Infant Leukemia: Let Baygon be Bygone

Leukemia that occurs in infancy is distinct from leukemia that occurs in older children: infants (<1 year of age at diagnosis) have a poor prognosis, are equally likely to have ALL or AML, and are more likely to be female; non-infants are more likely to have ALL, be male, and have a good prognosis. Epidemiologic studies suggest that high infant birthweight, prior maternal fetal loss, and maternal alcohol consumption during pregnancy are associated with an increased risk of younger age leukemias, although little data exist for infants only. Nearly 80% of infant cases have specific chromosomal rearrangements involving the MLL gene in their leukemia cells, and molecular evidence indicates that these abnormalities occur in utero. These MLL abnormalities are rarely found in older children with leukemia. Interestingly, therapy-related AMLs that are associated with the DNA topoisomerase II inhibitor drugs demonstrate MLL abnormalities similar to infants with de novo leukemia. This led to the speculation that infant leukemias could arise from maternal exposure to DNA topo 2 inhibitors [Ross et al, JNCI; 86:1678-80, 1994]. Natural and synthetic topo 2 inhibitors are found in certain foods, medications, and chemicals. A preliminary study of 84 infant cases found that increasing maternal consumption of dietary topo 2 inhibitors was associated with nearly a 10-fold increased risk of AML; no association was found with ALL [Ross et al, Cancer Causes Control; 7:581-90, 1996]. 

There are several current studies investigating the etiology of infant leukemia; the results of one study are described here. Alexander et al [Cancer Research 61:2542-46, 2001] report the results of a pilot case-control study of 136 infant cases and 266 controls recruited from Italy, Greece, Egypt, Brazil, Chile, China, Hong Kong, and Japan. Structured questionnaires were used to obtain maternal exposure data including lifestyle factors (alcohol and smoking use), medical drug use (prescription and other) and occupational and non-occupational exposures to pesticides, solvents, organic dusts, and ionizing radiation.  Status regarding MLL gene rearrangements for 85 cases was determined and utilized in the analysis. Known topo2 inhibitors evaluated included alcohol consumption (flavonoids) and cigarette smoke (benzene). DNA-damaging agents explored included dipryone (an NSAID) and metronidazole (used for treatment of vaginal infections), as well as pesticides and mosquitocidals. The authors found significant associations with maternal report of ingestion of several groups of drugs including herbal medicines and DNA-damaging drugs. The strongest associations were found with maternal report of dipyrone ingestion and use of Baygon, a pesticide. Interestingly, these positive associations were only found for the MLL positive leukemias (OR=5.84, 95%CI=2.09-16.30; OR=9.68, 95%CI=2.11-44.4), respectively. No associations were found with the MLL negative group. Further, most of these associations were more apparent with AML and not ALL.  Although the authors acknowledge these results are preliminary, they argue for further confirmation in other studies.

COMMENT: This is an important preliminary report. It is the first study of infant leukemia to examine maternal exposures with respect to MLL status. Although the numbers are small, these data provide support regarding a distinct etiology for MLL+/AML cases versus MLL+/ALL cases. Dipyrone, while not licensed for use in the United Kingdom and the United States, is used as an NSAID in several other countries. Maternal ingestion has also been associated with Wilms’ tumor in one study [Sharpe et al, Epidemiology 7:533-5, 1996]. Baygon is the brand name for a carbamate pesticide, propoxur. It is primarily used as a bait insecticide for the control of ants and cockroaches. While the possibly inhibitory activity of these agents on DNA topo 2 is unknown, there are some potential links that the authors discuss. Given these intriguing preliminary data, it will be important to take these observations into account in on-going studies. 

Infant Leukemia Continued: Folic Acid Metabolism

Folate metabolism plays a key role in DNA synthesis and abnormalities of folate are associated with abnormal hematopoiesis, cancer and vascular disease.  MTHFR (5,10-methylenetetrahydrofolate reductase) is an enzyme which shunts methyl groups from DNA synthesis to methylation pathways with the concomitant conversion of homocysteine to methionine. This gene is polymorphic, with a low-function variant termed C677T being associated with accumulation of homocysteine and increased risk of vascular diseases. The low-function variant is frequent in the population; 80% of Caucasians have at least one copy, and 40% have 2 low-function alleles.  In a study performed on children with ALL enrolled on United Kingdom Children’s Cancer Study Group protocols Wiemels et al Proc. Natl. Acad. Sci.  98:4004-9, 2001] have analyzed MTHFR allele frequencies. The data show a significantly reduced frequency of the low-function variant in infants with MLL-rearranged leukemia compared with cord blood controls, and compared with TEL-AML1 or hyperdiploid ALL. The authors examined an additional polymorphism in MTHFR (A1298C) and found a 4-fold reduced frequency of CC homozygotes in hyperdiploid leukemias only. 

COMMENT: These data indicate the importance of host factors in the etiology of ALL, and reinforce the fact that subsets of ALL differ etiologically.  This study provides important data but focuses on a small number of cases (37 MLL-rearranged leukemias; 78 TEL-AML1 leukemias, 138 hyperdiploid leukemias), even though this study incorporated the activities of a national cooperative group over 6 years.  The data need to be repeated in additional large-scale studies, but raise interesting questions regarding infant and maternal exposures and the etiology of MLL-rearranged leukemias. 

Gestational age and neuroblastoma: what’s weeks got to do with it?

Neuroblastoma is the most common malignancy in children under the age of 1 year, with an incidence rate of approximately 58 per million infants [Gurney J Pediatric Hematol Oncol 5:428-32, 1997]. Few risk factors that have been identified for neuroblastoma. Factors that have been associated with neuroblastoma in past studies include high birth weight (> 4000 grams), maternal alcohol use during pregnancy, use of specific medications (diuretics, antiemetics, sex hormones) during pregnancy, and certain parental occupational exposures including hydrocarbons. A few studies have found that preterm birth may be protective. In this report by Buck et al [Paediatric and Perinatal Epidemiology 15:47-53, 2001], 155 children under the age of 5 with histologically confirmed neuroblastoma were identified between 1976 and 1987 from the New York State Cancer Registry. Controls, frequency matched on year of birth, were randomly selected from livebirth registries; a total of 310 controls were included in the analysis. Parents of cases and controls were interviewed by telephone using a questionnaire that focused primarily on medication use during pregnancy and occupational history; data were supplemented with information from birth certificates. The authors report that pre- and post-term gestations (< 37 weeks and > 42 weeks) were associated with a statistically significant reduced risk of neuroblastoma, OR=0.4 (95% CI=0.1, 0.9) and OR=0.3 (0.1, 0.7), respectively. 1-min Apgar scores, reduced pelvic diameter, and birth injury (as recorded on the birth certificates) were associated with an increased risk of neuroblastoma, although the confidence intervals included 1. The authors suggest that birth trauma or injury may also be etiologically important. 

COMMENT: This study is based on a relatively small number of cases identified over an 11-year period; there are undoubtedly some difficulties in parental recall of specific exposures such as medication use and occupational exposures. Further, the authors do not address how events surrounding the birth of the child would be etiologically relevant, especially for cases diagnosed less than 1 year of age. It is possible that these events just increase surveillance and lead to diagnosis of the child. However, the strength of this study is the comparison of birth certificate information (gestational age, birth injury, etc) between cases and controls.  Using these types of data tends to increase the validity of the findings as they provide an unbiased source of information. Analysis of a large (approximately 600 cases) case-control study conducted in the Children’s Oncology Group is currently underway, and the investigators should consider exploring the factors reported here. 

War and peace- and leukemia.

In previous publications, Kinlen and colleagues have proposed that unusual mixing of rural and urban populations contributes to the etiology of leukemia [Kinlen LJ, Br J Cancer 1995;71:1-5; Kinlen et al, BMJ, 1993; 306: 743-48; Kinlen et al, BMJ 1995, 310; 763-68].  Kinlen and Balkwell [Lancet 357:858, 2001] now provide an additional example of this phenomenon in a study of Orkney and the Shetland Islands during World War II. Orkney and the Shetland Islands are small islands in the North of Scotland.  During World War II local people were outnumbered by the servicemen stationed there in case of an invasion from Norway after the German occupation in 1940. The Kinlen hypothesis proposes that exposure to infectious agents plays a role in the etiology of leukemia, and that population mixing allows contact between susceptible (rural) people and infected (urban) people, causing an abrupt increase in leukemia incidence. In agreement with this hypothesis, childhood leukemia increased 3.6-fold (p=0.001) in the wartime, but not in the postwar period. 

COMMENT: This is another fascinating study lending support for a role for infection in the etiology of childhood leukemia.  It should be noted that this report depends on only 9 cases of leukemia, and that these cases included biologically diverse malignancies (5 ALL, 2 AML and 2 unclassified).  It would be of interest to see these observations replicated by others, and of course to see some biological evidence for the nature of the infective agent. 

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