Early Studies of Anesthesia-Related Mortality

One of the earliest systematic approaches to anesthesia-related risk^{[21] [35] [43] [44] [45] [46] [47] [48] [49] [50] [51] [52] [53]} occurred in 1935, when Ruth^[54] helped establish the first anesthesia study commission to analyze perioperative deaths. The commissioners relied on voluntary submission of cases and determined the cause of death by majority vote. Both of these methodologies were deemed inadequate in subsequent years ( Table 24-3 ).

From a pragmatic standpoint, determining the cause of anesthesia-related mortality is important if information from the analysis can be used to improve subsequent care. This concept was best voiced by Sir Robert MacIntosh in 1948.^[55] He stated that many anesthesia-related deaths were preventable and that improved education was the best means to avoid unnecessary mortality. He went so far as to suggest that all development of new drugs be halted for 5 years to direct more attention to training young anesthetists. He attempted to focus attention on postoperative care, suggesting that more deaths occurred after patients were returned to the ward than intraoperatively. His final comments were related to the need to establish a formal inquiry mechanism for determining the cause of any death, because such an analysis could lead to improved practice. In many ways, the CEPOD was the culmination of such a philosophy.^[20]

A major advance in the analysis of anesthesia-related risk was the 1954 report by Beecher and Todd^[21] of anesthesia-related deaths at 10 institutions. Their study included 599,548 anesthesia procedures. The cause of mortality was determined at the local institution by consensus of a surgeon and the chief anesthetist of the institution. Each death was characterized as having one primary cause and could have multiple secondary causes. The overall chance of mortality from any cause was 1 per 75 cases. Anesthesia was the primary cause of mortality in 1 of 2680 procedures, and it was a primary or contributory cause of mortality in 1 of 1560 procedures. Surgical error

898

Data are present to show that death from anesthesia is of sufficient magnitude to constitute a public health problem. Anesthesia kills several times as many citizens each year out of the total population of the country as does poliomyelitis. Consideration of the millions of dollars rightly spent in attacking poliomyelitis and the next to nothing, comparatively, spent in anesthesia research makes clear an urgent need.

Dornette and Orth^[43] reported on deaths occurring in the operating room at their institution during a 12-year period from 1943 through 1954. During this period, 95 patients died on the operating table, 12 patients suffered an intraoperative cardiac arrest and died postoperatively, and 1 patient died in the recovery room. This resulted in a total of 108 deaths in 63,105 anesthesia procedures. In 19 cases, the death was judged to be the result of the operation. In 13 patients, preexisting disease was judged to be the cause of death. The patient's condition plus the operation was a cause of death in 29 instances. Anesthesia was the primary cause of death in 26 cases (2 of which were the result of an overdose of local anesthetic administered by a surgeon) and a partial cause in an additional 21 patients. The rate of mortality totally attributable to anesthesia was 1 in 2427 cases, and the rate of mortality totally or partially attributable to anesthesia was 1 in 1343 cases.

Dripps and colleagues^[46] at the University of Pennsylvania surveyed their experience during the 10-year period from 1947 through 1957. They identified 1285 operative deaths (i.e., death within 30 days) in approximately 120,000 anesthesia procedures, for a gross mortality rate of 1.1%. This definition includes late deaths, in contrast to many studies that have focused on the intraoperative period or the first 48 postoperative hours. After review of the hospital records, the investigators determined whether anesthesia was definitely or possibly contributory to each death. Among patients who underwent spinal anesthesia, mortality definitely related to anesthesia occurred in 1 of 1560 procedures, and mortality definitely or possibly related to anesthesia occurred in 1 of 780 procedures. Among those who underwent general anesthesia, mortality was definitely related to anesthesia in 1 of 536 cases and definitely or possibly related in 1 of 259 cases. Mortality was correlated with physical status with the use of the ASA Physical Status Classification scoring system. The higher mortality figures compared with other reports may be related to the higher physical acuity at this hospital. Importantly, deaths attributable to anesthesia did not occur among any of the 16,000 patients with ASA class I physical status.

Dripps and colleagues^[46] compared their data with those of Beecher and Todd.^[21] The principal intraoperative complications were hypotension and hypoxia. Virtually all patients who had spinal anesthesia had intraoperative complications, but postoperative complications were rare in this group. The investigators believed that anesthesia-related mortality had improved since the start of the study, mainly because of improvements in cardiac resuscitation, more rational transfusion therapy, standard use of recovery rooms, and efficient use of mechanical ventilators.

There were a number of reports from individual hospitals or a small group of hospitals during the subsequent 2 decades.^[49] The Baltimore Anesthesia Study Committee^[45] reviewed 1024 deaths occurring on the day of or the day after a surgical procedure to determine the potential contribution of anesthesia. This short period of observation (48 hours) contrasts with the 30-day period used by Dripps and coworkers.^[46] For each case reviewed, the committee determined whether anesthesia was the principal cause or one of several contributing factors. In 196 cases (19.2%), the committee voted that anesthesia management contributed to the death of the patient. Anesthesia was the principal cause of death in approximately one third of cases (64 of 196), with one half of those being related to improper management of the anesthetic. The researchers estimated the rate of operative mortality rate as 4 in 10,000 operations using these data and applying rates of surgery from the National Health Survey. In more than 50% of all cases studied, death occurred in the patient's room, leading the investigators to emphasize the need for routine use of postanesthesia care areas. In an attempt to educate practitioners and prevent duplication of preventable causes of death, each of the perioperative deaths was discussed.

Schapira and coworkers^[44] reported on mortality occurring within 24 hours after surgery between the years 1952 and 1956 at Montefiore Hospital in New York. They ascribed 27 deaths to anesthesia, including 18 in which anesthesia was a primary cause and 9 in which it was contributory. The overall prevalence of death partially or totally attributable to anesthesia was 1 in 1232 procedures.

Clifton and Hotton^[47] reported 162 deaths associated with anesthesia in 205,640 operations performed at the Royal Prince Alfred Hospital in Sydney, Australia, between 1952 and 1962. They calculated that the incidence of mortality totally attributable to anesthesia was 1 in 3955, that attributable to surgery was 1 in 2311, and that attributable to patient disease was 1 in 1996 procedures. One cause of postoperative mortality was respiratory insufficiency. The researchers argued that many of these complications would have been prevented by the use of a recovery unit. The potential safety advantages of a postanesthesia care unit (PACU) constituted a general theme in reports from the 1960s.

Dinnick^[56] reported 600 deaths associated with anesthesia as part of a series of investigations sponsored by the Association of Anaesthetists in London. This represents the second report from this group since a committee was established to collect clinical data in 1949. The first report included a review of 1000 fatalities and concluded that "in the majority of the reports there were departures from accepted practice." The initial report, published in 1956, found that regurgitation or vomiting was the main factor for death, whereas the second report, published in 1964, found that low blood volume was more important. Underventilation was the second most common finding, accounting for 25% of the cases.

899

Bodlander^[52] provided a follow-up report of anesthesia-related mortality at the Royal Prince Alfred Hospital for the years 1963 to 1972. The incidence of mortality totally attributed to anesthesia decreased to 1 in 1702 cases, whereas anesthesia contributed to mortality in 1 in 502 cases, compared with 1 in 1208 for the years 1952 to 1962. The reduction in anesthesia-related deaths was attributed to an increase in the number of qualified staff and the degree of supervision.

Marx and colleagues^[35] evaluated the incidence of death within 7 days after surgery among 34,145 consecutive patients at the Bronx Municipal Hospital Center between 1965 and 1969. A total of 645 patients died, and a death report form was constructed based on that developed by a committee of the New York Academy of Medicine. The deaths were then analyzed in relation to perioperative data available from a computer system. The patient's preexisting disease was considered to be the primary cause of death in 83% of cases, operation in 10%, and anesthesia in only 4% (1 of 1265 cases). Although mortality rose progressively with age, physical status correlated best with the incidence of mortality.

Marx and colleagues^[35] also determined the relationship between type of anesthesia and mortality. Regional anesthesia was associated with the lowest incidence of death, local anesthesia with the highest; the incidence for general anesthesia was intermediate. Although this relationship was significantly different between groups, the difference appeared to be related to patient risk factors, a finding supported by a study by Cohen and colleagues. ^[51A]

Farrow and coworkers^{[57] [58]} studied hospital mortality after 108,878 anesthesia procedures in Cardiff, Wales, between 1972 and 1977. The crude mortality rate was 2.2 per 100 patients. Mortality was greatest among patients older than 65 years. The mortality rate also increased with the severity of disease and the need for emergent operations.

Harrison^[51] evaluated mortality associated with 240,483 anesthesia procedures performed between 1967 and 1976 at Groote Schuur Hospital in Cape Town, South Africa. Data were collected prospectively, beginning in 1956. Anesthesia was the cause of death or major contributory factor in 0.22 cases per 1000 procedures, compared with 0.33 per 1000 operations in the previous 10 years. Anesthesia contributed to 2.2% of all of the deaths associated with surgery. The most common causes of anesthesia-related mortality, in order of frequency, were hypovolemia, respiratory inadequacy after neuromuscular blockade, complications of tracheal intubation, and inadequate postoperative care. Although the improvement in anesthesia-related mortality could not be directly attributed to specific improvements in care, four specific changes did occur: continuing improvement in routine monitoring, an increase in the ratio of consultants to registrars (British system residents), a decrease in the case load per anesthesiologist, and the introduction of recovery rooms and intensive care units (ICUs).

Studies before 1980 demonstrated steady improvements in anesthesia-related mortality. Studies performed throughout the world focused on identifying the causative factors for perioperative mortality. Several general themes emerged. Anesthesia represents a small but significant cause of perioperative mortality, perioperative respiratory complications represent a major complication, and elucidation of the causes of perioperative mortality and education about these causes should lead to improved outcomes.