Prediction of acute postoperative pain based on intraoperative nociception level (NOL) index values: the impact of machine learning-based analysis

Harnessing artificial intelligence for predicting and managing postoperative pain: a narrative literature review

PURPOSE OF REVIEW

This review examines recent research on artificial intelligence focusing on machine learning (ML) models for predicting postoperative pain outcomes. We also identify technical, ethical, and practical hurdles that demand continued investigation and research.

RECENT FINDINGS

Current ML models leverage diverse datasets, algorithmic techniques, and validation methods to identify predictive biomarkers, risk factors, and phenotypic signatures associated with increased acute and chronic postoperative pain and persistent opioid use. ML models demonstrate satisfactory performance to predict pain outcomes and their prognostic trajectories, identify modifiable risk factors and at-risk patients who benefit from targeted pain management strategies, and show promise in pain prevention applications. However, further evidence is needed to evaluate the reliability, generalizability, effectiveness, and safety of ML-driven approaches before their integration into perioperative pain management practices.

SUMMARY

Artificial intelligence (AI) has the potential to enhance perioperative pain management by providing more accurate predictive models and personalized interventions. By leveraging ML algorithms, clinicians can better identify at-risk patients and tailor treatment strategies accordingly. However, successful implementation needs to address challenges in data quality, algorithmic complexity, and ethical and practical considerations. Future research should focus on validating AI-driven interventions in clinical practice and fostering interdisciplinary collaboration to advance perioperative care.

Using the nociception level index to compare the intraoperative antinociceptive effect of propofol and sevoflurane during clinical and experimental noxious stimulus in patients under general anesthesia

STUDY

Propofol and sevoflurane are two anesthetic agents widely used to induce and maintain general anesthesia (GA). Their intrinsic antinociceptive properties remain unclear and are still debated.

OBJECTIVE

To determine whether propofol presents stronger antinociceptive properties than sevoflurane using intraoperative clinical and experimental noxious stimulations and evaluating postoperative pain outcomes.

DESIGN

A prospective randomized monocentric trial.

SETTING

Perioperative care.

PATIENTS

60 adult patients with ASA status I to III who underwent elective abdominal laparoscopic surgery under GA were randomized either in propofol or sevoflurane group to induce and maintain GA.

INTERVENTIONS

We used clinical and experimental noxious stimulations (intubation, tetanic stimulation) to assess the antinociceptive properties of propofol and sevoflurane in patients under GA and monitored using the NOL index, BIS index, heart rate, and mean arterial blood pressure.

MEASUREMENTS

We measured the difference in the NOL index alterations after intubation and tetanic stimulation during either intravenous anesthesia (propofol) or inhaled anesthesia (sevoflurane). We also intraoperatively measured the NOL index and remifentanil consumption and recorded postoperative pain scores and opioid consumption in the post-anesthesia care unit. Intraoperative management was standardized by targeting similar values of depth of anesthesia (BIS index), hemodynamic (HR and MAP), NOL index values (below the threshold of 20), same multimodal analgesia and type of surgery.

MAIN RESULTS

We found the antinociceptive properties of propofol and sevoflurane similar. The only minor difference was after tetanic stimulation: the delta NOL was higher in the sevoflurane group (39 ± 13 for the propofol group versus 47 ± 15 for sevoflurane; P = 0.04). Intraoperative and postoperative pain outcomes and opioid consumption were similar between groups.

CONCLUSIONS

Despite a precise intraoperative experimental and clinical protocol using the NOL index, propofol does not provide a higher level of antinociception during anesthesia or analgesia after surgery when compared to sevoflurane. Anesthesiologists may prefer propofol over sevoflurane to reduce PONV or anesthesia-related pollution, but not for superior antinociceptive properties.

Does Nociception Level Index-Guided Opioid Administration Reduce Intraoperative Opioid Consumption? A Systematic Review and Meta-Analysis

BACKGROUND

The nociception level (NOL) index is a quantitative parameter derived from physiological signals to measure intraoperative nociception. The aim of this systematic review and meta-analysis was to evaluate if NOL monitoring reduces intraoperative opioid use compared to conventional therapy (opioid administered at clinician discretion).

METHODS

This meta-analysis comprises randomized clinical trials comparing NOL-guided opioid administration to conventional therapy in adult patients undergoing any type of surgery. A systematic search of PubMed, Scopus, and CENTRAL databases was conducted. The primary outcome was intraoperative opioid consumption and the effect estimate of the NOL index was measured using the standardized mean difference (SMD) where 0.20 is considered a small and 0.80 a large effect size. A random-effects model with Hartung-Knapp-Sidik-Jonkman adjustment was applied to estimate the treatment effect. Heterogeneity was explored clinically and statistically (using the inconsistency I² statistic, prediction intervals, and influence analysis). The quality (certainty) of evidence was evaluated using the Grading of Recommendations, Assessment, Development, and Evaluations (GRADE) guidelines methodology.

RESULTS

This review comprised 9 trials (519 patients). The intraoperative opioid SMD (NOL monitoring versus conventional therapy) was -0.26 (95% confidence interval [CI], -0.82 to 0.30; P = .31; low certainty of evidence). We observed substantial clinical (intraoperative opioid regimens) and statistical heterogeneity with the I² statistic being 86% (95% CI, 75%-92%). The prediction interval was between -1.95 and 1.42 indicating where the SMD between NOL and conventional therapy would lie if a similar study were conducted in the future.

CONCLUSIONS

This meta-analysis does not provide evidence supporting the role of NOL monitoring in reducing intraoperative opioid consumption.

Progress in the validation of nociception monitoring in guiding intraoperative analgesic therapy

PURPOSE OF REVIEW

This article summarizes the current level of validation for several nociception monitors using a categorized validation process to facilitate the comparison of performance.

RECENT FINDINGS

Nociception monitors improve the detection of a shift in the nociception and antinociception balance during anesthesia, guiding perioperative analgesic therapy. A clear overview and comparison of the validation process for these monitors is missing.

RESULTS

Within a 2-year time-frame, we identified validation studies for four monitors [analgesia nociception index (ANI), nociception level monitor (NOL), surgical pleth index (SPI), and pupillometry]. We categorized these studies in one out of six mandatory validation steps: developmental studies, clinical validation studies, pharmacological validation studies, clinical utility studies, outcome improvement studies and economical evaluation studies. The current level of validation for most monitors is mainly focused on the first three categories, whereas ANI, NOL, and SPI advanced most in the availability of clinical utility studies and provide confirmation of a clinical outcome improvement. Analysis of economical value for public health effects is not yet publicly available for the studied monitors.

SUMMARY

This review proposes a stepwise structure for validation of new monitoring technology, which facilitates comparison between the level of validation of different devices and identifies the need for future research questions.

Predicting Chronic Pain and Treatment Outcomes Using Machine Learning Models Based on High-dimensional Clinical Data From a Large Retrospective Cohort

PURPOSE

To identify factors and indicators that affect chronic pain and pain relief, and to develop predictive models using machine learning.

METHODS

We analyzed the data of 67,028 outpatient cases and 11,310 valid samples with pain from a large retrospective cohort. We used decision tree, random forest, AdaBoost, neural network, and logistic regression to discover significant indicators and to predict pain and treatment relief.

FINDINGS

The random forest model had the highest accuracy, F1 value, precision, and recall rates for predicting pain relief. The main factors affecting pain and treatment relief included body mass index, blood pressure, age, body temperature, heart rate, pulse, and neutrophil/lymphocyte × platelet ratio. The logistic regression model had high sensitivity and specificity for predicting pain occurrence.

IMPLICATIONS

Machine learning models can be used to analyze the risk factors and predictors of chronic pain and pain relief, and to provide personalized and evidence-based pain management.

Predicting Postoperative Pain and Opioid Use with Machine Learning Applied to Longitudinal Electronic Health Record and Wearable Data

BACKGROUND

 Managing acute postoperative pain and minimizing chronic opioid use are crucial for patient recovery and long-term well-being.

OBJECTIVES

 This study explored using preoperative electronic health record (EHR) and wearable device data for machine-learning models that predict postoperative acute pain and chronic opioid use.

METHODS

 The study cohort consisted of approximately 347 All of Us Research Program participants who underwent one of eight surgical procedures and shared EHR and wearable device data. We developed four machine learning models and used the Shapley additive explanations (SHAP) technique to identify the most relevant predictors of acute pain and chronic opioid use.

RESULTS

 The stacking ensemble model achieved the highest accuracy in predicting acute pain (0.68) and chronic opioid use (0.89). The area under the curve score for severe pain versus other pain was highest (0.88) when predicting acute postoperative pain. Values of logistic regression, random forest, extreme gradient boosting, and stacking ensemble ranged from 0.74 to 0.90 when predicting postoperative chronic opioid use. Variables from wearable devices played a prominent role in predicting both outcomes.

CONCLUSION

 SHAP detection of individual risk factors for severe pain can help health care providers tailor pain management plans. Accurate prediction of postoperative chronic opioid use before surgery can help mitigate the risk for the outcomes we studied. Prediction can also reduce the chances of opioid overuse and dependence. Such mitigation can promote safer and more effective pain control for patients during their recovery.

Machine learning approach for predicting post-intubation hemodynamic instability (PIHI) index values: towards enhanced perioperative anesthesia quality and safety

BACKGROUND

Adequate preoperative evaluation of the post-intubation hemodynamic instability (PIHI) is crucial for accurate risk assessment and efficient anesthesia management. However, the incorporation of this evaluation within a predictive framework have been insufficiently addressed and executed. This study aims to developed a machine learning approach for preoperatively and precisely predicting the PIHI index values.

METHODS

In this retrospective study, the valid features were collected from 23,305 adult surgical patients at Peking Union Medical College Hospital between 2012 and 2020. Three hemodynamic response sequences including systolic pressure, diastolic pressure and heart rate, were utilized to design the post-intubation hemodynamic instability (PIHI) index by computing the integrated coefficient of variation (ICV) values. Different types of machine learning models were constructed to predict the ICV values, leveraging preoperative patient information and initiatory drug infusion. The models were trained and cross-validated based on balanced data using the SMOTETomek technique, and their performance was evaluated according to the mean absolute error (MAE), root mean square error (RMSE), mean absolute percentage error (MAPE) and R-squared index (R2).

RESULTS

The ICV values were proved to be consistent with the anesthetists' ratings with Spearman correlation coefficient of 0.877 (P < 0.001), affirming its capability to effectively capture the PIHI variations. The extra tree regression model outperformed the other models in predicting the ICV values with the smallest MAE (0.0512, 95% CI: 0.0511-0.0513), RMSE (0.0792, 95% CI: 0.0790-0.0794), and MAPE (0.2086, 95% CI: 0.2077-0.2095) and the largest R2 (0.9047, 95% CI: 0.9043-0.9052). It was found that the features of age and preoperative hemodynamic status were the most important features for accurately predicting the ICV values.

CONCLUSIONS

Our results demonstrate the potential of the machine learning approach in predicting PIHI index values, thereby preoperatively informing anesthetists the possible anesthetic risk and enabling the implementation of individualized and precise anesthesia interventions.

The role of surgical tissue injury and intraoperative sympathetic activation in postoperative immunosuppression after breast-conserving surgery versus mastectomy: a prospective observational study

BACKGROUND

Breast cancer is the second most common cause of death from cancer in women worldwide. Counterintuitively, large population-based retrospective trials report better survival after breast-conserving surgery (BCS) compared to mastectomy, corrected for tumour- and patient variables. More extensive surgical tissue injury and activation of the sympathetic nervous system by nociceptive stimuli are associated with immune suppression. We hypothesized that mastectomy causes a higher expression of plasma damage associated molecular patterns (DAMPs) and more intraoperative sympathetic activation which induce postoperative immune dysregulation. Immune suppression can lead to postoperative complications and affect tumour-free survival.

METHODS

In this prospective observational study, plasma DAMPs (HMGB1, HSP70, S100A8/A9 and S100A12), intraoperative sympathetic activation (Nociception Level (NOL) index from 0 to 100), and postoperative immune function (plasma cytokine concentrations and ex vivo cytokine production capacity) were compared in patients undergoing elective BCS (n = 20) versus mastectomy (n = 20).

RESULTS

Ex vivo cytokine production capacity of TNF, IL-6 and IL-1β was nearly absent in both groups one hour after surgery. Levels appeared recovered on postoperative day 3 (POD3), with significantly higher ex vivo production capacity of IL-1β after BCS (p = .041) compared to mastectomy. Plasma concentration of IL-6 was higher one hour after mastectomy (p = .045). Concentrations of plasma alarmins S100A8/A9 and S100A12 were significantly higher on POD3 after mastectomy (p = .003 and p = .041, respectively). Regression analysis showed a significantly lower percentage of NOL measurements ≤ 8 (absence of nociception) during mastectomy when corrected for norepinephrine equivalents (36% versus 45% respectively, p = .038). Percentage of NOL measurements ≤ 8 of all patients correlated with ex vivo cytokine production capacity of IL-1β and TNF on POD3 (r = .408; p = .011 and r = .500; p = .001, respectively).

CONCLUSIONS

This pilot study revealed substantial early postoperative immune suppression after BCS and mastectomy that appears to recover in the following days. Differences between BCS and mastectomy in release of DAMPs and intraoperative sympathetic activation could affect postoperative immune homeostasis and thereby contribute to the better survival reported after BCS in previous large population-based retrospective trials. These results endorse further exploration of (1) S100 alarmins as potential therapeutic targets in breast cancer surgery and (2) suppression of intraoperative sympathetic activation to substantiate the observed association with postoperative immune dysregulation.

Depth of Anesthesia and Nociception Monitoring: Current State and Vision For 2050

Anesthesia objectives have evolved into combining hypnosis, amnesia, analgesia, paralysis, and suppression of the sympathetic autonomic nervous system. Technological improvements have led to new monitoring strategies, aimed at translating a qualitative physiological state into quantitative metrics, but the optimal strategies for depth of anesthesia (DoA) and analgesia monitoring continue to stimulate debate. Historically, DoA monitoring used patient's movement as a surrogate of awareness. Pharmacokinetic models and metrics, including minimum alveolar concentration for inhaled anesthetics and target-controlled infusion models for intravenous anesthesia, provided further insights to clinicians, but electroencephalography and its derivatives (processed EEG; pEEG) offer the potential for personalization of anesthesia care. Current studies appear to affirm that pEEG monitoring decreases the quantity of anesthetics administered, diminishes postanesthesia care unit duration, and may reduce the occurrence of postoperative delirium (notwithstanding the difficulties of defining this condition). Major trials are underway to further elucidate the impact on postoperative cognitive dysfunction. In this manuscript, we discuss the Bispectral (BIS) index, Narcotrend monitor, Patient State Index, entropy-based monitoring, and Neurosense monitor, as well as middle latency evoked auditory potential, before exploring how these technologies could evolve in the upcoming years. In contrast to developments in pEEG monitors, nociception monitors remain by comparison underdeveloped and underutilized. Just as with anesthetic agents, excessive analgesia can lead to harmful side effects, whereas inadequate analgesia is associated with increased stress response, poorer hemodynamic conditions and coagulation, metabolic, and immune system dysregulation. Broadly, 3 distinct monitoring strategies have emerged: motor reflex, central nervous system, and autonomic nervous system monitoring. Generally, nociceptive monitors outperform basic clinical vital sign monitoring in reducing perioperative opioid use. This manuscript describes pupillometry, surgical pleth index, analgesia nociception index, and nociception level index, and suggest how future developments could impact their use. The final section of this review explores the profound implications of future monitoring technologies on anesthesiology practice and envisages 3 transformative scenarios: helping in creation of an optimal analgesic drug, the advent of bidirectional neuron-microelectronic interfaces, and the synergistic combination of hypnosis and virtual reality.

A novel non-invasive nociceptive monitoring approach fit for intracerebral surgery: a retrospective analysis

Background

Measuring depth of anesthesia during intracerebral surgery is an important task to guarantee patient safety, especially while the patient is fixated in a Mayfield-clamp. Processed electro-encephalography measurements have been established to monitor deep sedation. However, visualizing nociception has not been possible until recently and has not been evaluated for the neurosurgical setting. In this single-center, retrospective observational analysis, we routinely collected the nociceptive data via a nociception level monitor (NOL®) of 40 patients undergoing intracerebral tumor resection and aimed to determine if this monitoring technique is feasible and delivers relevant values to potentially base therapeutic decisions on.

Methods

Forty patients (age 56 ± 18 years) received total intravenous anesthesia and were non-invasively connected to the NOL® via a finger clip as well as a bispectral-index monitoring (BIS®) to confirm deep sedation. The measured nociception levels were retrospectively evaluated at specific time points of nociceptive stress (intubation, Mayfield-positioning, incision, extubation) and compared to standard vital signs.

Results

Nociceptive measurements were successfully performed in 35 patients. The largest increase in nociceptive stimulation occurred during intubation (NOL® 40 ± 16) followed by Mayfield positioning (NOL® 39 ± 16) and incision (NOL® 26 ± 12). Correlation with BIS measurements confirmed a sufficiently deep sedation during all analyzed time points (BIS 45 ± 13). Overall, patients showed an intraoperative NOL® score of 10 or less in 56% of total intervention time.

Conclusions

Nociceptive monitoring using the NOL® system during intracerebral surgery is feasible and might yield helpful information to support therapeutic decisions. This could help to reduce hyperanalgesia, facilitating shorter emergence periods and less postoperative complications. Prospective clinical studies are needed to further examine the potential benefits of this monitoring approach in a neurosurgical context.

Trial registration

German trial registry, registration number DRKS00029120.