Nassib G. Chamoun

An introduction to bispectral analysis for the electroencephalogram

The goal of much effort in recent years has been to provide a simplified interpretation of the electroencephalogram (EEG) for a variety of applications, including the diagnosis of neurological disorders and the intraoperative monitoring of anesthetic efficacy and cerebral ischemia. Although processed EEG variables have enjoyed limited success for specific applications, few acceptable standards have emerged. In part, this may be attributed to the fact that commonly usedsignal processing tools do not quantify all of the information available in the EEG. Power spectral analysis, for example, quantifies only power distribution as a function offrequency, ignoring phase information. It also makes the assumption that thesignal arises from alinear process, thereby ignoring potential interaction betweencomponents of the signal that are manifested asphase coupling, a common phenomenon in signals generated fromnonlinear sources such as the central nervous system (CNS). This tutorial describes bispectral analysis, a method of signal processing that quantifies the degree of phase coupling between the components of a signal such as the EEG. The basic theory underlying bispectral analysis is explained in detail, and information obtained from bispectral analysis is compared with that available from thepower spectrum. The concept of abispectral index is introduced. Finally, several model signals, as well as a representative clinical case, are analyzed using bispectral analysis, and the results are interpreted.

Hospital Stay and Mortality Are Increased in Patients Having a "Triple Low" of Low Blood Pressure, Low Bispectral Index, and Low Minimum Alveolar Concentration of Volatile Anesthesia

Background: 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. Methods: 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. Results: 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. Conclusions: 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.

Bispectral Analysis of the Electroencephalogram Correlates with Patient Movement to Skin Incision during Propofol/Nitrous Oxide Anesthesia

Background:Bispectral analysis Is a signal-processing technique that determines the harmonic and phase relations among the various frequencies in the electroencephalogram. Our purpose was to compare the accuracy of a bispectral descriptor, the bispectral index, with that of three power spectral variables (95% spectral edge, median frequency, and relative δ power) in predicting patient movement in response to skin incision during propofol-nitrous oxide anesthesia. Methods:Forty-four adult patients scheduled for elective noncranial surgery were studied. Gold cup electroencephalographlc electrodes were placed on each patient in a frontoparietal montage (Fpl, Fp2, P3, and P4) referred to C2, and the electroencephalogram was recorded continuously and processed off-line. Conventional frequency bands were used to describe power spectrum variables. Anesthesia was induced with propofol (1.5-3.0 mg-1 · kg-1) and maintained with 60% nitrous oxide in oxygen and with propofol at one of three randomized infusion rates (100, 200, or 300 µg · kg-1 · min-1). Inadequate anesthetic depth was denned as patient movement in response to a 2-cm skin incision at the planned site of surgery. Plasma propofol concentrations were measured within 2 min after skin incision. Results:Complete data were available for 38 patients, of whom 17 moved in response to skin incision. Analysis of the area under the receiver operating characteristic curves showed that only for bispectral index and drug dose group was there a significant predictive relation (area > 0.5). Furthermore, the bispectrum was significantly predictive even after stratification by dose group. Conclusions:The bispectral index of the electroencephalogram is a more accurate predictor of patient movement in response to skin incision during propofol-nitrous oxide anesthesia than are standard power spectrum parameters or plasma propofol concentrations.

Eeg bispectrum predicts movement during thiopental/isoflurane anesthesia

Objective. The objective of our study was to test the efficacy of the bispectral index (BIS) compared with spectral edge frequency (SEF), relative delta power, median frequency, and a combined univariate power spectral derivative in predicting movement to incision during isoflurane/oxygen anesthesia.Methods. A total of 42 consenting patients were assigned to 3 groups, isoflurane 0.75, 1.0, and 1.25 minimal alveolar concentration (MAC). Anesthesia was induced with thiopental and maintained with the appropriate end-tidal concentration of isoflurane. The electroencephalogram (EEG) was recorded using a microcomputer system, and data were analyzed off-line. The EEG during the 2 min before incision was analyzed. Following skin incision, each patient was carefully observed for 60 sec to detect occurrence of purposeful movement.Results. For all groups combined, there was a statistically significant difference for BIS (p<0.0001) and also for relative delta power (p<0.016) between movers and non-movers. There was a statistically significant difference between movers and nonmovers at 1.25 MAC isoflurane for BIS (p<0.01). There were no other significant differences for any other EEG variable at any concentration of isoflurane. No EEG variable showed a relationship to isoflurane concentration.Conclusions. When bispectral analysis of the EEG was used to develop a retrospectively determined index, there was an association of the index with movement. Thus, it may be a useful predictor of whether patients will move in response to skin incision during anesthesia with isoflurane/oxygen.

Broadly applicable risk stratification system for predicting duration of hospitalization and mortality

Background:Hospitals are increasingly required to publicly report outcomes, yet performance is best interpreted in the context of population and procedural risk. We sought to develop a risk-adjustment method using administrative claims data to assess both national-level and hospital-specific performance. Methods:A total of 35,179,507 patient stay records from 2001–2006 Medicare Provider Analysis and Review (MEDPAR) files were randomly divided into development and validation sets. Risk stratification indices (RSIs) for length of stay and mortality endpoints were derived from aggregate risk associated with individual diagnostic and procedure codes. Performance of RSIs were tested prospectively on the validation database, as well as a single institution registry of 103,324 adult surgical patients, and compared with the Charlson comorbidity index, which was designed to predict 1-yr mortality. The primary outcome was the C statistic indicating the discriminatory power of alternative risk-adjustment methods for prediction of outcome measures. Results:A single risk-stratification model predicted 30-day and 1-yr postdischarge mortality; separate risk-stratification models predicted length of stay and in-hospital mortality. The RSIs performed well on the national dataset (C statistics for median length of stay and 30-day mortality were 0.86 and 0.84). They performed significantly better than the Charlson comorbidity index on the Cleveland Clinic registry for all outcomes. The C statistics for the RSIs and Charlson comorbidity index were 0.89 versus 0.60 for median length of stay, 0.98 versus 0.65 for in-hospital mortality, 0.85 versus 0.76 for 30-day mortality, and 0.83 versus 0.77 for 1-yr mortality. Addition of demographic information only slightly improved performance of the RSI. Conclusion:RSI is a broadly applicable and robust system for assessing hospital length of stay and mortality for groups of surgical patients based solely on administrative data.

Impact of Present-on-admission Indicators on Risk-adjusted Hospital Mortality Measurement

Background:Benchmarking performance across hospitals requires proper adjustment for differences in baseline patient and procedural risk. Recently, a Risk Stratification Index was developed from Medicare data, which used all diagnosis and procedure codes associated with each stay, but did not distinguish present-on-admission (POA) diagnoses from hospital-acquired diagnoses. We sought to (1) develop and validate a risk index for in-hospital mortality using only POA diagnoses, principal procedures, and secondary procedures occurring before the date of the principal procedure (POARisk) and (2) compare hospital performance metrics obtained using the POARisk model with those obtained using a similarly derived model which ignored the timing of diagnoses and procedures (AllCodeRisk). Methods:We used the 2004–2009 California State Inpatient Database to develop, calibrate, and prospectively test our models (n = 24 million). Elastic net logistic regression was used to estimate the two risk indices. Agreement in hospital performance under the two respective risk models was assessed by comparing observed-to-expected mortality ratios; acceptable agreement was predefined as the AllCodeRisk-based observed-to-expected ratio within ±20% of the POARisk-based observed-to-expected ratio for more than 95% of hospitals. Results:After recalibration, goodness of fit (i.e., model calibration) within the 2009 data was excellent for both models. C-statistics were 0.958 and 0.981, respectively, for the POARisk and AllCodeRisk models. The AllCodeRisk-based observed-to-expected ratio was within ±20% of the POARisk-based observed-to-expected ratio for 89% of hospitals, which was slightly lower than the predefined limit of agreement. Conclusion:Consideration of POA coding meaningfully improved hospital performance measurement. The POARisk model should be used for risk adjustment when POA data are available.

The Position of Aspect

To the Editor:—I read with great interest the exchange of opinions regarding “The Media and the BIS Monitor,” which appeared in the June issue of ANESTHESIOLOGY. The letters highlighted some very important issues facing anesthesiologists and Aspect Medical Systems, especially in today’s healthcare environment. As you indicated in your response to Dr. Katz, Aspect has been very careful about not making specific claims that use of the BIS will reduce the incidence of intraoperative recall. We state in our marketing materials that the BIS Monitor can be used as a tool to “monitor,” “assess,” or “track” the risk of awareness based on results from numerous peer-reviewed publications demonstrating the relationship between BIS and sedation scores, memory function tests, and loss and return of consciousness in volunteers and surgical patients. It is not our intention to suggest that these studies prove that BIS monitoring will reduce the incidence of awareness. In fact, it is very important for anesthesiologists to understand that awareness can still occur when BIS monitors are used, although we believe the data clearly support the observation that awareness is most likely to happen when BIS values are high. Widespread publicity of awareness cases on national television, news magazines, and newspapers with the resulting reactions of the anesthesia community have represented a significant public relations challenge to Aspect. Our efforts to introduce the BIS to the anesthesia community have relied extensively on sound clinical research published in high-quality journals such as ANESTHESIOLOGY. To date, more than 45 full manuscripts and more than 300 abstracts have been published on BIS. Our sales and marketing programs have always highlighted the demonstrated drug savings and recovery benefits of BIS rather than focusing on prevention of awareness. The perceived benefit to Aspect from awareness-related publicity is overshadowed by the negative impact it has had on our relationship with the anesthesia community, which we have worked so hard to build over the past 13 yr. The letter from Dr. Katz is a good case in point. I sincerely hope that this letter provides a better understanding of Aspect’s position regarding the issue of awareness and BIS. Perhaps some day it will be possible to conduct a definitive clinical study to evaluate the impact of BIS monitoring on the incidence of awareness (although such a study would require randomization of approximately 50,000 patients to have adequate statistical power*), but until then, the potential efficacy of BIS monitoring in preventing awareness can only be reasonably inferred from the previously cited references. The scientifically demonstrated recovery benefits of BIS monitoring provide the most important reasons for anesthesia providers to consider using this technology. Nassib G. Chamoun President and CEO Aspect Medical Systems Natick, Massachusetts

Validation and Calibration of the Risk Stratification Index

Background: The Risk Stratification Index was developed from 35 million Medicare hospitalizations from 2001 to 2006 but has yet to be externally validated on an independent large national data set, nor has it been calibrated. Finally, the Medicare Analysis and Provider Review file now allows 25 rather than 9 diagnostic codes and 25 rather than 6 procedure codes and includes present-on-admission flags. The authors sought to validate the index on new data, test the impact of present-on-admission codes, test the impact of the expansion to 25 diagnostic and procedure codes, and calibrate the model. Methods: The authors applied the original index coefficients to 39,753,036 records from the 2007–2012 Medicare Analysis data set and calibrated the model. The authors compared their results with 25 diagnostic and 25 procedure codes, with results after restricting the model to the first 9 diagnostic and 6 procedure codes and to codes present on admission. Results: The original coefficients applied to the 2007–2012 data set yielded C statistics of 0.83 for 1-yr mortality, 0.84 for 30-day mortality, 0.94 for in-hospital mortality, and 0.86 for median length of stay—values nearly identical to those originally reported. Calibration equations performed well against observed outcomes. The 2007–2012 model discriminated similarly when codes were restricted to nine diagnostic and six procedure codes. Present-on-admission models were about 10% less predictive for in-hospital mortality and hospital length of stay but were comparably predictive for 30-day and 1-yr mortality. Conclusions: Risk stratification performance was largely unchanged by additional diagnostic and procedure codes and only slightly worsened by restricting analysis to codes present on admission. The Risk Stratification Index, after calibration, thus provides excellent discrimination and calibration for important health services outcomes and thus appears to be a good basis for making hospital comparisons.

Comparison of an Updated Risk Stratification Index to Hierarchical Condition Categories

Background: The Risk Stratification Index and the Hierarchical Condition Categories model baseline risk using comorbidities and procedures. The Hierarchical Condition categories are rederived yearly, whereas the Risk Stratification Index has not been rederived since 2010. The two models have yet to be directly compared. The authors thus rederived the Risk Stratification Index using recent data and compared their results to contemporaneous Hierarchical Condition Categories. Methods: The authors reimplemented procedures used to derive the original Risk Stratification Index derivation using the 2007 to 2011 Medicare Analysis and Provider review file. The Hierarchical Condition Categories were constructed on the entire data set using software provided by the Center for Medicare and Medicaid Services. C-Statistics were used to compare discrimination between the models. After calibration, accuracy for each model was evaluated by plotting observed against predicted event rates. Results: Discrimination of the Risk Stratification Index improved after rederivation. The Risk Stratification Index discriminated considerably better than the Hierarchical Condition Categories for in-hospital, 30-day, and 1-yr mortality and for hospital length-of-stay. Calibration plots for both models demonstrated linear predictive accuracy, but the Risk Stratification Index predictions had less variance. Conclusions: Risk Stratification discrimination and minimum-variance predictions make it superior to Hierarchical Condition Categories. The Risk Stratification Index provides a solid basis for care-quality metrics and for provider comparisons.