AIM: To investigate associations between perinatal risk factors and subsequent inflammatory bowel disease (IBD) in children and young adults.

METHODS: Record linked abstracts of birth registrations, maternity, day case and inpatient admissions in a defined population of southern England. Investigation of 20 perinatal factors relating to the maternity or the birth: maternal age, Crohn’s disease (CD) or ulcerative colitis (UC) in the mother, maternal social class, marital status, smoking in pregnancy, ABO blood group and rhesus status, pre-eclampsia, parity, the infant’s presentation at birth, caesarean delivery, forceps delivery, sex, number of babies delivered, gestational age, birthweight, head circumference, breastfeeding and Apgar scores at one and five minutes.

RESULTS: Maternity records were present for 180 children who subsequently developed IBD. Univariate analysis showed increased risks of CD among children of mothers with CD (P = 0.011, based on two cases of CD in both mother and child) and children of mothers who smoked during pregnancy. Multivariate analysis confirmed increased risks of CD among children of mothers who smoked (odds ratio = 2.04, 95% CI = 1.06-3.92) and for older mothers aged 35+ years (4.81, 2.32-9.98). Multivariate analysis showed that there were no significant associations between CD and 17 other perinatal risk factors investigated. It also showed that, for UC, there were no significant associations with the perinatal factors studied.

CONCLUSION: This study shows an association between CD in mother and child; and elevated risks of CD in children of older mothers and of mothers who smoked.

Both Crohn’s disease (CD) and ulcerative colitis (UC) are considered to be immune-mediated disorders, although the exact pathogenetic mechanisms are not yet clear. It is thought that a combination of environmental factors in genetically susceptible people lead to disordered immunity and chronic inflammation. Over the last 50 years, there have been large increases in the incidence of CD and UC in the UK, in other western countries[1-3], and more recent increases in Asia[4], which indicate changes over time in the environmental factors that can lead to inflammatory bowel disease (IBD).

In recent decades, there have been changes in the management of births, including large increases in caesarean deliveries, advances in neonatal medicine and substantial reductions in neonatal mortality. As perinatal risk factors have been associated with some immune-mediated diseases including asthma[5-7] and type 1 diabetes[8-10], perinatal risk factors and early life events may be relevant to other immune-mediated diseases, including IBD. One case-control study identified that infectious and non-infectious perinatal health events were linked with 40% of all cases of IBD in the study group[11]. A systematic review and meta analysis of 17 (mainly case-control) studies found a small but significant protective effect of breastfeeding against both CD and UC[12]. However, its authors commented that this finding was far from conclusive, and advocated the need for larger studies. There have been relatively few studies of other perinatal risk factors and IBD.

The aim of this study was to investigate associations between 20 perinatal risk factors and the subsequent development of IBD in children and young adults in a large geographically defined population of South East England. These perinatal risk factors include nine maternal characteristics, such as maternal age, parity, smoking during pregnancy, ABO blood group and social class, and 11 neonatal characteristics, including birthweight, gestational age, head circumference, breastfeeding and Apgar scores.

Ethical approval for analysis of the record linkage study data was obtained from the Central and South Bristol Multi-Centre Research Ethics Committee (04/Q2006/176).

We used the Oxford record linkage study (ORLS). The ORLS comprises abstracts of records of birth registrations, maternities, day cases and inpatient admissions in a defined geographical region of South East England around Oxford. The maternity data covered all National Health Service (NHS) hospitals in two health districts of the ORLS over the 20-year period from 1970 to 1989. The maternity data are linked to data on all inpatient and day case care for all clinical specialties in the ORLS up to 1999, in two health districts of the ORLS from 1970 to 1999 (population 0.9 million) and a further four adjacent districts (total population 1.9 million) from 1975 to 1999. We used the record linked data to identify cases of subsequent IBD in the children covered by the maternity data. We also used the linked data to identify records of IBD in the mothers, before and after childbirth. The original data comprising the ORLS were abstracted from hospital records by staff who were specifically trained for this purpose by senior clinicians.

We excluded maternity records in the ORLS for 985 abortions, 1560 stillbirths and 1567 early deaths that occurred within 30 d of birth. We also excluded 289 births in which the birthweight was recorded as < 1000 g, because most of these records had implausibly low values and/or substantial missing data for many of the perinatal risk factors that we were investigating. None of these excluded babies were subsequently identified as having IBD. After applying these exclusion criteria, a total of 248 659 births remained in the study.

Cases of CD and UC among offspring and mothers were identified using the following ICD codes on inpatient or day case records: 563.0 and 563.1 for CD and UC, respectively (in the ICD-8th revision), 555 and 556 (ICD-9) and K50 and K51 (ICD-10), when recorded in any diagnostic position on the hospital record. There were 114 and 66 children with both a maternity record and a subsequent admission for CD or UC, respectively. We compared 20 perinatal factors studied in these cases with those in the other 248 479 children without a record of inpatient admission or day case care for CD or UC. Parity was defined as the pregnant woman’s number of previous live and still births, as recorded on the ORLS maternity record. The length of “follow-up” for offspring ranged from 30 years for those born in 1970 to 10 years for those born in 1989, with an average follow-up of 18 years.

Statistical methods used included the chi-square test with Yates’ correction, odds ratios (ORs) and their 95% CI, and multivariate logistic regression. Statistical significance was accepted at the conventional 5% level. When using logistic regression, all perinatal risk factors that were significant (P < 0.05) in the univariate analysis were included in an initial model. Each of the factors that were not significant in the univariate analysis were then re-entered, one at a time, into the regression model. This approach was taken to test whether any perinatal factor that was not significant in the initial univariate analysis, became significant when assessed simultaneously with other significant factors in the multivariate analysis. Cases with missing data for the perinatal risk factors were excluded only for those risk terms that were included in the logistic regression model. Year of birth was routinely included in all of the models, as a potential confounder, because of the different periods of follow-up after different years of birth.

For the 114 and 66 children identified with CD and UC respectively, the age at first admission (mean ± SD) was 17.5 ± 5.1 years and 17.7 ± 6.0 years. Approximately half of the cases of both CD and UC had a first recorded admission in early adulthood (when aged 18 to 30 years) rather than in childhood (Table 1). A majority of cases were female; 59 (52%) for CD and 37 (56%) for UC.

A strength of our study is that we have investigated 20 perinatal risk factors for IBD, unlike other studies that have mostly investigated one, two or only a few factors. The study is based on a geographically defined population, covering prospective data collected over 30 years. Another important strength is that information about the perinatal risk factors and the main outcome measure - IBD in offspring - were collected independently of each other. They were subsequently brought together independently through systematic record linkage, such that information collected for each risk factor was not influenced by knowledge of the outcome measure. Our study is therefore not subject to potential interviewer and recall bias, e.g. about whether the mother smoked during pregnancy, which can affect studies based on interviews or self-reporting, and which provide much of the evidence about IBD and perinatal risk factors. The Oxford record linkage study has also been used as the basis of previous studies of perinatal risk factors[7,10,13,14].

The study has several limitations. There was variable follow-up after birth, with shorter durations of follow-up for those born in the more recent years of the study period. However, there was at least 10 years follow-up for all IBD cases among offspring. Maternities during the early years of the study period, and among younger mothers, had fewer years of pre-pregnancy inclusion to ascertain maternal IBD, while data for three of the 20 risk factors (social class, smoking and breastfeeding), were not available for the first four years of the study period. The study would not have identified offspring who were diagnosed with IBD after migrating out of the ORLS region, which would reduce the number of observed cases of IBD.

The identification of cases of IBD in the offspring was restricted to those who were admitted as inpatients or day cases. We will have missed some cases of IBD where the only inpatient or day case admission was for a diagnostic endoscopy and biopsy in patients with suspected IBD, and where the pathology results were not available to create a record of the diagnosis at the time of discharge. We will also have missed people without any day case or inpatient care. Migration over time in the Oxford region population would also have lowered our observed incidence of IBD, particularly among adults. Our age-specific cumulative incidence rates of 1.6 and 0.9 per 100 000 for CD and UC among 0-29 year olds, and 1.0 and 0.5 respectively among 0-19 year olds, are lower than those in some UK studies, but comparable with those in other UK studies[2,3,15-18].

There were two cases of CD in both the mother and child. Previous studies have identified associations between both maternal CD and maternal UC and IBD in offspring[19,20], which are part of a well-established genetic association of IBD in families[21-23]. There was some evidence of increased risks of CD, but not UC, among children whose mothers smoked during pregnancy. Previous studies have reported no association with smoking during pregnancy[19,24], although one case-control study found modest protection against both CD and UC[25]. Our data refers to smoking by the mother: in the general literature, there is strong evidence that active smoking increases the risk and perhaps the severity of CD[26,27]. Although we did not have information on smoking status of the IBD subjects themselves, it is of some interest that we found an association between maternal smoking and CD.

Some studies have identified slightly increased risks of IBD among children from lower socio-economic groups[11,20], others have reported reduced risks of CD among children from lower socio-economic backgrounds[18,28], and some have found no association between social background and IBD[29,30]. More generally, studies that have investigated the relationship between IBD and socio-economic group have often reported conflicting findings[31-37]. Overall, this indicates that any possible association between socio-economic background and subsequent IBD in children is probably quite weak. We did not find a significant association between social class and either CD or UC.

A systematic review and meta analysis of (mainly case-control) studies reported a small but significant protective effect of breastfeeding against subsequent IBD in offspring[12], although it concluded that new larger studies were required. More recent studies have shown little association between breastfeeding and IBD[19], or even increased risks of CD[30,38]. We found no association between breastfeeding at the time of discharge from hospital and IBD, although this includes those who subsequently discontinued breastfeeding after discharge, and is therefore an incomplete marker of breastfeeding.

Through the use of multivariate analysis, we found higher risks of CD, but not UC, among children of older mothers (aged 35+ years). This is consistent with a Swedish study that reported an increased risk of paediatric CD among female offspring born to older mothers[35], although other studies have identified no link between mother’s age and IBD in children[25,39]. Births among older mothers are sometimes associated with increased risks of prenatal medical and obstetric complications, intrapartum complications, perinatal and neonatal morbidity and mortality, as well as increased subsequent risks of various disorders. It is possible that children born to older mothers may be more exposed, or more susceptible, to factors associated with the aetiology of subsequent CD, but not with UC, in their children. It is also possible that our finding on maternal age, though significant, was a chance one, especially as our study investigated 20 perinatal factors. It is worth noting, however, that the finding was highly significant (P < 0.01 in multivariate analysis).

We found an indication of reduced risks of IBD, particularly UC, among children of mothers who were unmarried at the time of birth, which is consistent with findings from Sweden[11] and Australia[28].

It has been suggested that caesarean section might increase the risk of subsequent IBD in children, because there is less exposure to maternal bacteria than in vaginal delivery[40]. The reasoning behind this is that, according to the hygiene hypothesis, inadequate exposure to microorganisms in early life might result in higher levels of immune-mediated pathology in later life. Although one study found an increased risk of CD for elective caesarean sections[28], another found no association for either CD or UC[19]. We also found no association for either CD or UC.

We found no association between maternal parity and IBD. Although increased risks of IBD have been reported occasionally for first born[41,42], or subsequent siblings[43], most studies have found no association between birth order and IBD[11,20,25,30,39]. We also found no association between IBD and any of the other perinatal factors studied, including pre-eclampsia, birthweight, gestational age and Apgar score. These perinatal factors have not usually been associated with IBD in previous studies[11,19,25,28,30].

To summarise, of the 20 perinatal risk factors investigated in this study, we found that maternal CD, smoking during pregnancy and advanced maternal age were associated with increased risks of CD in offspring. For UC, there were no factors associated with increased risks after multivariate adjustment. This, and the fact that the few factors that were associated with CD had quite small effect sizes, suggests that perinatal risk factors have only a minor role in the aetiology of IBD.