Results
Population demographics, cuff width-to-tail circumference ratios, and height adjustment for the vertical distance between the tail and the scapulohumeral joint (i.e., heart level correction) for each horse are presented in Table 1. A total of 257 paired measurements including the NIBP, MT, FA, and TFA (n = 9 horses) and 124 paired measurements including the Ao (n = 4 horses) were obtained and compared. The ranges of invasive MAP in the different sites were as follows: 27-95 mmHg (Ao), 18-106 mmHg (TFA), 19-102 mmHg (FA), and 26-96 mmHg (MT). Results of comparisons between the MAP for NIBP, FA, TFA and MT using each individual horse’s heart level correction or a standardised correction based on the population median heart level correction of 27 mmHg are listed in Table 2. Using each horse’s individual heart level correction, the NIBP fulfilled the criteria for equivalence only when compared to the FA (p = 0.3); all other location comparisons were not equivalent. Using the standardised 27 mmHg heart level correction for all horses, equivalence was demonstrated between the NIBP and both the FA and MT, but not the TFA (Table 2). In general, the NIBP tended to underestimate the MAP invasive at all peripheral sites (Table 2). However, when using 70 mmHg in the FA as a clinically relevant cut-off to initiate treatment for hypotension, it was found that NIBP overestimated the MAP in the FA by 11 ± 5 mmHg (mean ± SD; range 2-18 mmHg) in 24 out of the 194 hypotensive readings (Figure 1), indicating a false negative rate of 12% (Table 3). Sensitivity, specificity, false positive, and false negative rates when using the individual or the standard heart level correction of 27 mmHg and different definitions of hypotension for NIBP compared against the FA are shown in Table 3. Using this approach, the optimum MAP target for the treatment of hypotension when using NIBP was 80 mmHg if using the 27 mmHg standard heart level correction for all horses (i.e., subtract 27 mmHg from NIBP raw values) or 107 mmHg if using uncorrected NIBP values.
The mean (95% CI) pressure gradients between the Ao and FA, TFA, MT and the coccygeal artery (measured via NIBP) were, respectively, -5 (-4, -7), 2 (0, 4), -3 (-4, -1) and -4 (-6, -2) mmHg. A visual representation of the differences in values between MAP invasive obtained from peripheral arterial sites relative to that obtained centrally at the Ao is shown in Figure 2. The coefficient of variation of MAP in the peripheral arteries relative to the Ao was smallest in the FA (126%), followed by the MT (257%) and the TFA (614%), which indicates that the FA might represent the best peripheral site for invasive arterial blood pressure measurement in dorsally-recumbent anaesthetised horses.