Confused Mice and Men: Can We Establish a Clinically Relevant Animal Model for Postoperative Delirium?

Christian S. Guay, MD

It is unlikely that we will ever be able to confidently diagnose a rodent with delirium. However, we can create animal models that exhibit delirium-like behavior. Leveraging multiple tests that assess different domains of cognition, the underlying mechanisms of postoperative delirium may become clearer, paving the way for targeted therapeutic interventions.

In this panel, “To Study the Underlying Mechanisms of Postoperative Delirium: Establishment of Clinically Relevant Animal Models” during the IARS 2021 Annual Meeting, experts in the basic science of postoperative delirium weighed in on an important question: can we develop valid animal models to study postoperative delirium, and if so, how?

Zhongcong Xie, MD, PhD, Professor of Anaesthesia at The Massachusetts General Hospital and Harvard Medical School, moderated the panel in addition to presenting some of his own work on the topic.

Postoperative delirium is a highly prevalent disorder defined as an acute confusion state characterized by a decline in attention and cognition after anesthesia and surgery. Its incidence ranges from 15-53% in the general surgical population and increases to 70-87% in the intensive care setting. Importantly, it is associated with increased morbidity and mortality as well as poor functional postoperative recovery and quality of life. Unfortunately, there is currently no consensus for a valid animal model to study the underlying mechanisms and possible interventions for postoperative delirium.

Dr. Xie highlighted one of the important challenges facing the field: clinical delirium assessments use multiple tests to assess several domains in patients who have no prior training, whereas most animal models use a single test that requires extensive training a priori. To overcome this challenge, his group developed a battery of tests for rodents inspired by the highly cited “Confusion Assessment Method” (CAM) developed for humans. A patient is considered delirious with the CAM if they exhibit a confusion state that is acute in onset and fluctuating in course (1), if they exhibit inattention (2), and if they exhibit either disorganized thinking (3) or altered level consciousness (4). Dr. Xie summarized the CAM scoring as “1 + 2 + 3 or 4 = delirium.” Each domain is assessed using multiple tests. Likewise, the “CAM in mice” incorporates multiple tests such as the buried food test, open field test and Y mazed test to assess the traditional CAM domains at multiple timepoints. Interestingly, older mice were shown to be more likely to test positive on the “CAM in mice” after surgery and anesthesia than younger mice, recapitulating the finding that elderly patients are at increased risk of developing postoperative delirium. A probiotic intervention was also demonstrated to mitigate the risk of testing positive in mice, paving the way for future studies in humans.

Mark Baxter, PhD, a neuroscientist working at the Icahn School of Medicine at Mount Sinai, began his presentation with a simple suggestion bearing deep implications for experimental design and interpretation: instead of trying to create animal models of postoperative delirium, we should develop models for postoperative delirium. The essence of his argument lies in the fact that we have no way of effectively communicating with nonhuman animals as we do with humans, and therefore cannot truly know whether they are delirious. So instead of pursuing the futile task of identifying and studying truly delirious animals, Dr. Baxter suggests we refocus our attention on creating behavioral tasks and animal models that are useful to study the disease process itself, a subtle but important distinction. Behavioral tasks that are currently in use appear to be sensitive but not specific for impairments characteristic of delirium. With this in mind, Dr. Baxter recommended using different behavioral models to provide converging lines of evidence, congruent with Dr. Xie’s “CAM in mice.”

Deborah Culley, MD, Department Chair of Anesthesiology and Critical Care at The University of Pennsylvania, echoed the other panelists’ views on the challenges of developing valid animal models to study postoperative delirium. It’s difficult to recreate the perioperative and intensive care environments, and perhaps even harder to assess nonhuman cognition in a clinically relevant model. Despite these challenges, she reviewed some of the literature detailing acute and persistent cognitive impairment in rodents and humans after anesthesia and surgery. A note of caution Dr. Culley shared with participants: if you want to study postoperative delirium in rodents, you cannot only use young animals! This point again highlights Dr. Xie’s findings with the “CAM in mice.”  Dr. Culley ended the panel with optimism that we can in fact establish clinically relevant animal models to study postoperative delirium, but that we mustn’t forget that old mice aren’t humans.