October 2011
Reviewer: Jeff C. Huffman, MD
Individual risk profiles for postoperative delirium after joint replacement surgery
Jain FA, Brooks JO 3rd, Larsen KA, et al
Psychosomatics 2011; 52(5):410-416
Background: Postoperative delirium is exceedingly common, and it is associated with complications, prolonged hospitalizations, and mortality. A somewhat limited literature has identified some risk factors for postsurgical delirium and has developed risk profiles for delirium in medical inpatients.
Methods: This study was an analysis of data from a randomized trial of olanzapine to prevent delirium in patients undergoing elective orthopedic surgery; the primary goal of this study was to develop risk profiles for delirium in this population. Subjects were over 65, or under 65 with a history of delirium; exclusion criteria included dementia, alcohol dependence, or use of antipsychotics. Delirium was assessed using a number of standardized instruments that were used to make DSM-IV diagnoses of delirium. To establish risk profiles for development of delirium, the authors used recursive receiver operator curve (ROC) analysis, a method that identifies predictor variables for an outcome within individual subsets of patients. Variables examined as delirium risk factors included demographic characteristics; medical status (e.g. anesthesiology classification [ASA] score); pre-operative blood pressure, WBC, and hematocrit [HCT]; operative variables (complexity of surgery, prophylactic olanzapine, intraoperative morphine dose); and perioperative variables (lowest oxygen saturation on day of surgery and POD #1, highest WBC/lowest HCT on day of surgery, and pain score on surgery day and POD#1).
Results: Overall, 400 subjects were enrolled; 113 became delirious. Overall risk factors for delirium were age, higher ASA score, lower pre-op blood pressure, and more intraoperative morphine, though perhaps the strongest predictor was receiving placebo rather than olanzapine. Given this, the risk profiles were split between placebo and olanzapine patients. In the placebo arm, older patients (>73 years old) with poor ASA score and lower oxygen saturation were at highest risk for delirium. In the olanzapine arm, older patients who had lower pre-op blood pressure and received greater than the mean amount of morphine intraoperatively had the greatest risk.
Commentary: These interesting findings can help to generate hypotheses in future work that examines delirium risk factors. Though some of the risk factors identified were unsurprising, the findings regarding lower pre-operative blood pressure are novel and potentially quite important—it seems possible that having low baseline blood pressure may predispose patients to poor cerebral oxygenation (especially if they are getting high-dose morphine or other agents that may lower blood pressure) that can cause delirium. Certainly having a worse overall medical/surgical status, and lower baseline oxygenation seem also to be important factors, with the latter again seeming to be a risk factor for poor cerebral function, especially in patients who have procedures/medications that can further impede oxygenation.
This analysis had a number of limitations that make it more of a hypothesis-generating study. Though the authors utilized a large study sample, characterized many important risk factors, and used well-grounded methods to identify risk factors, by splitting the cohort into smaller and smaller subgroups (while keeping the same large number of predictor variables), these somewhat exploratory findings will need to be examined in future work. Furthermore, it would have been useful to include additional variables known to be associated with delirium, such as narcotics, benzodiazepines, or anticholinergics administered in the perioperative period. The cohort was also limited to older patients getting elective orthopedic procedures.
In short, this is thoughtful, interesting work that identifies some interesting and novel potential risk factors for delirium, and seems to signal that paying close attention to perfusion/blood pressure and oxygenation in the operative and peri-operative period may be quite important.