DOI: 10.1213/ane.0000000000008179 ISSN: 0003-2999

A Novel Algorithm for Continuous Real-Time Cerebral Autoregulation Assessment Based on Mean Arterial Pressure and Cerebral Oxygen Saturation

Antonio Albanese, Zhengmin Ma, Rogier V. Immink, Hong Liu, Denise P. Veelo, Alexander P.J. Vlaar, Bin Zhang, Gregory Fischer, Zhongping Jian, Feras Hatib, Paul Benni, Charles W. Hogue
Show PDF Cite

BACKGROUND:

Continuous and real-time assessment of cerebral autoregulation can be of important clinical value to individualize blood pressure targets in perioperative settings. There is a high interindividual variability of the lower (LLA) and upper (ULA) limits of cerebral blood flow autoregulation, and exposure to blood pressure values outside of these limits has been associated with complications. We have developed a novel algorithm for continuous real-time assessment of cerebral autoregulation based on analysis of the dynamic interactions of mean arterial pressure (MAP) and near-infrared spectroscopy cerebral oxygen saturation (St

o
2 ) measurements. The algorithm generates an index, the cerebral autoregulation index (CAI), which characterizes the effectiveness of cerebral autoregulation on a 0 to 100 scale. The aim of this study is to validate the algorithm using data from animals and surgical patients.

METHODS:

MAP, cerebral St

o
2 , and cerebral laser-Doppler blood flow (CBF) data were collected as part of an animal study on a piglet model of controlled hypotension. Additionally, simultaneous MAP, cerebral St
o
2 , and transcranial Doppler cerebral blood flow velocity (CBFV) data were collected on patients in a multicenter prospective observational study during surgery. Individual plots of CBF/CBFV versus MAP were constructed retrospectively for both the animal and human data, and ground truth labels of cerebral autoregulation status were obtained by identifying on these curves the LLA and ULA values. CAI values were generated by postprocessing MAP and cerebral St
o
2 data through the algorithm. Receiver operating characteristic (ROC) analysis was then conducted to assess the capability of the algorithm to discriminate impaired autoregulation, where MAP is beyond the individual LLA/ULA limits, from intact autoregulation, where MAP is between LLA and ULA.

RESULTS:

Seventy-one patients were enrolled in the human study, and the ROC analysis showed an area under the ROC curve (AUC) (95% confidence interval) of 0.92 (0.89–0.94), with a sensitivity and specificity of 0.82 (0.76–0.87) and 0.94 (0.92–0.96), respectively, at the CAI threshold of 45. In addition, 10 female piglets underwent a controlled hypotension protocol where MAP was lowered below the LLA. The ROC analysis showed an AUC of 0.99 (0.98–1.00), with a sensitivity and specificity of 0.95 (0.90–0.99) and 0.96 (0.94–0.98), respectively.

CONCLUSIONS:

The study demonstrates that the CAI algorithm, using MAP and processed St

o
2 signals, is accurate in discriminating states of intact autoregulation from states of impaired autoregulation. This algorithm may allow for personalized cerebral autoregulation-oriented blood pressure management during surgery.

More from our Archive