The course of inflammatory bowel disease during pregnancy and postpartum: a prospective European ECCO-EpiCom Study of 209 pregnant women

Authors


Summary

Background

The impact of pregnancy on the course of IBD is still controversial.

Aim

To investigate the impact of pregnancy on IBD and to search for factors with potential impact on remission.

Methods

Pregnant IBD women from 12 European countries were enrolled between January 2003 and December 2006 and compared at conception (1:1) with nonpregnant IBD women. Data on disease course were prospectively collected at each trimester during pregnancy and in the postpartum (6 months) using a standardised questionnaire.

Results

A total of 209 pregnant IBD women were included: 92 with Crohn's disease (CD; median age 31 years, range 17–40) and 117 with ulcerative colitis (UC; median age 32 years, range 19–42). No statistically significant difference in disease course during pregnancy and postpartum was observed between pregnant and nonpregnant CD women. Longer disease duration in CD and immunosuppressive therapy were found to be risk factors for activity during pregnancy. Pregnant UC women were more likely than nonpregnant UC women to relapse both during pregnancy (RR 2.19; 95% CI: 1.25–3.97, 0.004) and postpartum (RR 6.22; 95% CI: 2.05–79.3, P = 0.0004). During pregnancy, relapse was mainly observed in the first (RR 8.80; 95% CI 2.05–79.3, P < 0.0004) and the second trimester (RR 2.84, 95% CI 1.2–7.45, P = 0.0098).

Conclusions

Pregnant women with Crohn's disease had a similar disease course both during pregnancy and after delivery as the nonpregnant women. In contrast, pregnant women with ulcerative colitis were at higher risk of relapse during pregnancy and in the postpartum than nonpregnant ulcerative colitis women.

Introduction

Crohn's disease (CD) and ulcerative colitis (UC) are chronic inflammatory diseases with an increasing incidence and prevalence in recent years.[1] The diseases affect mainly young patients; half of them are younger than 32 years of age at the time of diagnosis.[2-4] Pregnancy is a common event among these patients and 25% of women with inflammatory bowel disease (IBD) are expected to become pregnant after the diagnosis with IBD. These patients are often concerned about the influence of pregnancy on their disease.[5]

Although the effect of IBD on pregnancy outcomes is quite well studied,[5-11] there is a lack of studies on the effect of pregnancy on disease course, except for a few older, smaller and retrospective studies.[12-16] These few studies suggest that disease activity at conception is predictive of disease activity during pregnancy, with relapse seen in only one third of women with quiescent disease at conception.[14-18] Conversely, two thirds of patients with active disease at conception will have active disease during pregnancy. Nevertheless, the few previous studies suggest that pregnant women with IBD overall have a similar disease course as nonpregnant IBD controls.[15, 16] Factors with a potential impact on IBD course during pregnancy and postpartum are still not well defined.[12, 13, 15, 16]

Therefore, we decided to conduct a multicentre, prospective cohort study to investigate (i) the impact of pregnancy on the course of IBD during pregnancy and postpartum and (ii) factors with potential impact on maintenance of remission during pregnancy.

Materials and methods

Study design

This is a European, multicentre, prospective cohort study on pregnancy and IBD in which a cohort of pregnant IBD women were originally compared with cohorts of pregnant healthy women and nonpregnant women with IBD.[6] The study aim and protocol were initially proposed and discussed at a meeting of the ECCO in Amsterdam in 2002. After approval by the ECCO EpiCom, 32 centres from 12 European countries joined the study group.

The first part of the project investigating the influence of IBD on pregnancy outcome has recently been published by Bortoli and Pedersen et al.[6] This study is the second part of the project and focuses on the impact of pregnancy on IBD. It was designed as a non-inferiority study based on the hypothesis that pregnant women with IBD flare similar to nonpregnant women with IBD.[15, 16] A power calculation was not performed prior to study initiation.

Study population

Inflammatory bowel disease pregnant women from each participating centre were consecutively enrolled between January 2003 and December 2006 and followed up prospectively during pregnancy and 6 months after delivery or pregnancy termination. At the time of enrolment (conception/first trimester) of IBD pregnant women, a matched cohort of nonpregnant IBD women of same age, disease localisation and disease activity was collected.

Data acquisition

Electronic case report forms (CRFs) for index cases and controls were used to record the required data. The data were collected prospectively by trained physicians from each participating centre at inclusion, and then subsequently every trimester during pregnancy, as well as 6 months after delivery (postpartum) or pregnancy termination. The data were obtained at outpatient visits, by review of patients' medical records and supplied by telephone interview. Completed forms were sent electronically to the central database for data extraction and statistical evaluation. Data on demographics, smoking, disease characteristics, IBD medication, disease activity and pregnancy and newborn outcome were recorded.

Disease activity was assessed at each trimester during pregnancy (9 months: 1st, 2nd and 3rd trimesters) and postpartum (6 months: 1st and 2nd postpartum periods) using the Harvey-Bradshaw Index (HBI)[19] for CD patients and the Simple Clinical Colitis Activity Index (SCCAI)[20] for UC patients and classified into two categories as remission (HBI <5,[12, 19, 21] SCCAI ≤2[20, 22]) or active disease. Increase in disease activity (HBI >5 and SCCAI >2) with a duration of at least 1 week occurring during a trimester of pregnancy or in the postpartum period was considered a relapse.

Similarly, nonpregnant IBD controls were followed up prospectively every 3 months from enrolment up to 15 months after entry into the study. Information on demographics, smoking, disease characteristics and disease activity was collected as for pregnant IBD women.

Sample size

A sample size calculation for the proportions of patients with disease activity during pregnancy (same period for nonpregnant controls) (Table 3, All IBD: trimester 1–3) was retrospectively performed. From this table, the observed proportions of women with disease activity was 54/197 (27%) and 34/197 (17%), for pregnant and nonpregnant women respectively. With a significance level of 5%, and power of 80%, a total of 265 subjects per group were required to show a difference of 10%-points (27–17%). This was calculated using the following formula[23] for the large sample proportions:

display math

Where Zα/2=1.96, Zβ = 0.84, δ=0.1, p1 = 0.27 and p2 = 0.17. With 197 subjects per group, the power to detect the difference of 10%-points was 67%.

Alternatively, using the estimated relative risk of 1.7 (derived from the incidence rates), and 17% as reference proportion, with n = 197/group, the power was 80% to detect the observed difference.

Statistical analysis

Standard descriptive statistics was performed, including frequency distributions for categorical data and calculation of median and range for continuous variables. Differences between groups were analysed by Mann–Whitney U-test for the continuous data and Chi-Square/Fisher's exact test for categorical data.

Relative risk (RR) rates were obtained by comparing the incidence risk (IR) rates of disease activity between pregnant and nonpregnant women (time periods: 1st, 2nd, 3rd trimesters during pregnancy and 1st and 2nd postpartum). Kaplan–Meier curves with accompanied logrank test were used for time to relapse. Cox proportional hazard model with adjusted hazard ratio (HR) was used to analyse the effect of predictors of time to relapse.

The impact of all covariates, including disease-specific parameters, on IBD activity was analysed by a standard logistic regression model for the cases only. Due to sparse data for some of the potential prognostic factors, the association between the parameters of interest and all prognostic factors was additionally investigated by Fisher's Exact Test (for the cases only). P values were two sided and P < 0.05 was considered statistically significant.

Descriptive and comparative analyses were done by spss version 20 and sas version 9.2.

Ethical considerations

The Ethical Committee of the Coordinating Center in Rho, Italy approved the study protocol. All patients included in this study signed an informed consent.

Results

A total number of 520 pregnant women with IBD were enrolled: 244 with CD, 264 with UC and 12 with indeterminate colitis (IC). However, only 217 were matched by age, localisation and disease activity at conception to nonpregnant IBD women for the control cohort and considered to be eligible for the study. Four IBD women were subsequently excluded due to missing data on disease activity. Women with IC (4) were not included in the data analysis due to their small number. Hence, the final study population comprised 209 IBD pregnant women, 92 with CD and 117 with UC, and 209 nonpregnant IBD women (Figure 1). Patients and controls recruited from each participating country are indicated in Table S1. Patients' demographic and clinical characteristics are outlined in Table 1. One hundred and twenty-five (60%) pregnant women were enrolled from Italian centres and 84 (40%) from other European centres. No significant difference in disease characteristics between Italian and non-Italian patients was observed in UC. In CD, more non-Italian pregnant women had rectovaginal/perianal fistula prior to enrolment than those from Italy (6% vs. 32%, P = 0.001), otherwise no differences were seen.

Table 1. Demographic and clinical characteristics of pregnant and nonpregnant women with inflammatory bowel disease at study enrolment
 CD (n = 92)CD controls (n = 92) P UC (n = 117)UC controls (n = 117)P value
  1. CD, Crohn's disease; UC, ulcerative colitis.

  2. a

    Median, range.

Agea31 (17–40)31 (16–46)0.5432 (19–42)32 (18–42)0.56
Smoking (%)13 (14)19 (21)0.216 (5)13 (11)0.55
Disease duration in monthsa83 (1–307)68 (1–276)0.1460 (1–234)61 (1–291)0.87
Disease location/extension (%)
Small bowel26 (28)32 (35)0.560.60
Colon14 (15)15 (16)  
Colon + small bowel52 (57)45 (49)  
Extensive colitis 32 (27)33 (28) 
Left-sided colitis 37 (32)43 (37) 
Proctosigmoiditis 48 (41)41 (35) 
Previous intestinal surgery (%)33 (36)25 (27)0.203 (3)3 (3)0.99
Previous perianal/rectovaginal fistula (%)15 (16)12 (13)0.40 
Figure 1.

Patients' enrolment. Pregnant women with inflammatory bowel disease (IBD) 1:1 with nonpregnant IBD controls at conception matched during first trimester by age, diagnosis, disease activity and localisation; CD, Crohn's disease; UC, ulcerative colitis; IC, indeterminate colitis.

Treatment at conception and during pregnancy

Seventy-one (77%) pregnant women with CD received any therapy at conception and 60 (65%) continued the therapy during pregnancy. In UC, 98 (84%) women received any therapy at conception and 100 (86%) were treated during pregnancy. No significant difference in treatment between Italian and non-Italian UC patients was observed. In CD, women from Italian centres received less immunosuppressive therapy and more 5-ASA than patients from non-Italian centres (23% vs. 39%, P = 0.005 and 66% vs. 36%, P = 0.004 respectively). The proportion of women treated with particular medications is shown in Table 2.

Table 2. Therapy at conception and during pregnancy in women with Crohn′s disease (CD) and ulcerative colitis (UC)
 Conception1st trimester2nd trimester3rd trimester
CDUCCDUCCDUCCDUC
  1. 5-ASA, 5-aminosalicylic acid.

Corticosteroids systemic (%)3 (3)4 (3)4 (4)5 (4)5 (5)7 (6)3 (3)4 (3)
Corticosteroids topical (%)8 (9)20 (17)2 (2)6 (5)2 (2)7 (6)2 (2)4 (3)
5-ASA/sulfasalazine (%)49 (53)79 (68)37 (40)72 (62)33 (36)70 (60)30 (33)68 (58)
5-ASA topical (%)8 (9)24 (21)1 (1)28 (24)1 (1)28 (24)1 (1)23 (20)
Azathioprine (%)27 (29)9 (8)23 (25)8 (7)19 (21)6 (5)17 (19)6 (5)
Infliximab (%)2 (2)
Ciclosporin (%)1 (1)

Three pregnant women had interventions during the study period. A woman with CD had an incision and drainage of a perianal abscess during her third trimester. The patient delivered at term, a healthy 3300 grams baby boy. Another woman with CD underwent an incision of an abdominal abscess at 13th week of gestation. Both women were in remission at conception. In UC, one woman had a colectomy with an ileal-pouch anal anastomosis 3 months after delivery due to high-grade colonic dysplasia.

Disease activity in Crohn's disease

Eighty-two (89%) pregnant women with CD as well as nonpregnant women with CD were in remission and 10 (11%) had active disease at conception. Of the pregnant patients in remission, 66 (81%) were still in remission by the end of pregnancy, while 15 (19%) had relapsed. In one woman, data on disease activity during pregnancy were missing. Among nonpregnant patients, 65 (82%) women maintained remission and 14 (18%) had become disease active by the end of third observational trimester. Three women had incomplete data on disease activity. The evolution of disease activity during pregnancy and in the 6-month postpartum period according to disease activity at conception and in the third trimester, respectively, is shown in Figure 2. No significant difference in disease activity at conception, during pregnancy and during the first postpartum period between Italian and non-Italian CD pregnant and CD nonpregnant women was observed. In the second postpartum period, Italian CD pregnant women were less active than non-Italian (7% vs. 37%, P = 0.002).

Figure 2.

(a) Evolution of disease activity during pregnancy/9-month follow-up period in pregnant women with Crohn's disease (CD) and their nonpregnant CD controls with respect to disease activity at conception. Missing data on disease activity during pregnancy: in two pregnant women and in three controls. Percentage is calculated from valid data. (b) Evolution of disease activity after delivery (4th–5th observational postpartum trimesters) in pregnant women with CD and their nonpregnant CD controls with respect to disease activity at the end of pregnancy/3rd observational trimester. Missing data on disease activity: in three and eight pregnant women at 3rd trimester and postpartum, respectively; in four and eight controls at 3rd trimester and postpartum respectively. Percentage is calculated from valid data.

Regarding women with quiescent disease at conception, no significant difference in risk of relapse during pregnancy and in the postpartum period was revealed between pregnant CD women and nonpregnant women (Table 3).

Table 3. Disease activity during pregnancy and postpartum period in pregnant women with inflammatory bowel disease (IBD) compared with nonpregnant IBD controls
PeriodPregnantNonpregnant 
N/nPYIncidence ratesN/nPYIncidence ratesRelative riskP value
IR (95% CI)IR (95% CI)RR (95% CI)
  1. N, number of subjects; n, number of subjects with disease activity; PY, patient years measured as trimesters.

  2. Incidence rates (IR) express the incidence rate of disease activity: IR through trimester 1–5 of pregnancy and postpartum period; IR through trimester 1–3 of pregnancy period; IR trough trimester 4 (1st) to 5 (2nd) of postpartum period (PP) and IR per each trimester (1st, 2nd, 3rd trimesters during pregnancy and 1st and 2nd postpartum).

  3. a

    Statistically significant results.

  4. b

    Crohn's disease (excluding colonic disease).

All IBD
Trimester 1–5197/767790.09 (0.08–012)197/408510.05 (0.03–0.06)2.08 (1.4–3.12)0.001a
Trimester 1–3197/545230.10 (0.08–013)197/345580.06 (0.04–0.09)1.69 (1.08–2.68)0.02a
Trimester 4–5138/222560.09 (0.05–0.13)151/62930.02 (0.01–0.04)4.20 (1.65–12.7)0.001a
Trimester 11977196020.03 (0.02–0.05)197/36610.005 (0.00–0.01)6.95 (2.05–36.7)0.001a
Trimester 2177/284610.06 (0.04–0.09)190/185020.04 (0.02–0.06)1.69 (0.9–3.25)0.08
Trimester 3149/72890.02 (0.01–0.05)171/133530.04 (0.02–0.06)0.66 (0.22–1.77)0.38
Trimester 4 pp138/151980.08 (0.04–0.12)151/41780.02 (0.01–0.06)3.37 (1.07–14)0.02
Trimester 5 pp138/71180.06 (0.02–0.12)142/21420.01 (0.00–0.05)4.21 (0.8–41.6)0.06
Crohn's disease
Trimester 1–582/263420.08 (0.05–011)82/183450.05 (0.03–0.08)1.46 (0.77–2.82)0.22
Trimester 1–382/152280.07 (0.04–011)82/142280.06 (0.04–0.10)1.07 (0.48–2.4)0.86
Trimester 4–563/111140.09 (0.05–0.17)61/41170.03 (0.00–0.09)2.82 (0.84–12.2)0.07
Trimester 182/42650.02 (0.00–0.04)82/12670.004 (0.00–0.02)4.03 (0.4–198)0.22
Trimester 277/81990.04 (0.02–0.08)78/102050.05 (0.01–0.09)0.82 (0.28–2.32)0.69
Trimester 369/31320.02 (0.00–0.07)68/31380.02 (0.02–0.06)1.05 (0.14–7.81)0.96
Trimester 4 pp63/8990.08 (0.03–0.16)61/3760.04 (0.00–0.12)2.05 (0.49–12)0.30
Trimester 5 pp51/3510.06 (0.01–0.17)56/1560.18 (0.00–0.10)3.29 (0.26–173)0.33
Crohn's diseaseb
Trimester 1–573/223040.07 (0.05–011)73/163070.05 (0.03–0.08)1.39 (0.7–2.83)0.32
Trimester 1–373/132020.06 (0.03–0.11)73/122020.06 (0.03–0.10)1.08 (0.46–0.61)0.85
Trimester 4–556/91020.09 (0.04–0.17)55/41050.04 (0.01–0.10)2.32 (0.65–10.3)0.16
Trimester 173/42360.02 (0.00–0.04)73/12380.01 (0.00–0.02)4.03 (0.4–199)0.22
Trimester 268/71770.04 (0.02–0.08)69/91830.05 (0.02–0.09)0.8 (0.25–2.41)0.68
Trimester 361/21170.02 (0.00–0.06)60/21200.02 (0.00–0.06)0.03 (0.07–14.1)0.98
Trimester 4 pp56/7860.08 (0.03–0.17)55/3700.04 (0.01–0.13)1.80 (10.4–11.4)0.37
Trimester 5 pp46/2460.04 (0.01–0.16)50/1500.02 (0.00–0.11)2.17 (0.11–12.2)0.58
Ulcerative colitis
Trimester 1–5111/494180.12 (0.09–0.15)111/224860.05 (0.03–0.07)2.59 (1.54–4.5)0.001a
Trimester 1–3111/392830.14 (0.10–0.19)111/203180.06 (0.03–0.10)2.19 (1.25–3.97)0.004a
Trimester 4–571/101350.07 (0.04–0.14)86/21680.01 (0.01–0.04)6.22 (1.33–58.4)0.008a
Trimester 1111/153220.05 (0.03–0.08)111/23780.005 (0.00–0.02)8.80 (2.05–79.3)0.001a
Trimester 296/202510.08 (0.05–0.12)108/82850.03 (0.02–0.06)2.84 (1.2–7.45)0.009a
Trimester 376/41500.03 (0.01–0.07)99/102070.05 (0.00–0.09)0.55 (0.13–1.91)0.33
Trimester 4 pp71/6920.07 (0.02–0.14)86/1980.01 (0.01–0.06)6.39 (0.78–294)0.06
Trimester 5 pp64/4640.06 (0.02–0.16)82/1820.01 (0.00–0.07)5.13 (0.51–252)0.14

The cumulative probability of maintaining remission during pregnancy was 81% in pregnant CD women and 82% in controls (P = 0.20), whereas in the 6 months postpartum, the cumulative probability was 71% in CD pregnant women and 78% in nonpregnant CD controls (P = 0.20) (Figure 3).

Figure 3.

Kaplan–Meier curve for survival of remission of Crohn′s disease (CD) pregnant women and CD nonpregnant controls during pregnancy and in the postpartum period. Only patients in remission at conception/enrolment were included.

Survival analyses identified longer disease duration (>5 years) to be a risk factor of disease relapse during pregnancy and postpartum (HR 2.20; 95% CI: 1.05–4.61, P = 0.04). No impact of pregnancy and other factors such as smoking, history of surgery, perianal disease and disease localisation on disease activity was observed (Table 4).

Table 4. Risk factors of disease activity in pregnant and nonpregnant women with inflammatory bowel disease through five trimesters
 Risk factor for disease activityHR (95% CI)P value
  1. L1: Small bowel; L2: Colon; L3: Small bowel and colon; E1: Proctitis; E2: Left sided colitis; E3: Extensive colitis.

  2. Hazard ratios were adjusted for the factors common for both pregnant and nonpregnant (smoking, previous intestinal surgery, disease location/extension and previous perianal/rectovaginal fistula in case of CD).

  3. a

    Statistically significant results.

Crohn's diseasePregnancy (pregnant vs. nonpregnant)1.50 (0.81–2.79)0.20
 Disease duration (≥5 years)2.20 (1.05–4.61)0.04a
 Disease localisation (L1 ± L3 vs. L2)0.64 (0.27–1.48)0.29
 Previous perianal/rectovaginal fistula1.74 (0.76–3.98)0.19
 Previous intestinal surgery0.080 (0.41–1.56)0.50
 Smoking at conception0.62 (0.24–1.61)0.33
Ulcerative colitisPregnancy (pregnant vs. nonpregnant)2.74 (1.61–4.65)0.001a
 Disease duration (≥5 years)0.87 (0.54–1.40)0.56
 Disease localisation (E3 vs. E2 ± E1)0.91 (0.53–1.56)0.72
 Smoking at conception1.15 (0.46–2.91)0.76

Disease activity in ulcerative colitis

One hundred and eleven (95%) pregnant and nonpregnant women with UC were in remission and six (5%) had active disease at conception.

Among 111 pregnant UC patients in remission, 72 (65%) women maintained remission while 39 (35%) experienced disease activity during pregnancy. Among nonpregnant women, 89 (82%) were still in remission and 20 (18%) had relapsed by the end of their third observational trimester. Two nonpregnant women had incomplete data on disease activity in third trimester. The evolution of disease activity during pregnancy and in the 6-month postpartum period according to disease activity at conception and, respectively, in the third trimester is shown in Figure 4. No significant difference in disease activity at conception, during pregnancy and postpartum between Italian and non-Italian pregnant and nonpregnant UC women was observed.

Figure 4.

(a) Evolution of disease activity during pregnancy/9-month follow-up in pregnant women with ulcerative colitis (UC) and their nonpregnant UC controls with respect to disease activity at conception. Missing data on disease activity during pregnancy: in two controls. Percentage is calculated from valid data. (b) Evolution of disease activity after delivery (4th–5th observational postpartum trimesters) in pregnant women with UC and their nonpregnant UC controls with respect to disease activity at the end of pregnancy/3rd observational trimester. Missing data on disease activity: in four pregnant women at 3rd trimester and postpartum; in two and seven controls at 3rd trimester and postpartum respectively. Percentage is calculated from valid data.

However, there was a significantly higher risk of disease relapse both during pregnancy, particularly in the first and second trimester, and postpartum period in pregnant UC women as compared with nonpregnant UC women (Table 3).

Similarly, pregnant women with UC had a significantly lower cumulative probability of maintaining remission during pregnancy and postpartum as compared with their controls with 65% vs. 82% maintaining remission during pregnancy (P < 0.0001) and 60% vs. 81% maintaining remission during the 6 months after pregnancy termination (P < 0.0001) (Figure 5). Survival analyses identified pregnancy to be a risk factor for disease relapse in women with UC (HR 2.74; 95% CI: 1.61–4.65, P = 0.002) (Table 4).

Figure 5.

Kaplan–Meier curve for survival of remission of ulcerative colitis (UC) pregnant women and UC nonpregnant controls during pregnancy and in the postpartum period. Only patients in remission at conception/enrolment were included.

Factors with potential impact on disease course (among patients in remission)

In CD, pregnant women with longer disease duration (≥5 years) had an increased risk of relapse at any time during the pregnancy than women with shorter CD duration (OR 14.3, 95% CI: 1.0–194. 7, P = 0.04). Similarly, women on immunosuppressive therapy with or without 5-ASA were more likely to relapse during pregnancy than those treated only with 5-ASA preparations (OR 11.7; 95% CI: 1.5–90.3, P = 0.02). Among patients with UC, none of the assessed factors predicted disease course during pregnancy or postpartum.

Pregnancy outcome

There were 77 (84%) live births in women with CD, 70 at-term and 7 preterm deliveries. Eight (9%) pregnancies ended in spontaneous and 5 (5%) in therapeutic abortions. Two women had missing data on pregnancy outcome. In UC, 109 (92%) women had live births, of which 100 (85%) were at-term and 9 (8%) preterm. Seven women (6%) had spontaneous abortions and one woman had missing data on pregnancy outcome.

Congenital abnormalities were reported in three newborns of CD women (one kidney abnormality and two cases of hipospady). No abnormalities were observed in newborns of women with UC. Comparison results of pregnancy outcome between CD and UC show no difference in live birth (P = 0.06), preterm deliveries (P = 0.72), abortions (P = 0.06) and birthweight (P = 0.13) in both groups. More congenital abnormalities (P = 0.04) and caesarean sections (P = 0.03) were registered in CD.

Pregnancy outcome in 72 (61%) UC and 65 (72%) CD women with no any activity during pregnancy was compared with pregnancy outcome in 45 (39%) UC and 25 (28%) CD women with activity at any time during pregnancy. The results show no statistically significant difference outcome of pregnancy between these two groups in both UC and CD, regarding abortions (P = 0.70 and P = 0.30), preterm delivery (P = 0.80 and P = 0.34), means of delivery (P = 0.55 and P = 0.38), live births (P = 0.70 and P = 0.30) and birthweight (P = 0.41 and P = 0.56). There was no significant difference in congenital abnormalities between CD women with and without active disease during pregnancy (P = 1.00).

Excluded population

From a total number of 520 enrolled women, 311 (60%) were excluded due to missing nonpregnant women for the control cohort (n = 303; 58%), missing data on disease activity (n = 4; 1%) and due to diagnosis of IC (n = 4; 1%).

Demographic and clinical characteristics of the excluded pregnant IBD women are outlined in Table 5.

Table 5. Demographic and clinical characteristics of excluded pregnant women with inflammatory bowel disease
 CD included (n = 92)CD excluded (n = 152)P valueUC included (n = 117)UC excluded (n = 151)P value
  1. CD, Crohn's disease; UC, ulcerative colitis.

  2. Bold: statistically significant result.

  3. a

    Median, range.

Agea31 (17–40)30 (16–43)0.4432 (19–42)31 (19–41)0.49
Smoking (%)13 (14)28 (18) 6 (5)14 (9)0.25
Disease duration in monthsa83 (1–307)72 (1–468)0.3760 (1–234)61 (1–291)0.73
Disease location/extension (%)
Small bowel26 (28)46 (30)0.640.11
Colon14 (15)28 (18)  
Colon + small bowel52 (57)76 (50)  
Extensive colitis 32 (27)63 (42) 
Left-sided colitis 37 (32)39 (26) 
Proctosigmoiditis 48 (41)48 (32) 
Missing2 (2)    
Previous intestinal surgery (%)33 (36)57 (35)0.163 (3)5 (3)0.25
Previous perianal/rectovaginal Fistula (%)15 (16)17 (11)0.34
Disease activity at conception10 (11)16 (11)1.006 (1)23 (15)0.09
Therapy at conception (%)
Any IBD therapy at conception77 (84)110 (72) 0.03 93 (82)120 (80)1.00
Discontinuing therapy during conception/first trimester23 (29)22 (15) 0.04 8 (7)18 (12)0.64
Corticosteroids systemic3 (3)9 (6)0.554 (3)9 (6)0.42
5-ASA/sulfasalazine49 (53)76 (50)0.6979 (68)97 (64)0.89
5-ASA topical8 (9)7 (5)0.2924 (21)37 (25)0.53
Azathioprine27 (29)29 (29)0.469 (8)17 (11)0.39
Methotrexate1 (1)
Infliximab2 (2)7 (5)
Adalimumab1 (1)
Ciclosporin1 (1)

No significant differences in disease characteristics (including disease activity) between included and excluded pregnant women were observed except for CD included pregnant women who had more IBD therapy at conception (P = 0.03) and discontinued their therapy more often during the first trimester than those excluded from the analyses (P = 0.04).

Discussion

This is the first prospective cohort study on course of IBD in pregnant women. In this study, IBD pregnant women were matched 1:1 at conception/during first trimester with nonpregnant IBD women and follow-up was performed prospectively every trimester during pregnancy and 6 months postpartum. The study shows that CD pregnant women have a similar disease outcome during pregnancy as CD nonpregnant women, suggesting that pregnancy has no influence on the disease course in CD. A trend to an increased risk of relapse in CD in postpartum was observed, however, excluding the women with colonic disease from the analysis, shows no significant difference in relapse risk between cases and controls, Table 3. This may indicate that colonic CD to some extent behaves like UC.

In contrast, pregnant women with UC were more likely to relapse both during pregnancy, particularly during the first and second trimester, and in the first 6 months postpartum. This is the first prospective study to confirm higher relapse rate in UC patients during pregnancy.

Logistic regression analysis identified longer disease duration (≥5 years) and the use of immunosuppressive therapy (±5-ASA) to be risk factors for relapse during pregnancy in CD. Longer disease duration has generally been considered to be a predictor for a complicated disease course in CD.[24, 25] Furthermore, the use of immunosuppressive therapy indicates that the CD women might have had a more complicated disease. No factor with potential impact on disease activity during pregnancy was observed in women with UC.

The study has several important strengths. First, large-scale prospective cohort studies of pregnancy in IBD are lacking and this design currently offers the best evidence level[26] when it comes to studying disease outcome in pregnant women with IBD. Second, this prospective multicentre study is the first to examine the influence of multiple factors on the risk of relapse. Third, cases were matched to a cohort population by three parameters (age, disease location and disease activity) known to influence disease outcome, making the results more robust.

Previous studies have suggested that for women who conceive while in remission, the risk of relapse during pregnancy is approximately 30%. Conversely, two thirds of patients with active disease at conception are suggested to remain active during pregnancy.[12, 14-17] However, these data are derived from older, small and retrospective studies[14-16] and none was controlled for relapse risk factors, apart from smoking[12] and nutritional status[13] at conception.

The majority of CD (89%) and UC (95%) women in our study were in remission at conception. This suggests that conception was mainly planned in accordance with current guidelines.[26, 27] Moreover, 4/5 of CD and 2/3 of UC women who were in remission at conception remained in remission during pregnancy, which for CD is better than what has been reported in earlier studies.[14, 16] Furthermore, 80% of CD and UC women also maintained remission into the postpartum period. In contrast, when the disease was active at conception, 56% of CD and 33% of UC women had persistently active disease during pregnancy. However, the number of women with active disease at conception in our study was too small to draw any reliable conclusions. The majority of pregnant women in our study with CD (77%) and UC (84%) took medication (mainly 5-ASA and/or immunomodulators) at the time of conception, whereas 65% continued the therapy during pregnancy in CD and 86% in UC; all of which could have contributed to the favourable results in our study. Improved adherence to medication during IBD pregnancies has been observed in recent studies.[28, 29] The finding of a predominant use of 5-ASA preparations during pregnancy is in accordance with studies showing its safety for the foetus.[30] A minor use of immunosuppressive (29%) and biological (2%) therapy during pregnancy was registered in our cohort. This indicates that the disease was rather mild and well controlled at conception in the majority of women with limited need for immunosuppressive or biological therapy. Still thiopurines have been shown to be efficient in maintenance of remission during pregnancy and theratogenic effects have not been a great concern.[31-34] Likewise, an increasing experience with the use of anti-TNF drugs during pregnancy has not suggested any significant short-term complications in newborns.[35-40]

Ulcerative colitis pregnant women in our study had a tendency to relapse during the first and second trimester, which is consistent with earlier findings.[15] The reason why pregnant UC women relapse more in the first and second trimesters and why they had decreased remission rates during pregnancy and postpartum, when compared with non-IBD pregnant women, remains unclear. It can be speculated if the shift of T helper 1 (Th1) to T helper 2 (Th2) is the reason for UC patients to relapse during pregnancy. Pregnancy is a complex immunological state in which a bias towards Th2 protects the foetus.[41, 42]

In addition, it is possible that nonpregnant women might have had a milder disease or have been given different treatment; but because disease severity and therapy data on controls were not collected, such explanations cannot be confirmed. However, in accordance with our findings, preliminary results from a recent Dutch study by de Lima et al.[26] on disease relapse rates during pregnancy showed that relapse rates in women with UC were higher compared with CD. This centre has a well-organised preconception control of IBD women entering into a pregnancy at a well-controlled disease activity stage of the disease.[27] Furthermore, cessation of smoking in UC women might have had an impact on disease activity during pregnancy. However, data on smoking were collected only for conception period and we do not have data on the rate of smoking cessation during pregnancy.

Some previous studies investigating the impact of pregnancy on long-term disease outcome have shown a milder disease course with fewer relapses in the years following pregnancy.[13, 43] As our study followed the women only for 6 months postpartum, this finding could not be verified.

Despite increased risk of disease activity in UC, pregnancy outcomes in active UC patients were quite favourable comparing with UC without any activity during pregnancy. In the parallel study of Bortoli et al.,[6] the authors found that pregnancy outcome in both CD and UC did not differ from pregnancy outcome in healthy controls. Disease activity in the same study was found to be a factor risk for low birth weight. However, this was not the case in this study.

The major limitation of our study was a large exclusion of originally enrolled patients due, primarily, to a lack of comparable nonpregnant IBD women for the control cohort. The multiple criteria we required for matching may have contributed to drop-out. Importantly, included and excluded women were similar on all parameters. Another potential limitation was the lack of recording of data on medication taken by IBD pregnant women in the postpartum period; data on medication in IBD nonpregnant controls and data on disease behaviour in CD patients. Immunosuppressive medication use in pregnant women with CD was found to increase the risk of relapse during pregnancy. However, the medication use was not possible to search for in the controls due to lack of these data. This raises the concern about selection bias, although patients and controls were entered in the study consecutively. Lastly, a limitation was the lack of a sample size calculation prior to study start, even though this is the largest prospective cohort study of pregnant IBD women hitherto performed. A retrospective power analyses revealed that both the CD and the UC study populations were somewhat underpowered to detect differences between pregnant and nonpregnant patients. This did, however, not hinder the observation of differences among pregnant and nonpregnant UC patients, whereas the absence of an observed difference in disease activity between pregnant CD women and controls does not exclude that such a difference may actually exist, although probably of limited magnitude.

Also, it should be noted that as the included IBD population had rather mild disease activity (low use of immunosuppressives and biological); results might not be generalised to the whole population.

In conclusion, this is the first prospective European cohort study performed among pregnant women with IBD with the aim of examining disease outcome and factors influencing it. The majority of women were in remission at conception and received maintenance therapy. CD women had a similar disease course – both during pregnancy and postpartum – as their age and disease matched nonpregnant CD women, whereas pregnant UC women had a higher risk of relapse during the first and the second trimesters of their pregnancy and postpartum than nonpregnant UC women.

Authorship

Guarantor of the article: N. Pedersen and P. Munkholm.

Author contributions: The study was initiated by ECCO EpiCom. AB, NP and an ECCO collaborative group performed the data collection. NP, DD and JKJ (Statistician) performed the statistical analyses. NP and TJ drafted the manuscript, which was critically revised by all co-authors. All authors approved the final version of the manuscript.

Acknowledgements

Declaration of personal interests: None.

Declaration of funding interests: Unrestricted grant support has been received from Aage and Johanne Louise-Hansens; Aase and Ejnar Danielsen Funds in Denmark and the ECCO. The study sponsors have no contribution in the study design, analysis, interpretation of data and publication.

Ancillary