Effect of machine learning-guided haemodynamic optimization on postoperative free flap perfusion in reconstructive maxillofacial surgery: A study protocol.

ABSTRACT

AIMS: Intraoperative hypotension and liberal fluid haemodynamic therapy are associated with postoperative medical and surgical complications in maxillofacial free flap surgery. The novel haemodynamic parameter hypotension prediction index (HPI) has shown good performance in predicting hypotension by analysing arterial pressure waveform in various types of surgery. HPI-based haemodynamic protocols were able to reduce the duration and depth of hypotension. We will try to determine whether haemodynamic therapy based on HPI can improve postoperative flap perfusion and tissue oxygenation by improving intraoperative mean arterial pressure and reducing fluid infusion. METHODS: We present here a study protocol for a single centre, randomized, controlled trial (n = 42) in maxillofacial patients undergoing free flap surgery. Patients will be randomized into an intervention or a control group. In the intervention, group haemodynamic optimization will be guided by machine learning algorithm and functional haemodynamic parameters presented by the HemoSphere platform (Edwards Lifesciences, Irvine, CA, USA), most importantly, HPI. Tissue oxygen saturation of the free flap will be monitored noninvasively by near-infrared spectroscopy during the first 24 h postoperatively. The primary outcome will be the average value of tissue oxygen saturation in the first 24 h postoperatively.