Reader Questions “Triple Low” Conclusions

Nathaniel F. Simon, MD; Daniel I. Sessler, MD

To the Editor

I was alarmed at several conclusions reported in the Winter 2010 edition in the article synopsizing scientific papers from the 2009 ASA Annual Meeting. Specifically, the section entitled “Miscellaneous ‘Triple Low'” contained statements like, “Increased duration of [“triple-low”] increased the incidence of 30-day readmission and postoperative mortality” and “low BIS levels further increased relative mortality.” I do not have to tell you that these studies were retrospective analyses and cannot provide any association of variables other than correlation, certainly not causation. Furthermore, this notion of associating low bispectral index values with specific outcomes is in its nascence, with much remaining to be discovered. The “triple-low” phenomenon, in fact, exemplifies this, as one cannot simultaneously solve the “problem” of low anesthetic concentration and low BIS. To make a statement explicitly identifying a causative link between the two is not only poor form, it is scientifically and medicolegally irresponsible.

Nathaniel F. Simon, MD
Sacramento, CA

In Reply:

As Simon notes, it would be poor form and irresponsible to make statements explicitly identifying a causative link based on observational data. We thus did no such thing.

All of our published and presented work related to Triple Low clearly identifies the observational nature of our work and that we have only identified associations. We were equally clear that Aspect Medical (now Covidien) funded the project and that Aspect employees were involved in the analysis.

Rather than relying on the editorial (which we did not write or review), I encourage readers to review the abstracts on which it was based (A6, A880, and A354). Note use of words such as “predictor” and “associated.” Nowhere do we imply causality.

Low mean arterial pressure, low minimum alveolar concentration, and low bispectral index are reported to be independent predictors of mortality. For example, the association of low bispectral index and mortality is the subject of an article and editorial in the May issue of Anesthesiology. That combinations of 2 or 3 of these factors would also predict adverse outcomes is thus unsurprising. The purpose of our analysis is to show that specific combinations are especially predictive and, therefore, potential targets for intervention. Based on our current analysis, we hope to soon start a prospective evaluation.

Daniel I. Sessler, MD
Professor and Chair, Department of Outcomes Research
The Cleveland Clinic, Cleveland, OH