A-fib Screening Yields Promising Results, But Uncertainty Remains: mSToPS and VITAL-AF

Monitoring with a wearable ECG patch was tied to improved clinical outcomes, but a single-lead strategy was not.

A-fib Screening Yields Promising Results, But Uncertainty Remains: mSToPS and VITAL-AF

Assessing A-fib with a wearable ECG patch is feasible and results in lower rates of adverse clinical outcomes over 3 years compared with usual care, according to results of a recent trial. However, another study found that single-lead ECG screening during routine visits did not improve diagnosis.

In the mHealth Screening to Prevent Strokes (mSToPS) trial, patients who wore a Zio patch heart monitor (iRhythm) had a lower risk of a composite of death, stroke, systemic embolism, or MI compared with patients followed in the usual manner, with a greater difference observed in the subset of patients diagnosed with A-fib. That comparison was not randomized, however.

In the other study, the cluster-randomized VITAL-AF trial, screening patients during primary-care visits with the AliveCor single-lead ECG was feasible but did not increase the rate of A-fib diagnosis compared with usual care. There was a hint that such a strategy might be useful in the oldest patients.

The results of both investigations, presented during the virtual American Heart Association (AHA) 2020 Scientific Sessions this week, suggest that additional research is needed to answer questions about where screening for A-fib might make sense, and how it should be done.


Introducing the rationale for the mSToPS study, Steven Steinhubl, MD (Scripps Research Translational Institute, La Jolla, CA), noted that even though much attention is paid to preventing stroke in patients with A-fib, the arrhythmia is also an independent risk factor for HF and CV mortality.

A-fib is often not recognized until a serious clinical event like stroke or HF occurs, suggesting that earlier screening might be able to improve patient outcomes, Steinhubl said.

The primary objective of the trial, conducted among people covered by Aetna insurance, was to see whether immediate monitoring with the self-applied Zio ECG patch—sent by mail to participants—increased new diagnoses of A-fib by 4 months relative to delayed monitoring with the patch. It did, according to results published in JAMA in July 2018.

A secondary objective was to assess the relationship between monitoring with the patch (either immediate or delayed) and 3-year clinical outcomes when compared with an observational cohort of patients receiving usual care. The analysis presented by Steinhubl at AHA 2020 included 1,718 participants randomized to one of the monitoring groups and 3,371 matched observational controls. Overall, mean participant age was 74 years, and 41% were women. Median CHA2DS2-VASc score was 3.

Through a mean follow-up of 29 months, the cumulative probability of receiving a new A-fib diagnosis was higher in the monitored group (11.4% vs 7.7%; P < 0.01). About one-third of the new diagnoses in the monitored group were made via the patch. There was no difference between the monitored and control groups in the rate of initiation of anticoagulation (45.2% vs 44.0%; P = 0.84).

During follow-up, the rate of death, stroke, systemic embolism, or MI was lower in the monitored group (4.5 vs 5.5 per 100 person-years; HR 0.79; P < 0.01). There was a more substantial difference in the subset of patients diagnosed with A-fib (8.4 vs 13.8 per 100 person-years; HR 0.53; P < 0.01), driven mostly by an advantage among patients diagnosed via the patch. “It’s important to highlight that these analyses are likely confounded by spectrum bias and that, therefore, [they’re] being shown purely for mechanistic consideration only,” Steinhubl said.

In terms of safety, there was a lower rate of hospitalization for bleeding in the monitored group (0.32 vs 0.71 per 100 person-years; adjusted incident rate ratio 0.47; 95% CI 0.26-0.85).

Several limitations need to be kept in mind when interpreting these results, Steinhubl said: the potential for unmeasured confounding; the collection of endpoints based on claims data, limiting follow-up to the length of health plan enrollment; the fact that treatment decisions were based on physician preference; and the low rate of anticoagulation initiation in both groups.

“Active screening for A-fib as part of a pragmatic, direct-to-participant, nationwide study, was associated with a significant improvement in clinical outcomes and safety at 3 years relative to routine care,” he said. “However, independent replication of these results [is] required in order to be confident that aggressive pursuit of diagnosing atrial fibrillation in people at high risk, but without symptoms, is warranted.”


Steven Lubitz, MD (Massachusetts General Hospital, Boston), and colleagues evaluated a different approach to A-fib screening in the VITAL-AF cluster-randomized trial: single-lead ECG screening during routine primary-care visits. Screening in primary care clinics may be optimal because it can identify people who might benefit from treatment at a mass scale and enable doctors to initiate treatments efficiently, Lubitz said.

There are limited and conflicting randomized data testing this idea, however, which is reflected in variations in professional recommendations. The US Preventive Services Task Force (USPSTF), for one, has concluded that there is insufficient evidence to recommend for or against ECG screening in asymptomatic adults with an intermediate or high risk of CVD.

VITAL-AF included 16 hospital- and community-based practices within the Massachusetts General primary care network randomized to offer screening for A-fib with the AliveCor single-lead ECG at routine visits or not. Of the 15,397 patients (with 38,891 encounters) in the intervention arm, 91% were screened. That compares with just 2% of the 15,325 patients (with 40,459 encounters) in the control group.

Despite that difference, systematic screening did not increase the rate of newly diagnosed A-fib at 1 year (1.74% vs 1.60%; P = 0.33). An analysis stratified by age, however, suggested a larger increase in diagnosed A-fib with routine screening in patients 85 and older (absolute risk difference 1.88%; number-needed-to-screen: 53).

The proportion of patients with newly diagnosed A-fib in whom anticoagulation was initiated was slightly over 70% in both arms of the trial, with no difference between groups (P = 0.61).

Even though the study showed that routine A-fib screening using this approach is feasible in primary care, screening everybody age 65 and older “is not a very efficient way to identify undiagnosed atrial fibrillation. However, it’s likely that it’s much more efficient to screen individuals who are at very high risk of atrial fibrillation,” Lubitz told TCTMD, pointing to patients 85 and older as a possible group to target.

More Randomized Data Needed

In discussing the results of mSToPS and VITAL-AF, as well as SEARCH-AF, Ben Freedman, MBBS, PhD (Charles Perkins Centre, University of Sydney, Australia), pointed out that in order for groups like the USPSTF to recommend routine ECG screening for A-fib, they would need more data on the prognosis of screen-detected A-fib, whether screening detects more A-fib than usual care, and whether stroke is reduced with screening. All three studies provided information on the frequency with which A-fib is detected, and mSToPS provided information on the potential impact of screening on stroke risk, although none of the presentations provided much insight into the prognosis of screening-detected A-fib, Freedman said.

Regarding VITAL-AF, Freedman noted that the diagnosis of A-fib in the control arm was “very good,” something seen in the recently reported D2AF trial that also showed no increase in A-fib diagnosis with primary-care screening. “So I think these are not negative trials, [but rather] just the triumph of usual care,” he said. “You don’t need a screening strategy if you’re already doing a really good job with single time point screening.”

As for mSToPS, “this result on prognosis is important but only hypothesis-generating in my view,” Freedman said, pointing to the trial’s limitations. “Stroke, MI, [and] death may be better with a strategy of screening versus no screening as in mSToPS, but to be confident we need a randomized controlled trial to convince the USPSTF and payers, and there are trials that are underway.” He pointed to STROKESTOP, SAFER, GUARD-AF, and HEARTLINE.

“We need more data from randomized controlled trials before we can really say screening is worthwhile,” he concluded.

Todd Neale is the Associate News Editor for TCTMD and a Senior Medical Journalist. He got his start in journalism at …

  • Lubitz SA. Screening for atrial fibrillation in older adults at primary care visits using single lead electrocardiograms: the VITAL-AF trial. Presented at: AHA 2020. November 16, 2020.

  • Steinhubl SR. 3-year clinical outcomes in a nationwide pragmatic clinical trial of atrial fibrillation screening–mHealth Screening to Prevent Strokes (mSToPS). Presented at: AHA 2020. November 16, 2020.

  • VITAL-AF was funded by the Bristol-Myers Squibb/Pfizer Alliance.
  • Lubitz reports receiving grants from the US National Institutes of Health, the AHA, Bayer AG, Boehringer Ingelheim, Fitbit, and IBM; and consulting for Bayer AG and Blackstone Life Sciences.
  • Freedman reports relationships with Bayer, Bristol-Myers Squibb/Pfizer, Daiichi Sankyo, Omron, and AliveCor.