Effect of Continuous Electrocardiogram Monitoring on Detection of Undiagnosed Atrial Fibrillation After Hospitalization for Cardiac Surgery

Key Points Question Can continuous cardiac rhythm monitoring beyond hospital discharge enhance atrial fibrillation (AF) detection among cardiac surgical patients? Findings In this randomized clinical trial of 336 cardiac surgical patients with risk factors for stroke, use of continuous cardiac rhythm monitoring with wearable sensors increased the rate of AF detection by 17.9% within 30 days of hospital discharge compared with usual care. Meaning Among cardiac surgical patients with risk factors for stroke and AF lasting less than 24 hours postoperatively, continuous cardiac rhythm monitoring significantly improved the rate of AF detection during the first 30 days after hospital discharge compared with usual care.


Introduction
Postoperative atrial fibrillation (POAF) after cardiac surgery occurs in 30% to 50% of patients during their hospital stay, with the incidence peaking at 3 to 5 days and decreasing afterwards. 1 However, the incidence of POAF after discharge from cardiac surgery is not well defined. Most studies [2][3][4][5][6][7][8][9][10][11] have been limited to the hospitalization phase only and were small, nonrandomized, or included patients with antecedent atrial fibrillation (AF) and used limited monitoring for AF. Among cardiac surgical patients who experienced little to no POAF during hospitalization, their risk of experiencing AF after discharge is unknown. 12,13 Quantifying the risk of ongoing POAF after hospitalization is an important but unanswered issue because many patients have elevated stroke risk (ie, CHA 2 DS 2 -VASc [congestive heart failure, hypertension, age Ն75 years, diabetes, prior stroke or transient ischemic attack, vascular disease, age 65-74 years, female sex] score Ն2); hence, diagnosing AF in this population may be actionable. 14 In fact, POAF is associated with a 2-fold increase in early adverse outcomes such as stroke and death, 1-3 and the risk of adverse events is known to persist even during long-term follow-up. 1,2 Because there are sparse clinical trial data, guidelines provide little direction for clinicians on the optimal duration of monitoring, follow-up, or treatment for patients after hospitalization for cardiac surgery, particularly if they are in sinus rhythm at discharge. 15,16 Randomized data guiding detection and management of POAF after hospitalization for cardiac surgery are lacking. Accordingly, we conducted a randomized clinical trial comparing continuous 30-day electrocardiography (ECG) monitoring with usual care for the detection of AF or atrial flutter (AFL) after hospitalization for cardiac surgery.

Trial Design and Setting
The Post-Surgical Enhanced Monitoring for Cardiac Arrhythmias and Atrial Fibrillation (SEARCH-AF) trial was an investigator-initiated, prospective, open-label, multicenter, randomized clinical trial at 10 tertiary care cardiac surgical centers in Canada. Each site's research ethics board approved the study, and all participants provided written informed consent. The trial protocol, protocol amendments, and statistical analysis plan are available in Supplement 1. The statistical analysis plan was written without knowledge of outcome data. This report was prepared in accordance with the Consolidated Standards of Reporting Trials (CONSORT) reporting guideline for randomized trials.

Study Population
Patients were eligible if they were aged 18 years or older, underwent cardiac surgery (coronary artery bypass grafting [CABG] or valve repair or replacement with or without CABG), had no history of AF or AFL before surgery, had AF or AFL lasting less than 24 hours during hospitalization after surgery, and had a CHA 2 DS 2 -VASC score of greater than or equal to 4 or greater than or equal to 2 plus at least 1 additional factor associated with the risk developing POAF (chronic obstructive pulmonary disease, sleep apnea, impaired renal function, left atrial enlargement, elevated body mass index, or combined CABG with valve repair or replacement). Patients were excluded if they had a presurgical history of AF or AFL, had AF or AFL lasting 24 hours or longer after cardiac surgery, were in AF or AFL at the time of randomization, were hospitalized for 10 days or longer at the time of randomization, or received a mechanical valve or had other reason to be on oral anticoagulation. Race/ethnicity was assessed to examine its potential association with the incidence of POAF. Data on race/ethnicity were collected from review of medical records. Detailed inclusion and exclusion criteria are listed in eTable 1 in Supplement 2.

Trial Procedures
Patients were randomly assigned in a 1:1 ratio to the monitoring group (continuous 30-day ECG monitoring) or to usual care with no mandated monitoring. Randomization lists were prepared by the trial statistician (K.E.T.) and were computer-generated using random permuted blocks.
Randomization was stratified by study site and by the type of cardiac surgery performed (isolated CABG or valve replacement or repair with or without CABG). Randomization occurred between the third postoperative day and discharge from hospital. In each site, study coordinators (with oversight from the site principal investigator) enrolled patients in the trial and administered the trial interventions.
Patients randomized to the monitoring group received up to 30 days of continuous ECG monitoring with a wearable, adhesive patch monitor. Until September 30, 2018, the SEEQ system (Medtronic) was used. Afterward, the CardioSTATsystem (Icentia) was used because SEEQ manufacture was halted. Maximum monitoring duration for a single SEEQ or CardioSTAT patch was 7.5 and 14 days, respectively. Patients in the monitoring group received 4 SEEQ or 2 CardioSTAT patches, which provided up to 30 and 28 days of ECG monitoring, respectively. Before hospital discharge, the patch was applied on the anterior left chest, away from the sternotomy site. Patients were provided instruction on monitor use and were contacted weekly (SEEQ) or biweekly (CardioSTAT) to ensure compliance. Used patches were returned to the manufacturer for analysis. A report including all arrhythmic episodes with ECG strips was provided to site investigators and subsequently forwarded to the patient's primary care physician, cardiologist, and cardiac surgeon at the end of the 28-day or 30-day monitoring period.
Patients randomized to usual care did not undergo any protocol-mandated ECG monitoring within the first 30 days after randomization. This practice is consistent with current guidelines, which make no recommendations for routine monitoring. [15][16][17] However, if clinically indicated, patients in the usual care group could undergo standard ECG or Holter monitoring within the first 30 days of randomization. If AF or AFL was documented on a 12-lead ECG, performance of additional monitoring, such as Holter monitoring, was suggested. Any AF or AFL detected by an ECG, Holter, or any other type of monitoring performed in the usual care group was included for end point assessment to avoid negatively biasing the event rate in this group. In both groups, therapeutic decisions, such as initiation of oral anticoagulation therapy, were left to local physicians' discretion.
Patients were followed at 1 to 3 months and 6 months after discharge from surgery on an in-person basis. A telephone-based follow-up was conducted at 9 months after discharge from surgery. At 6 months, patients in both study groups underwent 14 days of continuous cardiac rhythm monitoring (SEEQ or CardioSTAT patch).

Outcomes
The primary outcome was documentation of cumulative AF or AFL duration of 6 minutes or longer or documentation of AF or AFL by a single 12-lead ECG within 30 days after randomization. A list of prespecified secondary outcomes is provided in eTable 2 in Supplement 2. Rhythm-based outcomes, major adverse cardiovascular events, and major bleeding were adjudicated by a committee of physicians blinded to treatment assignment.

Statistical Analysis
The trial was designed to test the hypothesis of superiority of continuous ECG monitoring over usual care in the detection of the primary outcome within 30 days after randomization. The rate of the primary outcome was estimated to be 3% during 30 days in the usual care group. 18 To detect a 7% absolute increase in primary outcome detection with a strategy of continuous cardiac rhythm monitoring, a total of 388 patients (194 in each group) would provide the trial with 80% power at a 2-sided α = .05. Assuming a 2% attrition rate, the final sample size was 396 patients.
Outcome analyses were performed on the intention-to-treat population, which analyzed patients according to their randomization status. The difference in the primary outcome between the 2 study groups was reported as the absolute rate difference with associated 95% CIs. Comparison of the proportion of patients between the 2 groups with the primary outcome was performed with the Pearson χ 2 test. The Pearson χ 2 test was used to compare the proportions between groups for secondary outcomes. A subgroup analysis was performed involving 5 prespecified subgroups (sex, age, CHA 2 DS 2 -VASc score, surgery type [isolated CABG vs valve repair or replacement with or without CABG], and left atrial size) to assess for effect modification between the 2 randomization groups. A Kaplan-Meier curve of the primary outcome was generated for illustrative purposes.
A per-protocol analysis was prespecified and consisted of all randomized patients who received their allocated intervention without significant deviations in the assigned treatment during the first 30 days after randomization. For patients randomized to the continuous cardiac rhythm monitoring group, the per-protocol cohort was defined by patients who wore the sensor for 24 hours or longer within the first 30 days after randomization. For patients randomized to the usual care group, the per-protocol cohort was defined by patients who did not wear a continuous cardiac rhythm monitor (SEEQ or CardioSTAT monitor) during the first 30 days after randomization. Statistical analyses were performed with R statistical software version 4.0.2 (R Project for Statistical Computing).

Study Population
Between pandemic, all sites mandated institutional restrictions on surgical procedures and research recruitment. On July 17, 2020, enrollment in the study was terminated, at which time 336 patients were enrolled, accounting for 85% of the planned sample size. The last protocol-mandated clinical follow-up occurred on August 31, 2020, and the last follow-up related to results of cardiac rhythm monitoring occurred on September 11, 2020. The median (IQR) time from surgery to randomization was 5 (4-6) days, and the median (IQR) time from randomization to discharge was 0 (0-1) days.
There was no statistically significant difference in the baseline characteristics of the 2 assigned groups ( Table 1). The median (IQR) CHA 2 DS 2 -VASc score was 4.0 points (3.0-4.0 points); 255 patients (75.9%) underwent CABG alone, 39 patients (11.6%) underwent valve repair or replacement alone, and 42 patients (12.5%) underwent both CABG and valve repair or replacement. At hospital discharge, amiodarone was prescribed for 13 patients (7 in the monitoring group and 6 in the usual care group). None of the patients had a presurgical history of AF or AFL by design, and 318 patients A total of 2803 patients were screened for the study and 336 patients underwent randomization. In the group that was assigned to receive continuous monitoring, 8 patients did not undergo any monitoring and 2 patients completed the monitoring but the data were not received by the monitoring center. In the usual care group, no patient underwent continuous monitoring and 4 patients underwent non-protocol-mandated Holter monitoring. During follow-up, 6 patients withdrew consent in both groups. There were 6 and 11 patients who were lost to follow-up in the intervention and usual care group, respectively. A total of 163 patients in the monitoring group and 173 patients in the usual care group were included in the primary analysis. A detailed list of reasons for screen failure is provided in eTable 3 in Supplement 2. CHA 2 DS 2 -VASc indicates congestive heart failure, hypertension, age 75 years or older, diabetes, prior stroke or transient ischemic attack, vascular disease, age 65 to 74 years, and female sex.

Primary Outcome
In the continuous monitoring group, the primary outcome was detected in 32 of 163 patients (19.6%) compared with 3 of 173 patients (1.7%) in the usual care group, with an absolute difference of 17.9% (95% CI, 11.5%-24.3%; P < .001) ( Table 2 and Figure 2).The number needed to screen to detect the primary outcome was 6 patients (95% CI, 4-9 patients). Of the 32 patients in the monitoring group in whom the primary outcome occurred, 30 of them had at least 1 episode of detected AF or AFL lasting for 6 minutes or longer.
In the monitoring group, the primary outcome was detected by continuous monitoring in 30 of 32 patients. For the other 2 patients, AF was detected by 12-lead ECG. The first episode of AF or AFL lasting for 6 minutes or longer was detected in 22 patients (73.3%) during the first week of monitoring, in 6 patients (20.0%) during the second week, and in 2 patients (6.7%) during the third week. The cumulative duration of detected AF or AFL episodes and the number of patients with detected AF or AFL lasting for 6 minutes or longer, stratified by monitoring weeks, is shown in     During the first 45 days after discharge, a total of 11 patients (3.3%) were prescribed oral anticoagulation therapy (

Subgroup Analysis and Per-Protocol Analysis
The treatment effect of the study intervention was similar between the prespecified subgroups.

Discussion
In this randomized clinical trial, we found that continuous monitoring during the first 30 days after hospitalization for cardiac surgery detected significantly more POAF than usual care among patients No. of patients with cumulative AF or AFL lasting for ≥6 min detected by monitoring in the intervention group (n = 163) Box and whisker plots of the cumulative duration of detected AF or AFL durations for patients randomized to continuous cardiac rhythm monitoring within the first 30 days of randomization. The upper and lower ends of the box represent the first and third quartile. The line in the box represents the median. In week 2, the median was 0.2 minutes. In week 1, the whisker above the upper quartile is drawn up to the largest observed point that falls within 1.5 times of the interquartile range (IQR). In week 2, the whisker is 1.5 times of the IQR. The whiskers below the lower quartile are not illustrated because they cross zero. Outliers are presented as circles located outside the whiskers. There were 10 patients who had AF lasting 6 minutes or longer in multiple monitoring weeks. which is assumed to resolve in days. 19 Although POAF after cardiac surgery is a common clinical problem, most studies 20,21 have only described its incidence during hospitalization. The incidence of POAF occurring weeks to months after surgery is not well-defined; prior studies 3-11 are few, retrospective, or nonrandomized. This represents an important unknown for clinicians managing these patients. In our study, continuous cardiac rhythm monitoring detected AF in nearly 1 of 5 patients within 30 days after discharge and was markedly higher than the rate of detected AF in the usual care group. This indicates that a substantial proportion of POAF is subclinical and would not be diagnosed without the use of continuous cardiac rhythm monitoring. In this study, patients had AF lasting less than 24 hours during hospitalization and would likely not have prompted any further monitoring or treatment under normal clinical circumstances. In this patient population, continuous ECG monitoring was essential to uncover the ongoing risk and extent of outpatient POAF. In particular, the occurrence of POAF was clustered within the first 2 weeks after discharge in our study cohort.
In our study, the rate of AF detection with continuous cardiac rhythm monitoring was substantially higher than that in patients in the nonsurgical setting and with a similar risk profile. The CHA 2 DS 2 -VASc score is associated with development of AF in the nonsurgical setting. 22,23 Among patients with a CHA 2 DS 2 -VASc score similar to that for our cohort (median, 4.0 points), AF incidences of 0.9 to 1.5 cases per 100 patient-years were reported, using limited modalities to ascertain AF. 22,23 In a study 24 using continuous monitors in a nonsurgical population with a CHA 2 DS 2 -VASc score similar to that of our population, the detection rate of new-onset AF was 6.2% after 30 days.
Compared with these studies, our 30-day AF detection rate of 19.6% was much higher than that in a general, matched population.
The rate of oral anticoagulation therapy in our patients was low. This was consistent with published studies 25,26 showing that rates of oral anticoagulation therapy after cardiac surgery were low despite a high incidence of in-hospital POAF. Reluctance to prescribe anticoagulation therapy may be related to concerns over bleeding, the belief that POAF is self-limiting, or uncertainty regarding the duration of AF that would justify the use of oral anticoagulation for stroke prevention in this patient population. In our population with a median CHA 2 DS 2 -VASc score of 4 and AF lasting 6 minutes or longer, the annual stroke risk is estimated to be 1.3%, 27 which would meet the threshold for oral anticoagulation therapy. Reluctance to prescribe oral anticoagulation therapy may reflect uncertainty in clinical practice guidelines as current recommendations for oral anticoagulation therapy after discharge are limited by lack of randomized trial data, especially if patients are in sinus rhythm. 15,16 Our data may help inform these guidelines in terms of the optimal method for POAF detection after discharge.
Our findings suggest that the incidence of POAF may decrease rapidly over time because the rate of AF detection markedly declined after the first 2 weeks of hospital discharge. However, previous studies 1, 13,14,20,28 found an association between POAF and cardiovascular risk over longterm follow-up, when, presumably, POAF had largely resolved. Future research is needed to assess whether early anticoagulation may mitigate adverse short-term and long-term cardiovascular outcomes. 29

Limitations
This study has limitations that should be addressed. The primary end point is limited to 30 days and does not estimate the long-term, ongoing risk of POAF. Among patients with POAF lasting less than