Introduction

Sedation is an important element of intensive care. This concept includes two main components, i.e. procedural sedation and deliberate prolonged oppression of the patient’s consciousness. The first option is designed to reduce patient discomfort during invasive procedures. Therefore, this sedation is short-term. At the same time, the goals of prolonged sedation are described in the guidelines of the Federation of Anesthesiologists and Reanimatologists (2020) [1]. These ones include, among other things, prevention and treatment of delirium, other cognitive and psycho-emotional disorders, synchronization with ventilator, and reduction of ICU- and hospital-stay. In addition, sedation issues are inextricably associated with pain management, immobilization of patients and sleep disorders [2].

Thus, it is obvious that there are many situations when sedation is necessary. There are literature data on protocol-based sedation reducing the risk of nosocomial infection, circulatory depression, number of sedatives used, duration of mechanical ventilation and ICU-stay [2]. On the other hand, there are two types of unsafe sedation, i.e. insufficient oppression of consciousness and, on the contrary, excessive sedation up to drug-induced coma. In the first case, agitation will increase tissue oxygen consumption, result desynchronization with ventilator in intubated patients and increase the risk of barotrauma, self-extubation or removal of vascular catheters. In the second case, the course of disease can be complicated by various events requiring correction of treatment (Fig. 1) [3].

Thus, physicians need to have the means of objective control of sedation to achieve therapeutic goals and prevent complications [4]. Moreover, modern guidelines on sedation in intensive care units suggest personal target level of sedation for each patient, its intermittent nature, adjusting the target level over time and choosing the least deep level of sufficient sedation [2]. However, these recommendations are useless without objective assessment of sedation.

To assess the quality of sedation, various methods have now been proposed. These ones include clinical assessment, sedation scales and hardware control (BIS monitoring, EEG entropy monitoring, auditory evoked potentials). National guidelines on sedation therapy indicate the RASS score (Richmond Agitation-Sedation Scale) as the best option [1]. However, we consider the following problems of objective control:

1. How to carry out objective control at night and while the patient is sleeping? On the one hand, the goal of sedation is achieved if the patient sleeps at night. On the other hand, how to evaluate the role of physiological decrease in functional brain activity to depression of consciousness? Should we reduce the doses of sedatives at night to avoid excessive depression of consciousness and associated undesirable consequences (Fig. 1)?

2. Empirical assessment based on clinical data can be misleading, since it depends on personal experience and subjective feelings [5].

3. There are difficulties in making decisions about adjusting the dose of sedatives, since the score is not a dynamically changing characteristic described momentary level of sedation.

4. There is conflicting evidence on benefits of instrumental assessment in the ICU.

Fig. 1. Adverse consequences of excessive sedation.

The last point requires additional discussion in the context of sedation control using BIS monitoring. This method was proposed in 1994 and subsequently used to assess consciousness suppression during general anesthesia [6, 7]. At the same time, BIS monitoring has been actively used in intensive care units since 2002 [7]. According to various studies, BIS monitoring can be an adequate alternative to RASS assessment. It is valuable to assess pain syndrome, reduces drug consumption, prevents insufficient or, conversely, excessive sedation and reduces treatment costs [7—11]. In addition, BIS monitoring in intensive care units can be useful when scoring is not possible, for example, in patients receiving muscle relaxants. However, the authors emphasize that this method should not be the only measure of sedation depth [12]. A systematic review of 37 studies (8 ones included into further analysis) was published in 2016. The authors revealed significant correlation between BIS and RASS scores and emphasized that this method should not be the only method [13].

On the other hand, some authors including meta-analysis published in 2018 found no significant advantages of BIS monitoring [14, 15]. This meta-analysis has significantly changed the attitude towards BIS monitoring. In fact, sedation control with BIS monitoring has ceased to be used in intensive care units. At the same time, the authors indicate uncertain results, since they were obtained in a few studies [14]. In addition, one of the studies cited above [7] was excluded from meta-analysis because BIS monitoring was used together with clinical assessment of sedation in this report. There is an opinion that BIS monitoring and scoring systems assess different aspects of sedation. Electromyogram can influence BIS value that also limits the use of this method in intensive care units [15]. Thus, these studies are not likely to discredit the method, but point to specific limitations.

Despite refusal to use BIS monitoring in intensive care units following these publications, recent data evidence that such monitoring can still be beneficial [2, 16]. Moreover, a systematic review protocol devoted to BIS monitoring in ventilated patients was published in 2020 [17]. A meta-analysis devoted to correlation between BIS and clinical sedation scales was published 2 years later. The authors revealed advisability of BIS monitoring when the scales are not applicable [18]. Finally, the FAR guidelines recommend instrumental assessment of sedation in patients receiving muscle relaxants (evidence level 2, class of recommendation B). Moreover, the authors emphasize advisability of neuromonitoring in the 2022 National Intensive Care Guidelines (perioperative management of geriatric patients, prevention and treatment of delirium) [19]. Thus, we can conclude certain interest in hardware monitoring of sedation in intensive care units in the world professional community, while the data on this assessment cannot be called unambiguous. Therefore, we also wanted to look into this issue. We believe that our study will provide a fresh look at the prospects for continuous hardware assessment of sedation in intensive care units considering all limitations of this method.

The purpose of the study was to improve intensive care safety using BIS monitoring.

To achieve this purpose, we formulated the following objectives:

— to determine the target level of sedation according to BIS within the target RASS scores;

— to compare propofol consumption in study groups;

— to assess the impact of BIS monitoring on the incidence of adverse events;

— to assess the feasibility of BIS monitoring to control sedation at night.

Material and methods

The local ethics committee approved this study (protocol No. 7 dated 12/09/2022).

There were 69 patients. Inclusion criteria were ICU-stay and the need for prolonged sedation (comfortable mechanical ventilation or therapy of delirium). Importantly, sedation was not used to ensure sleep in both groups, but precisely for the above-described indications. At the same time, one of the objectives was analysis of the need for sedation during nocturnal sleep. Non-inclusion criteria: age over 70 years, individual intolerance to propofol, Parkinson’s disease. Exclusion criteria: various states requiring invasive diagnosis and treatment or transportation outside the intensive care unit for diagnosis, acute cerebrovascular accident or acute myocardial infarction.

All eligible patients were randomized into 2 groups. The first group included 38 patients (sedation control by BIS monitoring + RASS score in accordance with the recommendations [1]). The control group enrolled 31 patients (RASS score alone). There were no significant between-group differences in severity of organ dysfunction.

BIS monitoring allows you to evaluate functional activity of the brain. The method is based on an integral indicator of four EEG characteristics. This value is expressed as a numerical value from 0 to 100 (dimensionless unit) where 0 is absence of bioelectrical cortical activity, 100 is fully functional activity [16].

In the 1^st group, we focused on BIS score 70—80 (moderate sedation, the patient responds to loud speech). In the 2^nd group, target RASS scores were -1…-2. Statistical analysis was performed using the Jamovi software package (version 2.3.18). We chose 95% CI (confidence interval) and p-value<0.05.

At the first stage, we analyzed normality of data distribution using Shapiro-Wilk test. The p-values for group 1 (BIS) and group 2 (RASS) were 0.401 and 0.356, respectively. Thus, distribution was normal. We applied t-test, Spearman and Kendall correlation coefficients for further comparison of groups.

Results and discussion

So, we studied 69 patients divided into 2 groups. In the first group (n=38), we assessed sedation using BIS monitoring, in the second group (n=31) — using RASS score. Data distribution was normal in both groups. Study groups were comparable regarding characteristics of patients and comorbidities (Table 1, 2). Mean time of mechanical ventilation is presented in Table 1. We understand that this indicator does not fully describe the actual duration of respiratory support in patients with various diseases (Table 2), so it is given here as an additional criterion for group comparability. Sedation time is not shown in the table, since it closely corresponded to the time of respiratory support in patients on mechanical ventilation. In patients without mechanical ventilation (5 and 4 ones in both groups, respectively), this value did not exceed 2 days.

Table 1. Patient characteristic

Table 2. Diseases in both groups

When analyzing the data, we chose propofol consumption as one of the main indicators. We deliberately abandoned dose calculation in mg/kg/h in favor of daily estimate. Indeed, the first measure reflects the current consumption varying when choosing the depth of sedation, and the second one describes global need for the drug. On the other hand, it was necessary to provide comparison of indicators in different patients. Therefore, we used dose calculation in mg/kg/day to estimate propofol consumption because this approach eliminates the influence of anthropometric parameters on final results. Mean propofol consumption was 32.37±3.42 and 38.42±3.13 mg/kg/day, respectively. Thus, consumption decreased by 15.7% (Student’s t-test for two independent samples -7.65, p<0.01). Thus, lower propofol consumption in the BIS monitoring group is not accidental (Fig. 2). In our opinion, this difference is associated with higher dosage in the RASS group for “certainly effective sedation”, when the doctor has a subjective feeling of the need for higher dose to eliminate adverse events and spontaneous psychomotor agitation following insufficient sedation. Another aspect is higher drug consumption at night, when the number of personnel and alertness are reduced.

Fig. 2. Between-group propofol consumption differences (mg/kg/day).

In the next stage, we assessed what BIS would be sufficient for purposes of sedation without excessive depression of consciousness (i.e., the patient is conscious, possibly drowsy or asleep, but easily aroused in response to verbal stimulation). This target level of sedation was 77±6.1% (data on moderate sedation within 60—80% [16]). Importantly, there are no literature data on technique of BIS-oriented sedation in intensive care units, and some authors emphasize no advantages of such sedation. Therefore, no indications on target BIS scores are present. We believe that our value can be considered a guideline for sedation control in intensive care units.

Finally, we assessed the correlation between BIS and RASS scores. Spearman’s rank correlation coefficient was 0.827 (p<0.01), Kendall’s correlation coefficient — 0.711 (p<0.01) (Fig. 3). The target BIS score 77±6.1% corresponds mainly to RASS scores from –1 to –2 with a few cases of more superficial or deeper sedation (one case of RASS score 0 and two cases of RASS score 3). Mean RASS scores in this range, including these three outliers, was 1.7±0.7. Thus, BIS monitoring data significantly correlate with RASS scores despite deviating BIS values in some cases. This result is consistent with available literature data [11]. On the one hand, this suggests that continuous monitoring of sedation may be useful along with routine scoring. On the other hand, instrumental method has obvious advantages:

Fig. 3. Correlation between RASS and BIS (blue — trend line, red dotted line — mean BIS, colored area — range of target values (BIS mean deviation)).

— BIS monitoring allows detecting changes in sedation, which may go unnoticed if they occur between the moments of RASS assessment;

— BIS monitoring is devoid of subjective component that is inevitable in clinical assessment; this factor contributes to stable sedation;

— nurses can observe the level of consciousness using BIS monitoring and involving a doctor to make tactical decisions if necessary;

— there is no need to involve the patient in evaluation of sedation, for example, by waking up when he is sleeping. That is, BIS monitoring can also be useful for assessing sedation during nocturnal sleep.

Of course, these findings and conclusions require further verification on large samples. However, even now it seems to us that instrumental continuous assessment of consciousness has certain prospects in intensive care units for sedation control. Perhaps, this method is not justifiably unpopular.

Importantly, there were no adverse events associated with excessive depression of consciousness in the 1^st group. At the same time, two patients in the 2^nd group required mechanical ventilation due to excessive sedation at night.

Conclusion

We determined the target BIS value of sedation (77.0±6.1%). The dose of drugs for this sedation was lower by 15.9% compared to the RASS scoring group. It is difficult to assess the impact of BIS monitoring on the incidence of adverse events, since there were no such cases in the BIS group, and only two events occurred in the control group. Nevertheless, it seems unlikely to us that BIS monitoring can increase the incidence of complications considering the correlation and higher dynamics of BIS monitoring.

Thus, BIS monitoring of sedation in intensive care units optimizes treatment and increases its safety.

The authors declare no conflicts of interest.