Predictors of blood pressure reductions with a second measurement in individuals with uncontrolled blood pressure in primary care clinics

Hypertension is a leading risk factor for morbidity and mortality worldwide. Accurate measurement of office blood pressure (BP) is critical to diagnosis and determining the efficacy of therapy. Current guidelines recommend more than one in-office BP measurement to improve accuracy.¹ Multiple office BP readings result in lower averaged BP, better correlation with ambulatory BP, and improved cardiovascular risk prediction.^{2, 3} This is thought to be due to higher initial BP from the alerting reaction and a partial contribution of regression to the mean with subsequent measurements.^{3, 4} However, features that contribute to changes in BP with additional in-office measurements remain unclear. We studied primary care clinic patients with an initial elevated BP to determine predictors of BP changes with a second measurement.

Institutional Review Board approval was obtained through Vanderbilt University Medical Center (VUMC) including a waived requirement for written informed consent. Over 11 weeks at sixteen VUMC primary care clinics, as a quality initiative clinic personnel were asked to take a second consecutive BP measurement in patients with an initial systolic BP >140 mmHg or diastolic BP >90 mmHg. BP measurement was performed with Welch Allen wall-mounted auscultatory or Phillips SureSign oscillometric devices. Medical histories and demographic data were obtained from the medical record, and laboratory values within 90 days of the BP measurement were included.

Of 2,272 patient encounters for which initial BP was >140 mmHg systolic or >90 mmHg diastolic and a second BP measurement was obtained, 714 encounters were excluded for no laboratory data available within 90 days, 3 encounters were excluded due to no body mass index data, and 2 for heart rate <10 beats per minute which was deemed artifactual. This left 1,553 patient encounters for analysis. Two group comparisons were performed with the Mann-Whitney test. Multivariable linear regression was used to determine predictors of systolic BP changes while controlling for covariates. Statistical testing was performed using GraphPad Prism version 9.3.1.

Demographic data for participants are shown in Supplemental Table 1. In these individuals with elevated BP, a second BP measurement resulted in a 6.2 mmHg lower systolic (P<0.0001; Mann-Whitney) and 3.0 mmHg lower diastolic BP (P<0.0001; Mann-Whitney) (Supplemental Table 1). There were no significant differences in the change between the first and second measurement based on whether BP was measured via auscultatory or oscillometric methods (systolic mean±SEM −6.7±11.7 vs −5.9±10.5, respectively, P=0.19; diastolic −3.3±9.0 vs −2.8±8.9, P=0.62). Multivariable linear regression revealed features associated with decreased systolic BP with a second measurement including higher first BP reading, younger age, male sex, and Asian race (Table 1). Of note, there were no significant associations with method of BP measurement, use of particular antihypertensive therapies, or heart rate (Table 1).

Table 1:

Multivariable linear regression of predictors of the change in systolic BP from the first to second readings.

	Estimated Difference in SBP (mm Hg)	95% CI	P value
First BP measurement (mm Hg)	−0.29	−0.33 to −0.25	<0.0001
Auscultatory vs oscillometric measurement	−0.72	−1.83 to 0.39	0.2013
Age (years)	0.050	0.0060 to 0.094	0.0258
Male vs female sex	−1.31	−2.39 to −0.23	0.0173
Black or African American vs white race	0.19	−1.36 to 1.74	0.8065
Other vs white race	−0.20	−3.31 to 2.91	0.9001
Asian vs white race	−3.59	−6.62 to −0.55	0.0205
Unknown vs non-Hispanic/Latino ethnicity	1.28	−1.72 to 4.28	0.4023
Hispanic/Latino vs non-Hispanic/Latino ethnicity	−0.39	−4.84 to 4.07	0.8652
Former vs never smoker	−0.38	−1.59 to 0.83	0.5354
Current vs never smoker	0.23	−1.61 to 2.07	0.8039
Unknown vs never smoker	−0.32	−6.78 to 6.13	0.9219
BMI (kg/m2)	−0.028	−0.11 to 0.051	0.4916
Pre-existing hypertension vs none	−0.27	−1.58 to 1.05	0.6920
Heart rate (beats per minute)	0.023	−0.015 to 0.061	0.2345
Thiazide diuretics	0.56	−1.02 to 2.14	0.4853
Potassium-sparing diuretics	0.33	−2.80 to 3.44	0.8375
Loop diuretics	−2.19	−5.34 to 0.95	0.1720
Beta blockers	0.57	−3.44 to 4.60	0.7789
ACE inhibitors	1.40	−0.21 to 3.01	0.0886
Angiotensin receptor blockers	−2.45	−5.83 to 0.92	0.1537
Calcium channel blockers	1.39	−1.53 to 4.32	0.3496
Alpha adrenergic blockers	10.73	−9.52 to 31.00	0.2988
Alpha beta adrenergic blockers	2.70	−0.56 to 5.95	0.1040
Vasodilators	−3.22	−24.02 to 17.57	0.7612
SGLT2 inhibitors	−0.65	−5.77 to 4.47	0.8037
Creatinine	0.86	−0.36 to 2.07	0.1664
Potassium	−1.07	−2.45 to 0.31	0.1283

ACE=angiotensin converting enzyme; BMI=body mass index; SBP=systolic blood pressure; SGLT2=sodium-glucose co-transporter 2. Regression model incorporates all variables included in the Table.

Although guidelines recommend more than one in-office BP measurement for hypertension diagnosis or treatment decisions, real-world logistical challenges have kept this from being performed uniformly.^{1, 3, 5} We found a lower systolic and diastolic BP with a second reading, consistent with prior studies.^{3, 4} We also present novel evidence that higher initial BP, male sex, younger age, and Asian race are predictors of BP decreases with a second measurement in individuals with elevated first readings. Higher initial BP resulting in a greater decrease with a second reading is consistent with a prior retrospective study³ and likely reflects increased sympathetic outflow as part of the alerting reaction that raises the initial BP which declines with subsequent measurements.⁴ The alerting reaction is a psychological stress reaction to the initial BP measurement with associated increased sympathetic outflow.⁶ Our finding that male sex is associated with a greater BP decrease with repeat measurements is also consistent with this, given that males exhibit increased sympathetic nervous system activity compared to females.^{7, 8} The association of younger age with larger systolic BP decreases on second measurement also likely reflects a greater contribution of sympathetic outflow in young individuals that is overcome by increased arterial stiffness with older age.⁹ While baseline heart rate did not predict changes in systolic or diastolic BP, additional studies of heart rate variability or muscle sympathetic nerve activity will be required to determine the direct association of sympathetic outflow with initial and subsequent BPs. Importantly, association of these demographic features would not be expected if BP reductions were simply a function of regression to the mean, suggesting that physiological factors contribute to this observation. Finally, there was no difference in the change in BP from the first to the second readings with auscultatory or oscillometric devices, indicating that clinics implementing ascertainment of two BP measurements would expect to observe the same decreases in BP using either type of device.

A prior study identified a lower proportion of African Americans with decreased BP on additional measurements compared to sustained hypertensives.⁴ We did not observe this relationship, though our study population consisted of limited number of African American participants. We did observe a novel association of Asian race with BP decreases on subsequent readings. This warrants further investigation including in cohorts with greater representation of Asian race, particularly given the increased hypertension and cardiovascular disease risk amongst those of South Asian ancestry.¹⁰

Limitations of our study include that only patients with initial BP >140/90 mmHg were included, so further studies will be needed to determine if findings are similar in those with initial normotensive BPs. In addition, our analysis does not include comparison to out of office BP readings, which are also recommended for hypertension diagnosis and to guide treatment decisions.¹ Finally, there was not a defined period of rest before obtaining the first BP measurement, so there could be a contribution to the reduction in BP with the second measurement from additional time for rest.

Overall, our analysis confirms prior findings that obtaining a second BP measurement results in a lower average BP reading.^{3, 4} In addition, certain features such as younger age, male sex, Asian race, and higher elevated first BP reading predict decreases with a second reading, which are consistent with an excessive alerting response. Our findings also suggest that certain populations may derive particular benefit from a second measurement including younger males, Asians, and those with a more elevated first BP reading which can help avoid inappropriate hypertension diagnoses and/or over-treatment.