Low mean arterial pressure (MAP) and deep hypnosis have been associated with complications and mortality. The normal response to high minimum alveolar concentration (MAC) fraction of anesthetics is hypotension and low Bispectral Index (BIS) scores. Low MAP and/or BIS at lower MAC fractions may represent anesthetic sensitivity. The authors sought to characterize the effect of the triple low state (low MAP and low BIS during a low MAC fraction) on duration of hospitalization and 30-day all-cause mortality.
Mean intraoperative MAP, BIS, and MAC were determined for 24,120 noncardiac surgery patients at the Cleveland Clinic, Cleveland, Ohio. The hazard ratios associated with combinations of MAP, BIS, and MAC values greater or less than a reference value were determined. The authors also evaluated the association between cumulative triple low minutes, and excess length-of-stay and 30-day mortality.
Means (±SD) defining the reference, low, and high states were 87 ± 5 mmHg (MAP), 46 ± 4 (BIS), and 0.56 ± 0.11 (MAC). Triple lows were associated with prolonged length of stay (hazard ratio 1.5, 95% CI 1.3-1.7). Thirty-day mortality was doubled in double low combinations and quadrupled in the triple low group. Triple low duration ≥60 min quadrupled 30-day mortality compared with ≤15 min. Excess length of stay increased progressively from ≤15 min to ≥60 min of triple low.
The occurrence of low MAP during low MAC fraction was a strong and highly significant predictor for mortality. When these occurrences were combined with low BIS, mortality risk was even greater. The values defining the triple low state were well within the range that many anesthesiologists tolerate routinely.
What We Already Know about This Topic
Anesthesiologists continue to refine factors associated with morbidity and mortality after surgery.
It is hoped identification of such factors will lead to treatments that may greatly reduce adverse outcomes during the perioperative period.
What This Article Tells Us That Is New
In this retrospective review of a large database from a single institution, the occurrence of low mean arterial pressure during low minimum alveolar concentration fraction was a strong and highly significant predictor for mortality, and when combined with low bispectral index, the mortality risk was even greater. Additional studies are needed to validate the triple low as an indicator of perioperative mortality.
THERE is increasing evidence that intraoperative anesthetic management influences long-term outcomes. For example, perioperative outcomes are improved by maintaining intraoperative normothermia,1guided fluid management,2,–,4minimizing blood transfusion,5,6and possibly restricting the storage time of transfused blood.7Two additional factors have been independently associated with postoperative mortality: low mean arterial pressure (MAP) and deep hypnotic level.8
There are various ways to characterize hypnotic level during general anesthesia, with electroencephalographic analysis being the most common. The best validated of these approaches is the Bispectral Index (BIS). BIS values range from 0 to 100, with 100 indicating full alertness and values less than 45 indicating deep anesthesia; optimal intraoperative BIS values are thought to range from 45 to 60.9Deep hypnosis, characterized by cumulative time with BIS less than 45, has been independently associated with poor postoperative outcomes in a number of higher risk populations, including the elderly,10patients with cancer,11those undergoing cardiac procedures,12and those at risk for intraoperative awareness.10
Volatile anesthetics reduce myocardial contractility and are vasodilators; thus, they provoke dose-dependent hypotension. Intraoperative MAP typically is maintained at approximately 85 mmHg, but values range widely among patients and procedures. Low intraoperative MAP is associated with increased risk of stroke,13,14myocardial infarction,15and 1-year mortality.8,10
The potency of a volatile anesthetic is characterized by its minimum alveolar concentration (MAC), which is the alveolar partial pressure at which 50% of patients move in response to skin incision. MAC varies among anesthetics, but the MAC fraction accurately characterizes relative dose for any volatile anesthetic. Anesthesia usually is administered to achieve an initial target expired MAC, and is then adjusted based on patient hemodynamic responses.
The expected response to high MAC fractions of anesthetics is hypotension and lower BIS values (indicating deeper hypnosis and suppression of brain electrical activity). In contrast, low MAP and/or BIS in patients receiving low anesthetic MAC fractions is atypical and may help identify patients who are unusually sensitive to anesthesia and at risk for complications. Thus, the combination of low MAP, BIS, and MAC (a “triple low”) may be associated with especially poor outcomes. We tested the hypothesis that a triple low of MAP, BIS, and MAC is associated with prolonged duration of hospitalization and increased 30-day all-cause mortality.
In our initial analysis, we defined a reference state consisting of patients whose average MAP, BIS, and MAC values were each within one SD of the population means. The remaining patients were classified into nonoverlapping groups characterized by whether the case average MAP, BIS, and MAC values were greater or less than the population average for each variable. Subsequently, we evaluated the association between cumulative minutes in the triple low condition, defined by MAP less than 75 mmHg, BIS less than 45, and MAC less than 0.80, and excess length of stay and relative risk of 30-day mortality.
Materials and Methods
The Cleveland Clinic Perioperative Health Documentation System is a clinical registry that includes the entire electronic anesthesia record, data from various administrative databases, and portions of the electronic medical record. Perioperative variables were collected prospectively concurrently with patient care from our electronic anesthesia record and other electronic systems. Mortality status was obtained from the United States Social Security Death Index. Use of the perioperative registry for this retrospective cohort analysis was approved by the Institutional Review Board, Cleveland Clinic, Cleveland, Ohio.
We included patients (≥16 yr old) who had noncardiac surgery at the Cleveland Clinic Main Campus between January 6, 2005, and December 31, 2009. Patients were included in our analysis when they had BIS monitoring and a single volatile anesthetic identified by nonzero concentrations of only one agent from incision to end of case. Total intravenous anesthesia cases were identified when all three volatile agents had zero concentration from incision to end of case and nonzero propofol recorded on the electronic record during the same period. When a given patient had more than one operation on different days, only data from the most recent surgical date was included. We also excluded emergency surgery and cases lacking essential clinical or endpoint information.
General anesthesia for adult noncardiac surgery at the Cleveland Clinic usually is induced with a small amount of fentanyl (typically 100–150 μg) and propofol (1–3 mg/kg); anesthesia usually is then maintained with a volatile anesthetic in a mixture of air and oxygen with only a small amount of additional opioid if needed. However, the Cleveland Clinic is a large teaching institution, so of course there is considerable patient-to-patient variability in anesthetic management based on provider preference.
Data Extraction and Analysis
Mean arterial pressure, BIS, and end-tidal volatile anesthetic concentration, propofol use, duration of hospitalization, and 30-day all-cause mortality were extracted from the registry. We also extracted age, sex, body mass index, American Society of Anesthesiologists Physical Status scores, and International Classification of Diseases, version 9 billing codes.
Mean arterial pressure values were recorded at 1-min intervals when an arterial catheter was used, as it was in approximately one half the cases; blood pressures were otherwise recorded oscillometrically at intervals of 2–5 min. MAP values were assumed to be artifactual and were excluded when the recorded value was less than 30 mmHg or more than 250 mmHg. BIS values were recorded at 1-minute intervals in more recent cases (41% of the total) or at 15-min intervals in older cases. The BIS and MAP values assigned to a given minute were the most recent values within the past 20 min and were otherwise considered to be missing.
Minimum alveolar concentration equivalents were calculated from end-tidal volatile anesthetic partial pressures using a 1 MAC-equivalent concentration of desflurane (6.6%), sevoflurane (1.8%), and isoflurane (1.17%).16Nitrous oxide use is recorded in our registry, but for technical reasons the concentration is not; thus, nitrous oxide was not included in our calculation of MAC fraction. However, in our case-based modeling, we included a binary variable indicating whether nitrous oxide was used in the procedure, and we attempted to account for the MAC-sparing effect of opioids and residual propofol by including case-average estimates of the effect-site concentration‡‡of fentanyl equivalents and propofol in our models.
The principal procedure was identified from International Classification of Diseases, version 9 billing codes and classified into the following surgical groups: general, gynecology, urology, neurology, orthopedic, abdominal, head and neck, vascular, thoracic, and other. Body mass index was divided into quintiles. Race was classified as into three tiers: Caucasian, African American, or Other. We also classified age into 6 decade levels: ⩽40, 41–50, 51–60, 61–70, 71–80, >80 yr. The American Society of Anesthesiologists Physical Status was grouped into scores of 1 and 2 versus ≥3.
For each included patient, we calculated average MAP, BIS, and MAC from the beginning to end of anesthesia. We defined a reference state consisting of patients whose average MAP, BIS, and MAC values were each within one SD of the population means (the data were nearly normally distributed). The remaining patients were classified into nonoverlapping groups characterized by whether the case average MAP, BIS, and MAC values were greater or less than the population average for each variable.
State categories were defined relative to the average reference threshold for each variable. Single lows were when patients exhibited any one of the three single low case-based averages of MAP, BIS, or MAC. Patients were assigned to one of three double low categories when two of the three MAP, BIS, and MAC values were less than their respective reference thresholds. Similarly, patients were assigned to the single triple low category when each value was less than the reference threshold.
Cox proportional hazards regression was used to create a model to predict 30-day postoperative mortality. Similarly, logistic regression was used to create a model to predict whether a patient's postoperative hospital length of stay would be longer (or not) than expected compared with the diagnostic related group-adjusted national average length of stay for the primary surgery as identified from the stay-based administrative record (ClinTrac, 3M, Minneapolis, MN). Each model included a single state variable based on case-average MAP, BIS, and MAC and also identified significant predictors (using forward conditional selection) from among demographic predictors (age, gender, race, body mass index, American Society of Anesthesiologists Physical Status); intraoperative factors (case-average estimates of blood concentration of propofol and fentanyl equivalents, estimated blood loss and administered erythrocyte volume, a variable indicating maintenance agent type (isoflurane, sevoflurane, desflurane), a binary variable indicating whether nitrous oxide was used or not in the procedure, case duration); and components of the Risk Stratification Indices for 30-day mortality and LOS (composite risk stratifications from International Classification of Diseases, version 9 diagnosis and procedure codes17) ranked in quintiles. P < 0.05 was considered statistically significant.
Averaging values over an entire case can conceal potentially important short periods of concomitant low MAP, BIS, and MAC. Thus, we conducted a second analysis based on cumulative minutes in the triple low condition for each patient, with no requirement that the minutes be contiguous.
In the case-based analysis, thresholds were defined by the population means. For the cumulative duration analysis, we constructed three-dimensional plots of mortality as a function of cumulative minutes at different MAP and BIS thresholds at various MAC fractions. This analysis suggested that triple low thresholds of MAP less than 75 mmHg, BIS less than 45, and MAC less than 0.8 discriminated well between patients who survived 30 postoperative days and those who did not. In choosing these thresholds, we considered values that might provide a high potential for reduced mortality without an excessive number of “false alarms.” Thus, we used these values for our formal analysis of cumulative minutes under triple low conditions.
Cumulative (not necessarily contiguous) minutes in a triple low state (MAP less than 75 mmHg, BIS less than 45, and MAC fraction less than 0.8) were calculated for each patient. Patients were partitioned into groups based on their cumulative triple low state duration: 0, 1–15, 16–30, 31–45, 46–60, and more than 60 min. Analysis of variance was used to test whether the incidence of mortality and excess length of stay were statistically significantly different between duration groups. P < 0.05 (after Bonferroni correction for multiple comparisons) was considered statistically significant.
Among the 103,324 surgical procedures in our registry at the time of analysis, we excluded 28,231 because only the most recent surgery was considered for each patient; 35,686 because BIS monitoring was not used; 6,810 because the primary anesthetic was not a single volatile agent; 123 because patients were younger than 16 yr old; and 8,354 because of critical missing data. Consequently, 24,120 patients were available for the case-based analysis.
Among included patients, isoflurane was the volatile anesthetic in 27% of the cases, sevoflurane in 45%, and desflurane in 28%. Nitrous oxide was used in 38% of patients, but in many or most cases, only briefly during emergence. Overall 30-day postoperative mortality was 0.8%; most deaths (0.5%) occurred in the hospital. An additional 5,188 patients were omitted from the length-of-stay analyses because they were outpatients.
The averages that define the reference state and low and high values were 87 ± 5 mmHg for MAP, 46 ± 4 for BIS, and a MAC fraction of 0.56 ± 0.11 (table 1). Approximately 6% of the patients were categorized as exhibiting a case average triple low condition.
The triple low combination was associated with the largest risk of a significantly prolonged length of stay (relative risk [hazard ratio] 1.5, 95% CI 1.3–1.7; table 2). Triple high values were not associated with a significant increase in 30-day mortality (table 3). The only single low value that was associated with increased mortality was low MAC. In contrast, all three double low combinations were associated with a roughly 2-fold mortality increase, with mortality being significantly increased for two of the three combinations. Mortality in the triple low group was quadrupled (table 1). The relative risks for mortality in each group are shown in figure 1. Age-adjusting MAC fractions did not perceptibly alter relative risks; inclusion of nitrous oxide use in the statistical model also did not substantively alter the results; and finally, use of a regional block did not have any important effect of relative risks (data not shown). Tables 2and 3do not consistently exhibit increased hazard ratios at higher levels of RSI; this may be due to risk-transference among the variables in the models.
We first plotted 30-day all-cause mortality as a function of cumulative (not necessarily contiguous) minutes at various thresholds for MAP and BIS at MAC fraction thresholds of 0.6, 0.7, and 0.8. At each MAC fraction, mortality increased as a function of cumulative duration at lower MAP and BIS thresholds. At cumulative durations exceeding 15 min, mortality increased substantially when the MAP threshold was less than 70 mmHg and the BIS threshold was less than 45; the combination of the two was especially associated with increased mortality (fig. 2).
The numbers of patients spending 0, 1–15, 16–30, 31–45, 46–60, and more than 60 min in the triple low state (MAP less than 75 mmHg, BIS less than 45, and MAC fraction less than 0.80) were 8,691, 7,858, 3,536, 1,573, 907, and 1,555, respectively. Thirty-day all-cause mortality was significantly increased from baseline (no triple low minutes) when cumulative triple low duration was 31–45 min and when it exceeded 60 min (fig. 3).
The fraction of patients requiring hospitalization for longer than the national average for a given procedure (i.e. , excess length of stay) increased significantly as the duration of triple low minutes increased and was significantly greater than baseline (no triple low minutes) at all times exceeding 30 min (fig. 4). Inclusion of nitrous oxide use in the statistical model did not substantively alter the results.
Although previous work in certain at-risk populations suggests that low MAP8,10,13,–,15and BIS8,10,–,12are associated independently with various poor postoperative outcomes, including mortality, we did not find that either alone was associated with 30-day mortality in our general population. In fact, isolated low BIS was associated with a (nonsignificant) reduction in mortality. Isolated low MAC fraction, which is much less studied, was associated with increased mortality. Considering the three single low groups together, there was no overall association with either mortality or duration of hospitalization.
In contrast to the minimal relationship between mortality and isolated low values of MAP, BIS, and MAC fraction, case-based double lows taken as a group were associated with a roughly 2-fold increase in 30-day postoperative mortality. Mortality was more than quadrupled in patients who demonstrated case-based triple lows. Thus, the combination of low MAP, BIS, and MAC fraction, which occurred in approximately 6% of the study population, was an ominous predictor of postoperative mortality.
The overall limited association of isolated case average low MAP, BIS, and MAC fraction suggests that no single measure, with the possible exception of low MAC, is sufficiently robust to adequately account for patient variability and clinical complexity. For example, consider three potential causes of low BIS: (1) Low BIS is the normal response to generous doses of volatile anesthetics. When high concentrations of volatile anesthetics are given to healthy patients who hemodynamically tolerate large doses, BIS should be low and would not be expected to be associated with poor prognosis. Our results are consistent with this theory in that isolated low BIS was associated with a (nonsignificant) reduction in mortality. (2) An alternative cause of low BIS is anesthetic sensitivity. This group is identified by the combination of low BIS and low MAC fraction. This is an atypical response because low MAC fraction should be associated with high BIS. That BIS was in fact low in some patients with low MAC fractions suggests an abnormal sensitivity to volatile anesthesia, potentially because of underlying illness. As might be expected, this double low combination was associated with twice the mortality of the reference group. (3) A third potential cause of low BIS is inadequate brain perfusion, resulting in ischemic suppression of brain metabolism. Brain hypoperfusion may especially occur in a fraction of patients who demonstrate low BIS combined with low MAP. The individual autoregulatory MAP threshold for critically reduced brain blood flow remains unknown but surely varies widely among patients, and it is likely that values that generally are well tolerated in healthy individuals are completely inadequate in some patients. Inadequate cerebral perfusion is perhaps the most interesting putative cause of low BIS because it is potentially amenable to hemodynamic intervention, such as giving vasopressors or fluids to improve MAP and brain perfusion.
The thresholds defining “low” and “high” values in our case-based analysis were objective and based on mean values for each measure (excepting the reference patients, who were excluded from each group). As shown in figure 2, 30-day mortality progressively increases as the thresholds decrease, especially for MAP and BIS. At MAC fractions between 0.6 and 0.8, mortality becomes extreme when the MAP threshold was set to less than 70 mmHg or when the BIS threshold was set to less than 45. Note the “wall of death” rising at the right and left rear portions of the lower images in figure 2.
For our cumulative minute analysis, we used thresholds of MAP less than 75 mmHg, BIS less than 45, and MAC less than 0.8 because these values discriminated well between patients who survived 30 postoperative days and those who did not. (These values were also chosen because in a prospective study they would generate a high potential benefit of intervention and a modest number of triple low events that would not overwhelm clinicians.) Mortality generally increased as cumulative minutes of triple low increased beyond 15 min and was substantially (roughly 4-fold) greater at cumulative durations exceeding 60 min. Duration of hospitalization also increased progressively as a function of cumulative triple low duration beyond 15 min. A fascinating aspect of our findings is that the thresholds identified in this analysis were otherwise unremarkable and, at least individually, would not concern most anesthesiologists. However, combined they were strong predictors of prolonged hospitalization and mortality.
Excess duration of hospitalization normally would be considered an “intermediate outcome” compared with mortality. However, duration of hospitalization was prolonged only in the triple low patients and even then not by much. In contrast, there were substantial and highly clinically important mortality differences in patients demonstrating case-based double and triple lows. However, we note that discordance between these two outcomes is not necessarily inconsistent because patients who die early may have shorter hospitalizations than comparable patients who recover normally.
One strength of our analysis is that we included more than 24,000 patients. Thus, we had more than 1,500 patients in seven of our eight categorical groups of “low” and “high” combinations and more than 8,000 in the reference group. It is apparent that starting with a much smaller number of patients, say 1,000–2,000, would be inadequate and result in type 2 statistical errors.
Our approach differs somewhat from previous analyses in that we extracted a reference group that was within one SD of the mean for MAP, BIS, and MAC fraction. Our “low” and “high” groups thus exclude the most typical patients. The extent to which our various low and high groups differ from a simple split at the mean or an arbitrary value depends critically on the size of the reference group. A larger window, say 1.5 SD, would result in smaller low and high groups, but augment differences between low and high pairs for each measure; conversely, a smaller window would reduce differences. Using our definition, we identified a high-risk triple low population, based solely on patient response to anesthesia, which represents 6% of the patients undergoing surgery at our institution.
Our results indicate that two double low combinations and a triple low of MAP, BIS, and MAC strongly predict postoperative mortality. However, as in all registry analyses, it is impossible to make causal conclusions from these observations. Our statistical models were adjusted for baseline comorbidity and procedural intensity using the Risk Stratification Index.17Nonetheless, prolonged hospitalization and increased mortality with double and triple lows to a large extent surely reflects selection of patients whose underlying illness makes them susceptible to anesthesia. If comorbidity is the full explanation, intervention is unlikely to improve outcome.
However, it is worth considering that components of the triple low state usually can be controlled with common anesthetic interventions. For example, BIS can be increased by reducing volatile anesthetic administration, and MAP can be increased by giving vasopressors or fluids. Our time-based analysis demonstrating a significant association between cumulative duration in the triple low state and increased mortality suggests a target for therapeutic intervention. To the extent that remaining in a triple low state worsens outcomes, rather than just predicts bad outcomes, clinician intervention in response to triple low events might reduce mortality. This theory is being tested in a randomized trial in which clinicians are alerted (or not) to triple low events (clinical trial NCT00998894). The thresholds for this study are identical to those in the second, time-based, analysis.
Limitations of our study include that our findings apply only to patients who were given volatile anesthesia. Volatile anesthesia is by far the most common approach at the Cleveland Clinic, and few patients managed this way are given propofol after induction. In addition, most are given only small amounts of opioid analgesia. Nonetheless, it remains possible that some of the patients with low MAC fractions of volatile anesthetic may have been given substantial amounts of propofol or opioids, rather than being atypically sensitive to anesthesia. Of course, MAC fraction is but one component of the triple low state, and patients in this study given mostly intravenous drugs tend to have high BIS and well-sustained MAP (data not shown). Thus, substituting propofol or opioids for volatile anesthetic will not, per se , generate a triple low state.
We know which patients were given nitrous oxide, but a limitation of our registry is that nitrous oxide concentration is not recorded; in addition, a given MAC fraction of nitrous oxide has less effect on BIS than do volatile anesthetics. Thus, we made no attempt to include nitrous oxide in our MAC fraction estimates. However, nitrous oxide was not used in most cases, and inclusion of nitrous oxide in our statistical models had only minimal effect on the results.
In summary, the combination of low MAC and low MAP was a strong and highly statistically significant predictor for mortality. When combined with low BIS, relative risk adjusted mortality was even greater. Thus, the combination of low MAC, low MAP, and low BIS, a triple low, is an ominous predictor of excessive hospital length of stay and postoperative mortality. This association is especially concerning because the threshold and average low values for each state were well within the range that many anesthesiologists tolerate routinely.
The authors gratefully acknowledge the contributions of Armin Schubert, M.D. (Chair, Department of Anesthesiology, Ochsner Health System, New Orleans, Louisiana), and Maged Argalious, M.D. (Professional Staff, Department of General Anesthesia, Cleveland Clinic, Cleveland, Ohio), who conceived and developed the Cleveland Clinic's Perioperative Health registry. The authors also thank Eric K. Christiansen, M.B.A. (Anesthesiology Institute, Cleveland Clinic), who led extraction of data from the registry.