Fragmented Sleep Enhances Postoperative Neuroinflammation but Not Cognitive Dysfunction.

Abstract

Sleep is integral to biologic function, and sleep disruption can result in both physiological and psychologic dysfunction including cognitive decline. Surgery activates the innate immune system, inducing neuroinflammatory changes that interfere with cognition. Because surgical patients with sleep disorders have an increased likelihood of exhibiting postoperative delirium, an acute form of cognitive decline, we investigated the contribution of perioperative sleep fragmentation (SF) to the neuroinflammatory and cognitive responses of surgery. The effects of 24-hour SF and surgery were explored in adult C57BL/6J male mice. The SF procedure started at 7 AM with cages being placed on a large platform orbital shaker that cycled every 120 seconds (30 seconds on/90 seconds off) for 24 hours. In separate cohorts, stabilized tibial fracture was performed either before or after the 24-hour SF procedure and assessed for systemic and hippocampal inflammation and cognition. SF-induced nonhippocampal memory dysfunction (mean ± standard deviation [SD] of the difference in time spent between novel and familiar object for control was 4.7 ± 1.4 seconds, n = 8 versus SF -0.5 ± 0.2 seconds, n = 11, yielding an estimated treatment effect of 5.2 seconds [95% confidence interval {CI}, 2.6-7.7]; P < .001) and increased systemic interleukin-6 (median [25%-75% quartile] for control 0.0 [0.0-2.4] pg/mL versus 9.7 [6.3-12.9] pg/mL, n = 8/group, yielding an estimated treatment effect of 9.7 pg/mL [95% CI, 5.8-11.8]; P < .0001). SF reduced freezing time in hippocampal-dependent memory test (mean ± SD for control 49.3% ± 5.8% versus for SF 32.9% ± 5.8%, n = 10/group, estimated treatment effect = 16.4% [95% CI, 11.0-21.8]; P < .0001). Although surgery also reduced freezing time (mean ± SD for control 49.3% ± 5.8% versus for surgery 30.3% ± 3.3%, n = 10/group, estimated treatment effect = 19.0% [95% CI, 14.6-23.4]; P < .0001), memory impairment was not further exacerbated by combining SF with surgery. One day after SF, there was an increase in hippocampal messenger RNA expression of tumor necrosis factor-α (relative quantitation [RQ] 5.12-fold, n = 5/group [95% CI, 1.64-15.97]; P < .01), and 1 day after surgery, there was an increase in messenger RNA interleukin-6 (RQ 4.64-fold, n = 5 [95% CI, 1.48-14.56]; P < .05) and tumor necrosis factor-α (RQ 5.54-fold, n = 5 [95% CI, 2.92-10.51]; P < .01). These increments were more pronounced when either pre- or postoperative SF was combined with surgery. Although SF and surgery can independently produce significant memory impairment, perioperative SF significantly increased hippocampal inflammation without further cognitive impairment. The dissociation between neuroinflammation and cognitive decline may relate to the use of a sole memory paradigm that does not capture other aspects of cognition, especially learning.