New Risk for Postoperative Delirium: Obstructive Sleep Apnea

Our multiple previous columns on delirium (see list at the end of this column) have mentioned multiple risk factors for the development of delirium in hospitalized patients. But a new risk factor for postoperative delirium has just been uncovered: obstructive sleep apnea (Flink 2012). The authors prospectively evaluated 106 nondemented elderly patients undergoing elective knee arthroplasty for delirium and found delirium in 25% of cases. The incidence of delirium in the 15 patients who had known obstructive sleep apnea (OSA) was 53%, compared with 20% in those without known OSA. In fact, in multivariate analyses, the only independent risk factor that emerged for delirium in this population was OSA. Patients with OSA were more than 4 times more likely to develop delirium. It was a well-done study in which the CAM (Confusion Assessment Method) and the DRS-R-98 (Delirium Rating Scale-Revised-98) were utilized on post-op days #2 and #3 (most patients were discharged by day #3) to identify patients with delirium and measure its severity. The incidence was highest on post-op day #2 as has been seen in prior studies and two-thirds of the cases had improved by post-op day #3.

The authors readily state that this study should be “hypothesis-generating” rather than showing a definitive cause and effect relationship between OSA and delirium but the findings are nevertheless striking and thought provoking.

The authors speculate on the potential mechanisms linking OSA and delirium. Obviously hypoxemia is suspected as a major factor and they speculate it might reduce ATP and synthesis of cholinergic precursors (reduced cholinergic activity has been supported as a mechanism of delirium in research settings). The accompanying editorial (Bateman 2012) notes that patients with OSA are most vulnerable to hypoxia on postoperative nights 2 and 3. Flink and colleagues also raise the possibility that inflammatory factors might play a role. Given the slight delay in the appearance of postoperative delirium one might also wonder about the possible role that “REM rebound” might play and its relationship to both OSA and postoperative delirium.

Recently, a clinical research group in the Netherlands had developed and validated a risk model for predicting delirium in hip fracture patients (Moerman 2012). Items considered in that model included prior episodes of delirium, presence of dementia, age, clock drawing, hearing and vision impairments, problems with ADL’s, and alcohol or substance abuse. The tool was pretty good at predicting delirium in this patient population. This population was clearly different from that in the Flink study but it would be interesting to go back and see whether OSA might have been identified in the Netherlands population.

In our August 17, 2010 Patient Safety Tip of the Week “Preoperative Consultation – Time to Change” we noted the relative ineffectiveness of the typical preoperative assessments done today. Instead of the intense focus on potential cardiac complications, we instead advocated for more focus on identifying frailty and risk factors for things like postoperative delirium and obstructive sleep apnea. Little did we know that those two might be interrelated! Admittedly there are a whole host of risk factors for delirium that should be considered in many patient populations. But now it makes sense to also screen for OSA, particularly in nondemented patient cohorts about to undergo elective surgical procedures. Screening for OSA is doubly important, given all the attention we’ve given to managing opiate therapy in the postoperative period (see links below).


Given the simplicity of screening for OSA with tools like the STOP-Bang questionnaire, there is really no good reason that such screening should not be part of the preoperative assessment, whether being done by an anesthesiologist, internist, hospitalist, or surgeon. It is incredible how often we send patients for cardiac stress testing preoperatively despite lack of clearcut evidence for its utility and cost-effectiveness in this population. Yet simple tests that can be done in a few minutes in the office, like the STOP-Bang questionnaire or the Timed-Up-And-Go Test (see our November 2011 What’s New in the Patient Safety World column “Timed Up-and-Go Test and Surgical Outcomes”), that are highly predictive of postoperative complications are seldom part of the preoperative evaluation.