Abstract

Objective

The introduction of the bispectral index (BIS) as a comprehensive indicator of depth of anesthesia has prompted research in the automated delivery of anesthetics. This trial aimed to evaluate the usefulness of the BIS as the feedback variable in a closed-loop anesthesia control system during cardiac surgery under cardiopulmonary bypass (CPB).

Methods

Thirty-four adult patients of American Society of Anesthesiologists grade I-III undergoing elective cardiac surgery under hypothermic CPB were evaluated prospectively in a tertiary care teaching hospital. The anesthesia was induced and maintained throughout surgery using a closed-loop anesthesia delivery system (502/DEL/2003) to control the BIS.

Results

The closed-loop system was functional during 96% of the anesthesia duration. The BIS was maintained at ± 10 of the target of 50 during 86% of the automated anesthesia. The closed-loop system was also functional during CPB despite a fall in BIS probably resulting from hypothermia. None of the patients had recall of events or experiences during the procedure.

Conclusion

The closed-loop control of anesthetic delivery adjusted to BIS is feasible and may be useful in open heart surgery under hypothermic CPB.

Keywords

cardiac surgical procedures; electroencephalography: bispectral index; feedback: closed-loop control;

---

1. Introduction

Conventionally, the depth of anesthesia during sur-gery is controlled by the attending anesthesiologist. In a closed-loop anesthesia delivery system (CLADS), depth of anesthesia is controlled automatically by microprocessors.  The  CLADS  functions  like  an  au-topilot maintaining the same efficiency throughout its use.¹ The crucial component of CLADS is a con-trolled variable to measure the depth of anesthesia wherein a target point can be set with control of the drug delivery system. This comprises an algorithm to translate the measured value of the controlled variable to a particular action for the drug delivery system, thus steering the controlled variable closer to the set target point. The closed loop is completed in the patient and the drug delivery system, usu-ally an infusion pump that drives the administration of drug.

Of the various electroencephalic-based indices,¹⁻⁵ the bispectral index (BIS) is the most promising mea-sure  for  indicating  depth  of  consciousness^6,7  be-cause of its relative ease of use and noninvasiveness. BIS is a dimensionless number scaled from 100 to 0, with 100 representing an awake encephalogram and 0 representing electrical silence. Although closed-loop anesthesia has been described in various sur-gical  settings,^1,3,5,8,9  it  has  not  been  evaluated  in open heart surgery where cardiopulmonary bypass (CPB), hypothermia and rewarming can alter anes-thetic depth unpredictably. Therefore, we aimed to evaluate a CLADS controlling the pharmacodynamic variable BIS in open heart surgery.

2. Methods

The  CLADS  was  a  patented  (502/DEL/2003)  drug  delivery system used for induction and maintenance of propofol anesthesia.⁹ BIS was obtained from an A-2000 monitor (version 3.0 rev 5, Aspect Medical Systems Inc., Norwood, MA, USA) and the drug de-livery system comprised an infusion system of Pilot-C (Fresenius, Paris, France). The “control algorithm” was based on the relation between various rates of propofol infusion (producing different plasma con-centrations) and BIS taking into consideration the pharmacokinetic variables (distribution and clear-ance) which were established in the developmen-tal  stage  of  CLADS.  The  controller  algorithm  was  implemented by an IBM compatible Pentium 4 PC which provided a user interface and received input from the patient’s BIS through serial ports (RS 232) and controlled output to the infusion system. The target BIS value and risk status of each individual patient  (generally  American  Society  of  Anesthesi-ologists  [ASA]  physical  status  IV,  New  York  Heart  Association III or above classified as high risk, ASA physical status I−III classified as low risk) were en-tered into the system before anesthesia. The con-troller algorithm could limit the maximum allowable rate of propofol infusion according to risk status of each patient. The controller algorithm received an update BIS data every 5 seconds and used the BIS error (actual BIS − target BIS) to calculate the rate of propofol infusion to steer the BIS closer to target.

After  obtaining  approval  of  the  Institutional  Ethics  Committee  and  patients’  written  informed  consent,  a  prospective  study  was  undertaken  in  adult patients aged 18−65 years, ASA physical sta-tus II−III, scheduled for elective open heart surgery under  CPB,  hypothermia  and  general  anesthesia.  Patients with weight more or less than 30% of ideal body weight, a neurological disorder, and those on any psychoactive medication, including alcohol were excluded.

Oral diazepam 5−10 mg was used for premedica-tion. Intraoperative monitoring included BIS, pulse oximetry,  continuous  electrocardiography,  neuro-muscular blockade, invasive blood pressure, central venous pressure and pulmonary artery parameters when  deemed  necessary.  Morphine  0.1−0.2 mg/kg i.v. was used to facilitate the invasive monitoring procedure. Fentanyl 3 μg/kg was administered be-fore induction. After the target BIS value was set at 50, propofol infusion was started. Based on height and  weight  of  the  patient  and  target  BIS,  CLADS  automatically calculated and titrated the initial and subsequent propofol infusion rates. The induction time was defined as the time required to achieve the target BIS after initiation of infusion.

Vecuronium  was  used  for  muscle  relaxation  to  maintain the “Train of Four” count ≤ 2. Fentanyl at 1 μg/kg/hr  was  given  to  provide  analgesia.  If  the  mean arterial pressure (MAP) or heart rate exceeded 25% of baseline, analgesia was supplemented with a 1 μg/kg bolus of fentanyl. In case hypertension or tachycardia persisted with BIS ≤ 50,  nitroglycerine infusion or esmolol was administered appropriately. In case of hypotension (MAP < 25% of baseline with normovolemia), inotropic support and/or a vasopres-sor were initiated. Similarly, atropine sulfate was used  to  treat  bradycardia.  During  CPB,  MAP  was  maintained at 50−80 mmHg. At the end of the sur-gery, CLADS was discontinued and the patient was transferred to the intensive care unit (ICU) for elec-tive mechanical ventilation. All patients were sub-jected  to  a  structured  interview  as  described  by  Nordstrom  et  al¹⁰  for  conscious  awareness  after  extubation of the trachea and 1 week later.

CLADS  is  designed  such  that  the  closed  loop  becomes complete and hence functional only dur-ing the periods of time when the BIS has adequate signal quality index (SQI > 30). The cumulative time during  which  CLADS  could  not  function  properly,  because of low SQI or malfunction of the infusion system such as leakage, blockage or disconnection, was deducted from the total duration of the pro-cedure  to  calculate  valid  CLADS  time.  The  per-formance of the system was assessed using methods described by Varvel et al¹¹ (Appendix). The percent-age of time BIS remained within ± 10 of the target BIS was calculated. Data are presented as mean ± standard  deviation  or  percentage.  Data  analysis  was performed using SPSS version 11.0 (SPSS Inc., Chicago, IL, USA).

3. Results

Thirty-four patients who fulfilled the inclusion crite-ria and were free of exclusion criteria were studied from  December  2006  to  March  2007.  The  charac-teristics and clinical data of patients are shown in Table 1. None of the patients had liver dysfunction preoperatively. Blood urea nitrogen was deranged (>  20  mg/dL)  in  two  patients,  although  the  urine  output was normal.

During  induction,  the  target  BIS  was  achieved  in 160 ± 51.2 seconds after initiation of propofol in-fusion, although loss of consciousness occurred at a BIS of 78.1 ± 4.9. The lowest BIS during induction was 40.2 ± 4.0 at 266.4 ± 52.8 seconds. This overshoot in  BIS  was  corrected  to  target  BIS  in  75.3  ± 49.1 seconds.  The  MAP  as  recorded  saw  a  peak  fall  of  24.2% during induction. Following intubation, skin incision  and  sternotomy,  BIS  exceeded  the  target  by  9.6  ± 5.6,  10.6 ±  5.2  and  7.9  ± 4.5,  respectively. CLADS responded to regain target BIS in 70.6 ± 35.4, 57.3 ± 28.4 and 45.1 ± 20.2 seconds, respectively.

CLADS was functional during 96.1% of the pro-cedure which lasted approximately 5.5 hours. Two patients required the aid of a pacemaker upon dis-continuation of CPB that increased the electromyo-gram  in  spite  of  adequate  SQI.  The  anesthesia  in  these patients was switched over to manual control. The  data  from  these  two  patients  were  excluded  from the analysis. All the BIS values from induction to discontinuation of propofol infusion are presented in Figures 1A and 1B.

During  CPB,  the  BIS  tended  to  drift  below  50  because of hypothermia [BIS median (IQR) = 39  (5.5)] during  which  CLADS  was  functional  and  propofol  infusion  was  cut  off  according  to  feedback.  The  requirements  of  propofol  and  fentanyl  during  an-esthesia are shown in Table 2. Excluding the dura-tion  of  CPB,  BIS  was  maintained  at  the  adequate  level of 40−60 during the majority of the procedure (86 ± 4.9%). The median performance error (MDPE), median  absolute  performance  error  and  Wobble  are shown in Table 3. Similarly, Table 4 shows the hemodynamic performance of the patients during CLADS. A pulmonary artery catheter was not used in  patients  who  underwent  repair  of  congenital  heart defects. Therefore, the data on cardiac index and pulmonary artery pressures indicate the aver-age of patients who underwent coronary artery by-pass  grafting  and  valvular  replacement  surgery  (n = 31).

Postoperatively, all the patients were mechani-cally ventilated electively for an average duration 126 G.D. Puri, P.J. Mathewof 489.3 ± 266.5 minutes and stayed in the postop-erative ICU for 1149.7 ± 342.1 minutes. None of the patients  had  recall  of  the  events  during  surgery  when  questioned  in  the  ICU  or  1  week  after  the  procedure.

4. Discussion

The  current  study  demonstrates  the  utility  of  a  closed-loop  feedback  control  system  for  propofol  anesthesia using BIS in cardiac surgery under CPB. The documented advantages of a closed-loop sys-tem in noncardiac surgery include decreased con-sumption of propofol, lesser overshoots of BIS and better hemodynamic performance.^1,4,5,9 The need to test the closed-loop system in extreme circum-stances to fully establish its safety, efficacy and util-ity has been stated.¹² Open heart surgery presents an extreme rearrangement of premorbid physiology and is therefore an appropriate context to evaluate these features.

CLADS was not functional  during  3.92%  of  the anesthesia time because of poor SQI resulting from electrocautery. However, these periods were brief and  did  not  interfere  with  the  conduct  of  the  closed-loop system. The MDPE of −2% indicated that the system had a slight negative bias, i.e., median measured BIS was 2% less than target BIS. A median absolute performance error value of 11.3% indicated that 50% of BIS values were within 11.3% of target BIS. Median Wobble was 8%. These figures are com-parable to the performance of systems reported by Struys et al,⁴ Absalom et al⁵ and Puri et al.⁹

CPB alters the pharmacokinetics of propofol as the physiological processes of distribution, metab-olism and elimination are affected by altered con-ditions of hemodilution, hypotension, hypothermia and non-pulsatile flow. The studies on pharmacoki-netics and elimination of propofol during CPB have yielded conflicting results.¹³⁻¹⁶ The lower BIS values during CPB could also result from brain cooling and reduced electrical activity¹⁷ apart from an increased level of free propofol blood concentrations result-ing  from  a  decrease  in  metabolism  and  altered  pharmacokinetics during hypothermic CPB.

Relying entirely on normal pharmacokinetics to control the depth of anesthesia may lead to inap-propriately high or low doses of propofol during open heart  surgery.  Therefore,  Schmidilin  et  al¹⁸  sug-gested that a controlled infusion of propofol with a target BIS value may be effective in overcoming the variability in propofol requirement during CPB. Our  system  incorporates  this  suggestion,  and  the  fact that none of the patients experienced aware-ness  validates  the  utility  of  this  system  for  open  heart procedures.

The heart rate and blood pressure were main-tained within ± 25% of the baseline during 83% and 89% of the pre-CPB period and 91% and 83% of the post-CPB period. These variations may be explained as the result of desirable hemodynamic manipula-tions  required  during  the  surgery,  e.g.,  a  higher  heart rate during post-CPB or a lower blood pres-sure during cardiac cannulation.

Our current trial was limited by its prospective and observational design. Also, the number of pa-tients studied was small as the aim was to evaluate the system in a demanding surgical setting. A ran-domized, controlled trial to assess the advantages of closed loop control of anesthesia over conven-tional manual control of anesthesia in open heart surgery is warranted.

In conclusion, we have demonstrated that auto-mated delivery of anesthetic based on BIS feedback from the patient is feasible and may be a valuable aid to the anesthesiologist to devote attention to better hemodynamic control during cardiac surgery.

Acknowledgments

This  study  was  supported  by  a  grant  from  the  Department of Information Technology, Ministry of Electronics  and  Communication,  Government  of  India, New Delhi.