International Journal of Cardiology 178 (2015) 247–252
Contents lists available at ScienceDirect
International Journal of Cardiology journal homepage: www.elsevier.com/locate/ijcard
Screening for atrial fibrillation in general practice: A national, cross-sectional study of an innovative technology☆ Gerard Bury a,⁎, Davina Swan a, Walter Cullen b, David Keane a, Helen Tobin a, Mairead Egan a, David Fitzmaurice c, Crea Carberry a, Cecily Kelleher a a b c
University College Dublin, Ireland University of Limerick, Ireland University of Birmingham, UK
a r t i c l e
i n f o
Article history: Received 3 December 2013 Received in revised form 17 September 2014 Accepted 18 October 2014 Available online 9 November 2014 Keywords: Atrial fibrillation General practice Screening Three-lead ECG
a b s t r a c t Background: To test the use of three lead monitoring as a screening tool for atrial fibrillation (AF) in general practice. AF is responsible for up to a quarter of all strokes and is often asymptomatic until a stroke occurs. Methods: 26 randomly selected general practices identified 80 randomly selected patients aged 70 or older from their database and excluded those known to have AF, those with clinical issues or who had not attended for three years. Up to 40 eligible patients/practice were invited to attend for screening. A 2 min three-lead ECG was recorded and collected centrally for expert cardiology assessment. Risk factor data was gathered. Outcomes: (i) point prevalence of AF, (ii) proportion of ECG tracings which were adequate for interpretation, (iii) uptake rate by patients and (iv) acceptability of the screening process to patients and staff (reported separately). Results: Of 1447 current patients, 1003 were eligible for inclusion, 639 (64%) agreed to take part in screening and 566 (56%) completed screening. The point prevalence rate for AF was 10.3%–2.1% new cases (12 of 566 who were screened) and 9.5% existing cases (137 of 1447 eligible patients). Only four of 570 (0.7%) screening visits did not record a usable ECG and 11 (2.6%) three lead ECGs required a clarifying 12 lead ECG. Conclusions: Three lead screening for AF is feasible, effective and offers an alternative to pulse taking or 12 lead ECGs. The availability of this technology may facilitate more effective screening, leading to reduced stroke incidence. © 2014 Elsevier Ireland Ltd. All rights reserved.
1. Introduction Atrial fibrillation (AF) is a common arrhythmia in elderly patients. Its prevalence doubles every decade after 50 years of age, from 0.5% at age 50–59 years to almost 9% of people aged 80–89 years [1]. AF creates a five-fold risk of ischaemic stroke [2]. The stroke risk associated with AF also increases steeply with advancing age and with the presence of common co-morbidities such as hypertension and diabetes. AF is responsible for 25% of ischaemic strokes and AF-associated strokes are more likely to be recurrent, severe and fatal than strokes due to other causes [3,4]. Anticoagulation of AF patients can reduce stroke risk by approximately 60% [5]. However, AF is often asymptomatic and may be paroxysmal, making its detection challenging. AF is usually detected opportunistically; one recent study has reported a rate of newly detected AF of 30% among 1368 patients with risk factors for stroke, not known to have AF, who had continuous cardiac monitoring for extended periods [6]. A further continuous monitoring study of patients who had suffered a stroke or ☆ This research was supported by a competitive grant received from Ireland's national medical research agency, the Health Research Board (HRA / 2009 / 239). ⁎ Corresponding author.
http://dx.doi.org/10.1016/j.ijcard.2014.10.037 0167-5273/© 2014 Elsevier Ireland Ltd. All rights reserved.
transient ischaemic attack (TIA) showed that 95% of AF detected was paroxysmal and 86% of episodes lasted for less than 1 h [7]. Given the significant stroke related burden of ‘silent AF’, the need for introduction of new methods of monitoring has been highlighted [8]. As undiagnosed and perhaps asymptomatic, elderly patients often visit their GP for common co-morbidities; GPs have a unique opportunity to actively seek patients with AF [9]. Strategies to detect such cases include a range of opportunistic and systematic screening models and various clinical techniques ranging from simple pulse checks to 12 lead ECG with expert interpretation. The most effective method of identifying unsuspected AF remains unclear [10]. The 2007 SAFE study showed that structured AF screening of elderly patients in general practice improved detection compared to routine care [11]. SAFE further compared opportunistic pulse taking with systematic 12-lead ECG screening and found both screening strategies equally effective at detecting AF. SAFE concluded that opportunistic pulse taking remains the best available screening technique, mainly because of the practical difficulties in carrying out 12 lead ECGs in general practice. However, a pulse check used to screen for AF has a low specificity (70–81%), and for every five 12-lead ECGs subsequently
248
G. Bury et al. / International Journal of Cardiology 178 (2015) 247–252
conducted to confirm diagnosis, only one case will have AF [12,13]. By contrast, ECGs using less than 12 leads, e.g. bipolar ECGs, have been found to have a high sensitivity and high specificity for AF screening, when interpreted by someone with appropriate expertise [13]. GPs' expertise in interpreting ECGs can vary greatly, however [14]. Novel technologies are under examination to address the challenges of screening for AF in the community, including the use of iPhone ECG devices by pharmacists and general practice nurses and receptionists in studies which demonstrate potential benefits to these approaches [15–17]. In this study we explore the use of a novel technology in a screening initiative in a general practice setting, using three-lead monitoring. Three-lead ECG monitoring is available as an adjunct to many Advisory External Defibrillators (AEDs) including the model introduced to Irish general practices through the MERIT project (Medical Emergency Responders: Integration and Training) [18]. Since 2005 MERIT has been equipping and training Irish GPs to manage cardiac arrest in the community. Ireland has a well developed general practice system, with approximately 2600 GPs working at around 2000 sites around the country. Approximately a quarter of Irish general practices are members of the MERIT scheme, which provides a standard Laerdal/Phillips FR2+ AED to each of the 497 participating practices. The FR2+ AED has a 3lead ECG monitor as an integral component. The monitor generates a lead II rhythm strip, visible on the FR2+ screen during ECG recording; interpretative software is not included. The ECG tracing may subsequently be electronically downloaded from a data card in the FR2+ for more detailed review. This technology has never been systematically used for the purpose of AF screening. While 12-lead ECGs are time-consuming (10–15 min) and require some privacy to conduct, the 3-lead ECG is quick and easy to use and does not require the patient to undress. It may thus be more acceptable to GPs and patients than 12-lead ECG screening. Our study aims to test the feasibility and acceptability of this technology to screen older patients for AF in general practice. 2. Methods This was a multi-centred study, among computerised general practices in Ireland, of the feasibility and acceptability of FR2+ units to screen for atrial fibrillation among approximately 1000 patients aged 70 years or older, who were not known to have had atrial fibrillation at any time. Patients aged 70 years and over were selected because of higher risk and because all in this age group became entitled to free primary care services in Ireland in 2004. Although eligibility changed slightly in 2010, the vast majority of people aged 70 or older are registered with general practices for free primary care. Excellent and easily accessible databases therefore exist on this population in computerised general practices. Each of the 467 MERIT practices was sent an information leaflet in Jan 2011 and an invitation to participate, if computerised. An expression of interest form was returned by 166 eligible practices. We randomly selected 25 practices for participation and an additional 15 reserve practices to replace any practice which declined to participate or withdrew from the study. Within each practice, we obtained a list of all patients aged 70 years and over. From this we identified a random sample of 80 patients for the practice to review for study eligibility until they attained a quota of 40 eligible patients. Several practices had large numbers of deceased and past patients on their register and needed longer lists of randomly selected patients to identify 40 eligible patients. Two practices had less than 80 patients within this age group and so had all of these patients selected. Random selection of patients for each practice was achieved with computer generated random numbers (www. randomizer.org) to access entries in lists of patients. Each practice reviewed their list of randomly selected patients and the patients' medical charts, sequentially until they identified 40 eligible patients. Patients were eligible if they had attended at least once in the past three years. Patients were not eligible if they ever had atrial fibrillation or atrial flutter or if they had a terminal illness or cognitive impairment which might impact on informed consent. Practices could also exclude patients where they considered participation inadvisable because of other significant clinical problems. Once practices had completed the selection process, they reported to the research team the number of patients excluded and the reason for each exclusion. Three practices identified less than 40 eligible patients for the study. 2.1. Intervention Training of practice staff covered recording an ECG using the FR2+ ECG monitoring leads and uploading ECG recordings to a PC using a card reader with Review Express Connect (REC) software and Cutepdf writer.
A pack of materials was provided to practices, including: Ambu electrodes (R-00-S/25), a Lindy card-reader (Lindy USB 2.0 Multi-Card Reader Pro/Lindy USB Card Reader 6-in-1/Lindy USB 2.0 Card Reader 23-in-1), REC software, a study manual, stationery and Irish Heart Foundation patient information booklets on AF. The researchers kept in contact with practices for the duration of the study to provide technical and other support. Practices wrote to selected patients inviting them to participate in 3-lead ECG screening for atrial fibrillation. A study information leaflet was included with a consent form to be completed and returned in a pre-paid envelope to the practice; patients were asked to consent to anonymised use of medical record data and to take part in a three-lead screening exercise by their general practitioner. Non-respondents received a written or a telephone reminder from the practice. Practices could screen patients when they attended the surgery for other purposes and/or could arrange appointments with patients specifically for screening. Each practice organised the various aspects of the screening process (i.e., recording, uploading and analysing ECGs) among their staff. Patients had their ECG recorded for 2 min. For convenience the monitoring leads could be placed in standard positions on the patient's trunk or limbs. If placed on the trunk, an electrode was positioned below the left and the right clavicles, and on the left hip. If placed on the limbs, an electrode was positioned on the left and the right wrists, or backs of each hand, and on the left ankle. The ECG tracing on the data card in the FR2+ was uploaded to a PC. ECGs were emailed, or occasionally printed and posted, to the research team and analysed by the study cardiologist for the presence of AF and other arrhythmias. Results were posted to GPs. We rang practices to confirm receipt of all abnormal results. Practices returned ECG results to patients. Participating GPs were responsible for patients' subsequent clinical care. 2.2. Main outcome measures Our main outcome measures were: (i) the point prevalence rate of atrial fibrillation, (ii) the proportion of ECG tracings which were adequate for interpretation, (iii) the uptake rate by patients, and (iv) the acceptability of the screening process to patients and staff. We recorded the following baseline data for each patient from their medical record: demography, lifestyle, known risk factors for and complications of AF, other health conditions, prescribed medicines and therapeutic interventions. A GP or practice nurse within each practice assisted researchers with data extraction. At patients' screening visit, the GP or practice nurse recorded the patient's selfreported current risk factors for AF, health status and current smoking and alcohol use. GPs took responsibility for immediate analysis of patients' ECGs and completed a form noting if AF and/or other arrhythmias were present. Forms were returned to the research team and were not seen by the study cardiologist. Acceptability surveys were carried out among staff and patients and are reported separately. We analysed the point prevalence rate of AF on an intention to treat basis. For continuous variables, the median and interquartile range were reported, and Mann and Whitney U tests were used. Two-tailed tests were applied. For categorical variables, chi square tests were used. A probability value of b0.05 was regarded as significant. SPSS (PASW) version 20 was used to analyse the data. The study was approved by the Research Ethics Committee of the Irish College of General Practitioners and funded by the Health Research Board, Ireland. The authors have no competing interests.
3. Results 3.1. Participant flow 3.1.1. Practices Of 25 practices initially invited to take part, seven did not complete the study. One practice immediately declined to participate, two withdrew after the training session, and four proved unresponsive prior to patient recruitment. Eight additional reserve practices were recruited; of 33 randomly selected practices invited to participate, 26 (79%) completed the study. Almost half (n = 12; 46%) of the 26 practices which completed the study served an urban patient population, one-fifth (n = 5; 19%) served a rural patient population, and one-third (n = 9; 35%) categorised themselves as serving a mixed population. One-fifth (n = 5; 19%) of practices consisted of single-handed GPs. Just one practice did not have a practice nurse. 3.1.2. Patients Fig. 1 describes the flow of participants in the study. In all, 2200 patients aged 70 years and over were randomly selected from the 26 practice registers and reviewed by practices for study eligibility. Practices excluded 754 patients who were not current patients and an additional
G. Bury et al. / International Journal of Cardiology 178 (2015) 247–252
443 patients for known AF/atrial flutter, terminal illness, cognitive impairment or unsuitability for screening leaving 1003 patients eligible for inclusion. Exclusion data is missing for two practices and is less than 80 missing cases. 1003 current patients were invited for 3-lead ECG screening and had baseline chart reviews carried out. Almost two-thirds (n = 639; 64%) of patients accepted the invitation for screening. There was a wide range of uptake across practices (median 64%; range 33–96%; IQR 20%). Of the 639, 25 patients proven to have a documented history of AF (20) or exclusion criterion (2) on chart review or developed new AF (3), leaving 614 eligible, consenting patients of whom 566 (56%) completed screening. 3.2. Patient characteristics Table 1 gives details of the age and sex of patients. Patients invited for screening had a median age of 76 years (range 70–98) and were significantly younger than current patients excluded by practices for not meeting the study criteria or other unsuitability for screening (U = 282,923; p b 0.001). Following the invitation to be screened,
younger patients were more likely to accept the screening offer than older patients (U = 116,697; p b 0.001). Men were more likely than women to accept screening (χ2 = 10.751; df = 1; p b 0.001). Among the 614 eligible patients who completed screening, 88.3% (542) were entitled at the time to free primary care (General Medical Services eligible). Table 2 outlines the prevalence of conditions, known to be associated with atrial fibrillation, among the 614 patients who completed screening. The most common associated conditions were hypertension (n = 296, 48.2%), ischaemic heart disease (n = 78, 12.7%), diabetes (n = 65, 10.6%) and obesity (n = 49, 8.0%). Thirty patients (4.9%) had a history of stroke or transient ischaemic attack (TIA); 33 (5.4%) had a previous myocardial infarction (MI) and seven patients (1.1%) had heart failure. 3.3. Prevalence of atrial fibrillation 3.3.1. Baseline AF prevalence Of the 1447 current patients randomly selected from practice registers, 111 cases of previously identified AF and atrial flutter (7.7%) were
2200 patients ≥ 70 years of age randomly selected and reviewed for eligibility (26 practices)* 754excluded as not a current patient:
443 excluded: AF/AFL** n=111 Cognitive Impairment n=73 Terminal Illness Other exclusion
1447 current patients
n=21 n=238
1003 patients invited for screening
3 known AF cases
364 declined (36%)
639 consented (64%)
3 AF cases newly diagnosed prior to screening
249
20 known AF cases identified during baseline data collection from patients’ charts
2 patients invited in error
614 eligible patients (61%)
566 had 3-lead ECG (56%)
12 new cases of AF (10) and AFL (2) identified on 3-lead ECG (1.2%)
* Exclusion data is missing for 2 practices: missing data is ≤ 80 cases **AFL atrial flutter Fig. 1. Flow of patients through the study. * Exclusion data is missing for 2 practices: missing data is ≤80 cases **AFL atrial flutter.
250
G. Bury et al. / International Journal of Cardiology 178 (2015) 247–252
Table 1 Patient demography. Group
Total
Data available for
Male (%)
Mean age
Median age
Age range
Excluded patientsa Invited patients Consented Declined Screened New AF cases
443 1003 639 362 566 12
329 (74%) 954 (95%) 639 (100%) 315 (87%) 566 (100%) 12 (100%)
122 (37) 352 (37) 257 (40) 93 (30) 229 (40) 11 (92)
83 77 78 78 78 79
83 76 76 77 77 81
71–101 70–98 70–95 70–98 70–94 73–88
a
Current patients excluded by practices during patient selection.
identified by practices during review for study eligibility. A further 20 previously identified AF cases were found by the research team when collecting baseline data from participating patients' charts. The baseline prevalence of previously identified AF among the sample of current patients was thus 9.3% (n = 134) (see Table 3).
3.3.2. Newly identified AF Five hundred and seventy eligible patients (57%) attended their GP surgery for AF screening. In total, 566 patients had 3-lead ECGs recorded and analysed by the cardiologist. A 3-lead ECG could not be registered for one patient despite two separate attempts and a 12-lead ECG was recorded instead. Three patients had 3-lead ECGs recorded which failed to save to the data card. Twelve new cases of AF (n = 10) and atrial flutter (n = 2) were identified on the 3-lead ECG by the cardiologist, yielding a 2.1% point prevalence among the screened group (1.2% based on ITT). The median CHA2DS2-VASc score among the 12 new AF cases was 4 (range 1–7). Nine of the 12 cases had AF/atrial flutter confirmed by subsequent testing. Six of the nine patients were anti-coagulated with warfarin, two were prescribed aspirin and one dabigatran; and one patient was already on clopidogrel and aspirin due to PCI within the past year. The 12 new AF cases identified on 3-lead ECG tended to be older than the non-AF cases in the screened group, but this was not statistically significant (U = 4237; p = 0.074). The 12 cases were, however, significantly more likely to be male than the non-AF cases screened (χ2 = 11.162; df = 1; p b 0.001). Of 570 patients who attended for screening, four (0.7%) could not be completed because of technical issues with the FR2+. Of 566 tests carried out, all were of sufficient quality to allow analysis. In all cases in which AF was identified, the GP was prompted to confirm with a 12 lead ECG; in 11 (1.9%) other cases a 12 lead ECG was requested to facilitate interpretation — in four cases for low amplitude readings, in four
Table 2 Prevalence of risk factors for AF among patients who completed screening (n = 614). Conditions
Hypertension Ischaemic heart disease/coronary artery disease Diabetes Obesity Chronic obstructive pulmonary disease Other heart surgery & cardiac procedures Pneumonia Stroke Hyperthyroidism Transient ischaemic attack Echocardiographic abnormalities Heart failure Pulmonary embolism Cardiomyopathy Sleep apnoea Lung cancer Lung surgery
Patients n
%
296 138 65 49 29 19 17 16 15 14 8 7 5 5 5 2 1
48.2 22.5 10.6 8.0 4.7 3.1 2.8 2.6 2.4 2.3 1.3 1.1 0.8 0.8 0.8 0.3 0.2
cases because of baseline artefacts and in three cases because of tachycardias. The total prevalence of AF among the sample of current patients was therefore 149/1447 (10.3%). 4. Discussion The key outcome measures of the study include: (i) The point prevalence rate of atrial fibrillation among screened patients was 2.1%. Among the 1447 patients initially identified as possible screening subjects, the total prevalence of AF was 10.3% (new and existing cases). (ii) Four of 570 (0.7%) patients who attended for screening had problems which prevented them from being screened on the day in question and in 11 (1.9%) cases, a 12 lead ECG was advised for clarification of the rhythm. (iii) The screening uptake rate by patients was 56.6% of the 1003 patients invited; 64% consented to take part. (iv) The acceptability of the screening process to patients and staff is reported separately but is very high. This study has identified a study sample which is likely to be highly representative of patients aged 70 and older currently attending Irish general practices. Patient selection used randomised samples from computerised records and then narrowed this to current patients (1447) and those who met inclusion criteria (1003). Participating GPs excluded 238 patients on the basis of their knowledge of the patients' clinical, social or personal circumstances (other than AF, terminal illness or significant cognitive impairment) and these exclusions may be a source of some uncertainty about uptake of screening offers among the entire older population. In all, 64% of eligible patients responded positively to a written invitation to be screened and 56% completed the process. While this is a significant response to a completely unfamiliar screening process, further research is needed to identify the barriers and facilitators to participation in AF screening. This is particularly relevant if repeat screening is to be offered; the current study provides a point prevalence rate but can offer no insight into incidence rates or the most appropriate frequency of screening. Three-lead ECG screening using FR2 + defibrillator/monitors yielded a 2.1% point prevalence rate of newly identified AF among patients over 70 years of age who underwent screening. This detection rate compares very favourably with the detection rates in similar populations using 12-lead ECGs. Among patients over 65 years of age who underwent systematic 12-lead ECG screening in the SAFE study, 2.2% (52/2357) had newly identified AF on ECG [11]. A similar study in Sweden of 75–76 year olds identified 1.3% new AF cases among patients with no known AF (10/767) who underwent a 12-lead ECG [10]. A key finding of this study was the high rate of known or previously documented AF in the general practices under study. Further research is underway to document the risk profiles and treatment strategies being used for these patients. The finding not only highlights the very significant issues raised by AF in Irish general practice, in terms of detection and treatment but also demonstrates the potential impact of effective use of existing records to monitor at-risk groups. Overall 2.6% of rhythm strips reviewed by the cardiologist required a 12-lead ECG for rhythm clarification on the basis that the rhythm strip suggested issues best explored with a 12 lead ECG — none of these abnormalities was suspected to be AF and a detailed breakdown of the rhythm abnormalities identified in this population will be published separately. This suggests that this technology may indeed be an effective and efficient screening method for AF. Although just three-quarters of the AF cases identified on the 3-lead ECG were confirmed by subsequent testing, this is likely to be due to paroxysmal arrhythmia in some cases. Almost 57% of our invited patients attended for AF screening, which is similar to the attendance rate for ECG screening in the SAFE study
G. Bury et al. / International Journal of Cardiology 178 (2015) 247–252
251
Table 3 AF prevalence rates. Denominator
AF cases
%
Total prevalence Baseline prevalence (known AF from exclusion report) Baseline prevalence (all known AF cases)
1447 (current patients) 1447 (current patients) 1447 (current patients)
10.3 7.7 9.3
Point prevalence (newly identified AF)
566 (screened patients)
149 (all known and new AF cases) 111 (exclusion report from practices) 134 (all known AF cases, including those identified through data collection from charts and reported by GP/PN) 12 (new AF cases)
(53%), which was also conducted within a primary care setting [11]. The Swedish study conducted 12-lead ECG screening in an outpatient setting and had 64% of invited patients attend [10]. This higher uptake may be due to cultural differences or differences in patients' perceptions of secondary versus primary care, with the specialist nature of the former possibly being more highly valued. We found a large range of uptake across our 26 practices, suggesting that general practice setting and/or GP–patient relations may be important variables in screening uptake. We found that older patients and women were less likely to take up the offer of screening. This is of concern given that older and female patients are at higher risk of stroke if they have AF. The Swedish study found that those who declined screening had more cardiovascular risk factors than those who accepted [10] Research also suggests that patients have little understanding of AF and its risks [19,20]. Public awareness and other initiatives are thus needed to educate and encourage the most high risk groups to engage with screening. Our study shows that the FR2+ units used in this study deliver high quality rhythm strips, suitable for clinical analysis. The technology is suitable for use in either structured screening exercises such as this one, or in case-finding exercises, based on risk factors or the finding of an irregular pulse. A further study is currently underway to examine the preferences for these strategies among a broader spectrum of general practices. The units used in this exercise were FR2+ defibrillators which have an in-built 3-lead monitoring capacity — however their primary and core role is as defibrillators in emergencies. A free-standing 3-lead monitoring unit would remove any possible impingement on the availability of the FR2+ and might also contain analytical software for rhythm analysis. We currently have no information about the extent to which other general practices in Ireland or other countries use AEDs equipped with three-lead monitoring capability but a study to explore this is now underway in Ireland. 4.1. Study limitations This study consisted of a single ECG screening of patients. It therefore may have missed cases of paroxysmal AF. The Swedish screening study identified 7% new AF cases among patients undergoing extended intermittent ECG screening using a hand-held ECG recorder [10]. Although the screened group were at particular risk of stroke (CHADS2 score ≥ 2), it underlines the challenge involved in detecting paroxysmal AF. Patients may have been reluctant to take part in a research study as opposed to an established screening programme. Practices may also have been more or less encouraging of their patients in availing of this screening opportunity. Practices excluded a substantial number of patients for discretionary reasons, usually because patients were immobile or in a nursing home and attending the surgery would be difficult. This screening modality is tailored to a well-structured general practice system with high quality registers of at-risk individuals (in this case populations of those aged 70 and older) and the availability of AEDs with three-lead monitoring. It also used a centralised specialist diagnostic service to confirm the presence of AF. While these elements may be available or potentially available in the wider Irish healthcare system or in other similarly organised healthcare systems, they may not be common components of primary care systems elsewhere. The use of such
2.1
an approach to population screening will therefore depend on the availability of similar resources or on access to other novel technologies including traditional telemedicine approaches. It remains unclear how often patients should be screened for optimum stroke prevention. 5. Conclusion Three-lead ECG screening using FR2+ defibrillator/monitors yields a comparatively high detection rate for newly identified AF among patients aged 70 years of age or older in general practice. This technology offers a novel screening tool with the potential to systematically improve the detection of AF in primary care. Conflicts of interest All authors take responsibility for all aspects of the reliability and freedom from bias of the data presented and their discussed interpretation. The authors have no conflicts of interest. Acknowledgements The authors wish to thank all the general practitioners, practice nurses and other members of practice staff who contributed their time and expertise to this study. References [1] W.B. Kannel, P.A. Wolf, E.J. Benjamin, D. Levy, Prevalence, incidence, prognosis, and predisposing conditions for atrial fibrillation: population-based estimates, Am. J. Cardiol. 82 (8A) (1998) 2N–9N. [2] P.A. Wolf, R.D. Abbott, W.B. Kannel, Atrial fibrillation as an independent risk factor for stroke: the Framingham study, Stroke 22 (1991) 983–988. [3] C. Marini, F. De Santis, S. Sacco, T. Russo, L. Olivieri, R. Totaro, A. Carolei, Contribution of atrial fibrillation to incidence and outcome of ischemic stroke: results from a population-based study, Stroke 36 (2005) 1115–1119. [4] N. Hannon, O. Sheehan, L. Kelly, et al., Stroke associated with atrial fibrillation: incidence and early outcomes in the north Dublin population stroke study, Cerebrovasc. Dis. 29 (2010) 43–49. [5] R.G. Hart, L.A. Pearce, M.I. Aguilar, Meta-analysis: antithrombotic therapy to prevent stroke in patients who have nonvalvular atrial fibrillation, Ann. Intern. Med. 146 (12) (2007) 857–867. [6] P.D. Ziegler, T.V. Glotzer, D.G. Daoud, D.E. Singer, et al., Detection of previously undiagnosed AF in patients with stroke risk factors and usefulness of continuous monitoring in primary stroke prevention, Am. J. Cardiol. 110 (2012) 1309–1314. [7] L.A. Sposato, F.R. Klein, A. Jauregui, M. Ferrua, et al., Newly diagnosed AF after acute ischemic stroke and TIA: importance of immediate and prolonged continuous cardiac monitoring, J. Stroke Cerebrovasc. Dis. 21 (2012) 210–216. [8] T.V. Glotzer, P.D. Ziegler, Silent AF as a stroke risk factor and anticoagulant indication, Can. J. Cardiol. 29 (2013) S14–S23. [9] Atrial Fibrillation in Primary Care (AFIP): Bringing Atrial Fibrillation Practice Closer to Guidelines. A Tool for Primary Care Physicians. (2012) [10] J. Engdahl, L. Andersson, M. Mirskaya, M. Rosenqvist, Stepwise screening of atrial fibrillation in a 75-year-old population: implications for stroke prevention, Circulation (2013) 930–937. [11] D.A. Fitzmaurice, F.D.R. Hobbs, S. Jowett, et al., Screening versus routine practice in detection of atrial fibrillation in patients aged 65 or over: a cluster randomised controlled trial, BMJ 335 (7616) (2007) 383. [12] G. Cooke, J. Doust, S. Sanders, Is pulse palpation helpful in detecting atrial fibrillation? A systematic review, J. Fam. Pract. 55 (2) (2006). [13] K. Harris, D. Edwards, J. Mant, How can we best detect atrial fibrillation? J. R. Coll. Physicians. Edinb. 42 (S18) (2012) 5–22. [14] J. Mant, D.A. Fitzmaurice, F.D.R. Hobbs, et al., Accuracy of diagnosing atrial fibrillation on electrocardiogram by primary care practitioners and interpretative diagnostic
252
G. Bury et al. / International Journal of Cardiology 178 (2015) 247–252
software: analysis of data from screening for atrial fibrillation in the elderly (SAFE) trial, BMJ 335 (7616) (2007) 380. [15] N. Lowres, L. Neubeck, G. Salkeld, I. Krass, et al., Feasibility and cost-effectiveness of stroke prevention through community screening for atrial fibrillation using iPhone ECG in pharmacies. The SEARCH-AF study, Thromb. Haemost. 111 (6) (2014) 1167–1176. [16] J. Orchard, S.B. Freedman, N. Lowres, D. Peiris, L. Neubeck, iPhone ECG screening by practice nurses and receptionists for atrial fibrillation in general practice: the GPSEARCH qualitative pilot study, Aust. Fam. Physician 43 (5) (2014) 315–319. [17] J.K. Lau, N. Lowres, L. Neubeck, D.B. Brieger, R.W. Sy, C.D. Galloway, D.E. Albert, S.B. Freedman, iPhone ECG application for community screening to detect silent atrial fibrillation: a novel technology to prevent stroke, Int. J. Cardiol. 165 (1) (2013) 193–194.
[18] G. Bury, M. Headon, M. Dixon, M. Egan, Cardiac arrest in Irish general practice: an observational study from 426 general practices, Resuscitation 80 (2009) 1244–1247. [19] G.Y.H. Lip, S. Kamath, M. Jafri, A. Mohammed, D. Bareford, Ethnic differences in patient perceptions of atrial fibrillation and anticoagulation therapy: the West Birmingham atrial fibrillation project, Stroke 33 (2002) 238–242. [20] E. Aliot, G. Breithardt, J. Brugada, et al., An international survey of physician and patient understanding, perception and attitudes to atrial fibrillation and its contribution to cardiovascular disease morbidity and mortality, Europace 12 (2010) 626–633.
Contents lists available at ScienceDirect
International Journal of Cardiology journal homepage: www.elsevier.com/locate/ijcard
Screening for atrial fibrillation in general practice: A national, cross-sectional study of an innovative technology☆ Gerard Bury a,⁎, Davina Swan a, Walter Cullen b, David Keane a, Helen Tobin a, Mairead Egan a, David Fitzmaurice c, Crea Carberry a, Cecily Kelleher a a b c
University College Dublin, Ireland University of Limerick, Ireland University of Birmingham, UK
a r t i c l e
i n f o
Article history: Received 3 December 2013 Received in revised form 17 September 2014 Accepted 18 October 2014 Available online 9 November 2014 Keywords: Atrial fibrillation General practice Screening Three-lead ECG
a b s t r a c t Background: To test the use of three lead monitoring as a screening tool for atrial fibrillation (AF) in general practice. AF is responsible for up to a quarter of all strokes and is often asymptomatic until a stroke occurs. Methods: 26 randomly selected general practices identified 80 randomly selected patients aged 70 or older from their database and excluded those known to have AF, those with clinical issues or who had not attended for three years. Up to 40 eligible patients/practice were invited to attend for screening. A 2 min three-lead ECG was recorded and collected centrally for expert cardiology assessment. Risk factor data was gathered. Outcomes: (i) point prevalence of AF, (ii) proportion of ECG tracings which were adequate for interpretation, (iii) uptake rate by patients and (iv) acceptability of the screening process to patients and staff (reported separately). Results: Of 1447 current patients, 1003 were eligible for inclusion, 639 (64%) agreed to take part in screening and 566 (56%) completed screening. The point prevalence rate for AF was 10.3%–2.1% new cases (12 of 566 who were screened) and 9.5% existing cases (137 of 1447 eligible patients). Only four of 570 (0.7%) screening visits did not record a usable ECG and 11 (2.6%) three lead ECGs required a clarifying 12 lead ECG. Conclusions: Three lead screening for AF is feasible, effective and offers an alternative to pulse taking or 12 lead ECGs. The availability of this technology may facilitate more effective screening, leading to reduced stroke incidence. © 2014 Elsevier Ireland Ltd. All rights reserved.
1. Introduction Atrial fibrillation (AF) is a common arrhythmia in elderly patients. Its prevalence doubles every decade after 50 years of age, from 0.5% at age 50–59 years to almost 9% of people aged 80–89 years [1]. AF creates a five-fold risk of ischaemic stroke [2]. The stroke risk associated with AF also increases steeply with advancing age and with the presence of common co-morbidities such as hypertension and diabetes. AF is responsible for 25% of ischaemic strokes and AF-associated strokes are more likely to be recurrent, severe and fatal than strokes due to other causes [3,4]. Anticoagulation of AF patients can reduce stroke risk by approximately 60% [5]. However, AF is often asymptomatic and may be paroxysmal, making its detection challenging. AF is usually detected opportunistically; one recent study has reported a rate of newly detected AF of 30% among 1368 patients with risk factors for stroke, not known to have AF, who had continuous cardiac monitoring for extended periods [6]. A further continuous monitoring study of patients who had suffered a stroke or ☆ This research was supported by a competitive grant received from Ireland's national medical research agency, the Health Research Board (HRA / 2009 / 239). ⁎ Corresponding author.
http://dx.doi.org/10.1016/j.ijcard.2014.10.037 0167-5273/© 2014 Elsevier Ireland Ltd. All rights reserved.
transient ischaemic attack (TIA) showed that 95% of AF detected was paroxysmal and 86% of episodes lasted for less than 1 h [7]. Given the significant stroke related burden of ‘silent AF’, the need for introduction of new methods of monitoring has been highlighted [8]. As undiagnosed and perhaps asymptomatic, elderly patients often visit their GP for common co-morbidities; GPs have a unique opportunity to actively seek patients with AF [9]. Strategies to detect such cases include a range of opportunistic and systematic screening models and various clinical techniques ranging from simple pulse checks to 12 lead ECG with expert interpretation. The most effective method of identifying unsuspected AF remains unclear [10]. The 2007 SAFE study showed that structured AF screening of elderly patients in general practice improved detection compared to routine care [11]. SAFE further compared opportunistic pulse taking with systematic 12-lead ECG screening and found both screening strategies equally effective at detecting AF. SAFE concluded that opportunistic pulse taking remains the best available screening technique, mainly because of the practical difficulties in carrying out 12 lead ECGs in general practice. However, a pulse check used to screen for AF has a low specificity (70–81%), and for every five 12-lead ECGs subsequently
248
G. Bury et al. / International Journal of Cardiology 178 (2015) 247–252
conducted to confirm diagnosis, only one case will have AF [12,13]. By contrast, ECGs using less than 12 leads, e.g. bipolar ECGs, have been found to have a high sensitivity and high specificity for AF screening, when interpreted by someone with appropriate expertise [13]. GPs' expertise in interpreting ECGs can vary greatly, however [14]. Novel technologies are under examination to address the challenges of screening for AF in the community, including the use of iPhone ECG devices by pharmacists and general practice nurses and receptionists in studies which demonstrate potential benefits to these approaches [15–17]. In this study we explore the use of a novel technology in a screening initiative in a general practice setting, using three-lead monitoring. Three-lead ECG monitoring is available as an adjunct to many Advisory External Defibrillators (AEDs) including the model introduced to Irish general practices through the MERIT project (Medical Emergency Responders: Integration and Training) [18]. Since 2005 MERIT has been equipping and training Irish GPs to manage cardiac arrest in the community. Ireland has a well developed general practice system, with approximately 2600 GPs working at around 2000 sites around the country. Approximately a quarter of Irish general practices are members of the MERIT scheme, which provides a standard Laerdal/Phillips FR2+ AED to each of the 497 participating practices. The FR2+ AED has a 3lead ECG monitor as an integral component. The monitor generates a lead II rhythm strip, visible on the FR2+ screen during ECG recording; interpretative software is not included. The ECG tracing may subsequently be electronically downloaded from a data card in the FR2+ for more detailed review. This technology has never been systematically used for the purpose of AF screening. While 12-lead ECGs are time-consuming (10–15 min) and require some privacy to conduct, the 3-lead ECG is quick and easy to use and does not require the patient to undress. It may thus be more acceptable to GPs and patients than 12-lead ECG screening. Our study aims to test the feasibility and acceptability of this technology to screen older patients for AF in general practice. 2. Methods This was a multi-centred study, among computerised general practices in Ireland, of the feasibility and acceptability of FR2+ units to screen for atrial fibrillation among approximately 1000 patients aged 70 years or older, who were not known to have had atrial fibrillation at any time. Patients aged 70 years and over were selected because of higher risk and because all in this age group became entitled to free primary care services in Ireland in 2004. Although eligibility changed slightly in 2010, the vast majority of people aged 70 or older are registered with general practices for free primary care. Excellent and easily accessible databases therefore exist on this population in computerised general practices. Each of the 467 MERIT practices was sent an information leaflet in Jan 2011 and an invitation to participate, if computerised. An expression of interest form was returned by 166 eligible practices. We randomly selected 25 practices for participation and an additional 15 reserve practices to replace any practice which declined to participate or withdrew from the study. Within each practice, we obtained a list of all patients aged 70 years and over. From this we identified a random sample of 80 patients for the practice to review for study eligibility until they attained a quota of 40 eligible patients. Several practices had large numbers of deceased and past patients on their register and needed longer lists of randomly selected patients to identify 40 eligible patients. Two practices had less than 80 patients within this age group and so had all of these patients selected. Random selection of patients for each practice was achieved with computer generated random numbers (www. randomizer.org) to access entries in lists of patients. Each practice reviewed their list of randomly selected patients and the patients' medical charts, sequentially until they identified 40 eligible patients. Patients were eligible if they had attended at least once in the past three years. Patients were not eligible if they ever had atrial fibrillation or atrial flutter or if they had a terminal illness or cognitive impairment which might impact on informed consent. Practices could also exclude patients where they considered participation inadvisable because of other significant clinical problems. Once practices had completed the selection process, they reported to the research team the number of patients excluded and the reason for each exclusion. Three practices identified less than 40 eligible patients for the study. 2.1. Intervention Training of practice staff covered recording an ECG using the FR2+ ECG monitoring leads and uploading ECG recordings to a PC using a card reader with Review Express Connect (REC) software and Cutepdf writer.
A pack of materials was provided to practices, including: Ambu electrodes (R-00-S/25), a Lindy card-reader (Lindy USB 2.0 Multi-Card Reader Pro/Lindy USB Card Reader 6-in-1/Lindy USB 2.0 Card Reader 23-in-1), REC software, a study manual, stationery and Irish Heart Foundation patient information booklets on AF. The researchers kept in contact with practices for the duration of the study to provide technical and other support. Practices wrote to selected patients inviting them to participate in 3-lead ECG screening for atrial fibrillation. A study information leaflet was included with a consent form to be completed and returned in a pre-paid envelope to the practice; patients were asked to consent to anonymised use of medical record data and to take part in a three-lead screening exercise by their general practitioner. Non-respondents received a written or a telephone reminder from the practice. Practices could screen patients when they attended the surgery for other purposes and/or could arrange appointments with patients specifically for screening. Each practice organised the various aspects of the screening process (i.e., recording, uploading and analysing ECGs) among their staff. Patients had their ECG recorded for 2 min. For convenience the monitoring leads could be placed in standard positions on the patient's trunk or limbs. If placed on the trunk, an electrode was positioned below the left and the right clavicles, and on the left hip. If placed on the limbs, an electrode was positioned on the left and the right wrists, or backs of each hand, and on the left ankle. The ECG tracing on the data card in the FR2+ was uploaded to a PC. ECGs were emailed, or occasionally printed and posted, to the research team and analysed by the study cardiologist for the presence of AF and other arrhythmias. Results were posted to GPs. We rang practices to confirm receipt of all abnormal results. Practices returned ECG results to patients. Participating GPs were responsible for patients' subsequent clinical care. 2.2. Main outcome measures Our main outcome measures were: (i) the point prevalence rate of atrial fibrillation, (ii) the proportion of ECG tracings which were adequate for interpretation, (iii) the uptake rate by patients, and (iv) the acceptability of the screening process to patients and staff. We recorded the following baseline data for each patient from their medical record: demography, lifestyle, known risk factors for and complications of AF, other health conditions, prescribed medicines and therapeutic interventions. A GP or practice nurse within each practice assisted researchers with data extraction. At patients' screening visit, the GP or practice nurse recorded the patient's selfreported current risk factors for AF, health status and current smoking and alcohol use. GPs took responsibility for immediate analysis of patients' ECGs and completed a form noting if AF and/or other arrhythmias were present. Forms were returned to the research team and were not seen by the study cardiologist. Acceptability surveys were carried out among staff and patients and are reported separately. We analysed the point prevalence rate of AF on an intention to treat basis. For continuous variables, the median and interquartile range were reported, and Mann and Whitney U tests were used. Two-tailed tests were applied. For categorical variables, chi square tests were used. A probability value of b0.05 was regarded as significant. SPSS (PASW) version 20 was used to analyse the data. The study was approved by the Research Ethics Committee of the Irish College of General Practitioners and funded by the Health Research Board, Ireland. The authors have no competing interests.
3. Results 3.1. Participant flow 3.1.1. Practices Of 25 practices initially invited to take part, seven did not complete the study. One practice immediately declined to participate, two withdrew after the training session, and four proved unresponsive prior to patient recruitment. Eight additional reserve practices were recruited; of 33 randomly selected practices invited to participate, 26 (79%) completed the study. Almost half (n = 12; 46%) of the 26 practices which completed the study served an urban patient population, one-fifth (n = 5; 19%) served a rural patient population, and one-third (n = 9; 35%) categorised themselves as serving a mixed population. One-fifth (n = 5; 19%) of practices consisted of single-handed GPs. Just one practice did not have a practice nurse. 3.1.2. Patients Fig. 1 describes the flow of participants in the study. In all, 2200 patients aged 70 years and over were randomly selected from the 26 practice registers and reviewed by practices for study eligibility. Practices excluded 754 patients who were not current patients and an additional
G. Bury et al. / International Journal of Cardiology 178 (2015) 247–252
443 patients for known AF/atrial flutter, terminal illness, cognitive impairment or unsuitability for screening leaving 1003 patients eligible for inclusion. Exclusion data is missing for two practices and is less than 80 missing cases. 1003 current patients were invited for 3-lead ECG screening and had baseline chart reviews carried out. Almost two-thirds (n = 639; 64%) of patients accepted the invitation for screening. There was a wide range of uptake across practices (median 64%; range 33–96%; IQR 20%). Of the 639, 25 patients proven to have a documented history of AF (20) or exclusion criterion (2) on chart review or developed new AF (3), leaving 614 eligible, consenting patients of whom 566 (56%) completed screening. 3.2. Patient characteristics Table 1 gives details of the age and sex of patients. Patients invited for screening had a median age of 76 years (range 70–98) and were significantly younger than current patients excluded by practices for not meeting the study criteria or other unsuitability for screening (U = 282,923; p b 0.001). Following the invitation to be screened,
younger patients were more likely to accept the screening offer than older patients (U = 116,697; p b 0.001). Men were more likely than women to accept screening (χ2 = 10.751; df = 1; p b 0.001). Among the 614 eligible patients who completed screening, 88.3% (542) were entitled at the time to free primary care (General Medical Services eligible). Table 2 outlines the prevalence of conditions, known to be associated with atrial fibrillation, among the 614 patients who completed screening. The most common associated conditions were hypertension (n = 296, 48.2%), ischaemic heart disease (n = 78, 12.7%), diabetes (n = 65, 10.6%) and obesity (n = 49, 8.0%). Thirty patients (4.9%) had a history of stroke or transient ischaemic attack (TIA); 33 (5.4%) had a previous myocardial infarction (MI) and seven patients (1.1%) had heart failure. 3.3. Prevalence of atrial fibrillation 3.3.1. Baseline AF prevalence Of the 1447 current patients randomly selected from practice registers, 111 cases of previously identified AF and atrial flutter (7.7%) were
2200 patients ≥ 70 years of age randomly selected and reviewed for eligibility (26 practices)* 754excluded as not a current patient:
443 excluded: AF/AFL** n=111 Cognitive Impairment n=73 Terminal Illness Other exclusion
1447 current patients
n=21 n=238
1003 patients invited for screening
3 known AF cases
364 declined (36%)
639 consented (64%)
3 AF cases newly diagnosed prior to screening
249
20 known AF cases identified during baseline data collection from patients’ charts
2 patients invited in error
614 eligible patients (61%)
566 had 3-lead ECG (56%)
12 new cases of AF (10) and AFL (2) identified on 3-lead ECG (1.2%)
* Exclusion data is missing for 2 practices: missing data is ≤ 80 cases **AFL atrial flutter Fig. 1. Flow of patients through the study. * Exclusion data is missing for 2 practices: missing data is ≤80 cases **AFL atrial flutter.
250
G. Bury et al. / International Journal of Cardiology 178 (2015) 247–252
Table 1 Patient demography. Group
Total
Data available for
Male (%)
Mean age
Median age
Age range
Excluded patientsa Invited patients Consented Declined Screened New AF cases
443 1003 639 362 566 12
329 (74%) 954 (95%) 639 (100%) 315 (87%) 566 (100%) 12 (100%)
122 (37) 352 (37) 257 (40) 93 (30) 229 (40) 11 (92)
83 77 78 78 78 79
83 76 76 77 77 81
71–101 70–98 70–95 70–98 70–94 73–88
a
Current patients excluded by practices during patient selection.
identified by practices during review for study eligibility. A further 20 previously identified AF cases were found by the research team when collecting baseline data from participating patients' charts. The baseline prevalence of previously identified AF among the sample of current patients was thus 9.3% (n = 134) (see Table 3).
3.3.2. Newly identified AF Five hundred and seventy eligible patients (57%) attended their GP surgery for AF screening. In total, 566 patients had 3-lead ECGs recorded and analysed by the cardiologist. A 3-lead ECG could not be registered for one patient despite two separate attempts and a 12-lead ECG was recorded instead. Three patients had 3-lead ECGs recorded which failed to save to the data card. Twelve new cases of AF (n = 10) and atrial flutter (n = 2) were identified on the 3-lead ECG by the cardiologist, yielding a 2.1% point prevalence among the screened group (1.2% based on ITT). The median CHA2DS2-VASc score among the 12 new AF cases was 4 (range 1–7). Nine of the 12 cases had AF/atrial flutter confirmed by subsequent testing. Six of the nine patients were anti-coagulated with warfarin, two were prescribed aspirin and one dabigatran; and one patient was already on clopidogrel and aspirin due to PCI within the past year. The 12 new AF cases identified on 3-lead ECG tended to be older than the non-AF cases in the screened group, but this was not statistically significant (U = 4237; p = 0.074). The 12 cases were, however, significantly more likely to be male than the non-AF cases screened (χ2 = 11.162; df = 1; p b 0.001). Of 570 patients who attended for screening, four (0.7%) could not be completed because of technical issues with the FR2+. Of 566 tests carried out, all were of sufficient quality to allow analysis. In all cases in which AF was identified, the GP was prompted to confirm with a 12 lead ECG; in 11 (1.9%) other cases a 12 lead ECG was requested to facilitate interpretation — in four cases for low amplitude readings, in four
Table 2 Prevalence of risk factors for AF among patients who completed screening (n = 614). Conditions
Hypertension Ischaemic heart disease/coronary artery disease Diabetes Obesity Chronic obstructive pulmonary disease Other heart surgery & cardiac procedures Pneumonia Stroke Hyperthyroidism Transient ischaemic attack Echocardiographic abnormalities Heart failure Pulmonary embolism Cardiomyopathy Sleep apnoea Lung cancer Lung surgery
Patients n
%
296 138 65 49 29 19 17 16 15 14 8 7 5 5 5 2 1
48.2 22.5 10.6 8.0 4.7 3.1 2.8 2.6 2.4 2.3 1.3 1.1 0.8 0.8 0.8 0.3 0.2
cases because of baseline artefacts and in three cases because of tachycardias. The total prevalence of AF among the sample of current patients was therefore 149/1447 (10.3%). 4. Discussion The key outcome measures of the study include: (i) The point prevalence rate of atrial fibrillation among screened patients was 2.1%. Among the 1447 patients initially identified as possible screening subjects, the total prevalence of AF was 10.3% (new and existing cases). (ii) Four of 570 (0.7%) patients who attended for screening had problems which prevented them from being screened on the day in question and in 11 (1.9%) cases, a 12 lead ECG was advised for clarification of the rhythm. (iii) The screening uptake rate by patients was 56.6% of the 1003 patients invited; 64% consented to take part. (iv) The acceptability of the screening process to patients and staff is reported separately but is very high. This study has identified a study sample which is likely to be highly representative of patients aged 70 and older currently attending Irish general practices. Patient selection used randomised samples from computerised records and then narrowed this to current patients (1447) and those who met inclusion criteria (1003). Participating GPs excluded 238 patients on the basis of their knowledge of the patients' clinical, social or personal circumstances (other than AF, terminal illness or significant cognitive impairment) and these exclusions may be a source of some uncertainty about uptake of screening offers among the entire older population. In all, 64% of eligible patients responded positively to a written invitation to be screened and 56% completed the process. While this is a significant response to a completely unfamiliar screening process, further research is needed to identify the barriers and facilitators to participation in AF screening. This is particularly relevant if repeat screening is to be offered; the current study provides a point prevalence rate but can offer no insight into incidence rates or the most appropriate frequency of screening. Three-lead ECG screening using FR2 + defibrillator/monitors yielded a 2.1% point prevalence rate of newly identified AF among patients over 70 years of age who underwent screening. This detection rate compares very favourably with the detection rates in similar populations using 12-lead ECGs. Among patients over 65 years of age who underwent systematic 12-lead ECG screening in the SAFE study, 2.2% (52/2357) had newly identified AF on ECG [11]. A similar study in Sweden of 75–76 year olds identified 1.3% new AF cases among patients with no known AF (10/767) who underwent a 12-lead ECG [10]. A key finding of this study was the high rate of known or previously documented AF in the general practices under study. Further research is underway to document the risk profiles and treatment strategies being used for these patients. The finding not only highlights the very significant issues raised by AF in Irish general practice, in terms of detection and treatment but also demonstrates the potential impact of effective use of existing records to monitor at-risk groups. Overall 2.6% of rhythm strips reviewed by the cardiologist required a 12-lead ECG for rhythm clarification on the basis that the rhythm strip suggested issues best explored with a 12 lead ECG — none of these abnormalities was suspected to be AF and a detailed breakdown of the rhythm abnormalities identified in this population will be published separately. This suggests that this technology may indeed be an effective and efficient screening method for AF. Although just three-quarters of the AF cases identified on the 3-lead ECG were confirmed by subsequent testing, this is likely to be due to paroxysmal arrhythmia in some cases. Almost 57% of our invited patients attended for AF screening, which is similar to the attendance rate for ECG screening in the SAFE study
G. Bury et al. / International Journal of Cardiology 178 (2015) 247–252
251
Table 3 AF prevalence rates. Denominator
AF cases
%
Total prevalence Baseline prevalence (known AF from exclusion report) Baseline prevalence (all known AF cases)
1447 (current patients) 1447 (current patients) 1447 (current patients)
10.3 7.7 9.3
Point prevalence (newly identified AF)
566 (screened patients)
149 (all known and new AF cases) 111 (exclusion report from practices) 134 (all known AF cases, including those identified through data collection from charts and reported by GP/PN) 12 (new AF cases)
(53%), which was also conducted within a primary care setting [11]. The Swedish study conducted 12-lead ECG screening in an outpatient setting and had 64% of invited patients attend [10]. This higher uptake may be due to cultural differences or differences in patients' perceptions of secondary versus primary care, with the specialist nature of the former possibly being more highly valued. We found a large range of uptake across our 26 practices, suggesting that general practice setting and/or GP–patient relations may be important variables in screening uptake. We found that older patients and women were less likely to take up the offer of screening. This is of concern given that older and female patients are at higher risk of stroke if they have AF. The Swedish study found that those who declined screening had more cardiovascular risk factors than those who accepted [10] Research also suggests that patients have little understanding of AF and its risks [19,20]. Public awareness and other initiatives are thus needed to educate and encourage the most high risk groups to engage with screening. Our study shows that the FR2+ units used in this study deliver high quality rhythm strips, suitable for clinical analysis. The technology is suitable for use in either structured screening exercises such as this one, or in case-finding exercises, based on risk factors or the finding of an irregular pulse. A further study is currently underway to examine the preferences for these strategies among a broader spectrum of general practices. The units used in this exercise were FR2+ defibrillators which have an in-built 3-lead monitoring capacity — however their primary and core role is as defibrillators in emergencies. A free-standing 3-lead monitoring unit would remove any possible impingement on the availability of the FR2+ and might also contain analytical software for rhythm analysis. We currently have no information about the extent to which other general practices in Ireland or other countries use AEDs equipped with three-lead monitoring capability but a study to explore this is now underway in Ireland. 4.1. Study limitations This study consisted of a single ECG screening of patients. It therefore may have missed cases of paroxysmal AF. The Swedish screening study identified 7% new AF cases among patients undergoing extended intermittent ECG screening using a hand-held ECG recorder [10]. Although the screened group were at particular risk of stroke (CHADS2 score ≥ 2), it underlines the challenge involved in detecting paroxysmal AF. Patients may have been reluctant to take part in a research study as opposed to an established screening programme. Practices may also have been more or less encouraging of their patients in availing of this screening opportunity. Practices excluded a substantial number of patients for discretionary reasons, usually because patients were immobile or in a nursing home and attending the surgery would be difficult. This screening modality is tailored to a well-structured general practice system with high quality registers of at-risk individuals (in this case populations of those aged 70 and older) and the availability of AEDs with three-lead monitoring. It also used a centralised specialist diagnostic service to confirm the presence of AF. While these elements may be available or potentially available in the wider Irish healthcare system or in other similarly organised healthcare systems, they may not be common components of primary care systems elsewhere. The use of such
2.1
an approach to population screening will therefore depend on the availability of similar resources or on access to other novel technologies including traditional telemedicine approaches. It remains unclear how often patients should be screened for optimum stroke prevention. 5. Conclusion Three-lead ECG screening using FR2+ defibrillator/monitors yields a comparatively high detection rate for newly identified AF among patients aged 70 years of age or older in general practice. This technology offers a novel screening tool with the potential to systematically improve the detection of AF in primary care. Conflicts of interest All authors take responsibility for all aspects of the reliability and freedom from bias of the data presented and their discussed interpretation. The authors have no conflicts of interest. Acknowledgements The authors wish to thank all the general practitioners, practice nurses and other members of practice staff who contributed their time and expertise to this study. References [1] W.B. Kannel, P.A. Wolf, E.J. Benjamin, D. Levy, Prevalence, incidence, prognosis, and predisposing conditions for atrial fibrillation: population-based estimates, Am. J. Cardiol. 82 (8A) (1998) 2N–9N. [2] P.A. Wolf, R.D. Abbott, W.B. Kannel, Atrial fibrillation as an independent risk factor for stroke: the Framingham study, Stroke 22 (1991) 983–988. [3] C. Marini, F. De Santis, S. Sacco, T. Russo, L. Olivieri, R. Totaro, A. Carolei, Contribution of atrial fibrillation to incidence and outcome of ischemic stroke: results from a population-based study, Stroke 36 (2005) 1115–1119. [4] N. Hannon, O. Sheehan, L. Kelly, et al., Stroke associated with atrial fibrillation: incidence and early outcomes in the north Dublin population stroke study, Cerebrovasc. Dis. 29 (2010) 43–49. [5] R.G. Hart, L.A. Pearce, M.I. Aguilar, Meta-analysis: antithrombotic therapy to prevent stroke in patients who have nonvalvular atrial fibrillation, Ann. Intern. Med. 146 (12) (2007) 857–867. [6] P.D. Ziegler, T.V. Glotzer, D.G. Daoud, D.E. Singer, et al., Detection of previously undiagnosed AF in patients with stroke risk factors and usefulness of continuous monitoring in primary stroke prevention, Am. J. Cardiol. 110 (2012) 1309–1314. [7] L.A. Sposato, F.R. Klein, A. Jauregui, M. Ferrua, et al., Newly diagnosed AF after acute ischemic stroke and TIA: importance of immediate and prolonged continuous cardiac monitoring, J. Stroke Cerebrovasc. Dis. 21 (2012) 210–216. [8] T.V. Glotzer, P.D. Ziegler, Silent AF as a stroke risk factor and anticoagulant indication, Can. J. Cardiol. 29 (2013) S14–S23. [9] Atrial Fibrillation in Primary Care (AFIP): Bringing Atrial Fibrillation Practice Closer to Guidelines. A Tool for Primary Care Physicians. (2012) [10] J. Engdahl, L. Andersson, M. Mirskaya, M. Rosenqvist, Stepwise screening of atrial fibrillation in a 75-year-old population: implications for stroke prevention, Circulation (2013) 930–937. [11] D.A. Fitzmaurice, F.D.R. Hobbs, S. Jowett, et al., Screening versus routine practice in detection of atrial fibrillation in patients aged 65 or over: a cluster randomised controlled trial, BMJ 335 (7616) (2007) 383. [12] G. Cooke, J. Doust, S. Sanders, Is pulse palpation helpful in detecting atrial fibrillation? A systematic review, J. Fam. Pract. 55 (2) (2006). [13] K. Harris, D. Edwards, J. Mant, How can we best detect atrial fibrillation? J. R. Coll. Physicians. Edinb. 42 (S18) (2012) 5–22. [14] J. Mant, D.A. Fitzmaurice, F.D.R. Hobbs, et al., Accuracy of diagnosing atrial fibrillation on electrocardiogram by primary care practitioners and interpretative diagnostic
252
G. Bury et al. / International Journal of Cardiology 178 (2015) 247–252
software: analysis of data from screening for atrial fibrillation in the elderly (SAFE) trial, BMJ 335 (7616) (2007) 380. [15] N. Lowres, L. Neubeck, G. Salkeld, I. Krass, et al., Feasibility and cost-effectiveness of stroke prevention through community screening for atrial fibrillation using iPhone ECG in pharmacies. The SEARCH-AF study, Thromb. Haemost. 111 (6) (2014) 1167–1176. [16] J. Orchard, S.B. Freedman, N. Lowres, D. Peiris, L. Neubeck, iPhone ECG screening by practice nurses and receptionists for atrial fibrillation in general practice: the GPSEARCH qualitative pilot study, Aust. Fam. Physician 43 (5) (2014) 315–319. [17] J.K. Lau, N. Lowres, L. Neubeck, D.B. Brieger, R.W. Sy, C.D. Galloway, D.E. Albert, S.B. Freedman, iPhone ECG application for community screening to detect silent atrial fibrillation: a novel technology to prevent stroke, Int. J. Cardiol. 165 (1) (2013) 193–194.
[18] G. Bury, M. Headon, M. Dixon, M. Egan, Cardiac arrest in Irish general practice: an observational study from 426 general practices, Resuscitation 80 (2009) 1244–1247. [19] G.Y.H. Lip, S. Kamath, M. Jafri, A. Mohammed, D. Bareford, Ethnic differences in patient perceptions of atrial fibrillation and anticoagulation therapy: the West Birmingham atrial fibrillation project, Stroke 33 (2002) 238–242. [20] E. Aliot, G. Breithardt, J. Brugada, et al., An international survey of physician and patient understanding, perception and attitudes to atrial fibrillation and its contribution to cardiovascular disease morbidity and mortality, Europace 12 (2010) 626–633.
