Screening for atrial fibrillation in high-risk individuals : a stroke preventive effort

Abstract

INTRODUCTION: A condition can be suitable for population screening if the disease is an important health problem, can be detected by an appropriate test and has a silent/latent stage during which treatment could be initiated in order to prevent consequences of the disease. Atrial fibrillation (AF) increases the risk of ischemic stroke 5-fold, a risk that can be reduced by at least 64% by the use of stroke-preventive, oral anticoagulant (OAC) therapy, and there are several non-invasive methods that can be used to diagnose AF. Hence it could be argued that AF is suitable for population screening. The aim of this thesis is to study the feasibility of systematic screening for AF in a high risk population and the possibility to initiate stroke-preventive anticoagulant therapy. In addition, this thesis also aims to compare different biomarkers and their association with AF, and to study if the biomarker NT-proBNP is increased in screening-detected AF. Finally, different methods for the detection of AF in a population that has already suffered an ischemic stroke or transient ischemic attack will be explored.

METHODS AND RESULTS: In study I all residents born in 1936/37, n=28,768, and who resided in two Swedish regions, were randomized 1:1 either to a control group or to attend a screening study. Of the 13,331 individuals who were invited, 7,173 (54%) participated. New AF was detected in 218 participants (3.0 %, 95% confidence interval (CI) 2.7-3.5) using twice daily intermittent ECG recordings for 14 days. In addition, 149 participants were known to have AF, but were not treated with oral anticoagulant therapy. Of the screened population, 5.1% (95 % CI 4.6-5.7) had untreated AF. Initiation of OAC treatment was made in 93% of participants with newly detected AF and in 3.7% (95% CI 3.3-4.2) of the screened population.

Study II collected five different biomarkers: i) N-terminal pro-B-type Natriuretic Peptide (NT-proBNP) indicative of atrial stretch, ii) high-sensitive (hs) Troponin (Tn) as a marker of myocyte damage, iii) Growth differentiation factor -15 (GDF-15), which is a marker of oxidative stress, iv) the renal biomarker cystatin C and v) the inflammatory marker hs C-reactive protein (CRP), in two Swedish cohorts, the Uppsala Longitudinal Study of Adult Men (ULSAM), n=883 and the Prospective Investigation of the Vasculature in Uppsala Seniors (PIVUS), n=978. Risk factors for cardiovascular disease were assessed at the baseline visit. After a median follow-up of 12.6 years in ULSAM and 10 years in PIVUS, 113 (12.8%) patients developed AF in ULSAM and 148 (15.1%) in PIVUS. Unadjusted Cox regression analysis showed significant association for all biomarkers with regards to the development of AF; however, in analysis adjusted for cardiovascular risk factors and the other biomarkers, only NT-proBNP remained significantly, p < 0.001, associated with incident AF with a hazard ratio of 2.05 (1.62-2.59) in ULSAM and 1.56 (1.30-1.86) in PIVUS per 1 standard deviation increase. NT-proBNP significantly increased risk prediction, both with regards to risk factors for cardiovascular disease and the established CHARGE-AF risk score.

In study III the last 815 consecutive participants from study I had a sample of NT-proBNP analysed bedside, and 71 individuals with newly detected AF had NT-proBNP taken at cardiology follow-up. Participants with newly detected AF, n=96, had significantly higher NT-proBNP (median 330 ng/L, interquartile range (IQR)121;634), compared to individuals in whom AF was not detected, n=742, median NT-proBNP 171 ng/L (IQR 95;283), p < 0.001. NT-proBNP remained significantly associated with screening-detected AF after multivariable logistic regression, p < 0.001. A cut-off of NT-proBNP 125 ng/L was shown to have a sensitivity of 75% and a negative predictive value of 92% for screening-detected AF.

In study IV, 41 elderly patients (mean 76.3 +/- 5.4 years) with a prior ischemic stroke/transient ischemic attack (TIA) and no prior diagnosis of AF were included and three methods of AF detection: 24-h Holter, 30-days intermittent ECG recording, and an implantable loop recorder (ILR), were initiated in parallel after a median of 6.8 +/- 4.3 days. One patient was excluded due to a brain tumour. AF was detected in 14/40 patients, intermittent ECG detected AF in 1 patient, whereas ILR fared significantly better, p < 0.001, and detected all 14 cases. Average time to AF detection was 14.7 +/- 11.6 months.

CONCLUSIONS: Systematic AF screening detected a significant proportion of screened individuals with untreated atrial fibrillation. The degree of initiation of stroke-preventive therapy was high, particularly in individuals with newly detected AF. The biomarker NT-proBNP was shown in two cohort studies to be the only biomarker out of five that was significantly associated with incipient AF after adjustment for other biomarkers and clinical risk factors. Individuals with screening-detected atrial fibrillation had higher levels of NT-proBNP. In elderly patients with an ischemic stroke/TIA, parallel investigation of three methods for AF detection revealed that long-term monitoring using an implantable cardiac monitor detected significantly more AF compared to shorter time-span methods.