Pre- and perinatal risk factors for childhood leukaemia and other malignancies: a Scottish case control study

Abstract

The considerable burden of childhood cancer, in terms of both morbidity and mortality, has generated extensive aetiological research over recent decades. Despite these efforts, apart from ionizing radiation, the causes of childhood leukaemia and other malignancies remain largely unexplained (Doll, 1989). The identification of risk factors has frequently focused on the vulnerable period of intrauterine growth and development, birth and neonatal life. The population-based Scottish case control study of childhood leukaemia and cancer has investigated potential risk factors occurring in the prenatal and perinatal period based on data abstracted from the hospital records of both mothers and neonates. A key aspect was to test prior hypotheses as well as identify associations for separate diagnostic subgroups and search for patterns of risk within and across distinct subsets of childhood malignancy. SUBJECTS AND METHODSThe cases studied were children (0-14 years) diagnosed with leukaemia or another malignancy during 1991-1994 while living in Scotland. The childhood population of Scotland at the 1991 census was 959 268. Pathological confirmation of diagnosis was required and cross-checks with the Scottish Cancer Registry and the UK National Register of Childhood Cancers (Stiller, 1995) optimized completeness of case ascertainment. Population-based controls matched on age (to within one calendar month), sex and health board area of residence were randomly selected from all eligible children registered for primary care. This sampling frame is considered appropriate representation of the general population in this age group (Roberts et al, 1995). An optimum two controls per case were selected and a full description of the study methodology is published elsewhere (McKinney et al, 1995). Scotland is a participating centre in the UK Children's Cancer Study (UKCCS).The study was approved by the Local Research Ethics Committees of the 15 Health Boards in Scotland. Consultant clinicians or general practitioners gave consent to approach and interview mothers who gave their signed permission for information to be abstracted from their medical notes. In Scotland, 76 hospitals and maternity units gave access to medical records relating to births from 1976 to 1994. Records for births occurring in England and Wales (3.4%) were obtained by post.Information from obstetric, delivery and neonatal records was recorded by two trained abstractors onto a specifically designed and highly structured standard form which has been validated in a previous study (Roman et al, 1997). Consistency and data quality were maximized through a programme of continual monitoring, coding checks and duplicate abstractions by the senior midwife abstractor (EF).Demographic details of the mother were collected at interview. Each child was assigned a deprivation score (five categories representing quintiles of the Scottish population) according to Carstairs and Morris (1991) from the home address at the time of birth and its validated postcode. This was based on 19...

Association of early life factors and acute lymphoblastic leukaemia in childhood: historical cohort study

“…Our finding of around a two-fold greater risk of childhood ALL among children born into surroundings where there is little overcrowding supports the hypothesis that childhood ALL is caused by an inappropriate immunological response to a common infection or infections in children whose immune systems are not 'programmed' by early exposure to these infections (Greaves and Alexander, 1993;Greaves, 1997). In support of this hypothesis, several studies have reported that early exposure to infection (van Steensel-Moll et al, 1986;Kaatsch et al, 1996;McKinney et al, 1999) is associated with a reduced risk of childhood ALL. However, contrary to what may be expected if this hypothesis is correct, we and several other investigators (Shaw et al, 1984;Shu et al, 1988;Kaye et al, 1991;McKinney et al, 1999;Schuz et al, 1999;Rosenbaum et al, 2000) did not find that later-born infants had a lower risk of ALL than first-born infants.…”

“…Most previous reports failed to find an association between ALL and length of gestation (Kaye et al, 1991;Kaatsch et al, 1996;McKinney et al, 1999), although one study has shown a similar, but not statistically significant, relationship (Dockerty et al, 1999). Some studies have reported positive associations between history of previous foetal loss and childhood leukaemia (van Steensel-Moll et al, 1985;Kaye et al, 1991), others found no association (Schuz et al, 1999) while in the current study, as in a Chinese case -control study (Shu et al, 1988) an inverse association was found.…”

“…As in this study, some investigators have shown that higher socioeconomic position, is associated with higher risk of ALL (McWhirter, 1982;Kaye et al, 1991;Dockerty et al, 1999) but others have failed to observe this association (Shaw et al, 1984;van SteenselMoll et al, 1985van SteenselMoll et al, , 1986Shu et al, 1988;Kaatsch et al, 1996;McKinney et al, 1999;Rosenbaum et al, 2000) and some have found the opposite pattern, both for childhood ALL (Brondum et al, 1999;Dockerty et al, 1999) and infant leukaemia . In some of these studies subjects from higher socio-economic and better educational backgrounds are overrepresented among controls, reducing the ability to detect a positive association between socio-economic position and childhood ALL.…”

“…To date, the findings on birth weight and risk of childhood leukaemia have been inconsistent. Several studies have reported positive relationships with ALL (Robison et al, 1987;Kaye et al, 1991;Westergaard et al, 1997) while others (McKinney et al, 1987(McKinney et al, , 1999Roman et al, 1997;Dockerty et al, 1999), found no association or a relationship with both high and low birth weight (Schuz et al, 1999). It is possible that selection bias, particularly among controls, may have resulted in the failure of some investigators to observe a positive association as controls are likely to be of higher socio-economic position and thus have infants of higher average birth weight, than the general population (Kogan, 1995).…”

“…Recently there has been growing interest in the relationship between prenatal and early life influences on risk of childhood leukaemia (Van Steensel-Moll et al, 1985;Shu et al, 1988;Kaye et al, 1991;Westergaard et al, 1997;McKinney et al, 1999). It is clear that mutations occurring in-utero in lymphoid stem cells contribute to the development of childhood leukaemia (Greaves, 1999;Wiemels et al, 1999) but it is not known which prenatal exposures result in these mutations or how they may be prevented.…”