OUP user menu

Pregnancy outcome and management of women with an implantable cardioverter defibrillator: a single centre experience

Pia K. Schuler, Anna Herrey, Angie Wade, Ruth Brooks, Donald Peebles, Pier Lambiase, Fiona Walker
DOI: http://dx.doi.org/10.1093/europace/eus172 1740-1745 First published online: 28 June 2012


Aims With improved survival of patients with congenital and inherited heart disease, there is now a younger cohort of patients with an implantable cardioverter defibrillator (ICD) for the prevention and treatment of ventricular dysrrhythmias. Young women with such disorders often wish to embark on pregnancy, but pregnancy outcome data for this group is sparse. We therefore evaluated pregnancy outcome in patients with heart disease and an ICD in situ.

Methods and results A retrospective analysis was performed on all women with an ICD in situ, who had pregnancy care provided by the specialist maternal cardiology service at University College London Hospitals. Data for 19 pregnancies in 14 women were collected. The underlying cardiac diagnoses were congenital heart disease (one), familial hypertrophic cardiomyopathy (eight), familial dilated cardiomyopathy (one), inherited long QT syndrome (one), and idiopathic cardiac arrest (one). Three women had moderate impairment of the left ventricular systolic function (ejection fraction <45%), in the remainder it was normal. Nine ICD implants were for primary prevention of sudden cardiac death (64%) and five for secondary prevention (36%). Of the 19 pregnancies, 18 continued beyond 24 weeks gestation with 18 live births. In eight pregnancies there were medical or device-related complications (42.9%) as follows: arrhythmias (four) (21.1%), heart failure (two) (9.1%), ICD shocks (one) (5.3%), atrial lead fracture (one) (5.3%), and lead-related thrombus (one) (5.3%). There were no inappropriate device shocks or therapies.

Conclusions Women with heart disease and an ICD implant can have a good outcome during pregnancy but medical and device complications are not uncommon.

  • Implantable cardioverter defibrillator
  • Pregnancy
  • Heart disease


It is now 30 years since the first implantable cardioverter defibrillator (ICD) was used for the treatment and prevention of life-threatening arrhythmias (ARRs). Subsequently, the number of device implants has increased each year,1 with ∼4500 implants in 2008 in the UK alone.2 Although, the majority of implants are in patients with acquired heart disease, the indications for ICD implantation have expanded to include younger age groups, in particular those with inherited and congenital heart disorders.3 For these younger patient cohorts, survival to reproductive maturity and beyond is now the norm,4 leading to a new and unique group of young women with an ICD in situ who wish to embark on pregnancy.

Pregnancy outcome for such patients is limited to case reports and retrospective multi-centre series with small numbers.513 Moreover, the haemodynamic and electrophysiological impact of pregnancy in these patients is also not well documented. We therefore report our single centre experience of pregnancy outcome in women with heart disease and an ICD implant.


All patients with heart disease and an ICD implant, under the care of the specialist multidisciplinary maternal cardiology team at University College London Hospitals NHS Trust (UCLH NHS Trust), between September 2001 and October 2010, were identified from the maternal cardiology database. The specialist team includes input from a grown-up congenital heart disease cardiologist with device expertise, two obstetricians, an anaesthetist, a haematologist, and clinical nurse specialist.

Follow-up during pregnancy

All patients with an ICD have their antenatal care and delivery in the high-risk pregnancy unit at UCLH. Patients are reviewed in the pre-conceptual counselling clinic where possible, or as soon as pregnancy is confirmed. Thereafter patients are reviewed at 14–16 weeks gestation and the frequency review and device interrogation is individualized and determined by symptoms, complexity of the underlying cardiac disorder, and ARR history. At 32–34 weeks a written delivery plan is distributed electronically to all members of the specialist team, labour ward, and to the patient by letter.

Device management during pregnancy, labour, and delivery

During the antenatal period, device parameters remain unchanged from the non-pregnant state unless there is an indication for re-programming e.g. inappropriate device therapies.

The devices were programmed with a ventricular tachycardia (VT) therapy (mean 171–217 per minute) and ventricular fibrillation (VF) zone accordingly. The supraventricular tachycardia (SVT) discriminators employed included rhythm ID for the six Guidant devices, onset and stability algorithms for Medtronic (n= 11) and morphology/stability for the two St Jude devices.

For delivery programming is as follows:

  • Vaginal delivery (VD)—device remains in full therapy mode.

  • Elective caesarean section—device is programmed to ‘monitor only’ (safe mode) to avoid electrical interference from diathermy and re-programmed to full therapy mode immediately postoperatively.14,15

Bipolar diathermy is requested for operative deliveries and a magnet is available on labour ward and in operating theatre for use in event of inappropriate ICD shocks (magnet placement disables ARR detection and prevents shock delivery).

Outcome measures

Maternal, obstetric, and neonatal outcomes were assessed. Maternal cardiac complications including death, heart failure (HF), ARR, thromboembolism (TE), or hospitalization for a cardiac indication were recorded. Device therapies and complications were also documented.


Continuous variables describing the 14 women are summarized by mean and range. Pregnancy outcomes are summarized as percentages within categories and confidence intervals adjusted for multiple pregnancies within women using random effects models using M1Win v2.2.


Baseline characteristics

From a total of 580 pregnancies in women with heart disease, there were 19 pregnancies in 14 women with an ICD implant (Table 1). The mean age at conception was 33 years (range 22–42 years) and the mean time interval between ICD implantation and first pregnancy was 3.8 years (range 1–9 years). The underlying cardiac diagnoses (Figure 1) were congenital heart disease [repaired Tetralogy of Fallot (ToF)] (one), inherited heart disease (nine) [hypertrophic cardiomyopathy (HCM) (seven), HCM with Ross (one)], familial dilated cardiomyopathy (one), long QT syndrome (LQTS) (three), and idiopathic VF arrest (one).

View this table:
Table 1

Demographics and clinical characteristics

Figure 1

Maternal diagnoses. HCM, hypertrophic cardiomyopathy (eight), Tetralogy of Fallot (one); DCM, dilated cardiomyopathy (one); SCA, sudden cardiac death (one); LQTS, long QT syndrome (three).

Nine device implants were for primary prevention of sudden cardiac death (SCD) (9/14 = 64%) [syncope (four), family history SCD (four), non-sustained ventricular tachycardia (nsVT) (one)] and five for secondary prevention (35%) [aborted cardiac arrest (two), sustained VT (three)].

Three patients had moderate impairment of left ventricular systolic function [ejection fraction (EF) <45%], the remainder had normal left ventricular (LV) systolic function.

Drug therapy was used in 18 pregnancies [95% (70, 93%)] including beta-blockers (12, 86%), diuretics (8, 56%), aspirin (7, 50%), and low molecular weight heparin (LMWH) at therapeutic dose (3, 35%).

Thirteen patients had transvenous endocardial leads with pre-pectoral devices and one had an abdominal device with a tunnelled transvenous endocardial lead placement. There were 3 single ventricular lead devices and 11 dual lead (atrial and ventricular lead) systems (Table 2, history of ICDs and ARR s).

View this table:
Table 2

History of implantable cardioverter defibrillator s and arrhythmias

Pregnancy outcome

Of the 19 pregnancies 18 pregnancies continued beyond 24 weeks gestation resulting in 18 live births. There was one termination of pregnancy at 10 weeks gestation for severe foetal chromosomal abnormality. The mean gestational age at delivery was 38 weeks (range 32–40 weeks). There were three pre-term deliveries prior to 36 of 40 weeks, for symptomatic palpitations (one), left ventricular failure (one), and spontaneous rupture of membranes (one). The mean birth weight was 2.9 kg (range 1.6–3.8 kg). The mode of delivery was per vagina (VD) in 12 [69%, confidence interval (CI) 43, 87%] of the 17 pregnancies for which mode of delivery was recorded (one termination, one lost to follow-up). The remaining five were by caesarean section (CS). Three CS were elective [breech presentation (one), symptomatic sinus tachycardia (one), and previous CS (one)] and two were emergency sections for failure to progress in labour with foetal distress.

Maternal outcome

Deaths—There were no maternal deaths.

Heart failure (two)—HF was precipitated in two pregnancies (one patient). In one patient with non-obstructive hypertrophic cardiomyopathy HF was precipitated during both her first (P1) and second (P2) pregnancies during the second trimester (9.1% CI 1.3, 43.5%). Her baseline LV function was impaired with EF 40–45% prior to both P1 and P2. It remained stable throughout both pregnancies and she responded well to diuretic therapy on both occasions, with uncomplicated induced vaginal deliveries at 36 and 37 weeks, respectively.

Thromboembolism (one)—There was one major TE complication of ICD lead thrombus in a patient with HCM and an ICD for primary prevention of SCD (5.3% CI 0.7, 30.3%). Biventricular function was good but routine echocardiography [transthoracic echocardiography (TTE)] during the second trimester identified a 13 × 15 mm thrombus attached to the ventricular lead in close proximity to the tricuspid valve. She was treated with Dalteparin (8500 IU bd). A thrombophilia screen diagnosed factor V Leiden deficiency (heterozygous) and a homozygous C677T MTHFR polymorphism. On further enquiry a positive family history of venous thromboembolism (VTE) was identified. She remained well for the remainder of pregnancy and the thrombus did not change in size. She underwent elective CS at 36 weeks gestation (previous CS). The evening dose of LMWH was omitted the day prior to CS, prophylactic dose LMWH was re-started 5 h post-CS, and therapeutic dosing 24 h post-operatively. One week post-partum she presented with leg swelling and breathlessness. TTE confirmed an increase in thrombus size at 41 × 18 mm with obstruction to tricuspid valve inflow (Figure 2). She underwent urgent surgical removal of the thrombus and ICD system (generator and leads) via sternotomy and made an uneventful recovery.

Figure 2

Implantable cardioverter defibrillator lead thrombus: thrombus (*) straddling at the implantable cardioverter defibrillator lead, obstructing the inflow of the right ventricle.

Arrhythmia episodes, device therapies, device-related complications

There were four ARR episodes (21.1%, CI 8.1; 44.7%), Table 3. Two episodes were symptomatic and two were detected on device interrogation only. The symptomatic ARR s were as follows: VF (one) and SVT (one). The patient with VF (patient 13) had LQTS and had an appropriate device shock of 31 J at 20-week gestation. Although the ICD implant was for primary prevention, she had been non-compliant with beta-blocker therapy during pregnancy. There were no adverse maternal or foetal effects and she had an uncomplicated delivery at 37-week gestation. Of note she experienced a second appropriate ICD shock 3-month post-partum, again having been non-compliant with beta-blocker treatment. There were no inappropriate device shocks or therapies during any pregnancy or for up to 6-month post-partum. The symptomatic SVT was in patient 5 with HCM, and it was appropriately detected by the device without need for device therapy. She subsequently commenced beta-blocker therapy. There were two asymptomatic ARR episodes detected by the ICD of sustained VT (patient 8) and nsVT (patient 9), respectively. The sustained VT occurred in two salvos during the third trimester in a patient with HCM and was detected at the time of device interrogation prior to CS. Both salvos were successfully treated with anti-tachycardia pacing (ATP). The episode of nsVT occurred in a HCM patient in the third trimester, with appropriate device detection and monitoring.

View this table:
Table 3

Total arrhythmias during pregnancies

One non-pacemaker-dependent patient with HCM and a previous Ross surgery developed high atrial lead impedance in the second trimester. An atrial lead fracture was confirmed (5.3%, 95% CI 0.7; 30.3%). The device was programmed from DDD to VVI without further complications.


Owing to improved survival in patients with congenital and inherited heart disease,4 there is a new and unique cohort of women reaching reproductive age,16,17 some with an ICD in situ for the treatment and prevention of ventricular dysrhythmias and SCD. Published pregnancy outcome data for these patients are minimal10,12,18 and yet their general acceptance of risk for the experience of motherhood, means there are likely to be increasing number of such patients. Our single centre experience demonstrates that pregnancy outcome overall is generally good but medical and/or device complications are frequent, occurring in 8 of the 19 pregnancies (42.9%; CI 21.4; 67.5%). The most serious complication was a thrombotic event, with the development of a thrombus on a ventricular lead in a patient with previously undiagnosed thrombophillia. Arrhythmia episodes were also not uncommon (21.1%, CI 8.1; 44.7%) especially as most device implants in this cohort were for primary prevention; however, all ARR s were appropriately detected and/or treated by the device and there were no inappropriate device therapies.

With the additional pro-arrhythmic and haemodynamic changes of pregnancy, one might expect an increased incidence of ARR and/or device therapies in these pre-disposed patients.1921 Even in the non-pregnant state there is an appropriate ICD shock rate of 22% over 2 years, in the congenital heart disease population, with one study reporting an annualized shock rate of 7.7% in adults with ToF with an ICD for primary prevention of SCD.22 Similar rates of appropriate device shocks (4% per annum) are reported in patients with familial HCM in a population with an ICD for primary prevention (86%).23,24 In our cohort there was one appropriate device shock for VF in a patient with LQTS. There were no adverse maternal effects and the foetus as expected was unaffected (the foetal myocardium has a high fibrillation threshold and very little shock energy is transmitted to the foetus25,26). Our low ICD shock rate contrasts the largest reported pregnancy series of Natale et al., where 18% (8 of 44 patients) experienced a device shock.10 This difference in shock rate is most likely explained by the fact that in Natale's series all of the device implants were for secondary prevention of malignant ARR or SCD (25 and 75%, respectively) compared with only 35% in our own cohort. Of note there was no difference in use of anti-arrhythmics between the two series at 57 and 58%, respectively, although perhaps a higher rate of anti-arrhythmic drug therapy might have been expected in their higher risk cohort. Although drugs are best avoided during pregnancy,27 many anti-arrhythmics are FDA category B or C, and generally considered safe28,29 [with the exception of Amiodarone and Atenolol (category D)]. The benefit of continuing therapy therefore more often than not outweighs risk, particularly when the indication is life-threatening ARR. One patient in our study highlights the importance of adjunctive beta-blocker therapy specifically in LQTS. Her ICD implant was for primary prevention of SCD but she experienced an appropriate device shock for sustained VT/VF on two occasions, both of which were temporally related to non-compliance with beta-blockers.

There were no inappropriate device shocks in our patient cohort. This is reassuring, particularly in light of the fact that there is a background inappropriate device discharge rate of 21% per year in those with congenital heart disease3,30 and 5% in patients with familial HCM23 in the non-pregnant state. Moreover in 14 of our 19 pregnancies (74%, CI 47, 90%) there were risk factors for inappropriate ICD shocks (<35 years old, prior history of atrial fibrillation). Our low inappropriate shock rate may be explained by the fact that all ARR s even if non-sustained are treated promptly, and euvolaemia is maintained with the use of concomitant diuretics. Diuretics were continued and or up-titrated in four pregnancies (28%) and newly prescribed in one pregnancy, in order to treat symptoms of dyspnea. In HCM patients such symptoms often occur in the absence of clinically manifest signs of pulmonary congestion but represent increasing pulmonary venous hypertension, which responds well to small doses of furosemide.

Lead complications31,32 occurred in two patients in our cohort (one lead fracture and one lead thrombus; 5.3%, CI 0.7, 30.3%). Such complications are often difficult and potentially hazardous to treat in pregnancy.24 Lead explantation and re-implantation has an associated procedural risk (radiation, wound infection, tamponade), which must be carefully considered and balanced against the risks of a conservative approach (device re-programming and medical therapy), which may also be efficacious. Such management decisions are case specific, but for our patient with lead fracture, ventricular lead function was normal and device re-programming was possible. The second more serious device-related complication of lead thrombus is reportedly not uncommon with a 25% incidence (detected by transoesophageal echocardiography) in one study26 but thrombi are usually small and subclinical.3335 Our patient had an undiagnosed thrombophilia (Factor V Leiden deficiency) and developed a large thrombus on the ventricular lead (Figure 2), which increased in size despite therapeutic anti-coagulation, eventually causing obstruction to tricuspid valve inflow. The patient had been non-compliant with aspirin therapy but whether this treatment alone would have prevented thrombus formation in the presence of this inherited thrombophilia is doubtful. Her case highlights the need for a detailed history and VTE risk assessment in all patients with heart disease, but especially in those with intra-cardiac hardware36,37 because LMWH treatment may then be indicated.3840

The majority of pregnancies in our series were delivered vaginally which is the preferred mode for the majority of women with heart disease41 (69%, CI 42, 87%; 0.88). Of the five CS deliveries, four were for obstetric reasons and only one for a cardiac indication (dyspnoea and persistent sinus tachycardia). In our series patients with an ICD implant had a planned date for delivery, which allowed appropriate personnel to be available should the device need to be re-programmed in the event of an operative delivery. In the event of an emergency CS, the device remains in full therapy mode with bipolar diathermy used for cautery. A magnet is also available in the operating theatre for de-activating the device in the event of inappropriate device shocks.14,15,41


The outcome of pregnancy in women with heart disease and an ICD implant is good, but medical and or device complications are frequent. They therefore represent a high-risk pregnancy group who require specialist multidisciplinary team antenatal care.


P.K.S. has research grants from the European Society of Cardiology and the Swiss National Foundation.


We wish to thank Adam Simpson, cardiac electrophysiologist, for help with the ICD programming evaluation.

Conflict of interest: none declared.


View Abstract