California eatin’

We have previously discussed results from the Northern California Childhood Leukemia Study (NCCLS) [see C3: Vol. 13, No. 5]. In this ongoing study, cases of acute lymphoblastic and acute myeloid leukemia (ALL and AML, respectively) diagnosed between 0 and 14 years of age are recruited at northern California hospitals. Controls are recruited through the state birth registry and matched to cases on age, sex, Hispanic ethnicity, and maternal race. Two new analyses from the study deal with maternal diet in the year prior to pregnancy (to capture consumption patterns in early pregnancy) [Jensen CD et al. (2004) Cancer Causes and Control 15: 559-570] and with the diet of the child in the first two years of life [Kwan ML et al. (2004) Am J Epi 180(11): 1098-1107]. In both instances dietary data were collected through a food frequency questionnaire completed by the mother and analyzed using conditional logistic regression.

The analysis of maternal diet was focused on ALL and produced copious results. Most interesting were those regarding food groups and specific nutrients, rather than individual foods. Comparing the diets of 138 matched case-control pairs, the investigators found significant inverse associations of vegetable consumption (Odds Ratio (OR)= 0.53; 95% confidence interval (CI): 0.33-0.85; p for trend = 0.008) and of consumption of protein sources (OR = 0.40; 95% CI = 0.18-0.90; p for trend = 0.03). There was also a borderline significant inverse association of fruit consumption (OR = 0.71; 95% CI: 0.49-1.04; p for trend = 0.08). The data did not support the a priori hypothesis that cured meat consumption would be associated with ALL (OR = 0.71; 95% CI: 0.44-1.15; p for trend = 0.16). Among the specific nutrients, provitamin A carotenoids and the antioxidant glutathione were inversely associated with ALL; ORs were 0.65 (95% CI: 0.42-1.01; p for trend = 0.05) and 0.42 (95% CI: 0.16-1.10; p for trend = 0.08), respectively. The study did not confirm previously reported findings that maternal folic acid consumption is inversely associated with ALL [see C3: Vol. 12, No. 6].

For child’s diet, there were 328 matched case-control sets available for analysis. The questionnaire inquired about consumption of the limited number of foods and food groups, as follows: hot dogs/lunch meats, beef/hamburger, vegetables, oranges/bananas, apples/grapes, orange juice, fruit juice, milk, soda, and vitamins. Significant inverse trends were apparent for orange/banana and orange juice consumption for all leukemias diagnosed at 2-14 years of age. The ORs comparing regular to rare or no consumption of these foods were 0.49 (95% CI: 0.26-0.94; p for trend = 0.02) and 0.54 (95% CI: 0.31-0.94; p for trend 0.04), respectively. No other significant associations were noted.

COMMENT: The role of diet in the etiology of childhood leukemia is a burgeoning line of inquiry. This analysis of maternal diet is the most comprehensive yet, while that of the child’s diet is entirely novel. Diet is complex, so that one may always quibble with a particular analytic strategy, but the design of the NCCLS is methodologically sound. These results suggest that fruit and vegetable consumption, of both mother and child, may reduce the risk of childhood leukemia. Limitations of these analyses include small sample sizes (138 cases in the first analysis, and 328 in the second) and the large number of comparisons performed. Further, the lack of confirmation of an inverse association of maternal folic acid consumption with leukemia is perhaps not surprising given that several genes mediate folate bioavailability. As the NCCLS has collected DNA from participating mothers and children, it would be worthwhile for the investigators to eventually present the results of an analysis of folic acid consumption by genotype.

Logan G. Spector

Fanconi Anemia- X Marks the Spot!

Fanconi anemia (FA) is a genetically heterogeneous disease. At least 11 genes are believed to be involved in the disorder, 8 of which have been cloned previously. Meetei AR et al [Nature Genetics 2004, 36:1219-1224] now report that the gene for FA complementation group B (FANCB) is located on the X chromosome. This is a surprising finding, as X-linked inheritance has not been described in FA. The paper reports robust proof that the candidate locus described as FANCB is truly the correct gene. Mutations in the gene were found in 4 individuals identified as complementation group B. Transfection of lymphoblasts from individuals with FA-B with the gene corrected the abnormal phenotype. Mutations were identified in the healthy mothers and one sister of two cases. As expected with X-linked inheritance, all affected cases are male. X-inactivation in mothers carrying the mutation is skewed in blood and in fibroblasts, favoring expression of the normal allele, as has been described in other X-linked disorders, e.g. Wiskott-Aldrich syndrome. The female relatives were all clinically well with no stigmata of FA and a normal response to mitomycin C, with no evidence of mosaicism. This finding indicates a proliferative disadvantage for cells carrying the mutant allele.

COMMENT: The finding of X-linked inheritance in FA has important implications for genetic counseling. Although FA-B is rare, cases can now be accurately diagnosed. Only male siblings of a proband will be affected but sisters and maternal aunts may be carriers and 50% of their male offspring will be affected.

Stella M. Davies

Diploid is Best!

Mosaic variegated aneuploidy (MVA) is a rare recessive condition characterized by growth retardation, microcephaly, childhood cancer and constitutional mosaicism for chromosomal gains and losses. The proportion of aneuploid cells is usually greater than 25%, much higher than occurs in normal people. The abnormalities are typically monosomies and trisomies, and the risk of cancer is high, with reports of rhabdomyosarcoma, Wilms’ tumor and leukemia. Hanks S et al [Nature Genetics 2004; 36:1159-1161] screened genes involved in the human spindle checkpoint for mutations in affected cases and found biallelic mutations in the gene BUB1B in five families. In 2 of those families, children developed embryonal rhabdomyosarcoma, one at 5 months of age and one at 7 years.

COMMENT: MVA is a rare condition that might be overlooked. Children with MVA typically present with severe intrauterine growth retardation, microcephaly, eye abnormalities, mild dysmorphism and variable developmental delay. Karyoptypic studies show multiple different aneuploidies in different cells. Not all of the cases investigated in this study had abnormalities in BUB1B, suggesting that a similar phenotype might be produced by mutation in other genes in the spindle checkpoint pathway. Increased awareness of the disorder amongst clinicians will likely identify additional cases.

Stella M. Davies

Birth weight and childhood acute lymphoblastic leukemia (ALL)- need we say more?

High birth weight has been associated with an increased risk of childhood ALL, although there have been a few studies that have shown no association. Hjalgrim LL et al [JNCI 2004; 96: 1549-1556] studied 1905 children with ALL and 299 children with AML diagnosed between Jan 1, 1984 and Dec 31, 1999 in the countries of Norway, Denmark, Sweden, and Iceland. Cases were identified through the population-based Nordic Society of Paediatric Haematology and Oncology acute leukemia database. Five control subjects, identified through birth registries in the respective countries, were individually matched to cases on sex, birth month, and birth year. Information on birth weight and gestational age was abstracted for all cases and controls. The authors confirmed the association between high birth weight and ALL; infants that weighed more than 4000g at birth were at an 18% increased risk of developing ALL (OR=1.18, 95%CI =1.03-1.34). Moreover, when birth weight was examined as a continuous variable and adjusted for gestational age, birth order and parental age, individual subtypes of ALL showed a different pattern: the risk of B-precursor ALL increased by 30% (95% CI=1.16.-1.47) per kg increase in birth weight, while the risk of T-cell ALL was non-significantly increased (OR=1.14, 95%CI=0.80-1.62). For AML, there was a suggestion of a U-shaped risk; infants with birth weights less than 2500g (OR=1.56, 95%CI=0.71-3.44) and greater than 4000g (OR=1.17, 95% CI=0.83-1.65) were at a slightly higher risk than infants with birth weights 2500-3999g, although neither OR reached statistical significance. Increasing birth order was associated with a decreased risk of ALL (OR=0.91 per one position increase in birth order 95% CI=0.85-0.97), which was most notable for B-precursor ALL. Although the authors found no obvious patterns with age at diagnosis, it was notable that the magnitude of risk associated with birth weight > 4500g was greatest in children diagnosed in the first year of life (OR=2.20, 95% CI=0.65-7.50). Finally, in a comparison with sibling birth weights, the authors found that ALL case patients did not weigh more at birth than their siblings, suggesting a strong familial component of high birth weight in children with leukemia.

COMMENT: This study joins several others that have evaluated the association between high birth weight and childhood leukemia. A meta-analysis of previous studies examined over 10,000 patients and concluded that birth weight of more than 4000g increases the risk of childhood ALL by 26% (95% CI=1.17-1.37) [Hjalgrim LL et al Am J Epi 2003; 158:724-735]. While it is recognized that there will always be some studies that do not find such an association, we feel that the data are sufficiently consistent to conclude that high birth weight is a risk factor for childhood ALL, particularly for cases diagnosed at younger ages. The next step, of course, is to determine why. Hjalgrim et al note hypotheses regarding this relationship including a) children with higher birth weight have higher circulating levels of growth factors, including insulin-like growth factor-1, which may increase proliferative stress on pre-leukemic cells [Ross JA, et al Cancer Causes & Control 1996; 7:553-559]; or b) children with higher birth weight may have more cells at risk of malignant transformation [Albanes D & Winick M. JNCI 1988; 80:772-774]. The third hypothesis states that pre-leukemic cells may secrete growth factors that increase birth weight [Ross JA et al, 1996; Vorwerk P, Mol Pathol 2002; 55:40-44]. Hjalgrim et al suggest that the evidence of a familial association with birth weight argues against the latter hypothesis. Nonetheless, there is an increasing body of evidence that links high birth weight to a number of malignancies, including breast cancer, and it will be important to go beyond these descriptive studies into investigating the reasons why.

Julie A. Ross
 

C3 Quarterly Newsletter
Children's Cancer Research Fund
Epidemiology Research Unit
Division of Pediatric Epidemiology
Clinical Research
University of Minnesota
420 Delaware St. SE, Box 422
Minneapolis, MN 55455
pedsepi@umn.edu

Editors: 
Stella M. Davies, MD, PhD, and Julie A. Ross, PhD