Abstract

OBJECTIVES

Postoperative atrial fibrillation (POAF) occurs in 25–40% of patients following cardiac surgery, and is associated with a significant increased risk of stroke and mortality. Routine surveillance is not performed post-discharge; however, recurrence of POAF can occur in up to 30% of patients discharged in sinus rhythm. This study aimed to determine the feasibility of patients self-monitoring with an iPhone handheld electrocardiogram (iECG) to identify recurrence of POAF in the post-discharge period following cardiac surgery.

METHODS

Patients with POAF following cardiac surgery were eligible for participation if they had no prior history of atrial fibrillation (AF) and were discharged home in stable sinus rhythm. Participants were provided with an iECG and asked to record a 30-s iECG, four times per day for 4 weeks post-discharge. iECGs were automatically transmitted to a secure server, and reviewed for the presence of AF by the research team and a validated algorithm. All participants also received brief education on AF.

RESULTS

Forty-two participants completed the intervention (mean age 69 ± 9 years, 80% male). Self-monitoring for POAF recurrence using an iECG was feasible and acceptable, and participants felt empowered. Self-monitoring identified 24% (95% confidence interval, 12–39%) with an AF recurrence within 17 days of hospital discharge. These participants were significantly younger than those without AF recurrence (64 ± 7 vs 70 ± 10 years; P = 0.025), and had a significantly lower CHA 2 DS 2 -VASc score (2.3 ± 1.2 vs 3.7 ± 2.3; P = 0.007). However, 80% were at high enough stroke risk to warrant consideration of anticoagulation, i.e. CHA 2 DS 2 -VASc score ≥2. Only 30% of recurrences were associated with palpitations. Participation also improved AF knowledge from 6.4 ± 1.8 to 7.3 ± 1.8 ( P = 0.02), of a total score of 10.

CONCLUSIONS

Providing patients with an iECG is a non-invasive, inexpensive, convenient and feasible way to monitor for AF recurrence in post-cardiac surgery patients. It also provides a mechanism to provide knowledge about the condition and also potentially reduce anxiety. The success of patients using this technology also has implications for extending the use of iECG self-monitoring to other patient groups such as those undergoing antiarrhythmic interventions for AF.

INTRODUCTION

Postoperative atrial fibrillation (POAF) occurs in 25–40% of patients following cardiac surgery [ 1 , 2 ]. It is associated with a significantly increased risk of stroke both before and after discharge [ 3 ], and significantly increased short- and long-term mortality [ 4 ]. Although often thought to be transient, POAF is associated with a significant risk of recurrence, with 18% experiencing atrial fibrillation (AF) recurrence in the first year postoperatively [ 5 ], and 25% by 5 years [ 6 ]. However, the majority of recurrences may go undetected by both patients and their physicians, as it has been shown that only a third of all AF episodes are symptomatic with palpitations [ 7 ], and stroke risk is significantly elevated in patients with asymptomatic AF [ 7 ]. Despite this, guidelines for management of POAF do not currently recommend surveillance or monitoring for patients with new POAF who revert to stable sinus rhythm prior to discharge [ 8 ].

If monitoring is performed in the post-discharge period, recurrence of AF can be detected in >30% of patients with transient POAF discharged in sinus rhythm [ 9–11 ].

To date, monitoring studies have utilized the 12-lead ECG, Holter monitors, telemetry or loop recorders [ 9–11 ]; which require health professionals to review results to identify recurrences. New portable technology has the potential to enable individuals to monitor themselves for AF post-discharge. There are a variety of validated devices including: handheld portable ECGs such as the AliveCor heart monitor [ 12 ], Zenicor EKG [ 13 ] and Mydiagnostick [ 14 ]; modified blood pressure monitors such as Microlife BPA 200 [ 15 ] and Omron M6 [ 16 ]; and iPhone photo plethysmographs [ 17 ]. Additional smartphone-based ECG devices exist including the QardioCore and ECG Check; however, the AliveCor heart monitor (iECG) is the only validated smartphone ECG published in the literature. We have demonstrated the efficacy of health professionals using an iECG to detect unknown AF [ 18 ]. The iECG has an automated AF detection algorithm that we validated with recordings both in a clinic setting (98% sensitivity, 97% specificity) [ 12 ], and in community pharmacies (98.5% sensitivity and 91.4% specificity) [ 18 ]. This accuracy makes it an ideal device to detect asymptomatic or unrecognized AF.

We performed this study to determine the feasibility of patients self-monitoring with an iECG to identify recurrence of AF in the post-discharge period following cardiac surgery, and to determine if providing a brief inpatient AF education programme improves patient knowledge of AF and its related health risks, symptoms and medical management.

MATERIALS AND METHODS

Study design

This feasibility study used a cross-sectional study design to investigate patient self-monitoring to detect the recurrence of AF post-hospital discharge; and a pre–post study design to evaluate a brief educational module on AF (ACTRN12614000383662). The study was conducted between March 2014 and July 2015, at Royal North Shore Hospital (tertiary teaching hospital), and North Shore Private Hospital in Sydney, Australia. This study was approved by the Northern Sydney Local Health District Human Resource Ethics Committee (HREC/13/HAWKE/415) and the North Shore Private Hospital Ethics Committee (NSPHEC 2014-006).

The study protocol has been described in detail elsewhere [ 19 ]. In brief, we recruited cardiothoracic surgery patients who experienced a transient episode of POAF following cardiac surgery; with no history of AF prior to admission; who reverted or were cardioverted to stable sinus rhythm prior to discharge; and were ≥18 years old. Patients were approached and assessed during their inpatient admission. The baseline assessment included demographic data, medical history, postoperative complications, AF knowledge (assessed using two open-ended questions, and a modified version of the Atrial Fibrillation Knowledge Scale [ 20 ]) ( Supplementary Material 1 ), and a 30-s lead-1 ECG to determine baseline heart rhythm (assessed using the iECG). All participants received the intervention, as this was a feasibility study with no requirement for randomization or blinding. All medical treatment, including the decision to prescribe oral anticoagulation (OAC), was at the discretion of the treating specialists and was not influenced by the research team.

All participants received brief one-on-one education regarding AF, symptoms and health risks associated with AF particularly in relation to stroke, and were provided with take-home educational materials. Participants were also provided with an iPhone and an AliveCor Heart monitor (iECG) (Fig. 1 ) for a 4-week period post-hospital discharge. Participants were requested to record a 30-s iECG four times a day during the study period, and take additional iECGs if AF symptoms were experienced, recording any symptoms in a diary. Each iECG was automatically transmitted to a secure server, where the recordings were reviewed by our Research Assistant, and analysed by a validated algorithm [ 12 ] for the presence of AF. If AF was identified, the participant was contacted to arrange follow-up with their treating physician, and the participant's treating physicians and specialists were advised of the findings. Participants were telephoned once or twice during the 4-week period to ensure no difficulties using the iECG. On completion, AF knowledge and postoperative complications were reassessed. Participants were also invited to participate in a semi-structured interview designed to explore the experience of using the iECG, the benefits and challenges of the programme and factors related to future sustainability.

Figure 1:

AliveCor heart monitor.

Figure 1:

AliveCor heart monitor.

Outcomes

The primary and secondary outcomes of the study, as previously reported [ 19 ], are outlined in Table 1 .

Table 1:

Outcome measures

Primary outcome 
Feasibility of patient self-monitoring for AF recurrence using an iECG, evaluated using:
  • Acceptability and willingness to participate in the programme (measured using recruitment data including the reasons provided for non-participation)

  • Participant ability to learn to use the iECG technology and successfully take their own recordings

  • Participant compliance to the intervention (measured by the number of iECG recordings they recorded over the 4 weeks—i.e. did they achieve the requested target of 3–4 recordings each day)

  • The ability of the iECG and automated AF algorithm to identify recurrences of AF (calculated using sensitivity and specificity; compared with a cardiologist interpretation of the iECG in combination with the noise-reduced iECG, and 12-lead ECG or Holter monitor where performed)

 
Secondary outcomes 
1. The proportion of participants identified with recurrent AF 
2. Estimated stroke risk of participants identified with recurrent AF (using the CHA 2 DS 2 -VASc score).  
3. Participant knowledge of AF (measured using two open-ended questions, and a modified version of the Atrial Fibrillation Knowledge Scale [ 20 ]).  
4. Qualitative data regarding acceptability of use of the iECG (participant interviews) 
Primary outcome 
Feasibility of patient self-monitoring for AF recurrence using an iECG, evaluated using:
  • Acceptability and willingness to participate in the programme (measured using recruitment data including the reasons provided for non-participation)

  • Participant ability to learn to use the iECG technology and successfully take their own recordings

  • Participant compliance to the intervention (measured by the number of iECG recordings they recorded over the 4 weeks—i.e. did they achieve the requested target of 3–4 recordings each day)

  • The ability of the iECG and automated AF algorithm to identify recurrences of AF (calculated using sensitivity and specificity; compared with a cardiologist interpretation of the iECG in combination with the noise-reduced iECG, and 12-lead ECG or Holter monitor where performed)

 
Secondary outcomes 
1. The proportion of participants identified with recurrent AF 
2. Estimated stroke risk of participants identified with recurrent AF (using the CHA 2 DS 2 -VASc score).  
3. Participant knowledge of AF (measured using two open-ended questions, and a modified version of the Atrial Fibrillation Knowledge Scale [ 20 ]).  
4. Qualitative data regarding acceptability of use of the iECG (participant interviews) 

Statistical considerations

Primary analyses were conducted using SPSS for Windows (Version 19.0). New episodes of AF were expressed as true positives divided by the total number screened with accompanying binomial 95% confidence intervals (CIs) calculated using Clopper–Pearson methodology. χ2 -tests were used to assess any associations between AF recurrence and age group, gender or AF risk factors. AF knowledge level was compared pre- and postintervention using paired t -tests. Continuous variables are reported as means ± standard deviations (SD), and categorical variables as numbers and percentages. Within subject differences between baseline and follow-up were analysed using Wilcoxon signed-rank tests for non-parametric variables (two-tailed P < 0.05 considered significant). Analysis was limited to complete cases to avoid artificially increasing precision around the estimates by imputing values or carrying baseline values forward. To ensure an analysis of completers was appropriate, Little's Missing Completely At Random (MCAR) test was performed to ensure data were at least missing at random, and that any missingness could be ignored. This test returned a non-significant P -value of 0.641; therefore, the hypothesis that data were missing at random could not be rejected, and missing data were assumed to be MCAR and to not have influenced the results. A power calculation was not performed, as the primary outcome was feasibility and acceptability. A sample size of 50 participants was chosen to maximize the probability of reaching data saturation during thematic analysis of the interviews and during review of process measures such as reasons for declining participation. Interview data were transcribed verbatim and were independently reviewed and analysed by two reviewers using general interpretive methods [ 21 ].

RESULTS

Participants

In total, 44 participants were recruited to the study, and 42 completed the intervention (Fig. 2 ). The mean age of participants was 69 ± 9 years (±SD), and 80% were male. The baseline characteristics are presented in Table 2 . During the study period, 7 participants were readmitted to hospital for the following reasons: recurrence of AF ( n = 2), pleural or pericardial effusion ( n = 3), fever and hypertension ( n = 1) and a planned abdominal aortic aneurysm repair ( n = 1).

Table 2:

Participant characteristics ( n = 44)

Characteristic Number 
Gender 
 Male 35 80 
 Female 20 
Surgical procedure 
 CABG 21 48 
 Valve 14 
 CABG + valve 11 25 
 Aortic surgery ± valve 11 
 Atrial septal defect repair 
Comorbidities 
 Stroke/TIA/thromboembolism 16 
 Diabetes 13 30 
 Hypertension 33 69 
 Congestive heart failure 
 Vascular disease 35 80 
Level of education 
 Did not complete high school 20 45 
 Completed high school 20 
 Tertiary education 15 34 
 Mean (SD) Range 
Age (years) 69 (9) (45–85) 
Postoperative length of stay (days) 13 (9) (6–39) 
Stroke risk (CHA 2 DS 2 -VASc score)  3.4 (1.7) (0–7) 
BMI 29 (6) (21–54) 
Characteristic Number 
Gender 
 Male 35 80 
 Female 20 
Surgical procedure 
 CABG 21 48 
 Valve 14 
 CABG + valve 11 25 
 Aortic surgery ± valve 11 
 Atrial septal defect repair 
Comorbidities 
 Stroke/TIA/thromboembolism 16 
 Diabetes 13 30 
 Hypertension 33 69 
 Congestive heart failure 
 Vascular disease 35 80 
Level of education 
 Did not complete high school 20 45 
 Completed high school 20 
 Tertiary education 15 34 
 Mean (SD) Range 
Age (years) 69 (9) (45–85) 
Postoperative length of stay (days) 13 (9) (6–39) 
Stroke risk (CHA 2 DS 2 -VASc score)  3.4 (1.7) (0–7) 
BMI 29 (6) (21–54) 

CABG, coronary artery bypass graft; TIA, transient ischaemic attack; SD, standard deviation; CHA 2 DS 2 -VASc: [C: congestive heart failure/left ventricular dysfunction, H for high blood pressure, A 2 : age >75 years, D: diabetes, S 2 : stroke/transient ischaemic attack/thromboembolism, V: vascular disease (coronary artery disease, myocardial infarction, peripheral artery disease, aortic plaque), A: age 65–74 years, Sc: female gender]; BMI: body mass index.

Figure 2:

Study flow. AF: atrial fibrillation; CVA: cerebrovascular accident; iECG: iphone handheld electrocardiogram (AliveCor Heart Monitor).

Figure 2:

Study flow. AF: atrial fibrillation; CVA: cerebrovascular accident; iECG: iphone handheld electrocardiogram (AliveCor Heart Monitor).

Feasibility

Ability of the iPhone handheld electrocardiogram to identify recurrences of atrial fibrillation

Patient self-monitoring using an iECG to detect recurrences of POAF is feasible, and detected AF recurrence in 10/42 participants (outlined in detail in the following section). During the study, 3481 iECGs were recorded, of which 146 (4%) were non-diagnostic due to interference caused by issues such as hand tremor or poor mobile-phone reception: 3 rural participants with poor mobile reception accounted for the majority (88/146) of these readings. Therefore, in total 3335 diagnostic iECGs were recorded over the study period, with representative examples shown in Fig. 3 . The automated AF algorithm had high accuracy for detection of AF, with sensitivity of 94.6% (95% CI, 85.1–98.9) and specificity 92.9% (95% CI, 92.0–93.8). The majority of false-positive iECGs were associated with low-voltage p-waves and QRS complexes, atrial ectopy and left bundle branch block. Interestingly, 81% of participants were also identified with atrial and/or ventricular ectopy, or sinus arrhythmia during the study period.

Figure 3:

iECG recordings. ( A ) Sinus rhythm. ( B ) Atrial fibrillation.

Figure 3:

iECG recordings. ( A ) Sinus rhythm. ( B ) Atrial fibrillation.

Acceptability

Participation in the programme was well accepted by the majority of patients. Only 14 (24%) of the 58 patients approached declined participation, the majority of those who refused stating they were feeling too overwhelmed post-surgery to participate in a research study (Fig. 2 ). The majority of patients (9/14) who declined were female with a mean age of 71 ± 11 years.

Ability to learn and use iPhone handheld electrocardiogram

The iECG was reported to be easy to use by 95% of participants. Participants easily adapted to using the iECG technology, and surprisingly age was not a barrier (age range 45–85 years). For participants with higher education levels and previous smartphone experience, a 5- to 10-min practice session was required to successfully learn to use the iECG. Each participant was also provided with a user manual, and followed up with a short visit the following day to ensure the correct technique in recording iECGs. For participants with lower education level and no previous smartphone experience, a longer training session of up to 20 min was required to learn to use the iECG, including additional tasks such as switching the phone on and off, charging the phone and how to open and close an ‘app’. However, lack of previous smartphone experience did not impede the ability to self-monitor in the discharge period: ‘I thought I might [have trouble with the technology] because I don't have those sorts of iPhones and things, but I handled it quite well and didn't have any problems’ [Patient #40] . Only 2 participants reported they needed a ‘familiarization period’: ‘I had a bit of trouble using it for about a week but after that it was fine’ [Patient #19].

Compliance

Overall, there was good compliance with the requested iECG recording schedule of 3–4 iECGs per day for 4 weeks. iECGs were recorded for a mean of 29 ± 5 days (range 9–46), with 86% of participants recording iECGs for 27 days or more, and only 2 participants (5%) recording for <21 days. A mean of 2.8 ± 0.9 iECGs were recorded per day. Participants reported that self-monitoring was not onerous, took only a short time and they were able to incorporate it into their daily routine such as with meals, while taking blood pressure recordings and by having it by their bedside. They stated they had ample time to complete the recordings as they were in a period of convalescence with reduced work and social commitments.

Benefits

Participants felt empowered by the ability to monitor their heart rate and rhythm daily, and the feedback on the presence or absence of AF, ‘it was absolutely fantastic because I would never have been able to keep a track of what was happening’ [Patient #30]. For some patients, the iECG recording confirmed an abnormality in their heart rhythm, ‘it was only through using the iPhone that I discovered I had episodes of atrial fibrillation before going back to hospital’ [Patient #42]. Patients also reported peace of mind and reassurance that their heart rhythm was being monitored once they had been discharged from hospital.

Barriers

In some rural areas, poor mobile-phone reception affected the quality of iECG readings, resulting in participants having to move to a position with better mobile reception and re-record the iECG. Movement artefact also affected the quality of iECG recordings; therefore, participants had to ensure their fingers were correctly covering the electrodes, and that they remained still for the duration of the recording. Another common problem encountered by participants was forgetting to charge the phone.

Recurrence of atrial fibrillation post-discharge

Incidence and stroke risk

Within 3 weeks of discharge (range 1–17 days), self-monitoring with the iECG detected a POAF recurrence in 10/42 participants, equating to 24% (95% CI, 12–39%). These participants (mean age 64 ± 7 years) were on average 7 years younger than those without AF recurrence (mean age 70 ± 10 years) P = 0.025. No association was noted for gender nor comorbidities including body mass index (BMI). Additionally, the mean CHA 2 DS 2 -VASc score of participants with an AF recurrence was lower (2.3 ± 1.2) than those without recurrence (3.7 ± 2.3) P = 0.007. However, 80% of participants (8/10) with an AF recurrence had high enough stroke risk (i.e. CHA 2 DS 2 -VASc score ≥2) to recommend anticoagulation.

Symptoms

Symptoms were not a reliable indicator of AF recurrence, as the majority of AF episodes were either asymptomatic or associated with atypical symptoms, with only 30% associated with palpitations (Table 3 ). Additionally, 25% of all participants reported palpitations which were not AF-related, but due to atrial or ventricular ectopic beats.

Table 3:

POAF recurrences ( n = 10)

Symptoms experienced (per participant) Number (%) Timeframe to first recurrence post-discharge (range in days)  CHA 2 DS 2 -VASc score (range)  
Asymptomatic 4 (40) 1–8 2–3 
Palpitations 3 (30) 1–6 1–2 
Dizziness or fatigue or dyspnoea 2 (22) 1–17 1–5 
Palpitations or dizziness or fatigue 1 (11) 
Symptoms experienced (per participant) Number (%) Timeframe to first recurrence post-discharge (range in days)  CHA 2 DS 2 -VASc score (range)  
Asymptomatic 4 (40) 1–8 2–3 
Palpitations 3 (30) 1–6 1–2 
Dizziness or fatigue or dyspnoea 2 (22) 1–17 1–5 
Palpitations or dizziness or fatigue 1 (11) 

POAF, postoperative atrial fibrillation; CHA 2 DS 2 -VASc: [C: congestive heart failure/left ventricular dysfunction, H for high blood pressure, A 2 : age >75 years, D: diabetes, S 2 : stroke/transient ischaemic attack/thromboembolism, V: vascular disease (coronary artery disease, myocardial infarction, peripheral artery disease, aortic plaque), A: age 65–74 years, Sc: female gender].

Follow-up interventions for participants with atrial fibrillation recurrence

During the study period, 2/10 participants were readmitted to hospital for AF recurrence: one was started on OAC, cardioverted to sinus rhythm and subsequently taken off anticoagulant, and the other was medically managed with sotalol and OAC. An additional 2/10 participants experienced further AF recurrences: both managed with increased doses of metoprolol, and one of these commenced on OAC. The remaining 6/10 participants were screened post study using either a 12-lead ECG or a 24-h Holter monitor; however, no further paroxysmal AF was detected: 2 of these participants remained on metoprolol, and 1 on OAC, though this was indicated for a mechanical valve.

Of the 10 patients with recurrent AF detected by the iECG, 8 had a CHA 2 DS 2 -VASc score ≥2, and only 3/8 remained on OAC after 1 month, with 1 of the 3 requiring warfarin because of a mechanical valve. In general, those without further recurrence were not prescribed OAC.

Patient knowledge of atrial fibrillation

Overall, participation in the study increased participant knowledge of AF. Results for the AF knowledge questionnaire (total score out of 10) improved from a mean of 6.4 ± 1.8 to 7.3 ± 1.8 ( P = 0.02). The greatest improvement within the AF questionnaire was noted in questions related to ‘AF medications' (Questions 2, 5 and 6) and questions related to ‘understanding of AF’ (Questions 4 and 10) (Fig. 4 ). When the AF knowledge questionnaire was combined with the two short answer questions (total score out of 14), a significant improvement in knowledge was noted from a mean of 8.3 ± 2.6 to 9.5 ± 2.8 ( P = 0.01). The addition of the short answer questions to the AF knowledge questionnaire improved the internal consistency reliability from low to moderate: Cronbach's α increased from 0.41 to 0.56 at baseline and from 0.43 to 0.63 at follow-up.

Figure 4:

AF Knowledge Questionnaire.

Figure 4:

AF Knowledge Questionnaire.

Understanding of atrial fibrillation

Following the intervention, both knowledge of AF and symptom awareness improved, as measured by the short answer questions (Fig. 5 ). At baseline, <50% of participants were able to correctly describe AF, with the majority of these responses focused on AF as ‘a rapid and/or irregular heart rhythm’ or ‘an electrical disorder of the heart affecting function’. Only 3% were aware of an association between AF and increased risk of stroke, and 27% were unable to provide any response regarding AF. Postintervention, ∼75% of participants were able to correctly describe AF, with additional descriptors of ‘the upper chambers are not beating and pumping properly’; and the proportion able to describe an associated stroke risk increased from 3% to 20%. However, 13% of participants still reported no knowledge of AF.

Figure 5:

Short Answer Questions.

Figure 5:

Short Answer Questions.

Symptom awareness

Prior to the intervention, knowledge of AF symptoms was poor: only 34% of participants reported that palpitations were a symptom of AF; 23% could not report any symptoms; and 48% mentioned at least one of the atypical symptoms. Patients often related this question to their own experience in hospital. At follow-up, knowledge had improved, with 63% aware that palpitations were a symptom of AF and 78% able to state at least one atypical symptom. However, only 40% were able to correctly identify ≥2 of the AF symptoms, scoring ‘good’ in the short answer question (Fig. 5 ); and 20% of participants were unable to state any symptoms of AF.

DISCUSSION

Self-monitoring for POAF recurrence using an iECG is feasible and acceptable, and empowers patients in management of their health. Self-monitoring identified 24% (95% CI, 12–39%) with an AF recurrence within 17 days of hospital discharge post-cardiac surgery. Surprisingly, these patients were on average younger and they had a lower CHA 2 DS 2 -VASc score than those without AF recurrence. However, 80% were at high enough stroke risk to warrant consideration of anticoagulation. Of the 7 participants with non-valvular AF recurrence and CHA 2 DS 2 -VASc score ≥2, only 2 were placed on oral anticoagulant post-discharge, highlighting the issues surrounding the uncertainty and interpretation of stroke risk by treating physicians, and the inconsistency in duration or type of OAC prescription post-discharge for patients with a transient episode of POAF. Reliance on the patient reporting symptoms to detect AF recurrence is not sufficient, as only 30% of AF recurrences were associated with palpitations. Therefore, the majority of AF recurrences may not have been identified without additional monitoring. Additionally, the majority of palpitation episodes experienced were not AF recurrences, but due to atrial and ventricular ectopic beats.

Monitoring for AF post-discharge is important, as it is not possible to predict which patients will have a recurrence. Methods for predicting the risk of developing POAF, such as using the CHA 2 DS 2 -VASc score [ 22 ], or identifying those with a higher BMI [ 23 ], do not appear useful for determining POAF recurrence risk, as the CHA 2 DS 2 -VASc score was lower for those with a recurrence and there was no association noted for BMI. There are many options for self-monitoring that can be utilized, other than the iECG. Using pulse palpation as a self-monitoring tool is an option that would be inexpensive, but it is likely to face some problems in this population as 81% of participants experienced episodes of atrial ectopy, ventricular ectopy or sinus arrhythmia during the study; therefore, many false positives could be detected, and it is likely there could be a high proportion of false negatives. Accuracy of pulse palpation for detection of AF is low, even when used by health professionals, with a specificity of only 72% [ 24 ], and one could not hope for even this level of accuracy with an elderly lay population. Likewise, devices such as modified blood pressure cuffs, which rely on RR interval irregularity, may also have poor specificity in this population [ 15 , 16 ]. The iECG algorithm uses both RR irregularity and the presence of p-waves in the analysis phase; thus, accuracy is less likely to be influenced by the high incidence of ectopy and sinus arrhythmia.

The diagnostic accuracy of handheld ECG devices is comparable with the traditional methods of transtelephonic and 24-h Holter monitoring. Over a 3-month period, there was excellent diagnostic agreement between repeated iECG recordings and traditional transtelephonic monitoring for post-ablation patients ( κ statistic 0.82) [ 25 ]. Interestingly, 92% of participants preferred the iECG to transtelephonic monitors, which reflects the positive acceptability identified in our study [ 25 ]. Other recent studies have also demonstrated that repeated intermittent handheld ECG (Zenicor EKG) recordings are superior to a 24-h monitor to screen patients with ambiguous symptoms: self-recorded ECGs for 28 days yielded an additional 9.5% with unknown AF [ 13 ]. Continuous monitoring with an implanted device is likely to detect more episodes of paroxysmal AF than intermittent iECG readings. However, implanted devices are relatively expensive, require a minor surgical procedure, though becoming less invasive with the newer generation of smaller devices, and need health professional monitoring. In contrast, self-monitoring with an iECG empowers the patient and, when used with the automated ‘on-phone AF algorithm’, will reduce the need for a health professional. Furthermore, the iECG is compatible with both Android and iPhones; so patients could be provided with an AliveCor Heart Monitor case and have the app set up on their own phone.

The main limitation for obtaining diagnostic quality iECGs was the interference caused by poor mobile reception in some rural areas. This was not a widespread problem, with only 3/22 rural participants experiencing this problem. It is possible to overcome this issue by activating ‘airplane mode’ to record each iECG; however, this technique was not routinely implemented in this study. Likewise, interference from hand tremors can also be overcome by placing the iECG directly over the chest in contact with the skin. An additional limitation is that self-monitoring occurred for only 1 month post-discharge; therefore, additional AF recurrences may have been detected if monitoring had been extended past this time.

Additional benefits noted from participation in the study were patient empowerment and increased AF knowledge. Knowledge improvements were noted in three key areas of importance for this population: i.e. understanding of AF and the associated stroke risk; AF symptom awareness; and medication management, specifically anticoagulation. However, the timing of AF education in the inpatient setting post-cardiac surgery may not be ideal, as this is a busy time for many patients with priority being given to inpatient treatment and investigations post-surgery, and management of post-surgical pain. To overcome barriers associated with the inpatient setting, it is possible that a patient education module could be introduced onto the smartphone or device, in an acceptable and effective manner, which allows the patient to work through the education at their own pace. This may further empower the patient in management of their health and understanding of AF.

The results of this study should be interpreted with caution, as the sample size was small due to its design as a feasibility study. Patients with a recurrence were found to be younger ( P = 0.025) and had a lower CHA 2 DS 2 -VASc score ( P = 0.007) than those without a recurrence; however, it is possible this is due to the small number of patients. Additionally, it is possible that improvements in AF knowledge were influenced by a ‘memory effect’, resulting in better results the second time the questionnaire was completed.

However, it appears that self-monitoring with an iECG is feasible and acceptable in post-cardiac surgery patients. It is also likely self-monitoring with the iECG would be feasible in other populations, such as patients who have had catheter or surgical ablation, or other antiarrhythmic interventions including pharmacological therapy. It would also be feasible to use this technique in future studies investigating whether POAF may occur after discharge post-cardiac surgery in the absence of an inpatient episode of POAF.

CONCLUSION

Providing patients with an iECG case to attach to a smartphone is a non-invasive, inexpensive, convenient and feasible way to monitor for AF recurrence in post-cardiac surgery patients. It also provides a mechanism to provide knowledge about the condition and also potentially reduce anxiety. The success of patients using this technology also has implications for extending the use of iECG self-monitoring to other patient groups such as those undergoing anti-arrhythmic interventions for AF.

SUPPLEMENTARY MATERIAL

Supplementary material is available at EJCTS online .

Funding

This work was supported by a competitive grant from the Cardiothoracic Surgery Research and Education Fund, Sydney Medical School Foundation, University of Sydney. AliveCor provided ECG Heart Monitors for study purposes: the investigators are not affiliated with, nor have any financial or other interest in AliveCor. Nicole Lowres was funded by a National Heart Foundation Postgraduate Scholarship (PP12S6990). Lis Neubeck is an NHMRC early career fellow (APP1036763). The funding sources had no role in study design; collection, analysis or interpretation of data; writing of the report; or the decision to submit the article for publication. Researchers were independent from the funders and sponsors of the study.

Conflict of interest: Saul Ben Freedman reports grants, personal fees and non-financial support from Bayer Pharma AG outside the submitted work, grants and non-financial support from Boehringer Ingelheim outside the submitted work, grants and personal fees from BMS/Pfizer outside the submitted work, personal fees from Servier outside the submitted work, personal fees from Astra-Zeneca outside the submitted work and consulting fees from Gilead outside the submitted work. Lis Neubeck has received grants and honoraria from BMS/Pfizer outside the submitted work.

ACKNOWLEDGEMENTS

We thank the staff of the Royal North Shore Hospital and North Shore Private Hospital for their assistance and support during this study.

REFERENCES

1
Helgadottir
S
,
Sigurdsson
M
,
Ingvarsdottir
I
,
Arnar
D
,
Gudbjartsson
T
.
Atrial fibrillation following cardiac surgery: risk analysis and long-term survival
.
J Cardiothorac Surg
 
2012
;
7
:
87
.
2
Mozaffarian
D
,
Marchioli
R
,
Macchia
A
,
Silletta
MG
,
Ferrazzi
P
,
Gardner
TJ
et al
.
Fish oil and postoperative atrial fibrillation: the Omega-3 Fatty Acids for Prevention of Post-operative Atrial Fibrillation (OPERA) randomized trial
.
JAMA
 
2012
;
308
:
2001
11
.
3
Gialdini
G
,
Nearing
K
,
Bhave
PD
,
Bonuccelli
U
,
Iadecola
C
,
Healey
JS
et al
.
Perioperative atrial fibrillation and the long-term risk of ischemic stroke
.
JAMA
 
2014
;
312
:
616
22
.
4
Phan
K
,
Ha
HSK
,
Phan
S
,
Medi
C
,
Thomas
SP
,
Yan
TD
.
New-onset atrial fibrillation following coronary bypass surgery predicts long-term mortality: a systematic review and meta-analysis
.
Eur J Cardiothorac Surg
 
2015
;
48
:
817
24
.
5
Melduni
RM
,
Schaff
HV
,
Bailey
KR
,
Cha
SS
,
Ammash
NM
,
Seward
JB
et al
.
Implications of new-onset atrial fibrillation after cardiac surgery on long-term prognosis: a community-based study
.
Am Heart J
 
2015
;
170
:
659
68
.
6
Ahlsson
A
,
Fengsrud
E
,
Bodin
L
,
Englund
A
.
Postoperative atrial fibrillation in patients undergoing aortocoronary bypass surgery carries an eightfold risk of future atrial fibrillation and a doubled cardiovascular mortality
.
Eur J Cardiothorac Surg
 
2010
;
37
:
1353
9
.
7
Healey
JS
,
Connolly
SJ
,
Gold
MR
,
Israel
CW
,
Van Gelder
IC
,
Capucci
A
et al
.
Subclinical atrial fibrillation and the risk of stroke
.
N Engl J Med
 
2012
;
366
:
120
9
.
8
Dunning
J
,
Treasure
T
,
Versteegh
M
,
Nashef
SAM
.
Guidelines on the prevention and management of de novo atrial fibrillation after cardiac and thoracic surgery
.
Eur J Cardiothorac Surg
 
2006
;
30
:
852
72
.
9
Giaccardi
M
,
Macchi
C
,
Colella
A
,
Polcaro
P
,
Zipoli
R
,
Cecchi
F
et al
.
Postacute rehabilitation after coronary surgery: the effect of preoperative physical activity on the incidence of paroxysmal atrial fibrillation
.
Am J Phys Med Rehabil
 
2011
;
90
:
308
15
.
10
Cioffi
G
,
Mureddu
G
,
Cemin
C
,
Russo
TE
,
Pellegrini
A
,
Terrasi
F
et al
.
Characterization of post-discharge atrial fibrillation following open-heart surgery in uncomplicated patients referred to an early rehabilitation program
.
Italian Heart J
 
2001
;
2
:
519
28
.
11
Lomivorotov
VV
,
Efremov
SM
,
Pokushalov
EA
,
Romanov
AB
,
Ponomarev
DN
,
Cherniavsky
AM
et al
.
Randomized trial of fish oil infusion to prevent atrial fibrillation after cardiac surgery: data from an implantable continuous cardiac monitor
.
J Cardiothorac Vasc Anesth
 
2014
;
28
:
1278
84
.
12
Lau
JK
,
Lowres
N
,
Neubeck
L
,
Brieger
DB
,
Sy
RW
,
Galloway
CD
et al
.
iPhone ECG application for community screening to detect silent atrial fibrillation: a novel technology to prevent stroke [Research Letter]
.
Int J Cardiol
 
2013
;
165
:
193
4
.
13
Hendrikx
T
,
Rosenqvist
M
,
Wester
P
,
Sandstrom
H
,
Hornsten
R
.
Intermittent short ECG recording is more effective than 24-hour Holter ECG in detection of arrhythmias
.
BMC Cardiovasc Disord
 
2014
;
14
:
41
.
14
Tieleman
R
,
Plantinga
Y
,
Rinkes
D
,
Bartels
G
,
Posma
JL
,
Cator
R
et al
.
Validation and clinical use of a novel diagnostic device for screening of atrial fibrillation
.
Europace
 
2014
;
16
:
1291
5
.
15
Wiesel
J
,
Arbesfeld
B
,
Schechter
D
.
Comparison of the Microlife blood pressure monitor with the Omron blood pressure monitor for detecting atrial fibrillation
.
Am J Cardiol
 
2014
;
114
:
1046
8
.
16
Marazzi
G
,
Iellamo
F
,
Volterrani
M
,
Lombardo
M
,
Pelliccia
F
,
Righi
D
et al
.
Comparison of Microlife BP A200 Plus and Omron M6 blood pressure monitors to detect atrial fibrillation in hypertensive patients
.
Adv Ther
 
2012
;
29
:
64
70
.
17
McManus
DD
,
Lee
J
,
Maitas
O
,
Esa
N
,
Pidikiti
R
,
Carlucci
A
et al
.
A novel application for the detection of an irregular pulse using an iPhone 4S in patients with atrial fibrillation
.
Heart Rhythm
 
2013
;
10
:
315
9
.
18
Lowres
N
,
Neubeck
L
,
Salkeld
G
,
Krass
I
,
McLachlan
AJ
,
Redfern
J
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
 
2014
;
111
:
1167
76
.
19
Lowres
N
,
Freedman
SB
,
Gallagher
R
,
Kirkness
A
,
Marshman
D
,
Orchard
J
et al
.
Identifying postoperative atrial fibrillation in cardiac surgical patients posthospital discharge, using iPhone ECG: a study protocol
.
BMJ Open
 
2015
;
5
.
20
Hendriks
JML
,
Crijns
HJGM
,
Tieleman
RG
,
Vrijhoef
HJM
.
The atrial fibrillation knowledge scale: development, validation and results
.
Int J Cardiol
 
2013
;
168
:
1422
8
.
21
Elo
S
,
Kääriäinen
M
,
Kanste
O
,
Pölkki
T
,
Utriainen
K
,
Kyngäs
H
.
Qualitative content analysis: a focus on trustworthiness
.
Sage Open
 
2014
; .
22
Baker
WL
,
Coleman
CI
,
White
CM
,
Kluger
J
.
Use of preoperative CHA2 DS2-VASc score to predict the risk of atrial fibrillation after cardiothoracic surgery: a nested case-control study from the Atrial Fibrillation Suppression Trials (AFIST) I, II, and III
.
Pharmacotherapy
 
2013
;
33
:
489
95
.
23
Bramer
S
,
van Straten
AHM
,
Soliman Hamad
MA
,
Berreklouw
E
,
van den Broek
KC
,
Maessen
JG
.
Body mass index predicts new-onset atrial fibrillation after cardiac surgery
.
Eur J Cardiothorac Surg
 
2011
;
40
:
1185
90
.
24
Cooke
G
,
Doust
J
,
Sanders
S
.
Is pulse palpation helpful in detecting atrial fibrillation? A systematic review
.
J Fam Pract
 
2006
;
55
:
130
4
.
25
Tarakji
KG
,
Wazni
OM
,
Callahan
T
,
Kanj
M
,
Hakim
AH
,
Wolski
K
et al
.
Using a novel wireless system for monitoring patients after the atrial fibrillation ablation procedure: the iTransmit study
.
Heart Rhythm
 
2015
;
12
:
554
9
.

Author notes

Presented at the 29th Annual Meeting of the European Association for Cardio-Thoracic Surgery, Amsterdam, Netherlands, 3–7 October 2015.