Prediction of hemodynamic reactivity using dynamic variations of Analgesia/Nociception Index (∆ANI).

Abstract

The Analgesia/Nociception Index (ANI), a 0-100 non-invasive index calculated from heart rate variability, reflects the analgesia/nociception balance during general anesthesia. We hypothesized that dynamic variations of ANI (∆ANI) would provide better performance than static values to predict hemodynamic reactivity during desflurane/remifentanil general anesthesia. One hundred and twenty-eight patients undergoing ear-nose-throat or lower limb orthopedic surgery were analyzed in this prospective observational study. The ANI, heart rate and systolic blood pressure were recorded before induction, at skin incision, during procedure and at emergence from general anesthesia. Changes in these variables were recorded after 1min for ANI (ANI1min) and 5min for heart rate and systolic blood pressure. The dynamic variation of ANI at the different time points was defined as: ∆ANI=(ANI1min-ANI)/([ANI+ANI1min]/2). Receiver-operating characteristic (ROC) curves were built to evaluate the performance of ANI, ANI1min and ∆ANI to predict hemodynamic reactivity (increase by more than 20% in heart rate and/or systolic blood pressure within 5min). For the prediction of hemodynamic reactivity, better performance was observed with ∆ANI (area under ROC curve (AUC ROC)=0.90) in comparison to ANI (ROC AUC=0.50) and ANI1min (ROC AUC=0.77). A ∆ANI threshold of -19% predicts hemodynamic reactivity with 85% [95% CI 77-91] sensitivity and 85% [95% CI 81-89] specificity. Dynamic variations of ANI provide better performance than static values to predict hemodynamic reactivity during desflurane/remifentanil general anesthesia. These findings may be of interest for the individual adaptation of remifentanil doses guided by ∆ANI during general anesthesia, although this remains to be demonstrated.