Anesthesia-Related Mortality Studies after 1980

Whereas studies conducted before 1980 typically focused on one institution or a small group of institutions, studies since then have frequently been performed on a national basis. Improvements in anesthesia-related and overall mortality have made the analysis of a single institution inadequate because of the small sample size. For example, Holland^[59] reported the deaths occurring within 24 hours after an anesthesia procedure in New South Wales, Australia. A committee of six anesthesiologists, three surgeons, an obstetrician, a general practitioner, and a medical administrator was established in 1960 and reviewed all such cases except during a 3-year period between mid-1980 and mid-1983. Four categories were established to define the relationship of anesthesia to operative morbidity and mortality ( Table 24-4 ). Between 1960 and 1985, information was obtainable on 92% to 96% of all cases. Twenty-five percent of the 5262 deaths in that period were deemed attributable in whole or in part to the anesthesia procedure. The incidence of anesthesia-attributable deaths decreased from 1 in 5500 procedures performed in 1960 to 1 in 10,250 in 1970 and then to 1 in 26,000 in 1984 ( Table 24-5 ). Based on these estimates, the investigators asserted that it was at least five times safer to undergo anesthesia in 1984 compared with 1960, particularly for healthy individuals. A subsequent follow-up report stated that factors under the control of the anesthesiologist caused or contributed to perioperative mortality at a rate of 1 in 20,000 operations.^[60]

TABLE 24-4 -- Edwards classification of the relationship of anesthesia to operative morbidity and mortality*

Category	Definition
I	When it is reasonably certain that the event or death was caused by the anesthetic agent or technique of administration or in other ways coming directly within the anesthetist's province
II	Similar to type I cases, but ones in which there is some element of doubt about whether the agent or technique was entirely responsible for the result
III	Cases in which the patient's adverse event or death was caused by the anesthetic and the surgical technique
IV	Events entirely referable to surgical technique
From Holland R: Anaesthetic mortality in New South Wales. Br J Anaesth 59:834, 1987.

900

The New South Wales Committee determined the primary error in management that led to each perioperative death. The most frequent primary cause was inadequate preparation of the patient, followed by a wrong choice of anesthetic drug. From the committee's standpoint, a wrong choice of agent included administering a renally excreted agent to a patient with chronic renal failure. The third most common error was inadequate crisis management during the initial decade of the study, but this cause was much less important during the period from 1983 through 1985. Inadequate postoperative management was the second most common cause of problems during the final years of the study.

The New South Wales study of anesthesia-related mortality evaluated the contribution of the anesthetist on perioperative mortality. Four groups of providers were identified: specialists, nonspecialists, CRNAs, and residents. The absolute number of anesthesia-related deaths decreased in all groups but was most pronounced for the nonspecialists. During the period from 1960 through 1969, the resident medical officer frequently provided anesthesia. During this same period, it was found that residents contributed significantly to the mortality observed among "good-risk" patients, which led to a phasing out of the resident medical officer as a member of the anesthesia work force.

Under the direction of the French Ministry of Health, Tiret and colleagues^[61] carried out a prospective survey of complications associated with anesthesia in France between 1978 and 1982 from a representative sample of 198,103 anesthesia procedures chosen at random from hospitals throughout the country. The sample included a survey of 460 public and private hospitals. The investigators evaluated the occurrence of death or coma within 24 hours after surgery. The opinions of the participating anesthesiologist and that of the National Committee of Assessors were determined, and the latter was accepted if there was disagreement. In the group studied, 268 patients had major complications, 67 patients died, and 16 patients had persistent coma. Death was totally related to anesthesia for 1 in 13,207 procedures and partially related for 1 in 3810 ( Table 24-6 ). Sixty-two percent of the coma cases were deemed totally attributable to

TABLE 24-6 -- Incidence of complications partially or totally related to anesthesia

Complications	Partially Related	Totally Related	Total *
All complications	1:1887	1:1215	1:739
Death	1:3810	1:13,207	1:1957
Death and coma	1:3415	1:7924	1:2387
From Tiret L, Desmonts JM, Hatton F, Vourc'h G: Complications associated with anaesthesia—A prospective survey in France. Can Anaesth Soc J 33:336–344, 1986.

*Total number of anesthetics: 198,103.

One of the most important findings of the survey was that postanesthesia respiratory depression was the largest cause of death and coma that were totally attributable to anesthesia ( Table 24-7 ). Almost all of the patients who had respiratory depression leading to a major complication had received narcotics and muscle relaxants that had not been reversed. They also reported a high incidence of anaphylactoid shock, which the investigators contended was caused primarily by the use of Althesin and succinylcholine. There was no category of drug overdose, which might have been a more appropriate label for some of these cases.

The study by Tiret and colleagues^[61] had the advantage of collecting data prospectively, allowing more accurate estimation of overall mortality than many of the other studies completed. A major limitation of the study was that only deaths occurring within 24 hours after surgery were included, ignoring late deaths that were the direct result of intraoperative complications.

Tikkanen and Hovi-Viander^[62] studied death associated with anesthesia and surgery in Finland and compared the results in 1986 with those collected in 1975. Mortality related to anesthesia decreased during the 9-year period; the incidence of anesthesia-related mortality was 0.15 per 10,000 procedures in 1986.

Lunn and colleagues^{[63] [64] [65]} published two reports on anesthesia-related surgical mortality in the United Kingdom. When a death occurred in a hospital within 6 days after surgery, a questionnaire was sent to the patient's anesthetist and surgeon. For 59.3% of the 4034 reported deaths, both the surgeon and the anesthetist returned the forms. The replies were reviewed anonymously by two assessors, and differences of opinion were determined by arbitration. After review, further details were obtained if the reply indicated that anesthesia was at least partly responsible for the death. The second report was based on an analysis of 197 reports of death within 6 days after anesthesia during 1981. In this report, 43% of the deaths were found by the assessors to have

901

The pioneering work of Lunn and others led to the development of the CEPOD, which assessed almost 1 million cases of anesthesia during a 1-year period in 1987 in three large regions of the United Kingdom.^{[20] [66]} Unique to this study was the establishment of "crown privilege" by the government to allow total confidentiality:

The Secretary of State is satisfied that the disclosure of documents about individual cases prepared for the Enquiry into Perioperative Deaths would be against the public interest and would undermine the whole basis of a confidential study. The data or information sent to the Confidential Enquiry into Perioperative Deaths is therefore protected from subpoena...

Deaths occurring within 30 days after surgery were included in the study. There were 4034 deaths in an estimated 485,850 operations, resulting in a crude mortality rate of 0.7% to 0.8%. Surgery contributed totally or partially in 30% of all cases. Progression of the presenting disease contributed to death in 67.5% of the cases, and progress of an intercurrent disease was relevant in 44.3%. Anesthesia was considered the sole cause of death in only three individuals, for a rate of 1 in 185,000 cases, and anesthesia was contributory in 410 deaths, for a rate of 7 in 10,000 cases ( Table 24-8 ).

There are several potential causes for the improvement in mortality between the CEPOD study and previous studies, including cases from the same group. One explanation is that improvement in care led to improvement in outcome. Many of the deaths that would previously have been classified as "anesthesia totally contributory" were later classified as "anesthesia partially contributory."

An important aspect of the CEPOD study was that it established anesthesia- and surgery-related factors that contributed to mortality. The five most common causes of death are shown in Table 24-9 . Of the 410 perioperative deaths, there were 9 cases of aspiration or vomiting and 18 cases of cardiac arrest. A large proportion of elderly women had fractures of the femoral neck. The death rate was inversely related to the seniority of the operating surgeon and to preoperative preparation. The operating surgeon was a consultant in only 19% of the orthopedic cases, compared with 47% overall.

The CEPOD study also provided important information regarding anesthesia practice. Patients were seen preoperatively in more than 80% of the cases but postoperatively in less than 50%. Although the electrocardiogram was monitored in 97% of cases, core temperature was assessed in only 7%. Muscle relaxants were used in more than 50% of the cases, but a nerve stimulator was used in only 14%.

The assessors concluded that avoidable factors were present in about 20% of the perioperative deaths. Contributing factors for anesthesiologists and surgeons

TABLE 24-9 -- Most common clinical causes of death in the Confidential Enquiry into Perioperative Deaths

Cause of Death	Percent of Total
Bronchopneumonia	13.5
Congestive heart failure	10.8
Myocardial infarction	8.4
Pulmonary embolism	7.8
Respiratory failure	6.5
Adapted from Buck N, Devlin HB, Lunn JL: Report of a Confidential Enquiry into Perioperative Deaths, Nuffield Provincial Hospitals Trust. London, The King's Fund Publishing House, 1987.

902

Several large national studies have been published since the CEPOD. Pedersen and colleagues^[67] performed a series of studies in the late 1980s in Denmark to look at factors attributable to anesthesia that led to serious morbidity or mortality. They performed a prospective study of 7306 anesthesia procedures. In a method similar to that used in earlier studies, three anesthetists reviewed the records of complications and determined whether the cause was attributable to anesthesia (no distinction was made between totally and partially contributing factors). Complications attributable to anesthesia occurred in 43 patients (1 of 170), and 3 patients (1 of 2500) died. Complications in the 43 patients, in order of incidence, included cardiovascular collapse in 16 (37%), severe postoperative headache after regional anesthesia in 9 (21%), and awareness under anesthesia in 8 (19%). The researchers determined that 37% of the anesthesia-related morbidity was preventable. The three deaths occurred in severely ill patients (ASA physical status class III or greater), and two of these deaths were judged to be preventable.

Cohen and coworkers^[51A] developed a methodology for studying anesthesia outcome in four teaching hospitals in Canada. They developed a new anesthesia record to serve as a data collection instrument. Research nurses reviewed all inpatient records within 72 hours after surgery, conducted interviews of the patients using a standardized instrument, and conducted a telephone survey of most of the outpatients. Assessment of the contribution of anesthesia to each complication was determined using the classification of the New South Wales Committee on Anaesthesia Mortality (see Table 24-4 ). In a total of 6914 anesthesia procedures in adults, there were no deaths directly attributable to anesthesia.

Lagasse^[19] reviewed perioperative deaths (i.e., deaths occurring within 2 days after surgery) at a suburban university hospital network between 1992 and 1994 and an urban university hospital network between 1995 and 1999. There were a total of 347 deaths in 184,472 cases. Anesthesia-related mortality (i.e., death to which error by an anesthesia practitioner contributed) occurred in 1 of every 12,641 procedures in the suburban setting and in 1 of 13,322 procedures in the urban setting. Mortality increased with increasing ASA physical status ( Fig. 24-2 ). In reviewing data over the previous decade, the investigators estimated that the anesthesia-related mortality rate had remained stable at approximately 1 death per 13,000 procedures, with wide variation making it impossible to detect trends in anesthesia safety.