Friday, 14 March 2008

The B-unAware Study: Anesthesia Awareness and the Bispectral Index

NEJM March 13. This study reports a comparison of BIS monitoring compared with end-tidal volatile agent monitoring. It found no difference in the incidence of awareness. There is an associated editorial. This finding is in contrast with the B-Aware study that found a 75% reduction in awareness when using BIS monitoring in high risk patients. The study included 1900 patients and was an RCT.

Comments on the methods:
The study's inclusion criteria to be 'at risk of awareness' are very liberal, especially the minor criteria. This was probably not a patient sample that was at significant risk of awareness. Specifically, table 1 shows the following inclusion characteristics: 44% of patients were on beta-blocker, 50% had moderate limited exercise tolerance, and 44% had obesity (each of these are a minor inclusion criteria). This suggests that the study is under-powered to detect a difference by using BIS monitoring as an intervention. The wide inclusion criteria seems to explain how a single hospital quickly completed a study of nearly 2000 patients (in only 14 months).


  1. I could not agree more. There are more than a few reasons to believe this study is of dubious value.

    1) The population group is not really "high risk" so to expect 1% awareness seems odd. It is no wonder they ended up with 0.2% awareness, the group was general population.

    2) They did not share the drug mix, so no way to know if the BIS group received similar mix drugs as the ETAG group. So comparing MAC is not as relevant if we don't know the mix of drugs. Also they did not indicate if they used versed.

    ALso, it seems like they overdosed their patients...which would explain why you had less awareness. It is no doubt that overdosing would lead to less awareness.

    3) Look at the cases of awareness cited. They have absolutely NO idea when the awareness happened. In a couple instances they said awareness was for the ENTIRE CASE! Also, for patient A-4, they don't even have consistent BIS DATA!! How could this be included??

    4) They rarely adhered to their stated protocols. Only 26% of the time for the ETAG. So is this really a legitimate alternative??

    I am frankly shocked that the NEJM would allow a study like this to enter their journal. Seems like there might have been some backroom dealings.

  2. Good points. Thanks for making these comments. I expect the NEJM will get many letters about the study.

  3. FYI. Comment #1 was posted verbatim on other blogs from an individual who has never participated before in any these discussion groups/blogs. This individual is widely thought to be a biased representative of Aspect Medical initiating a corporate public relations campaign rather than participating in a truly objective, scientific-based discussion.

  4. Dear 2nd Anonymous commentator,
    Please limit your comments to the data or science of this study.


  6. I agree with Richard Haillwell that the study was probably underpowered with respect to its generous inclusion criteria. Unfortunately, to study awareness stringently is difficult given the low incidence.

    To Anonymous

    1. I would agree that the population isn't completely high risk. However, a good chunk of the patients were still getting cardiac surgery, more so than in the general population as a whole. Maybe if the true incidence of awareness is 1:14500 and not 1:1000, then 0.2% truly is high risk?

    2. On what grounds can you say that the patients were overdosed? No clinician could draw those conclusions from the data. There's no evidence presented of excess hypotension, cardiac events, excessively slow emergence, etc. Unless you are drawing that conclusion from the BIS number alone, which no clinician would (or should) do.

    As to the mix of drugs, yes that would have been nice to see, but the study was looking at the use of ETAG and BIS, not trying to control for the ways clinicans might give midaz with these devices. If not having a bis monitor causes me to give more midaz to make me feel better, leading to lower awareness, then well not having a bis will decrease awareness. Or, if having a BIS monitor makes me feel safe and causes me to give less midaz, leading to more awareness, then having a BIS monitor will lead to more awareness. In any case, it doesn't change the fact that the use of the BIS monitor did not lower awareness.

    That being said, it would be interesting to see how use/nonuse of a bis monitor affects drug usage and outcomes.

    3. Of course they can't specify when awareness occurred in most of the patients. Do you think you could based on a tale of hearing a voice and feeling pain? Most awareness tales are vague enough that it is impossible to pinpoint the time it occurred.

    4. It would have been nice if they had a defined protocol. That being said, maybe the clinicians were trying to avoid "overdosing" their patients, thus leading to lower ETAG concentrations. There were a lot of ASA 3 and 4 patients that might not have tolerated higher concentrations. Or maybe they were becoming dangerously hypertensive and the physician felt that a higher level of gas was needed (these MACs are standard numbers, but you need to control for age, temperature, etc)

    The study isn't perfect, but it's reasonable. Certainly I don't suspect backroom dealings.

  7. Great study. Reinforces dubious data presented in past supporting BIS monitoring. Hard to believe a single BIS reading is predictive of awareness given complex mechanisims of general anesthesia. Also highlights how when a medical company is pushing a device any data they get even remotely near is tarnished.

  8. The following points may be relevant (though I won't claim credit since they were pointed out to me)

    1) As pointed out already, the study was underpowered. It used the power calculation for the B-aware study, expecting a rate of awareness of 1%. However, the patient groups weren't really equivalent, with the NEJM group not particularly at risk. In fact, you could just about recruit every patient that goes through a tertiary institution going by that criteria.

    2) The inclusion criteria in the NEJM study enabled the patients to be "overdosed". If you give patients a MAC of 0.8, you would expect no awareness. MAC awake is around 0.3 for most volatiles. By contrast, the mean MAC in the B-aware study was 0.6, with a significant number experiencing intraoperative hypotension. Giving some of these patients a MAC of 0.8 would probably have killed them.

    3) Significant chunks of data were missing from many of the patients in both arms.

    4) The first patient that had "awareness" in the BIS group looks like they had their anaesthesia titrated to MAC rather than BIS. Their MAC stayed within their defined range for their study, but BIS stayed will below.

    The reality is this study does not draw any conclusions and will not change my practice. I will still use a BIS in a patient that is at high risk of awareness (as defined by proven criteria). I will still use all the information available to me including BIS, MAC, vital signs, in determining depth of anaesthesia... you don't turn the gas up if the BIS is 75 but the PR is 40 and the BP is 70/20. The editorial seems to imply we did that.

  9. It is clear from the picture of the BIS shown that EMG was not selected as the secondary trace.

    Although not FDA approved, trending EMG as a secondary trace makes the BIS device enormously useful even in patients with Botox in their frontalis muscle. Frowning is rarely, if ever, seen with an EMG spike.

    The difference between BIS/EMG monitoring and BIS alone is analogous to the difference between driving one's vehicle while looking through the windshield versus the rear view mirror.