Implantable Cardioverter-Defibrillators and Pregnancy
A Safe Combination?
Background The purpose of this investigation was to evaluate the outcome of pregnancy in women with implantable cardioverter-defibrillators (ICDs).
Methods and Results A multicenter retrospective analysis was performed on women with an ICD who became pregnant. Data were collected on 44 patients. The mean age and ejection fraction at the time of the initial ICD implant were 25.6±4.9 years and 49.8±9.7%, respectively. The mean follow-up after the ICD implant was 4.8±2.8 years. Forty-two women had abdominally implanted generators, and 2 had a prepectoral device. Thirty had epicardial lead sensing systems, and 14 had transvenous. Thirty-six (82%) experienced no complications, and 8 (18%) had a medical or device-related complication. The ICD-related problems included tenderness at the ICD pocket scar (2 patients), generator migration (1), and pericarditis secondary to the epicardial patches (1). Medical complications were pulmonary embolism (1), therapeutic abortion (1), worsening hyperthyroidism (1), congestive heart failure (1), and weight loss (1). Thirty-seven women delivered vaginally, and 7 underwent cesarean section. Thirty-nine babies were born healthy, 1 was stillborn, 2 were small for gestational age, 1 had transient hypoglycemia, and 1 woman had a therapeutic abortion unrelated to the ICD. During pregnancy, 33 women received no ICD therapy, 8 had 1 shock, 1 had 5 discharges, 1 had 11 shocks, and 1 had 5 shocks. The total number of shocks during pregnancy ranged from 0 to 11, with an average of 0.66±1.9 discharges. Five women had 7 additional pregnancies without an ICD shock.
Conclusions The mere presence of an ICD should not defer a women from becoming pregnant unless she has an underlying structural cardiac disease that is considered a contraindication. Pregnancy does not increase the risk of major ICD-related complications or result in a high number of ICD discharges.
For young women with life-threatening cardiac arrhythmias, the use of the implantable cardioverter-defibrillator (ICD) has allowed an increased number of them to reach their reproductive years. Although cardiac disease complicates ≈1% of all pregnancies,1 only a small portion of these are represented by women with arrhythmias. The effects of various pharmacological agents to treat arrhythmias during pregnancy has been evaluated2 ; however, little is known regarding the outcome of pregnancy in women with ICDs. In fact, only two cases have been reported of women with ICDs who have successfully completed a pregnancy.3 4 Given this, no general guidelines or recommendations are available for physicians to offer patients with ICDs who want to become pregnant.
The purpose of this study was to evaluate a large cohort of women who had had an ICD implanted because of a malignant ventricular arrhythmia to determine whether pregnancy was associated with an increased risk of maternal and fetal complications, an increased number of ICD discharges, and a higher risk of lead- and generator-related complications. A multicenter joint project was therefore undertaken to address these issues.
Any woman of child-bearing age who had an ICD implanted for a life-threatening ventricular arrhythmia either before or during a pregnancy was enrolled in the study. Although this analysis was principally retrospective because most women had completed one or more pregnancies at the time of entrance, several were followed prospectively through delivery. Private and academic centers caring for such patients were identified by telephone. A total of 93 institutions with a large population of patients with ICDs were contacted. Thirty-eight reported having cared for 1 or more such patients. Although a total of 54 patients meeting entry criteria were identified, data were collected on only 44 patients.
Data were collected in the form of questionnaire completed by the treating electrophysiologist or other designated health care personnel. The information that was requested was age and ejection fraction at time of initial ICD implantation and current age and ejection fraction; indication for ICD implantation; other cardiac and medical disorders; device information (generator model, implant position, lead sensing system [transvenous versus epicardial, number of generator and lead replacements]); total number of pregnancies and number of pregnancies with an ICD in place; time of device implantation in relationship to pregnancies; mode of delivery; device status at time of delivery; maternal and fetal complications; number of ICD firings during each pregnancy, delivery, and 24- to 48-hour postpartum period; total number of ICD discharges in a lifetime; and antiarrhythmic medication taken before and during pregnancy.
The number of ICD discharges in a lifetime does not include shocks given during device testing or pregnancy. Similarly, in the calculation of the total number of shocks received during pregnancy, the discharges administered when an ICD was implanted or a generator was replaced during pregnancy were not included. However, shocks given for device testing during pregnancy were included in the assessment of the safety of ICD firing on fetal outcome.
Each ICD was implanted according to the standard practice of the electrophysiologist performing the procedure; follow-up was also according to local practice. Each pregnancy was considered high risk, and care was provided by both the obstetrician and electrophysiologist accordingly.
Data are presented as mean±SD. The χ2 test was used to determine the statistical significance between proportions.
General demographic and clinical characteristics are summarized in the Table⇓. Forty-four women with an ICD who completed a pregnancy or was currently pregnant were enrolled in the study. Their collective current average age is 30.4±4.4 years. The mean age at the time of first ICD implantation was 25.6±4.9 years, and the mean follow-up since that implantation was 4.8±2.8 years. The mean ejection fraction at the time of the first ICD implantation was 49.8±9.7%. The present mean ejection fraction is 51.4±9.5%. Their underlying cardiac diseases were categorized into long-QT syndrome (13 patients), idiopathic ventricular fibrillation (17), cardiomyopathy (8), congenital heart disease (3), coronary artery disease with an ischemic cardiomyopathy (1), idiopathic hypertrophic subaortic stenosis (1), and right ventricular dysplasia (1). The clinical presentation and indication for ICD implantation were sudden cardiac death in 33 patients, ventricular tachycardia in 9 patients, and ventricular tachycardia with syncope in 2 patients. Pertinent past medical history in this cohort included the following: left stellectomy (patient 1), pacemaker implantation (patients 3, 6, and 20), atrioventricular (AV) node modification for AV node reentry tachycardia (patient 34), mitral valve prolapse (patients 5 and 9), mitral regurgitation (patients 11 and 33), cardiac surgery for congenital or coronary disease (patients 19, 20, and 30), recurrent pericarditis (patient 24), collagen vascular disorder (patient 25), and hyperthyroidism (patients 23 and 30).
A wide range of ICDs were implanted in this population, including those from CPI (models 1500, 1520, 1550, 1600, PRX 1700, AID-BR), Ventritex (model V100), Medtronic (models 7217B, Jewel 7219D, 7216) and Teletronics (model 4203). The total number of generator replacements averaged 1.9±1.2 and ranged from one to six. Only 2 patients had the generator implanted in the prepectoral position (patients 10 and 37); all others were located in the abdomen. Thirty patients had epicardial lead sensing systems, and 14 had transvenous lead systems.
Device and Lead Complications During and After Pregnancy
Two patients (patients 18 and 28) reported mild tenderness at the ICD pocket scar as the abdomen expanded; 2 patients experienced migration of the generator onto the iliac bone (patient 30) during pregnancy and into the left lower abdominal quadrant (patient 42) nearly 3 years after the pregnancy; and 1 patient had pericarditis attributed to the epicardial patches (patient 17). No generator erosion or lead fractures were reported during pregnancy or the postpregnancy mean follow-up of 2.4±1.2 years.
The total number of pregnancies after ICD implantation by 1 woman ranged from one to three, with the majority (39 patients) having had only one pregnancy. With reference to the first pregnancy after ICD implantation, 36 women (82%) had an uneventful experience, and 8 (18%) had a medical or device-related complication. In terms of the medical complications, 1 had a pulmonary embolism during the fifth month of pregnancy (patient 2); 1 had an induced therapeutic abortion, unrelated to the presence of the ICD or an arrhythmia (patient 4); 1 developed worsening signs and symptoms of preexisting hyperthyroidism, resulting in tachyarrhythmias that abated after adjustment of her thyroid medication (patient 30); 1 with a history of a cardiomyopathy developed congestive heart failure with a reduction in her EF, all of which resolved after delivery (patient 21); and the patient with severe coronary artery disease had a 9-lb weight loss (patient 20). In addition, 2 women had the ICD implanted while pregnant (patients 6 and 15), and 1 had the generator replaced during the fifth month because it was at the end of life (patient 17).
Thirty-seven women delivered vaginally, and 7 underwent cesarean section. The indications for cesarean section included cephalic disproportion (patient 6), fetal distress (patient 24), repetitive variable heart rate decelerations (patient 22), oligohydramnios (patient 17), severe coronary artery disease such that labor or induction was thought to be too stressful (patient 20), failed induction (patient 34), and unknown (patient 5). The device was on during delivery in 28 women and was off in 16 patients. In only 1 of the women who had a cesarean section, the device was left on; in all others, it was turned off.
Regarding fetal outcome after the first pregnancy, 37 babies were born healthy; 2 others were born healthy but with long-QT syndrome and are doing well on β-blocker therapy (mothers 10 and 12); 1 was stillborn due to umbilical cord strangulation, but it was known that the heartbeat was absent the day before delivery (mother 15); 2 were small for gestational age, which was thought to be secondary to antiarrhythmic therapy (mothers 24 and 41); 1 had transient hypoglycemia due to sotalol the mother had been taking (mother 17); and 1 woman had a therapeutic abortion (patient 4). Among the 11 women experiencing shocks during pregnancy, 9 babies were born healthy, 1 was stillborn (mother 15) and proven to be healthy after the last ICD shock, and 1 had transient hypoglycemia thought to be secondary to sotalol therapy.
Five women had seven additional pregnancies. Three women had an uncomplicated second pregnancy and delivered healthy children (patients 15, 16, and 43). The patient who had an abortion during the first pregnancy became pregnant two additional times (patient 4). The second was complicated by nausea, vomiting, and weight loss secondary to amiodarone. At 5 weeks before the estimated date of delivery, she experienced spontaneous rupture of the membranes and delivered the baby. The child had a ventricular septal defect and experienced transient sinus bradycardia after birth attributed to the amiodarone. Her third pregnancy resulted in a spontaneous abortion unrelated to the presence of the ICD or an arrhythmia. The remaining patient also had two additional pregnancies (patient 7). The second was uncomplicated, and the baby was born healthy; the third pregnancy ended in a miscarriage, but this was unrelated to the ICD or an arrhythmia. No ICD discharges occurred during any of these subsequent pregnancies.
During the first pregnancy, 33 women experienced no ICD discharge, 8 received a single shock (patients 2, 6, 8, 11, 14, 17, 21, and 35); 1 experienced 5 firings in response to new onset atrial fibrillation (patient 15); and 2 had 11 and 5 discharges, respectively, for monomorphic ventricular tachycardia (patients 20 and 44). The latter were the patients with severe coronary artery disease and dilated cardiomyopathy. In 3 women, 1 of the ICD shocks received during pregnancy was for device testing at the time of ICD implantation (patients 6 and 14) or generator replacement (patient 17). During delivery, in the women in whom the ICD remained active, none received any shocks. External defibrillation was not required in the 16 in whom the ICD was turned off. In the 24- to 48-hour period after delivery, 1 patient experienced an ICD discharge for ventricular fibrillation (patient 8); all others had no firings. Overall, the total number of ICD discharges during pregnancy ranged from none to 11, with an average of 0.66±1.9 shocks (0.07 shock per month). The total number of discharges in a lifetime for this population ranged from none to 42, with a mean of 3.1±6.7 shocks. Given an average follow-up period of 4.8 years minus the 9 months of pregnancy, this averages to 0.79 shock per year (0.06 shock per month).
Thirty-one women had been on one or more antiarrhythmic medications before becoming pregnant, and 14 had never taken any pharmacological therapeutics. During pregnancy, 25 women took antiarrhythmic drugs, and 19 took none. The therapies used included β-blocker alone (16 patients), β-blocker and digoxin (4), β-blocker and mexiletine (1), quinidine and digoxin (1), sotalol (2), and amiodarone (1). Two of the women on β-blocker therapy took it for only a brief duration of their pregnancy (patients 30 and 31). As shown in the Table⇑, for women on antiarrhythmic medications while pregnant, either the mother or the fetus experienced additional medical complications. The complications that could be attributed to the antiarrhythmic therapy included weight loss by the mother (patient 4), ventricular septal defect and sinus bradycardia of the baby (patient 4), spontaneous abortion during pregnancy number 3 (patient 4), oligohydramnios and transient fetal hypoglycemia (patient 17), variable fetal heart rate deceleration (patient 22), fetal distress (patient 24), and small fetal size (patients 24 and 41).
In women with congenital or other structural heart disease, pregnancy outcomes have been shown to be related to the functional cardiac status, pulmonary vascular resistance, and degree of residual impairment after surgical correction. In general, young women with ICDs are advised against becoming pregnant because it has not been known what effects the ICD may have on the pregnancy or how pregnancy might affect the device, its function, or the underlying rhythm disorder. However, this clinical experience of 44 women with ICDs revealed that the majority completed and tolerated pregnancy and delivery without serious complications. In our series, the incidence of cesarean section, spontaneous abortion, preterm delivery, and fetal growth retardation were similar to that expected in the general population.5 Therefore, a woman who is in clinically stable condition and desires to have a child should not be routinely discouraged from doing so based on the presence of the ICD. An increased risk of device-related complications due to the pregnancy itself was not evident. Despite the concern regarding generator erosion as the abdomen expands, 96% of the women in this study had abdominally implanted generators and not one erosion occurred. Regarding the one generator migration, we can postulate that this may have been due to the increasing abdominal girth, but other causes cannot be excluded. Nevertheless, these problems will become less important in the future as more generators are implanted in the prepectoral position.
A second major concern has been that of lead fracture, which could conceivably occur with either the epicardial or transvenous systems. As the diaphragm elevates and the abdominal girth expands secondary to fetal growth, stress increases on the epicardial or transvenous system. In addition, the contractions associated with labor are intense enough to affect the muscles located in the upper torso and shoulder and therefore could apply stress particularly to the transvenous system. Despite this, however, no lead fractures were reported, not even in the 14 patients with a transvenous system.
Women who are otherwise healthy may experience an increased incidence of supraventricular tachycardias during pregnancy,6 most of which are due to reentrant mechanisms.7 The same can apply to ventricular tachycardia, which may be triggered during pregnancy as a result of hemodynamic changes and autonomic nervous system alterations8 9 In our cohort in whom there was an underlying electrical disorder or structural cardiac disease before pregnancy, it might be expected that the patients would experience an increased number of ventricular arrhythmias and hence more frequent ICD firings due to the stress of the pregnancy. However, no increase in ICD discharges compared with baseline occurred in this population when reviewed collectively. When assessed on an individual basis, only 3 women experienced more shocks during their pregnancy compared with their nonpregnant state. The first was the woman with new-onset atrial fibrillation. Since the implant of the ICD, she had five shocks while pregnant as a result of inappropriate discharging of the device for the atrial arrhythmia. The remaining 2 were patients with severe structural heart disease who had a total of 15 and 13 firings, 11 and 5 of which were for monomorphic ventricular tachycardia during pregnancy. It is possible that the diffuse electrical disorder present in the majority of patients in this study is probably less affected by the physiological changes of pregnancy and could explain why an increased number of ICD firings did not occur.
Surprisingly, the hormonal and autonomic changes and the strong uterine contractions during delivery did not precipitate any arrhythmias or ICD firings. In this respect, even though the status (on/off) of the ICD at the time of delivery appears to have no effect on the overall outcome, we recommend leaving the device “on” during vaginal deliveries. After all, if an arrhythmia develops, more prompt therapeutic intervention can be delivered than that by external defibrillation. Certainly, in the case of the cesarean section, the device must be “off” because cautery is involved.
As far as the safety of the fetus is concerned, recurrence of ventricular tachycardia could be dangerous because it may result in hypotension and hypoperfusion and the energy of the ICD shock could be shunted to the fetal heart, with unpredictable effects. However, in the 11 women in whom ICD therapy was delivered, no adverse fetal outcomes were evident as a result of the discharges. It is conceivable that hemodynamic changes are only transient due to the rapid termination of ventricular tachyarrhythmias by the ICD. In addition, it is unlikely that ICD firings could cause life-threatening fetal arrhythmias given that the fetal heart has a high fibrillation threshold and that the amount of current reaching the uterus should be small because therapy from internal defibrillation is very directed.10
Although it is undesirable for any woman to take medication while pregnant, 25 (56%) of the patients in this study were on antiarrhythmic therapy. This, however, is not surprising considering this patient population is more ill given their various underlying electrical and structural heart disease. Of interest, more patients on antiarrhythmic drugs experienced one or more medical complications during their pregnancy or the fetus suffered some minor adverse outcome. Whether these women were simply more ill and therefore more likely to suffer complications during pregnancy cannot be excluded. It may be advisable to consider and attempt discontinuing any antiarrhythmic medication that the patient may be taking before or during the pregnancy to avoid these adverse effects.
To our knowledge, this is the largest clinical series reported on pregnancy outcomes in ICD recipients. Based on our results with 44 patients, the presence of an ICD poses negligible risks to a pregnant woman or her fetus. Therefore, it seems unreasonable to prohibit pregnancy in patients with an ICD unless the decision is based on the underlying cardiac status and severity of the structural heart disease. Such patients should be made aware of the clinical experience to date, which will help them to reach an informed decision regarding pregnancy. If vaginal delivery is anticipated, the ICD should remain active as it poses no additional risks and can provide therapy quickly if needed. The number of women reaching childbearing years and desiring to have children after ICD implantation will continue to increase. This analysis should provide useful guidelines for counseling women with preexisting ICDs about their prospects for a successful pregnancy and the effect of pregnancy on their underlying rhythm disorder.
Kelly Anderson, MD, Barb Miklos, RN, University of Pittsburgh Medical Center (Pa); Robert F. Hanich, MD, Debbie Jones, RN, Asheville Cardiology Associates, Asheville, NC; Gust Bardy, MD, Jill Anderson, RN, University Hospital, Seattle, Wash; John M. Herre, MD, Barbara McGuire, RN, Sentara Norfolk General Hospital, Norfolk, Va; Hugh Calkins, MD, Misty Fales, RN, John Hopkins Hospital, Baltimore, Md; L. Bing Liem, MD, Jan Peterson, RN, Falk Cardiovascular Research Center, Stanford, Calif; Robert Chilton, Md, Audie Murphy VA Medical Center, San Antonio, Tex; Frank Marcus, MD, Cathy Gean, RN, University of Arizona/Health Science Center (Tucson); John DiMarco, MD-PhD, University of Virginia Health Sciences Center (Charlottesville); Thomas Mattioni, MD, Sue Welch, RN, Arizona Heart Institute (Phoenix); David Fitzgerald, MD, Nancy Sherrill, RN, Bowman Grey/Baptist Hospital, Winston-Salem, NC; Raul Mitrani, MD, J. Koehlinger, RN, Krannert Institute of Cardiology, Indianapolis, Ind; Charles Gottlieb, MD, Mary Alice Roth, RN, Philadelphia Heart Institute, Philadelphia, Pa; Brian Olshansky, MD, Loyola University Medical Center, Maywood, Ill; Thomas Guarnieri, MD, MidAtlantic Cardiovascular Consultants, Baltimore, Md; Ennio Pisanó, MD, Raffaele Fanelli, MD, Division of Cardiology, Casa Sallievo della Sofferenza, San Giovanni Rotondo, Foggia, Italy; Michael Reiter, MD, University of Colorado USC (Denver); Paul Troup, MD, Birmingham, Ala; Jeffrey Rottman, MD, Jewish Hospital, St Louis, Mo; Paul Walter, MD, Nancy Romeiko, RN, Emory University Hospital, Atlanta, Ga; Rodolphe Ruffy, MD, University of Utah (Salt Lake City); Roger Winkle, MD, Paula Smith, RN, Cardiovascular Med. and Cardiac Arrhythmias, Palo Alto, Calif; Jeremy Ruskin, MD, Mary Guy, RN, Massachusetts General Hospital, Boston, Mass; Stephen Winters, MD, Debra Hibbard, RN, Morristown Memorial Hospital, Morristown, NJ; William Sanders, MD, Peg Herbst, RN, University of North Carolina (Chapel Hill); Mark Wish, MD, Kim Hill, RN, Arrhythmia Associates, Fairfax, Va; David Steinhaus, MD, Debbie Cardinal, RN, Mary Waters, RN, Mid America Heart Institute, Kansas City, Mo; Adam Strickberger, MD, University of Michigan Medical Center (Ann Arbor); Charles Swerdlow, MD, Kim Dawson, RN, Cedars-Sinai Medical Center, Los Angeles, Calif; Jose Texiera, MD, Cardiology Associates, Rapid City, SD.
↵1 Coinvestigators are listed in “Appendix.”
- Received February 27, 1997.
- Revision received June 5, 1997.
- Accepted June 19, 1997.
- Copyright © 1997 by American Heart Association
Davies GAL, Herbert WNP. Cardiac disease in pregnancy. Cardiol Rev. 1995;3;5:262-272.
Cox JL, Gardner MJ. Treatment of cardiac arrhythmias during pregnancy. Prog Cardiovasc Dis. 1993;2;36:137-178.
Isaacs JD, Mulholland DH, Hess LW, Albert JR, Martin RW. Pregnancy in a woman with an automatic implantable cardioverter defibrillator: a case report. J Reprod Med. 1993;36;6:487-488.
James DK, Stirrat GM. Pregnancy and Risk: The Basis for Rational Management. New York, NY: John Wiley & Sons; 1988.
Brodsky M, Doria R, Allen B, Sato D, Thomas G, Sada M. New-onset ventricular tachycardia during pregnancy. Am Heart J. 1992;123;4:933-941.
Page RL. Treatment of arrhythmias during pregnancy. Am Heart J. 1995;130;4:871-876.