Accurate diagnosis of arrhythmias is improved with longer monitoring duration but can risk delayed diagnosis. We compared diagnostic yield, outcomes, and resource utilization by arrhythmia monitoring strategy in 330 matched adults (mean age 64 years, 40% women, and 30% non-White) without previously documented atrial fibrillation or atrial flutter (AF/AFL) who received ambulatory electrocardiographic monitoring by 14-day Zio XT (patch-based continuous monitor), 24-hour Holter, or 30-day event monitor (external loop recorder) between October 2011 and May 2014. Patients were matched by age, gender, site, likelihood of receiving Zio XT patch, and indication for monitoring, and subsequently followed for monitoring results, management changes, clinical outcomes, and resource utilization. AF/AFL ≥30 seconds was noted in 6% receiving Zio XT versus 0% by Holter (p = 0.04) and 3% by event monitor (p = 0.07). Nonsustained ventricular tachycardia was noted in 24% for Zio XT patch versus 8% (p <0.001) for Holter and 4% (p <0.001) for event monitor. No significant differences between monitoring strategies in outcomes or resource utilization were observed. Prolonged monitoring with 14-day Zio XT patch or 30-day event monitor was superior to 24-hour Holter in detecting new AF/AFL but not different from each other. Documented nonsustained ventricular tachycardia was more frequent with Zio XT than 24-hour Holter and 30-day event monitor without apparent increased risk of adverse outcomes or excess utilization. In conclusion, additional efforts are needed to further personalize electrocardiographic monitoring strategies that optimize clinical management and outcomes.
Ambulatory electrocardiographic (AECG) monitoring is used to diagnose arrhythmias and their correlation with symptoms. The type of AECG monitor selected is driven by the frequency and severity of symptoms and by the level of concern for high-risk arrhythmias. Each technology’s operating characteristics (i.e., ability to capture arrhythmias, diagnostic accuracy, and timeliness of reporting) can also impact subsequent resource utilization. Longer duration monitors have higher diagnostic yield for detection of atrial fibrillation or atrial flutter (AF/AFL), , as well high-risk arrhythmias. However, previous studies primarily focused on the duration and approach to AECG monitoring for diagnosing AF/AFL in selected scenarios (e.g., cryptogenic stroke). We evaluated the diagnostic yield, outcomes, and resource utilization in matched adults without previous documented AF/AFL who received AECG monitoring using a 14-day patch-based continuous monitor, 24-hour Holter, or 30-day event monitor for various clinical indications.
The source population was derived from 2 integrated health care delivery systems that provide comprehensive care for >4.5 million members in Kaiser Permanente Northern California (KPNC) and >4.6 million members in Kaiser Permanente Southern California (KPSC). The collective membership is highly representative of the local surrounding and statewide population in terms of age, gender, and race/ethnicity. This study was approved by the KPNC and KPSC institutional review boards, and waiver of informed consent was obtained due to the observational nature of this data-only study.
The Kaiser Permanente Real-World Heart Monitoring Strategy Evaluation, Treatment Patterns and Health Metrics in Atrial Fibrillation (KP-RHYTHM) Study first identified all adult members who underwent up to 14 days of continuous AECG monitoring using the Zio XT patch (iRhythm Technologies, Inc., San Francisco, California) between October 2011 and May 2014. After mailing the monitor back to iRhythm Technologies, Inc., the recording was analyzed using proprietary algorithms with verification of algorithm output and report creation by certified cardiac technicians. , The final report is then transferred to the treating physician.
Because we aimed to characterize yield and clinical utility of Zio XT patch monitoring versus other AECG monitoring strategies, we separately identified additional KPNC and KPSC patients who were monitored by a 24-hour Holter or cellular transmission-enabled 30-day event monitor during the same study period. Patients were excluded if they were <18 years old at the time of monitoring, had <12 months of previous continuous membership and pharmacy benefit, disenrolled after AECG monitoring, died during AECG monitoring, had a previous organ transplant, did not have any analyzable wear time during the monitoring period if receiving a Zio XT patch, or had previous documented AF/AFL. If the same patient received multiple monitoring modalities during the study period, we analyzed data from the first AECG monitoring approach.
We ascertained the primary indication for AECG monitoring by board-certified physician review of clinical notes within 30 days before monitoring for suspected AF/AFL, presence of ischemic stroke and/or transient ischemic attack, syncope or presyncope, ventricular tachycardia or ventricular fibrillation, heart block, palpitations, shortness of breath/dyspnea, chest pain or other chest symptoms other than palpitations, bradycardia, or other indications. In patients who were monitored by Zio XT patch, other than suspected AF/AFL, all other primary indications were obtained using information from adjudicated monitor reports.
Physician adjudicators also reviewed AECG monitoring reports, imaging reports, and clinical notes up to 30 days after monitoring to ascertain arrythmias that were detected (i.e., AF/AFL, other supraventricular tachycardias, sustained or nonsustained ventricular tachycardia or fibrillation, sinus pauses >3 seconds, other pauses >5 seconds during AF/AFL, other pauses, and high-grade heart block). For detected AF/AFL, we also ascertained the duration of the arrythmia.
In our cohort of adults without previous documented AF/AFL, we performed matching using a 1:1:1 matching strategy for each of the three AECG monitoring strategies. Specifically, we created sets of patient triplets of individuals receiving a Zio XT patch, 24-hour Holter, or 30-day event monitor that were matched by age (±3 years), gender, Kaiser Permanente region, and a high-dimensional propensity score (hdPS) difference (≤0.01) in the predicted likelihood of receiving a Zio XT patch. We employed the nearest neighbor approach when matching by hdPS. We evaluated the balance of the hdPS by comparing the standardized mean difference in both pair of comparison groups (i.e., Zio XT patch vs 24-hour Holter and Zio XT patch vs 30-day event monitor), and the mean difference was <0.10.
We evaluated diagnostic yield between the AECG monitoring strategies by comparing frequencies in detected newly diagnosed AF/AFL, other supraventricular tachycardia, sustained ventricular tachycardia or fibrillation, nonsustained ventricular tachycardia, sinus pauses >3 seconds, other pauses >5 seconds for adults in AF/AFL, other symptomatic pauses, and high-grade heart block from monitoring results reports. In patients with previously detected AF/AFL, we compared the duration of arrythmia that was detected and categorized the duration as permanent AF/AFL, ≥30 seconds AF/AFL, <30 seconds AF/AFL (if described for Holters and event monitors), or unknown duration (if no duration was mentioned for Holters and event monitors).
To assess any change in clinical management that might have occurred due to the different AECG monitoring strategies, we identified separate subsets of patients without previous exposure to oral vitamin K antagonists, direct oral anticoagulants, or antiplatelet agents; rate control agents (beta blocker, calcium channel blocker, and digoxin); antiarrhythmic agents (procainamide, disopyramide, flecainide acetate, propafenone, sotalol, dofetilide, and amiodarone); implanted pacemakers; or implantable cardioverter defibrillators. We then compared across monitoring strategies in each of the subsets the number and proportion of patients who initiated each of the new management strategies within 90 days after monitoring.
The clinical outcomes of interest included all-cause death, hospitalization for ischemic stroke or transient ischemic attack, other arterial thromboembolic event, and hospitalization for heart failure within 90 days after monitoring. Deaths were comprehensively ascertained from electronic medical record (EMR) data, Social Security Administration vital status files, and California state death certificate files. We searched health system hospital discharge and emergency department databases for principal discharge diagnoses using previously validated International Classification of Diseases, Ninth and Tenth Edition codes for ischemic stroke or transient ischemic attack, arterial thrombotic event, and heart failure (codes available on request). ,
To assess for any differential resource utilization by AECG monitoring strategy, we measured any hospitalization, ambulatory visit (emergency department or outpatient clinic), or receipt of coronary revascularization within 90 days after the index monitoring episode.
We identified age, self-reported gender, and self-reported race (White, Black, Asian/Pacific Islander, and so on) and Hispanic ethnicity from health plan databases. Systolic and diastolic blood pressures and body mass index were identified from outpatient clinic visits. We identified stroke risk factors including previous ischemic stroke, heart failure, diabetes, hypertension, and reduced estimated glomerular filtration rate from EMR data using previously described methods. We calculated each patient’s predicted risk of thromboembolism using the Anticoagulation and Risk Factors in Atrial Fibrillation (ATRIA) score that has been shown to more accurately predict the risk of ischemic stroke than CHA 2 DS 2 -VASc in the participating sites. The ATRIA stroke risk score ranges from 0 to 15 (low risk = 0 to 5, moderate risk = 6, and high risk ≥7).
All analyses were conducted using SAS software, version 9.4 (SAS Institute Inc, Cary, North Carolina). We performed pairwise comparisons, separately comparing matched patients receiving either 24-Holter or 30-day event monitoring with patients receiving Zio XT patch as the reference group on patient characteristics before monitoring, indications for monitoring, diagnostic yield of the monitoring, clinical management, clinical outcomes, and resource utilization; we used the Kruskal-Wallis test to compare median values for continuous variables, chi-square test for categoric variables, and Fisher’s exact test for any categoric values with cells <5. A 2-sided p value <0.05 was considered significant.
The sponsor provided data from the analyzed Zio XT patches but had no role in the study design, analysis and interpretation of the data, writing the report, or in the decision to submit the article for publication.
We identified 3,530 eligible adults monitored by Zio XT patches, 23,113 by 30-day event monitors, and 74,534 by 24-hour Holters between October 2011 and May 2014 in the participating health care delivery systems. Among those without previous documented AF/AFL, we successfully matched 110 adults who were monitored with a Zio XT patch to 110 adults receiving a 24-hour Holter and separately to 110 adults receiving a 30-day event monitor. Premonitoring baseline characteristics of the 110 matched triplets were well matched ( Table 1 ). Overall, median (interquartile range) age was 66 (57 to 75) years, 40% were women, 30% were persons of color, and the median (interquartile range) ATRIA stroke risk score was 3 (1 to 6). The primary monitoring indications were primarily for syncope/presyncope (26%), palpitations (25%), and ventricular tachycardia or fibrillation (24%) ( Table 2 ).
| Variable | Overall | Zio XT Patch | 24-hour Holter | p Value Zio XT Patch vs 14-dayHolter | 30-day Event Monitor | p Value Zio XT Patch vs 30-day Event Monitor |
|---|---|---|---|---|---|---|
| (N=330) | (N=110) | (N=110) | (N=110) | |||
| Median (IQR) Age (years) | 66 (57-75) | 66 (57-75) | 65 (57-75) | 0.99 | 66 (57-75) | 0.95 |
| Women | 132 (40%) | 44 (40%) | 44 (40%) | 1.00 | 44 (40%) | 1.00 |
| Race | 0.74 | 0.68 | ||||
| White | 228 (69%) | 80 (73%) | 74 (67%) | 74 (67%) | ||
| Black | 35 (11%) | 10 (9%) | 10 (9%) | 15 (14%) | ||
| Asian/Pacific Islander | 40 (12%) | 13 (12%) | 15 (14%) | 12 (11%) | ||
| Hispanic ethnicity | 43 (13%) | 16 (15%) | 12 (11%) | 0.72 | 15 (14%) | 0.98 |
| ATRIA Risk Score, median (IQR) | 3 (1-6) | 3 (1-6) | 3 (1-6) | 0.67 | 3 (1-6) | 0.63 |
| Heart failure | 22 (7%) | 6 (6%) | 7 (6%) | 0.77 | 9 (8%) | 0.42 |
| Hypertension | 193 (59%) | 61 (56%) | 65 (59%) | 0.59 | 67 (61%) | 0.41 |
| Proteinuria | 24 (7%) | 8 (7%) | 5 (5%) | 0.39 | 11 (10%) | 0.47 |
| Ischemic stroke or transient ischemic attack | 24 (7%) | 11 (10%) | 4 (4%) | 0.06 | 9 (8%) | 0.64 |
| Systolic blood pressure (mm Hg) | 0.13 | 0.54 | ||||
| 160-179 | 5 (2%) | 3 (3%) | 1 (1%) | 1 (1%) | ||
| 140-159 | 35 (11%) | 16 (15%) | 7 (6%) | 12 (11%) | ||
| 130-139 | 79 (24%) | 24 (22%) | 36 (33%) | 19 (17%) | ||
| 121-129 | 77 (23%) | 26 (24%) | 23 (21%) | 28 (26%) | ||
| ≤120 | 132 (40%) | 40 (37%) | 43 (39%) | 49 (45%) | ||
| Diastolic blood pressure (mm Hg) | 0.76 | 0.16 | ||||
| 100-109 | 1 (0.3%) | 1 (1%) | 0 (0%) | 0 (0%) | ||
| 90-99 | 12 (4%) | 4 (4%) | 7 (6%) | 1 (1%) | ||
| 85-89 | 11 (3%) | 5 (5%) | 5 (5%) | 1 (1%) | ||
| 81-84 | 26 (8%) | 10 (9%) | 9 (8%) | 7 (6%) | ||
| ≤80 | 278 (84%) | 89 (82%) | 89 (81%) | 100 (92%) | ||
| Body mass index (kg/m 2 ) | 0.12 | 0.06 | ||||
| <18.5 | 7 (2%) | 4 (4%) | 2 (2%) | 1 (1%) | ||
| 18.5-25.0 | 114 (35%) | 47 (43%) | 33 (30%) | 34 (31%) | ||
| 25.0-29.9 | 108 (33%) | 34 (31%) | 41 (37%) | 33 (30%) | ||
| 30.0-39.9 | 85 (26%) | 23 (21%) | 28 (26%) | 34 (31%) | ||
| ≥40 | 14 (4%) | 1 (1%) | 6 (6%) | 7 (6%) |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
