The outcome of pregnancies in women with cystic fibrosis—single centre experience 1998–2011

Authors


Correspondence: JG Thorpe-Beeston, Chelsea and Westminster NHS Foundation Trust, 369 Fulham Road, London, SW10 9NH, UK. Email guytb@btopenworld.com

Abstract

Objective

To describe the maternal and fetal outcomes of pregnancies in women with cystic fibrosis.

Design

Retrospective study.

Setting

Single obstetric hospital and adult cystic fibrosis centre.

Methods

Retrospective case-note review of pregnant women with cystic fibrosis referred for antenatal care and delivery.

Main outcome measures

Maternal and fetal outcomes, mode of delivery, lung function and pregnancy complications.

Results

Forty-eight pregnancies were studied in 41 women. There were two miscarriages, 44 singleton pregnancies and two sets of twins. All babies were liveborn and survived. The mean gestational age at delivery was 35.9 ± 3.3 weeks. There were no fetal abnormalities or terminations of pregnancy. The median birthweight centile was 31.9 (interquartile range 14.9–55.6). Twenty-five (52.1%) of the women had pancreatic insufficiency and 17 (35.4%) required insulin. There was a positive correlation between booking predicted forced expiratory volume in 1 second (FEV1) and gestational age at delivery (P < 0.01). Women with FEV1 ≤60% were more likely to deliver earlier and by caesarean section compared with women with FEV1 >60% (35.0 ± 3.2 weeks versus 37.1 ± 3.0 weeks; P = 0.02 and 75.0% versus 25.0%; P = 0.01). Three of the seven women with an FEV1 <40% died within 18 months of delivery. Four of the eight women with FEV1 40–50% died between 2 and 8 years after delivery.

Conclusion

Pregnancy for women with cystic fibrosis is possible and results in favourable maternal and fetal outcomes, but the incidence of preterm delivery and caesarean section is increased. Women with pre-existing poor lung function should be counselled antenatally to ensure that they understand the implications of their shortened life-expectancy and parenthood.

Introduction

Cystic fibrosis (CF) is an autosomal recessive disease caused by mutations of a gene on chromosome 7. The defects result in abnormalities of the CF transmembrane conductance regulator protein, reduced chloride conductance results in thickened secretions in the respiratory and gastrointestinal tracts, ultimately leading to irreversible organ damage. The gene prevalence is approximately 1:25 in the Caucasian population and in the UK there are over 9000 people affected by the condition.[1] Considerable advances in the diagnosis and management of CF have led to a much improved quality of life and life expectancy. One study has suggested that the life expectancy for a child born with CF in 2000 will be more than 50 years.[2] In 1984 Coffman et al.[3] reported the difficulties of sexual adaptation experienced by young single women with CF, but in a recent study of attitudes to fertility issues among adults with CF, 72% of women and 85% of men interviewed felt that it was important to have children.[4] There is little evidence that the fertility of women with CF is compromised and so all sexually active women should be advised to use contraception unless they wish to fall pregnant.[5] It is reasonable to hypothesise that the cervical mucus plug may be altered in some women and therefore consideration may be given to assisted conception techniques if there is an undue delay in conception. It is possible that severely ill people with reduced body mass index may have reduced fertility secondary to anovulation. Approximately 97% of men with CF are infertile as the result of abnormal development of the vas deferens, epididymis and seminal vesicles, but with the help of advanced microsurgical techniques and in vitro fertilisation (IVF) they may still achieve biological paternity.[6, 7] Parenthood is therefore becoming a normal expectation, rather than an exception.

Normal pregnancy is associated with significant changes in respiratory, cardiovascular and endocrine metabolism. Although healthy women generally adjust easily to these changes, in women with CF they impose a significant challenge on potentially already compromised lung, cardiac and pancreatic function. The first pregnancy in a mother with CF was reported in 1960. The woman laboured prematurely and died 6 weeks postpartum.[8] Several subsequent reports have documented the outcome of pregnancy in CF and in a multi-centre study over 200 pregnancies were reviewed by Kent and Farquharson in 1993.[9] The overall rate of miscarriage was 4.6%, 13.8% of pregnancies were electively terminated and 81.6% of pregnancies progressed beyond 20 weeks of gestation. The rate of preterm delivery was 24.3%. The perinatal death rate was 7.9% and 13.6% of mothers died within 2 years of delivery. Predictably, poor maternal outcomes were related to the maternal pulmonary status. This has led some authors to advocate termination of pregnancy or avoidance of pregnancy for women with poor lung function, forced expiratory volume in 1 second (FEV1) <60% predicted, forced vital capacity <50%, Shwachman or Taussig clinical score below 80.[10-13] However, there are reports of successful pregnancy outcomes in women with poor lung function.[14-16]

The crucial issue that multiple studies have attempted to address is whether pregnancy has an adverse effect on the disease course in women with CF. Cohen et al.[13] in 1980 described a 10% maternal death rate within 6 months of delivery and a 15% loss within 2 years. Three small case–control studies were subsequently reported.[17-19] In the first series although statistical significance was not reached, the pregnant mothers had better lung function than the controls prepregnancy and yet had a greater decline after delivery.[17] The second study contained only seven women and the authors concluded that pregnancy has little adverse effect on women with stable CF, but they still cautioned that, although pregnancy is a viable option in women with CF, pregnancy should only be recommended to individuals with milder disease.[18] In the third study, 33 women were matched and pregnancy outcomes and maternal lung function were studied. The authors suggested that those with good lung function were unaffected by pregnancy, but that pregnancy may directly affect women with poor lung function.[19] The most significant study to date of 680 women identified to be pregnant in 1990 by the North American CF Foundation database studied the 10-year postpregnancy survival. It concluded that even in women with very poor lung function their survival was not worse and indeed may have been improved compared with a nonpregnant matched population.[20]

The aim of our study was to report the outcomes of pregnancies in women with CF in a single centre over a relatively short time-frame. This is in contrast to previous studies (Table 1), most of which have suffered from being either multicentre studies or from analysing data over a 20-year or longer period during which time it is likely that medical care will have evolved.[19, 21-26]

Table 1. Studies reporting the outcome of pregnancies in women with cystic fibrosis
Ref.Study typeDatePregn. (n)Live births (%)Prem. rate (%)Misc. (%)TOP (%)GA (weeks)BW (kg)Diab. (%)
  1. Pregn., pregnancy; Prem. rate, prematurity rate; Misc., miscarriage; GA, gestational age; BW, birthweight; Diab., diabetes.

Cohen et al.[13]MS1975–761297527519
Barak et al.[14]MS1977–200411100270045
Frangolias et al.[18]S19937100430034.62.8529
Edenborough et al.[19]MS1977–96727032102036.92.914
Goss et al.[20]MS1985–9768074617
Gilljam et al.[21]MS1963–9892807128403.219
Gillet et al.[23]MS1980–99758525772.73
Odegaard et al.[23]MS1977–988075241410
Canny et al.[24]S1963–9038895011393.211
Jankelson et al.[25]S1977–9613100500036.123
Cheng et al.[26]MS1989–0443844497373.0930
Lau et al.[28]S1995–002095250543
Boyd et al.[30]MS200184741710029
Thorpe-Beeston et al.S1998–20114896464035.93.035.4
Combined outcomes 1975–20111392782471337.13.021

Methods

Data were collected retrospectively following a review of the maternity and medical case-notes from pregnant women under the care of the Royal Brompton Hospital Adult CF Centre who were referred for antenatal care and delivery to the Chelsea and Westminster Hospital between 1998 and 2011. In addition to the mother's demographic details, data were collected concerning the maternal body mass index, genotype, pancreatic insufficiency, diabetes status and the most recent spirometric lung function and microbial colonisation. The obstetric data collected included the gestational age at delivery, mode of delivery, birthweight and sex of the neonate. Birthweight centiles were calculated using a reference range based on gestational age at delivery and birthweight.[27]

Statistical analysis

Normality of the distribution of continuous data was examined with the Kolmogorov–Smirnoff test. Descriptive data were expressed as mean ± standard deviation (SD) or as median (interquartile range) for normally and not normally distributed data respectively. Comparisons between groups were performed using Student's t test/Mann–Whitney and chi-square test for numerical and categorical data, respectively. Univariate regression analyses were used to determine whether there was an association between the maternal respiratory variables and gestational age at delivery or birthweight percentile. The statistical analyses were performed using the SPSS version 12.0 (SPSS Inc., Chicago, IL, USA).

Results

Forty-eight pregnancies were studied in 41 women with CF between 1998 and 2011. Six women had two liveborn babies and there were two twin pregnancies. Two women miscarried before 11 weeks of gestation. Five pregnancies were conceived by in vitro fertilisation. The demographic details are shown in Table 2 and further details provided in Table 3. The gestation at delivery was 35.9 ± 3.3 weeks, range 26.0–40.4 weeks). There were no fetal abnormalities. A further three women requested termination of pregnancy for social reasons and were managed at their referring hospitals. These women were not included in the study group. There were 48 liveborn babies and there were no losses in the postnatal and neonatal periods. Twenty-three (50%) women in the study delivered vaginally. The median birthweight centile at delivery was 31.9 (14.9–55.6).

Table 2. Demographic details of the study population (n = 48)
  1. FEV1, forced expiratory volume in 1 second.

Maternal age (years), mean ± SD 29.5 ± 5.2
Parity, n (%)
Nulliparous38 (79.2)
Multiparous10 (20.8)
Body mass index (kg/m2), mean ± SD 21.9 ± 3.6
Body mass index <20, n (%) 14 (29)
Lung function
FEV1% at booking, mean ± SD60.9 ± 19.8
FEV1 <60% at booking, n (%)25 (52.0%)
FEV1 <40% at booking, n (%) 5 (10.4%)
FEV1% postdelivery, mean ± SD58.4 ± 20.0
Pancreatic insufficiency, n (%) 26 (54.2%)
Pseudomonas aeruginosa positive, n (%) 33 (68.7%)
Table 3. Demographic, lung function and pregnancy outcomes of the study participants
PatientsMaternal age (years)FEV1 (%)Pancreatic insuff.DiabetesMode of deliverySexBW (g)GA at delivery (weeks)BW centilesLong-term outcomes
  1. FEV1, forced expiratory volume at 1 second; GA, gestational age; F, female; M, male; Misc., miscarriage; BW, birthweight.

13947YesYesVaginalF260037.415.1Died 3.5 years
23160YesNoVaginalM310040.019.4 
32822NoNoCaesareanF70026.015.3Died 2 days
42732NoNoCaesareanM79026.041.1Died 7 days
53865YesYesCaesareanM330038.756.0 
64026NoNoVaginalM302039.025.2 
73045YesYesCaesareanM140033.60.6Died 1year
83750NoNoVaginalF270038.410.9 
924102NoYesCaesareanF270038.014.9Died 8 years
10 Twins3952NoNoCaesarean

M

F

1690

1800

34.1

34.1

2.9

6.2

 
11
Preg. 13171YesYesVaginalM240037.94.0 
Preg. 23360YesYesVaginalM269037.00.0 
1224101YesNoVaginalF295039.01.8 
133159YesNoCaesareanF290037.440.4 
142660YesYesCaesareanF265036.045.5 
15
Preg 13585NoNoVaginalM340040.4 Misc.38.8 
Preg 23980 
161740     Misc.  
172856NoNoCaesareanM280035.084.2Transplant 2 years
182545YesYesCaesareanM226032.491.5Died 8 years
193644YesYesCaesareanM260037.118.7Died 8 years
20 Twins3276NoNoCaesarean

M

F

1470

1420

31.4

31.4

19.5

14.2

 
213353YesNoCaesareanF260037.020.7 
223261NoNoVaginalM280037.923.6 
233485NoNoCaesareanM390038.197.5 
243391YesYesVaginalF130029.068.9 
252650NoNoCaesareanM280037.039.1 
262958NoNoVaginalM230034.348.8 
272235YesYesCaesareanM220033.169.9Died 1 year
28
Preg 12553YesYesCaesareanF210035.013.2 
Preg 22848YesYesCaesareanM260035.065.4 
2934106NoNoVaginalF320039.041.0 
30
Preg 12592YesNoVaginalM353040.056.7 
Preg 22877YesNoVaginalM320038.746.2 
313067YesNoVaginalM390039.093.4 
322253YesYesVaginalF228037.04.7 
33
Preg 13557NoNoCaesareanM187034.011.1 
Preg 23757NoNoVaginalM298036.079.5 
34
Preg 13071NoNoVaginalF240036.118.4 
Preg 23365NoNoVaginalF290037.049.6 
353750YesYesCaesareanF180032.049.7 
361763NoNoVaginalM286038.915.5 
37
Preg 12338YesYesCaesareanM189034.47.7 
Preg 22072YesYesVaginalM186032.349.0 
382953.6YesNoVaginalM329037.085.1 
393153NoNoVaginalM228034.054.6 
402386YesYesCaesareanF280038.021.6 
413181NoNoCaesareanM340037.090.9 

Twenty-five women (52.1%) were pancreatically insufficient prepregnancy and 17 women (35.4%) required insulin. There was no significant difference in gestational age at delivery and birthweight percentile between women requiring insulin and those that did not (P = 0.43 and P = 0.24, respectively).

There was a strong positive association between booking FEV1 and gestational age at delivery (P < 0.01) (Figure 1). Women with FEV1 ≤60% were more likely to deliver earlier and by caesarean section compared with women with FEV1 >60% (35.0 ± 3.2 weeks versus 37.1 ± 3.0 weeks; P = 0.02 and 75.0% versus 25.0%; = 0.01, respectively). However, there was no association between booking FEV1 and birthweight centile (P = 0.37). Similarly, there was no significant association between forced vital capacity% and gestational age at delivery (P = 0.25) or birthweight centile (P = 0.76).

Figure 1.

Forced expiratory volume in 1 second with gestational age at delivery illustrating the individual values and the regression line (P < 0.01).

Eight women died within the follow-up period and one has undergone transplantation. Three of five women with an FEV1 <40% died within 18 months of delivery. Four of nine women with FEV1 40–50% died between 2 and 8 years after delivery.

Discussion

In this series of 46 pregnancies that continued beyond the first trimester there were no fetal losses. The three women with the worst prepregnancy lung function were offered termination of pregnancy on more than one occasion, but it was the mothers' wish in each case to continue the pregnancy. Sadly all three died within 18 months of delivery. It is likely that as clinical experience grows, fewer women will opt for termination for medical indications, but it reinforces the need for very careful prepregnancy counselling. The results of this series on the one hand reinforce the view that those with poor lung function may have a short period of parenthood and some would therefore counsel against pregnancy, but others may be encouraged that all the women were able to endure the pregnancy and deliver a liveborn infant. Women with CF are well aware of their limited life-expectancy, but it essential that these issues are addressed prepregnancy, possibly formally at the woman's annual CF review. Appropriate contraception should be encouraged to avoid unplanned pregnancies.

Although 46% of babies were delivered prematurely (delivered before 37 weeks), in all of them the outcome was good. There were two cases of spontaneous preterm labour, which would be consistent with the rate in the general population. In all other pregnancies prematurity was iatrogenic as the result of concern regarding maternal wellbeing. In all of these women antenatal steroids were administered. Seven women delivered two liveborn babies and it is possible that their results may have introduced some bias; however, in all cases their individual lung function deteriorated between pregnancies. It seems clear therefore that women with CF, even those with severe disease are likely to deliver a liveborn baby and with the benefit of neonatal care the outcome for these children is excellent. These findings are in agreement with the outcomes reported by Lau et al.[28] in a smaller study, but over a similar period, between 1995 and 2009. Table 1 summarises the main results of 14 studies. The rate of prematurity appears greater in the more recent studies, perhaps as a consequence of increasingly good outcomes for premature babies and that women with worse lung function are now falling pregnant. The management of CF has changed hugely over the last four decades and the prognosis for the condition has improved markedly. Our data have the benefit of representing the largest single-centre experience, with a consistency of management and approach being adopted. The mean age of women in our study was 29.5 years. This is older than earlier studies report.[19, 21, 22, 26] However, as expected, the older age of this cohort was also linked to an overall worse prepregnancy lung function. The mean FEV1 in our study was 61%, lower than older series (68–83%), but similar to a recent study by Cheng et al. (61%).[18, 19, 22, 26] The studies addressing the long-term outcome following pregnancy in women with CF have uniformly reported a higher mortality rate for women with worse pre-existing lung function. The critical issue though is whether the pregnancy has detrimentally affected lung function and hastened maternal demise. Early studies were clear in their interpretation of the poor maternal outcomes and Larsen[10] recommended termination of pregnancy in women with FEV1 <50%. Subsequently Cohen et al.[13] reported an 18% mortality rate within 24 months of delivery in women with severely compromised lung function. Two further studies have also counselled against pregnancy in women with poor lung function.[19, 22] However, the results from this study and others confirm that pregnancy with a good fetal outcome is possible in women with severely compromised lung function.[20, 21, 26, 28] In counselling women antenatally, it is helpful to be able to provide some information on the likely outcome and course of their pregnancy. We therefore compared the obstetric outcome in women with an FEV1 <60% with outcomes in women with better lung function. Those with poor lung function had a significantly increased risk of operative and preterm delivery. With the advances in neonatal care, however, the outcome for these babies is generally very good, but the implications are potentially more serious for those born extremely prematurely.

A large study from the Cystic Fibrosis Foundation National Patient Registry analysed the outcome of pregnant women with CF and found no difference in the rates of hospitalisation, use of home intravenous antibiotics, supplemental oxygen and feeding in the year after pregnancy.[29] A matched parallel-cohort study of women from the same national registry concluded that women with CF who became pregnant were initially healthier and had a better 10-year survival rate than matched women with CF who did not become pregnant. Further analysis of the data concluded that even in women with severe CF (FEV1 <40%) or diabetes mellitus, pregnancy was not harmful.[20] The key issue therefore is that women with CF undertaking a pregnancy, particularly those with poor lung function, must recognise that their days as a parent may be limited. Twenty percent of mothers with CF will be dead by their child's tenth birthday and for those with an FEV1 <40%, 40% will have died. Our findings are broadly in agreement with this large study and 50% of women with an FEV1 <50% died within 8 years. Long-term issues, such as the support of the spouse and family, should be discussed frankly before undertaking pregnancies so that the enormity of these issues has time to be assimilated.

Five women were able to achieve a pregnancy with the assistance of IVF. Although there is little evidence that women with CF and reasonable lung function have impaired fertility, in view of their reduced overall life expectancy it would seem reasonable that any delay in their failing to achieve a wanted pregnancy should initiate prompt referral to an assisted conception unit. A population-based cohort study of males and females registered with the UK Cystic Fibrosis Database suggested that few individuals with CF sought fertility treatment (1% male, 0.5% female).[30] It is clear from the successful outcomes in this study that referral for fertility treatment may be associated with a good outcome. Considerable care should be taken before embarking on possible fertility treatment in the cohort with poor lung function and the shortened life expectancy should be considered. It may be appropriate to seek the views of the local ethical committee. Hypothetical concerns regarding the altered nature of the cervical mucus would seem to make intrauterine insemination a possible first option, before embarking on IVF. Regrettably in our series one woman who had undergone IVF initially had a triplet pregnancy, before a spontaneous reduction to a twin pregnancy occurred. It is unfortunate that the IVF centre failed to appreciate the particular medical issues in her case.

Approximately a third of the women in our series required insulin, a similar proportion as reported by Cheng et al.[26]. Insulin-dependent diabetes is associated with poor pregnancy outcome and tight glucose control is recommended. All women with CF who are not diabetic before pregnancy should have a full glucose tolerance test because the prevalence of gestational diabetes has been reported to be as high as 14%.[21] If preterm delivery is contemplated and antenatal steroids are administered, temporary episodes of hyperglycaemia may occur and a sliding scale insulin regimen may be required to maintain glucose control. Interestingly the mean birthweight centile in our series was 31.9 despite the significant number of diabetic mothers. We hypothesise that this could be because of good diabetic control, possibly secondary to the large number of clinic visits. It is likely though that there are other metabolic factors at work, including possibly the increased work of breathing and the difficulty that some women have in maintaining their calorific requirements. Another common complication is recurrent chest infection. Sixty-nine percent of the women in our series were known to be colonised by Pseudomonas, a common finding. One woman in our series was colonised by Burkholderia cepacia and she sadly died shortly after delivery. However, although the infection has been associated with some poor outcomes, the largest study to date has failed to confirm an association between Burkholderia cepacia colonisation and survival after pregnancy.[20, 31] Although women with CF who become pregnant do not appear to suffer long-term loss of lung function, during the course of their pregnancy they require more hospital visits, suffer more respiratory exacerbations, and require more diabetic surveillance and increased physiotherapy input.[32, 33] This may prove a considerable burden and again these issues should ideally be explained before pregnancy. Respiratory infections should be aggressively managed as continued infection may lead to increased loss of lung function. Many antibiotics have been widely used in pregnancy, but it is sensible to continually review drug regimens. A very helpful summary has been published.[34] Our policy is to routinely admit women with poor lung function at approximately 32–34 weeks of gestation for intense physiotherapy and possible antibiotic therapy before anticipated delivery.

Half of women were delivered by caesarean section. In four (17.4%) women the fetal presentation was breech. A higher number of abnormal presentations is to be expected as the mean gestational age at delivery of these pregnancies was 36 weeks. Despite this the caesarean section rate was higher than in previous studies, which reported rates of 21–27.3%.[19, 23, 24] This may reflect the poorer overall prepregnancy lung function of the women in this series. The lung function of women delivered by caesarean section was significantly worse than in those delivering vaginally. All the women had been seen antenatally by senior obstetric anaesthetists. In all but one woman anaesthesia was regional, the exception being one woman with extremely poor lung function (FEV1 = 22%) who was unable to tolerate positioning for an epidural, required a general anaesthetic and sadly died within 72 hours of delivery.

Modern obstetric and CF management, together with expert neonatal and anaesthetic services, have combined to enable women with CF to have successful pregnancies without apparent significant deterioration of their lung function. There is though a significant increase in treatment burden during and following pregnancy. Pre-existing lung function will determine to a great degree the time some women may spend as parents.

Disclosure of interests

The authors have no conflicts of interest to declare.

Contribution to authorship

JGTB and SM conceived the idea and study design, collected and analysed the data. JGTB prepared the manuscript. DB and SM reviewed and edited the manuscript. All authors contributed to the interpretation of the data and finalising the article to be published.

Details of ethics approval

No ethical approval was required.

Funding

None.

Acknowledgements

The authors thank Mina Savvidou for her statistical advice.

Ancillary