Bispectral Index Monitoring: Ability to Detect Deep Sedation during Endoscopy

Background: Clinical practice guidelines recommend monitoring the depth of anesthesia during endoscopic examination of the gastrointestinal tract using sedation scales, despite their subjective nature, while the use of the bispectral index, an objective measure, during sedation, remains controversial. The main objective of this study was to assess the ability of bispectral index monitoring to characterize the depth of anesthesia during endoscopy. Methods: We conducted a cross-sectional study to assess the performance of the bispectral index using data from a multicentre clinical trial with 180 patients undergoing scheduled colonoscopies. Sedation was monitored using the bispectral index and Ramsay Sedation Scale. Data on sedation were recorded at ﬁ ve-time points (t1 to t5) during the colonoscopy. Results: Bispectral values were signiﬁ cantly associated with Ramsay scores (rho, -0.73; p < 0.0001). In regression analysis, each unit increase in bispectral value was associated with a reduction in the risk of a high Ramsay score (> 3) at all points (OR 0.922; 95% CI: 0.865–0.979; p < 0.0001 at t1). Receiver operating characteristic curve analysis found areas under the curve of 0.8272 for a bispectral index cut-oﬀ for deep sedation of 69.76 (sensitivity, 95.35%; negative predictive value, 97.53%) when reaching the colic ﬂ exure (t2) and 0.8399 for a cut-oﬀ of 69.29 (sensitivity, 96.15%; negative predictive value, 98.81%) at the end of the colonoscopy (t5). Conclusion: Bispectral index monitoring enables objective real-time reliable assessment of sedation. It enables easy continuous monitoring with a very good performance for detecting deep

. Nonetheless, given the technical and methodological characteristics of these monitoring systems, the results may be less reliable during sedation, thus hindering the generalization of their use.
The 2018 guidelines of the European Society of Anaesthesiology on the management of sedation and analgesia in adults [2] indicated the need for continuous clinical observation, this being the basic level of clinical monitoring required during and after any procedural sedation (very good consensus: level of evidence B, grade of recommendation strong [2]).For this, the depth of sedation must be regularly assessed using one of the validated scales for assessing response to verbal and tactile stimuli [6,7].It was also concluded that processed electroencephalography

Background
In recent decades, the development of advanced diagnostic techniques and colorectal cancer screening programs has led to a steady increase in the number of gastrointestinal procedures worldwide.In this context, deep sedation and anaesthetist-led analgesia have become important, enabling better conditions for performing examinations, with a low rate of adverse events and good levels of patient satisfaction [1][2][3][4].
Monitoring of hypnosis enables the objective assessment of the depth of anesthesia, making it possible to tailor the anesthesia to each patient.Indeed, the ef icacy of this approach has been well demonstrated for general anesthesia https://doi.org/10.29328/journal.acgh.1001042 monitoring could be considered in sedated patients (very good consensus: level of evidence B, grade of recommendation weak [2]), though its role remains controversial.
The main objective of this study was to assess the ability of a BIS monitoring system to characterize the depth of anesthesia in patients under deep sedation.The evidence available in this context is limited, so we propose the study with the largest sample size published to date investigating different and relevant time points during the colonoscopy.

Methods
We designed a cross-sectional diagnostic accuracy study comparing a BIS monitoring system to the Ramsay Sedation Scale as a reference standard.We recruited patients undergoing scheduled colonoscopies under sedation and analgesia.The clinical research ethics committee of Galdakao Hospital (Bizkaia, Spain) evaluated the research project and approved the protocol on 18 January 2018 (protocol 01/18).This study is based on data from a clinical trial.It was conducted in accordance with the Declaration of Helsinki [8] and in compliance with data protection regulations, ensuring that personal data was handled in such a way that no data collected could not be associated with identi ied or identi iable individuals (Spanish Organic Law 15/1999, 13-12, Personal Data Protection).The trial was registered on ClinicalTrials.gov(identi ier NCT03453359).All participants provided written informed consent before their inclusion.
The target population was adult patients undergoing scheduled colonoscopies in Galdakao Hospital (Bizkaia, Spain) or Medaro Hospital (Gipuzkoa, Spain).The inclusion criteria were: age between 18 and 85 years, indication for scheduled colonoscopy, American Society of Anesthesiologists (ASA) physical status class I, II, or III, body mass index under 35 kg/m 2 , and no neurological dysfunction.We excluded patients with allergies to the drugs used, those with neurological conditions, moderate-to-severe kidney or liver failure, moderate-to-severe lung disease, and longterm opioid users.Sedation was carried out by one of ive anesthetists involved in the study, all of whom had extensive experience in this ield.
The primary outcome of this study was depth of sedation.We also recorded data on patient characteristics including age, sex, body weight and height, ASA class, potentially relevant comorbidities at baseline, and procedure duration and medication administered.
All patients were monitored noninvasively (Monitor In inity® Delta, Dräger Medical Systems Inc, USA), following the recommendations in sedation guidelines of international societies [1,2,4].In addition, all patients were assessed using both the RSS (Table 1) and a BIS monitoring system (BIS TM Quatro sensor, for the BIS TM VISTA monitor, Aspect Medical System).Notably, the BIS values were recorded separately from the RSS scores, the former by a trained observer and the latter by the anesthetist.
Both RSS scores and BIS values were recorded at ive-time points during the procedure: t1, at the start of the colonoscopy, t2, when the endoscope reached the right colic lexure; t3, at the start of endoscope removal; t4, during resection of the irst polyp, and t5, at the end of the procedure.We accepted BIS readings provided that they had a signal quality index > 50 and were obtained from electroencephalographic traces with no electromyogenic artifacts.
Total intravenous sedation was provided using a computer-controlled infusion system with an Alaris TM PK syringe pump, following the Marsh model for propofol, with plasma concentrations between 1 and 3 μg/ml.Remifentanil was administered by continuous infusion with an Alaris TIVA syringe pump at doses between 0.05 and 0.15 μg/kg/ min, except in the case of patients who were over 70 years old or ASA class III, for whom the initial delivery rate was reduced to 0.02 μg/kg/min.The target level of sedation was a BIS value of 66 to 75 in the BIS group and an RSS score of 3 to 5 in the controls.
In a previous observational study at Galdakao Hospital, 60% of patients had BIS index values of between 66 and 75, and with long, complex procedures, sedation tends to be deeper.We estimated that we would need at least 83 patients per group to detect an around 20% difference in the percentage of patients under deep sedation, and speci ically, to test the main hypothesis that in the control group, the percentage would be 70% while in the BIS group, this percentage would be at least 20% higher.Assuming a loss to follow-up of 20%, we therefore needed 90 patients per group to achieve a level of signi icance (alpha) of 5% and statistical power (beta) of 80%.These calculations were performed using nQuery Advisor version 7.0.
Qualitative variables were expressed as frequencies and percentages and continuous variables as means and standard deviations.Percentages were compared using the chi-squared test (or Fisher's exact test, when the expected frequencies were less than 5), and differences between the means for continuous variables were examined using Student's t-tests or the nonparametric Wilcoxon test, depending on the type of distribution.Correlations between BIS and RSS were assessed using Pearson's correlation coef icient and the in luence of the BIS on having an RSS score > 3, which is the cut-off point that distinguishes deep sedation from mildto-moderate sedation, was analyzed using a simple logistic regression model at each of the ive-time points during the intervention.The predictive performance of each of the models was assessed using the area under the receiver operating characteristic curve.The threshold for statistical signi icance was set at p < 0.05.All the statistical analysis was carried out using SAS V9.4 (SAS Institute, Inc., Carey, NC).

Results
By the end of 2018, a total of 206 patients who underwent sedation for scheduled colonoscopy had been recruited but 26 were excluded for not meeting the selection criteria.We were left with 180 patients for inclusion in the analysis, reaching the required sample size.Data concerning the demographic and clinical characteristics of patients in each group and the interventions they received are summarised in Table 2.
A total of 829 BIS value and RSS score measurements were obtained.Overall, 109 patients underwent polyp resection (60.56% of all the patients included).A Pearson's correlation test yielded an r -value of 0.73 (Table 3), indicating a strong correlation between RSS scores and BIS values.

Discussion
The results of this study indicate that the BIS index performs very well in identifying patients with deep sedation, allowing us to assess the level of sedation objectively, reliably, and in real-time, without interrupting the procedure.They demonstrate that BIS readings are correlated with RSS scores during sedation using propofol and remifentanil.Further, this study determines cut-off BIS values indicative of deep sedation in scheduled colonoscopies, with high predictive accuracy.
Several studies have concluded that there is no evidence to support the routine use of the BIS index in endoscopic procedures [9][10][11][12][13].Nonetheless, motivated by the goal of improving patient safety and optimizing the care provided to patients under deep sedation, we believe that the use of this system for monitoring deep sedation should be considered, especially when using propofol [2].BIS monitoring minimizes complications, is associated with a high level of satisfaction among patients and endoscopists [14], and enables more effective titration with a corresponding reduction duration of sedation [15][16][17], the use of such a monitoring system is particularly advisable in the case of complex examinations [1].
In endoscopic sedation, previous research has indicated that the BIS index is strongly correlated with sedation scales, of which the RSS is one of the most widely used [15,18,19].This association has also been observed in other clinical settings, including intensive care [20][21][22][23], palliative sedation [24], and pediatric analgesia [25].Nonetheless, few studies have focused on the BIS values obtained in endoscopic sedation, values that might enable us to predict and minimize adverse cardiopulmonary events.Bower, et al. [26] and Yu, et al. [27] de ined optimal BIS cut-off values of 82 and 91, respectively, for maintaining moderate sedation; but did not de ine cut-offs for deep sedation.Qadeer, et al. in 2008 [9] published the results from elective ambulatory endoscopic procedures, showing BIS monitoring to have poor sensitivity and accuracy for detecting deep sedation, but the study had some limitations.More recently, cut-off points have been established for respiratory depression under deep sedation; however, the indings cannot be generalized due to the small sample size [28].Our study focuses on one of the procedures most commonly used for the diagnosis and treatment of gastrointestinal disorders.Every year, 13 to 15 million colonoscopies are performed in the United States of America [29] and more than 540,000 in the United Kingdom [30], in part due to colorectal cancer screening programs and population aging.Thus, the identi ication of deep sedation with this monitoring, established in our study at BIS values between 64 and 72, allows us to detect patients at risk of airway obstruction and respiratory depression.
The strengths of our study include the results being analyzed for ive speci ic time points during a colonoscopy, and to our knowledge, this study is the irst to conduct this type of analysis.Previous authors have de ined the parameters obtained every 30 seconds to 3 minutes [9,11,15,26]; however, such regular intervals are not linked to the stage of the procedure or requirement for sedation in given patients.Further, frequent stimulation to assess sedation level might lead to the administration of more medication and higher sedation scores, without increasing the validity of the results.We believe that assessing the indings as a function of the stage of the endoscopic procedure allows us to obtain more relevant and practical information, potentially useful for optimizing the level of sedation and hence improving the satisfaction of both patients and health professionals involved in sedation and endoscopic procedures.Our results suggest that BIS monitoring could reduce the incidence and severity of respiratory events during scheduled colonoscopies by early identifying patients most susceptible to these events.This would allow clinicians to take appropriate preventive measures in each case.
Nonetheless, our study also has some limitations.First, the population analyzed only included ASA class I to III patients who underwent elective endoscopic procedures; that is, we did not assess more complex patients, who might have had different BIS values.This exclusion of complex cases may restrict the external validity of the indings, though it allowed us to obtain conclusive data for the population studied.A second limitation is related to the recording of the depth of the anesthesia.While BIS monitoring provides a continuous objective measure of this level, the interpretation of monitoring based on the RSS may vary with the observer carrying out the measurement.Nevertheless, this scale has been validated, it is widely recommended by working groups, and its use has become widespread in daily clinical practice.Further, the same small group of anesthetists and observers collected all the data for the study, reducing the impact of inter-observer variability.

Strengths and limitations of this study
This study determines cut-off BIS values indicative of deep sedation in scheduled colonoscopies.
This study evaluates BIS values obtained at the ive time points during the colonoscopy, in which the stimulus varies considerably, and correlates them with the RSS score.
One limitation is that patients with multiple pathologies were not evaluated.

Conclusion
Our indings indicate that BIS monitoring provides an accurate reliable measurement of the level of sedation, with very good performance for detecting deep sedation in scheduled endoscopic procedures.

Figure 1 :
Figure 1: Relationship between bispectral (BIS) values and Ramsay Sedation Scale (RSS) scores.Box diagram of the BIS values obtained at the fi ve time points during the colonoscopy by RSS score.

Table 2 :
Descriptive analysis of patient characteristics.
N: Number; %: Percentage; *Results expressed as mean (standard deviation).ASA: American Society of Anesthesiologists; † Diagnosis of relevant comorbid conditions including hypertension, diabetes, high cholesterol, infl ammatory bowel disease, mild asthma, mild chronic obstructive pulmonary disease, hyperthyroidism, gastritis, and mild kidney disease.

Table 3 :
Correlation between the bispectral index and Ramsay Sedation Scale over the course of the colonoscopy.

Table 4 :
Results of the logistic regression assessing changes in BIS index refl ecting changes in RSS score during the colonoscopy (t1 to t5).

Table 5 :
Bispectral index cut-off values for deep sedation as assessed by the Ramsay Sedation Scale score > 3 (vs.≤ 3) at each of the fi ve measurement time points during the colonoscopy.

Table 6 :
Procedure-specifi c validity estimates for the bispectral (BIS) monitoring system for predicting deep sedation.