Purpose:

To review the evidence on screening for atrial fibrillation (AF) in older adults for populations and settings relevant to primary care in the United States.

Data Sources:

MEDLINE, the Cochrane Library, and trial registries through October 5, 2020; bibliographies from retrieved articles, outside experts, and surveillance of the literature through October 31, 2021.

Study Selection:

Two investigators independently selected English-language studies using a priori defined criteria. We included trials that evaluated the benefits or harms of screening for AF in adults age 50 years or older without known symptoms, diagnosis of AF, or previous stroke compared with no screening or usual care. We included studies of screening with devices feasible or referable from primary care settings. For treatment benefits and harms, we included trials of anticoagulation treatment for primary stroke prevention (warfarin or direct oral anticoagulants [DOACs]) compared with placebo or no treatment among persons with AF. Eligible outcomes included diagnostic yield, test accuracy, all-cause mortality, stroke, stroke-related morbidity, quality of life, and harms from screening or treatment. We also included systematic reviews reporting on anticoagulation benefits or harms and observational studies reporting harms. We excluded studies with poor methodological quality and studies conducted in developing countries.

Data Extraction and Analysis:

One investigator extracted data and a second checked accuracy. Two reviewers independently rated methodological quality for all included studies using predefined criteria. When at least three similar studies were available, we conducted meta-analyses.

Data Synthesis:

We included 26 studies (in 33 publications). One randomized, controlled trial (RCT) designed to evaluate health outcomes randomized all residents aged 75 or 76 years within two geographic regions (N=28,768) to twice-daily ECG screening for 2 weeks or to no screening. Of those invited to screening, 51.3 percent participated. At a median followup of 6.9 years, the rate of composite endpoint events (ischemic stroke, hemorrhagic stroke, systemic embolism, all-cause mortality, and bleeding leading to hospitalization) was significantly lower in the invitation-to-screening group (5.45 events/100 person years) compared with the control group (5.68 event/100 person years) with an unadjusted hazard ratio (HR) of 0.96 (95% confidence interval [CI], 0.92 to 1.00; p=0.045). No significant differences were observed between the invitation-to-screening group and the control group for any of the individual outcomes contributing to the composite endpoint. Two additional RCTs reported health outcomes, but data were limited.

In eight RCTs (n=86,145) evaluating various ECG screening strategies, more cases of AF were detected when compared with no screening (risk difference range 0.06% to 4.8% over 4 to 12 months); statistically significant larger differences between groups were observed for studies using intermittent or continuous ECG compared with one-time testing approaches. No differences in cases detected were observed in two RCTs (n=12,867) comparing one-time ECG screening to pulse palpation reminders. In seven studies of test accuracy (n=4,497) for various one-time screening strategies (single- or 12-lead ECG, oscillometric blood pressure monitors with AF detection algorithms), sensitivity ranged from 0.80 to 1.00, and specificity ranged from 0.76 to 1.00 when compared with a 12-lead ECG interpreted by one or more cardiologists. In a population of 1,000 persons with a 1.3 percent prevalence of previously unknown AF, this would result in between 0 to 9 false-negative tests and 0 to 237 false-positive tests.

Four RCTs (N=43,633) and one cohort study (n=5,214) reported potential harms of screening. Increased detection of non-AF arrhythmias (1 RCT, 1 cohort) and increased initiation of anticoagulation, antiarrhythmics, and procedures were observed (2 RCTs, 1 cohort) for screening compared with no screening, but the clinical consequences of these findings are not known. Skin irritation from continuous ECG patch ranged from 1.2 to 1.5 percent of participants (2 RCTs). Limited data exist regarding the impact of screening on anxiety (1 RCT) and bleeding outcomes (2 RCTs) compared with no screening.

Warfarin was associated with a reduced risk of ischemic stroke (pooled risk ratio [RR], 0.32 [95% CI, 0.20 to 0.51]; 5 RCTs; n=2,415) and all-cause mortality (pooled RR, 0.68 [95% CI, 0.50 to 0.93]) compared with placebo over a mean of 1.5 years in populations with clinical, mostly persistent AF that was not screen detected. For a population of 1,000 adults with an annual stroke risk of 4 percent, this translates to an absolute reduction of 28 ischemic strokes and 16 deaths per year. DOACs were also associated with lower incidence of stroke (adjusted odds ratios [ORs] ranged from 0.32 to 0.44) in indirect comparisons with placebo or no treatment. The pooled RR for major bleeding for warfarin compared with placebo was 1.8 (95% CI, 0.85 to 3.7; 5 RCTs; 2,415 participants), and the adjusted ORs for major bleeding for DOACs compared with placebo or no treatment ranged from 1.38 to 2.21 but did not exclude a null effect. In one observational study of 26,628 participants, the adjusted hazard ratio for time to first bleeding event for participants receiving anticoagulation over 2 years was 1.7 (95% CI, 1.3 to 2.3) compared with those not receiving anticoagulation.

Limitations:

The only study designed to assess the direct benefits and harms of screening had poor fidelity, did not exclude persons with known AF at baseline, and had some risk of bias due to outcome ascertainment. Trials of warfarin treatment were focused on persons with clinical and persistent AF and were limited to 1.5 years. No studies of anticoagulation treatment focused on screen-detected populations were identified. We did not assess the comparative effectiveness or harms of various anticoagulation treatments.

Conclusions:

The available direct evidence for health outcomes has numerous limitations, precluding a definitive conclusion about screening benefits and harms. Screening with intermittent or continuous ECG strategies in primary care settings can detect more cases of previously unknown AF compared with no screening, but spot one-time ECG screening may not detect more cases than pulse palpation reminders. In low-prevalence settings, spot one-time screening tests may generate more false-positive than true-positive results. In persons with clinically detected AF, warfarin and DOACs reduce the risk of first stroke and all-cause mortality compared with placebo; the evidence also suggests they increase the risk of major bleeding, although estimates for this harm are imprecise. No trials have assessed the benefits and harms of anticoagulation treatment among screen-detected populations.