Paediatric acute lymphoblastic leukaemia and caesarean section: A report from the United Kingdom Childhood Cancer Study (UKCCS)

Abstract Background Reports have suggested that children born by caesarean initiated before labour onset may be at increased risk of developing acute lymphoblastic leukaemia (ALL). However, with most data being derived from case‐control study interviews, information on the underpinning reasons for caesarean section is sparse, and evidence is conflicting. Objectives Use clinical records compiled at the time of delivery to investigate the association between childhood ALL and caesarean delivery; examining timing in relation to labour onset, and reasons for the procedure. Methods Data are from the UK Childhood Cancer Study, a population‐based case‐control study conducted in the 1990s, when caesarean section rates were relatively low, in England, Scotland, and Wales. Children with ALL were individually matched to two controls on sex, date of birth, and region of residence. Information on mode of delivery and complications was abstracted from obstetric records. Odds ratios (OR) and 95% confidence intervals (CI) were calculated using logistic regression models adjusted for matching variables and relevant covariates. Results Around 75% of the 1034 cases and 1914 controls were born through unassisted vaginal delivery. Caesarean delivery was as frequent in cases and controls (OR 1.07, 95% CI 0.84, 1.36). No association was observed between ALL and caesarean delivery either during or before labour, with adjusted ORs of 1.08 (95% CI 0.78, 1.48) and 1.09 (95% CI 0.78, 1.53), respectively. For B‐cell ALL, the ORs were 1.14 (95% CI 0.81, 1.59) for caesarean during labour and 1.21 (95% CI 0.85, 1.72) for prelabour. The underpinning reasons for caesarean delivery differed between cases and controls; with preeclampsia, although very rare, being more common amongst cases born by caesarean (OR 8.91, 95% CI 1.48, 53.42). Conclusions Our obstetric record‐based study found no significant evidence that caesarean delivery increased the risk of childhood ALL, either overall or when carried out before labour.


| BACKG ROU N D
Acute lymphoblastic leukaemia (ALL) is the commonest paediatric malignancy, representing around a third of cancers diagnosed in children (0-14 years) in high-income countries. The characteristic incidence peak between 2 and 5 years of age 1 has provided the foundation for several aetiological hypotheses, most notably, concerning a potential role of exogenous factors on the emergent immune system and subsequent risk of ALL. 2 In this context, it has been suggested that because children born by caesarean delivery may not be subject to the same hormonal and microbial challenges as those born vaginally, their immune development could be altered, which in turn could increase their risk of ALL. Although findings from the majority of studies examining the relationship with caesarean delivery have provided little support for this hypothesis, [3][4][5][6][7][8][9] some investigators observed that children delivered by caesarean, 10 especially if performed before the onset of labour, 11,12 may be at increased ALL risk. A further case-control study, finding no overall association with caesarean delivery, either before or during labour, reported an increased risk with prelabour caesarean among children diagnosed with ALL before three years of age 13 ; and a Californian birth record linkage study reported a 20% increased risk of ALL with caesarean delivery (pre-and post-labour combined) in children aged 2-4 years. 11 The Childhood Leukaemia International Consortium (CLIC) pooled analysis used algorithms to classify the, mostly self-reported, mode of delivery across the 13 included studies. 14 Four of these asked mothers about the reason for the caesarean; categorizing caesareans as "prelabour" if the reason given was either previous caesarean or multiple birth. 14 Likewise, the Californian record linkage study used "elective" as a marker for prelabour. 11 With a view to investigating the association between caesarean section and ALL in more depth, this report presents findings from a detailed examination of obstetric records collected during the United Kingdom Childhood Cancer Study (UKCCS). 15 Conducted in the early 1990s, the median year of birth was 1989, a time when caesareans represented only 12% of deliveries in the UK, 16

| ME THODS
Data are from a population-based case-control study specifically designed to examine the potential aetiological role of a range of potential risk factors, including perinatal and reproductive events (UKCCS). 15,19 This report used obstetric records of 1034 mothers of children diagnosed with ALL (cases) in England and 1914 mothers of children without cancer (controls); each case being individually matched to at least one control on sex, date of birth, and region of residence. Data relating to mode of delivery, timing (before or during labour), and reasons for caesarean delivery were extracted directly from obstetric records. Detailed information about abstraction methods is presented elsewhere. 15 Caesarean delivery was classified as prelabour if the medical record stated that labour had not started before it was performed.

| RE SULTS
Characteristics of cases and controls are presented in Table 1. As expected, children who developed ALL were, on average, slightly heavier at birth than controls, mostly due to B-cell ALL (adjusted OR 4000 g or more 1.32, 95% CI 1.02, 1.71). Children with B-ALL tended to be first-born more often than controls (OR 1.17, 95% CI 0.99, 1.38).
No differences between cases and controls were evident for gestational age or area-based deprivation.
Around 88% of cases and controls were delivered vaginally and 12% by caesarean (

Study question
Does birth by caesarean delivery increase the risk of childhood leukaemia?

What is already known
Previous reports are conflicting, but some have suggested an increased risk of childhood leukaemia following caesarean delivery initiated before labour onset.

What this study adds
Obstetric records were used to look at type of delivery, time of labour onset, and reason for caesarean delivery.
No strong evidence of an association between leukaemia and caesarean delivery was found, either generally or before labour onset. However, the reasons for the caesarean differed between children with and without leukaemia; preeclampsia occurring more frequently in mothers of affected children.
TA B L E 1 Numbers of cases and controls, odds ratios (OR), and 95% confidence interval (CI), distributed by baseline and delivery characteristics: UKCCS, acute lymphoblastic leukaemia Year of birth

TA B L E 2
Main reasons for caesarean delivery recorded in obstetrical records, ordered according to the overall frequency among controls

| COMMENT
Based on reliable clinical data collected at the time of birth, our findings do not provide strong support for the previously reported association between caesarean delivery and childhood ALL. No evidence of an effect on childhood ALL was observed either overall, before or during labour, or between planned and emergencies caesarean deliveries. Furthermore, the weak positive association between caesarean and B-ALL was not specific to those conducted prelabour. No association specific to hyperdiploid ALL, or to any age group, was noted.
In our study, 12% of cases and controls were delivered by caesarean. This comparatively low proportion is consistent with contemporaneous national data, 16 and with the range observed in previous childhood cancer case-control studies (7%-38% as summarized by Marcotte and colleagues 14 ). Lying within the WHO ideal range of 10%-15%, 17 our findings provide an important benchmark predating the dramatic rise in caesarean deliveries observed in many countries. 17,18 The pooled CLIC analysis reported a 23% increased risk of ALL in children born by prelabour caesarean, these findings being based on a definition that did not account for all indications. 14 Whilst our findings for B-ALL are broadly compatible with such an association, no evidence for a prelabour-specific or age-specific effect was observed. 11,13 More recently, another study reported a 2.67 (95% CI 1.09, 6.57) increased risk of ALL in children born by caesarean delivery, but the findings are difficult to interpret due to data exclusions (eg multiple pregnancies, delivery before 37 weeks, emergency deliveries). 20  It has been hypothesized that a positive association between prelabour caesarean and childhood ALL could be related to a lack of exposure to vaginal microbiota 14 ; and other known caesarean-related hormonal and epigenetic mechanisms could also be implicated. 21 In order to properly investigate the relation between caesarean and childhood ALL and eliminate confounding, it is crucial to also account for the reason for the caesarean delivery. For instance, one could hypothesize that neonates who develop ALL in utero may suffer foetal distress in late pregnancy, therefore prompting an emergency caesarean; the possibility of indication bias warrants more studies with relevant information from medical records. Moreover, one cannot yet rule out confounding by indication on maternal health-related events.
While based on small numbers, our results suggest that preeclampsia could be such a confounder. Birth order, being related to caesarean indication, preeclampsia and childhood ALL, also appears as a likely confounder, as does macrosomia, which used to be an indication for caesarean delivery. 22 Interestingly, in our study, adjusting for birth weight was the only factor to increase the point estimates.Whether caesarean could be causally related to childhood ALL, or whether they could share a common cause, also needs to be addressed.
Caesarean deliveries have become increasingly common, in some countries worryingly so. 17 Clearly studying the potential adverse effects that caesareans could have on subsequent child (and maternal) health is complex, and needs to look in detail at the underpinning reasons for the procedures. In order to gain insight into potential mechanisms, future studies will need to access comprehensive clinical data with information on delivery mode and indication for caesarean.

ACK N OWLED G M ENTS
The UKCCS is supported by Bloodwise grant number 15037.

CO N FLI C T O F I NTE R E S T
None declared.