Systems and methods for monitoring nociception and analgesia during administration of general anesthesia - Patents.com

Abstract

Systems and methods for patient monitoring and / or treatment are provided. The system may include one or more sensors configured to measure electroencephalogram (EEG) signals and electrodermal characteristic signals of a patient administered at least one anesthetic agent and at least one analgesic agent during a medical surgical procedure. The system may further include a processor operatively coupled to the one or more sensors. The processor receives the EEG signals and the electrodermal characteristic signals and uses the EEG signals or the electrodermal characteristic signals to generate a report indicative of a nociceptive state of the patient in real time during the medical surgical procedure while the patient is administered the at least one anesthetic agent and is administered a pain management plan based on a target post-operative outcome.

Description

[Technical field]

[0001] CROSS-REFERENCE TO RELATED APPLICATIONS This application is based on and claims priority to U.S. Provisional Patent Application No. 63 / 320,535 (filed March 16, 2022), the entire contents of which are incorporated herein by reference for all purposes. [Background technology]

[0002] TECHNICAL FIELD The present disclosure relates generally to the medical field, and more specifically to the field of medical treatment planning and / or patient monitoring.

[0003] Opioid-related overdoses cause tens of thousands of deaths each year, making it one of the largest unsolved public health problems in the United States. Surgery is one of the major risk factors for developing opioid dependence, as most surgical patients are prescribed opioids to manage postoperative pain. Postoperative pain is greatly influenced by how well intraoperative nociception is controlled during surgery and general anesthesia. Therefore, effective control of nociception during general anesthesia and surgery is important to reduce postoperative opioid requirements while relieving postoperative pain.

[0004] Surgical nociception is typically managed by administering opioid analgesics, but these drugs must be administered with care. Excessive opioid administration can cause respiratory depression and oversedation, which can increase postoperative hospital stays, and can also lead to central sensitization, which can increase downstream opioid requirements. On the other hand, ineffective control of surgical nociception can increase postoperative pain, which can increase postoperative opioid requirements. Currently, anesthesiologists rely on changes in heart rate (HR) and blood pressure (BP) to understand nociception and titrate opioids and other medications to minimize these changes. Unfortunately, HR and BP are affected by numerous intraoperative factors, including blood loss, anesthetic drugs, and antihypertensive drugs, which make them unreliable indicators of nociception. Thus, there is a clear need for improved methods to monitor surgical nociception during general anesthesia. Fortunately, HR and BP are not the only physiological variables that can be used to monitor and track nociception. Electrodermal activity (EDA) also changes through autonomic mechanisms in response to nociceptive stimuli, and this EDA is also known as skin conductance response (SCR). Electroencephalography (EEG) fluctuations can also track arousal and nociception. Recently, our research team has identified a novel and robust EEG signature that is specific to opioid drugs, which may be used to monitor the effects of opioid drugs differently from other anesthetics.

[0005] Up to 10% of surgical patients develop opioid dependence after surgery. Poor control of acute postoperative pain can lead to chronic pain, which may require long-term opioid use to control. Opioids are the first-line treatment for acute postoperative pain, but are highly addictive. Anesthesiologists, intensivists, and nurses have no tools to help monitor patients' nociception during surgery and / or intensive care, nor do they have any tools to help monitor the effectiveness of opioid analgesics administered to treat nociception, leading to oversedation and suboptimal postoperative pain outcomes. Summary of the Invention

[0006] The present disclosure addresses the above deficiencies by describing systems and methods for monitoring patient parameters during a medical procedure, e.g., intraoperatively, and / or for patient treatment planning to reduce the potential for unwanted risks associated with any procedure or post-procedure care, such as the potential for unwanted risks when administering opioid drugs. In one non-limiting example, a system and method is provided for patient planning and / or for monitoring patients to better understand nociception and develop plans for managing post-operative use of drugs and other procedures, such as administration of opioids. In one non-limiting example, a system may be used or configured to acquire patient information during general anesthesia and / or sedation and identify from the information signatures that are highly correlated with opioid drug concentrations and applicable to titration of opioids, independent of reliance on sedative-hypnotic drugs. Such signatures may be used along with other physiological features to monitor nociception and analgesia during anesthesia and intensive care. Other physiological characteristics can be combined with monitoring indices that can be used to improve postoperative pain and opioid requirements. Surgical nociceptive monitors can be used to reduce postoperative opioid requirements. In another non-limiting example, an anesthesiologist is provided with a monitor that allows the clinician to create both a surgical and / or postoperative plan that minimizes postoperative pain and postoperative opioid requirements by integrated measures of nociceptive control based on neurophysiological markers indicative of arousal and nociception (using EEG as one non-limiting example) and autonomic markers (using EDA as one non-limiting example).

[0007] In one aspect of the present disclosure, an intra-operative patient monitoring system is described, the system comprising one or more sensors configured to measure electroencephalogram (EEG) signals and electrodermal characteristic signals of a patient administered at least one anesthetic agent and at least one analgesic agent during a medical surgical procedure, and a processor operatively coupled to the one or more sensors, the processor configured to receive the EEG signals and electrodermal characteristic signals, monitor a nociceptive state of the patient during the medical surgical procedure in real time using the EEG signals or the electrodermal characteristic signals, and generate a post-operative pain management plan using at least one of the nociceptive state of the patient during the medical surgical procedure and the at least one analgesic agent administered to the patient for the medical surgical procedure.

[0008] Another aspect of the present disclosure describes an intra-operative patient monitoring system comprising one or more sensors configured to measure electroencephalogram (EEG) signals and electrodermal characteristic signals of a patient administered at least one anesthetic agent and at least one analgesic agent during a medical surgical procedure, and a processor operatively coupled to the one or more sensors, wherein the processor is configured to receive the EEG signals and the electrodermal characteristic signals and use the EEG signals or the electrodermal characteristic signals to generate a report indicative of a nociceptive state of the patient in real time while the patient is administered the at least one anesthetic agent during the medical surgical procedure.

[0009] Some embodiments of the present disclosure will be described below with reference to the accompanying drawings. When a person skilled in the art reads the following description together with the drawings, it will be clear how some embodiments of the present disclosure can be implemented. The drawings are for detailed explanation and are not intended to show the details of the structure of the embodiments in more detail than is necessary for a basic understanding of the teachings of the present disclosure. [Brief description of the drawings]

[0010] [Figure 1]FIG. 1 illustrates a closed-loop monitoring and control system of the present disclosure. [Diagram 2] (A) Predicted effect site concentration (ESC) for a representative subject; (B) Correspondence between fentanyl exposure and changes in the spectrogram; (C) Correspondence between fentanyl exposure and power spectrum changes for a representative subject, with significant increases in theta band (4-8 Hz) and slow wave / delta band (0-4 Hz). Increases in theta power correspond to increases in fentanyl concentration; (D) Mixed effects model constructed for subjects with 95% CI, detailing the relationship between changes in theta power and fentanyl concentration. [Diagram 3] (A) shows the appearance of theta oscillations (4-8 Hz) during loss of consciousness, which are specific to fentanyl. Theta signals clearly decrease during intubation, suggesting that they arise in response to noxious stimuli. (B) is a spectrogram showing distinct changes during loss of consciousness at 25 minutes and during intubation at 28 minutes. [Figure 4] (A) Fentanyl concentrations estimated using the Pk / Pd model, (B) θ band power correlates closely with reaction time across protocols, (C) K-fold cross validation predicts prediction error comparing theta power and Pk / Pd models and the combined model. [Diagram 5] (A) Traditional spectral analysis and state-space methods showing EEG signatures in the slow-wave, theta, and alpha bands during general anesthesia with sevoflurane and fentanyl. (B) AIC analysis results using a two-oscillator state-space model and a three-oscillator state-space model, which show that the three-oscillator model consistently outperforms the two-oscillator model, suggesting that the fentanyl theta component can be detected and extracted separately from the background of sevoflurane-derived slow-wave and alpha oscillations. [Figure 6A]FIG. 1 shows an EEG spectrogram during surgery under general anesthesia, with annotations indicating when propofol is administered and when lidocaine infusion is stopped. [Figure 6B] FIG. 1 shows that EDASCR is clearly decreased following one bolus of propofol, consistent with the presence of an antinociceptive agent. [Figure 6C] FIG. 13 shows that after surgery, the lidocaine infusion to provide analgesia is stopped and the patient remains unconscious, but EDASCR is significantly increased, suggesting elevated nociception. [Figure 6D] Figure 1 shows the evaluation of epochs specific to potential nociceptive stimuli using event records obtained from the patient's anesthesia record. Nociceptive epochs showed increased SCR compared to baseline, and this pattern was consistent throughout the duration of general anesthesia, despite reduced SCR levels likely due to the presence of anesthetic agents. [Figure 7] (A) shows an EEG spectrogram where a patient has consistent alpha oscillations associated with unconsciousness (left) versus a disruption in alpha oscillations associated with nociceptive / arousal states (right).(B) shows that alpha oscillations in a patient's EEG recording can be extracted using the state space model described above, and that components thereof, including the illustrated amplitude or envelope, or frequency or phase of the oscillations, can be used to further explore the disruption in various stimuli. [Figure 8] (A) shows how alpha power is observed to decrease with increasing SCR when a lumbar puncture is performed, (B) is a spectrogram showing alpha power fluctuations alongside the SCR response, and (C) is a cross-correlation plot, which reveals a similar correspondence between alpha power and SCR across nociceptive events and across subjects (mean=0.5135, SD=0.0558). [Figure 9]FIG. 1 illustrates an overview of prospective data collection for models, algorithms and index trends. [Figure 10] FIG. 1 illustrates a schematic representation of an exemplary approach to developing an operating room monitor as a comparison to the approach of aspects of the present disclosure for developing algorithms for optimizing and calibrating intraoperative monitoring variables for post-operative outcomes. [Figure 11] (A) Mean intraoperative fentanyl concentrations in three representative patients, and (B) median intraoperative theta EEG power in the same three representative patients. [Figure 12] (A) shows the total amount of opioids required in the first 24 hours after surgery for the same three representative patients, and (B) shows the maximum pain experienced by the same three representative patients in the first 24 hours after surgery. [Figure 13] FIG. 1 shows median perioperative skin conductance values ​​for the same representative three patients. [Figure 14] (A) shows the perioperative EEG alpha oscillation amplitude variability for the same three representative patients, and (B) shows the perioperative EEG alpha oscillation instantaneous frequency variability for the same three representative patients. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0011] The following describes a system and method for monitoring nociception and analgesia using a combination of physiological measurements including electroencephalography (EEG) and electrodermal activity (EDA). In one non-limiting example, separate sensors are used to obtain EEG and EDA data. In one embodiment, one or more EEG sensors are applied to the patient's scalp. Alternatively, one or more EEG sensors are applied to the patient's forehead. In one non-limiting example, one or more EDA sensors are applied to the surface of the patient's skin, such as the palm of the hand. In one embodiment, one or more EDA sensors are placed on the patient's forehead. In another embodiment, one sensor can be used to obtain EEG and EDA measurements. In one embodiment, the one sensor for measuring EEG and EDA is placed on the patient's forehead.

[0012] These embodiments employ novel methods for extracting information from the signals regarding the patient's autonomic response to nociception, the patient's cerebral response to nociception, and the patient's pharmacological response to opioid analgesics.

[0013] EEG-based anesthesia monitors have been commercially available on the market for several decades, but existing devices focus on monitoring the patient's level of consciousness. However, monitoring nociception (i.e., the physiological response to receiving noxious stimuli) has not been possible with existing technology. Some advantages of each aspect of the present invention include: 1) processing using novel EEG signatures of opioid drugs; 2) processing using novel sensors corresponding to electrodermal activity; 3) processing using novel signatures of cerebral responses to nociception; 4) integrating all of these features into one quantitative system for predicting postoperative outcomes of interest, including postoperative pain and opioid consumption; and 5) calibrating this integrated monitoring capability to minimize postoperative pain and opioid consumption.

[0014] We conducted three studies to support the above concept: 1) a prospective laboratory study to discover opioid EEG signatures in n=25 human patients, 2) a retrospective analysis of clinical EEG data screening thousands of surgeries under general anesthesia, and 3) a retrospective cohort study including tens of thousands of cases examining the relationship between patient baseline variables, intraoperative opioid administration, postoperative pain, and opioid outcomes.

[0015] Referring to FIG. 1, an example system 110 of the present invention is shown. The system 110 includes a patient monitoring device 112, such as a physiological monitoring device, which may include an electroencephalogram (EEG) sensor array. However, the patient monitoring device 112 may include other monitoring systems, such as mechanisms for monitoring electrodermal activity (EDA), such as mechanisms for measuring arousal to external stimuli, or cardiovascular monitors, including electrocardiogram and blood pressure monitors. In a non-limiting example, the EEG and EDA electrodes are integrated into the same sensor of the patient monitoring device 112. Alternatively, the EEG and EDA sensor arrays may be separate. In a preferred embodiment, the patient monitoring device 112 is placed on the surface of the patient's forehead. In another alternative embodiment, the EEG sensor array may be placed on the patient's scalp or forehead, and the EDA sensor array may be placed on the patient's palm.

[0016] Patient monitoring device 112 is communicatively coupled to monitoring system 116 via cable 114. Wireless connections between these components may be provided in place of cable 114 or similar connections. As shown, monitoring system 116 may also be coupled to a dedicated analysis system 118. Alternatively, monitoring system 116 and analysis system 118 may be integrated.

[0017] In one non-limiting example, monitoring system 116 may be configured to receive and combine raw signals acquired by the combined array of EEG and EDA sensors described above and display the raw signals as EEG waveforms, EEG spectrograms, and / or EDA skin conductance response (SCR) data. Correspondingly, analysis system 118 may receive EEG waveforms from monitoring system 116, analyze the EEG waveforms and signatures contained therein based on a selected anesthetic compound as described below, determine a patient's condition based on the analyzed EEG waveforms and signatures, and generate a report, which may be generated, for example, as a printed report or, preferably, as a real-time display of signature information and determined condition. However, the functions of monitoring system 116 and analysis system 118 may be combined into one and the same system.

[0018] In one non-limiting example, analysis system 118 can further generate a post-operative pain management plan based on signals received from patient monitoring devices. As described in more detail below, effect site concentration (ESC) signatures of certain analgesics (e.g., opioids) can provide information regarding the patient's nociceptive state. Furthermore, changes in nociceptive state can be detected by changes in EDA, which arise due to neurally mediated effects on sweat gland permeability, which can be observed as changes in skin resistance, resulting in small electrical currents, or as potential differences between different parts of the skin.

[0019] Alternatively, in one non-limiting example, intraoperative plans can be generated by a system calibrated to target and optimize postoperative outcomes, including, but not limited to, postoperative pain, opioid requirements, cognitive recovery time, and respiratory depression. The system can be configured to characterize the relationship between intraoperative opioid administration and postoperative pain and opioid requirements. For example, the relationship can be expressed by increasing intraoperative opioid administration and decreasing postoperative pain, decreasing opioid requirements, less respiratory depression, and / or decreasing hospital stay.

[0020] In one non-limiting example, Table 1 provides a non-exclusive list of postoperative pain management outcomes that may include maximum pain score during post-anesthesia care unit (PACU) stay, cumulative opioid dose administered during PACU stay, frequency of uncontrolled pain at 24 hours, new cases of chronic pain diagnosis between 3 months and 1 year, total opioids administered at 24 hours during hospitalization, opioid prescriptions at 30, 90, and 180 days postoperatively, frequency of new persistent opioid use at 90 and 180 days, maximum pain score during the first 24 hours of hospitalization, incidence of opioid-related complications in the PACU (postoperative nausea and vomiting (PONV), sedation, and respiratory depression), length of stay (LOS) in PACU, length of hospital stay, 30-day readmission rate, and 30-day mortality.

[0021] Table 1 shows the predicted effect of adding 100 mcg fentanyl or 500 mcg hydromorphone to the observed intraoperative exposure of each patient in our study population. These analyses suggest that a 100 mcg increase in intraoperative fentanyl corresponds to a mean reduction in maximum PACU pain score of 0.26 points, and a 500 mcg increase in intraoperative hydromorphone corresponds to a mean reduction in maximum PACU pain score of 0.12 points.

[0022] For postoperative opioid administration, a 100 mcg increase in fentanyl intraoperatively was associated with a mean reduction in postoperative opioid morphine milligram equivalents (MMEs) of 0.44 in the PACU (-16.0%), 3.2 at 24 hours (-29.3%), and 6.6 during hospitalization (-14.5%). Conversely, a 500 mcg increase in hydromorphone intraoperatively was associated with a mean reduction in opioid MMEs of 0.26 in the PACU (-9.3%), 1.9 at 24 hours (-17.7%), and 1 during hospitalization (-2.2%).

[0023] A 100 mcg increase in intraoperative fentanyl was equivalent to a reduction in length of stay of 8.2 hours (-12.5%), whereas a 500 mcg increase in intraoperative hydromorphone was equivalent to an increase in length of stay of 4.2 hours (+6.3%).With regard to opioid prescriptions, a 100 mcg increase in intraoperative fentanyl was equivalent to 20.7 fewer prescriptions per 1000 at 30 days (-8.3%), 22.6 fewer prescriptions per 1000 at 90 days (-8.6%), and 23.1 fewer prescriptions per 1000 at 180 days (-8.3%), in addition to a 16.9 fewer prescriptions per 1000 persistent opioid use (10.2%). A 500 mcg increase in intraoperative hydromorphone was equivalent to an increase of 11.3 opioid prescriptions per 1000 at 30 days (+4.5%), an increase of 11.3 opioid prescriptions per 1000 at 90 days (+4.3%), an increase of 11.2 opioid prescriptions per 1000 at 180 days (+4.0%), and an increase of 4.3 opioid prescriptions per 1000 at 180 days (+2.6%), plus persistent use. [Table 1] Table 1 Counterfactual predictions For non-pain score outcomes, predicted change and 95% confidence intervals are also reported as percentages. PACU: Post-anesthesia care unit; MME: Morphine milligram equivalents; PONV: Post-operative nausea and vomiting.

[0024] In one non-limiting example, the system may be configured to select or identify at least one of the most highly correlated postoperative pain and opioid requirement variables. The system may be further calibrated by constraining the postoperative pain and opioid requirement outcomes to acceptable ranges. The system may also be refined by subspaces and ranges of modifiable physiological variables that result in acceptable postoperative outcomes.

[0025] As described below, system 110 may also include a drug delivery system 120. Drug delivery system 120 may be coupled to analysis system 118 and monitoring system 116 such that system 110 is a closed-loop monitoring and control system. As described below, such a closed-loop monitoring and control system of the present invention is capable of a wide range of operations, but also includes a user interface 122 that allows a user to configure the closed-loop monitoring and control system, receive feedback from the closed-loop monitoring and control system, and, if necessary, reconfigure and / or override the closed-loop monitoring and control system.

[0026] In one non-limiting example, the drug delivery system 120 may include multiple subsystems dedicated to administering any one of anesthetics, analgesics (opioid and non-opioid), vasoconstrictors, antihypertensives, opioid antagonists, analgesic and anesthetic adjuvants, or combinations thereof, and may also employ CYP3A4 inducers and inhibitors that can alter the dose-response characteristics of any of the above therapeutic agents.

[0027] In one non-limiting example, the anesthetic agent can include nitrous oxide (laughing gas), propofol, desflurane, isoflurane, and sevoflurane.

[0028] In one non-limiting example, opioid analgesics can include fentanyl, hydromorphone, morphine, methadone, oxycodone, meperidine, remifentanil, codeine, hydrocodone, oxymorphone, sufentanil, alfentanil, nalbuphine, buprenorphine, butorphanol, levorphanol, pentazocine, tramadol, tapentadol, dihydrocodeine, opium, and opium benzoate.

[0029] In one non-limiting example, the non-opioid analgesic is aspirin, celecoxib, diclofenac, diflunisal, etodolac, fenoprofen, flurbiprofen, ibuprofen, indomethacin, ketorolac, ketoprofen, magnesium salicylate, meclofenamic acid, mefenamic acid, meloxicam, nabumetone, naproxen, oxaprozin, piroxicam, salsalate, sulindac, tolmetin, acetaminophen, gabapentin, pregabalin, carbamazepine, oxcarbamazepine, valproic acid, topiramate, dexamethasone, progesterone, riboflavin ... These can include rednisone, amitriptyline, nortriptyline, doxepin, clomipramine, duloxetine, venlafaxine, milnacipran, desvenlafaxine, lamotrigine, cyclobenzaprine, methocarbamol, baclofen, tizanidine, clonidine, propranolol, verapamil, almotriptan, eletriptan, frovatriptan, naratriptan, rizatriptan, sumatriptan, zolmitriptan, ketamine, lidocaine, pamidronate, zoledronic acid, denosumab, capsaicin, and diclofenac.

[0030] In one non-limiting example, vasoconstrictors can include dopamine (100 mcg / kg / min), ephedrine, epinephrine (1 mcg / kg / min), norepinephrine (1 mcg / kg / min), phenylephrine (10 mcg / kg / min), and vasopressin (0.4 mcg / kg / min).

[0031] In one non-limiting example, CYP3A4 inducers can include apalutamide, carbamazepine, enzalutamide, fosphenytoin, lumacaftor, lumacaftor-ivacaftor, mitotane, phenobarbital, phenytoin, primidone, rifampin, rifampicin, bexarotene, bosentan, cenobamate, dabrafenib, dexamethasone, efavirenz, elagolix, eslicarbazepine, etravirine, lorlatinib, modafinil, nafcillin, pexidartinib, rifabutin, rifapentine, St. John's wort, nevirapine, and griseofulvin.

[0032] In one non-limiting example, a CYP3A4 inhibitor is atazanavir, ceritinib, clarithromycin, cobicistat, darunavir, idelalisib, indinavir, itraconazole, ketoconazole, lonafarnib, lopinavir, mifepristone, nefazodone, nelfinavir, ombitasvir-paritaprevir-ritonavir, ombitasvir-paritaprevir-ritonavir-dasabuvir, posaconazole, ritonavir, saquinavir, tucatinib, voriconazole, amiodarone, These can include aprepitant, berotralstat, cimetidine, conivaptan, crizotinib, cyclosporine, diltiazem, duvelisib, dronedarone, erythromycin, fedratinib, fluconazole, fosamprenavir, fosaprepitant, fosnetupitant-palonosetron, imatinib, isavuconazole, isavuconazonium sulfate, lefamulin, letermovir, netupitant, nilotinib, ribociclib, verapamil, and metronidazole.

[0033] In one non-limiting example, antihypertensive drugs can include esmolol, metoprolol, propranolol, labetalol, nicardipine, clevidipine, hydralazine, nitroglycerin (glyceryl trinitrate), nitroprusside, fenoldopam, verapamil, and diltiazem.

[0034] In one non-limiting example, opioid antagonists can include Naloxone (Naxolone), Naltrexone, Methylnaltrexone, and Arimopan (Alyimopam).

[0035] In one non-limiting example, analgesic and anesthetic adjuvants can include diclofenac, ibuprofen, indomethacin, ketorolac, meloxicam, acetaminophen, lidocaine, ketamine, dexmedetomidine, esmolol, magnesium (magnesium sulfate), and dexamethasone.

[0036] As a non-limiting example, fentanyl administration and EEG signatures are shown in Figures 2A-2D. Subjects (n=25) were administered up to 3 boluses at 2-minute intervals after the baseline phase, and concentration levels were calculated by pharmacokinetic-pharmacodynamic (Pk / Pd) modeling. Figure 2B shows the change in the EEG signal due to fentanyl ESC. The effect site concentration (ESC) of fentanyl was estimated by pharmacokinetic-pharmacodynamic (PK / PD) modeling, and Figure 2A shows the effect site concentration (ESC) of fentanyl for a representative subject. EEG can be analyzed using multitaper spectral analysis. In the spectrogram of a representative subject in Figure 2B, it can be seen how the EEG power in the theta band (4-8 Hz) and slow wave / delta band (0-4 Hz) increases as the fentanyl concentration increases. The same oscillations can be seen in the spectra shown at different time points in Figure 2C. Figure 2D shows a strong association between EEG and fentanyl concentrations for all n = 25 subjects [slope: 0.55 (CI: 0.25-0.80), marginal R 2 (marginal R 2 )=0.151, conditional R 2(conditional R 2 ) = 0.744], this association is illustrated by a linear mixed-effects model representing fentanyl concentration as a function of theta power. As shown in Figure 2D, when fentanyl effect-site concentrations range from 0 to approximately 15 ng / mL, theta power can range from approximately -10 to 15 dB.

[0037] Further details of the systems and methods are provided in the examples below.

[0038] <Example> 1.a. Novel opioid-specific EEG signatures tracking patient state detectable during general anesthesia (polypharmacy). Retrospective EEG data from n=6 patients who received fentanyl to induce general anesthesia during cardiac surgery uncovered novel fentanyl-induced EEG signatures. Spectral features specific to fentanyl sedation and unconsciousness are shown in Figures 3A-3B. More specifically, Figures 3A-3B show EEG spectrograms from a representative subject, which show elevated slow wave band (0.1-1 Hz), delta band (1-4 Hz), and theta band power (4-8 Hz), consistent with an anesthetized state. Theta band signal appears as a distinct oscillation specific to fentanyl.

[0039] Prospectively recorded EEG data were recorded in n=26 individuals who received fentanyl prior to induction of general anesthesia and were administered a structured auditory behavioral response task. After a 5-minute baseline period, patients received two or three boluses of fentanyl at 2 mcg per kg ideal body weight, spaced 2 minutes apart (shown in Figure 2A). Figures 4A-4C are visualizations of the theta power model and the model using fentanyl Pk / Pd concentration to predict behavioral reaction time. In Figure 4A, the patient's reaction time to repeated auditory stimuli was used to gauge the patient's level of arousal. Figure 4B shows how the EEG theta power signature tracks the patient's behavioral reaction time. We investigated how the EEG theta signature performs as a predictor of behavioral response compared to and in combination with predicted fentanyl Pk / Pd effect-site concentrations (Figure 4C). Linear regression models were fitted using either fentanyl (predicted effect-site) concentrations, EEG theta concentrations, or a combination of both fentanyl and EEG theta, and model mean square prediction error (MSE) was characterized using k-fold cross validation, leaving out one patient in each cross-validation run. The MSE of the EEG theta model was 4% lower than that of the fentanyl effect-site concentration model, and the MSE of the combined model was 14% lower than that of the fentanyl effect-site concentration model. Using a generalized likelihood ratio test, the MSE of the combined model was P << 0.001 when all data were used. These data suggest that EEG theta power, alone or in combination with drug information, is a strong predictor of opioid-induced patient state and a better predictor than predicted Pk / Pd effect-site concentration alone. EEG theta power, alone or in combination with drug information, can also provide information about an individual patient's real-time individualized drug response to opioid analgesics.

[0040] In one non-limiting example, we investigated whether the fentanyl-derived EEG theta signature could be extracted during general anesthesia (polypharmacy) when other drugs such as propofol or sevoflurane are administered to maintain unconsciousness. In particular, propofol and sevoflurane produce large slow-wave and alpha oscillations that can obscure nearby fentanyl theta oscillations. A model comparison analysis was performed to determine whether the data contained two (slow wave, alpha) or three (slow wave, theta, alpha) oscillations. To compare the models, we used the class of "state-space oscillator" models, as well as either two or three oscillators and AIC. In one non-limiting example, the state-space oscillator model proposed by Matsuda and Komaki (2017) and Beck, He, Gutierrez, and Purdon (2022) can be used. Figure 5A shows EEG spectra in a representative subject during general anesthesia with sevoflurane and fentanyl, analyzed using traditional spectral analysis and state-space methods. Both techniques show clear peaks in the slow-wave, theta, and alpha bands. Figure 5B shows AIC analysis performed using the two-oscillator state-space model and the three-oscillator state-space model (Figure 5B) for n=30 subjects receiving sevoflurane and fentanyl. The analysis suggested that the fentanyl theta component could be detected and extracted separately from the background of sevoflurane-derived slow-wave and alpha oscillations. In other retrospective analyses (not shown), similar theta oscillations were observed when remifentanil was used with propofol. These data support the use of opioid EEG signatures as markers of nociceptive control and opioid drug effects during general anesthesia and other settings in which multiple drugs are administered with opioids.

[0041] 1.b.Electrodermal activity recorded from a forehead sensor simultaneously with EEG tracks surgical nociception. EDA tracks autonomic changes induced by nociceptive or emotional stimuli. Typically, EDA is recorded from the palmar surface, which has a high density of sweat glands, but the forehead has a comparable high density of sweat glands, allowing EDA to be measured on the forehead simultaneously with EEG. This simultaneous measurement can be done by any number of techniques known in the art, including, for example, administering a known current to EEG electrodes and measuring the resulting voltage change to infer skin conductance. Figures 6A-6D show observations of skin conductance responses (SCRs) during nociceptive events during surgery. That is, EDA information was measured from a forehead sensor simultaneously with EEG during general anesthesia, which is shown in Figures 6A-6C. Figures 6B-6C show representative forehead EDA data for one subject undergoing general anesthesia during surgery. In FIG. 6B, after induction of general anesthesia with propofol, the EDA skin conductance response (SCR) is clearly decreased and remains below about 5 microsiemens thereafter. In FIG. 6C, the lidocaine infusion to provide analgesia is stopped and the patient remains unconscious, but the EDASCR is clearly increased, rising from above about 5 microsiemens during surgery and general anesthesia to a high level of about 35 microsiemens, suggesting an increase in nociception. In FIG. 6D, pre-operative, intra-operative and post-operative EDA levels are compared in three representative subjects. During surgery, suspected nociceptive events (e.g., intubation, initial incision, active surgery, etc.) were seen to result in a clearly higher SCR than non-nociceptive periods (e.g., periods before the start of active surgery, etc.). EDA decreases during general anesthesia but increases after awakening. These preliminary data demonstrate that it is possible to simultaneously record EDA and EEG from the forehead and confirm that EDA changes can track surgical nociception using a combined sensor.

[0042] In one non-limiting example, EEG and EDA can be monitored simultaneously during surgery to inform intraoperative fentanyl titration. If SCR decreases after induction of general anesthesia with propofol in FIG. 6B and increases after lidocaine injection in FIG. 6C, this can suggest an upper range for fentanyl titration. For example, if conductance is below a certain value, e.g., below 5 microsiemens, this can suggest an optimal intraoperative fentanyl dose that minimizes intraoperative nociception and postoperative pain. Alternatively, the optimal fentanyl dose can correspond to a range of conductance values.

[0043] 1.c. Changes in alpha oscillation amplitude track arousal states and correlate with EDA; a novel method to accurately extract alpha fluctuations. Anesthesia-induced frontal alpha band (8-12 Hz) power and amplitude changes are known to be due to changes in arousal states, with alpha power increasing as anesthetic dose increases from sedation through unconsciousness and decreasing upon arousal and / or noxious stimulation.

[0044] Figures 7A-7B show the extraction of alpha oscillations and alpha amplitude in a surgical patient under general anesthesia. Thus, changes in alpha power can be used to track nociception-related arousal states. Figure 7A shows a typical EEG spectrogram during general anesthesia, in which alpha power fluctuates. The cause is presumed to be due to underlying changes in arousal states or nociception. Some state-space signal processing methods have been developed that can extract instantaneous fluctuations in alpha amplitude (Figure 7B) as described in Beck, He, Gutierrez, and Purdon (2022), as well as extract phase and instantaneous frequency information. Alternatively, other standard techniques can be used to estimate alpha amplitude, phase, or instantaneous frequency, such as applying a bandpass filter corresponding to alpha frequency, performing a Hilbert transform, and using bandpass (real part) and Hilbert transform (imaginary part) processed signals. Changes in alpha power or alpha frequency or alpha amplitude can be indicative of nociceptive-related arousal and can be used to titrate opioid or other non-opioid analgesics, for example, by selecting a dose that minimizes changes in alpha power, frequency or amplitude.

[0045] Figures 8A-8B compare the skin conductance response (SCR) and alpha power, both individual and averaged across the three patients in the study of Figure 6D, which correlates highly with the simultaneously recorded EDA (Figures 8A-8B). The correlation between alpha amplitude and EDA helps to validate the above literature linking alpha fluctuations to nociception (Figure 8C).

[0046] In one non-limiting example, the disclosed systems and methods are used to generate an intraoperative nociceptive management plan to maximize postoperative outcomes by constructing monitoring indices or variables and calibrating said monitoring indices or variables using models that characterize the relationship between intraoperative variables, anesthetic information, patient baseline, demographic variables, and postoperative outcomes, including but not limited to postoperative pain, opioid requirements, cognitive recovery time, and respiratory depression. Such models can be constructed using any number of techniques, including but not limited to machine learning models, deep learning networks, or regression models. These models can consider additional confounding variables or covariates that can introduce bias into the prediction of postoperative outcomes, including patient baseline, demographic variables, medical history, anesthesia records, and other clinical variables, which can be obtained from electronic medical record systems. In one non-limiting example, the models can be constructed and calibrated using acquired data including all of the above measurements and variables and / or systematically collected data from patients.

[0047] Figure 9 is a schematic description of the variables to be measured and their timeline in the clinical observational study to collect data for model, algorithm, and index development. Preoperative measurements include pain numeric rating, psychomotor vigilance test, TabCAT Brain Health Assessment (BHA), and Hospital Anxiety and Depression Scale (HADS). Intraoperative measurements include EEG and EDA, as well as heart rate variability (HRV), blood pressure (BP) waveform data, and records of all medications and clinical events. Postoperative measurements in the postoperative recovery room (PACU) continue to include EEG and EDA recordings, as well as pain scores, every 15 minutes for the first hour and hourly thereafter. PVT, BHA, and HADS are measured every hour until discharge from the PACU. Opioid-related adverse events, Modified Aldrete Score (hyperpnea), PACU and hospital prescription analgesic administration and opioids administered, and 90-day and 180-day opioid consumption will be obtained from the medical records. Patient baseline, demographic variables, medical history, anesthesia records, and other clinical variables can be obtained from the electronic medical record system (e.g., EPIC).

[0048] 10, some approaches 1002 for developing operating room monitors focus on physiological variables 1006 and expert evaluation 1008 in an operating room setting, but do not explicitly consider the relationship between intraoperative variables and postoperative outcome 1012 in the construction or calibration of the monitoring system, algorithm, or monitoring index 1010. Thus, an operating room monitor developed using such an exemplary approach 1002 may not achieve the desired postoperative outcome 1012 because the monitoring index 1010 value or its range of values ​​bears no or little relationship to the desired postoperative outcome 1012. Instead, an approach 1004 described in an aspect of the present disclosure uses data acquired, for example as described in FIG. 9, to operate a monitoring system, algorithm, and monitoring index 1016 derived from physiological variables 1014 and patient records to achieve the desired postoperative outcome 1018. As described herein, postoperative planning based on the surgical process can be provided to reduce postoperative pain, opioid requirements, respiratory depression, and postoperative hospital stay compared to postoperative planning that does not consider opioids and other medications administered during the surgical procedure.

[0049] In one non-limiting example, the disclosed system and method was considered in the context of three female patients, Patient 1, Patient 2, and Patient 3. Patient 1 was 48 years old, Patient 2 was 51 years old, and Patient 3 was 52 years old, all of whom underwent surgery under general anesthesia with different levels of intraoperative analgesia and different postoperative outcomes. The mean perioperative fentanyl concentration was calculated for each patient, as were the mean theta EEG power, skin conductance, and variability in alpha EEG amplitude and instantaneous alpha EEG frequency. The mean perioperative fentanyl concentration was calculated by retrieving medication information from the electronic medical record system and calculating effect-site concentrations using the Pk / Pd model described by McClain and Hug (Clinical Pharmacology & Therapeutics 1980). Theta power over the surgical period was calculated in four-second windows over the surgical period using a multitaper spectral estimation method. Theta power was estimated for each window using multiple tapers with a center frequency of 6 Hz covering theta band from 4 to 8 Hz, and a midpoint report of these values ​​was generated for the duration of surgery. Electrode conductance was estimated by demodulating the 7.5 nA and 78 Hz test current used to measure electrode impedance. These data were filtered using a 6 Hz bandwidth filter and then a Hilbert transform was applied to estimate the amplitude of the 78 Hz signal. This estimated amplitude was used to calculate impedance and conductance. A midpoint report of these values ​​was generated for the duration of surgery. Alpha amplitude and instantaneous frequency were calculated by applying a Hilbert transform to the EEG data bandpass filtered to 8 to 12 Hz with a 2 Hz filter transition window. Postoperative outcomes were also examined for each patient, i.e., total opioid use in the first 24 hours after surgery and maximum pain score in the first 24 hours after surgery. These data were obtained from the electronic medical records.

[0050] Figure 11 shows the mean fentanyl concentration and median theta power for each of patients 1, 2, and 3. Patient 1 received the highest fentanyl concentration, patient 2 received a lower concentration of fentanyl, while patient 3 received no fentanyl. The median theta power for each patient shows the corresponding trends for each patient, with patient 1 having the highest theta power, patient 2 having a lower theta power, and patient 3 having the lowest theta power, as can be seen in Figure 2D.

[0051] FIG. 12 illustrates the postoperative outcomes of the patients listed above. Consistent with the modeling and counterfactual analysis presented above, patient 1, who had the highest fentanyl concentration and the highest corresponding theta power, required the lowest amount of postoperative opioids for pain control in the first 24 hours after surgery of the three patients and also had the lowest maximum pain score in the first 24 hours after surgery of the three patients. Patient 3, who had the lowest fentanyl concentration and the lowest corresponding theta power, required the highest amount of postoperative opioids for pain control in the first 24 hours after surgery of the three patients and also had the highest maximum pain score in the first 24 hours after surgery of the three patients. Patient 2, whose fentanyl concentration and corresponding theta power are between those of patients 1 and 3, required the highest amount of postoperative opioids for pain control in the first 24 hours after surgery between those of patients 1 and 3 and had a maximum pain score in the first 24 hours after surgery between those of patients 1 and 3. As shown in Figure 4, theta power, alone or in combination with estimated drug concentrations using the Pk / Pd model, provides an accurate assessment of a patient's individual response to opioids. Given the above modeling and counterfactual analysis, theta power, alone or in combination with estimated drug concentrations, can be used to predict postoperative outcomes or to guide analgesic treatment to achieve a desired postoperative outcome.

[0052] Skin conductance for each patient is shown in Figure 13. Consistent with the above, maintaining skin conductance below some threshold may help optimize post-operative pain and opioid requirements. The variability in alpha amplitude and alpha instantaneous frequency for each patient is shown in Figure 14. Consistent with the above, maintaining the variability in alpha amplitude and alpha instantaneous frequency below some threshold may help optimize post-operative pain and opioid requirements.

[0053] By taking multiple related measurements including multiple complex patient baseline variables, medical history, multiple administered medications, as well as opioid-induced EEG theta oscillations, skin conductance, and alpha band EEG amplitude and instantaneous frequency variations, among other variables, the disclosed systems and methods of each aspect that model the relationships between the above variables and summarize information to guide intraoperative analgesia treatment or postoperative pain management are advantageous and highly desirable. However, the complexity of this high-dimensional information, which must often be interpreted in real time, makes it unlikely that a clinician would consistently achieve optimal results without the above systems and methods for processing such multiple information.

[0054] It will be apparent to those skilled in the art that, although the invention disclosed herein has been described in connection with particular embodiments and examples, the invention is not limited thereto, and that numerous other embodiments, examples, uses, improvements, and derivatives of the embodiments, examples, and uses are intended to be encompassed within the scope of the appended claims.

[0055] Various features and advantages of the invention are set forth in the following claims.

Claims

1. one or more sensors configured to measure electroencephalogram (EEG) signals and electrodermal characteristic signals of a patient administered at least one anesthetic agent and at least one pain relieving agent during a medical surgical procedure; a processor operatively coupled to the one or more sensors; An intraoperative patient monitoring system comprising: The processor, receiving the EEG signal and the electrodermal characteristic signal; using the EEG signal or the electrodermal characteristic signal to monitor the nociceptive state of the patient in real time during the medical-surgical procedure; generating a post-operative pain management plan using at least one of the nociceptive state of the patient during the medical-surgical procedure and the at least one pain relieving agent administered to the patient for the medical-surgical procedure; An intraoperative patient monitoring system comprising:

2. The processor is further configured to identify a signature from the EEG signal that correlates highly with an effect site concentration (ESC) of the at least one analgesic agent in the patient. The intraoperative patient monitoring system of claim 1.

3. the at least one analgesic agent comprises an opioid; 3. The intraoperative patient monitoring system according to claim 1 or 2.

4. The opioid comprises fentanyl. The intraoperative patient monitoring system according to claim 3.

5. The post-operative pain management plan includes a plurality of post-operative outcomes. The intraoperative patient monitoring system of claim 1.

6. The plurality of outcomes includes at least one of postoperative pain, opioid requirement, cognitive recovery time, or respiratory depression. The intraoperative patient monitoring system of claim 5.

7. The processor is further configured to achieve a target post-operative outcome comprising at least one of reduced post-operative pain, reduced opioid requirements, reduced cognitive recovery time, and mitigated respiratory depression based on the patient's real-time nociceptive state during the medical surgical procedure. The intraoperative patient monitoring system of claim 1.

8. one or more sensors configured to measure electroencephalogram (EEG) signals and electrodermal characteristic signals of a patient administered at least one anesthetic agent and at least one pain relieving agent during a medical surgical procedure; a processor operatively coupled to the one or more sensors; An intraoperative patient monitoring system comprising: The processor, receiving the EEG signal and the electrodermal characteristic signal; using the EEG signals or the electrodermal characteristic signals to generate a report indicative of the patient's nociceptive state in real time while the patient is administered the at least one pain relieving agent during the medical-surgical procedure. An intraoperative patient monitoring system comprising:

9. The processor is further configured to identify a signature from the EEG signal that correlates highly with an effect site concentration (ESC) of the at least one analgesic agent in the patient. The intraoperative patient monitoring system of claim 8.

10. the at least one analgesic agent comprises an opioid; 10. The intraoperative patient monitoring system according to claim 8 or 9.

11. The opioid comprises fentanyl. The intraoperative patient monitoring system of claim 10.

12. the report indicating the real-time nociceptive state of the patient during the medical surgical procedure further comprises an indication of a titration of at least one of the dosage of the at least one anesthetic agent or the dosage of the at least one analgesic agent administered to the patient. The intraoperative patient monitoring system of claim 9.

Images