Postoperative cognitive dysfunction: true deterioration versus random variation

Background: Postoperative cognitive dysfunction (POCD) is a common complication, especially in the elderly. The aim of this study was to describe how variability in neuropsychological testing could lead to the detection of cognitive improvement and poor consistency of POCD between postoperative test sessions. Methods: In four published studies performed by the ISPOCD group, we included 2536 patients and 359 healthy controls. Cognitive function was assessed using neuropsychological tests preoperatively and at 7 days and 3 months thereafter, comparing the changes between those at baseline with those after surgery. Postoperative cognitive dysfunction was defined as a Z score greater than 2, and we also defined a corresponding improvement as a Z score less than -2. Consistency of POCD between postoperative test sessions was analyzed and we also assessed test-retest variability using data from healthy control subjects. Results: Improvement in cognitive function was found in 4.2-8.7% of patients after 1 week and in 5.0-7.8% after 3 months. The ratio between incidence of dysfunction and improvement varied in patients between 3.3 and 6.2 early after major surgery. Of those patients who displayed POCD at the 3-month test, 30-48% also had POCD at the previous test at 1 week. The test-retest reliability was between 0.56 and 0.90, except for the error score in Concept Shifting Test, where the values were 0.20 and 0.37. Conclusion: Variability in neuropsychological test data contributes to a low consistency between postoperative test sessions but it does not explain the detection of cognitive dysfunction after major surgery.