monitored with pulse oximetry (solid line) and for 378 patients who were not (dashed line).

puter. All data were entered into a mainframe computer (IBM 4361) with specially designed management software that constructed an event history data base. The computer software used for statistical analysis included SCIBAS [24] and BMDP (BMDP Statistical Software, Inc., Los Angeles, CA). Patient data, events with hypoxaemia, changes in patient care and results from the questionnaire underwent chi-square testing or Fisher's exact test. WMS and RT data were subjected first to the Mann-Whitney test and second, to stratification by age, weight, ASA physical status and type of anaesthesia. P < 0.05 was considered statistically significant. RESULTS

Initially, we enrolled 861 patients in the study for preoperative cognitive testing (WMS I and RT I). Of these, 736 (86%) completed the postoperative testing (WMS II and RT II). The other 125 patients (14%) did not complete the study: five needed to undergo a second operation, 63 refused to participate after the initial testing and 57 were discharged from hospital before postoperative testing could be administered. These 125 eliminated patients were distributed equally between the two groups and did not differ significantly from the 736 study participants in patient characteristics—risk factors (including ASA physical status), type of anaesthesia or surgical procedure and frequency of postoperative complications (10% for the total population vs 11 % for the

eliminated group (ns)). In the subsample (the casecontrol part of the study), 23 patients did not return for the 3-month follow-up test, 14 being in the case group and nine in the control group. The study population comprised 410 women and 326 men, distributed equally between the two study groups. Although patients in group I (with oximetry) (n = 358) were significantly heavier than group II (without oximetry) (n = 378) (table I), there were no other differences in risk factors or type of anaesthesia or surgical procedure (table II). During anaesthesia, hypoxaemia was diagnosed in 7.8% of the patients in group I (with oximetry) and in 0.3% in group II (without oximetry) (P < 0.00005). In the recovery room, 11.7 % of patients in group I and 0.5% of those in group II were diagnosed as having experienced hypoxaemia (P < 0.00005). The staff stated that in 24.9% of the patients the readings from the oximeter prompted one or more changes in treatment. The flow rate of supplementary oxygen given in the recovery room was significantly greater in group I, with 45 % of the oximeter monitored patients receiving more than 3 litre min"1 of oxygen, compared with 35 % of patients in group II (P < 0.01). The proportion of patients discharged from the recovery room with a request for supplementary oxygen was 13% in group I and 3 % in group II (P < 0.00005). Preoperative scores on the WMS I and RT I did not differ between groups (tables III, IV). Administration of the postoperative cognitive test occurred

Downloaded from http://bja.oxfordjournals.org/ at Georgetown University on September 6, 2014

Patients (%)

70-

BRITISH JOURNAL OF ANAESTHESIA

344

TABLE V. Subjective reports {by questionnaire) of cognitive deficits. Group I: Monitored with pulse oximeter; group II: no pulse oximeter monitoring. In addition to the listed questions, the patients were asked about former problems with memory, concentration and irritability Group I (« = 345) No.

Main questions After discharge from hospital and returning to your home have you experienced problems with recent memory? (i.e. what to buy when shopping, telephone numbers, appointments) After discharge from hospital and returning to your home have you experienced problems with ability to concentrate? (i.e. when reading your newspaper, attention to the news in the television) After discharge from hospital and returning to your home have you experienced problems with irritability? (chiding about nothing, fly into temper for nothing)

WMS II

WMS III

5 days after anaesthesia (median value; mean 5.5 days, range 2-16 days); this value did not differ between groups. For both the total score and the scores for each subtest, the median difference between preoperative (WMS I) and postoperative (WMS II) memory test results did not differ significantly between group I and group II (table III); however, the distribution of the individual differences in total scores on the two WMS tests varied widely in both study groups (fig. 1). For 17% of the patients, total WMS II scores were either 10 points greater or 10 points less than the WMS I score—a percentage that did not differ significantly between the two groups (fig. 1). The continuous reaction times tested after operation were of greater duration by a few milliseconds than before operation, but these differences were not statistically significant, nor were there significant differences between group I and group II (table IV). Strati-

No.

%

P

24

7.0

41

11.0

0.07

36

10.4

35

9.4

0.60

27

7.8

42

11.2

0.12

fication of data by age, weight, ASA physical status and type of anaesthesia did not reveal any statistically significant difference in cognitive function between groups. The questionnaire was returned by 97.7 % of the study population 45 days after anaesthesia (median value; mean 50 days, range 40-106 days) (table V). For all patients, including the patients in the case-control part of the study, subjective perception of cognitive dysfunction did not correlate with WMS or RT test scores. In the subsample (the case-control part of the study), 40 cases and 45 controls were tested 96 days after anaesthesia (median value; mean 97 days, range 65-140 days). There were no significant differences in the case group's characteristics concerning monitoring with or without pulse oximetry, age and type of anaesthesia, compared to the main study group (736 patients). The case group's total median WMS score on the 3-month follow-up test returned to the preoperative value. The interindividual difference for the case group was within the range of the interindividual difference for the control group (fig- 2). DISCUSSION

In the past 30 years, several investigators have examined cognitive function after anaesthesia and surgery [1-3,25-30]. More recent studies have concentrated on differences between the effects of general anaesthesia and those of regional anaesthesia [8,31-37] and neuropsychological alteration after cardiac operation [28,38,39]. Some investigators have concluded that anaesthetic technique has no major influence on cognitive function beyond the first 3-A days after operation [29-31]. Our results suggest that use of pulse oximetry monitoring during anaesthesia and in the recovery room does not affect cognitive outcome. Catastrophic hypoxaemia is a well known cause of brain dysfunction [6, 7, 40]. However, several neurological studies making use of the WMS tests have demonstrated that mild to moderate hypoxaemia is also associated with cognitive dysfunctions [11-14,41]. Did the patients in the present study experience hypoxaemia? Paradoxically, it seems as if

Downloaded from http://bja.oxfordjournals.org/ at Georgetown University on September 6, 2014

WMS

FIG. 2. Median change (ranges) in Wechsler Memory Scale (WMS) test scores for patients with suspected postoperative cognitive dysfunction ("cases") (x, solid line) and for control patients ( • , dashed line). "Cases" designates patients whose WMS scores 5 days (median values) after surgery were 10 points less than preoperative scores. Control patients were selected randomly from the remaining patients and their data represent the normal postoperative variation in cognitive scores. Change in WMS score was measured at three times: before surgery (base line, WMS I); 5 days (median value) after surgery (WMS II); and 96 days (median value) after surgery (WMS III).

%

Group II (n = 374)

PULSE OXIMETRY AND COGNITIVE DYSFUNCTION

less than 10 % if an increase of 3 points (5 %) in the total WMS score of the pulse oximetry group had been achieved. This agrees well with the required minimum statistical power of 90 % established in the design of the trial. To test the hypothesis that noncatastrophic perioperative hypoxaemia affects the patient's subjective perception of mental function, a larger group of patients than selected by us would be required, or we would need to apply objective tests to patients at special risk of hypoxaemia-induced cognitive dysfunction. Patients with systemic diseases (ASA physical status III or IV) or elderly patients already suffering cognitive dysfunction may be more sensitive to the putative effects of noncatastrophic hypoxaemia. The fact that stratification of data by age and ASA physical status did not reveal any difference between groups does not argue against this hypothesis, as only 7 % of our study patients were classified ASA III or IV (table II) and only 43 % of the patients were older than 60 yr. The scores of individual patients varied widely between preoperative and postoperative tests, this tendency being most evident with the WMS (fig. 1). The decrease in the total postoperative WMS scores of more than 50 % of the patients was nearly equalled by the increase in scores of the others. Normal biological variation and many confounding factors may influence the results of postoperative cognitive tests. An increase in scores could have resulted from previous exposure to a given test and from the recovery from illness (the patient would be free of pain or less anxious). A decrease in scores could result from a residual or direct effect of anaesthesia; from sedation, intraoperative hyperventilation or analgesics; or from surgery itself and the hormonal or metabolic sequelae [3,26,27,30,37,44]. To reduce the influence of these confounding factors, our postoperative cognitive tests were administered at the day of discharge from hospital or at the latest on the 7th day after operation. The WMS was chosen to reduce the effects of practice, as this test exists in well standardized parallel forms. Furthermore, before the first RT test, a learning session familiarized the patient with the procedure. Finally, we excluded patients with suspected or confirmed cancer as a consequence of our own pilot study and other previous reports [45, 46] suggesting that cancer itself is associated with cognitive disorders in addition to well known affective changes such as depression. Group analysis of the results may conceal important individual variation in the case-control part of the study [47]. We followed the patients who had the greatest decrease in postoperative test scores and included a control group to document normal postoperative variation in cognitive scores (fig. 2). Three months after anaesthesia and surgery, the case group's median WMS score had returned to the preoperative value. This observation is supported by results from other recent studies [37-39, 44]. However, a practice effect may have influenced the follow-up test. The fact that the WMS test exists in only two parallel forms obliged us to reuse the preoperative WMS form at the follow-up. We believe this bias did not influence the retest which

Downloaded from http://bja.oxfordjournals.org/ at Georgetown University on September 6, 2014

the patients in the pulse oximetry group suffered more hypoxaemia than the control group, in both the operating theatre and the recovery room. However, in reality hypoxaemia was discovered much more frequently in the pulse oximetry group and was treated, while in the control group it was infrequently detected and treated. From our previous blinded observational studies, we know that 75 % and 95 % of hypoxaemic episodes in the operating theatre and recovery room remain undiscovered [15, 17]. In the operating theatre 53% and in the recovery room 55 % of these patients had one or more episodes of mild hypoxaemia with oxyhaemoglobin saturation (SpOi) between 86% and 90%. Severe hypoxaemia with SpO2 less than 81 % was observed in 20 % of the patients in the operating theatre and 13 % in recovery [15, 17]. Findings of other studies during anaesthesia and postoperative recovery support our results [16, 18, 20]. In a recent, randomized study we have demonstrated that monitoring with pulse oximetry could substantially reduce the incidence, duration and severity of hypoxaemia during and after anaesthesia [19]. In this study, the incidence of hypoxaemia (5pOz 81-85%) in patients monitored with oximetry was three to five times less than in the control group. In the control group 7% of the patients in the recovery room experienced SpO2 < 80%, while none had such decreased values in the oximeter group [19]. Similar observations were reported in other studies [18, 20]. In the present study and in our previous reports, oximetry monitoring resulted in several changes of patient care [21, 22]. In group I (with oximetry) flow rates of supplementary oxygen in the recovery room were increased significantly and significantly more patients were discharged from recovery receiving supplementary oxygen compared with group II (without oximetry). Thus ample evidence points to a substantial difference in the incidence and severity of hypoxaemia in patients monitored with pulse oximetry compared with patients not monitored in the operating theatre and recovery room. In our study, this was not reflected in differences in the results of objective psychological tests completed by our patients. Perhaps the postulated effect of oximetry and treatment with oxygen was confounded by the well known fact that hypoxaemia is a common occurrence several days after operation—that is, at a time when none of our patients was monitored with pulse oximetry [42, 43]. In the present study, the difference between the two study groups at 6 weeks after anaesthesia regarding subjective perception of a cognitive dysfunction (problems with recent memory) approached, but did not reach, statistical significance (P = 0.07) (table V). If this difference did not arise by chance, it could represent a type II error because of inclusion of too few patients in the study. On the basis of the responses to the questionnaire, our probability of having a type II error is 53%. It is impossible to determine the probability of a type II error (P) for the objective test (the WMS and the RT test) as there was no meaningful numerical difference between the two groups. However, with 736 patients in the study, the probability of a type II error was

345

BRITISH JOURNAL OF ANAESTHESIA

346

of this study: T. Pedersen, B. Chraemmer-Jorgensen, L. Heslet, F. Wiberg-Jergensen, B. D. Pedersen, L. S. Rasmussen, O. Ravlo, N. H. Rasmussen, T. Kann, T. Jorgensen, and B. Daneskjold Samsoe. Finally, we thank M. Davidsen, M.sc, and E. Henriksen, industrial engineer (Department of Medical Statistics) for statistical assistance, and the Ohmeda Company (Louisville, CO) and Radiometer (Copenhagen, Denmark) for providing pulse oximeters. Supported by grants from the Gangsted Foundation; the King Christian Xth Foundation; the Skt. Lukas Stiftelsen's Foundation; The Research Foundation for Ringkebing, Ribe, and Sonderjylland Counties; the S. & W. Foundation; The Danish Society of Anaesthesiology Foundation; the Lilly Benthine Lund Foundation; the Lundbeck Foundation; the Jacob and Olga Madsen Foundation; the Svend and Ina Hansen Foundation; the Danish Medical Research Council; and the Anesthesia Patient Safety Foundation (U.S.A.). REFERENCES 1. Bedford PD. Adverse cerebral effects of anaesthesia on old people. Lancet 1955; 2: 257-263. 2. Simpson BR, Williams M, Scott JF, Smith AC. The effects of anaesthesia and elective surgery on old people. Lancet 1961;2: 887-893. 3. Drummond GB. The assessment of postoperative mental function. British Journal of Anaesthesia 1975; 47: 130-142. 4. Knill RL, Novick TY, Skinner MI. Idiopathic postoperative delirium associated with longterm cognitive impairment. Canadian Journal of Anaesthesia 1991; 38: A54. 5. Chung F. Postoperative cerebral dysfunction (delirium). In: McLesky C, ed. Geriatric Anaesthesiology. Baltimore: Williams and Wilkins, 1993 (in press). 6. Adams JH. Hypoxic brain damage. British Journal of Anaesthesia 1975; 47: 121-129. 7. Hardy CA, Fischbach HP. Delayed postanoxic encephalopathy. Anesthesiology 1975; 43: 694-695. 8. Hole A, Terjsesen T, Breivik H. Epidural versus general anaesthesia for total hip arthroplasty in elderly patients. Ada Anaesthesiologica Scandinavica 1980; 24: 279-287. 9. Berggren D, Gustafson Y, Eriksson B, Bucht G, Hansson L-I, Reiz S, Winblad B. Postoperative confusion after anesthesia in elderly patients with femoral neck fractures. Anesthesia and Analgesia 1987; 66: 497-504. 10. Rosenberg J, Kehlet H. Postoperative mental confusion associated with postoperative hypoxemia. Anesthesiology 1992; 77: A215. 11. Prigatano GP, Parsons O, Wright E, Levin DC, Hawryluk G. Neuropsychological test performance in mildly hypoxemic patients with chronic obstructive pulmonary disease. Journal of Consulting and Clinical Psychology 1983; 51: 108-116. 12. Vaernes RJ, Owe JO, Myking O. Central nervous reactions to a 6.5-hour altitude exposure at 3048 meters. Aviation, Space, and Environmental Medicine 1984; 55: 921-926. 13. Berry DT, Webb WB, Block AJ, Bauer RM, Switzer DA. Nocturnal hypoxia and neuropsychological variables. Journal of Clinical and Experimental

14. 15. 16. 17.

ACKNOWLEDGMENTS We thank the following nurses who administered the psychological tests: R.Jensen, S. Bang, B. Sieben, R. Madsen, S. Rasmussen, E. Bundgaard, H. Sarensen, H. Andersen, K. Rude, K. R. Senderborg, and A. Carstensen. We also thank Ms M. Dragsted and Ms N. Meyer for accurate and careful entering of data, and Ms G. Blom for secretarial assistance. We are grateful to the other members of the Danish Pulse Oximetry Study Group for their support and participation in the planning

18. 19.

20.

Neuropsychology

1986; 8:

229-238. Gibson GE, Pulsinelli W, Blass JP, Duffy TE. Brain dysfunction in mild to moderate hypoxia. American Journal of Medicine 1981; 70: 1247-1254. Moller JT, Wittrup M, Johansen SH. Hypoxemia in the postanesthesia care unit: An observer study. Anesthesiology 1990; 73: 890-895. Brown LT, Purcell GJ, Traugott FM. Hypoxaemia during postoperative recovery using continuous pulse oximetry. Anaesthesia and Intensive Care 1990; 18: 509-516. Moller JT, Johannessen NW, Berg H, Espersen K, Larsen LE. Hypoxaemia during anaesthesia—an observer study. British Journal of Anaesthesia 1991; 66: 437^44. Cote CJ, Goldstein EA, Cote MA, Hoaglin DC, Ryan JF. A single-blind study of pulse oximetry in children. Anesthesiology 1988; 68: 184-188. Moller JT, Jensen PF, Johannessen NW, Espersen K. Hypoxaemia is reduced by pulse oximetry monitoring in the operating theatre and in the recovery room. British Journal of Anaesthesia 1992; 68: 146-150. Cote CJ, Rolf N, Liu LMP, Goudsouzian NG, Ryan JF, Zaslavsky A, Gore R, Todres D, Vassallo S, Polaner D,

Downloaded from http://bja.oxfordjournals.org/ at Georgetown University on September 6, 2014

occurred after 3 months. This is supported by the fact that the control group experienced only a slight and insignificant improvement in test score at the follow-up test (WMS III) (fig. 2). One may speculate that application of a broader range of neuropsychological assessments than used in the present study could have detected varying deficits of an enduring nature. Using a broad range of tests, several investigators have described moderate to severe cognitive dysfunction lasting for several months after coronary bypass surgery [28, 38, 39]. Interestingly, several of these cognitive dysfunctions were detected with the WMS test [28,38]. Although 9% of our total population thought that mental functioning had deteriorated, this perceived deficit was not confirmed by objective tests. Another study reported similar results (a perceived but unconfirmed cognitive deficit) 3 months after anaesthesia [44]. This finding suggests that some of the widely used psychological tests may not be sensitive enough to detect changes in cognitive function reported by a patient. The drop-out rate of subjects in our study (14%) is a cause of concern, especially if the loss were caused by postoperative complications that could influence cognitive function. However, analysis of the patient data, risk factors, types of anaesthesia and surgical procedures, and preoperative cognitive test results revealed no significant difference between the study population and the patients who were eliminated from study. Most importantly, the incidence of postoperative complications was the same for the two groups, suggesting that loss of subjects did not bias our results. We believe the present study raises several questions that future studies should examine. What caused 9% of our patients to perceive that their memories were impaired 6 weeks after anaesthesia and operation? Is it possible that hypoxemia, which occurs commonly during the first days after operation [42,43], played a role in the cognitive impairment reported by our patients ? Future studies must assess the possible contribution of postoperative hypoxaemia to cognitive dysfunction. May some other objective tests of cognition confirm the subjective impressions reported by our patients? Should future studies concentrate on elderly patients sensitive to cognitive impairment and undergoing major operations under general anaesthesia? In summary, for our 736 patients, subjective and objective measures did not indicate less late postoperative cognitive impairment after perioperative monitoring with pulse oximetry, and the 3-month follow-up test did not produce any objective signs of major cognitive dysfunction in any patient.

PULSE OXIMETRY AND COGNITIVE DYSFUNCTION

21.

22.

23. 24.

26. 27

28.

29. 30. 31.

32.

33. 34.

35. 36.

37.

38.

39. 40. 41. 42. 43. 44.

45. 46. 47.

surgery during spinal and general anaesthesia. Anaesthesia 1985; 40: 672-676, Chung F, Meier R, Lautenschlager E, Carmichael FJ, Chung A. General or spinal anesthesia: Which is better in the elderly? Anesthesiology 1987; 67: 422-427. Ghoneim MM, Hinrichs JV, O'Hara MW, Mehta MP, Pathak D, Kumar V, Clark CR. Comparison of psychologic and cognitive functions after general or regional anesthesia. Anesthesiology 1988; 69: 507-515. Hughes D, Bowes JB, Brown MW. Changes in memory following general or spinal anaesthesia for hip arthroplasty. Anaesthesia 1988; 43: 114-117. Asbjern J, Jakobsen BW, Pilegaard HK, Blom L, 0stergaard A, Brandt MR. Mental function in elderly men after surgery during epidural analgesia. Ada Anaesthesiologica Scandinavica 1989; 33: 369-373. Nielson WR, Gelb AW, Casey JE, Penny FJ, Merchant RN, Manninen PH. Long-term cognitive and social sequelae of general versus regional anesthesia during arthroplasty in the elderly. Anesthesiology 1990; 72: 1103-1109. Shaw PJ, Bates D, Cartlidge NEF, French JM, Heaviside D, Julian DG, Shaw DA. Neurologic and neuropsychological morbidity following major surgery: Comparison of coronary artery bypass and peripheral vascular surgery. Stroke 1987; 18: 700-707. Hammeke TA, Hastings JE. Neuropsychologic alterations after cardiac operation. Journal of Thoracic and Cardiovascular Surgery 1988; 96: 326-331. Volpe BT, Hirst W. The characterization of an amnesic syndrome following hypoxic ischemic injury. Archives of Neurology 1983; 40: 436-440. West JB. Human psychology at extreme altitudes on Mount Everest. Science 1984; 223: 784-788. Marshall BE, Wyche MQ. Hypoxemia during and after anesthesia. Anesthesiology 1972; 37: 178-209. Hanning CD. Prolonged postoperative oxygen therapy. British Journal of Anaesthesia 1992; 69: 115-116. Jones MJT, Piggott SE, Vaughan RS, Bayer AJ, Newcombe RG, Twining TC, Pathy J, Rosen M. Cognitive and functional competence after anaesthesia in patients aged over 60: Controlled trial of general and regional anaesthesia for elective hip or knee replacement. British Medical Journal 1990; 300: 1683-1687. Wood MM, Cousins MJ. Iatrogenic neurotoxicity in cancer patients. Pain 1989; 39: 1-3. Massie MJ, Holland JC. The cancer patient with pain: Psychiatric complications and their management. Medical Clinics of North America 1987; 71: 243-258. Hickey S, Asbury AJ, Millar K. Psychomotor recovery after outpatient anaesthesia: Individual impairment may be masked by group analysis. British Journal of Anaesthesia 1991; 66: 345-352.

Downloaded from http://bja.oxfordjournals.org/ at Georgetown University on September 6, 2014

25.

Alifimoff JK. A single-blind study of combined pulse oximetry and capnography in children. Anesthesiology 1991; 74: 980-987. Moller JT, Pedersen T, Rasmussen LS, Jensen PF, Pedersen BD, Ravlo O, Rasmussen NH, Espersen K, Johannessen NW, Cooper JB, Gravenstein JS, Chraemmer-Jorgensen B, Wiberg-Jorgensen F, Djernes M, Heslet L, Johansen SH. Randomized evaluation of pulse oximetry in 20,802 patients. I. Design, demography, pulse oximetry failure and overall complication rate. Anesthesiology 1993; 78: 436—444. Moller JT, Johannessen NH, Espersen K, Ravlo O, Pedersen BD, Jensen PF, Rasmussen NH, Rasmussen LS, Pedersen T, Cooper JB, Gravenstein JS, Chraemmer-Jorgensen B, Djernes M, Wiberg-Jergensen F, Heslet L, Johansen SH. Randomized evaluation of pulse oximetry in 20,802 patients. II. Perioperative events and postoperative complications. Anesthesiology 1993; 78: 445-453. Wechsler D. A standardized memory scale for clinical use. Journal of Psychology 1945; 19: 87-95. Mortensen SL, Larsen ME. A data care system for scientific purposes. In: Barra JR, ed. Recent Developments in Statistics. Proceedings of the European Meeting of Statisticians, Grenoble, September 6—11, 1976. Amsterdam: NorthHolland Publ. Co., 1977; 705-707. Gruvstad M, Kebbon L, Lof BA. Changes in mental functions after induced hypotension. Ada Psychiatrica Scandinavica 1962; 37 (Suppl.): 163. Davison LA, Steinhelber JC, Eger El n, Stevens WC. Psychological effects of halothane and isoflurane anesthesia. Anesthesiology 1975; 43: 313-324. Thompson GE, Miller RD, Stevens WC, Murray WR. Hypotensive anesthesia for total hip arthroplasty: A study of blood loss and organ function (brain, heart, liver and kidney). Anesthesiology 1978; 48: 91-96. Townes BD, Bashein G, Hornbein TF, Coppel DB, Goldstein DE, Davis KB, Nessly ML, Bledsoe SW, Veith RC, Ivey TD, Cohen MA. Neurobehavioral outcomes in cardiac operations. A prospective controlled study. Journal of Thoracic and Cardiovascular Surgery 1989; 98: 774—782. Chung F, Seyone C, Dyck B, Chung A, Ong D, Taylor A, Stone R. Age-related cognitive recovery after general anesthesia. Anesthesia and Analgesia 1990; 71: 217-224. Smith C, Carter M, Sebel P, Yate P. Mental function after general anaesthesia for transurethral procedures. British Journal of Anaesthesia 1991; 67: 262-268. Riis J, Lomholt B, Haxholdt O, Kehlet H, Valentine N, Danielsen U, Dyrberg V. Immediate and long-term mental recovery from general versus epidural anaesthesia in elderly patients. Ada Anaesthesiologica Scandinavica 1983; 27: 44-49. Bigler D, Adelhej B, Petring OU, Pederson NO, Busch P, Kalhke P. Mental function and morbidity after acute hip

347