(VCCEP) Tier 1 Pilot Submission for BENZENE - Tera

4.0, 95% CI = 1.8-9.3) but not ALL (Shu et al., 1988). This is the only study to separate the two
main forms of acute childhood leukemia in children and reported an increase in the leukemia
type associated with benzene exposure in adults. There was no quantification of benzene air
concentrations and exposures to confounding chemicals could have occurred. There was also
no attempt to characterize children’s direct chemical exposures that could have occurred in the
home or from the general environment. Further, recall bias could have resulted in a
misclassification of exposures or cases in this study. McKinney et al. (1991) also conducted a
case-control study to evaluate maternal and paternal exposures to benzene and the relationship
with childhood leukemia (subtype not identified). Few risk factors were identified for the mother,
but paternal exposures to benzene during the pre-conception period were significantly
associated with childhood leukemia (OR = 5.81, 95%CI = 1.67-26.44) (McKinney et al., 1991).
This study was potentially limited by a small numbers of cases, inadequately quantified benzene
exposures, and potential exposures with confounding chemicals. As with most case-control
studies, recall bias could have occurred.
Various other studies of childhood leukemia have failed to show an association with potential
benzene exposures of parental occupations. Feingold et al, (1992) conducted a case-control
study of childhood ALL and parental occupation. These investigators reported no association
between maternal exposure to benzene and ALL (Feingold et al., 1992). Kaatsch et al. (1998)
conducted a case control study of childhood leukemia. Though the majority of the leukemias
were ALL, there were 147 cases of childhood ANLL included in the analysis (but not analyzed
separately). No association between parental exposure to benzene and childhood leukemia
was found (Kaatsch et al., 1998). Shu et al. (1999) examined childhood ALL with paternal or
maternal exposure to benzene or ‘petroleum products’ (including during pregnancy) in a casecontrol
study. There was no relationship reported (Shu et al., 1999). Feychtling et al. (2001)
conducted a case control study on paternal occupation and childhood cancers including
leukemia (all combined). Paternal exposure to benzene (RR = 1.23, 95% CI = 0.39-3.85) was
not positively associated with an increased leukemia risk (Feychting et al., 2001). Infante-
Rivard, et al. (2005) conducted a recent case-control study of childhood ALL and maternal
exposure to various solvents. The OR for benzene was 0.82 (95% CI = 0.22-3.06) for exposure
during the 2 years before birth, including pregnancy and 1.39 (0.31-6.25) for exposure during
pregnancy (Infante-Rivard et al., 2005). The significance of this study, like the others listed in
this section, was limited by small numbers, inadequate exposure assessment and exposures
with confounding chemicals. Potential recall bias could have resulted in a misclassification of
cases or exposures and significant changes in the study findings.
In general, there is consistent support in the scientific and medical literature that parental
exposure to benzene is not a causative factor in the development of childhood ALL. However,
the type of leukemia most clearly associated with benzene exposure in adults is AML. It should
be pointed out that the etiological risk factors for AML and ALL in both adults and children are
different. Only one study was available that specifically addressed childhood AML. The other
studies either combined all forms of childhood leukemia into a single group or evaluated ALL
singly. Both approaches would limit the ability of these investigations to detect a positive
association between benzene and childhood AML, while increasing the likelihood of finding an
association despite differing etiologies for these types of childhood leukemia. In Shu et al.,
(1988), such an association was reported, but has not been independently confirmed.
Additionally, 23% of the leukemia cases occurred in children 10-14 years old. Clinical literature
indicates that the latency period for t-AML is approximately 5-7 years. The latency period for
benzene induced AML is less well characterized and could be longer, but the possibility exists
that cases of leukemia occurring in older children were unrelated to in utero exposures.
Therefore, data presented in Shu et al, (1988) should be interpreted cautiously.
Benzene VCCEP Submission
March 2006
56