Brain Functional Connectivity Related To Post-op Atrial Fibrillation

Patients who develop atrial fibrillation after cardiac surgery have greater resting-state functional connectivity in certain areas of the brain than those who did not develop this sometimes devastating side effect, a small study has concluded.

These findings not only potentially provide insight into the central mechanisms responsible for autonomic derangement after cardiac surgery, but may ultimately open the door to the use of noninvasive cerebral stimulation technologies in cardiac surgery.

According to Nathan Waldron, MD, MHS, Adult Cardiothoracic Anesthesiology Fellow at Duke University Health System, in Durham, N.C., although autonomic imbalance has been implicated in the development of postoperative atrial fibrillation, the basis of this imbalance is poorly understood. Nevertheless, central autonomic integration has been associated with a diverse group of functionally interconnected brain networks. This led the investigators to delve deeper into resting-state functional connectivity in cardiac surgical patients, hypothesizing that patients with postoperative atrial fibrillation would demonstrate altered presurgical functional connectivity in the midcingulate and insular cortical regions of the brain.

It certainly helped that the department was rich with functional MRI (fMRI) data. “[Study co-investigator] Dr. Joseph Mathew is somewhat unique in that he’s a cardiac anesthesiologist, but he’s also very interested in neurocognitive function after surgery,” Dr. Waldron said. “And as part of some of his studies, he has an fMRI database that is powered to look at various things, including cerebral connectivity on the whole-brain level.”

Using this set of baseline, resting-state fMRI data, the investigators retrospectively evaluated 16 age-matched patients undergoing valvular procedures and/or major aortic surgery. After processing and normalization, cross-correlation coefficients were calculated for the bilateral midcingulate and insular cortices. Correlations between patient pairs were then compared for functional connectivity differences between groups.

But No Amygdala Connectivity Relationship

As reported at the 2017 annual meeting of the Society of Cardiovascular Anesthesiologists (abstract SCA13), after eliminating four patients with artifact-corrupted scans, a total of 12 patients were included in the analysis, six of whom developed postoperative atrial fibrillation. The analysis revealed that patients with postoperative atrial fibrillation had greater mean connectivity between the bilateral insula (R-value = 0.79+0.23 vs. 0.52+0.12; P=0.008) than those who did not. What’s more, patients with postoperative atrial fibrillation also demonstrated greater connectivity between the bilateral midcingulate cortices (R value=0.85+0.06 vs. 0.68+0.18; P=0.025).

“Interestingly, that trend mostly persists six weeks after surgery,” he continued. “So patients who had atrial fibrillation still had higher connectivity between their right insula and left midcingulate cortex, as well as between their bilateral midcingulate cortices.”

However, the investigators were surprised to find no relationship between amygdala connectivity and development of postoperative atrial fibrillation. “This may have to do with the size of the region on fMRI, or the ability to detect average changes in the actual 3-D map that we have,” Dr. Waldron said.

Although preliminary, these results open the door to many future possibilities. “It allows us to continue moving forward,” he said. “And perhaps—if this trend continues to develop and we replicate it in a larger cohort of patients—this could potentially be an impetus to look at noninvasive cerebral stimulation technologies.

“I think these data are too preliminary for us to apply them in a clinically meaningful way right now,” Dr. Waldron said.

“But I think it will hopefully be the beginning of a larger body of evidence showing that the autonomic dysfunction that underpins postoperative and ambulatory atrial fibrillation may actually generate from the brain itself, and not just be a manifestation of the intrinsic cardiac system or spinal cord level change.” The study continues to accrue patients.

Daniel T. Bainbridge, MD, associate professor and director of cardiac anesthesia at the University of Western Ontario, in London, Ontario, agreed that there is quite a bit more work to be done, as fascinating as the research may be.

He also raised some issues of practicality surrounding the technology, including the price tag of fMRI. “The other thing to consider is that whenever you have a new technology, we need to ask whether it offers good discrimination in terms of predicting the outcome. And how does it improve over previous methods of discrimination?” he asked.

“Because we don’t only want to know if this is a good way to determine atrial fibrillation, but if it adds to what we already know. If you’re spending a lot of money and it gives the same results as clinical tests, then why are we spending the money?”

Should future research elucidate these functional connectivity relationships, Dr. Bainbridge continued, the next issue would be developing relevant pharmacologic agents.

“The treatment question is not clear-cut, because we find that agents that work directly on the brain tend to impact other areas as well,” he added. “Most drugs that cross the blood–brain barrier have adverse effects, which is why we treat atrial fibrillation with peripheral sympathetic blockers.

“But the ultimate question is whether or not it really matters at all,” Dr. Bainbridge said. “Some people would argue that it doesn’t matter because if the patient gets atrial fibrillation we’ll treat it, and if the patient doesn’t get atrial fibrillation, we don’t treat it. We’re still stuck in this debate of whether prophylactic treatment for atrial fibrillation improves patient outcomes. And I don’t think that’s been quite settled yet.”