Screening for silent cardiac disease in an elderly population aiming at stroke reduction

Abstract

INTRODUCTION: Atrial Fibrillation is the most common clinical arrhythmia and affects over 30 million people worldwide. Atrial fibrillation is a leading cause of morbidity and mortality worldwide with death, stroke and heart failure being the most feared complications. It is a well-known risk factor for stroke, a risk that can be reduced by about two-thirds in high-risk individuals with oral anticoagulant treatment. Atrial fibrillation patients can be asymptomatic, and many individuals-especially the elderly-are unaware that they have atrial fibrillation. Because atrial fibrillation is a common, chronic and often asymptomatic disease, with increased risk of serious but to certain extent preventable complications, it seems to meet most of the World Health Organization’s criteria for population screening. Screening for atrial fibrillation and initiation of oral anticoagulant therapy could potentially prevent stroke from atrial fibrillation in high-risk individuals, but the optimal screening program and strategies are not yet defined. Several large-scale, randomised studies are ongoing with the aim to answer the question if we should screen for atrial fibrillation, the STROKESTOP I and STROKESTOP II studies being among them.

AIMS: I. to study atrial fibrillation detection and predictors for new atrial fibrillation by using the biomarker NT-proBNP in a stepwise screening of a high-risk population and to study oral anticoagulant treatment uptake as well as one-year adherence to treatment in individuals diagnosed with new atrial fibrillation during screening. II. to study atrial fibrillation detection using pulse palpation and compare it to single-lead ECG and to study if symptoms of palpitations are associated with atrial fibrillation yield in screening. III. to analyse geographic and sociodemographic disparities in the uptake of the STROKESTOP II study and compare the results between STROKESTOP I and STROKESTOP II after the intervention of adding two screening sites in STROKESTOP II. IV. to study the potential yield of detected disease in non-participants and to compare characteristics of non-participants to participants.

METHODS: In study I all 75/76-year-olds in the Stockholm region (n=28,712) were randomized 1:1 to be invited to participate in the STROKESTOP II study-an atrial fibrillation screening study-or to serve as control group. Participants without previous atrial fibrillation had NT-proBNP measured and were stratified into low-risk group with NT-proBNP <125 ng/L or a high-risk group with NT-proBNP ≥125 ng/L. The highrisk group was offered 2 weeks intermittent ECG-screening whereas the low-risk group performed only one single-lead ECG recording at the screening visit. Participants diagnosed with atrial fibrillation were referred to a cardiologist for assessment and started on OAC unless contraindicated. In study II the participants from the STROKESTOP II study that did not have prior atrial fibrillation diagnosis were included. Healthcare professionals at the screening visit palpated their radial pulse for 30 seconds and recorded it as regular or irregular. Thereafter a 30-second single-lead ECG was registered. Patients also answered the yes/no question if they had felt palpitations. In study III data from the intention-to-screen group in the STROKESTOP II study were included and the same data variables from the STROKESTOP I participants was available. In the STROKESTOP II study, two additional screening sites were used compared to in the STORKESTOP I study, closer to low-income neigbourhoods. Invitee’s residential parish was used for geo-mapping analysis of the geographical disparities in participation, using hierarchical Bayes methods. Individual data for participants and non-participants were obtained from Statistics Sweden with respect to following socioeconomic variables: educational level, disposable income, immigrant and marital status. In study IV anonymised individual data for non-participants and participants were obtained from Swedish registries regarding socioeconomic factors, medical history and dispensed medicines. A random forest was trained to predict propensity scores for participation. The propensity scores were used to infer on potential screening yield of disease among non-participants.

RESULTS: In total, 6,868 individuals accepted the screening invitation to participate in the STROKESTOP II study. Participants without prior atrial fibrillation diagnosis were 6,315 (91.9%). New atrial fibrillation was diagnosed in 4.4% (95% CI 3.7-5.1) of the participants in the high-risk group. In all participants without prior atrial fibrillation new atrial fibrillation was detected in 2.6% (95% CI 2.2-3.0). The screening procedure resulted in an increase in known prevalence from 8.1% to 10.5% among all participants. Oral anticoagulant treatment was initiated in 94.5% of the participants with new atrial fibrillation diagnosis and at one-year follow-up 96% were still adherent to the treatment. Of the 6,159 participants included in the study II, 461 (7.5%) had irregular pulse. Twenty-two (4.8%) of those with irregular pulse were diagnosed with atrial fibrillation on single-lead ECG rhythm strip. Among those with regular pulse, 6 (0.1%) cases of new atrial fibrillation were found. The sensitivity of the pulse palpation test was 78.6% and positive predictive value 4.8%. The proportion of newly diagnosed atrial fibrillation was not different between those with and without history of palpitations. Higher participation was observed in those with higher education, high income, among non-immigrants and married individuals in study III. Participation between STROKESTOP I and STROKESTOP II improved significantly where additional screening sites were introduced. These improvements were generally significant, in each population group according to socio-demographic characteristics. In study IV the atrial fibrillation prevalence was 17.9% among non-participants compared to 8.9% in non-participants. Non-participants were in poorer health, had lower socioeconomic status and were less mobile than participants. The most important factors to predict non-attendance were low income, hospitalisations and higher CHA2DS2-VASc scores. The weighted estimates suggested that the estimated yield of untreated AF was 3.6% in non-participants compared to 3.3% in the participants.

CONCLUSIONS: In an NT-proBNP-stratified systematic screening for atrial fibrillation, 4.4% of the high-risk participants were diagnosed with new atrial fibrillation. Oral anticoagulant treatment initiation was well accepted in the group diagnosed with new atrial fibrillation. Pulse palpation was inferior to single-lead ECG when screening for atrial fibrillation and we would advocate the use of single-lead ECG rather than pulse palpation when screening for atrial fibrillation. Palpitations did not predict new diagnosis of atrial fibrillation. Decentralisation of screening sites in an atrial fibrillation-screening-program yielded a significantly positive impact on screening uptake. Addition of local screening sites in areas with low uptake had beneficial impact on participation across a wide spectrum of socio-demographic groups. Importantly, decentralised screening increased substantially the screening uptake in deprived areas. The potential yield of untreated atrial fibrillation detected by screening was estimated higher among non-participants than among participants. The non-participants had higher CHA2DS2-VASc scores, underlining their high stroke-risk and probable benefit from attending screening.