Validity of a protocol to estimate patients’ pre-morbid basal blood pressure^*

* This work was presented at the 41st ANZICS/ACCCN Intensive Care ASM 2016 at Perth, Australia.

Figures & data

Figure 1. Patient flow diagram.

Figure 2. Frequency histogram showing normal distribution of the measured basal MAP and the estimated basal MAP.

Table 1. Cohort characteristics and key results.

Key variables
Number of patients (n)	137
Age, mean (SD), y	66 (12)
Female gender, % (n)	93 (68%)
Co-morbidities
Hypertension	107 (78%)
Chronic kidney disease	14 (10%)
Diabetes mellitus	31 (23%)
Stroke	8 (6%)
Peripheral vascular disease	6 (4%)
Number of BP readings per patient during ABPM^a, median [IQR]	9 [8–10]
Measured^a basal SBP, mean (SD), mmHg	123 (18)
Measured^a basal SBP range, mmHg	84–170
Measured^a basal DBP, mean (SD), mmHg	71 (11)
Measured^a basal DBP range, mmHg	48–112
Measured^a basal MAP, mean (SD), mmHg	88 (12)
Measured^a basal MAP range, mmHg	64–130
Number of previous available BP readings per patient, median [IQR]	4 [3–5]
Estimated^b basal SBP, mean (SD), mmHg	124 (13)
Estimated^b basal SBP range, mmHg	83–157
Estimated^b basal DBP, mean (SD), mmHg	70 (8)
Estimated^b basal DBP range, mmHg	49–100
Estimated^b basal MAP, mean (SD), mmHg	88 (8)
Estimated^b basal MAP range, mmHg	63–118
Difference between measured and estimated basal MAP, mean (SD), mmHg	0.3 (12)

SD: standard deviation; SBP: systolic blood pressure; DBP: diastolic blood pressure; MAP: mean arterial pressure; BP: blood pressure; IQR: interquartile range.

^a Measured by nighttime ambulatory BP monitoring.

^b Estimated using a preset protocol (https://www.anzctr.org.au/Trial/Registration/TrialReview.aspx?id =364139).

Figure 3. Box plots of the measured and estimated basal systolic BP (A), diastolic BP (B) and mean arterial BP (C). These are vertical box-and-whisker plots, where the central line within the box represents the median value, the two ends of boxes represent the first and the third quartile values, and the two ends of whiskers represent the upper and lower adjacent values.

Figure 4. Correlation between the measured and estimated basal systolic BP (A), diastolic BP (B) and mean arterial BP (C). These plots show strength and direction of linear relationship between estimated and measured basal BP values. The Pearson’s r statistic of >0.4 in all the three panels represents fair correlation.

Figure 5. Bland-Altman plots for the measured and estimated basal systolic BP (A), diastolic BP (B) and mean arterial BP (C). These plots demonstrate mean bias and upper and lower agreement limits in each panel. The shaded portion represents the pre-designated precision limits of ±15.7 mmHg, which corresponds to a standard deviation of 8 mmHg [15,16] and are considered acceptable to validate a new BP measuring device against the gold standard.

Table 2. Comparison between measured and estimated basal MAP in various subgroups.

	Measured basal MAP,mean (SD), mmHg	Estimated basal MAP,mean (SD), mmHg	P-value (t-test)
Gender
Male (n = 44)	90 (13)	88 (8)	0.28
Female (n = 93)	88 (12)	89 (8)	0.61
Chronic hypertension
No (n = 30)	84 (12)	86 (7)	0.42
Yes (n = 107)	90 (12)	89 (8)	0.44
Chronic kidney disease
No (n = 123)	88 (12)	88 (8)	0.64
Yes (n = 14)	93 (11)	86 (7)	0.02
Diabetes mellitus
No (n = 106)	88 (12)	89 (8)	0.79
Yes (n = 31)	89 (13)	86 (8)	0.36
Previous stroke
No (n = 129)	88 (11)	88 (8)	0.82
Yes (n = 8)	98 (18)	89 (14)	0.04

SD: standard deviation; MAP: mean arterial pressure.

Table 3. Univariate analysis of variables and the bias between measured and estimated basal MAP.

Variables	Regression coefficient (95% CI)	R-squared value	P value
Number of recent clinic BP readings	0.72 (−1.03 to 2.46)	0.49%	0.42
Number of nighttime BP readings during ABPM	0.49 (−0.58 to 1.56)	0.60%	0.38
Age (≥70 vs. <70)	5.8 (2–9.7)	6%	0.003
Female (vs. Male)	−2.6 (−6.8 to 1.5)	1%	0.21
Hypertension (vs. those without)	2.6 (−2.1 to 7.3)	0.9%	0.28
Chronic Kidney Disease (vs. those without)	7.7 (1.4–14)	4%	0.02
Diabetes (vs. those without)	2.6 (−2 to 7.3)	0.9%	0.26
Stroke (vs. those without)	9 (0.9–17.2)	3%	0.03

Table 4. Bias and agreement between measured and estimated basal MAP in different scenarios.

Different estimation methods	Bias or meandifference (mmHg)	P value[Two-tailed paired t-test]	Limits of agreement (mmHg)	Regression coefficient of the bias^a	R-squared value^a	p value^a
Protocol-estimated basal MAP	0.3	0.76	−22 to 22	0.5	16%	<0.001
Standard estimated MAP of 65 mmHg	23.5	<0.0001	0 to 47	1.5	58%	<0.001
Standard estimated MAP of 70 mmHg	18.5	<0.0001	−5 to 42	1.5	58%	<0.001

^aLinear regression of bias with the average of measured and estimated basal MAP.

O'Brien E, Atkins N. A comparison of the British Hypertension Society and Association for the Advancement of Medical Instrumentation protocols for validating blood pressure measuring devices: can the two be reconciled? J Hypertens. 1994;12:1089–1094.

 PubMed Web of Science ®Google Scholar

Lee M, Weinberg L, Pearce B, et al. Agreement between radial and femoral arterial blood pressure measurements during orthotopic liver transplantation. Crit Care Resusc. 2015;17:101–107.

 PubMed Web of Science ®Google Scholar