術中低血圧予測モニター

Predictive Hemodynamic Monitoring to Prevent Intraoperative Hypotension during Noncardiac Surgery

はじめに

　術中低血圧（intraoperative hypotension：IOH）による術後合併症増加，予後悪化は重要な課題であり，高齢患者，心血管系併存症を有する患者，フレイル患者が増加している今日，IOH回避は重要性を増している。この課題に対してAnesthesia Patient Safety Foundation（APSF），PeriOperative Quality Initiative（POQI）から文書が発出されており，前者では “encourage monitoring that not only provides early detection but ideally predicts hemodynamic instability”^1），後者ではopen research questionとして “Do predictive algorithms or closed-loop control prevent organ injury?” と記載されている^2）。これらのステートメントにおける “predict” および “predictive” を最初に実用化した指標が，今回紹介する低血圧予測インデックス（hypotension prediction index：HPI）である。なお，予測機能を実現するために用いられている機械学習は人工知能の一領域で，明示的な規則に基づいたプログラムではなく，コンピューターが自律的，段階的にデータから学習するプロセスと説明されている^3）。

　Organ damage caused by intraoperative hypotension（IOH）is a serious clinical problem, and the prediction of IOH is attracting attention as a means of preventing IOH. Software that can predict the occurrence of IOH up to 15 minutes in the future by analyzing arterial pulse contour characteristics（e.g., Acumen IQ^TM software, BD, USA）has been put to practical use. This narrative review provides an overview of the Acumen IQ^TM software and the results of clinical studies in which this software was used. The Acumen IQ^TM software displays a hypotension prediction index（HPI）, the dynamic arterial elastance, and the systolic slope, and it provides an HPI alert to assist intervention when IOH is anticipated. The accuracy of the HPI in predicting the occurrence of IOH has been evaluated using the receiver operator characteristics curve, and an accuracy value close to 0.9 has been reported for predicting the occurrence of IOH up to 15 minutes ahead, while a positive predictive value of approx. 0.3 has been reported. Although many reports indicate that the HPI significantly reduces the duration and time weighted average of IOH, further investigation is needed to determine whether using the HPI reduces organ damage from IOH and improves patient prognoses.