Pregnancy outcome after maternal organ transplantation in Sweden

Authors


Professor B. Källén, Tornblad Institute, Biskopsgatan 7, SE-223 62 Lund, Sweden.

Abstract

Objective  To study pregnancy outcome before and after organ transplantation.

Design  Registry study.

Setting  Swedish Health Registers.

Population  All births in Sweden 1973–2002.

Methods  Women who had organ transplantation were identified from the Hospital Discharge Register and their deliveries before and after the transplantation were identified from the Medical Birth Register. Abortions requiring hospitalisation were identified from the Hospital Discharge Register. Outcomes were compared with those in the total population. Adjustments were made for maternal age, parity and smoking habits.

Main outcome measures  Miscarriage, preterm delivery, low birthweight, small for gestational age, congenital malformations, infant death, pre-eclampsia and placental abruption.

Results  A total of 980 infants born before and 152 born after transplantation were identified. A high frequency of pre-eclampsia (22%), preterm birth (46%), low birthweight (41%), small for gestational age (16%) and infant death (5% before the age of one) were found for deliveries after transplantation but similar frequencies were found also among deliveries a few years before transplantation. No significant increase in congenital malformation rate was seen. Among 15 infants born after liver transplantation, 2 were malformed.

Conclusions  Pregnancies after organ transplantation have an increased risk of complications but this is similar to pregnancies occurring before the transplantation.

INTRODUCTION

Many papers have been published on pregnancies after organ transplantation. Most refer to renal transplantation1,2 and include small numbers of women, treated at one transplantation centre,3 national series4,5 or cases collected from transplantation registries.6–8 Less data are available on pregnancy outcome after liver transplantation.8,9–13

Most large studies found a high rate of preterm births, low birthweight and intrauterine growth retardation but no increase in congenital malformations. It has been suggested that use of cyclosporine as an immunosuppressant increases the risk for preterm birth more than the use of prednisolone and azathioprine.5

As new immunosuppressive drugs are introduced continued surveillance is important. We present here data on infants born after maternal organ transplantation in Sweden. For renal transplants, we also compare infant characteristics with those of infants born before renal transplantation. We have seen no similar comparisons.

METHODS

Women who have undergone organ transplantation in Sweden were identified from the Swedish Hospital Discharge Register. It contains discharge diagnoses from Swedish hospitals. From 1987 it covers the whole country, before that year some counties did not participate. Organ transplantation is performed in a limited number of hospitals, which all reported to the Hospital Discharge Register from the early 1970s onwards. From this register, women with an operation code indicating an organ transplantation were selected up to 2002, but information was incomplete for 2002. Each patient is identified with her personal identification number (PIN), a unique number allocated to each person living in Sweden.

The Medical Birth Registry contains data on nearly all deliveries that have taken place in Sweden since 1973. The year 2002 is still incomplete. This register is based on copies of medical documents from antenatal care centres, delivery units and the paediatric examination of the newborn infants.14,15 Women are identified with the abovementioned PIN to enable linkage with the Hospital Discharge Register. In this way, all infants born by the women who had an organ transplantation were identified, including both those born before and those born after the transplantation.

These infants were also searched for in the Register of Congenital Malformations16 in order to get further information on possible congenital malformations. The infants born from 1984 and later after transplantation were also searched for in the Hospital Discharge Register (from 1984 up to and including 2001) in order to identify morbidity resulting in hospitalisation among the infants.

Various characteristics were studied: type of organ transplantation, maternal age, parity, smoking habits in early pregnancy (information obtained at the first antenatal care centre visit), maternal diagnoses including diabetes and pre-eclampsia, gestational duration, infant survival, birthweight, sex, multiple birth, presence of infant diagnoses of congenital malformations and other infant and child diagnoses (neonatal and in the Hospital Discharge Register).

Characteristics of women who gave birth after organ transplantation were analysed for deliveries up to the end of 2001 (the year 2002 was incomplete at the time of analysis) and comprised 141 women with 143 infants (two twin pairs).

The data on pregnancy outcome of transplanted women were compared with corresponding data for all deliveries recorded in the Medical Birth Register.

Women who had been in hospital care for miscarriage were identified from the Hospital Discharge Register (1987–2001). The expected numbers of miscarriages and of deliveries were calculated using the Hospital Discharge Register and the Medical Birth Register, respectively, and the official information from Statistics Sweden on the age distribution of all women in Sweden (one-year classes) for each year. The proportion of pregnancies was estimated before and after transplantation and compared with corresponding proportions in the population. For pregnancies occurring after transplantation, truncation was made at the year of death of deceased women.

The statistical analysis was performed as Mantel–Haenszel odds ratio (OR) analyses with adjustment for variables as described. We used Miettinen's technique for estimate of 95% confidence intervals (95% CI). Intrauterine growth was evaluated using a birthweight for gestational age standard graph,17 based on data from the Medical Birth Registry.

Drug use among women who gave birth after 1995 was identified by scrutiny of medical records.

RESULTS

A total of 976 deliveries occurred before and a further 149 occurred following organ transplantation. There were four and three twin pregnancies among deliveries before and after organ transplantation, respectively. The distribution of the different types of organ transplantation is given in Table 1.

Table 1.  Deliveries before and after organ transplantation according to organ(s) transplanted.
Number of deliveries
Organ transplantationBefore transplantAfter transplant
Kidney643130
Liver7714
Heart2320
Liver + pancreas10
Kidney + liver41
Kidney + pancreas174
Kidney + heart20
Total976149
Number of infants979152

The expected number of deliveries before transplantation (1987–2001) (adjusted for year of delivery and maternal age) was 444 and the observed number 205 and after transplantation the expected number was 337 and the observed number 149. Thus, only about half of the expected number of deliveries occurred both before and following transplantation. Before transplantation, 34 miscarriages were identified—the expected number is 25 (risk ratio = 1.4, 95% CI 0.93–1.9). After transplantation, the observed number was 22 and the expected number 59 (risk ratio = 0.37, 95% CI 0.23–0.56).

The odds ratio for a miscarriage among all identified pregnancies (miscarriages + deliveries) adjusted for year of abortion/delivery and maternal age was 2.2 (95% CI 1.6–3.2) for women before transplantation and 3.2 (95% CI 2.1–5.0) after transplantation. These two odds ratios do not differ significantly (z= 1.28, P= 0.10).

Women who gave birth after organ transplantation were older and had lower parity than expected (Table 2). Information on smoking in early pregnancy was not available until in 1983, and this analysis was therefore restricted to the period 1983–2001 (n= 128). There was an excess of smokers among women who had organ transplantation but statistical significance was only reached for women smoking <10 cigarettes/day.

Table 2.  Characteristics of women giving birth after organ transplantation.
VariableOR*95% CI
  • *

    Odds ratios (OR) adjusted for year of birth, maternal age, parity and (1983–2001) for smoking in early pregnancy, with 95% confidence intervals (95% CI).

Maternal age
–190.530.17–1.6
20–240.460.28–0.75
25–290.720.50–1.0
30–341.40.95–1.9
35–392.81.9–4.3
40–443.01.4–6.6
 
Parity
12.82.0–43.9
20.720.50–1.0
30.470.27–0.82
4+0.260.10–0.66
 
Smoking
None1.00reference
<10 cigarettes/day1.81.2–3.0
10+ cigarettes/day0.780.31–2.0
Any1.50.94–2.3

Among 801 women who had a renal transplantation, 41 (5%) had a diagnosis of preconceptional diabetes. Among 324 women who had another type of organ transplantation, only one had a diabetes diagnosis (0.3%).

Pre-eclampsia occurred at a high rate in pregnancies after kidney transplantation: among 135 such women, 30 had pre-eclampsia (22%). After liver transplantation, an even higher frequency was seen (5/15 = 33%) but the confidence interval is large (95% CI 12–62%). Among women who had a delivery within five years before kidney transplantation, the frequency of pre-eclampsia was also high (16/78 = 17%). The odds ratio for having pre-eclampsia, adjusted for year of birth, maternal age, parity and (after 1982) for maternal smoking, was 9.6 (95% CI 6.8–14) for women who gave birth after renal transplantation, 12 (95% CI 8.0–19) for women who gave birth within five years before transplantation and 6.5 (95% CI 4.4–9.4) for women who gave birth five years or more before renal transplantation.

Only few women experienced a placental abruption, none of them had pre-eclampsia. After organ transplantation, only one woman (expected number 3.8 after adjustment for year of birth, maternal age, parity and smoking habits) had a diagnosis of abruption. Within five years before organ transplantation, five women had an abruption (expected number 1.5).

Among singleton pregnancies after transplantation, there was a higher frequency of preterm delivery (<37 weeks) and low birthweight (<2500 g) and of infants that were small for date (<−2 SD) (Table 3). These effects were more pronounced when the mother had only a kidney transplant than when she had only a liver transplant. The OR for preterm birth after adjustment for year of birth, maternal age, parity and (after 1982) smoking habits was 13 (95% CI 10–17), for births <32 weeks of gestation (n= 19) it was 16 (95% CI 11–22). The OR for birthweight <2500 g was 16 (95% CI 12–20) and for <1500 g (n= 14) 13 (95% CI 8.5–20). The OR for small for gestational age was 6.4 (95% CI 4.2–9.5).

Table 3.  Preterm birth (<37 weeks), low birthweight (<2500 g) and small for date (<−2 SD) among singleton infants born before and after maternal organ transplantation.
 Born before transplantationBorn after transplantation
Preterm birth
All202/965 = 21%60/139 = 46%
Only kidney154/637 = 24%58/126 = 46%
Only liver26/214 = 12%4/10 = 29%
 
Low birthweight
All182/965 = 19%58/143 = 41%
Only kidney150/635 = 24%52/126 = 41%
Only liver14/215 = 6.5%2/12 = 17%
 
Small for date
All90/982 = 9.4%24/143 = 17%
Only kidney73/634 = 24%23/126 = 18%
Only liver9/214 = 4.2%0/12 = 0.0%

Among 73 singleton infants born within five years after the transplantation, 34 were preterm (47%). Among 72 singleton infants born five years or more after transplantation, 32 were preterm (44%). These two rates obviously do not differ.

Also among infants born before the organ transplantation, a high percentage were born preterm, with low birthweight, and were small for date (Table 3), more pronounced for infants born of women who later had a kidney transplant than for infants whose mothers later had a liver transplant.

Figure 1 shows ORs for the infant being born preterm and for being small for date according to the time elapsed between delivery and transplantation. Definite risk increases were seen long before the transplantation, even though the risk increased the closer to the transplantation the pregnancy occurred.

Figure 1.

Odds ratios (OR) with 95% confidence intervals (95% CI) for preterm birth (<32 weeks, <37 weeks) or small for gestational age (SGA). ORs are adjusted for year of birth, maternal age, parity and (after 1982) maternal smoking habits.

A congenital malformation was identified (using both the Medical Birth Registry and the Register of Congenital Malformations) in 57 among 976 infants born before organ transplantation, 5.8%, and 10 among 149 infants born after organ transplantation, 6.7%, but two of the 10 infants were twins.

The two rates do not differ (P= 0.71). In six of the malformed infants born before transplantation, the mother had diabetes before pregnancy. One of the women who had a malformed infant after transplantation had diabetes.

The odds ratio for having any congenital malformation, adjusting for year of birth, maternal age and parity (and after 1982) for smoking, is 1.6 (95% CI 0.79–3.1) for infants born after transplantation and 1.2 (95% CI 0.91–1.7) for infants born before transplantation.

Table 4 specifies 10 infants who had congenital malformations, born after organ transplantation in the mother. Two were born after liver transplantation (among 15 such infants), one of them had a serious and complex malformation. That woman had been treated in early pregnancy with mycofenolate mofetil (MMF), tacrolimus, prednisolone and ursodeoxycholic acid. MMF was introduced in Sweden in 1995 and we searched the medical records for all women who had an infant born after transplantation during 1995–2002: among 32 such women, none further had got MMF. Most had a combination of cyclosporine, prednisolone and azathioprine.

Table 4.  List of congenital malformations identified among 149 infants born after maternal organ transplantation. Maternal age at delivery (years).
Maternal ageYear of transplantYear of birthOrgan transplantCongenital malformation
3519781983KidneyVentricular septum defect
2919761985KidneyUnspecified urinary malformation
4019831992KidneyDown with ECD
2619941996KidneyPartial trisomy 16
1919891996KidneyTwins: both unspecified cardiac defects
2519751998KidneyPreauricular tag
2819871999KidneyTricuspidal malformation
3619972001LiverVentricular septal defect
2220002002LiverOesophageal atresia + complex cardiac defect + iris anomaly

Table 5 shows the number of infants and children who died before the age of one year and odds ratios for any death, adjusted for year of birth, maternal age and parity. The risk of death is markedly increased but is so already before the transplantation.

Table 5.  Survival of infants born before and after organ transplantation.
StatusRelationship to transplantation
≥5 years before<5 years beforeAfter
  • *

    Odds ratios (OR) adjusted for year of birth, maternal age, parity and (1983–2001) for smoking in early pregnancy, with 95% confidence intervals (95% CI).

Alive817118141
Stillborn12142
Dead <7 days932
Dead 7 days to 1 month251
Dead 1 month to 1 year333
Total number of infants843143149
Percent dead3.117.55.4
OR*3.216.79.3
95% CI2.1–4.811–265.0–17

Among the 123 infants born before the end of 2001 (and therefore followed up in the Hospital Discharge Register), most (39 of 74) had been hospitalised because of various infections. The odds ratio for being hospitalised for any infection at any age was 2.1 (95% CI 1.5–3.0), adjusted for year of birth, maternal age and parity and maternal smoking. When stratification was also made for gestational duration (24–27 weeks, 28–31 weeks, 32–36 weeks, 37–38 weeks and 39+ weeks), the odds ratio decreased to 1.2 (95% CI 0.77–1.9) and was no longer significant.

Nine infants were hospitalised for asthma (7%), one for allergic purpura, one for coeliac disease and one for an undefined malabsorption condition. One infant had a diagnosis of a benign tumour, located to the long bones of the lower limbs. Two further infants with congenital malformations were identified from the Hospital Discharge Register. One had a patent ductus arteriosus (born after 29 gestational weeks), the other a lacrimal duct stenosis.

Discussion

Women who had organ transplantation were less fertile compared with all women in the population, both before and after the transplantation, reasonably an effect of severe maternal disease. There was an apparent increase in miscarriage rate both before and after transplantation but this finding is difficult to evaluate. The study refers to women who were hospitalised for a miscarriage and hospitalisation may be more likely when the woman has a chronic disease or has had a transplant. Furthermore, no information is known about induced abortions, which may also be more common in chronically sick women than in the general population.

The outcome data in the present study agree well with what is known in the literature: a very high rate of preterm birth, of low birthweight and of small for gestational age. The advantage of the present study is that it represents a total population and that the outcome data were obtained from a medical birth register, based on original medical record data. This fact made it possible to get better risk estimates than when outcome data obtained by individual follow up are compared with general population data. It also made it possible to adjust for certain characteristics of the pregnant woman in the transplanted group, like age, parity and smoking habits. We verified the observation made repeatedly in the literature of an increased risk for pre-eclampsia but a similar risk existed in pregnancies during the years before transplantation. Even many years before transplantation, an increased risk for pre-eclampsia was found.

In the study, no significant increase in congenital malformation rate was seen, which is also in agreement with most publications in the field. However, among the 15 infants born after maternal liver transplantation, there were two with a congenital malformation, one of which was complex and serious: oesophageal atresia with a heart defect and an iris malformation. This woman was the only one who had been treated with MMF. This may be a coincidence. Only few pregnancies exposed for MMF are published in the literature.18 We know about one further case of MMF exposure in early pregnancy with gastroschisis in the infant (reported as an adverse drug reaction). This case was not included in the study because the liver transplantation was made abroad and was therefore not recorded in the Hospital Discharge Register.

When delivery outcome was studied among women who later had a renal transplantation, an excess of preterm delivery, low birthweight, intrauterine growth retardation, and infant death was seen before the transplant, similar odds ratios were seen for infants born a few years before and after the transplant. This indicates that the major cause for these pregnancy outcomes is disease morbidity of the woman. There are numerous publications showing the ill effect of chronic renal failure on pregnancy outcome.19

The increased fetal/neonatal mortality and the high rate of preterm deliveries in patients already before kidney transplantation have not been described previously and are reasonably the result of chronic renal disease. It seems valuable to consider the reproductive prognosis of these women when evaluating them for organ transplantation.

Another concern about treatment with immunosuppressive drugs during pregnancy is the possibility that they can affect the immunological mechanisms of the fetus, resulting in morbidity in later life. We found no clear-cut such effect. It is true that the children, when followed up, appeared to have been hospitalised for infections in a higher frequency than other children, but this seemed to be mainly an effect of preterm birth. Larger studies with longer follow up are needed in order to exclude the possibility of such late morbidity and also of an increased risk for malignancy.

Accepted 30 October 2004

Ancillary