HYPOTENSION PREDICTABILITY INDEX, A NEW TOOL TO PREDICT HYPOTENSION AFTER CARDIAC SURGERY

Abstract

<h3>Introduction</h3> Cardiac surgery is prone to hypotensive periods due to the frequent, unpredictable shifts in fluid status and vasomotor tone. Hypotension has been shown to occur as high as 99% of the cases in the perioperative period and is clearly linked to increased morbidity (acute kidney injury, stroke, myocardial ischaemia) and possibly mortality as well. Hypotension Predictability Index (HPI) a newly developed tool which utilizes the benefits of machine learning to predict hypotension minutes before its occurrence in different surgical situations. However, there is less data available regarding the postoperative period after cardiac surgery. The aim of our study was to monitor HPI in postoperative cardiac surgical patients and validate its effectiveness in predicting hypotension in this cohort. <h3>Methods</h3> In this pilot observational study, 10 cardiac surgical patients were sequentially chosen to be equipped with a Flowtrac IQ pressure transducer (Edwards Lifesciences), which contains the HPI protocol. The Flowtrac IQ can be used as a regular pressure transducer but pairing with a Hemosphere Monitor (Edwards) allows the display and recording of the HPI and blood pressure data. The patients did not receive any additional intervention based on HPI, postoperative care was carried out as per our Intensive Care Unit's protocol. Blood pressure and HPI values from the first 12 post-operative hours were collected and analysed retrospectively. Hypotension was defined as a decrease in mean arterial pressure below 65 mmHg for at least a minute as suggested by previous studies. The data set was analysed in its entirety. The total number of hypotensive events was calculated and analysed in terms of absolute duration. HPI and blood pressure data from the preceding 5 minutes were analysed with Receiver Operating Characteristics (ROC) analysis for correlation. <h3>Results</h3> Patients mean age was 64,4±11,3 years with an Euroscore of 1,7±1,1%. 70% were male. 80% of patient underwent CABG procedure, 20% had AVR, while 1 patient had an atrial mass removal. 90% had at least one hypotensive episode in the postoperative 12 hours. An average of 8.1±11.2 hypotensive events were recorded during the observed period with a mean duration of 4.1±4.0 minutes, which resulted in an average total hypotensive length of 44.2±66,4 minutes. HPI was able to predict the occurrence of hypotension with a sensitivity of 95,1% and specificity of 92,4%. Our data suggests that a cut off value of 92 might provide the best sensitivity and specificity, which is higher than cut off of 85 suggested by the manufacturer. <h3>Conclusions</h3> Hypotension is a frequent occurrence during cardiac surgery, which can be seen in even our small randomly chosen cohort. HPI was able to predict the incidence of hypotension with a good sensitivity and specificity in the post-operative 12 hours after cardiac surgery. This suggests that HPI can be a valuable tool in preventing post-operative hypotension and its complications.