Circulation 60,475 • Volume 13, No. 1 • Spring 1998

Lancet Article Addresses a Different Type of Safety Question; Elderly Suffer Prolonged Postoperative Cognitive Dysfunction

J.S. Gravenstein, M.D.

A long-standing interest of the Anesthesia Patient Safety Foundation has reached a significant milestone with the publication of an important article on prolonged postoperative cognitive dysfunction in the elderly.1 It was confirmed that 10% of elderly patients (7% age 60-70 and 14% of those over age 70) who receive routine anesthetics, without hypoxemia or hypotension, demonstrate persistent decreased cognitive function three months postoperatively.

Concerns that anesthesia potentially may have deleterious effects on CNS function and, specifically, intellectual abilities have existed for some time. The story that led to this particular just-published investigation of this subject starts with the large Danish study of pulse oximetry in which Moller and his collaborators had set out to test the hypothesis that undetected hypoxemia in the perioperative period carries a measurable risk and that the use of pulse oximetry would facilitate the timely detection and treatment of hypoxemia.2,3 That study was conceived because of the concerns of anesthesiologists in general, and members of the APSF in particular, about the clinical consequences of hypoxemia. There was enough clinical experience to support the belief that pulse oximetry had an important role to play in the prevention of disasters linked to inadequate ventilation. Through the organizational efforts of the APSF, several companies made over 50 pulse oximeters available and, with great luck, the project found a capable leader in Dr. Jakob Moller who organized and conducted the actual investigation. Eventually, some 20,000 patients were enrolled in the prospective study of the impact of pulse oximetry on outcome in surgical patients. Pulse oximetry was not shown statistically to affect the length of stay or perioperative mortality, even though many instances of hypoxemia were discovered and treated in the patients monitored. The authors concluded that far, far more patients would have to be studied in order to demonstrate an effect on mortality. There were indications that morbidity may have been reduced because patients who were monitored with pulse oximetry were less likely to suffer evidence of myocardial ischemia as compared to those not so monitored, presumably because the monitored patients were treated for hypoxemia in contrast to their unmonitored control counterparts.

The most feared complication of anesthesia disasters is hypoxic brain damage. While there was no evidence of significant, obvious brain damage among the 20,000 Danish patients, the possibility was considered that some might have suffered minor cerebral impairment secondary to hypoxemia. The investigators, therefore, examined a subset of 736 patients both preoperatively and postoperatively with psychologic tests designed to detect subtle changes in intellectual function. Half of these patients had been in the oximetry group, half not.4 No difference in psychologic function was detected with these particular tests after anesthesia compared to before anesthesia, whether the patient was monitored with pulse oximetry or not. A questionnaire mailed to the patients six weeks after their surgeries inquired about subjective symptoms. Nearly 10% of these patients complained that their memory was not as good as it had been preoperatively. This complaint was slightly more common in patients not monitored with pulse oximetry than in those monitored. This observation supported anecdotes of patients who complained that after general anesthesia, for weeks or months they noticed an impaired recent memory.

Dose-Response for Hypoxemia?

While these data of self-reported decrements in memory after anesthesia were tantalizing, they were inadequate. Critics pointed out that better psychologic tests might have yielded different results, that the number of patients studied was too small, that there was no control group of volunteers who took the same tests at the same intervals but without undergoing a surgical procedure under anesthesia, and finally, that self-reported psychologic and intellectual problems are difficult to interpret. The concerns of the investigators about hypoxemic damage to the brain, however, were not assuaged by these studies. Some argued that if severe hypoxemia ravages the brain, then mild hypoxemia should cause milder damage to cerebral function, in other words, there should be a dose-response curve for hypoxic brain damage. The concept of such a potential dose response curve fueled the drive to launch a large, prospective study of patients even though several recent studies had failed to discover prolonged cognitive changes after anesthesia.

At the World Congress of the World Federated Societies of Anesthesiologists in the Netherlands in 1992, a small group of interested investigators representing several countries discussed the issue and decided to launch a prospective study of prolonged cognitive dysfunction. Only elderly patients (60 years and older) were to be examined and the patients were to have at least two hours of general anesthesia for surgery other than that involving brain and heart. Two international experts of cognitive function in the elderly (Drs. Rabbitt and Jolles) were recruited to design the “mother of all psychologic tests” for this purpose and Dr. Jakob Moller, who had distinguished himself as the organizer of the Danish pulse oximetry study, was persuaded to lead the effort. An elaborate organization was established and funding was secured. Again, the Anesthesia Patient Safety Foundation lent crucial support to the effort which eventually involved institutions in Denmark, France, Germany, Great Britain, Greece, Spain, The Netherlands, and the United States. The bulk of the funding was provided by the European Commission and regional foundations and charitable funds. Industry helped significantly by providing financial help (Datex) and many of the monitors (Criticon, Datascope, Nellcor, and Protocol). The psychologic tests were designed to be culturally neutral, and all countries collected appropriately matched control data from elderly volunteers who did not undergo surgery or anesthesia.

Prolonged (3-Month) CNS Deficit Found “Disturbing”

Of 1,218 patients enrolled, 947 completed the study which included a battery of psychologic tests preoperatively, about one week after the operation, and again about 3 months postoperatively. That approximately 25% of the patients had measurable cognitive dysfunction a week after their operation was not unexpected. Lingering drug effects, hormonal and nutritional influences, and the strange hospital environment can well be imagined to conspire to cloud the mind of some elderly patients within this time frame. That close to 10 % had such dysfunction 3 months postoperatively, however, was disturbing. This compared to an incidence of about 3% having cognitive deterioration among 321 healthy control subjects in the same age range who were tested at the same intervals but who had neither anesthesia nor surgery. The differences between the controls and the patients were highly statistically significant. Of the many factors examined in a search for a mechanism potentially responsible for the prolonged postoperative cognitive dysfunction, hypoxemia and hypotension were singled out for special attention. Hypoxemia as measured by continuous pulse oximetry was defined as SpO2 < 80% for >2 min, and was assessed during the operation, in the recovery room, and for the first three postoperative nights. Blood pressure was monitored by oscillometric sphygmomanometry during anesthesia, in the recovery room and every 30 minutes during the first postoperative night. Hypotension was defined as a systolic blood pressure < 60% of control lasting for >5 minutes. A notable result was that neither hypoxemia nor hypotension was correlated with the occurrence of prolonged cognitive dysfunction. Of course, this does not mean that hypoxemia and hypotension should be ignored; even values less extreme than the thresholds used in the study may cause damage in a susceptible patient. Age was a prominent risk factor; the incidence of prolonged cognitive dysfunction in patients in their sixties being 7%, while it was 14% in those older than 70 years of age. The results are likely to lead to a vigorous examination of the issues and many questions will be raised. Is the occurrence of prolonged cognitive dysfunction related to the type of anesthesia or to specific drugs? Is it more common in some operations than others? Will it improve over additional postoperative time? Does it occur in younger patients or after shorter procedures? Does it have anything to do with anesthesia and surgery at all or could it be a consequence of other events? What is the significance of the observation that a decline in activities of daily living was observed in many in whom cognitive dysfunction had been found?

90% of Elderly Showed No Deficit, But…

Physicians and nurses may want to study the details of the Lancet article which is likely to be quoted in the lay press. It will not be wrong to point out to our patients that 90% of elderly patients showed no evidence of prolonged cognitive dysfunction after anesthesia and surgery. For the 10% of patients who do suffer cognitive dysfunction, and especially for those who are near the threshold of losing their ability to function independently, the worry arises that the effects of surgery and anesthesia might hasten the day when they lose their independence. This ground-breaking study raises many issues that will have to be examined in future research.

Dr. Gravenstein, a founding member of the APSF Executive Committee, is Graduate Research Professor, University of Florida, Gainesville.


1. Moller JT, Fluitmans P, Rasmussen LS, Houx P, Rasmussen H, Canet J, Rabbitt P, Jolles J, Larsen K, Hanning CD, Langeron O, Johnson T, Lauven PM, Kristensen PA, Biedler A, Beem HV, Fraidakis O, Beneken JEW, Gravenstein JS: Prolonged Postoperative Cognitive Dysfunction in the Elderly. Lancet 1998;351: 857-861.

2. Moller JT, Pedersen T, Rasmussen LS, Jensen PF, Pedersen BD, Ravlo O, Rasmussen NH, Espersen K, Johannessen NW, Cooper JB, Gravenstein JS, Chraemmer Jrrgensen B, Wiberg Jrrgensen 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.

3. Moller JT, Johannessen NH, Espersen K, Ravlo O, Pedersen BD, Jensen PF, Rasmussen NH, Rasmussen LS, Pedersen T, Cooper JB, Gravenstein JS, Chraemmer Jrrgensen B, Djernes M, Wiberg Jrrgensen 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.

4. Moller, JT, Svennild I, Johannessen NW, Jensen PF, Espersen K, Gravenstein JS, Cooper JB, Djernes M, Johansen SH: Perioperative monitoring with pulse oximetry. Brit. J Anaesthesia 71;340-347, 1993.