Pregnancy Outcomes for Kidney Transplant Recipients With Transplantation as a Child

Importance  Pregnancy outcomes for women who received a kidney transplant in childhood are uncertain.

Objectives  To report pregnancy outcomes for women with kidney transplantation in childhood (aged <18 years; child-tx mothers) and to compare them with those for women who received a kidney transplant in adulthood (aged ≥18 years; adult-tx mothers).

Design, Setting, and Participants  Observational cohort study in Australia and New Zealand of all women with a functioning kidney transplant included in the Australia and New Zealand Dialysis and Transplant Registry for whom at least 1 pregnancy was reported between January 1, 1963, and December 31, 2012.

Main Outcomes and Measures  Pregnancy outcomes including live birth rates, gestational age, and proportion of babies who are small for gestational age.

Results  A total of 101 pregnancies in 66 child-tx mothers and 626 pregnancies in 401 adult-tx mothers were reported. At the time of pregnancy, child-tx mothers had a mean age of 25 (95% CI, 24-26) years with a functioning transplant for 10 (95% CI, 9-11) years, while adult-tx mothers had a mean age of 31 (95% CI, 31-31) years with a functioning transplant for 6 (95% CI, 5-6) years (both P < .001). Live births occurred in 76% of pregnancies in child-tx mothers and 77% of pregnancies in adult-tx mothers. The mean gestational ages were similar between child-tx and adult-tx mothers (35 [95% CI, 33-37] vs 36 [95% CI, 35-36] weeks, respectively; P = .68). The incidence of prematurity (<37 weeks’ gestation) was also similar (child-tx mothers, 45%; adult-tx mothers, 53%). A similar proportion of preterm babies born to child-tx and adult-tx mothers were small for gestational age (22% vs 10%, respectively; odds ratio [OR] = 2.53 [95% CI, 1.13-5.69]). Term babies born to child-tx and adult-tx mothers were frequently small for gestational age (57% vs 38%, respectively; OR = 2.16 [95% CI, 1.23-3.81]), both significantly more frequently than babies born at term in the general population (child-tx mothers, OR = 11.93 [95% CI, 5.56-25.61]; adult-tx mothers, OR = 5.52 [95% CI, 2.56-11.89]).

Conclusions and Relevance  Pregnancy outcomes for child-tx mothers are similar to those for adult-tx mothers, with no difference in the rate of live births, gestational age, or small for gestational age. Regardless of when women received their kidney transplant, their pregnancies are likely to result in a live, albeit preterm, birth. This work should provide comfort to child-tx mothers and their physicians that their early onset of kidney failure and longer period of transplantation and immunosuppression do not adversely affect their pregnancy outcomes compared with adult-tx mothers.

Children receiving a kidney transplant are now expected to survive and reach reproductive age. Kidney transplantation improves fertility in women of reproductive age with end-stage kidney disease (ESKD). Previous studies have shown that pregnancy in kidney transplant recipients does not cause long-term harm to either patient or graft in the vast majority of cases.^1,2 It has also been established that babies born to women with transplants are likely to be born prematurely and/or small for gestational age.³ However, previous studies have reported outcomes for all women with transplants, regardless of age at transplantation, and it is unclear whether their findings apply to women who received the transplant as children. Childhood recipients are likely to have had lengthy exposure to immunosuppression by the time pregnancy occurs and may also have had exposure to immunosuppression, including corticosteroids, during puberty. It is unknown how these factors may affect pregnancy outcomes. To our knowledge, no work to date has looked at whether pregnancy outcomes differ for women with kidney transplantation in childhood (child-tx mothers) compared with those with kidney transplantation in adulthood (adult-tx mothers).

This study aims to describe pregnancies occurring in child-tx mothers and to compare the pregnancy outcomes of child-tx vs adult-tx mothers to test the hypothesis that transplantation as a child may alter pregnancy outcomes.

This study used data from the Australia and New Zealand Dialysis and Transplant (ANZDATA) Registry, a national registry that collects demographic, treatment, and outcome data annually for all renal replacement patients from all renal units in Australia and New Zealand, with patient inclusion rates greater than 99%. Institutional review board approval was not required as this was a registry study of deidentified data from consenting patients (written or oral informed consent depending on unit) collected expressly for the purpose of such analyses. We included all female patients who received a transplant with a pregnancy reported between January 1, 1963, and December 31, 2012. The following patient parameters were recorded: age, weight, cause of ESKD, age at ESKD onset, time since transplantation, source of transplanted kidney, smoking status, prepregnancy immunosuppression, serum creatinine level, and estimated glomerular filtration rate (Modification of Diet in Renal Disease formula).⁴ Immunosuppression details are not specifically recorded at the time of pregnancy but are routinely collected at months 1, 2, 3, and 6 and at years 1, 2, 3, 5, 7, 10, 15, 20, 25, 30, and 35 after transplantation. Estimated date of conception, maternal complications (gestational diabetes mellitus and preeclampsia), gestational age, and birth weight were reported for pregnancies from January 1, 2001, onward. Pregnancy outcomes included live birth, stillbirth, spontaneous abortion (<20 weeks’ gestation), and termination.

We compared the pregnancy outcomes of kidney recipients with transplantation before age 18 years (child-tx mothers) with those who received the transplant at age 18 years or older (adult-tx mothers). Primary outcomes of interest include live birth rate, gestational age at birth, and birth weight.

Categorical variables were reported as frequencies and percentages, with mean and standard deviation reported for continuous variables. Categorical data were compared by the χ² test, Fisher exact test, or logistic regression as appropriate. Continuous data were compared by the t test or linear regression as appropriate. Statistical analyses were performed using Stata version 12 statistical software (StataCorp LP).

Between January 1, 1963, and December 31, 2012, 341 female patients received a kidney transplant in childhood and reached age 18 years with a functioning graft. Of these 341 women, 66 have had 101 pregnancies; 36 (55%) had a single pregnancy, 25 (38%) had 2 pregnancies, and 5 (8%) had 3 pregnancies. This compared with 626 pregnancies in 401 adult-tx mothers, of whom 215 (55%) had a single pregnancy, 118 (30%) had 2 pregnancies, 45 (12%) had 3 pregnancies, and 10 (3%) had 4 pregnancies.

The mean age of child-tx mothers at first pregnancy was 24 (95% CI, 22-25) years, 6 years younger than adult-tx mothers’ mean age of 30 (95% CI 30-31) years (P < .001) (Table 1). The mean weight of child-tx mothers at pregnancy was 58 (95% CI, 56-60) kg, 6 kg lower than the mean weight of adult-tx mothers at pregnancy of 64 (95% CI, 62-65) kg (P < .001). As the mean height of child-tx mothers was not collected at the time of pregnancy, body mass index cannot be calculated.

The mean age at onset of ESKD for child-tx mothers was 13 (95% CI, 12-14) years. The cause of ESKD is shown in Table 1. At the time of pregnancy, child-tx mothers had a functioning transplant for a mean of 10 (95% CI, 9-11) years, 4 years longer than the mean of 6 (95% CI, 5-6) years for adult-tx mothers (P < .001). There were no pregnancies in the first year after transplantation in the child-tx mothers, consistent with the American Society of Transplantation’s recommendations.⁵

A significantly higher frequency of living-donor transplants was evident in child-tx mothers than in adult-tx mothers (47% vs 34%, respectively; odds ratio [OR] = 1.56 [95% CI, 1.02-2.39]; P = .04).

There were no significant differences in the prepregnancy serum creatinine level of child-tx mothers compared with adult-tx mothers (1.30 [95% CI, 1.21-1.38] vs 1.28 [95% CI, 1.24-1.31] mg/dL [to convert to micromoles per liter, multiply by 88.4]; P = .68) (Table 1). The American Society of Transplantation recommends that pregnancy only occur with stable and adequate graft function defined in part as a creatinine level lower than 1.50 mg/dL.⁵ Prepregnancy serum creatinine level exceeded 1.50 mg/dL in 22% of both child-tx mothers and adult-tx mothers. The mean prepregnancy serum creatinine levels for these mothers were 1.95 (95% CI, 1.75-2.14) mg/dL and 1.83 (95% CI, 1.76-1.90) mg/dL, respectively. Excluding terminations, the live birth rate was 70% (12 of 15 pregnancies) for child-tx mothers in whom the prepregnancy serum creatinine level was higher than 1.50 mg/dL, compared with 92% (49 of 53 pregnancies) for those in whom the prepregnancy serum creatinine level was lower than 1.50 mg/dL (OR = 0.20 [95% CI, 0.09-0.47]). Excluding terminations, the live birth rate was 83% (76 of 92 pregnancies) for adult-tx mothers in whom the prepregnancy serum creatinine level was higher than 1.50 mg/dL, compared with 88% (336 of 380 pregnancies) for those in whom the prepregnancy serum creatinine level was lower than 1.50 mg/dL (OR = 0.67 [95% CI, 0.30-1.48]).

The serum creatinine level at 5 years post partum was significantly higher than prepregnancy and 1-year postpregnancy values. The serum creatinine level of child-tx mothers was 1.39 (95% CI, 1.24-1.55) mg/dL at 1 year post partum and 1.58 (1.40-1.76) mg/dL at 5 years post partum. The serum creatinine level of adult-tx mothers was 1.31 (95% CI, 1.27-1.35) mg/dL at 1 year post partum and 1.49 (1.41-1.57) mg/dL at 5 years post partum (Table 1).

Of 61 child-tx mothers over 5 years post partum, 34 (56%) had a serum creatinine level increase of less than 20%, 15 (25%) had an increase between 20% and 50%, and 12 (20%) had an increase exceeding 50%. The pattern was very similar for 334 adult-tx mothers over 5 years post partum: 204 (61%) had a serum creatinine level increase of less than 20%, 83 (25%) had an increase between 20% and 50%, and 47 (14%) had an increase of more than 50% (Figure). Women with a prepregnancy serum creatinine level higher than 1.50 mg/dL were not more likely to experience a serum creatinine level increase of more than 20% at 5 years post partum than those with a prepregnancy creatinine level lower than 1.50 mg/dL (OR = 1.03 [95% CI, 0.28-3.86] for child-tx mothers; OR = 1.42 [95% CI, 0.81-2.49] for adult-tx mothers).

Pregnancy complication data were collected from 2001 onward. These data were available for the pregnancies of 29 child-tx mothers and 215 adult-tx mothers. Preeclampsia occurred in 8 child-tx mothers (28%) and 60 adult-tx mothers (28%). Gestational diabetes occurred in no child-tx mothers and in 8 adult-tx mothers (4%). Baseline creatinine level was not significantly different for women who developed preeclampsia compared with those who did not (1.29 [95% CI, 1.21-1.37] vs 1.29 [95% CI, 1.26-1.32] mg/dL, respectively; P = .81).

Among babies born to child-tx mothers with preeclampsia, the mean gestational age was 31 (95% CI, 27-36) weeks (n = 8) and the mean birth weight was 1794 (95% CI, 795-2794) g (n = 5). For babies born to adult-tx mothers with preeclampsia, the mean gestational age was 35 (95% CI, 34-36) weeks (n = 57) and the mean birth weight was 2325 (95% CI, 2133-2517) g (n = 54).

Immunosuppression data were collected by ANZDATA at routine intervals, but not specifically at the time of pregnancy (see Methods). The most recent immunosuppression regimen recorded prior to pregnancy was used as a guide to prepregnancy immunosuppression, but this may not reflect the immunosuppression used at the time of conception or during pregnancy.

Of pregnancies occurring since 2000, immunosuppression data were available for 29 of the 30 pregnancies in child-tx mothers and 238 of the 242 pregnancies in adult-tx mothers. Of the 29 pregnancies in child-tx mothers, the immunosuppression regimens in the period preceding pregnancy were steroid free in 7 (24%), were tacrolimus based in 7 (24%), were cyclosporine based in 12 (41%), included azathioprine in 24 (83%), and included mycophenolate mofetil in 3 (10%). Of the 238 pregnancies in adult-tx mothers, the immunosuppression regimens in the period preceding pregnancy were steroid free in 73 (31%), were tacrolimus based in 93 (39%), were cyclosporine based in 129 (53%), included azathioprine in 178 (75%), and included mycophenolate in 44 (18%). There was no statistically significant difference in the use of prednisone, tacrolimus, cyclosporine, azathioprine, or mycophenolate between child-tx and adult-tx mothers.

There was no evidence of an era effect on outcomes of pregnancy in child-tx or adult-tx mothers. The live birth rate for pregnancies in child-tx mothers who received their graft after 1994 (n = 13; 11 live births [85%]) was not different from that in child-tx mothers who received their transplant prior to 1994 (n = 88; n = 66 live births [75%, or 89% excluding the 14 terminations]). Calcineurin inhibitor use was similarly not associated with live birth rate (P = .15).

The 101 pregnancies in child-tx mothers had the following outcomes: 77 live births (76%), 14 terminations (14%), 5 stillbirths (5%), 4 spontaneous abortions (4%), and 1 outcome of other (<1%) (Table 2). The 626 pregnancies in adult-tx mothers followed a similar pattern: 485 live births (77%), 65 terminations (10%), 61 spontaneous abortions (10%), 14 stillbirths (2%), and 1 outcome of other (<1%) (Table 2).

During the 5 decades for which pregnancy outcomes data have been collected, the largest change has occurred in the termination rate, which decreased from 33% of all pregnancies between 1963 and 1972 to 3% of all pregnancies between 2003 and 2012. This held true both for child-tx mothers, in whom terminations decreased from 50% to 0%, and for adult-tx mothers, in whom terminations decreased from 30% to 3%.

The mean maternal age at termination for child-tx mothers was 22 (95% CI, 20-24) years, significantly younger than the mean maternal age for any other outcome of 25 (95% CI, 24-26) years (P = .01).The mean maternal age at termination for adult-tx mothers was 30 (95% CI, 29-32) years, not significantly different from the mean maternal age for other outcomes of 31 (95% CI, 31-31) years (P = .22).

Excluding terminations, the live birth rates were 89% and 86% for child-tx and adult-tx mothers, respectively.

Maternal age has been shown to affect live birth rates.³ We looked at the effect of maternal age and age group at transplantation (child vs adult) on live birth rates. Neither of these variables significantly influenced live birth rates (Table 3).

Gestational age has been collected since 2001. Mean gestational age for babies was not different between child-tx mothers (n = 29) and adult-tx mothers (n = 198) (35 [95% CI, 33-37] vs 36 [95% CI, 35-36] weeks, respectively; P = .68) (Table 2). The incidence of prematurity (<37 weeks’ gestation) was 45% in child-tx mothers and 53% in adult-tx mothers. The risks from prematurity are at their greatest at a gestational age of less than 32 weeks. In total, 24% of babies born to child-tx mothers were born at less than 32 weeks compared with 16% of babies born to adult-tx mothers (OR = 1.68 [95% CI, 0.66-4.26]; P = .27). Time since transplantation has been shown to affect gestational age.³ In multivariable analysis, neither time since transplantation nor age group at transplantation (child vs adult) was shown to significantly influence gestational age (Table 4).

Birth weight has been collected for babies born to transplant recipients since 2001. The mean birth weight of babies born to child-tx mothers was 2365 (95% CI, 2030-2701) g (n = 23) (Table 2). Babies with child-tx mothers were small for gestational age (<10th percentile) in 57% born at term compared with 22% born prematurely (OR = 4.70 [95% CI, 2.54-8.71]).

The mean birth weight of babies born to adult-tx mothers was 2545 (95% CI, 2433-2656) g (n = 173) (Table 2). Babies with adult-tx mothers were small for gestational age (<10th percentile) in 38% born at term compared with 10% born prematurely (OR = 5.52 [95% CI, 2.56-11.89]).

There was no significant difference between the birth weight of babies born to child-tx mothers compared with adult-tx mothers (P = .28) (Table 2). Preterm babies born to child-tx mothers were not significantly more likely to be born small for gestational age (<10th percentile) (22%) than babies born to adult-tx mothers (10%) or the general population (10%) (OR = 2.53 [95% CI, 1.13-5.69]; P = .22). Term babies born to child-tx mothers were not significantly more likely to be born small for gestational age than babies born to adult-tx mothers (57% vs 38%, respectively; OR = 2.16 [95% CI, 1.23-3.81]; P = .24); however, both were more likely to be born small for gestational age than babies born among the general population (OR = 11.93 [95% CI, 5.56-25.61] for term babies born to child-tx mothers and OR = 5.52 [95% CI, 2.56-11.89] for term babies born to adult-tx mothers; P < .001 for both).

In multivariable analysis, gestational age and age at transplantation were both shown to significantly influence expected birth weight (Table 4). Each additional week of gestation was associated with an increase in birth weight of 135 g (Table 4). Babies born to child-tx mothers were expected to be 314 g lighter than babies born to adult-tx mothers (Table 4).

Understanding the pregnancy outcomes for women who received a transplant as a child is of increasing importance given the growing number of such women. At the end of 2012 there were 275 transplant patients younger than 18 years in Australia and New Zealand, of whom 107 were female. The vast majority of these children are expected to survive and reach reproductive age.⁶ The situation is similar in the United States. In 2011, 280 children received a kidney transplant in the United States, bringing the number of children with a functioning graft to 5469.⁶ It is not known how many of these children were female, but it is likely that one-third to one-half, or between 1823 and 2735, were female. This article provides the first look at the outcomes from pregnancy that patients who underwent transplantation as children can expect. These data will be of use to such patients and their clinicians in deciding whether to attempt pregnancy and, once achieved, what outcomes to expect.

We reported 101 pregnancies in 66 women who underwent transplantation as children and compared their outcomes with those of 626 pregnancies in 401 women who underwent transplantation as adults. Women who received their transplants as children had their first child at a younger age, but with a longer interval between transplantation and pregnancy, compared with women who received their transplants as adults. This finding is likely due to a combination of factors. Those who undergo transplantation as a child face inevitable gaps between transplantation and reaching reproductive age, desiring pregnancy, and becoming a parent. For women with transplantation as adults, the ability to bear children may have been a significant driver of their decision to pursue transplantation, prompting earlier efforts to achieve pregnancy after transplantation.

The mean age at first pregnancy in the Australian general population is 28 years.⁷ Women with transplantation as a child had their first pregnancy at 24 years, considerably younger than their healthy peers. It is possible that this reflects a desire to become pregnant with a well-functioning graft rather than wait and risk a poorly functioning graft that may not be able to support a pregnancy, a desire to see their children to adulthood despite their own reduced life expectancy, and/or differences in contraception counseling.

The live birth rate for child-tx mothers was 76%, or 89% excluding terminations. This compares favorably with the live birth rate of adult-tx mothers of 77%, or 86% excluding terminations, and suggests that pregnancy in women with transplantation as children had similar outcomes as pregnancy in women with transplantation as adults.

There was no significant difference in the gestational age at which babies were born to child-tx mothers compared with adult-tx mothers (35 and 36 weeks, respectively), nor was there a significant difference in the percentage of babies born before 32 weeks’ gestation, when morbidity from prematurity is highest. The absolute rate of babies born before 32 weeks’ gestation was very high at 24% for child-tx mothers and 16% for adult-tx mothers compared with the general Australian population at 1.5%.⁷

As expected, gestational age was a significant predictor of birth weight, with each additional week of gestation associated with an increase of 135 g in birth weight. Our multivariable analysis found that babies born to child-tx mothers were 314 g lighter than babies born to adult-tx mothers, a difference equivalent to more than 2 fewer weeks of gestation.

Babies born preterm to either child-tx mothers or adult-tx mothers were no more likely to be born small for gestational age than babies born in the general population (P = .22). However, babies born at term were significantly more likely to be born small for gestational age, with 57% of term babies born to child-tx mothers and 38% of term babies born to adult-tx mothers being small for gestational age (ORs of 11.93 and 5.52, respectively). Thus, babies born to mothers who have received a kidney transplant tend to be born either prematurely or small for gestational age, not both. The reason is unknown; however, hypotheses include obstetricians choosing to delay delivery for babies expected to be small, or higher rates of late placental failure.

Preeclampsia was reported in 28% of pregnancies for both child-tx mothers and adult-tx mothers, 4 to 5 times higher than in the general population.⁸ Babies born to child-tx mothers with preeclampsia appear to be born early and small compared with those born to adult-tx mothers with preeclampsia. However, caution should be applied to these results given the small absolute number of child-tx mothers with preeclampsia.

The findings of this article should be considered in the context of its limitations. First, this study relies on registry data for transplant recipient outcomes. Registry data provision is voluntary and can lead to underreporting of pregnancy and other outcomes. Second, immunosuppression data were available only from the most recent annual data collection period prior to pregnancy and may not reflect the immunosuppression used at conception or during pregnancy. Third, a number of important pregnancy outcomes, including birth weight and gestational age, have only been collected since 2001, which limits analyses of these measures; others, including cesarean delivery rates, have not been collected and so cannot be incorporated into the analysis.

To our knowledge, this study is the first to look at pregnancy outcomes for women who received a kidney transplant as a child. We have shown that outcomes are similar to those of women who received a transplant as an adult, with no difference in the rate of live births, mean gestational age, or mean birth weight. Regardless of when women received their kidney transplant, their pregnancies are likely to result in a live birth and a preterm birth. For both child-tx mothers and adult-tx mothers, babies born preterm are no more likely to be small for gestational age than those in the general population, while babies born at term are significantly more likely to be small for gestational age. This work has shown that outcomes for child-tx mothers are similar to outcomes for adult-tx mothers and should provide comfort to such mothers and their physicians that their early onset of kidney failure and longer period of posttransplant exposure to immunosuppression do not adversely affect their pregnancy outcomes.

Corresponding Author: Melanie L. Wyld, MBBS, MBA, MPH, Royal Prince Alfred Hospital, 50 Missenden Rd, Camperdown, New South Wales 2050, Australia (melwyld@gmail.com).

Accepted for Publication: December 8, 2014.

Published Online: February 2, 2015. doi:10.1001/jamapediatrics.2014.3626.

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

Study concept and design: All authors.

Acquisition, analysis, or interpretation of data: Wyld, Clayton, Kennedy.

Drafting of the manuscript: Wyld, Clayton, Alexander, Chadban.

Critical revision of the manuscript for important intellectual content: Wyld, Clayton, Kennedy, Alexander.

Statistical analysis: Wyld, Clayton.

Administrative, technical, or material support: Clayton.

Study supervision: Clayton, Alexander, Chadban.

Conflict of Interest Disclosures: None reported.