Agreement of high definition oscillometry with direct arterial blood pressure measurement at different blood pressure ranges in horses under general anaesthesia

ObjectiveTo determine the agreement of high definition oscillometry (HDO) with direct arterial blood pressure measurements in normotensive, hypotensive and hypertensive horses during general anaesthesia. Study designExperimental study. AnimalsSeven healthy warmblood horses, aged 3-11years, weighing 470-565kg. MethodsMeasurements from a HDO device with the cuff placed around the base of the tail were compared with pressures measured invasively from the facial artery. High blood pressures were induced by intravenous (IV) administration of dobutamine (5g kg(-1)minute(-1)) over ten minutes followed by norepinephrine (0.1mgkg(-1) IV) and low pressures by increasing the inspired fraction of isoflurane and administration of nitroglycerine (0.05mgkg(-1) IV). For analysis three pressure levels were determined: high (MAP>110mmHg), normal (60mmHg<MAP<110mmHg) and low (MAP< 60mmHg). Bland-Altman analysis was used for comparison between non-invasive and invasive measurements. ResultsA total of 245 paired measurements of systolic (SAP), mean (MAP) and diastolic (DAP) pressures were obtained. The HDO device underestimated blood pressure at hypertensive and normotensive levels and overestimated blood pressure at hypotensive levels. Best agreement was obtained for SAP and MAP within normotensive limits. At normotension, biasstandard deviation for SAP, MAP and DAP were 0.1 +/- 19.4mmHg, 0.5 +/- 14.0, 4.7 +/- 15.6, respectively. At high pressure levels bias and SD were 26.1 +/- 37.3 (SAP), 4.2 +/- 19.4 (MAP), 1.5 +/- 16.8 (DAP) and at low pressures -20.0 +/- 20.9 (SAP), -11.4 +/- 19.6 (MAP), -4.7 +/- 20.1 (DAP), with HDO measurements at a MAP <50mmHg often failing. Conclusion and clinical relevanceGood agreement with invasive arterial blood pressures was obtained with HDO at normotensive levels in horses. At high and low pressure ranges HDO was unreliable. Therefore, if haemodynamic instability is expected, invasive measurement remains preferable.