Population screening for AF may reduce negative outcomes

Widespread screening for atrial fibrillation (AF) may reduce clinical outcomes, suggested findings of the STROKESTOP trial.

Individuals aged 75 or 76 years (mean age 76 years, 54.6 percent female) residing in Stockholm or Halland County, Sweden, were randomized 1:1 into a screening or control group (n=13,979 and 13,996, respectively). In the screening group, individuals without prior AF history had single-lead ECG recordings done twice/day for 14 days and underwent systematic follow-up if AF was detected. Patients with prior AF who were not on oral anticoagulant (OAC) therapy also underwent systematic follow-up. All participants were followed up for ≥5.6 years.

Mean CHA₂DS₂-VASc score at baseline was 3.5. About 15 percent of individuals in the screening and control groups, respectively, had diabetes, 7.5 and 7.8 percent, respectively, had a history of heart failure (HF), 11.1 and 10.8 percent had a history of ischaemic stroke, systemic embolism, or transient ischaemic attack (TIA), 35.5 and 35.6 percent had hypertension, 12.1 and 11.7 percent had vascular disease, and 9.4 and 9.6 percent were on OACs.

Among individuals randomized to the screening group, 51.3 percent (n=7,165) chose to undergo screening. Those who chose to undergo screening were younger compared with those who chose not to participate (mean age 75.8 vs 76.2 years), had a lower CHA₂DS₂-VASc score (mean 3.3 vs 3.7), were less likely to be on OACs (8.2 percent vs 10.5 percent), and had significantly fewer comorbidities (diabetes: 11.6 percent vs 18.9 percent; HF: 4.8 percent vs 10.3 percent; ischaemic stroke, systemic embolism, or TIA: 8.8 percent vs 13.5 percent; hypertension: 31.6 percent vs 39.6 percent; vascular disease: 9.1 percent vs 14.4 percent; p<0.001 for all). [EHRA 2021, Late Breaking Clinical Trials]

In this group, the diagnosis of AF increased following screening, from about 12.1 to 14.0 percent, while there was no significant change in diagnosis rates in the control group (about 12.8 percent).

The incidence of a composite of ischaemic stroke, systemic embolism, all-cause mortality, haemorrhagic stroke, or hospitalization for bleeding was significantly lower in the screening vs the control group (4,456 vs 4,616 events; hazard ratio [HR], 0.96; p=0.045).

“We needed to invite 91 individuals [for screening] in order to prevent one event,” presented study author Dr Emma Svennberg from the Karolinska Institute, Stockholm, Sweden, at EHRA 2021.

This reduction appeared to be driven by the significantly reduced risk of ischaemic stroke among those who chose to participate in screening compared with controls (HR, 0.76; p<0.001).

However, the participants who chose to undergo screening were generally healthier than the non-participants, reminded Svennberg.

“Population-based screening for AF provided a net clinical benefit in an elderly population,” she concluded. “Efforts should be made to increase participation in AF screening. Non-participants were at the highest risk of adverse events,” she added.

While the event rates were lower than anticipated, the survival curves began to separate after 4 years, suggesting relevant long-term effects, noted discussant Professor Renate Schnabel from the University Heart Center Hamburg, Germany.

Where to go from here?

“[AF] is often asymptomatic and first manifests with its complications,” remarked Schnabel.

The current ESC guidelines recommend opportunistic screening in individuals aged ≥65 years but the evidence for the benefit of systematic screening is weaker, she added.

“[T]he future guidelines have gained strong evidence to judge on systematic AF screening by prolonged intermittent handheld device,” said Schnabel. “How to implement AF screening, including systematic screening in healthcare systems across Europe and beyond, remains an open question,” she said.

“We will [also] need to see whether large-scale systematic screening approaches, such as in the STROKESTOP study, will be bypassed by increasingly common consumer-led screening through wearables and apps,” she continued.

Schnabel pointed out that these findings should encourage cost-effectiveness analyses of systematic screening programmes, which until now have “largely relied on assumptions.”