Predicting Adverse Postoperative Outcomes in Patients Aged 80 Years and Older

abstract & commentary

Synopsis: Preoperative comorbidities were more important than intraoperative events in predicting the 25% adverse postoperative outcomes in this group of geriatric surgical patients who sustained an in-hospital mortality rate of 4.6%. Neurological and cardiovascular complications were the leading causes of morbidity. The only intraoperative event shown to be predictive of complications was the use of vasoactive agents.

Source: Liu LL, et al. J Am Geriatr Soc 2000;48:405-412.

Liu and colleagues at the university of california-San Francisco reviewed medical records from all noncardiac surgical admissions at two teaching hospitals for patients aged 80 years and older over a one-year period in 1995 (identified from an operating room database). Ambulatory surgery and cardiothoracic cases were excluded; 367 patients who underwent 410 procedures were studied. Anesthesia records, type of surgery, intraoperative drug records, and O2 saturation were also studied. Adverse outcomes were verified by a reviewer blinded to the patient’s preoperative status. Discharge disposition and length of stay were also evaluated.

Results from both hospitals were combined together when demographics and preoperative predictors were found to be similar. A total of 86% had one or more preoperative risk factors, the most common being hypertension (50%), coronary artery disease (30%), preexisting neurological disease (29%), and pulmonary disease (23%). Fully 30% had a history of smoking. Only 14% had no preoperative conditions; 41% had three or more conditions. The most common procedures performed were orthopedic (hip fracture repair and hip or knee arthroplasties) and exploratory laparotomies, with 68% receiving general anesthesia.

Predictors of postoperative deaths included history of congestive heart failure, coronary artery disease, neurologic disease or use of vasopressors intraoperatively; these same predictors were associated with nonfatal adverse outcomes along with a history of arrythmia and urgent/emergent surgery. Age, gender, type of surgery, and number of comorbid conditions were not associated with adverse outcomes by univariate analysis. Choice of general vs. regional anesthesia were not clearly associated with any difference in complications either. Further analysis by multivariate logistic regression models found that CHF, arrythmia, and history of neurologic disease increased the odds of any adverse postoperative event.

Increased hospital length of stay was predicted mainly by preoperative conditions (history of CHF, coronary artery disease, and cerebral vascular accidents) and urgent/emergent surgery. The fairly large number of adverse intraoperative events (91 events with 80 procedures, or 19% of patients) did not affect length of stay except for intraoperative vasopressor use. If a patient had a postoperative acute MI, CHF, arrythmia, or the need for a second operation the stay was also prolonged. Postoperative delirium was the most common neurologic adverse outcome, occurring in nearly 15% of the patients.

Discharge placements showed 55% went home and 45% went to skilled nursing facilities; the latter group was 33% more likely to have suffered a postoperative complication than the group discharged to home (15% complications). Liu et al note that the economic effect of these placements and the extra hospital days are significant, along with the unmeasured cost to caregivers and quality of life for the older patients.

In conclusion, they note that previous studies have not examined geriatric surgical outcomes related to intraoperative events, and the current review does not suggest any predictive associations except for the use of vasoactive agents. However, only nine of these events were observed out of 91 adverse events, so they acknowledge small numbers do not lead to strong conclusions. Liu et al urge that more studies be done to identify optimal management of CHF and cognitive dysfunction to improve postoperative outcomes, and that geriatric surgical patients without significant comorbidities be considered in a "low risk" group for cost-saving reductions in treatment and length of stay.

COMMENT BY MARY ELINA FERRIS, MD

With the fastest growing segment of the elderly population in the eighth decade of age and older, more medical research is needed for this group that has been previously neglected. Clinicians are often called upon to advise elderly patients and their families about surgical risks, and the preoperative history and physical is an important opportunity to adjust any modifiable health factors that could contribute to improved outcomes. It has long been known in geriatric medicine that age alone does not predict adverse surgical outcomes, but rather that the underlying state of health is more important. However, with advancing age the inevitable cumulative comorbidities make it difficult to distinguish a lower risk group since nearly all patients have some potential medical complications.

This study is aimed at establishing postoperative noncardiac surgical morbidity and mortality in patients aged 80 years and older in the current environment of improved perioperative care and outcomes. It originated in the department of anesthesia, so the emphasis was also on identifying any intraoperative and anesthetic management issues that might contribute to adverse outcomes, but its findings are interesting for all those who care for the elderly and are part of the decisions surrounding surgical choices and preparation.

Its weakness may be that the two hospitals studied are part of a prestigious medical school with many referral cases (although 1 hospital is more community-based than the other), biasing the sample toward a sicker or more complicated group of elderly patients. On the other hand, the intrahospital care delivered may have been superior to community hospitals that do not have the luxury of resident housestaff and the oversight of multiple faculty and geriatric services. Why the 1995 calendar year was chosen and not a more recent year was not explained; over the past five years, advances in treatment may make their conclusions less valid. Inpatient medical chart review as the only source of information has limitations and may not have provided complete information on the patients, particularly when so much testing is performed in the outpatient setting and when many complications may occur after discharge.

The findings emphasize the importance of preoperative screening and treatment of any modifiable cardiac, pulmonary, and neurologic conditions. Age alone is not a predictor of adverse surgical outcomes, but rather the pre-existing state of health of the elderly person.

The high rate of postoperative delirium (15%), which prolonged the hospital stays by two days, merits further attention; at least one-third of the patients had preexisting neurologic disease but their mental status before surgery is unknown. Preoperative cognitive impairment has been shown in previous studies to increase the risk of in-hospital delirium,1 which subsequently increases the risk of mortality and subsequent institutional placement. A large Danish study of 1218 noncardiac surgical patients older than age 60 showed an even higher rate of 26% postoperative cognitive impairment one week after surgery, which persisted in 10% three months later, most commonly in the oldest patients.2 Future areas needing study include the role of pharmacologic agents and postoperative pain management in the etiology of these costly deliriums.

An accompanying article in the same issue from geriatric experts at RAND Corporation and UCLA identifies 21 target conditions in the vulnerable elderly for quality of care improvement.3 Hospitalization and surgery ranked no. 6, with pharmacologic problems no. 1, and dementia and delirium no. 3. Clearly, all these areas need more attention for our older patients and will help us improve the care we deliver.

References

1. Lipowski Z. N Engl J Med 1989;320:578-582.

2. Moller JT, et al. Lancet 1998;351:857-861.

3. Sloss EM, et al. J Am Geriatr Soc 2000;48:363-369.