Feasibility of Identifying a Chinese Study Population and Eligible Pediatric Cancer Cases, and Linking the Two Groups to Assess the Role of Periconceptional Folic Acid Supplements in Risk of Pediatric Cancer.

Few risk factors have been consistently linked with pediatric cancer. During the past 15 years, small casecontrol epidemiologic studies in Caucasian populations have reported reduced risks of pediatric leukemia and brain tumors in offspring of mothers treated with folic acid and iron supplements in the prenatal and periconceptional period.

We describe a feasibility study to evaluate the potential for prospective assessment of periconceptional and prenatal folic acid supplements and other postulated risk factors in risk of pediatric leukemia and other pediatric cancers. If the feasibility study is successful, we propose to carry out a prospective record linkage Children and Families Cohort Study (CFCS) (N=243,779) to compare risk of pediatric leukemia, brain tumors and other pediatric cancers in offspring of Chinese mothers who took vs did not take periconceptional folic acid supplements during 1993-1995 in a community-based intervention trial that assessed the effects on occurrence of neural tube defects. The original investigators last followed up the population in 2000-2001 The Chinese CDC, the U.S. CDC, and NCI collaborators carried out small pilot studies (N=560 families) during 2011-2013 demonstrating that CFCS families can be successfully identified and traced, that mothers and offspring would participate in single and in multi-day interviews, and that referral and community hospitals (N=8) could locate and abstract medical records of potentially eligible pediatric cancer cases diagnosed during 1994 to the present.

The proposed feasibility study involves three components. The first component will test success in re-creating the original CFCS offspring list in the 21 counties in three provinces. Data collected in the original 1993-1995 study and 2000-2001 follow-up to be used in recreating the list includes the original study ID number, sex, birth dates of offspring, mothers and fathers, place of birth (village, township, and hospital), residence at birth, and other identifying information; names of subjects and family members are not available on the original study lists. This component will be undertaken in collaboration with county- and provincial-level maternal and child health organizations that will coordinate with public health bureaus, family planning, and public security agencies. The goal of this component is to re-identify all of the CFCS offspring in 5 counties in Hebei Province and in 9 counties in Jiangsu and Zhejiang Provinces, and in samples of 500 randomly selected CFCS offspring in each of 7 counties in Jiangsu and Zhejiang Provinces.

The second component will identify all hospitals where children from the 21-county study population would have been diagnosed or treated for pediatric cancer during 1994 to the present. An expanded sample of these hospitals (N=10 beyond the 8 hospitals included in our recent pilot study) will be contacted and asked to collaborate in the proposed study. Tasks will include identification and abstraction of medical records of pediatric cancer cases occurring in children who are potentially CFCS offspring (born during 1994-1996 in any of the 21 CFCS study counties). Mothers of the pediatric cancer patients will be interviewed and asked to provide identifying information to ascertain if the offspring were in the CFCS population, and to provide consent (together with their offspring s assent) for the collaborators to have access to the medical record abstract data and to match the offspring with the CFCS population list. The investigators will then attempt to match the potentially eligible pediatric cancer cases with the CFCS list in the 5 counties in Hebei Province and the 9 counties in Jiangsu and Zhejiang Provinces.

The third component will identifying deaths among offspring to ascertain those that definitely, probably or possibly might be due to childhood cancer. For deaths among CFCS offspring identified during the 2000-2001cohort follow-up by the original study team, cause of death will be ascertained from death certificates if available or from medical records or interview of primary care physicians if death certificates are not available. For deaths among offspring who died after 2000-2001,review of death certificates among children born during 1994-1996 in the 21 counties will be undertaken. Identifying information will be sought from the death certificate to match the deceased child to the CFCS. If the identifying information is insufficient for a match, for those deaths definitely, probably or possibly due to childhood cancer, the mother of the deceased child will be traced and interviewed to obtain additional identifying information, permission to review medical records for hospitalizations near to the time of death, and contact information for the child s primary care doctor. For those children who died after 2000-2001and are determined to definitely or probably be members of the CFCS, a provisional determination will be made if the deceased offspring definitely, probably, possibly or did not have a diagnosis of childhood cancer. The provisional diagnoses will be reviewed by the expert pediatric hematologists/oncologists/neurosurgeons for a final determination and assessment of level of certainty.