Pregnancy in cardiac transplant recipients

Purpose: Successful pregnancy following cardiac transplantation has been described, although outcome data from individual centers are relatively sparse. We investigated maternal and fetal outcomes including change in left ventricular (LV) function and calcineurin inhibitor (CNI) dose in women who became pregnant from our institution. Methods: We identified every female patient <49 years at the time of transplant who survived >3 months post‐surgery, between 1985 and 2014. Those who conceived had a review of their medical records and transplant charts. Those currently alive were interviewed. Results: There were 22 pregnancies in 17 women with 20 live births (91%). Mean time from transplantation was 98±62.4 months. Rejection complicated one pregnancy, and LV function remained normal in all others. Hypertension complicated 3 (13.6%), preeclampsia 3 (13.6%), and cholestasis 1 (4.5%). Mean birthweight was 2447±608 grams at 34.1±3.6 weeks. Four women died following pregnancy. A significant increase in total daily dose of tacrolimus and cyclosporine A was required to maintain therapeutic levels through pregnancy (CyA, P<.001; and Tac, P=.001), with no deterioration in serum creatinine. Conclusions: We report a 91% live birth rate post‐cardiac transplantation. Meticulous individualized care with frequent monitoring of CNI levels and LV function is necessary to optimize the maternal and fetal outcomes.

calcineurin inhibitor doses in pregnancy; cardiac transplantation; fetal outcome; maternal outcome; pregnancy; pregnancy after cardiac transplant; pregnancy after heart transplant

Between 2003 and 2013, over 4500 cardiac transplants were performed worldwide in women of childbearing age (personal correspondence, based on the International Society of Heart and Lung Transplantation [ISHLT] transplant registry data, March 2015). Females account for approximately 25% of heart transplant recipients with a 5‐year survival of 69%.[1] Additionally, there is a growing population of surviving young women with complex congenital heart disease and end‐stage heart failure being referred for consideration of transplantation.[2] Following transplantation, these women frequently regain fertility and may wish to consider pregnancy. Pregnancy in a heart transplant recipient is complex, with significant risks to the mother, graft, and fetus. Current ISHLT guidelines recommend avoiding pregnancy in the first 12 months following transplantation due to the higher risks of rejection and the aggressive immunosuppression regimen.[3]

Currently, the only active registry of pregnancy outcomes following solid organ transplantation is the National Transplant Pregnancy Register (NTPR) in the United States. The United Kingdom (UK) transplant pregnancy registry which was established in 1997 ceased to collect data in 2001.[4] Although several series have reported pregnancy outcomes in both single‐ and multicenter settings,[5] , [6] , [7] , [8] , [9] , [10] there is a paucity of outcome data from individual centers within the UK. Our center is one of only two within the UK performing pediatric cardiac transplantation, and transplantation in children and adults with complex congenital heart disease. Transplanted patients are followed in our program through adult life. The aim of this study was to report our center's experience of pregnancy in heart transplant recipients. Specifically, we aimed to report maternal and fetal outcomes, investigate rates of rejection, and document the changes in daily dose of calcineurin inhibitors (CNI) required to maintain therapeutic levels through the trimesters of pregnancy.

We identified female patients below the age of 49 years at the time of transplantation who survived >3 months post‐surgery between 1985, when our heart transplant program began, and 2014. Through a search of our local database, we identified every woman who had a pregnancy following transplantation. The medical records and transplant charts of these women were reviewed, and those currently alive were interviewed. Clinical outcomes, changes in serum creatinine (Cr), pregnancy‐related comorbidities, daily dose, and corresponding trough levels of CNIs were recorded. Rejection was defined as a fall from baseline left ventricular systolic function during pregnancy where no other cause was identified.

This study was considered an audit of clinical activity. All patient data were completely anonymized. According to our local research and development policy, no application for ethical approval was required.

SigmaPlot for Windows version 11.0 (Systat Software Inc., San Jose, CA, USA) and Microsoft Excel 2007 were used for data analysis. Where data were normally distributed, results are presented as means±standard deviations (SDs). Where data were not normally distributed, medians and interquartile ranges (IQRs) are reported. Groups were compared using Student's t‐test or Mann–Whitney rank‐sum test for numerical variables. To evaluate trends in results over time using matched data, repeated‐measures analysis of variance (ANOVA) on ranks or Friedman's repeated‐measures ANOVA was used. A P value of <.05 was considered statistically significant.

Over the 29‐year period, 1017 cardiac transplants were performed in 1000 patients; 304 were women and 141 met inclusion criteria. Of these, 17 women had 22 pregnancies resulting in 20 live births (91%). Four of the 22 pregnancies were planned (18%). One woman (6%) required fertility treatment to conceive. Four women (18%) had a second pregnancy.

The mean maternal age at the time of conception was 25.3±5.8 years (range 17–33 years). The mean time from transplantation to conception was 98±62 months (range 10–209 months). Six women were transplanted in childhood or adolescence. The indication for cardiac transplantation was dilated cardiomyopathy in 10 (59%) and congenital heart disease in 7 (41%). All women had normal left ventricular systolic function by transthoracic echocardiography prior to pregnancy. Five women were hypertensive prior to pregnancy, one woman had had an episode of rejection prior to pregnancy, and one had undergone combined heart and liver transplant. One woman had renal dysfunction secondary to CNI treatment, and one was on medication for epilepsy. Of the 17 women, 13 (76%) are currently alive. Table [NaN] summarizes the baseline maternal characteristics.

Maternal characteristics, n (%)

Maternal comorbidities during pregnancy included hypertension in three (13.6%), preeclampsia in three (13.6%), cholestasis of pregnancy in one (4.5%), and rejection with deterioration in left ventricular ejection fraction (LVEF) in one (4.5%).

The single reported episode of rejection was in a woman transplanted at the age of 25 years for peripartum cardiomyopathy. She conceived 10 months post‐transplantation, having already had one episode of rejection with recovery of left ventricular function. At 29 weeks of gestational age, she was hospitalized with another episode of rejection due to poor compliance with immunosuppression. Her LVEF fell to 30%–35%. No endomyocardial biopsy was performed. She was treated with high‐dose methylprednisolone and delivered by Cesarean section at 30 weeks due to graft dysfunction. Postpartum, her left ventricular systolic function normalized; however, she is one of four women in our cohort who have subsequently died.

One woman who conceived 24 months after transplantation had diagnostic coronary angiography at 21 weeks of gestation which showed minor atheroma.

There was a significant change in serum Cr through pregnancy (P=.015) as a result of the mid‐trimester drop and third‐trimester rise in mean levels (baseline Cr 108±31 μmol L−1, first‐trimester Cr 103±28 μmol L−1, second‐trimester Cr 98±33 μmol L−1, third‐trimester Cr 114±43 μmol L−1).

Cesarean section was the mode of delivery for 11 (55%) women. Five (25%) women had spontaneous vaginal delivery. The mode of delivery for the remaining four live births was unclear from the available records. The principal indication for Cesarean section was maternal hypertensive disorders. The three women with preeclampsia delivered at 27, 29, and 34 weeks of gestation.

Four women died following pregnancy. One woman died in the immediate postpartum period from atonic postpartum hemorrhage. One death was the result of graft coronary disease at 48 months postpartum. The other two deaths at 26 and 57 months postpartum were in women known to be poorly compliant with immunosuppression (Table [NaN] ).

Maternal outcomes, n (%)

Two pregnancies (9%) ended in first‐trimester spontaneous miscarriage. Both miscarriages occurred in the same woman between two other successful pregnancies. There were no pregnancy terminations and no congenital anomalies. Four babies required admission to the neonatal special care unit, but all were discharged home in good condition. Within the offspring, there were no congenital cardiac anomalies. None of the interviewed women reported childhood developmental delay. Fetal outcomes are reported in Table [NaN] .

Fetal outcomes, n (%)

The immunosuppression regimens for each of the 22 pregnancies are shown in Table [NaN] . Immunosuppression was cyclosporine A (CyA) based in the majority (64%). One woman had mycophenolate mofetil (MMF) and another azathioprine stopped when pregnancy was confirmed. Both spontaneous miscarriages occurred in the same woman treated with tacrolimus (Tac) monotherapy.

Immunosuppression regimens during pregnancy

Once pregnancy was confirmed, trough CNI levels were monitored more frequently, generally every 2 weeks. There was a significant increase in the total daily dose of CyA (CyA dose: pre‐pregnancy 220±43 mg, first trimester 232±40 mg, second trimester 250±40 mg, and third trimester 257±34 mg, P<.001) required to maintain therapeutic levels through pregnancy (CyA level: pre‐pregnancy 122±32 μg/L, first trimester 94±32 μg/L, second trimester 104±39 μg/L, and third trimester 107±31 μg/L, P=.03, driven by the difference between pre‐pregnancy and first‐trimester levels, P=.014; Fig. [NaN] ).

Similarly, there was a significant increase in the total daily dose of Tac (Tac dose: pre‐pregnancy 4.2±1.8 mg, first trimester 5.6±2.5 mg, second trimester 9±6.3 mg, and third trimester 8±5.4 mg, P=.001) required to ensure no significant change in plasma levels (Tac level: pre‐pregnancy 7.2±1.1 ng/mL, first trimester 6.2±1.8 ng/mL, second trimester 5.1±2.1 ng/mL, third trimester 7.6±1.2 P=.29; Fig. [NaN] ).

This large single‐center experience of pregnancy in cardiac transplant recipients over a 29‐year period has shown encouraging maternal and fetal outcomes with a higher live birth rate (91%) than reported by the NTPR in 2010 (62%).[11] The UK transplant pregnancy registry which collected data between 1994 and 2001 reported an 83% live birth rate among cardiothoracic recipients, of whom 11 were women with heart transplants. Our findings highlight that although pregnancy following cardiac transplantation is relatively infrequent, it has the potential to become more common with increasing numbers of surviving pediatric patients with congenital heart disease being referred for transplantation.

Preconception counseling should ideally begin during assessment for transplantation and continue through post‐transplant follow‐up, with prompt attention to the need for contraception. Preconception counseling in these women can expose many difficult medical, psychological, and ethical questions. In women whose initial cardiac diagnosis was an inherited cardiomyopathy or congenital heart disease, comprehensive genetic counseling and assessment is also recommended. We report a high rate of unplanned pregnancies at 82%. Similarly, the NTPR showed that more than 50% of pregnancies in solid organ recipients are unplanned.[12] Due to difficulties in patient recollection we were unable to provide accurate details on the number of women reliably using contraception at or around the time of conception. Our center routinely recommends an appropriate contraceptive agent post‐transplantation taking into account the individual risks of hormonal treatment.[13] We also recognize that some previous reports have raised concern about the interaction between immunosuppressive agents and the contraceptive pill, perhaps rendering it less effective.[14] Preconception counseling should also acknowledge the longer‐term maternal prognosis, graft, and maternal survival and discuss the implications of each. Despite many successful pregnancies worldwide, limited information is available on the longer‐term outcome of women after pregnancy, and in comparison with those who have not had a pregnancy. The NTPR data suggest that 70% of cardiac recipients maintain adequate graft function postpartum, with only two patients in this series experiencing graft loss within 2 years. Maternal death was documented in 16 women from a group of 57 (28%), but all beyond 2 years.[15]

It is accepted that the physiological changes of pregnancy are generally well tolerated in the denervated cardiac allograft, provided graft function prior to pregnancy is normal. Pregnancy in a cardiac transplant recipient, however, is a complex medical condition. These women have been shown to be at significantly higher risk of medical comorbidities in pregnancy.[12] , [16] The ISHLT guidelines recommend that women contemplating pregnancy should undergo full cardiac assessment 6 months prior to conception. This includes coronary angiography, echocardiography, and electrocardiography. Depending on the clinical circumstances, right heart catheterization and endomyocardial biopsy may also be required.[3] Vaccination is an important component of pre‐ and post‐transplant care. Prior to transplantation, vaccinations should be up to date to reduce the risks of preventable disease, as live vaccines cannot be given to post‐transplant patients.[17] Vaccination history should be reviewed during preconception counseling as vaccines against influenza, pneumococcus, hepatitis B, and tetanus can be administered safely if required.[16] Asymptomatic urinary tract infection (UTI) is common in pregnancy. UTI occurred in 11% of the NTPR cohort. Routine urine cultures should be obtained on a regular basis, generally monthly to screen for this condition.[15] , [18]

Arterial hypertension is the commonly encountered comorbidity post‐transplantation, mainly due to CNI nephrotoxicity, and hypertensive disorders of pregnancy are the most frequently encountered comorbidity.[16] Preconception counseling should include a review of immunosuppressive and other medical therapies including antihypertensive agents to allow timely changes or substitution of those contraindicated in pregnancy. Inhibitors of the renin–angiotensin–aldosterone system are considered harmful in pregnancy due to their potentially teratogenic effects in all trimesters.[19] Statins remain contraindicated and should be stopped.[20] , [21] Alternative agents for the management of hypertension include labetalol, methyldopa, hydralazine, and nifedipine, all of which have a good safety profile in pregnancy and may be considered in conjunction with obstetric advice. Once pregnancy is confirmed, frequent monitoring of blood pressure, urinalysis, and renal function and close surveillance for signs of preeclampsia is recommended.[15] Over the last 10 years, international guidelines for the diagnosis and management of hypertension and preeclampsia have evolved due to a greater understanding of the pathogenesis of this complex condition. Despite this, preeclampsia remains a leading cause of maternal and neonatal death, and delivery of the placenta remains the only “cure.”[22] , [23] , [24]

Our results have demonstrated a similar rate of preeclampsia as those recently published from the UK Obstetric Surveillance System (UKOSS) at 14%.[25] The NTPR data from 2010 reported a slightly higher rate of preeclampsia, 18% in 57 transplant recipients with 103 pregnancies.[11] Current data from developed countries show that preeclampsia complicates between 1.4% and 4% of otherwise healthy pregnancies, with an overall decline in numbers over the last 10 years.[26] Hypertension in pregnancy was found in 3.6%–9.1% of pregnancies in the population‐based study by Roberts et al.[26] In 1993, Scott et al. observed a 48% rate of pregnancy induced hypertension and a 24% rate of preeclampsia among 30 women with heart transplants.[8] The incidence of hypertension in pregnancy in the NTPR was higher than that in our cohort, 39%.[27] It appears that the rate of hypertensive disorders in pregnancy may be declining among the transplanted population. Possible contributors to the reduction in hypertensive disorders in pregnancy in the general population include the use of interventions such as low‐dose aspirin, calcium supplementation, and early delivery for mild hypertension to avoid progression to preeclampsia;[26] these may also translate to transplant recipients. None of the women in our cohort developed gestational diabetes. The prevalence of gestational diabetes within the non‐transplant population in the United States has been reported to be as high as 9.2%.[28] The incidence among pregnant cardiac recipients has been documented to be significantly lower, with rates between 1% and 2%.[12] , [25] The explanation for this is not clear.

Pregnancy is a hypercoagulable state which confers an increased risk of venous thromboembolism (VTE).[29] This continues for several weeks postpartum. It has also been shown that heart transplant recipients have a higher risk of VTE[16] and surveillance for this complication is vital. It is unknown whether the pregnant heart transplant recipient is at even higher risk; therefore, screening for VTE must be universal and include a personal and family history of VTE according to published guidelines.[30] This risk assessment should be repeated at every hospital admission to assess the need for prophylactic anticoagulation.[30]

Figures from the ISHLT registry indicate that the likelihood of developing rejection in a cardiac allograft has decreased over the past 10 years. Current rates of treated rejection in the first year are approximately 13%.[1] The NTPR previously reported an 11% rate of rejection in pregnancy. In several cases, an increase in the oral dose of immunosuppression was sufficient to treat rejection.[12] Surveillance for graft rejection in our group of pregnant women was through serial echocardiography. We report a lower rate of rejection in pregnancy with only one case (4.5%) which was attributable to poor compliance with immunosuppressive therapy. Two case reports have described pregnancy‐related sensitization to human leukocyte antigens (HLA). One woman who developed severe rejection and graft dysfunction 2 months after delivery required re‐transplantation 5 months postpartum.[31] The other case report described a spontaneous miscarriage at 8 weeks, followed by de novo HLA sensitization and antibody mediated rejection. This woman died 2 years later following a massive pulmonary embolism.[32] On the basis of these findings, it has been suggested that any transplant recipient who develops rejection during or after pregnancy should be tested for HLA‐specific antibodies.[32]

Outcome data on maternal and fetal risks during subsequent pregnancies in cardiac allograft recipients are sparse. Three of the four women in our cohort who had more than one pregnancy had an uncomplicated course and were alive at follow‐up. The fourth had two first‐trimester miscarriages followed by a successful pregnancy, but died due to postpartum hemorrhage following the birth of her second child. Branch et al. previously described 12 women who had one or two additional pregnancies. They found no significant increase in the incidence of maternal, fetal, or graft complications.[33] These findings are yet to be confirmed by larger series.

The mode of delivery in cardiac transplant recipients should be dictated by obstetric indications.[15] We report a relatively high rate of Cesarean section within our group at 53% compared to the previously reported 40% in the 2010 NTPR report and 45% in the UK transplant pregnancy registry.[11] , [34] It is important to recognize that our data span an era during which it was unclear how the transplanted heart would adapt to the hemodynamic changes of labor and delivery. This may partially explain these results. The majority of Cesarean section deliveries in our cohort were for obstetric indications, however.

It is widely recognized that maternal health during pregnancy determines overall pregnancy and fetal outcomes.[15] Hypertensive disorders in pregnancy increase the risks of intrauterine growth restriction, low birthweight, and prematurity.[16] Preterm delivery in our group was dictated by maternal rather than fetal comorbidities as described with an average birthweight of 2447 g at a mean gestational age of 34 weeks. Similarly, in a report from the French transplant centers, the mean birthweight of infants born to women with heart or heart and lung transplants was 1990 g, at a mean gestational age of 35 weeks.[9] Results are also in keeping with those from the UKOSS which saw a mean birthweight of 2364 g at 35 weeks of gestation[25] and the NTPR with a birthweight of 2600 g at 36.8 weeks of gestation.[11]

Immunosuppression post‐transplantation must be individualized. Most patients in our center begin with a CNI, an antiproliferative agent and steroid. This may be reduced to two‐ or single‐agent therapy over time depending on clinical progress. The ISHLT guidelines recommend that immunosuppressive therapy is reviewed and that antimetabolites (MMF) and proliferation signal inhibitors (sirolimus and everolimus) are stopped and/or substituted at least 6 weeks prior to conception due to their potentially teratogenic effects.[3] The expansion of circulating blood volume, changes in distribution of body fat, increase in glomerular filtration rate, threat of hyperemesis gravidarum, and changes in gut motility influence the bioavailability and serum levels of immunosuppressive agents during pregnancy.[15] Frequent monitoring of CNI trough levels, ideally every 4 weeks until 32 weeks, then every 2 weeks until 36 weeks, then weekly until delivery has been suggested, with appropriate dose adjustments to maintain adequate trough levels. This may be coordinated in conjunction with obstetric reviews.[3] , [15] We would suggest even more frequent CNI trough levels in the first two trimesters to minimize the risk of significantly reduced levels culminating in rejection. Postpartum weekly CNI trough levels are also required as the changes in maternal circulating plasma volume take several weeks to return to baseline, and further dose adjustments are often necessary. We have illustrated a significant increase in the total daily doses of CNIs required to maintain therapeutic levels through the trimesters of pregnancy, with no associated decline in renal function. Similar findings have been described in pregnant renal transplant recipients.[35] , [36] Kim et al. demonstrated that renal function improved slightly during pregnancy and returned to pre‐pregnant levels following delivery. They also showed that a 20%–25% dose increase in CNIs should be expected during the gestational period.[36] In cardiac patients, there are less conclusive data in this field, with conflicting reports on dose adjustments throughout the literature.[6] , [10] , [37]

All immunosuppressive agents cross the placenta with the potential for teratogenic effects. No congenital cardiac or other anomalies were reported among the offspring in our group. There were no longer‐term health or developmental problems reported among the children of the women who were interviewed. The risk of spontaneous miscarriage in cardiac transplant recipients has been reported to be in the order of 15%–20%,[16] only 5% higher than that of the general population.

The limitations of this study are its retrospective and single‐center nature. As several of the women in our group delivered in other institutions we were unable to obtain all information pertaining to labor and delivery, including details on proteinuria with respect to preeclampsia. Given that not all included patients were available for interview, we lack longer‐term outcome data for the offspring. Our results are nevertheless encouraging, in keeping with recent publications, and contribute to the growing body of literature in this field.

In conclusion, we report a favorable live birth rate among cardiac transplant recipients. Pregnancy in women following cardiac transplantation is of high risk, and a dedicated multidisciplinary team comprising a transplant physician, an obstetrician with experience in maternal and fetal medicine, and an obstetric anesthesiologist is necessary to optimize maternal, fetal, and neonatal outcomes. Given the longer‐term survival following cardiac transplantation, pre‐pregnancy counseling is important in highlighting these issues. Larger prospective studies are still required to determine the longer‐term effects of pregnancy on maternal and graft survival.

We gratefully acknowledge the participation of the patients who gave us their time in allowing this study to be conducted. We also extend thanks to Dr R. D'Souza (Assistant Professor of Maternal Fetal Medicine, University of Toronto and Mount Sinai Hospital, Toronto, Canada) for his obstetric contribution to the preparation of this manuscript.

The authors have no conflict of interests to disclose.

Graph: Mean total daily dose of CyA required to maintain therapeutic levels through the trimesters of pregnancy, n=14, P <.001

Graph: Mean total daily dose of tacrolimus (Tac) required to maintain therapeutic levels through the trimesters of pregnancy, n=8, P =.001