Response to Prognostic Significance of Between-Arm Blood Pressure Differences: Between-Arm Blood Pressure Difference and Mortality and the Interarm Blood Pressure Difference: Between-Arm Blood Pressure Differences Are Important
We thank Kleefstra et al1 for raising important questions about the methods and conclusions of our study.2 The blood pressure device that we used is now discontinued but is substantially equivalent to the HEM-442C Manual Inflation BP Monitor (John Winegardner, Omron Healthcare, written communication, 2005.). The automatic version HEM-432C, also known as the Omron M1 Classic Model, has been validated using the International Protocol of the Working Group on Blood Pressure Monitoring of the European Society of Hypertension.3
Reproducibility of blood pressure measurement is undoubtedly influenced by time. The longer we wait between measurements, the less reproducible the readings are likely to be. Thus, it is no surprise that reproducibility of measurements >1 year is less compared with those performed over 1 week. Because it is also likely that, over the longer term, progression of atherosclerotic disease or treatment of hypertension may influence the reproducibility of blood pressure measurements, the shorter-term reproducibility more likely reflects the true reproducibility of between-arm differences in blood pressure.
Our interest was in exploring the differences in blood pressure between arms and mortality. Accordingly, we measured 3 blood pressures in each arm and averaged the readings to compare the interarm differences. The order of measurement was not specified. We agree that concurrent measurements in both arms are less likely to reveal false differences, and a device that can capture concurrent measurements would be an important tool for future research. Nonetheless, if all that we obtained with sequential measurements were such false differences, then we would not have observed a mortality signal.
The relationship of minute-to-minute differences in blood pressure with mortality was not a hypothesis of our study. However, when we analyzed our data using the Cox model, with average systolic blood pressure and SD of the blood pressures as predictors of death, we found no evidence of the variability in blood pressure contributing to mortality.
Clark et al4 identified 31 studies of interarm blood pressure differences to test the prevalence and magnitude of between-arm blood pressure differences. Twenty studies were excluded because of the lack of simultaneous BP measurements, and only 4 studies (1 of which had 1000 normotensive subjects) were finally analyzed. Systolic differences between arms of ≥10 mm Hg were found in 19.6% and of ≥20 mm Hg in 4.3%. However, among the studies excluded, the respective interarm differences were 34.5% and 7.5%. Whether the excluded studies had greater prevalence of vascular disease when compared with the included studies was not analyzed. Therefore, whether the interarm differences were solely because of sequential measurements or other factors, such as obstructive arterial disease, is unclear from their report.
Clark et al4 assert that their finding that those with more cardiovascular disease have more interarm blood pressure difference is evidence of greater reproducibility. However, reproducibility, as we report it, is the agreement between interarm differences in BP on 2 occasions. This is quite different from the definition by Clark et al4 of reproducibility.
Clark et al4 cite methodologic concerns and wonder at the high prevalence of the interarm differences in blood pressure and the associated outcome. The prevalence of interarm differences is most likely dictated by underlying obstructive arterial disease, which can vary between populations. Another reason may be sequential measurements. In fact, studies reporting the vital outcomes by Clark et al4 have evaluated sequential—not simultaneous—interarm blood pressure differences.5,6
They reported that 13.6% patients attending “reregistration medicals” between the ages of 6 and 88 years had systolic BP differences of ≥20 mm Hg, and this was associated with mortality; mortality was not corrected for age or other vascular risk factors.7 The study of Aboyans et al7 was composed predominantly of a population of patients with vascular disease: those participating in Lipid Research Clinic protocol, Walking and Leg Circulation Study cohort, and those referred to the San Diego Veterans Hospital for noninvasive vascular testing. These populations are highly selected and not representative of the general population. Thus, we continue to remain cautious about the extrapolating of our findings to less selected populations with limited vascular disease. In the latter populations, large interarm differences in blood pressure may be false-positive signals for vascular disease and may not correlate with outcomes.
In summary, the short-term variability in blood pressure in our study was not associated with mortality. The reproducibility of interarm difference, and prognostic signal associated with it reinforce our belief that the arm designated for blood pressure measurement cannot be used interchangeably, at least in a population with a high prevalence of atherosclerotic vascular disease. We agree with Clark and Campbell8 that the interarm blood pressure difference and the associated mortality quite likely reflect underlying obstructive arterial disease and, in a population with a high prevalence of atherosclerotic vascular disease, it would be wise to measure blood pressure in both arms.
Kleefstra N, Houweling ST, Logtenberg SJJ, Bilo HJG. Prognostic significance of between-arm blood pressure differences: between-arm blood pressure difference and mortality. Hypertension. 2008; 52: e14.
Agarwal R, Bunaye Z, Bekele DM. Prognostic significance of between-arm blood pressure differences. Hypertension. 2008; 51: 657–662.
Clark CE, Powell RJ. The differential blood pressure sign in general practice: prevalence and prognostic value. Fam Pract. 2002; 19: 439–441.
Clark CE, Campbell JL. The interarm blood pressure difference. Hypertension. 2008; 52: e15.