November 16, 2020
5 min read
Lubitz SA, et al. LBS.06: To screen or not to screen, and then what? Studies of detection and treatment of AF. Presented at: American Heart Association Scientific Sessions; Nov. 13-17, 2020 (virtual meeting).
VITAL-AF was funded by Bristol-Myers Squibb/Pfizer. Lubitz reports he received grant funding from the AHA, Bayer, Boehringer Ingelheim, Fitbit, IBM and the NIH and consultant fees from Bayer and Blackstone Life Sciences. The mSToPS trial was investigator-initiated and supported by a grant from Janssen Pharmaceuticals. Steinhubl reports he received a research grant from Janssen and has financial ties with Dynosense, EasyG, Livongo and Otsuka.
More aggressive monitoring for undiagnosed atrial fibrillation may be beneficial in preventing future hospitalization, but the benefit may be limited to higher-risk patients, according to results from the VITAL-AF and mSToPS trials.
Presented at the virtual American Heart Association Scientific Sessions, these two trials evaluated point-of-care and active ECG monitoring methods for the detection of undiagnosed AF and evaluated their impact on guideline anticoagulation use and longer-term outcomes.
In VITAL-AF, investigators enrolled 30,722 patients (mean age, 74 years; 59% women; 83% white) from 16 practices within the Massachusetts General Hospital network to determine whether point-of-care rhythm assessment with a single-lead ECG (AliveCor) may increase newly diagnosed AF compared with a control group who received diagnosis at a follow-up visit.
“VITAL-AF examined the potential for patients with most likely persistent forms of AF to be diagnosed at the time of a primary care practice encounter using a point-of-care screen,” Steven A. Lubitz, MD, MPH, associate professor at Harvard Medical School and cardiac electrophysiologist at Massachusetts General Hospital, told Healio. “Other studies examining ambulatory or continuous rhythm monitoring or short-term repeated forms of monitoring are likely addressing a different question about paroxysmal forms of AF, for which we have less understanding of the potential risks of ischemic stroke.”
Overall, 3% of patients who were screened in the primary care setting were identified as having possible AF, 12% were unclassified and 83% were normal; there was no analysis for the remaining 2%. Of those with possible AF, approximately 50% received a same-day 12-lead ECG, compared with less than 10% of patients with an unclassified or normal screen.
Compared with individuals who were not screened in the point-of-care setting, single-lead ECG did not significantly affect new diagnosis of new AF in the overall population of the VITAL-AF trial (new diagnosis, 1.74% of screening group vs. 1.6% of control arm; P = .33). The incidence rate of new AF diagnosis per 100 person-years in the screening group was 2.59 compared with 2.35 in the control group (P = .27).
However, the association between point-of-care screening and new AF diagnosis was stronger among patients aged 85 years or older, with a risk difference of 1.88% (95% CI, 0.27-3.35) and a number needed to screen of 53 (95% CI, 30-337) compared with the control arm.
After analyzing the proportion of newly diagnosed AF in primary care encounters and the total number of event encounters, investigators determined that the likelihood of AF diagnosis in the primary care setting may be greater with single-lead ECG screening (100 events in 38,374 encounters) compared with the control group (72 events in 40,003 encounters; P = .02).
“VITAL-AF puts in perspective a strategy that has been endorsed in some, but not all, guidelines and that is that opportunistic screening at the time of a clinic encounter may be effective for identifying undiagnosed AF,” Lubitz said in an interview. “We know that if you look for AF, you can find AF, but one of the questions is what is the most efficient way for identifying AF that’s undiagnosed [but] high-risk enough to cause a potential stroke, and that remains an unanswered question.”
In a secondary analysis, researchers observed that although the rate of new anticoagulation initiation declined with increasing patient age, there was no significant difference between the intervention and control arms (P = .61).
“VITAL-AF definitively answers, in a contemporary U.S. practice like ours, the question of what the potential effect is of screening for AF using single-lead ECG,” Lubitz told Healio. “These results should help put in context existing recommendations for screening and identify new areas for future research.”
The second trial was a secondary analysis of the mSToPS study, in which investigators sought to determine whether AF screening via an ECG monitor (Zio patch, iRhythm Technologies) can improve clinical outcomes at 3 years after initiation of screening (mean age, 74 years; 41% women; median CHA2DS2 score, 3). The primary composite outcome was time to death or first stroke, systemic embolism or MI among participants with a diagnosis of AF at any time during the 3-year analysis and among the entire cohort at 3 years. The primary safety endpoint was incidence of hospitalization for bleeding.
As Healio previously reported at the 2018 American College of Cardiology Scientific Session, an ECG patch conferred improved diagnosis of AF, use of guideline-recommended therapies and increased health care utilization.
Steven R. Steinhubl
“The mSToPS trial was a unique, direct-to-participant, siteless design carried out within the Aetna health payer system and was primarily digital outreach recruitment and enrollment,” Steven R. Steinhubl, MD, director of digital medicine at Scripps Translational Science Institute and a cardiologist at Scripps Health, said during his presentation. “Individuals who agreed to participate were sent a Zio patch, which they placed on themselves and the results from the Zio patch were sent back to the participant, and if they agreed, were sent to their physician. No guidance was provided to the physicians as to what to do with the results. Just an explanation of the trial and a brief evaluation of what was found overall.”
Overall, 11.4% of actively monitored patients were diagnosed with AF compared with 7.7% of the matched control group (P < .01). However, 68% of AF diagnoses in the ECG patch group were attributed a clinical diagnosis, not active monitoring.
Anticoagulants were initiated in approximately 44% of the actively monitored cohort and 45.2% of matched controls (P = .84).
Investigators found that participants in the ECG patch group experienced fewer events (4.5 per 100 person-years) compared with the control arm (5.5 per 100 person-years; adjusted HR = 0.79; P < .01).
The association was stronger when researchers looked at only those who received a diagnosis of AF during the 3-year study period, with 8.4 events per 100 person-years in the actively monitored group and 13.8 events per 100 person-years in the control arm (aHR = 0.53; P < .01).
In addition, patients in the actively monitored group who received their AF diagnosis via ECG patch first experienced fewer occurrences of the composite endpoint (2.6 events per 100 person-years) compared with those who were actively monitored and diagnosed clinically and the control group (P for both < .01).
Moreover, incidence of the composite outcome among participants in the ECG patch group who were diagnosed clinically was not significantly different compared with the control group (10.9 vs. 13.8 events per 100 person-years; P = .21).
“When outcomes for the AF diagnosis subset are broken down by the mode of diagnosis, as expected, there is no difference between those who are clinically diagnosed, as their pathway in health care and diagnosis will be similar. It is only those diagnosed with ECG patch that experienced significantly better outcomes,” Steinhubl said during the presentation. “It is important to highlight that these analyses are likely confounded by spectrum bias, and it therefore is being shown purely for mechanistic consideration only.”
For the safety endpoint, patients who were actively monitored had a lower incident rate for total hospitalizations (adjusted incident rate [IR] = 0.69; 95% CI, 0.63-0.76; P < .01) and hospitalizations for bleeding (aIR = 0.47; 95% CI, 0.26-0.85; P = .01).
“We need to keep several limitations in mind. Foremost, the comparison between individuals choosing to participate in a randomized trial, making up a low single-digit percentage of individuals who received an invitation, and the observational cohort, may be biased by unmeasured confounding, despite adjusting for all comorbidities and baseline health care utilization,” Steinhubl said.
“Active screening for AF as part of a pragmatic direct-to-participant nationwide study was associated with significant improvement in clinical outcomes and safety at 3 years relative to routine care,” Steinhubl said during the presentation. “However, independent replication that these results are required in order to be confident that aggressive pursuit of diagnosing atrial fibrillation and people at high risk but without symptoms is warranted.”
- Steinhubl SR, et al. LBS.06: To screen or not to screen, and then what? Studies of detection and treatment of AF. Presented at: American Heart Association Scientific Sessions; Nov. 13-17, 2020 (virtual meeting).