Algorithm identifies predictors of pain after spine surgery

At the World Congress on Regional Anesthesia and Pain Medicine, researchers presented a validated predictive model that identified eight key factors for early pain after spine surgery. The findings also highlighted links between the use of non-opioid analgesics and intraoperative remifentanil with postoperative pain.

“Our machine-learning approach allows us [to extract] variables, such as the intraoperative use of non-opioid analgesics, that reliably affect postoperative pain and take into account the interdependencies of these variables in a real-life clinical scenario,” Martin Angst, MD, told Healio.com/Orthopedics. “As such, our findings are robust and stand a good chance to be valid when implemented into clinical practice.”

Angst and colleagues identified 1,008 patients who underwent spine surgery under general anesthesia at hospitals that participated in the Pain-Out Initiative between 2011 and 2013. Investigators used an elastic net algorithm to develop a predictive model of postoperative pain within the first 24 hours postoperatively. Using an independent group of patients, investigators validated the performance of the model. The stability of the model was characterized with bootstrap modeling for factor selection. Modifiable perioperative factors that impacted postoperative pain were identified with linear regression analysis of key factors determined by the predictive model.

Results showed eight out of the 30 potential factors were identified as robust predictive factors using bootstrap modeling. Investigators noted stable factors included sex, preoperative pain, duration of surgery, remifentanil use, ketamine use, non-opioid intraoperative pain medication use, volatile anesthetic use and post-anesthesia care unit morphine-equivalent consumption. Linear regression analysis results showed how key factors individually contributed to postoperative pain; whereas, Bayesian network analysis showed the significant interdependency between several factors. The analysis showed postoperative pain strongly depends on the number of classes of non-opioid analgesics used during surgery and use of intraoperative remifentanil. – by Monica Jaramillo

Reference:

Baca Q, et al. Paper 5527. Presented at: World Congress on Regional Anesthesia & Pain Medicine; April 19-21, 2018; New York.

Disclosure: Angst reports no relevant financial disclosures.