|
|
October 16-20, 2010 San Diego, CA Home Abstract Archive Search Abstracts 2010 Session Grid 2010 Meeting Website Policy Statements ASA Website Feedback 
Visit Anesthesiology.org |
| Previous Abstract | Next Abstract |
| Printable Version |
|
A1293
October 20, 2009 2:00 PM - 4:00 PM Room Area M |
| A Wavelet Based Algorithm for Steady-State Detection: Application in BIS Guided Anesthesia |
|
**
Ana Castro, B.S., Fernando G. Almeida, Ph.D., Eduarda Amadeu, M.D., Pedro Amorim, M.D., Catarina S. Nunes, Ph.D. Serviço de Anestesiologia, Hospital de Santo António, Porto, Portugal |
|
Background and Goal of Study: Modeling the behavior of a system considering inputs and outputs can benefit from steady-state detection (SS).
The goal of this study was to detect SS periods on propofol, remifentanil (inputs) and BIS index (output), under target controlled infusion (TCI) TIVA. An indicator of SS for BIS was calculated based on a wavelet transform technique for SS detection [1]. Material and Methods: Data were collected during 20 urological procedures under general anesthesia, with TCI of propofol and remifentanil using Schnider's [2] and Minto's [3] PKmodels. RugloopII software was used to collect data every 5s. Drugs' effect-site concentrations (Ce) were analyzed (15 cases), and SS periods were identified using an arithmetic rule: if the sum of the absolute difference between the target Ce and the estimated Ce of both drugs was lower than 0.01 for more than one minute, SS for the input was considered. BIS from 5 cases was, in a first approach, visually analyzed to identify periods considered as SS, which were later used to tune the wavelet based algorithm. This algorithm filters the signal, removes outliers and calculates the SS index [1]. The index varies from 0 (non -SS) to 1 (SS), with a smoothing period of 30s. The algorithm was then applied to the remaining 15 cases and its performance was visually evaluated. MATLAB R2007a was used for signal analysis and processing. (Data: Mean±SD) Results and Discussion: Data was collected from 20 patients, 13 male, age 54±14 years, weight 70±14kg, height 166±9cm. The results for the 15 patients demonstrated an adequate identification of the SS periods in both the input (drugs) and the output effect (BIS).[figure1]Fig. 1(c)/(d) presents the result of the filtered signal and the SS index in a patient of the test group. The SS index (30s smoothing) is consistent with the BIS trend. The drugs' SS periods are also presented in Fig.1(a) in a different color overlapping the original BIS trend; Fig.1(b) presents the drugs' targets, and the simultaneous measure of deviation used in the arithmetic rule. Drugs' Ce SS periods are not always accompanied by a SS in BIS. This may be explained as a response of the patient to outer stimuli, and not only to the administered drugs. Conclusion: We were able to develop a steady-state detector for drugs and BIS. Observational assessment of its performance suggests that a steady-state detector can be useful to model the effect of the drugs on BIS, and also to provide complementary information to the BIS value. The discrepancy at any given moment between input (drugs) steady-state and output (BIS) steady-state could be interpreted as an indicator of change in noxious stimuli. References: [1] Computers & Chemical Engineering, 2003, 27:569-78; [2] Anesthesiology, 1998, 88:1170-82; [3] Anesthesiology, 1997, 86:24-33. Acknowledgements: FCT; UISPA-IDMEC, Portugal. From Proceedings of the 2009 Annual Meeting of the American Society Anesthesiologists. |
Figure 1
 |