Effectiveness of Implanted Cardiac Rhythm Recorders With Electrocardiographic Monitoring for Detecting Arrhythmias in Pregnant Women With Symptomatic Arrhythmia and/or Structural Heart Disease: A Randomized Clinical Trial

Key Points

Question  Does the addition of an implantable loop recorder to 24-hour Holter electrocardiographic monitoring increase the yield of arrhythmias and effect on pregnancy outcome in pregnant women with structural heart disease and/or symptoms suggestive of arrhythmias?

Findings  In this randomized clinical trial of 40 pregnant women with symptoms of arrhythmia and/or structural heart disease, the implantable loop recorder detected arrhythmias that were not detected by prior 24-hour Holter monitoring, leading to a change in management in a significant proportion of patients.

Meaning  An implantable loop recorder in addition to 24-hour Holter monitoring was better than 24-hour Holter monitoring alone in detecting arrhythmias in pregnant women with structural heart disease and/or symptoms suggestive of arrhythmias.

Importance  Arrhythmias are an important cause of maternal morbidity and mortality but remain difficult to diagnose.

Objective  To compare implantable loop recorder (ILR) plus 24-hour Holter electrocardiographic (ECG) monitoring with standard 24-hour Holter ECG monitoring alone in terms of acceptability, ability to identify significant arrythmias, and effect on management and pregnancy outcome in women who were symptomatic or at high risk of arrythmia because of underlying structural heart disease.

Design, Setting, and Participants  This single-center, prospective randomized clinical trial recruited 40 consecutive patients from the Cardiac Disease and Maternity Clinic at Groote Schuur Hospital in Cape Town, South Africa. Pregnant patients with symptoms of arrhythmia and/or structural heart disease at risk of arrhythmia were included.

Intervention  Patients were randomized to standard care (SC; 24-hour Holter ECG monitoring [n = 20]) or standard care plus ILR (SC-ILR; 24-hour Holter ECG monitoring plus ILR [n = 20]). Only 17 consented to ILR insertion, and the 3 who declined ILR were allocated to the SC group.

Main Outcomes and Measures  Arrhythmias considered included atrial fibrillation, atrial flutter, premature ventricular complexes, supraventricular tachycardia, ventricular tachycardia, or ventricular fibrillation.

Results  Among the 40 women in this trial, the mean (SD) age was 28.4 (5.5) years. Holter monitoring detected arrhythmias in 3 of 23 patients (13%) in the SC group and 4 of 17 patients (24%) in the SC-ILR group compared with 9 of 17 patients (53%) patients who had arrhythmias detected by ILR. Seven patients (4 with supraventricular tachycardia, 1 with premature ventricular complexes, and 2 with paroxysmal atrial fibrillation recorded by ILR) did not have arrhythmias detected by 24-hour Holter monitoring. Three of these 7 patients (43%) had a change in management as a result of their ILR recordings. There were no maternal deaths. However, the SC group had a significantly lower mean (SD) gestational stage at delivery (35 [5] weeks vs 38 [2], P = .04).

Conclusions and Relevance  The ILR was better than 24-hour Holter monitoring in detecting arrhythmias, which led to a change in management for a significant proportion of patients. Our findings suggest that ILR may be beneficial for pregnant women at risk of arrhythmia.

Trial Registration  ClinicalTrials.gov Identifier: NCT02249195

Cardiovascular disease (CVD), the leading cause of maternal mortality worldwide,^1,2 can be explained by several contributing factors, including delaying pregnancy until later in life and an increase in comorbidities (eg, diabetes, hypertension, and obesity). More women with repaired congenital or valvular heart disease reach reproductive age and become pregnant. Persisting structural cardiac abnormalities are associated with arrhythmias, heart failure, and sudden death during pregnancy, with arrhythmias reportedly accounting for up to 31% of cases of maternal death.³

It remains difficult to conduct research (especially blinded randomized clinical trials) during pregnancy. In observational studies, such as the Cardiac Disease in Pregnancy (CARPREG),⁴ Zwangerschap bij Aangeboren Hartafwijkingen (ZAHARA),⁵ and Registry of Pregnancy and Cardiac Disease (ROPAC)⁶ studies, women with arrhythmias were often excluded. Specifically, no study, to our knowledge, has investigated diagnostic approaches for arrhythmias or whether its improved diagnostic yield would affect management and/or pregnancy outcome. This single-center, prospective randomized clinical trial compared implantable loop recorder (ILR) plus 24-hour Holter electrocardiographic (ECG) monitoring vs standard 24-hour Holter ECG monitoring alone in terms of acceptability, ability to identify significant arrythmias, and effects on management and pregnancy outcome in women who were symptomatic (eg, palpitations) or at high risk of arrythmia because of underlying structural heart disease (SHD).

In 2010, a dedicated weekly Cardiac Disease and Maternity Clinic was established at Groote Schuur Hospital, Cape Town, South Africa, to provide multidisciplinary care for pregnant women with suspected or previously known CVD (trial protocol in Supplement 1). Heart failure and maternal mortality were reduced through intensified management.^7,8 Patients were referred from primary or secondary care facilities or within the tertiary hospital (eFigure 1 in Supplement 2). All evaluable women were pregnant (second or third trimester) at the time of consultation and underwent clinical evaluation, ECG, and echocardiography before enrollment. Patients were stratified according to the modified World Health Organization risk classification for pregnant women with cardiac disease.⁹ All data were deidentified. Ethical approval was obtained at the University of Cape Town. All patients provided written informed consent before inclusion. This study followed the Consolidated Standards of Reporting Trials (CONSORT) reporting guideline.

Patients classified to modified World Health Organization risk groups II to IV were included. Patients were eligible if they presented with arrhythmic symptoms (eg, palpitations) or were thought to be at risk of arrhythmia because of underlying SHD. All eligible patients were randomized according to a computer-generated randomization list to receive either standard care (SC) or standard care plus an implantable loop recorder (SC-ILR). At entry, 24-hour Holter monitoring was performed with an ambulatory recorder (SEER Light recorder, General Electric) with a continuous 2-channel ECG recording available for analysis for all patients. Patients randomized to the SC-ILR group also received an ILR reading (REVEAL XT model 9529 ILR, Medtronic). The ILR readings were obtained at clinic visits or when patients were symptomatic. All Holter recordings and ILR readings were reviewed by an electrophysiologist (A.C.). Arrhythmias considered in this study included atrial fibrillation (AF), atrial flutter, premature ventricular complexes (PVCs), supraventricular tachycardia (SVT), ventricular tachycardia (VT), or ventricular fibrillation. Sinus tachycardia, a physiologic adaptation to pregnancy, was not included as an arrhythmia. The Figure shows the CONSORT flow diagram.

Baseline data, including age, parity, medical comorbidities, prior cardiac surgery or interventions, symptoms and signs, New York Heart Association (NYHA) functional class, medication, and ECG and echocardiographic findings were recorded at the first visit. Follow-up data were collected until 42 days after delivery. The frequency of visits depended on severity of disease. Arrhythmias detected by Holter or ILR were managed with treatment deemed to be safe in pregnancy. Mode of delivery and perinatal outcome were obtained. Obstetric events were defined as death, pregnancy-induced hypertension, and (pre)eclampsia. Neonatal events were defined as premature birth (<37 weeks’ gestation), low birth weight (<2500 g), and stillbirth (>20 weeks’ gestation, birth weight >500 g).

Statistical analyses were performed with GraphPad Prism software for Windows, version 7.03 (GraphPad Software Inc). Categorical variables were expressed as frequencies and percentages. Continuous data were expressed as mean (SD) or median (interquartile range) depending on data distribution. Comparison of means and proportions between subgroups at baseline was performed by independent, 2-tailed, unpaired t test and χ² statistics (or Fisher exact test when necessary), respectively, and when data were not normally distributed, a Mann-Whitney test was used. A 2-sided P < .05 was considered to be statistically significant.

The study randomized 20 women to the SC-ILR group and 20 to the SC group (mean [SD] age, 28.4 [5.5] years); however, only 17 consented to ILR insertion (SC-ILR group), and the 3 who declined ILR were allocated to the SC group. Thirty-five women (87%) presented with NYHA functional class I or II. Twenty-three women (57%) had a prior diagnosis of CVD. There was no significant difference between the SC and SC-ILR groups in mean (SD) maternal age (27.1 [4.9] vs 30.3 [5.6] years, P = .07), preexisting CVD (14 [61%] vs 9 [53%], P = .75), NYHA functional class (21 [91%] vs 14 [82%] with class I-II and 2 [9%] vs 3 [18%] with class III-IV, P = .63), blood pressure (systolic: 116 [12] vs 116 [14] mm Hg, P = .92; diastolic: 71 [9] vs 69 [11] mm Hg, P = .55), and mean (SD) gestational age at presentation (<12 weeks: 2 [10%] vs 2 [12%]; 12-24 weeks: 8 [38%] vs 8 [47%]; >24 weeks: 11 [52%] vs 7 [41%]; P = .79). Echocardiographic variables (left ventricular end diastolic diameter: 48.9 [7.9] vs 45.4 [6.1] mm, P = .14; left ventricular end systolic diameter: 34.6 [9.2] vs 33.1 [6.2] mm, P = .20; and left ventricular ejection fraction: 57.6% [13.2%] vs 59.2 [11.7%], P = .69) and rates of SHD (previously operated on CVD: 4 [17%] vs 2 [12%], P > .99) were similar in both groups (Table 1).

The Holter monitor identified arrhythmias (all PVCs) in 3 patients (13%) in the SC group (Table 2). In the SC-ILR group, Holter monitoring detected arrhythmias in 4 patients (24%; persistent AF in 1, PVCs in 2, and nonsustained VT in one), whereas the ILR identified arrhythmias in 9 of the 17 patients (53%; paroxysmal AF in 2, persistent AF in 1, PVCs in 3, and SVT in 4) (eFigure 2 in Supplement 2). In 7 of 9 patients (78%), arrhythmias were detected by the ILR that were not detected by Holter monitoring (4 with SVT, 2 with paroxysmal AF, and 1 with PVCs). All 4 patients with SVT had short, nonsustained episodes of atrial tachycardia or atrioventricular junctional–dependent reentrant tachycardia.

In the SC-ILR group, 3 of 7 patients (43%) had a change in management as a direct result of the ILR recordings (arrhythmias not detected by Holter monitoring). One patient with a perimembranous ventricular septal defect and paroxysmal AF (CHA₂DS₂VASC [congestive heart failure, hypertension, age ≥75 years (doubled), diabetes, stroke/transient ischemic attack/thromboembolism (doubled), vascular disease (prior myocardial infarction, peripheral artery disease, or aortic plaque), age 65-75 years, sex category (female)] score of 1) was given carvedilol for rate control of AF. Another patient with cardiac sarcoidosis with left ventricular dysfunction and paroxysmal AF (CHA₂DS₂VASC score of 1) underwent anticoagulation. Another patient with frequent symptomatic PVCs was admitted for the initiation of β-blockade and underwent investigations that excluded arrhythmogenic right ventricular cardiomyopathy. Three patients with PVCs in the SC group had a low PVC burden and did not require treatment.

The eTable in Supplement 2 summarizes maternal and fetal outcomes per group. No maternal deaths occurred within 42 days postpartum. Mean (SD) gestational stage at delivery was 35 (5) weeks for the SC group and 38 (2) weeks for SC-ILR group (P = .04).

This prospective randomized clinical trial that investigated the use of an ILR compared with 24-hour Holter ECG monitoring for the detection of arrhythmias during pregnancy found that arrhythmias were common in pregnant women with SHD who were symptomatic of arrhythmias (eg, palpitations). Arrhythmias were confirmed in 30% of patients with SHD. In this patient cohort, PVCs, SVTs, and AF accounted for all underlying arrhythmias.

Physiologic changes in pregnancy are known to increase the risk of arrhythmias.^10,11 Arrhythmias may occur for the first time in pregnancy. Pregnancy may also worsen preexisting arrhythmias.⁵ In our study, ILR monitoring significantly increased the yield of arrhythmias during pregnancy compared with 24-hour Holter monitoring alone. Holter monitoring may fail to detect infrequent episodes of paroxysmal AF, which were detected by ILR in this study. The efficacy of ILRs to detect arrhythmias in pregnant patients with SHD seems to be consistent with other population groups, such as patients with cryptogenic stroke and patients with high CHA₂DS₂VASC scores who are at risk of underlying AF.^12,13 The detection of arrhythmias is important because it has previously been reported that women who develop SVT in pregnancy are more likely to have adverse maternal and fetal outcomes.¹⁴

The study found that ILR was an acceptable diagnostic modality (safe and efficacious) in more than 85% of patients. Not only did the detection of arrhythmias by ILR result in a change of management in 43% of patients, but also the SC-ILR group had a significantly longer gestation period than the SC group.

The sample size (N = 40) of this novel study was small. Ideally, Holter monitoring should have been performed for more than 24 hours. However, prolonged ambulatory ECG monitoring is not standard practice universally. Although the study population had a high risk of developing arrhythmias, this was a heterogenous cohort of pregnant women as far as symptoms and underlying SHD were concerned. The significantly higher mean gestational stage at delivery in the SC-ILR group might suggest that factor(s) in addition to ILR monitoring play a role in the outcomes of pregnant women at risk of arrhythmias.

This prospective randomized clinical trial provides information on the feasibility, safety, and efficacy of extended monitoring using an ILR compared with 24-hour Holter monitoring alone in detecting underlying arrhythmias during pregnancy. The ILR detected arrhythmias that were not detected by prior 24-hour Holter monitoring, leading to a change in management for a significant proportion of patients. The preliminary nature of our findings could serve as the basis for future research in detection of arrhythmias in pregnant women with symptoms of arrhythmia or significant SHD. We suggest that the role of prolonged external ambulatory ECG monitoring devices be evaluated and compared with ILR.

Accepted for Publication: December 1, 2019.

Corresponding Author: Karen Sliwa, MD, PhD, Hatter Institute for Cardiovascular Research in Africa, Department of Medicine, Faculty of Health Sciences, University of Cape Town, Chris Barnard Building, Fourth Floor, Anzio Road, Observatory, Cape Town 7530, South Africa (karen.sliwa-hahnle@uct.ac.za).

Published Online: February 19, 2020. doi:10.1001/jamacardio.2019.5963

Author Contributions: Dr Azibani had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Sliwa, Ntsekhe, Chin.

Acquisition, analysis, or interpretation of data: Sliwa, Azibani, Johnson, Viljoen, Baard, Osman, Briton, Chin.

Drafting of the manuscript: Sliwa, Azibani, Johnson, Viljoen, Baard, Briton, Chin.

Critical revision of the manuscript for important intellectual content: Sliwa, Johnson, Viljoen, Osman, Ntsekhe, Chin.

Statistical analysis: Sliwa, Azibani, Viljoen, Chin.

Obtained funding: Ntsekhe.

Administrative, technical, or material support: Sliwa, Viljoen, Baard, Osman, Briton, Ntsekhe.

Supervision: Sliwa.

Conflict of Interest Disclosures: None reported.

Funding/Support: This study was supported by the Medtronic External Research Program, Maastricht, the Netherlands. Medtronic provided the injectable loop recorders free of charge.

Role of the Funder/Sponsor: Medtronic had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Data Sharing Statement: See Supplement 3.