The Impact of Inflammation and General Anesthesia on Memory and Executive Function in Mice

BACKGROUND: Perioperative neurocognitive disorders (PNDs) are complex, multifactorial conditions that are associated with poor long-term outcomes. Inflammation and exposure to general anesthetic drugs are likely contributing factors; however, the relative impact of each factor alone versus the combination of these factors remains poorly understood. The goal of this study was to compare the relative impact of inflammation, general anesthesia, and the combination of both factors on memory and executive function. METHODS: To induce neuroinflammation at the time of exposure to an anesthetic drug, adult male mice were treated with lipopolysaccharide (LPS) or vehicle. One day later, they were anesthetized with etomidate (or vehicle). Levels of proinflammatory cytokines were measured in the hippocampus and cortex 24 hours after LPS treatment. Recognition memory and executive function were assessed starting 24 hours after anesthesia using the novel object recognition assay and the puzzle box, respectively. Data are expressed as mean (or median) differences (95% confidence interval). RESULTS: LPS induced neuroinflammation, as indicated by elevated levels of proinflammatory cytokines, including interleukin-1 beta (LPS versus control, hippocampus: 3.49 pg/mg [2.06-4.92], P<.001; cortex: 2.60 pg/mg [0.83-4.40], P=.010) and tumor necrosis factor-alpha (hippocampus: 3.50 pg/mg [0.83-11.82], P=.002; cortex: 2.38 pg/mg [0.44-4.31], P=.021). Recognition memory was impaired in mice treated with LPS, as evinced by a lack of preference for the novel object (novel versus familiar: 1.03 seconds [-1.25 to 3.30], P=.689), but not in mice treated with etomidate alone (novel versus familiar: 2.38 seconds [0.15-4.60], P=.031). Mice cotreated with both LPS and etomidate also exhibited memory deficits (novel versus familiar: 1.40 seconds [-0.83 to 3.62], P=.383). In the puzzle box, mice treated with either LPS or etomidate alone showed no deficits. However, the combination of LPS and etomidate caused deficits in problem-solving tasks (door open task: -0.21 seconds [-0.40 to -0.01], P=.037; plug task: -0.30 seconds [-0.50 to -0.10], P <.001; log values versus control), indicating impaired executive function. CONCLUSIONS: Impairments in recognition memory were driven by inflammation. Deficits in executive function were only observed in mice cotreated with LPS and etomidate. Thus, an interplay between inflammation and etomidate anesthesia led to cognitive deficits that were not observed with either factor alone. These findings suggest that inflammation and anesthetic drugs may interact synergistically, or their combination may unmask covert or latent deficits induced by each factor alone, leading to PNDs.