A clinical prediction rule for delirium after elective noncardiac surgery

Objective: To develop and validate a clinical prediction rule for postoperative delirium using data available to clinicians preoperatively.

Design: Prospective cohort study.

Setting: General surgery, orthopedic surgery, and gynecology services at Brigham and Women's Hospital, Boston, Mass.

Patients: Consenting patients older than 50 years admitted for major elective noncardiac surgery between November 1, 1990, and March 15, 1992 (N = 1341).

Measurements: All patients underwent preoperative evaluations, including a medical history, physical examination, laboratory tests, and assessments of physical and cognitive function using the Specific Activity Scale and the Telephone Interview for Cognitive Status. Postoperative delirium was diagnosed using the Confusion Assessment Method or using data from the medical record and the hospital's nursing intensity index. Patients were followed up for the duration of hospitalization to determine major complication rates, length of stay, and discharge disposition.

Results: Postoperative delirium occurred in 117 (9%) of the 1341 patients studied. Independent correlates included age 70 years or older; self-reported alcohol abuse; poor cognitive status; poor functional status; markedly abnormal preoperative serum sodium, potassium, or glucose level; noncardiac thoracic surgery; and aortic aneurysm surgery. Using these seven preoperative factors, a simple predictive rule was developed. In an independent population, the rule stratified patients into groups with low (2%), medium (8%, 13%), and high (50%) rates of postoperative delirium. Patients who developed delirium had higher rates of major complications, longer lengths of stay, and higher rates of discharge to long-term care or rehabilitative facilities.

Conclusions: Using data available preoperatively, clinicians can stratify patients into risk groups for the development of delirium. Since delirium is associated with a variety of adverse outcomes, patients with substantial risk for this complication could be candidates for interventions to reduce the incidence of postoperative delirium and potentially improve overall surgical outcomes.