New models quantify how ‘nociception’ could help improve surgical pain management

The extent to which subconscious pain processing, or “nociception,” is properly managed by the anesthesiologist will have a direct bearing on the degree of postoperative side effects from medications he or she will experience and the need for additional pain management. But pain is a subjective sensation to measure even when patients are awake, let alone unconscious.

In a new study, researchers from MIT and Massachusetts General Hospital (MGH) describe a set of statistical models that objectively quantify nociception during surgery. Ultimately, they hope to help anesthesiologists optimize drug dosing and minimize postoperative pain and side effects.

The results are published in the journal Proceedings of the National Academy of Sciences.

The new models incorporate meticulously recorded data from 18,582 minutes of 101 abdominal surgeries on men and women at MGH. Led by former MIT graduate student Sandya Subramanian, now an assistant professor at UC Berkeley and UC San Francisco, the researchers collected and analyzed data from five physiological sensors while patients underwent a total of 49,878 separate “nociceptive stimuli” (such as incisions or cauterization).

The team also recorded the drugs administered, the amount and the timing of administration, to account for their effects on nociception or cardiovascular parameters. They then used all the data to develop a set of statistical models that worked well to retrospectively indicate the body’s response to nociceptive stimuli.

The team’s goal is to provide accurate, objective, physiologically-based information in real time to anesthesiologists who currently must rely heavily on intuition and past experience to decide how to administer pain medications during surgery.

If anesthesiologists administer too much of the drug, patients may experience side effects ranging from nausea to delirium. If they administer too little of the drug, patients may experience excessive pain upon awakening.

“Sandya’s work has helped us establish a principled method for understanding and measuring nociception (unconscious pain) during general anesthesia,” said the study’s senior author, Emery N. Brown, the Edward Hood Taplin Professor of Medical Engineering and Computational Neuroscience in the Picower Institute for Learning and Memory, the Institute for Medical Engineering and Science, and the Department of Brain and Cognitive Sciences at MIT. Brown is also an anesthesiologist at MGH and a professor at Harvard Medical School.

“Our next goal is to make the knowledge gained from Sandya’s studies reliable and practical so that anesthesiologists can use it during surgical procedures.”

Surgery and statistics

The research began as Subramanian’s doctoral dissertation project in Brown’s lab in 2017. The best previous attempts to objectively model nociception relied solely on the electrocardiogram (ECG, a proxy for heart rate variability) or other systems that could integrate more than one measure, but were either based on laboratory experiments using pain stimuli that were not comparable in intensity to surgical pain or validated by statistically aggregating only a few time points across multiple patient surgeries, Subramanian said.

“There is no other place to study surgical pain than the operating room,” Subramanian said.

“We wanted to not only develop algorithms from data from surgery, but also validate them in the context in which we want someone to use them. If we’re asking them to track nociception moment by moment during surgery, we need to validate it in the same way.”

So she and Brown worked to advance the state of the art by collecting multi-sensor data throughout surgeries and accounting for the confounding effects of the drugs being administered. They hoped to develop a model that could make accurate predictions that would hold true for the same patient throughout the operation.

Part of the team’s improvements came from tracking heart rate and skin conductance patterns. Changes in both of these physiological factors can be indications of the body’s primary “fight or flight” response to nociception or pain, but some medications used during surgery directly affect cardiovascular status, while skin conductance (or “EDA,” electrodermal activity) remains unchanged.

The study not only measures the ECG but also supplements it with PPG, an optical measure of heart rate (like the oxygen sensor in a smart watch), because ECG signals can sometimes be made noisy by all the electrical equipment buzzing around the operating room.

Similarly, Subramanian supplemented the EDA measurements with skin temperature measurements to ensure that changes in skin conductance due to sweating were due to nociception and not simply the patient’s temperature being too high. The study also tracked breathing.

The authors then performed statistical analyses to develop physiologically relevant indices from each of the cardiovascular and skin conductance signals. Once each index was established, further statistical analyses allowed them to be grouped together to produce models capable of making accurate, principled predictions about when nociception occurs and how the body responds.

Mastering nociception

In four versions of the model, Subramanian “supervised” them by providing them with information about when noxious stimuli occurred so that they could then learn the association between physiological measures and the incidence of pain-provoking events. In some of these trained versions, she left out the drug information, and in other versions, she used different statistical approaches (either “linear regression” or “random forest”).

In a fifth version of the model, based on a “state space” approach, she left it unsupervised, meaning it had to learn to infer moments of nociception based solely on physiological cues. She compared all five versions of her model to one of the current industry standards, an ECG-tracking model called ANI.

The results of each model can be visualized as a graph plotting the predicted nociception level over time. ANI is slightly better than chance but is implemented in real time. The unsupervised model performed better than ANI, but not as well as the supervised models.

The best-performing model was the one that incorporated drug information and used a “random forest” approach. Still, the authors note, the fact that the unsupervised model performed significantly better than the random model suggests that there is indeed an objectively detectable signature of the body’s nociceptive state even when looking at different patients.

“A state-space framework using multisensory physiological observations is effective in uncovering this implicit nociceptive state with a consistent definition across multiple subjects,” Subramanian, Brown, and their co-authors wrote.

“This is an important step toward defining a metric to track nociception without including nociceptive ‘ground truth’ information, that is most practical for scalability and implementation in clinical settings.”

Indeed, the next steps in the research are to increase the data sample and further refine the models so that they can eventually be put into practice in the operating room. This will require enabling them to predict nociception in real time, rather than as part of a post-hoc analysis. Once this is achieved, anesthesiologists or intensivists will be able to inform their decisions about analgesic dosing.

In the future, the model could inform closed-loop systems that automatically dose drugs under the supervision of the anesthesiologist.

“Our study is an important first step toward the development of objective markers to monitor surgical nociception,” the authors conclude.

“These markers will enable objective assessment of nociception in other complex clinical settings, such as the intensive care unit, and will catalyze the future development of closed-loop control systems for nociception.”

This article is republished with kind permission from MIT News (web.mit.edu/newsoffice/), a popular site covering the latest research, innovation, and teaching at MIT.