Abstract
Objective
Improved anesthetic agent delivery system (IAADS), a modification of closed-loop anesthesia delivery system (CLADS), is designed to deliver inhalational anesthetics and propofol through closed-loop control with bispectral index (BIS) as target. We compared the performance of IAADS with the manual control isoflurane administration during cardiac surgery.
Methods
Forty patients of ASA (American Society of Anesthesiologists) physical status class II–III, undergoing elective cardiac surgery with cardiopulmonary bypass (CPB) in a tertiary care hospital in India were randomized to receive isoflurane through a closed-loop system (IAADS group) or through a Tech 7 vaporizer adjusted manually (manual group) to achieve a target BIS of 50. Patients were induced with a propofol infusion and isoflurane was started after intubation. During CPB, patients received propofol; isoflurane was restarted after separation from CPB. The efficacy of IAADS in controlling depth of anesthesia and hemodynamic variations was compared with that of manual control.
Results
IAADS was able to maintain BIS within ± 10 of target for significantly longer period (84.6 ± 7.2% in IAADS group vs. 75.9 ± 11.2 in manual group, p < 0.01). Both overall performance, as assessed by global score (p < 0.01), and precision, as judged by median absolute performance error (MDAPE) (p < 0.04), were significantly better in the IAADS group. The IAADS group required significantly less propofol for induction (1.3 ± 0.4 mg/kg in IAADS vs. 1.6 ± 0.5 mg/kg in manual, p < 0.05) and less isoflurane during maintenance of anesthesia (3.3 ± 0.8 ml/h vs. 3.4 ± 0.9 ml/h, p < 0.01).
Conclusion
The present study proves the feasibility and efficacy of inhalation anesthetic administration through closed-loop control. This is the first system that has been developed to control intravenous and inhalational anesthetic agents in a closed-loop model using BIS.
Keywords
bispectral index; cardiac surgical procedures; closed-loop control drug delivery systems; electroencephalography; intraoperative monitoring;
1. Introduction
A better understanding of the pharmacokinetics and pharmacodynamics of anesthetic agents and the availability of accurate monitors for assessing depth of anesthesia, together with improvements in computer technology, have facilitated improved accuracy in the automated control of anesthetic agent delivery.1, 2, 3, 4 Closed-loop control allows greater accuracy than open-loop control, because it can cope with interindividual pharmacokinetic-dynamic differences.5
Most studies of closed-loop anesthesia have used propofol.1, 2, 3, 4, 6, 7 Studies on closed-loop administration of inhalational agents are few,8, 9 probably because of technical difficulties. Closed-loop anesthesia has been administered in open heart surgery,10 a clinical situation that is associated with an increased incidence of intraoperative awareness.11 A closed-loop anesthesia delivery system (CLADS), based on a target bispectral index (BIS) value, has been used for propofol administration in patients undergoing cardiac10 and noncardiac surgery6 and in varying circumstances.12 The system has been upgraded to also administer inhalation anesthetics (Improved Anesthetic Agent Delivery System, IAADS), with added safety measures. We studied the feasibility of using the new system in patients undergoing cardiac surgery and requiring cardiopulmonary bypass (CPB), and compared its performance with conventional manual control of anesthetic agent delivery by the attending anesthesiologist.
2. Methods
2.1. Description of the system
IAADS, the patented anesthetic delivery system [Indian Patent Application no. 2158/DEL/2007 and US Patent Application no.12/683.000 (PCT/IN08/000674)] developed at the Post Graduate Institute of Medical Education and Research (PGIMER), Chandigarh, India, can be used to administer propofol or isoflurane at the operator’s own will to control BIS. The control algorithm for propofol is based on the relation between various rates of propofol infusion, plasma propofol concentration and BIS. The algorithm for inhalational anesthetics is based on the relation between end-tidal concentrations and BIS established during the developmental stages of the system. A personal computer (PC) with Pentium 4 or higher processor was used to implement the control algorithm, provide a user interface and control communication through serial ports (RS 232) with the infusion system (Pilot-C, Fresenius, Brezins, France) and a Datex (Datex-Ohmeda, Helsinki, Finland) vital sign monitor. We used an Avance ventilator (Datex Ohmeda, software version 5.0, Helsinki, Finland) with a CO2 absorber and ascending bellows to provide constant-volume ventilation. The amount of acceptable leak in the circuit was ≤150 ml/min. Low fresh gas flows were used during the study. Isoflurane was injected into a heat and moisture exchanger placed in the expiratory limb of a circle system, via a 20 ml syringe (Dispovan, BD, Singapore) and 100 cm long polyethylene tube (Lectro Cath, Vygon, Gurgaon, India), by a Pilot C infusion pump (Fresenius Vial SA, Brezins, France) (Fig. 1). The stated volumetric accuracy of the infusion pump is ± 2 %.
Download full-size image
IAADS can operate in two different modes: manual and automatic. In the manual mode, the rate of propofol infusion is controlled manually via the keyboard of the PC. In the automatic mode, the system automatically controls the rate of propofol or isoflurane infusion on the basis of BIS feedback (Fig. 2). In addition, the user can limit the maximum allowable rate of drug infusion during induction and maintenance by setting the risk status suitable for the patient.
Download full-size image
2.2. Anesthesia protocol
After obtaining approval for the study from our institution’s ethics committee and written informed consent from the participating patients, 40 ASA physical status class II–III patients aged 18–65 years and scheduled for elective cardiac surgery under CPB were recruited into this prospective randomized trial. The trial was registered with Clinical Trials (www.clinicaltrials.gov-NCT01069562). The study was conducted in a tertiary care teaching hospital. Patients were randomly allocated to two groups, manual group and IAADS group, using computer-generated random numbers in sealed opaque envelopes. Preethy J Mathew generated the sequence for randomization and J. Sethu Madhavan enrolled participants for the interventions. Exclusion criteria were: weight more than ± 30% of the ideal body weight, pre-existing neurological disorder, ejection fraction <30%, anticipated difficult airway, known allergy to propofol and use of psychoactive medications including alcohol.
Premedication included oral diazepam 0.1 mg/kg the night before and on the morning of surgery. Intraoperative monitoring consisted of pulse oximetry, continuous electrocardiography, invasive blood pressure, central venous pressure and pulmonary artery pressure when deemed necessary. A BIS sensor (Aspect Medical Systems Inc., Newton, MA, USA) was placed on the patient’s forehead according to the manufacturer’s instructions, after cleaning with alcohol- based solutions, and attached to the BIS module of an Avance machine (the S/5 BIS module). Morphine 0.1–0.2 mg/kg was administered intravenously to facilitate the insertion of invasive lines. Induction and maintenance of anesthesia was performed according to the randomization sequence. After pre-oxygenation for 3 min anesthesia was induced with a propofol infusion after a bolus of fentanyl 3 μg/kg.
In the IAADS group, the initial and subsequent propofol infusion rates were calculated and delivered by IAADS, based on the height, weight and risk status of the patient and the target BIS. The target BIS value was set at 50. In the manual group, the initial and subsequent propofol infusion rates were set by an attending anesthesiologist to a target BIS of 50. The induction time was defined as the time required to achieve two consecutive BIS values of ≤55 after initiation of the propofol infusion. After intubation, fresh gas flow (FGF) of 100% O2 was set at 1000 ml/min in the manual group and 3% isoflurane delivered through a Tech 7 vaporizer. After 5 min, the FGF was reduced to 500 ml/min and subsequently isoflurane was titrated manually by adjusting the vaporizer dial setting to suit the targeted BIS. In the IAADS group, the FGF was set at 500 ml/min after intubation and isoflurane was started via infusion into the expiratory limb titrated by IAADS. The target BIS was 50 in both the groups. Vecuronium was used for muscle relaxation and normocapnia was maintained. Fentanyl infused at 1 μg/kg/h provided analgesia during surgery. An additional bolus of 1 μg/kg was administered before skin incision, before sternotomy and at the start of CPB. Analgesic supplement (fentanyl 1 μg/kg) was given when the mean arterial pressure (MAP) or heart rate (HR) exceeded 25% of the baseline with BIS in the target range. If hypertension or tachycardia persisted with a BIS ≤ 50, nitroglycerine infusion or esmolol was used appropriately. In the event of hypotension, inotropic support or vasopressor or both were initiated after ensuring normovolemia. Similarly, atropine sulfate was used to treat bradycardia (heart rate of 45 bpm) after excluding other treatable causes. Changes in the rate of isoflurane administration were done by the IAADS system in the closed-loop group. In the manual group, the attending anesthesiologist adjusted the dial setting to maintain BIS in the range of 45 to 55. The number of times that the isoflurane dial setting was changed was noted.
At the start of CPB, isoflurane administration was stopped and propofol infusion started, with the rate of delivery controlled by IAADS in the closed-loop group and by the attending anesthesiologist in the manual group. The MAP on CPB was maintained between 50 and 80 mmHg using phenylephrine/sodium nitroprusside as required. Systemic hypothermia up to 28°C was maintained during CPB and patients were actively re-warmed to 36°C before separation from CPB. At the end of CPB, propofol was stopped and isoflurane administered as per the pre-CPB protocol in both the groups. At the end of the surgery, isoflurane, fentanyl and vecuronium administration was discontinued and the patient shifted to the postsurgical intensive care unit for elective mechanical ventilation. All patients were subjected to a structured interview (as modified from Brice et al13 and described by Nordstrom et al14) for conscious awareness, after extubation of the trachea and approximately a week thereafter.
The BIS and hemodynamic variables were digitally recorded online every 10 s in both the groups for the duration of the study, and stored on a PC for offline analysis. The cumulative time during which IAADS could not function properly, because of malfunction of the infusion system such as leakage, blockage or disconnection, or low signal quality index, high EMG activity or electrical interference causing an error in BIS calculation, were excluded from the total duration of time IAADS was in use (IADDS time) to compute the valid IAADS time. In addition, objective signs of inadequate anesthesia such as sweating, tearing or patient movement under anesthesia were recorded.
2.3. Statistical analysis
The primary outcome measured to assess performance was the duration of time BIS was maintained in the target range of 50 ± 10 expressed as a percentage of the valid IAADS time. From a pilot study, we estimated that manual control maintained BIS in the range of ± 10 of the target, 70% of the time. To achieve a 20% improvement with IAADS, we calculated a sample size of 36 at 80% power and 5% type I error. A total of 40 patients were recruited, taking into account possible inadvertent patient attrition. Secondary outcome measures were assessed by calculating the median performance error (MDPE), median absolute performance error (MDAPE), wobble and divergence using the methods described by Varvel15 and co-workers (Appendix). The global score, which gives an idea of the overall performance of the closed-loop system, was calculated as the sum of MDAPE and wobble divided by the time BIS was within ± 10 of the target.7 The stability of hemodynamic control was assessed by the percentage of valid IAADS time for which the MAP and HR remained within ± 25% of baseline. Data for the pre-CPB period and post-CPB period were analyzed separately and the combination of these two durations is expressed as “off CPB” period.
Data are presented as mean ± SD. Continuous data were analyzed using the Student t test; categorical data were analyzed using the χ2 test. The total dose of propofol and fentanyl, and the total dosage of nitroglycerine, adrenaline and phenylephrine were compared using the Mann–Whitney U-test. Probability values ≤0.05 were considered significant. Data analysis was performed using SPSS version 15.0 (SPSS Science Inc., Chicago, IL, USA).
3. Results
All the 40 patients recruited completed the study protocol over the period January 2009 to June 2010. The demographic profile, baseline and surgical characteristics were similar in both groups (Table 1).
Although the induction time was significantly longer in the IAADS group compared with manual group (p = 0.03), the induction dose of propofol was significantly higher in the manual group, resulting in a greater fall in BIS during induction (Table 2). The maximum BIS after intubation was similar in both groups. During the maintenance phase, the consumption of isoflurane was significantly less in the IAADS group compared with the manual group. There was no significant difference in the amount of fentanyl infused during the entire procedure. The amount of propofol used during CPB and intraoperative inotrope and vasodilator requirements were also similar in both groups (Table 2).
Fresh gas flows were maintained at 500 ml/min in all patients. In the manual control group, the isoflurane dial setting was changed a median of two times (range, 1–4) during the pre-CPB period and two times (range, 0–4) in the post-CPB period.
Mean BIS in the two groups at various times during the procedure is depicted in Fig. 3. The closed-loop system was able to maintain BIS within target for a significantly longer period compared to manual control during isoflurane administration (Table 3). The system performed more accurately (as judged by lower MDAPE) and with less bias (as shown by lower MDPE values) compared with manual group. Although wobble was similar in both the groups, the global score, which reflects the overall performance of the system, was significantly better in the IAADS group than in the manual group (p = 0.02). HR and MAP were similar in both the groups (Table 3).
Download full-size image
Postoperative hemodynamic parameters were similar in both the groups. There was no significant difference in time to extubation between the two groups (10.5 ± 3.8 h in IAADS vs. 10.8 ± 3.7 h in manual, p = 0.81). None of the patients experienced intraoperative awareness. There were no objective signs of a lighter plane of anesthesia such as sweating, patient movement, etc.
4. Discussion
The closed-loop administration of intravenous anesthetics, predominantly using propofol, has been extensively reported in published studies. Liu et al7 demonstrated that the use of a closed-loop system was associated with reduced consumption of propofol, earlier recovery from anesthesia and less variation in BIS. There was also a high rate of manual target modification, which was taken care of by the computer in the closed-loop group, thereby allowing the anesthesiologist to concentrate on other critical areas. Nearly similar results were obtained by our indigenous system, CLADS while administering propofol in patients undergoing cardiac10 and noncardiac6 surgery. Now the spectrum of this system has been extended to include inhalational anesthetics.
Hemodynamic changes caused by anesthetic agents, both intravenous and inhalational, is a major concern for the anesthesiologist; and maintaining an adequate plane of anesthesia so as to avoid a stress response to painful surgical stimuli needs continuous titration of anesthetics. The IAADS, in the present study, performed better than manual control both during induction with propofol and during pre- and post-CPB maintenance with isoflurane. During induction, IAADS used significantly less propofol and caused fewer variations in BIS than occurred in the manual group. During maintenance, BIS was maintained in a narrow range for a significantly longer period. Hemodynamic parameters were within baseline for a longer period with IAADS than with manual control.
Use of inhalational anesthetics for the maintenance of anesthesia carries many advantages over the use of intravenous agents. Inhalational anesthetics have a predictable wash in and wash out compared with intravenous anesthetics. End-tidal anesthetic concentrations can be monitored continuously. Inhalational anesthetics are far more cost effective than intravenous agents when used with low flows.16 In cardiac patients, use of inhalational agent carries the added advantage of ischemic preconditioning, causing less elevation in cardiac enzymes,17 less incidence of arrhythmia18 and a better 1-year survival rate.19 Studies have shown that use of isoflurane-fentanyl based anesthesia for patients undergoing cardiac surgery was associated with faster extubation time compared with propofol-fentanyl based anesthesia.20
In contrast to the literature on intravenous anesthetics, that relating to closed-loop administration of inhalational anesthetics is scant.8 This is partly because of the technical difficulties associated with the servo-controlled delivery of inhalational anesthetics. Current options for administering closed-loop inhalational anesthetics include: (1) high-end machines like the Drager Zeus21 or the PhysioFlex22; (2) servo-motor controlled vaporizers8 that can be operated by the closed-loop system; and (3) injection of liquid anesthetic into the circuit through a syringe pump controlled via the computer. Drager Zeus uses end-tidal concentrations as a target of feedback (closed-loop control of end-tidal concentration), hence cannot account for interindividual pharmacokinetic-dynamic differences. Although infusion of inhalational anesthetic is simple and inexpensive,23 anesthesiologists are reluctant to use the injection technology; they are used to administering inhalational anesthetics through modern vaporizers, and administration through a syringe infusion pump requires precise calculation of the amount of liquid anesthetic required from time to time. By feeding such data into the computer and titrating it to BIS, the closed-loop system could overcome the disadvantages of both frequent titration and interindividual pharmacokinetic-dynamic differences, thus offering a new prospect in closed-loop inhalational anesthetic administration.
Of the available monitors used to assess the depth of anesthesia, extensive reviews have been published on BIS, showing its accuracy in predicting depth of anesthesia better than other EEG-based monitors like entropy24 or 95% spectral edge frequency.25 Use of BIS has been shown to decrease the anesthetic agent consumption,6 and to be associated with earlier awakening6 and a lower incidence of intraoperative awareness.26 By targeting BIS rather than the predicted plasma concentrations or end-tidal concentrations, our system could control the administration of both inhalational and intravenous anesthetics. This is probably the first system of its kind that can target administration of both inhalational and intravenous agents using BIS as the control variable. The BIS-guided IAADS provided clinically adequate anesthesia in all the studied patients.
Cost-effectiveness is an important measure of efficiency when assessing the quality of healthcare services. IAADS used significantly less propofol during induction and less isoflurane during maintenance when compared to the manual control. This was achieved because the IAADS carries out frequent small alterations to maintain the optimal depth of anesthesia, as judged by BIS, and thus avoids very deep or light planes of anesthesia. Although the difference in consumption of anesthetic agent appears small, the cumulative difference over long-duration and multiple surgeries can result in more cost-effective anesthetic administration. Moreover, the number of times isoflurane dial settings were changed manually translates to an involvement of anesthesia human resource in maintaining appropriate depth of anesthesia, making that resource unavailable at crucial steps during the surgery.
The following points may be considered when assessing the clinical relevance of this study. Investigations on the pharmacodynamic interaction between opiates and anesthetics on BIS values are necessary to document the effect of opiates on BIS value. Liu et al27 have devised a dual-loop algorithm in which both remifentanil and propofol can be administered in closed loop, in response to deviation from the target BIS. A similar algorithm for administering fentanyl is incorporated in the present closed-loop system, based on the sharpness of rise in BIS, HR and MAP, but was not tested in this study. The dosage of opioids was comparable in both the groups in our study, so the influence of opioids on BIS is unlikely to have affected the results. The effect of hypothermia on BIS is also controversial. We maintained normothermia during the pre- and post-CPB period and there was no significant difference in temperature between the two groups. The hypothermic CPB period was excluded from part of the analysis and hence would not have confounded the results.
The build-up of isoflurane concentration following induction and after starting isoflurane post-CPB coincides with the disposition kinetics of propofol. Estimating the propofol concentration and eliminating periods before complete disposition would have been ideal, but considering the short context-sensitive half-time of propofol and the low doses infused, we believe it would not be a confounding factor. The infusion of inhalational anesthetic into the breathing circuit has inherent problems, but with the safety features of IAADS that detect and respond to inspired and expired concentrations of inhalational anesthetics and end-tidal carbon dioxide, these were avoided.
Despite the above mentioned limitations, our study does prove the feasibility and efficacy of closed-loop control of anesthesia in cardiac surgery compared to manual control using inhalational anesthetics. In addition, this is the probably the first and the only system reported in literature that can control intravenous and inhalational anesthetic agents in a closed-loop model using BIS.
Acknowledgment
The Department of Information Technology, Government of India, New Delhi, funded the development and upgrading of the - Improved Anesthetic Agent Delivery System (IAADS).
Appendix 1.
Valid IAADS time = total anesthesia time – cumulative time for which IAADS did not function properly.
The secondary outcome measures were assessed by the parameters proposed by Varvel et al and were calculated according to the following equations:
Performance error ðPEÞ ¼ ½ðmeasured BIS target BISÞ=
target BIS 100
Median performance error ðMDPEÞ ¼ fmedian PEij;
j ¼ 1; : : :; Nig
Median absolute performance error ðMDAPEÞ
¼ median fjPEjij; j ¼ 1; : : :; Nig
Wobble ¼ median fjPEij MDPEij; j ¼ 1; : : :; Nig
where i is subject number, j is the jth (one) measurement of the
observation period, and N is the total number of measurements
during the observation period.
Global score ¼ fðMDAPE þ wobbleÞ=% of the time BIS value
was between 40 and 60g 100