Objective:
Efficient screening strategies are necessary for detecting atrial fibrillation (AF) aiming at stroke prevention. In 2013 the UK National Institute for Health and Care Excellence recommended opportunistic AF screening during routine office blood pressure (BP) measurement using an automated monitor with high diagnostic accuracy. This systematic review and meta-analysis evaluated the current evidence on the diagnostic accuracy of AF detection technologies implemented in automated devices for office, home, and ambulatory BP measurement.
Design and method:
A systematic PubMed search was conducted using the algorithm: “atrial fibrillation” AND (“blood pressure” OR “blood pressure monitor” OR “blood pressure device” OR “blood pressure measurement”) AND (detect∗ OR screen∗), until May 2022.
Results:
Initial search retrieved 423 articles of which 37 were relevant and 17 were included in meta-analysis. Three studies (n = 1465, weighted age 77.8 years, males 51%, AF prevalence 10.4%) evaluated the diagnostic accuracy of office BP monitors. Meta-analysis showed pooled sensitivity, specificity, and accuracy: 94% (95% CI 88–98%), 94% (89–98), and 94% (89–97), respectively. Eleven studies (n = 8343, weighted age 66.7 years, males 47%, AF prevalence 7%) evaluated the diagnostic accuracy of home BP monitors. Pooled sensitivity, specificity and accuracy were 98% (93–100%), 94% (90–97) and 95% (91–97), respectively. Three studies (n = 238, weighted age 68.0 years, males 62%, AF prevalence 12%) evaluated the diagnostic accuracy of ambulatory BP monitors. Pooled sensitivity, specificity and accuracy were 94% (81–100%), 82% (77–87) and 83% (78–88), respectively. Simultaneous (or shortly before/after BP measurement) 12- or single-lead electrocardiography (ECG) was used as reference method in studies evaluating office/home BP monitors. For ambulatory BP monitors, two studies used 24h-continuous ECG Holter monitoring and one study used pacemaker recordings (implanted for sick sinus syndrome). Analysis per individual was performed for the evaluation of office and home BP monitors (an individual was considered as AF positive if the majority of BP readings indicated AF, e.g., 2 of 3). For ambulatory BP monitors, analysis per reading was performed (vs. simultaneous ECG trace). With increasing AF prevalence across studies, meta-regression analysis showed higher positive predictive value and a trend towards lower negative predictive value.
Conclusions:
There is considerable evidence that AF detection during routine automated BP measurement in elderly people using office, home, or ambulatory BP monitors has high diagnostic accuracy. AF screening in the elderly during routine office and out-of-office BP measurement is feasible and can increase early detection of asymptomatic AF.
