Share this article on:

The Effect of Bispectral Index Monitoring on Long-Term Survival in the B-Aware Trial

Leslie, Kate MBBS, MD, MEpi, FANZCA*†; Myles, Paul S. MBBS, MD, MPH, FANZCA, FCARSCI, FRCA‡§‖; Forbes, Andrew MSc, PhD; Chan, Matthew T. V. MBBS, FANZCA#

doi: 10.1213/ANE.0b013e3181c3bfb2
Patient Safety: Research Reports
Chinese Language Editions

BACKGROUND: When anesthesia is titrated using bispectral index (BIS) monitoring, patients generally receive lower doses of hypnotic drugs. Intraoperative hypotension and organ toxicity might be avoided if lower doses of anesthetics are administered, but whether this translates into a reduction in serious morbidity or mortality remains controversial. The B-Aware Trial randomly allocated 2463 patients at high risk of awareness to BIS-guided anesthesia or routine care. We tested the hypothesis that the risks of death, myocardial infarction (MI), and stroke would be lower in patients allocated to BIS-guided management than in those allocated to routine care.

METHODS: The medical records of all patients who had not died within 30 days of surgery were reviewed. The date and cause of death and occurrence of MI or stroke were recorded. A telephone interview was then conducted with all surviving patients. The primary end point of the study was survival.

RESULTS: The median follow-up time was 4.1 (range: 0–6.5) years. Five hundred forty-eight patients (22.2%) had died since the index surgery, 220 patients (8.9%) had an MI, and 115 patients (4.7%) had a stroke. The risk of death in BIS patients was not significantly different than in routine care patients (hazard ratio = 0.86 [95% confidence interval {CI}: 0.72–1.01]; P = 0.07). However, propensity score analysis indicated that the hazard ratio for death in patients who recorded BIS values <40 for >5 min compared with other BIS-monitored patients was 1.41 (95% CI: 1.02–1.95; P = 0.039). In addition, the odds ratios for MI in patients who recorded BIS values <40 for >5 min compared with other BIS-monitored patients was 1.94 (95% CI: 1.12–3.35; P = 0.02) and the odds ratio for stroke was 3.23 (95% CI: 1.29–8.07; P = 0.01).

CONCLUSIONS: Monitoring with BIS and absence of BIS values <40 for >5 min were associated with improved survival and reduced morbidity in patients enrolled in the B-Aware Trial.

Published ahead of print November 12, 2009 Supplemental Digital Content is available in the text.

Authors' affiliations are listed at the end of the article.

Accepted for publication September 14, 2009.

Published ahead of print November 12, 2009

Supported by a Project Grant from the Australian and New Zealand College of Anaesthetists and a Direct Grant for Research from The Chinese University of Hong Kong (Project code 4450149).

Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal's Web site (www.anesthesia-analgesia.org).

Reprints will not be available from the author.

Address correspondence to Kate Leslie, MBBS, MD, MEpi, FANZCA, Department of Anaesthesia and Pain Management, Royal Melbourne Hospital, Parkville, Victoria 3050, Australia. Address e-mail to kate.leslie@mh.org.au.

When anesthesia is titrated using bispectral index (BIS) monitoring, patients generally receive lower doses of hypnotic drugs and, as a consequence, they emerge faster from anesthesia with less postoperative nausea and vomiting.1–3 Intraoperative hypotension and organ toxicity might also be avoided if lower doses of anesthetics are administered, but whether this translates into a reduction in serious morbidity or mortality remains controversial.4

Two observational studies have explored this issue (Table 1). Monk et al.5 studied 1046 patients undergoing noncardiac surgery. BIS values were recorded in all patients, but anesthesiologists were blinded to the BIS values. Cumulative deep hypnotic time was defined as the time that BIS was <45 and was a significant predictor of 1-yr mortality (relative risk = 1.24 [95% confidence interval {CI}: 1.06–1.44]). Similarly, Lindholm et al.6 studied 4087 patients undergoing noncardiac surgery whose anesthesia was titrated using BIS monitoring. Cumulative deep hypnotic time was a significant predictor of 2-yr mortality, but only when preexisting malignancy was excluded (hazard ratio = 1.18 [95% CI: 1.08–1.29]). The principal mode of anesthetic maintenance in these studies was based on inhaled anesthesia, and 30-day mortality was 0.7%.

Table 1

Table 1

The B-Aware Trial was a large multicenter trial that compared the incidence of awareness during anesthesia in patients receiving BIS-guided anesthesia or routine care.7 The trial differed from the aforementioned studies in several ways: 1) patients were randomly allocated to BIS-guided anesthesia or routine care; 2) 42% of patients were undergoing cardiac surgery; 3) 43% of patients received propofol for anesthetic maintenance; and 4) 30-day mortality was much higher (4.1%) because of the inclusion of high-risk patients. After publication of the data of Monk et al.5 in 2005, we planned a long-term follow-up and post hoc analysis of the B-Aware Trial patients to test the hypothesis that survival would be improved in patients receiving BIS-guided anesthesia over those undergoing routine care. In addition, we explored the effect of BIS monitoring on myocardial infarction (MI) and stroke in secondary analyses.

Back to Top | Article Outline

METHODS

The B-Aware Trial was a multicenter, randomized controlled trial of BIS monitoring to prevent awareness during general anesthesia in patients at high risk of awareness.7 In the original trial, 2463 surgical patients aged 18 yr or older who were at high risk of awareness (i.e., cesarean delivery, high-risk cardiac surgery, acute trauma with hypovolemia, rigid bronchoscopy, significant impairment of cardiovascular status and expected intraoperative hypotension requiring treatment, severe end-stage lung disease, history of awareness, anticipated difficult intubation, known or suspected heavy alcohol intake, chronic benzodiazepine or opioid use, or current protease inhibitor therapy) were randomized to BIS monitoring (target range: 40–60) or routine care. In all other respects, perioperative care was at the discretion of the anesthesiologist. For each patient allocated to the BIS group, BIS values were manually recorded by the anesthesiologists at 10-min intervals for the first hour, and every 30 min thereafter. The time-averaged mean BIS value was calculated from these values. In addition, the highest and lowest BIS readings that persisted for at least 5 min were recorded by the anesthesiologist. BIS was not monitored in the routine care group. Serious hypotension was defined as a systolic blood pressure <90 mm Hg and requiring treatment. Patients were interviewed about awareness on 3 occasions (2–6 h, 24–36 h, and 30 days postoperatively) and were followed up regarding complications with a medical record review and interview on postoperative day 30.

Ethics Committee approval was obtained at each participating site for long-term follow-up. In view of the patient consent obtained for inclusion in the original B-Aware Trial, and the opportunity for patients to refuse ongoing participation in this study at the time of follow-up, a second consent process was waived by all centers. The medical records of all patients who had not died within 30 days of surgery were reviewed. The date and cause of death and occurrence of MI or stroke were recorded. A telephone interview was attempted in all surviving patients. Verbal consent was obtained before asking patients about the occurrence of MI and stroke because the index surgery was using a standardized questionnaire. If the patient had died, the patient's relatives or family doctors were questioned, after verbal consent was obtained, using the same questionnaire. At least 3 attempts were made to contact patients using the contact details on the original case report form and the hospitals' information management systems. If these attempts failed, at least 3 attempts were made to contact the patients' relatives and/or family doctors.

The primary end point of the study was survival, which was recorded as the time to last confirmed contact with the patient (either during the awareness interview or long-term follow-up interview) or time of death. Secondary end points included MI and stroke. MI was defined by a typical increase in plasma troponin concentration with or without electrocardiographic changes. Stroke was defined as a new neurological deficit on examination with or without computerized tomographic or magnetic resonance imaging confirmation.

Back to Top | Article Outline

Data Analysis

Survival rates were computed for each category of each predictor and were expressed as deaths per 1000 person-years. Univariate Cox proportional hazard models were used to define hazard ratios and 95% CIs. All predictors were included in a multivariable Cox proportional hazard model, with the assessment of proportionality of hazard functions performed by separating follow-up time into tertiles with equal numbers of deaths (0–6 mo, 6 mo to 2 yr, and >2 yr) as described in Appendix 1 (see Supplemental Digital Content 1, http://links.lww.com/AA/A41). The effect of BIS monitoring was measured in 2 ways: 1) comparing BIS-randomized patients to routine care patients, and 2) separating the BIS-randomized patients into those who recorded BIS values <40 for >5 min and those who did not. Logistic regression was used to compute odds ratios and 95% CI for MI and stroke, with all predictors included in multivariate models. We did not use Cox proportional hazards models, because the date of these events was missing in many patients (e.g., the patient had forgotten, his/her relatives did not know, or the event was silent and was detected on later testing). The preplanned assessment of the interaction of each variable with BIS monitoring status was performed using interaction terms in the regression models. We also performed a propensity score analysis using inverse probability weighting to assess the effect of BIS <40 for >5 min versus other BIS monitoring on death, MI, and stroke. Details are provided in Appendix 2 (see Supplemental Digital Content 2, http://links.lww.com/AA/A42). Analyses were conducted using Stata 10.2 (Stata, College Station, TX). All P values are 2-sided and P < 0.05 was considered statistically significant.

Back to Top | Article Outline

RESULTS

Recruitment to the original B-Aware Trial occurred between September 2000 and December 2002. Long-term follow-up occurred between March 1, 2006 and October 21, 2007, with a median follow-up of 4.1 (range: 0–6.5) yr. No follow-up attempt was made for 301 patients because of lack of resources at the recruiting centers. The follow-up time for these patients was recorded as the time of the last awareness interview. At the time of the long-term follow-up, 108 patients could not be contacted and 2 patients declined further participation. The follow-up time of these patients was also recorded as the time of the last awareness interview. Long-term follow-up data were therefore obtained for 1947 (83%) of the study patients (Fig. 1).

Figure 1

Figure 1

Five hundred forty-eight patients (22.2%) had died since surgery (4.3% by Day 30, 10.8% by 1 yr, and 14.6% by 2 yr). The death rate per 1000 patient-years was 68 (95% CI: 63–74). The causes of death were cancer (39%), MI (9%), stroke (4%), other cardiovascular disease (13%), respiratory failure (9%), sepsis (9%), other causes (8%), and unknown (9%). Two hundred twenty patients (8.9%) had an MI and 115 patients (4.7%) had a stroke during the entire follow-up period.

The risk of death in patients randomly allocated to BIS monitoring was not significantly different from the risk of death in patients allocated to routine care (Table 2). The hazard ratio for patients with BIS <40 for >5 min compared with routine care patients also was not statistically significantly different. However, patients who were BIS monitored and in whom BIS values were never <40 for >5 min were at significantly lower risk of death than patients who were not monitored with BIS. In addition, propensity score analysis indicated that the hazard ratio for death in patients who recorded BIS values <40 for >5 min compared with other patients in the BIS group was 1.41 (95% CI: 1.02–1.95; P = 0.039) (Fig. 2 and Appendix 2). There were no interactions among variables in the model (i.e., the effect of BIS monitoring did not differ for different levels of the variables).

Table 2

Table 2

Figure 2

Figure 2

The risk of MI and stroke in patients was not significantly different in patients randomized to BIS monitoring or routine care (Tables 3 and 4). The odds ratios for patients with BIS <40 for >5 min compared with routine care patients also were not statistically significantly different. However, propensity score analysis indicated that patients who were BIS monitored and in whom BIS values were never <40 for >5 min were at significantly lower risk of MI and stroke than patients who were not BIS monitored. The odds ratios for MI in patients who recorded BIS values <40 for >5 min compared with other BIS-monitored patients was 1.94 (95% CI: 1.12–3.35; P = 0.02) and the odds ratio for stroke was 3.23 (95% CI: 1.29–8.07; P = 0.01). When assessing MI, the only significant interaction was between BIS monitoring and emergency surgery. Compared with standard care (no BIS monitoring), the odds ratio for MI was 0.69 (95% CI: 0.40–1.19) if BIS was never <40 for >5 min and 1.14 (95% CI: 0.96–2.86) if BIS was <40 for >5 min for nonemergency surgery. In contrast, the odds ratio for MI was 0.19 (95% CI: 0.04–0.88) if BIS was never <40 for >5 min and 0.37 (95% CI: 0.16–0.85) if BIS was <40 for >5 min for emergency surgery.

Table 3

Table 3

Table 4

Table 4

Serious hypotension was recorded in 56% of patients who were not BIS monitored, 54% of patients with BIS never <40 for >5 min, and 60% of patients with BIS <40 for >5 min (P = 0.06).

Back to Top | Article Outline

DISCUSSION

Monitoring with BIS and absence of BIS values <40 for >5 min were associated with improved survival and less serious morbidity in patients enrolled in the B-Aware Trial. One explanation for this result is that simply monitoring the depth of anesthesia is insufficient: anesthesiologists must also avoid low BIS values by careful titration of hypnotic drugs to affect long-term outcome. An alternative or additional explanation is that low BIS values reflect underlying disease processes and trauma and that these factors, rather than anesthetic dosing, affect survival.5

Our results are consistent with those of Monk et al.5 and Lindholm et al.,6 despite differences in study population and study design. The B-Aware Trial patients were sicker than patients in the other studies and had higher rates of postoperative death and serious morbidity. In addition, our patients were scheduled for cardiac and noncardiac surgery, and anesthesia was maintained with propofol as well as volatile anesthetics. Most importantly, our patients were randomly allocated to BIS monitoring or routine care, so we were able to compare patients with and without monitoring as well as assess the effect of anesthetic depth. Nevertheless, our results should be extrapolated to other patient groups with caution.

In the B-Aware Trial, anesthesiologists were requested to target BIS values between 40 and 60 during surgery in the BIS-monitored group and to record whether BIS values were <40 for >5 min.7 We therefore assessed the effect of BIS <40 on survival, rather than BIS <45 as previous studies have done.5,6 Furthermore, in our study, BIS values were recorded intermittently during surgery. For this reason, we used the dichotomous variable “BIS <40 for >5 min” as a secondary factor of interest in our analyses. Monk et al.5 and Lindholm et al.6 collected BIS values every 5 min and 1 min, respectively, and were able to calculate cumulative deep hypnotic time, whereas we were not.

The median follow-up time in our study (4.1 yr) was longer than in previous studies.5,6 A predictor (such as BIS monitoring) may not have the same effect at reducing mortality in patients who die shortly after surgery compared with those who die many years after surgery. For this reason, we conducted analyses looking at differing effects of BIS monitoring over time. These results did not reveal nonproportionality of hazards for BIS-monitored and non–BIS-monitored patients over time (P = 0.54) (details not reported here). Despite all these differences in study population and design, our results are consistent with the studies of Monk et al.5 and Lindholm et al.6

Other significant predictors of death in our study have been reported, including increasing age,8–10 increasing ASA physical status score,8,10,11 noncardiac surgery,8,12,13 emergency surgery,11 and severe intraoperative hypotension.5 Our result that shorter duration of surgery leads to higher mortality contradicts previous work9 but arises from the inclusion criteria (i.e., patients having short procedures for the investigation of cancer). Lower weight possibly reflects the presence of malignancy (which we did not specifically record).14 The retention of the “omission of nitrous oxide” predictor in the multivariate model likely reflects residual unidentified confounding,15 because anesthesiologists were likely to omit nitrous oxide in patients whom they perceived to be at high risk perioperatively.

Monitoring BIS and absence of BIS <40 for >5 min also reduced the incidence of MI and stroke during long-term follow-up in our patients. The rates of MI (8.9%) and stroke (4.7%) in our study are consistent with published data in patients who have or are at risk of ischemic heart disease, undergoing major cardiac and noncardiac surgery.13,16,17 Analysis of an interaction with emergency status suggests a benefit of BIS monitoring in emergency patients; this interaction requires confirmation in future studies.

A limitation of previous studies,5,6 and ours, is that patients were not randomized to different BIS values during surgery. Although we performed a detailed propensity score analysis of the effect of BIS <40 for >5 min, we were limited by the data available, and indeed, there will be unmeasured factors influencing BIS monitoring that may be alternate explanations for our findings. In particular, we did not record anesthesiologists' explanations for episodes of low BIS values, so we cannot comment on whether these episodes were intentional, unintentional, or out of the anesthesiologists' control. In addition, because anesthesia technique was at the discretion of the anesthesiologists and patients received various combinations of volatile anesthetics, IV hypnotics, and potent opioids, we cannot report a unifying measure of hypnotic drug delivery (such as minimum alveolar concentration equivalents) in our patients. We therefore cannot report whether patients with BIS values <40 for >5 min received more anesthesia or were more sensitive to anesthetic drugs. Similarly, we cannot draw conclusions about causality between BIS values and serious hypotension. These questions require illumination by an appropriately designed randomized trial.

Finally, BIS data were manually recorded by the anesthesiologists, which may have introduced bias. However, data collection for the B-Aware Trial was almost complete when the first report of an association between anesthetic depth and mortality was published.18 Furthermore, a systematic bias is unlikely given that the case report form was completed by hundreds of different anesthesiologists.

In conclusion, monitoring with BIS and absence of BIS values <40 for >5 min were associated with improved survival and reduced morbidity in patients enrolled in the B-Aware Trial. The exact relationship between anesthetic and mortality awaits confirmation by a suitably designed randomized controlled trial.

Back to Top | Article Outline

AUTHORS' AFFILIATIONS

From the *Department of Anaesthesia and Pain Management, Royal Melbourne Hospital; †Department of Pharmacology, University of Melbourne; ‡Department of Anaesthesia and Perioperative Medicine, Alfred Hospital; §Academic Board of Anaesthesia and Perioperative Medicine, Monash University, Melbourne; ‖National Health and Medical Research Council, Canberra; ¶Department of Epidemiology and Preventive Medicine, Monash University, Melbourne, Australia; and #Department of Anaesthesia and Intensive Care, The Chinese University of Hong Kong, Hong Kong, People's Republic of China.

Back to Top | Article Outline

DISCLOSURE

The original B-Aware Trial received sensors and some unrestricted funding from Aspect Medical Systems. Loaned BIS monitors were returned at the end of the original trial. Aspect Medical Systems had no role in the concept, design, data collection, data analysis, data interpretation, or writing of reports of either the original trial or the follow-up studies.

Back to Top | Article Outline

ACKNOWLEDGMENTS

We gratefully acknowledge the contributions of the B-Aware Trial Investigators who undertook the patient follow-ups for this study.

Back to Top | Article Outline

REFERENCES

1.Liu S. Effects of bispectral index monitoring on ambulatory anesthesia: a meta-analysis of randomized controlled trials and a cost analysis. Anesthesiology 2004;101:311–5
2.Leslie K, Myles P, Forbes A, Chan M, Short T, Swallow S. Recovery from bispectral index-guided anaesthesia in a large randomised controlled trial of patients at high risk of awareness. Anaesth Intensive Care 2005;33:443–51
3.Leslie K, Myles P, Chan M, Paech M, Peyton P, Forbes A, McKenzie D. Severe post-operative nausea and vomiting in a randomized trial of nitrous oxide-based versus nitrous oxide-free anesthesia. Br J Anaesth 2008;101:498–505
4.Cohen N. Anesthetic depth is not (yet) a predictor of mortality! Anesth Analg 2005;100:1–3
5.Monk T, Saini V, Weldon B, Sigl J. Anesthetic management and one-year mortality after noncardiac surgery. Anesth Analg 2005;100:4–10
6.Lindholm M, Traff S, Granath F, Greenwald S, Ekbom A, Lennmarken C, Sandin R. Mortality within 2 years after surgery in relation to low intraoperative bispectral index values and preexisting malignant disease. Anesth Analg 2009;108:508–12
7.Myles P, Leslie K, McNeil J, Forbes A, Chan M. A randomised controlled trial of BIS monitoring to prevent awareness during anaesthesia: the B-Aware randomised controlled trial. Lancet 2004;363:1757–63
8.McNicol L, Story D, Leslie K, Myles P, Fink M, Shelton A, Clavisi O, Poustie S. Postoperative complications and mortality in older patients having non-cardiac surgery at three Melbourne teaching hospitals. Med J Aust 2007;186:447–52
9.Turrentine F, Wang H, Simpson V, Jones R. Surgical risk factors, morbidity, and mortality in elderly patients. J Am Coll Surg 2006;203:865–77
10.Hamel M, Henderson W, Khuri S, Daley J. Surgical outcomes for patients aged 80 and older: morbidity and mortality from major noncardiac surgery. J Am Geriatr Soc 2005;53:424–9
11.Khuri S, Daley J, Henderson W, Hur K, Gibbs J, Barbour G, Demakis J, Irvin G III, Stremple J, Grover F, McDonald G, Passaro EJ, Fabri P, Spencer J, Hammermeister K, Aust J. Risk adjustment of the postoperative mortality rate for the comparative assessment of the quality of surgical care: results of the National Veterans Affairs Surgical Risk Study. J Am Coll Surg 1997;185:325–38
12.Dewey T, Brown D, Ryan W, Herbert M, Prince S, Mack M. Reliability of risk algorithms in predicting early and late operative outcomes in high-risk patients undergoing aortic valve replacement. J Thorac Cardiovasc Surg 2008;135:180–7
13.POISE Study Group, Devereaux PJ, Yang H, Yusuf S, Guyatt G, Leslie K, Villar JC, Xavier D, Chrolavicius S, Greenspan L, Pogue J, Pais P, Liu L, Xu S, Malaga G, Avezum A, Chan M, Montori VM, Jacka M, Choi P. Effects of extended-release metoprolol succinate in patients undergoing non-cardiac surgery (POISE trial): a randomised controlled trial. Lancet 2008;371:1839–47
14.Mullen J, Davenport D, Hutter M, Hosokawa P, Henderson W, Khuri S, Moorman D. Impact of body mass index on perioperative outcomes in patients undergoing major intra-abdominal cancer surgery. Ann Surg Oncol 2008;15:2164–72
15.Datta M. You cannot exclude the explanation you have not considered. Lancet 1993;342:345–7
16.Devereaux P, Goldman L, Cook D, Gilbert K, Leslie K, Guyatt G. Perioperative cardiac events in patients undergoing noncardiac surgery: a review of the magnitude of the problem, the pathophysiology of the events and methods to estimate and communicate risk. Can Med Assoc J 2005;173:627–34
17.Serruys P, Morice M, Kappetein A, Colombo A, Holmes D, Mack M, Stahle E, Feldman T, van den Brand M, Bass E, Van Dyck N, Leadley K, Dawkins K, Mohr F. Percutaneous coronary intervention versus coronary artery-bypass grafting for severe coronary artery disease. N Engl J Med 2009;360: 961–72
18.Weldon B, Mahla M, van der Aa M, Monk T. Advancing age and deeper intraoperative anesthetic levels are associated with higher first year death rates (abstract). Anesthesiology 2002;96:A1097
19.Robins J, Hernan M, Brumback B. Marginal structural models and causal inference in epidemiology. Epidemiology 2000;11: 550–60
    20.Forbes A, Shortreed S. Inverse probability weighted estimation of the marginal odds ratio: correspondence regarding ‘the performance of different propensity score methods for estimating marginal odds ratios.’ Stat Med 2008;27:5556–9
      21.Austin P, Mamdani M. A comparison of propensity score methods: a case study estimating the effectiveness of post AMI statin use. Stat Med 2008;25:2084–106

        Supplemental Digital Content

        Back to Top | Article Outline
        © 2010 International Anesthesia Research Society