Novel insights on association and reactivity of Bispectral Index, frontal electromyogram, and autonomic responses in nociception- sedation monitoring of critical care patients

Introduction

Sedation and analgesia are a crucial part of critical care, but optimizing these in non-communicative patients can be challenging. Deep sedation is common, with a prevalence of 35–68% in mechanically ventilated patients, and excessive sedation is associated with adverse outcomes, such as a higher mortality and longer ICU stays[1–5]. Several randomized studies have shown improved outcomes with strategies avoiding over-sedation, however insufficient sedation increases both patient agitation and staff work load, and may compromise patient safety[6, 7].

One of the main challenges in detecting and treating pain and stress in ICU patients is the lack of suitable monitors of nociception and analgesia[1, 6, 8, 9]. Assessing abstract concepts, such as pain and suffering, is based on observing behavioural and autonomic physiological responses. Of these, the latter might provide an objective monitoring medium[8–10], however, the basic physiological parameters (such as heart rate and blood pressure) alone are not accurate enough for pain assessment[10, 11].

Derived frontal electroencephalogram (EEG) and electromyogram (EMG) variables can be used as noninvasive neuromonitoring methods of sedation and anesthesia depth. The most widely used EEG derived variable is the Bispectral Index (BIS), which has been validated for perioperative sedation[12] and has showed positive results in monitoring ICU sedation[5–7, 10, 13, 14]. The Responsiveness Index (RI) is an EMG-derived variable, proposed for sedation monitoring in the ICU. To determine RI the frontal EMG is measured with a forehead sensor, EMG power is derived from each 0.5 s epoch, and finally RI is derived based on the EMG power time series of the last 60 minutes[15]. Both BIS and RI provide real-time monitoring with a simple scale from 0 to 100, with low values representing deep sedation and higher values representing increasing arousal[12, 15–19].

The physiological stress responses to pain (tachycardia, hypertension, diaphoresis) can be blunted in ICU patients, mainly due to medication (sedatives, analgesics, muscle relaxants, blood pressure medication)[8, 10]. The increase in the tone of the forehead muscles as a response to nociception is more resistant to medication, including neuromuscular blockade agents (NMBA)[7] Thus, frontal EMG should reflect the stress response to pain, even when classical signs (such as tachycardia and hypertension) are absent[10].

Short-term blood pressure variability (BPV) is an interesting variable for assessing nociception and analgesia, as the autonomic responses of heart rate and blood pressure are inherently linked to each other[10, 11, 20, 21]. The variability of heart rate and blood pressure, along with direct increases in heart rate and systolic blood pressure, have been linked to nociception[10, 22, 23].

The Surgical Pleth Index (SPI), a derived variable combining normalized pulse photoplethysmographic waveform amplitude (PPGA) and RR interval (RRI), monitors nociception by reflecting the changes in the balance of sympathetic and parasympathetic tone[9, 10, 20].

The aim of this explorative study was to test the performance of sedation monitoring variables derived from EEG, EMG and hemodynamic measurements (heart rate, blood pressure, BPV, SPI), against Richmond Agitation-Sedation scores (RASS), and their responsiveness to painful stimuli during critical care.

Methods

The inclusion criteria for this study were adult patients, with a planned or unplanned postoperative admission to the ICU, on mechanical ventilation via an endotracheal tube, and receiving continuous intravenous sedation (propofol, midazolam). Exclusion criteria were primary neurological disorders (including stroke, probable hypoxic brain injury, intracranial hemorrhage, and head injury with reduced level of consciousness), the continuous use of NMBA during monitoring, confirmed CNS infection, or a short data collection time (less than 12 hours). Sparing bolus administration of NMBA to facilitate ventilation were allowed, as frontal EMG is reasonably resistant to the effects of partial blockade. Patient recruitment and data collection took place from the 7th of May 2007 to the 1st of April 2009.

Of the 32 recruited patients, 30 were included (2 excluded due to short data collection time). During daytime one of two dedicated research nurses conducted computerized and standardized RASS assessments every 60 minutes, with increasing stimuli given every minute (first a 90 dB verbal command from headphones, followed by 105 dB white noise from headphones, followed by a peripheral train-of-four nerve stimulation, followed by peripheral nerve tetanic stimulation). All observed painful stimuli (including bronchoscopy, puncture for pleural drainage, suction of airways, and train-of-four/tetanic nerve stimulations), medications, and patient reactions to painful stimuli were annotated. Data on the cumulative dose of sedative drugs (including propofol and midazolam), opioid analgesia (including fentanyl, oxycodone, sufentanil and buprenorphine), and muscle relaxants were recorded for the whole monitoring period. During mechanical ventilation, a target RASS of -2 to 0 was used as a sedation guideline.

Monitoring methods

A BIS sensor was positioned in the standard position on the patient’s forehead, from which BIS and EMG values were monitored with the E-BIS module of GE Datex-Ohmeda S/5 monitoring system (BIS XP, algorithm version 4.0, smoothing rate 15 s). The Entropy sensor was positioned bilaterally on the forehead, above the BIS sensor[15]. The RI values were retrospectively calculated from the EEG/EMG signal obtained from the Entropy sensor and E-Entropy module (GE Healthcare, Helsinki, Finland). Quality of BIS and RI signals were controlled with automatic sensor checks, and both sensors were changed every 24 hours. Plethysmographic pulse waveform signal was acquired from the GE SpO₂ sensor and measurement module. SPI, and its subcomponents PPGA and RRI, were derived offline using the plethysmographic pulse waveform signal and the SPI program code of GE Carescape monitor (GE Healthcare, Helsinki, Finland). Invasive blood pressure data, including systolic (sysBP) and diastolic (diaBP) blood pressure, were monitored from a peripheral arterial line. Mean values of sysBP and diaBP were stored at 10 second intervals. The sysBP time series were used to derive average real variability (ARV)[24], a mathematical variable describing BPV:

$${ARV\left(sysBP\right)}_{m}=\frac{1}{N-1}\sum _{k=m-N}^{m-1}\left|{sysBP}_{k+1}-{sysBP}_{k}\right|$$

where N was 180, i.e., a 30-minute time-window was used. SysBP measurements greater than 280 mmHg or lower than 50 mmHg and an increase of over 100 mmHg in 10 seconds were discarded as artifacts.

All patients were monitored continuously with 3-lead ECG and a peripheral arterial line with invasive blood pressure monitoring, and all ECG results were reviewed offline by a cardiologist (J.S.). Monitoring data were captured with the S/5 Collect SW (GE Healthcare, Helsinki, Finland).

Statistical Methods

In this exploratory study, variables for RASS comparison were selected from variables used to monitor the depth of anesthesia or sedation (i.e., BIS, EMG, and RI). As the preliminary visual analysis suggested a close resemblance between RI and ARV, ARV results were compared against RASS too. Number of study patients shown in high case (N), while the number of measurements are shown in low case (n).

Associations of BIS, EMG, RI and ARV to RASS levels were analysed with prediction probability (P_K). P_K is an established statistical method for quantifying the ability of an anesthetic depth indicator to decrease consistently with deepening anesthesia. Although it is more commonly used for binary categories, such as comparing the variable value of the responsive state against its value in a non-responsive state, the methodology can be similarly applied to more than two-ordered categories[25]. In our material, a P_K of 1 would indicate that the variable value will monotonically decrease with deepening sedation from RASS + 2 to RASS − 5, whereas a P_K of 0.5 would indicate that the variable’s capability for predicting RASS is equal to flipping a coin. As the data contained multiple samples from each patient, with different amounts of samples for each patient, we used the random sampling without replacement method as proposed by Lüginbuhl et al[26]. The method randomly selects one sample from every patient and derives the P_K value using the jack-knife method for this subset of data[25]. This procedure is repeated 1000 times, and the presented P_K values are the medians of the 1000 subset P_K values. The value of each variable was recorded just prior to the start of each RASS assessment, so that the stimulus of the assessment itself does not affect the recorded variables.

To evaluate EMG influence on BIS and RI, we divided all RASS observations to low and high EMG groups according to a threshold value of 30 dB[27]. To evaluate the possible influence of the autonomous nervous system on BIS and RI, we divided all RASS observations into two equal sized groups by the median ARV value. In both analyses’, we derived P_K values separately for each group with the presented random subsampling method.

We were further interested in studying the responsiveness of the variables to painful stimuli at different RASS levels. For this analysis, we selected the mean value of each variable from a time period 2 to 5 minutes prior to each registered stimulus, and the mean value from a time period of 0 to 3 minutes after each stimulus. Successive stimuli occurring within 10 minutes of the earlier stimulus were not included in the analysis. For the analysis we selected all the variables used in the study, including systolic and diastolic blood pressures, SPI and SPI subcomponents of RRI and PPGA. Wilcoxon signed rank test with Bonferroni correction was applied to study whether pre and post stimuli values were from the same population, the type I error was set at 5% (two-sided) which resulted in a Bonferroni corrected limit of statistical significance at a = 0.0056.

All statistical analyses were performed with Matlab 9.5 (The MathWorks Inc., Natick, MA, USA).

Results

All demographic data and data from the ICU treatment period are presented in Table 1. Of all the monitored ECG data 19% were non-sinus rhythm (e.g. atrial fibrillation, atrial flutter, or pacemaker rhythm). To facilitate ventilation, bolus NMBA were administered sparingly in 12 (40%) patients. Of these, 11 patients received rocuronium with a cumulative median (range) dose of 60 mg (30–450 mg), and 1 patient received cisatracurium with a cumulative dose of 14 mg. The patient with the highest dose of NBMA (rocuronium 450 mg) had pulmonary hypertension and was treated in the ICU with nitrous oxide inhalation for acute respiratory distress syndrome after cardiac surgery.

Table 1

Demographic and clinical data of all study patients (N = 30). Values are given as median (range), or total number of patients (%, percentage of all patients in subgroup), as appropriate.
Parameter	Value
Age (years)	59 (30 to 80)
Gender, female/male (N, %)	12 (40%) / 18 (60%)
BMI (kg/m²)	27.8 (23.7 to 33.5)
Monitoring time (h)	50 (31 to 70)
Propofol infusion dose during monitoring period (mg/kg/h)	1.2 (0.0 to 3.9)
Emergency ICU admittance	6 (20%)
Planned postoperative ICU admittance	18 (60%)
ICU LOS (days)	17 (2 to 37)
Hospital LOS (days)	18 (2 to 42)
SOFA score on 1st day	8 (4 to 15)
Discharged to a high-dependency unit	4 (14%)
In-hospital death	4 (13%)
Main reason for admittance to ICU:
Gastro-intestinal	8 (27%)
Cardiac	6 (20%)
Pancreatitis	5 (17%)
Ruptured abdominal aortic aneurysm	4 (13%)
Infection	4 (13%)
Urologic	1 (3%)
Thoracic	2 (7%)
Electrocardiogram dominant rhythm:
Sinus rhythm	81%
Atrial fibrillation	12%
Pacemaker rhythm	7%
Abbreviations: BMI = Body mass index; ICU = intensive care unit; LOS = length of stay; SOFA = Sequential organ failure assessment.

Table 2 presents the results of P_K analysis of 406 pairs of RASS score and variable values (for RASS distribution, see Table 5). Of the tested variables, BIS, RI and EMG demonstrated a moderate association with RASS.

Table 2

Prediction probabilities (PK) for monitored parameters, compared against the Richmond Agitation-Sedation Score (RASS). PK was estimated from 1000 jack-knife samples, each including one parameter-RASS observation pair from each patient (N = 30). The table presents the median PK of 1000 subset PK values, with interquartile range (IQR). A total of 406 RASS assessments were available for analysis.
Parameter	P_K [IQR]
Bispectral Index (BIS)	0.776 [0.739, 0.808]
Frontal electromyogram power (EMG)	0.761 [0.719, 0.795]
Responsiveness Index (RI)	0.763 [0.728, 0.799]
Blood pressure average real variability (ARV)	0.549 [0.504, 0.596]

Random subsampled P_K values for BIS and RI in both low and high EMG groups, and low and high ARV groups, are presented in Table 3.

Table 3

Presenting the medians and interquartile ranges (IQR) of jack-knife prediction probabilities (PK) for Bispectral Index (BIS) and Responsiveness Index (RI), compared against the Richmond Agitation-Sedation Score (RASS). Comparisons are grouped into low and high EMG groups (upper part of table), and into low and high blood pressure averaged real variability (ARV) groups (lower part of table). A total of 1000 jack-knife PK estimates were derived, each estimate including one sample of N patients in the group.
	Low EMG	High EMG
Parameter	N = 24	N = 29
Bispectral Index (BIS)	0.716 [0.684, 0.749]	0.736 [0.695, 0.777]
Responsiveness Index (RI)	0.749 [0.718, 0.775]	0.716 [0.673, 0.755]
	Low ARV	High ARV
Parameter	N = 28	N = 27
Bispectral Index (BIS)	0.770 [0.736, 0.800]	0.789 [0.756, 0.821]
Responsiveness Index (RI)	0.723 [0.685, 0.761]	0.770 [0.742, 0.801]

Figure 1 presents violin plot diagrams of BIS versus RASS, first using the whole data set, then grouped into both low and high EMG groups, and low and high ARV groups. Similarly, Fig. 2 presents violin plot diagrams of RI versus RASS for the whole data set and grouped into both low and high EMG groups, and low and high ARV groups.

In Table 4 are presented the results of 524 stimulus-response pair analyses (for RASS distribution, see Table 5). The difference between pre and post stimulus variable values (∆ BIS, ∆ EMG, ∆ ARV) at different RASS levels is presented in Fig. 3.

Table 4

Responses of the tested parameters to stressful stimuli at different Richmond Agitation-Sedation score (RASS) levels (n samples), analysed with the Wilcoxon signed rank test with Bonferroni correction (corrected = 0.0056). A statistically significant p value indicates a consistent change after the stimulus (either decrease or increase), and is marked with an asterisk (*).
	RASS
	-5	-4	-3	-2	> -2
Parameter	n = 141	n = 163	n = 98	n = 60	n = 63
Bispectral Index (BIS)	0.5547	< 0.0001*	< 0.0001*	0.9069	0.3944
Frontal electromyogram power (EMG)	< 0.0001*	< 0.0001*	< 0.0001*	0.4866	0.3109
Responsiveness Index (RI)	0.4158	0.0339	< 0.0001*	0.0779	0.0587
Systolic blood pressure	0.2339	0.0673	0.0258	0.4223	0.2725
Diastolic blood pressure	0.1248	0.0177	0.0016*	0.1532	0.0827
Blood pressure average real variability (ARV)	0.4278	0.0006*	< 0.0001*	0.0059	0.0178
Surgical Pleth Index (SPI)	0.0696	0.0051*	0.9157	0.2226	0.5407
RR Interval (RRI)	0.9108	0.0001*	0.2602	0.3588	0.8214
Plethysmograph amplitude (PPGA)	0.0121	0.0906	0.8456	0.7000	0.1644

Table 5

Presenting the n for all the different analysed pairs at different RASS levels. In the first group (Table 2) are the pairs of RASS and the analysed variables from before the RASS assessment presented in Table 2, including Bispectral Index (BIS), frontal EMG, Responsiveness Index (RI), and Averaged Blood pressure Variability (ARV). In the second group (Table 3) are the BIS/RI and RASS pairs presented in Table 3, grouped into low and high EMG and ARV groups. In the final group (Table 4) are the stimulus-response pairs presented in Table 4, which represent the pairing of a painful stimulus and measured variable values following stimulation. For the Table 4 Stimulus-response pairs, RASS ≥-2 were pooled**.
		RASS (n)
	Pairs	-5	-4	-3	-2	-1	0	+ 1	+ 2
Table 2	Variable-RASS	104	100	98	51	35	11	6	1
Table 3	Low EMG power BIS/RI-RASS	52	46	19	4	1	-	-	-
	High EMG power BIS/RI-RASS	52	54	79	47	34	11	6	1
	Low ARV BIS/RI-RASS	36	54	57	20	8	2	2	-
	High ARV BIS/RI-RASS	55	31	33	24	25	6	4	1
Table 4	Stimulus-response	141	162	98	60	63**

The correlation of EMG power versus BIS is presented in Fig. 4, showing almost linear correlation in the BIS range of 40–95. At low EMG activity (EMG Power < 30 dB), the correlation with BIS is lost.

As an example, the continuous monitoring data of a single patient can be seen in Fig. 5, presenting EMG power, RI, systolic blood pressure and ARV.

Discussion

This explorative study shows that several easily measurable continuous physiological variables reflect the sedation level of ICU patients, as determined by the RASS scale, and also respond to painful stimuli in sedated, mechanically ventilated patients who are unable to report pain.

Of the studied variables, EEG and EMG derived variables were associated with RASS levels, as was demonstrated by the moderate P_K values of BIS, RI and EMG power. The variability of blood pressure, represented by ARV, showed no association with RASS levels. Interestingly, the P_K value of BIS was not substantially better than the P_K value of EMG power provided by the BIS monitor. It is a known fact that frontal EMG activity contaminates BIS values[28], and past BIS algorithm improvements have focused on decreasing the impact of EMG to BIS[13, 16]. The distribution of BIS values in this material is trimodal (Fig. 4), peaking at 38, 62 and 98. The actual depth of sedation is unlikely to follow a similar trimodal shape, and the reason for such a presentation is probably a characterization of earlier observations, where BIS “freezes” just below or above the recommended range of 40–60 for surgical anesthesia[29]. This could be related to the algorithmic switch between the assigned weight of different BIS subparameters[30]. Based on our data these switches may be triggered by the EMG value. Although the origin of frontal EMG activity remains obscure[31], anesthesiologists have utilized frontal EMG responsiveness for a long time in connection with painful stimuli[16, 32], and in later studies frontal EMG variability was found to be good classifier between somatic and non-somatic events during elective, noncardiac surgery with possible predictive power for movement respons[33].

Our results confirm the earlier findings of a correlation between BIS and frontal EMG in the ICU setting[27, 34], but contrary to earlier studies we demonstrated that this correlation could be a favorable property for BIS, as it seems to explain part of the association between BIS and RASS (Fig. 1). Riker and co-workers demonstrated a decreased correlation between BIS (version 3.2) and the Sedation-Agitation Scale (SAS) and the Visual Analog Scale (VAS), when EMG power was over 39 dB[34]. Tonner and co-workers compared one of the older BIS algorithms to a XP-level algorithm, demonstrating improved discrimination of Ramsay score levels with the XP-level system[27]. Our results confirm the results of Tonner and colleagues[27], who demonstrated enhanced discrimination between different sedation levels when EMG activity is over 30 dB (Kendall t = -0.38 vs. t = -0.26 for BIS XP).

Our study suggests that the RI reflects autonomous nervous system responses, and its reasonable capability to detect deep sedation is partly explained by the fact that sedative drugs attenuate those responses. Figure 2 reveals that RI is most often either 0 or 100, where the value 0 is more probable at RASS levels from − 5 to -3, while the value 100 is more probable at RASS levels higher than − 3. Moreover, the RI value of 100 tends to be less likely in the low ARV group. Thus, RI had difficulties in detecting light sedation (RASS levels − 2 or higher) in calm patients with little or no blood pressure fluctuation. Low RI values in arousable patients have been reported earlier by Walsh et al., and were explained to be caused by sleep or minimal clinical stimulation[18]. Our data do not support the sleep hypothesis, as a majority of the concurrent BIS values were over 70 (Fig. 1), whereas the BIS values during sleep are typically less[35]. Figure 5 demonstrates a visual similarity of the RI and ARV trends, supporting our hypothesis that RI is linked to the effects of the autonomous nervous system. A recent study by Wennervirta et al. demonstrated a significantly higher incidence of hypertension (systolic blood pressure over 160 mmHg) in critical care patients when sedation was targeted to a RI level of 40 to 80, when compared to patients with a sedation target of RASS − 3 to 0. This finding supports the hypothesis that RI mostly reflects sympathetic activity and has thus very limited applications in sedation titration[36].

Frontal EMG was the last remaining response to painful stimuli in the deepest sedation level (RASS scale − 5). BIS was also reactive in RASS levels − 4 and − 3, but we assume that these responses are mainly explained by the EMG activation. To our knowledge, this was the first study where the utility of SPI was assessed in the ICU setting. The capability of SPI to detect painful stimuli in ICU patients seems to be limited, and mostly explained, by the RRI response. It is, however, important to note that factors inherently affecting SPI were not excluded or controlled in this study, for instance atrial fibrillation, beta blockers, and pacemaker rhythm.

The study results are limited by the explorative nature of the study, and by the small sample size. All results should be treated as hypothesis-generating and need to be validated with further research. The RASS assessment followed a standardized flow chart, but with two research nurses some inter-rater variability is bound to remain. Also, the results of the parameters responsiveness to painful stimuli should be taken as a preliminary finding, and these shall be confirmed in a future study with stricter study protocol and standardized stimuli.

References

Barr J, Fraser GL, Puntillo K, Ely EW, Gélinas C, Dasta JF, Davidson JE, Devlin JW, Kress JP, Joffe AM et al: Clinical practice guidelines for the management of pain, agitation, and delirium in adult patients in the intensive care unit. Crit Care Med 2013, 41(1):263–306.
Devlin JW, Skrobik Y, Gélinas C, Needham DM, Slooter AJC, Pandharipande PP, Watson PL, Weinhouse GL, Nunnally ME, Rochwerg B et al: Clinical Practice Guidelines for the Prevention and Management of Pain, Agitation/Sedation, Delirium, Immobility, and Sleep Disruption in Adult Patients in the ICU. Crit Care Med 2018, 46(9):e825-e873.
Jackson DL, Proudfoot CW, Cann KF, Walsh TS: The incidence of sub-optimal sedation in the ICU: a systematic review. Crit Care 2009, 13(6):R204.
Tanaka LM, Azevedo LC, Park M, Schettino G, Nassar AP, Réa-Neto A, Tannous L, de Souza-Dantas VC, Torelly A, Lisboa T et al: Early sedation and clinical outcomes of mechanically ventilated patients: a prospective multicenter cohort study. Crit Care 2014, 18(4):R156.
Wang ZH, Chen H, Yang YL, Shi ZH, Guo QH, Li YW, Sun LP, Qiao W, Zhou GH, Yu RG et al: Bispectral Index Can Reliably Detect Deep Sedation in Mechanically Ventilated Patients: A Prospective Multicenter Validation Study. Anesth Analg 2017, 125(1):176–183.
Devlin JW, Skrobik Y, Gelinas C, Needham DM, Slooter AJC, Pandharipande PP, Watson PL, Weinhouse GL, Nunnally ME, Rochwerg B et al: Clinical Practice Guidelines for the Prevention and Management of Pain, Agitation/Sedation, Delirium, Immobility, and Sleep Disruption in Adult Patients in the ICU. Crit Care Med 2018, 46(9):e825-e873.
Fraser GL, Riker RR: Bispectral index monitoring in the intensive care unit provides more signal than noise. Pharmacotherapy 2005, 25(5 Pt 2):19s-27s.
Chen HJ, Chen YM: Pain assessment: validation of the physiologic indicators in the ventilated adult patient. Pain Manag Nurs 2015, 16(2):105–111.
Ledowski T: Objective monitoring of nociception: a review of current commercial solutions. Br J Anaesth 2019, 123(2):e312-e321.
Cowen R, Stasiowska MK, Laycock H, Bantel C: Assessing pain objectively: the use of physiological markers. Anaesthesia 2015, 70(7):828–847.
Kyle BN, McNeil DW: Autonomic arousal and experimentally induced pain: a critical review of the literature. Pain Res Manag 2014, 19(3):159–167.
Johansen JW, Sebel PS, Sigl JC: Clinical impact of hypnotic-titration guidelines based on EEG bispectral index (BIS) monitoring during routine anesthetic care. J Clin Anesth 2000, 12(6):433–443.
Johansen JW: Update on bispectral index monitoring. Best Pract Res Clin Anaesthesiol 2006, 20(1):81–99.
LeBlanc JM, Dasta JF, Kane-Gill SL: Role of the bispectral index in sedation monitoring in the ICU. Ann Pharmacother 2006, 40(3):490–500.
Lapinlampi TP, Viertiö-Oja HE, Helin M, Uutela KH, Särkelä MO, Vakkuri A, Young GB, Walsh TS: Algorithm for Quantifying Frontal EMG Responsiveness for Sedation Monitoring. Can J Neurol Sci 2014, 41(5):611–619.
Ball J: How useful is the bispectral index in the management of ICU patients? Minerva Anestesiol 2002, 68(4):248–251.
Kaila M, Everingham K, Lapinlampi P, Peltola P, Särkelä MO, Uutela K, Walsh TS: A randomized controlled proof-of-concept trial of early sedation management using Responsiveness Index monitoring in mechanically ventilated critically ill patients. Crit Care 2015, 19(1):333.
Walsh TS, Everingham K, Frame F, Lapinlampi TP, Särkelä MO, Uutela K, Viertiö-Oja HE: An evaluation of the validity and potential utility of facial electromyelogram Responsiveness Index for sedation monitoring in critically ill patients. J Crit Care 2014, 29(5):886.e881-887.
Walsh TS, Lapinlampi TP, Ramsay P, Särkelä MO, Uutela K, Viertiö-Oja HE: Responsiveness of the frontal EMG for monitoring the sedation state of critically ill patients. Br J Anaesth 2011, 107(5):710–718.
Laycock H, Bantel C: Objective Assessment of Acute Pain. J of Anesth & Clin Res 2016, 7(6):3.
Pagani M, Lombardi F, Guzzetti S, Rimoldi O, Furlan R, Pizzinelli P, Sandrone G, Malfatto G, Dell'Orto S, Piccaluga E et al: Power spectral analysis of heart rate and arterial pressure variabilities as a marker of sympatho-vagal interaction in man and conscious dog. Circ Res 1986, 59(2):178–193.
Jeanne M, Logier R, De Jonckheere J, Tavernier B: Heart rate variability during total intravenous anesthesia: effects of nociception and analgesia. Auton Neurosci 2009, 147(1–2):91–96.
Koenig J, Jarczok MN, Ellis RJ, Hillecke TK, Thayer JF: Heart rate variability and experimentally induced pain in healthy adults: a systematic review. Eur J Pain 2014, 18(3):301–314.
Mena L, Pintos S, Queipo NV, Aizpúrua JA, Maestre G, Sulbarán T: A reliable index for the prognostic significance of blood pressure variability. J Hypertens 2005, 23(3):505–511.
Smith WD, Dutton RC, Smith NT: Measuring the performance of anesthetic depth indicators. Anesthesiology 1996, 84(1):38–51.
Luginbühl M, Schumacher PM, Vuilleumier P, Vereecke H, Heyse B, Bouillon TW, Struys MM: Noxious stimulation response index: a novel anesthetic state index based on hypnotic-opioid interaction. Anesthesiology 2010, 112(4):872–880.
Tonner PH, Wei C, Bein B, Weiler N, Paris A, Scholz J: Comparison of two bispectral index algorithms in monitoring sedation in postoperative intensive care patients. Crit Care Med 2005, 33(3):580–584.
Vivien B, Di Maria S, Ouattara A, Langeron O, Coriat P, Riou B: Overestimation of Bispectral Index in sedated intensive care unit patients revealed by administration of muscle relaxant. Anesthesiology 2003, 99(1):9–17.
Vakkuri A, Yli-Hankala A, Talja P, Mustola S, Tolvanen-Laakso H, Sampson T, Viertiö-Oja H: Time-frequency balanced spectral entropy as a measure of anesthetic drug effect in central nervous system during sevoflurane, propofol, and thiopental anesthesia. Acta Anaesthesiol Scand 2004, 48(2):145–153.
Rampil IJ: A primer for EEG signal processing in anesthesia. Anesthesiology 1998, 89(4):980–1002.
Hight DF, Voss LJ, García PS, Sleigh JW: Electromyographic activation reveals cortical and sub-cortical dissociation during emergence from general anesthesia. J Clin Monit Comput 2017, 31(4):813–823.
Edmonds HL, Jr., Couture LJ, Stolzy SL, Paloheimo M: Quantitative surface electromyography in anesthesia and critical care. Int J Clin Monit Comput 1986, 3(2):135–145.
Mathews DM, Clark L, Johansen J, Matute E, Seshagiri CV: Increases in electroencephalogram and electromyogram variability are associated with an increased incidence of intraoperative somatic response. Anesth Analg 2012, 114(4):759–770.
Riker RR, Fraser GL, Simmons LE, Wilkins ML: Validating the Sedation-Agitation Scale with the Bispectral Index and Visual Analog Scale in adult ICU patients after cardiac surgery. Intensive Care Med 2001, 27(5):853–858.
Nieuwenhuijs D, Coleman EL, Douglas NJ, Drummond GB, Dahan A: Bispectral index values and spectral edge frequency at different stages of physiologic sleep. Anesth Analg 2002, 94(1):125–129, table of contents.
Wennervirta JE, Sarkela MOK, Kaila MM, Pettila V: Responsiveness Index versus the RASS-Based Method for Adjusting Sedation in Critically Ill Patients. Crit Care Res Pract 2021, 2021:6621555.

Novel insights on association and reactivity of Bispectral Index, frontal electromyogram, and autonomic responses in nociception- sedation monitoring of critical care patients

Abstract