BIS and Entropy Monitoring: Detailed Explanation
Both Bispectral Index (BIS) and entropy are processed EEG (pEEG) monitors that quantify the hypnotic effect of anesthetic drugs on the frontal cerebral cortex, providing numerical indices to guide anesthetic dosing and reduce risks of awareness and excessive anesthesia 1.
What These Monitors Measure
Bispectral Index (BIS)
BIS is a dimensionless number (0-100) derived from multiple EEG parameters using proprietary multivariate statistical algorithms 2. The technology processes raw EEG signals through:
- Power spectral analysis: Evaluates frequency distribution and power in the EEG signal
- Bispectral analysis: Assesses phase coupling and synchronization between different frequency components (the "SynchFastSlow" parameter)
- Time domain analysis: Includes burst suppression detection
- Beta ratio (BetaRatio): Relative power in beta frequency bands
- Spectral Edge Frequency 95% (SEF95): The frequency below which 95% of EEG power resides
The algorithm was empirically derived from EEG data of over 5,000 anesthetized patients 2. BIS values of 90-100 indicate an awake state, 60-80 represents light-to-moderate sedation, 40-60 indicates general anesthesia, and values below 40 suggest deep hypnotic states 3. A BIS of 0 represents complete cortical suppression (flat line).
Entropy Monitoring
Entropy uses Shannon's entropy calculations to assess EEG signal irregularity and complexity. It provides two values:
- State Entropy (SE): Derived from 0.8-32 Hz frequency range, primarily reflecting cortical EEG activity
- Response Entropy (RE): Derived from 0.8-47 Hz, includes both EEG and frontalis muscle EMG activity
The difference between RE and SE theoretically indicates facial muscle tone, which may signal inadequate analgesia or light anesthesia.
How the Calculations Differ
At surgical anesthesia levels (BIS 30-80), BIS correlates well with both SynchFastSlow (r=0.60) and burst-compensated SEF95 (r=0.78), but SEF95 shows significantly better correlation than the bispectral-derived parameters 4. This suggests that despite its name, BIS relies heavily on power spectral analysis rather than true bispectral analysis.
In the BIS range of 60-100, BetaRatio alone shows strong linear correlation with BIS (r=0.90) 4. At BIS values below 30, the burst suppression ratio becomes the dominant determinant of the index value.
Clinical Performance and Discrepancies
Accuracy for Consciousness Detection
BIS demonstrates 95.4% accuracy with only 0.5% false positive rate when compared to clinical judgment, while entropy shows 90.87% accuracy with 4.6% false positive rate 5. However, BIS exhibits at least a one-minute delay in displaying state changes in 90% of patients (18 of 20) 5.
Common Causes of Index Disagreement
BIS and entropy values disagree by ≥10 units in approximately 11% of recorded time during sevoflurane anesthesia 6. Specific EEG patterns causing discrepancies include:
- Burst suppression: SE reads higher than BIS (35 of 49 episodes) 6
- Delta-theta dominance: BIS reads higher than SE (141 of 157 episodes) 6
- Alpha or beta activity: SE consistently reads higher than BIS (all 49 episodes) 6
- High-frequency signals: Eye blinks, saccades, and EMG contamination cause falsely elevated indices suggesting consciousness when patients are actually unconscious 7
State entropy shows more false classifications at transitions between consciousness and unconsciousness, particularly after loss of consciousness during sevoflurane anesthesia (14% disagreement rate vs. 9% for BIS) 7.
Critical Limitation
Both monitors have good sensitivity for detecting the "alert" state but poor sensitivity for confirming the "asleep" state 3. This means falsely elevated values will cause clinicians to administer unnecessary anesthetic doses, potentially causing excessively deep anesthesia with associated hemodynamic complications.
Clinical Applications and Guidelines
Mandatory Use Scenarios
Processed EEG monitoring should be used when total intravenous anesthesia (TIVA) is administered together with neuromuscular blockade 1. This recommendation stems from NAP5 findings that the vast majority of definite/probable awareness cases occurred during anesthetics involving neuromuscular blocking drugs.
Recommended Use Scenarios
- TIVA without neuromuscular blockade (consider use)
- Inhalational anesthesia with neuromuscular blockade (may reduce awareness risk)
- Any case where reducing postoperative delirium or cognitive dysfunction is prioritized
Monitoring Protocol
Start monitoring before induction and continue at least until full recovery from neuromuscular blockade is confirmed 1. This is critical because nearly two-thirds of awareness cases in NAP5 occurred either before surgery start or after surgery end—periods when end-tidal anesthetic concentration monitoring provides no protection.
Essential Clinical Caveat
Clinicians must not rely solely on the numerical index values displayed by these monitors 1. Instead, develop competency in:
- Interpreting raw EEG waveforms
- Understanding power spectral displays
- Recognizing artifacts (electrocautery, eye movements, EMG contamination)
- Identifying pathological patterns (epileptiform activity, burst suppression)
Electrocautery causes both BIS and entropy to change, sometimes in opposite directions, though values typically return to baseline after cautery stops 6. This represents a technical artifact rather than a change in anesthetic depth.
Evidence for Clinical Outcomes
One adequately powered randomized controlled trial demonstrated BIS-guided anesthesia reduced confirmed intraoperative awareness from 11 cases to 2 cases (P=0.022) 3. Additional benefits include:
- Reduced risk of excessive anesthesia
- Improved early recovery times
- Decreased incidence of postoperative delirium
- Reduced postoperative cognitive dysfunction 1
The technology is drug-independent and responds to propofol, volatile anesthetics, and their combinations in an approximately linear fashion 2.