Low Bispectral Index Values and Death

Screaming headlines in the public press recently announced that research had solved the age-old riddle of which came first, “the chicken or the egg”—it was the chicken!1 This classic causality dilemma, together with its classic misreporting in the popular press, exemplifies the situation in which something that is associated with a phenomenon is asserted to be its underlying cause, and if repeated enough gathers its own truth. Other examples of the causality dilemma are determining which came first, reduced spending or economic downturn, or lack of work experience and unemployment. Although finding an association does not imply causality, it does not exclude it either: for example, the association between smoking and lung cancer has been found to be causal, but only after >20 years of research, which ranged from detailed epidemiology to the biochemistry of carcinogens and oncogenes.2,3 The processed electroencephalographic (EEG) indices that are available for monitoring depth of anesthesia identify deep general anesthesia with a low index value. In anesthesia, a causality dilemma exists as to whether low index values during general anesthesia lead to poor outcomes or whether already fragile health results in low index values.4–8 In this report, we discuss the relationships among low depth of anesthesia index values, anesthetic administration, and other factors that may influence the long-term outcomes from general anesthesia and explore the role of a novel concept we have named “relative anesthetic overdose” (Fig. 1).Figure 1: Model of low processed electroencephalographic (EEG) index values and their relationship to perioperative morbidity and mortality. Organ dysfunction = organ disease, genetic malfunction, or toxicity.RELATIVE ANESTHETIC OVERDOSE Anesthetic dose requirements vary widely from person to person, and the anesthesiologist must make an estimate of what constitutes adequate anesthesia for each individual patient. This will often be close to the end-tidal anesthetic gas concentration or predicted effect-site concentration that causes unconsciousness in 90% of patients.9–11 It follows that (just like Goldilocks, whose porridge was “too hot,” “too cold,” or “just right”) this concentration will be inadequate for only a few, just right for some, and excessive to the minimal needs of many patients. Even with adjustment for age, weight, and other drugs, it is not possible to predict the anesthetic requirement of individual patients.12 In addition, the right dose or concentration for a patient in one situation may prove to be excessive for the same patient in another situation.9,13–15 For example, lower doses or concentrations of hypnotic are required when a patient receives high-dose opioids than when the same patient receives low-dose opioids, or when a patient is severely hypotensive compared with when they are normotensive. This between- and within-patient variability can result in what we have called “relative anesthetic overdose,” which is the situation in which the dose of anesthetic that is appropriate for most patients results in excessively deep anesthesia in other patients; or when a dose that is appropriate for an individual patient in one situation is excessive for that patient in another situation. DEPTH OF ANESTHESIA MONITORS AND ANESTHETIC DOSING A monotonic relationship exists between processed EEG index values and the effect-site concentration of potent inhaled anesthetics or propofol.16 That is, although this relationship is not linear, increasing doses of these hypnotic drugs cause a decrease in the index value. Early studies identified that use of processed EEG monitoring was associated with the use of lower doses of anesthetic drugs and more rapid emergence from anesthesia when compared with routine care.17,18 The reason for this result was that processed EEG monitoring could identify patients who were receiving anesthetic doses in excess of their needs, irrespective of the cause. This means that there is an association among “no processed EEG titration,” “relative anesthetic overdose,” and “low index value” (Fig. 1). However, lack of processed EEG titration is not the only cause of relative anesthetic overdose. BIS AND POOR OUTCOMES All of the studies published to date have examined the association between low Bispectral Index (BIS) values and poor outcomes.4–8 In the first report on this subject, Monk et al.4 collected BIS data of 1046 noncardiac surgery patients participating in a study on cognitive dysfunction, with the attending anesthesiologists blinded to the BIS values. They created a new variable called “cumulative deep hypnotic time,” defined as the time that BIS was <45 during the anesthetic and demonstrated that this variable was significantly associated with 1-year mortality; relative risk per hour = 1.24 (95% confidence interval, 1.06–1.44). The possibility that increased anesthetic depth was associated with increased mortality was greeted with skepticism and calls for replication of the finding.19,20 This has been achieved in 3 subsequent studies.5–7 Lindholm et al.5 (noncardiac surgery patients) and Kertai et al.7 (cardiac surgery patients) reported hazard ratios for mortality per hour of BIS <45 of 1.18 (1.08–1.29) and 1.29 (1.12–1.49), respectively, and Leslie et al.6 (cardiac and noncardiac surgery patients) reported a hazard ratio for mortality for patients who recorded a BIS value of <40 for >5 minutes of 1.41 (1.02–1.95). Leslie et al.6 also reported that the hazard ratios for myocardial infarction (1.94 [1.12–3.35]) and stroke (3.23 [1.29–8.07]) were greater in patients with low BIS values (BIS <40 for >5 minutes) compared with those with optimal BIS values. Kertai et al.7 reported that the average volatile anesthetic concentration during maintenance was slightly lower in patients with low BIS values, suggesting increased sensitivity to anesthesia. In contrast, in a cohort of noncardiac surgery patients, Kertai et al.8 recently reported no association between BIS <45 and poor outcomes; hazard ratio = 1.03 (95% confidence interval, 0.93–1.14). The significant limitations of these observational studies were highlighted by the authors themselves and also emphasized in subsequent editorials.19,21–24 Because the studies were not randomized with respect to depth of anesthesia, significant selection bias may have been introduced as a result of the anesthesiologists intentionally targeting low BIS values, ignoring or not noticing low BIS values, or being unable to change low BIS values. The studies were also limited by lack of data and/or control over other factors that may have resulted in a poor outcome. Nevertheless, the association between “low BIS” and “death” is supported by most of these studies and we have incorporated this association into Figure 1. There is no evidence that there is a directly causal relationship between these elements, but there is also only one negative study and no evidence of decreased effect with each subsequent positive study.25 The unproven implication for clinical practice is that actions taken to avoid low processed EEG index values may improve outcomes for patients. However, other more important influences may be at play. OTHER FACTORS ASSOCIATED WITH POOR OUTCOMES, RELATIVE ANESTHETIC OVERDOSE, AND LOW PROCESSED EEG INDEX VALUES The foregoing studies4–8 identified more important associations with poor outcome than cumulative deep hypnotic time, namely, advancing age, comorbidities including malignancy, and intraoperative hypotension. These factors have been identified before26–28 and are included in Figure 1 under the headings “organ dysfunction” and “organ hypoperfusion.” Anesthetic agents are potent causes of arterial hypotension and any significant cerebral hypoperfusion may mean that a dose or concentration that was appropriate from an anesthetic depth perspective when the patient was normotensive is now excessive (i.e., a relative anesthetic overdose).15,29 Hypotension may also develop for other reasons such as bleeding, sepsis, or heart failure and these may also contribute to relative anesthetic overdose.30 Inadequate cerebral perfusion may also lead to low processed EEG index values and, independently of that, to poor outcomes.31 Therefore, arterial hypotension is an important confounding factor in the relationship between low index values and poor outcomes.4,6,19,22,23 In addition, the conditions leading to hypotension may be a valid alternative explanation for low BIS values and poor outcomes in these studies that do not invoke an independent effect of anesthetic depth or anesthetic agent toxicity. The unproven implication for clinical practice is that actions taken to correct hypotension or the causes of hypotension may improve outcomes for patients.32 The hypothesis that relative anesthetic overdose, as evidenced by low processed EEG index values, is detrimental in and of itself implies that anesthetic agents have a toxic and dose-dependent effect on the brain and other vital organs. This hypothesis is supported by work in young and elderly animals and humans.33 However, brain dysfunction may also result in relative anesthetic overdose, because concentrations delivered on the basis of normal brain function may be in excess of the requirements of someone with abnormal function.34,35 Patients with significant cerebral injury and disease also exhibit low index values and, independently of that, have poor outcomes.36 There is an obvious question if the hypothesis is true: What is the mechanism? At a molecular level, the deleterious effects of anesthesia are protean.37–42 Animal and some human studies have shown that anesthetics can provoke the inflammatory response, increase deposition of Alzheimer proteins, induce neuronal apoptosis, and cause prolonged postoperative cognitive dysfunction. Opioids can also induce angiogenesis and may decrease cancer survival times. The unproven clinical implication of inherent dose-dependent anesthetic toxicity is that a lower dose of anesthetic is better. In any case, one could make the case that unnecessarily deep anesthesia is wasteful. THE LOW MAC PARADOX Several studies have associated low inspired concentration of inhaled anesthetic with poor outcome.43,44 It is highly unlikely that low anesthetic concentrations are inherently harmful, except when awareness ensues, a condition that, although distressing, is not associated with increased postoperative mortality.45 Rather, low delivered concentrations represent the anesthesiologist's response to low processed EEG index values, hypotension, or other evidence of relative anesthetic overdose. Low minimal alveolar concentration is a marker of anesthetic intolerance and the fact that it remains associated with mortality in some multivariable models43,44 suggests that the problem may, unfortunately, not be controllable by the anesthesiologist. CONCLUSIONS Current evidence suggests that low processed EEG index values are associated with poor long-term outcomes.4–8 We have described the complex interplay among organ dysfunction, organ hypoperfusion, and our concept of “relative anesthetic overdose” in the production of low processed EEG index values and poor outcomes. We have also observed that low index values may be a marker of anesthetic intolerance rather than evidence of inherent anesthetic toxicity. The tantalizing question is whether preventing relative anesthetic overdose and/or low index values, while maintaining adequate organ perfusion, will make any difference. The simple answer to this complex question is a well-controlled randomized trial of a large group of patients at high risk of postoperative morbidity and mortality.19,22–24 Meanwhile, our advice is to “avoid hypoxia and hypotension … and unnecessarily deep anesthesia.” DISCLOSURES Name: Kate Leslie, MBBS, MD, M Epi, FANZCA. Contribution: This author helped design the study and write the manuscript. Attestation: Kate Leslie approved the final manuscript. Name: Timothy G. Short, MBChB, MD, FANZCA. Contribution: This author helped design the study and write the manuscript. Attestation: Timothy G. Short approved the final manuscript.