Predicting Effects of Surgery on Cognition


by Carl Sherman

July 25, 2016

Americans 65 and older undergo nearly 20 million surgeries each year, which frequently improve and sometimes save their lives. But the brain doesn’t always fare so well. Delirium in the immediate post-operative period is alarmingly frequent. More prolonged dysfunction is not rare. And there’s a substantial increase in subsequent dementia.

Clinicians have noted the association between surgery and cognitive compromise for decades, but only recently have researchers started paying serious attention. “This is really a bedside to bench story,” said Howard Fillit, director of the Alzheimer’s Drug Discovery Foundation.

The Foundation was a co-sponsor—the other was the New York Academy of Sciences—of what was perhaps the first conference devoted to surgery and cognition, held recently at the Academy.

“It’s a topic of tremendous importance,” said Sharon Inouye, professor of medicine at Harvard Medical School. Surgeries on older people increased by one-third from 2000 to 2010. Serious complications ensue 10–25 percent of the time, and “delirium is, according to some studies, the most frequent and devastating of these,” she said.

Besides delirium—acute confusion with disruption in attention and other cognitive functions —surgery can be followed by more sustained post-operative cognitive decline (POCD), and dementia months or years later. “The interrelationship between these three conditions remains unclear, and it’s crucial to gain understanding to prevent and manage them,” Inouye said.

Focus on delirium

Post-operative delirium occurs in 15–53 percent of older patients, depending on factors like age, comorbidity (the presence of additional disorders and diseases), and the specifics of the surgery and anesthesia. It is associated with increased risk of institutionalization and of death.

The apparent relationship with dementia is bi-directional. Pre-existing dementia or mild cognitive impairment are leading risk factors for delirium, and dementia rates increase, as much as 12-fold according to limited research, after a delirium episode.

“It’s not just dementia,” Inouye said. “The risk of decline in cognitive score is strongly suggested in various studies.”

Cognitive function after surgery typically follows a trajectory: an immediate dip, then gradual recovery. In one study of 225 cardiac surgery patients, those who hadn’t had delirium improved to above baseline function. Patients whose delirium lasted 3 days or longer still hadn’t recovered fully a year later, and those with briefer delirium fell between.

The Successful Aging after Elective Surgery (SAGES) study, which Inouye directs, followed 556 post-surgery patients for up to 36 months. Those without delirium were still above baseline cognitive function 3 years later; while the 24 percent who had had an episode declined gradually over this period, to a degree equivalent to mild cognitive impairment.

The implications of such studies are important, she said, in that "delirium is potentially preventable." One important step would be identifying people at heightened risk. Inouye said an analysis of SAGES data found that Alzheimer's disease biomarkers such as APOE4 failed to predict delirium. But neuroimaging of 136 SAGES patients suggested an association with pre-existing structural disconnection between hemispheres and in frontal-thalamic networks involved in memory and limbic function.

Inouye pointed out that while epidemiological associations and studies such as SAGES are highly suggestive, they don't establish causality between delirium, post-surgical cognitive decline, and dementia.

 Areas of uncertainty

Miles Berger, of Duke University, elaborated on this theme. "There are two ways to interpret data showing that long-term decline is worse in patients who had cognitive impairment soon after surgery. The alternative hypothesis is that some had pre-existing Alzheimer's pathology or a lack of cognitive reserve, and anesthesia and surgery unmasked it: They constitute a stress test for the aging brain."

 "Reverse causality" is possible, he said: "motor dysfunction in an early, preclinical phase of Alzheimer's could lead to falls needing surgery." Several large studies suggest that people with Alzheimer's disease (AD) have higher surgery rates.

Berger described research showing post-operative changes in biomarkers of processes underlying AD: a three-fold increase in CSF tau protein and in the tau/amyloid-beta ratio— into the range seen in AD—in the 24-hour period surrounding surgery.

"It was unclear whether this was caused by the surgery, the anesthesia, or both," but the fact that changes were the same with deep intracranial and peripheral central nervous system surgery pointed toward anesthesia.

More importantly, Berger said, "it's uncertain how long these changes persist, and what they mean functionally. They could be just an acute phase reaction."

A subsequent study, now in progress, might clarify some questions. Berger and colleagues are following these biomarkers and cognitive indices in 100 patients from before to a year after various surgical procedures. They also will do neuroimaging to compare connectivity in the surgery group and controls will also be done, he said.

Dissecting the cognitive effects of anesthesia from surgery is difficult, other presenters agreed. Joshua Mincer, of the Icahn School of Medicine at Mt. Sinai, described an ongoing study that may advance that effort.

Investigators are using fMRI to track resting state brain activity and cognitive scores in healthy elders before and up to a year after two hours of general anesthesia without surgery. In particular, they are assessing “dimensionality,” a measure of the complexity of resting state network activity, during the trajectory of recovery. Preliminary findings have not yet been published.

Seeking insights to protect the brain

 Michael S. Avidan, professor of anesthesiology at Washington University in St. Louis, described research using EEG to investigate (and perhaps mitigate) anesthesia’s effect on the brain.

A number of studies have linked EEG patterns during anesthesia with post-operative delirium. In particular, “burst suppression,” in which periods of flattened activity alternate with bursts of electrical discharge, has been associated with up to a fourfold increase in delirium.

In a meta-analysis of four studies, using EEG to guide anesthesia administration, cutting back when burst suppression occurred, reduced delirium incidence. Avidan described an ongoing randomized controlled trial of EEG-guided anesthesia titration, ultimately enrolling 1200 patients, to explore the question further.

“Is burst suppression a murderer, a mediator, or a mirror of brain vulnerability? We need to sort this out,” Avidan said. “If people are vulnerable, we need to intervene, while in healthy people [anesthesia depth] may not matter.”

EEG monitoring in the post-operative period might help distinguish delirium characterized by somnolence—a common, easily misdiagnosed form—from other conditions, such as sustained non-convulsive seizure, that require different treatment. “Continuous EEG can be an important adjunct to clinical assessment,” he said.

While evidence of association between surgery, anesthesia, delirium and lasting cognitive compromise has accumulated, the pathophysiology linking them is speculative. Edward R. Marcantonio, professor of medicine at Harvard Medical School, discussed a leading contender: neuroinflammation.

According to this hypothesis, “surgery, a major inflammatory event in the periphery, releases cytokines, which can cross the blood-brain barrier and activate microglia, setting up a process of neuroinflammation that leads to neuronal insult, even death. This can resolve quickly [as in the case of delirium], or persist, resulting in post-operative cognitive decline, even dementia.”

Marcantonio described several studies that support this hypothesis. Data from SAGES compared inflammatory biomarkers in elderly patients who did and did not develop post-surgical delirium. “There were two significant results,” he said: IL-6 was similar in both patient groups initially, but rose substantially in those who developed delirium; IL-2 was elevated throughout in this group. The latter cytokine, which regulates blood-brain barrier permeability, may be a marker of delirium risk, he said.

Another study found that C-reactive protein, a marker of active inflammation, was higher before and rose dramatically after surgery in patients who developed delirium. “Those in the highest CRP quartile had the most severe, prolonged [episodes],” he said.

“Delirium may be associated with a heightened inflammatory response to stress, and patients at risk may have higher levels of inflammation even before surgery.” A meta-analysis of animal studies linked inflammation after surgery to reduced cognitive performance, he said.

The ultimate goal is clinical translation, Marcantonio said. “Hopefully we can integrate biomarkers to identify individuals at risk of delirium and long-term [post-operative] cognitive dysfunction, and develop intervention strategies that protect the brain.”