Prevalence of systolic inter-arm differences in blood pressure for different primary care populations: systematic review and meta-analysis

Christopher E Clark, Rod S Taylor, Angela C Shore, John L Campbell

Various prevalence figures have been reported for inter-arm differences in blood pressure (IAD); variation may be explained by differing population vascular risk and by measurement method. To review the literature to derive robust estimates of IAD prevalence relevant to community populations. Systematic review and meta-analysis. MEDLINE, Embase, and CINAHL were searched for cross-sectional studies likely to represent general or primary care populations, reporting prevalence of IAD and employing a simultaneous method of measurement. Using study-level data, pooled estimates of mean prevalence of systolic IADs were calculated and compared using a random effects model. Eighty IAD studies were identified. Sixteen met inclusion criteria: pooled estimates of prevalence for systolic IAD ≥10 mmHg were 11.2% (95% confidence interval [CI] = 9.1 to 13.6) in hypertension, 7.4% (95% CI = 5.8 to 9.2) in diabetes, and 3.6% (95% CI = 2.3 to 5.0) for a general adult population (P<0.001 for subgroup differences). Differences persisted for higher cut-off values. Prevalences were lower for East Asian than for Western populations and were overestimated by sequential measurement where this could be compared with simultaneous measurement within studies (relative risk for IAD: 2.9 [95% CI = 2.1 to 4.1]). Studies with higher mean absolute systolic pressures had higher prevalences for a systolic IAD ≥10 mmHg (P = 0.04). Prevalences of IADs rise in relation to underlying cardiovascular comorbidities of the population studied, and are overestimated threefold when sequential measurement is used. Population-specific variation in prevalences of IAD should be taken into account in delivering clinical care and in planning future studies.