Aliment Pharmacol Ther 2011; 33: 333–339
Background Inflammatory bowel diseases are frequent in women at their optimum age for reproduction. Possible effects on pregnancy are therefore of interest.
Aim To assess pregnancy outcomes in 212 women: 135 of whom were diagnosed with inflammatory bowel disease before pregnancy and 77 after pregnancy.
Methods A clinical birth database was examined retrospectively. Odds ratios (ORs) for the main outcomes were calculated with 95% confidence intervals; these were adjusted for confounding factors.
Results Overall, pregnancy outcomes were good. No increased risk of preterm births or need for neonatal intensive care was observed. However, women with inflammatory bowel disease had more growth-retarded newborns, at an adjusted OR of 2.08 (1.26–3.44) and the mode of delivery was more frequently Caesarean section, with an OR of 2.75 (1.82–4.16). In the women who were diagnosed with inflammatory bowel disease after the pregnancy, we found no difference in the obstetric outcome as compared with the general obstetric population.
Conclusions Most women with inflammatory bowel disease have a normal pregnancy outcome and, overall, inflammatory bowel disease during pregnancy poses low-to-moderate risks, the main concern being increased risk of impaired foetal growth. Normal maternal weight gain during pregnancy, possibly indicating a more stable disease, appeared to protect against adverse outcomes.
Inflammatory bowel disease (IBD) affects patients at the age when they are most fertile; the highest incidence is in individuals in their 20s and 30s. The mean age at diagnosis is between 31.4 and 31.8 years.1, 2
For reasons of the nature of IBD and its early age of onset, there has been concern about its effects on pregnancy outcomes. Some studies have shown that IBD is associated with a possible threefold increase in the risk of preterm delivery3, 4 and two- to ninefold risk in the risk of low birth weight (LBW), i.e. under 2500 g.3, 5, 6 Accordingly, a four- to sixfold increase in the risk of foetal growth restriction [small for gestational age (SGA)] has been reported.7, 8 Furthermore, it has been reported that the risk of a Caesarean section (CS) being required is 1.5–9 times greater in women suffering from IBD.1, 3, 7 Reports have even suggested a sevenfold increase in the risk of stillbirth9 and two- to fourfold risk of congenital abnormalities for patients with ulcerative colitis (UC).3, 7
Only a few studies have been published that examine pregnancies before IBD diagnosis in women who eventually develop either UC or Crohn’s disease (CD). These studies suggest that pregnancy outcomes are affected prior to diagnosis. Specifically, the risk of preterm delivery has been reported as being up to 10 times greater10, 11 and the risk of LBW up to eight times greater.12
The objective of this study was to determine the outcomes of pregnancy in women with IBD and to compare the pregnancies of the same women in the time period prior to IBD diagnosis.
A retrospective analysis was conducted of an existing database containing information about patients with bowel diseases, focusing on women of reproductive age. The database comprises information on all patients (n = 1303) diagnosed or treated at Kuopio University hospital for IBD. Data pertaining to family history, the onset of the disease, type and anatomical location of the disease, medication, surgical treatment and complications were collected and updated as part of the hospital’s clinical work. The diagnosis of IBD was routinely confirmed by colonoscopy and biopsies. Data pertaining to all the pregnancies (n = 212) of women on the IBD register were collected from the Kuopio University Hospital birth register, which contains information on all women who gave birth at Kuopio University Hospital between 1989 and 2006. Women with a diagnosis of UC or CD before the year of delivery were included in the UC and CD groups respectively. If the diagnosis of IBD was made during the same year as the delivery or after the delivery, the women were included in the pre-UC and pre-CD groups. These groups were compared with the entire obstetric population, which acted as the reference group, and with each other. For further analysis, the UC and CD subgroups were pooled into an IBD group and the pre-UC and pre-CD subgroups into a pre-IBD group. Numbers of women in each study group are presented in Figure 1.
Exclusion criteria were (i) multiple pregnancies, i.e. twins, triplets and higher-order pregnancies and (ii) major foetal structural anomalies, because such pregnancies carry an unusually high risk of adverse outcome, and the effect of maternal IBD on these pregnancies would be difficult to distinguish. A total of 38 497 pregnancies were included in the analyses.
Information on maternal characteristics was based on data from self-administered questionnaires at approximately 20 weeks of pregnancy. When any data were missing, the gaps were filled in by interviews conducted by a nurse on the delivery ward. The questionnaire consisted of 75 background items. Information on maternal medication, pregnancy complications, pregnancy outcome and the neonatal period was collected in real time, as part of clinical work, by the nurses and midwives who undertook delivery and neonatal care. The Institutional Review Board accepted the study and childbearing women gave informed consent at the time of data collection. The ethics committee has accepted the database and given permission for its use for research purposes. The data were processed in such a way that individuals could not be identified.
Kuopio University Hospital is a tertiary-level obstetric referral centre, but it is also the only obstetric hospital in our district. During data collection, antenatal care was readily and easily accessible to all women. Home deliveries were exceptional, and therefore not included.
Definitions of the background and outcome variables are listed in supporting information Table S1. Estimation of gestational age was based on menstrual history and ascertained by measuring foetal crown–rump length by ultrasonography, in approximately 95% of cases, at 10–12 weeks of pregnancy. If a subject had two adverse pregnancy outcomes, for example LBW and preterm delivery, each was treated as an independent outcome and the subject was included in both categories. The pregnancy complications chorioamnionitis, placental abruption and placenta praevia were recorded when these obstetric diagnoses were made during the hospital stay. Abnormal foetal heart rate [cardiotocogram (CTG)] was recorded in the database by obstetricians. Meconium staining of the amniotic fluid during delivery was noted in the delivery reports and entered into the database by midwives. Apgar scores were assessed mainly by midwifes in uncomplicated deliveries and by paediatricians, when consulted.
Statistical differences between the subjects and the reference group were evaluated using Chi-squared tests (dichotomous variables), and Fisher’s exact test was applied when the minimal estimated expected value was less than five. A value of P < 0.05 was considered statistically significant. Continuous variables were analysed using two-tailed, pooled t-tests and standard deviations (±s.d.) were calculated. Logistic regression analysis, controlling for all possible clinically significant confounding factors, was performed using sas for Windows (SAS statistical package release 8, Cary, NC, USA). In the first analysis, unadjusted odds ratios (ORs) were calculated (OR). The logistic regression models included the variables: age if under 18 years, age if over 35 years, maternal BMI, smoking during pregnancy, primiparity, multiparity and scarred uterus. In the second and third models, the possible causal factors anaemia and poor maternal weight gain were added separately into the analysis. In the fourth model, anaemia and poor maternal weight gain were simultaneously included in the analysis. Confidence intervals (CIs) were evaluated at 95%.
At the University Hospital of Kuopio, the frequency of IBD was 0.35% of all pregnancies progressing beyond 22 weeks; this equated to one in 284 pregnancies. Women, who had a diagnosis of UC or CD prior to pregnancy (IBD group, n = 135) were significantly older than women in the reference group (29.9 ± 5.50 vs. 29.0 ± 5.5 years, P = 0.04). Maternal weight gain during pregnancy was significantly lower in the UC and IBD groups than in the reference group (12.3 ± s.d. 5.8 kg vs. 13.5 ± s.d. 5.6 kg, P = 0.05; 12.3 ± s.d. 6.1 kg vs. 13.5 ± s.d. 5.6 kg, P = 0.01 respectively), although at the beginning of pregnancy, there were no significant differences in the mean maternal BMIs.
Birth weight of newborns in the UC and IBD groups was significantly lower than it was in the reference group (3317 ± s.d. 658 g vs. 3506 ± s.d. 613 g, P = 0.003; 3340 ± s.d. 631 g vs. 3507 ± s.d. 613 g, P = 0.002 respectively). Newborns in the CD group were also smaller than those in the reference group but the difference was not statistically significant.
Reproductive risk factors, along with maternal and pregnancy characteristics, are shown in supporting information Table S2. Women in the UC and IBD groups had undergone more induced abortions prior to the current pregnancy than women in the reference group (19% vs. 11%, P = 0.009; 18% vs. 11%, P = 0.008 respectively). Primiparous women were less frequent in the UC group than in the reference group (28% vs. 41%, P = 0.01). Previous uterine operations were also more common in the UC and IBD groups than in the reference group (21% vs. 10%, P < 0.001; 18% vs. 10%, P = 0.002 respectively). Other maternal characteristics, specifically prior miscarriages, previous stillbirths, age ≥36 years, unmarried status, overweight, smoking, drinking alcohol and hypertension did not differ significantly between the study and reference groups.
Modes of delivery and pregnancy outcomes are shown in supporting information Table S3. Anaemia during the third trimester was more common in women in the UC and IBD groups than in the reference group (6% vs. 2%, P < 0.001; 6% vs. 2%, P < 0.01 respectively). Pregnancy morbidities of the study groups and the reference group were similar as regards gestational diabetes, pre-eclampsia, placenta praevia, chorioamnionitis and placental abruption.
Significantly higher incidence of adverse pregnancy outcomes were recorded in women with diagnosed UC or CD compared with the reference group (supporting information Table S3). CS was a more likely mode of delivery in the UC and IBD groups than in the reference group (34% vs. 17%, P = 0.01; 32% vs. 17%, P < 0.001 respectively). SGA was also more common in the UC and IBD groups than in the reference group (17% vs. 9%, P = 0.014; 16% vs. 9%, P = 0.015 respectively). The need for neonatal intensive care was higher in the CD group than in the reference group (20% vs. 9%, P = 0.02). There was no statistically significant difference in maternal or pregnancy characteristics or in pregnancy outcomes between the group of women diagnosed with IBD after pregnancy and the reference group. No difference between the study groups and the reference group was observed in the incidences of vacuum-assisted delivery, abnormal foetal CTG during delivery, umbilical vein pH at birth, Apgar scores at the age of 5 min, preterm births or LBW under 2500 g. Supporting information Table S4 lists the maternal and foetal anthropometric variables, weight gain and haemoglobin in the study groups compared with the reference group.
The results of the logistic regression analysis models comparing main pregnancy outcomes in the healthy reference mothers and the study groups are presented in Table 1. When adjusted for confounding factors and anaemia and low weight gain, women who had IBD were more likely to require delivery by CS (adjusted OR: 2.75; 95% CI: 1.82–4.16) and LBW for the gestational age (adjusted OR: 2.08; 95% CI: 1.26–3.44).
|Study group||Reference (n = 38 285), n (%)||UC (n = 95), n (%)||Pre-UC (n = 57), n (%)||CD (n = 40), n (%)||Pre-CD (n = 20), n (%)||IBD (n = 135), n (%)||Adjusted OR, IBD compared with reference†|
|Caesarean section||6508 (17)||32 (34)*||9 (16)||11 (28)||2 (10)||43 (32)*||2.75 (1.82–4.16)|
|Premature birth <37 pregnancy weeks||2352 (6)||9 (10)||3 (5)||3 (8)||0 (0)||12 (9)||1.23 (0.60–2.54)|
|Low birth weight <2500 g||1796 (5)||7 (7)||3 (5)||3 (8)||0 (0)||10 (7)||1.83 (0.88–3.80)|
|Small for gestational age||3609 (9)||16 (17)*||7 (12)||5 (13)||2 (10)||21 (16)*||2.08 (1.26–3.44)|
|Neonatal intensive care||3333 (9)||9 (10)||1 (2)||8 (20)*||3 (15)||17 (13)||1.64 (0.96–2.79)|
Of the 135 mothers who had IBD during pregnancy, 70 (52%) had drug treatment either for active disease or as a maintenance medication: mesalazine (mesalamine) in 60 (44%) patients, systemic prednisolone in 19 (14%), sulfasalazine in 10 (7%), azatioprin in 2 (2%) and ciclosporin in 1 (1%) patient.
In subgroup analyses, there was no difference in the incidence of anaemia (7% vs. 5% respectively, P = 0.53) in medicated women compared with women who had no medication; nor in the incidence of LBW under 2500 g (9% vs. 6% respectively, P = 0.59); nor in the incidence of birth weight SGA (14% vs. 17% respectively, P = 0.67).
Inflammatory bowel disease was a factor in 0.35% of all deliveries at the Kuopio University Hospital. Thus, the frequency of IBD is comparable to its prevalence in Europe.13 However, in a clinical setting, the criteria used for diagnosis may vary and some of the cases may go undiagnosed if clinical signs are not present or are mild. Although rare, the conditions involved are of concern to affected women and their doctors.
The main finding of this study was that the overall risk of an adverse obstetric outcome was moderately low, although a doubled risk of foetal growth limitation (SGA) was recorded in affected women. Second, our results showed that the effect on foetal growth was clearly associated with the onset of the disease, as there was no risk of impaired foetal growth in women who were diagnosed with IBD after pregnancy. In other words, genetic factors and foetal growth potential have no major role to play herein. Third, pregnancies proceeded normally with respect to most factors considered, although the mean haemoglobin concentration and maternal weight gain were lower in affected women than in the reference group and this difference was statistically significant. Although these differences were of limited clinical significance, it is noteworthy that the risk of third trimester maternal anaemia was more than three times higher in women with IBD. Fourth, Caesarean deliveries were more common for the affected group than the reference group. Finally, the need for neonatal intensive care was higher among infants born to mothers affected by CD.
A number of studies have suggested that women with IBD during pregnancy are exposed to a substantial risk of adverse outcomes.3, 6 A recent meta-analysis by Cornish et al.3 concluded that women with IBD included in studies published since 2000 were more likely to experience premature birth (OR: 1.99; 95% CI: 1.22–3.26),4, 5, 7, 14, 15 congenital abnormalities in the infant (OR: 2.39; 95% CI: 1.28–4.48)4, 7, 16 and CSs (OR: 1.54; 95% CI: 1.24–1.91),4, 5, 7, 14, 15 but not SGA (OR: 1.79; 95% CI: 0.66–4.82).4, 5, 7, 8, 17, 18 However, two studies that were also included in the meta-analysis did report an increased risk of SGA.7, 8 Premature births occurred up to 30% more often in the current study, but the difference was not statistically significant; in contrast we found that the risk of SGA was significantly increased even after adjusting for confounding factors.
The outcome differences between our results and those in previous studies may be because only women with pregnancies extending beyond 22 weeks were considered in this study. Pregnancies ending with miscarriages in the first or early second trimester were not included, and therefore it is possible that women suffering from severe bowel disease were excluded. On the other hand, the number of previous miscarriages was statistically similar in affected and unaffected women. Another possible explanation is that the disease activity at the time of conception or in the course of pregnancy was not comparable in the different studies, thus affecting the premature birth rates. However, the significantly increased risk of SGA, both unadjusted and adjusted, remains to be explained. Interestingly, the correction for low maternal weight gain and anaemia appeared to reduce somewhat, but not eliminate, the SGA risk. Medication during pregnancy seemed not to be the explanation for increased risk of SGA, as women with no medication had similarly increased risk. Unfortunately, we could not analyse the role of active disease in this retrospective study.
The results of the current study concording that women diagnosed with IBD after the pregnancy are based on a limited population and should be confirmed in future studies with larger populations, reaching higher statistical power.
The clinical importance of an increased rate of SGA in affected births is considerable, as an underweight full-term infant has a four times higher risk of severe morbidity compared with an infant of normal weight.19 Specifically, these infants have increased risks of asphyxia, meconium aspiration and hypoglycaemia.20 In addition, LBW of a full-term infant has been associated with increased incidence of the risk factors associated with coronary heart disease, such as insulin resistance, glucose intolerance, type II diabetes mellitus and elevated blood pressure in adulthood.21–23
Inflammatory bowel disease is usually preceded by a preclinical phase, when the patients are not yet diagnosed and have not yet sought medical help. The preclinical phase of the disease may, however, be associated with health issues. Thus, in this study, we investigated whether preclinical IBD has any impact on pregnancy outcome. We did not observe any differences in the pregnancy outcomes of women who were pregnant before the diagnosis of CD or UC and the reference group. A few other studies have, similarly, found no association between prediagnosis IBD and pregnancy outcome.9, 24 In this study, we had no opportunity to assess pregnancy outcome of the same women before and after IBD diagnosis, the setting possibly offering new perspectives to the problem.
On the other hand, our results contradict a few previous studies that have reported a higher preterm birth rate and lower mean birth weight of infants born to women with preclinical CD compared with a reference group.10–12 In addition, CS has been reported to be more common in women who eventually developed CD or UC than in an equivalent reference group.11 Furthermore, there is one report of a higher rate of preterm deliveries in prediagnosis CD and UC than in a reference population.10
We conclude that, in most cases, the pregnancy outcome of the IBD patients considered was good, with no increase in the rate of preterm births or neonatal intensive care, the associated problems of foetal growth restriction and more CSs being clinically less severe. This result may be partly due to the antenatal care that was readily and easily accessible to all women. Thus, the commonly held fears about childbirth recently reported among IBD patients25 are not justifiable on the basis of the results of the current study. However, pregnancies in women with IBD should be considered associated with a low-to-moderate risk, depending on the activity of the disease. Therefore, monitoring at specialised maternity clinics appears to be advisable, aiming at early recognition and adequate treatment of growth-restricted foetuses.
Declaration of personal interests: Kaisa Raatikainen has served as a speaker for Abbott OY. Markku Heikkinen is an advisory board member for Abbott and MSD Finland. Markku Pajala has served as a speaker for Abbott, MSD and Algol. Seppo Heinonen and Jyrki Mustonen have no competing interests. Declaration of funding interests: None.