Summary

Leukemia contributes about 1/3 of all cancers in children below the age of fifteen. This heterogeneous, multi-factorial disease of the hematopoietic system accounts for the largest proportion of all cancers in this age group, with acute lymphatic leukemia (ALL) being the most common subtype. Incidence rates show a slight increase with time.

There is increasing evidence that a first damaging event to the hematopoietic stem cells (1st hit) occurs during the prenatal phase and one or more postnatal hits are needed to transform the pre-leukemic clones into leukemia cells. In parallel to the body of research investigating the different indicators of molecular damage, therapeutic procedures have been continually developed and improved. Today individually optimized therapeutic plans for childhood ALL patients are available which result in a survival rate of more than 80%.

The aim of the international workshop, jointly organized by ICNIRP, WHO and BfS was to bring together experts from different disciplines to summarize the current knowledge about the role of genetic and environmental risk factors for childhood leukemia. It became clear that, in spite of the many available epidemiological studies, knowledge about the causes of leukemia is still rather incomplete.

Genetic risk factors
Studies on monozygotic twins and the fact that most ALLs occur in the age between 2 and 5 years, suggested an increased inherited susceptibility to damage by environmental exposures in ALL patients. For example, the chromosomal rearrangement TEL-AML1 (t12;21) is commonly associated with ALL, but is also found in approximately 1% of children who do not develop leukemia. Studies currently in progress are focusing on candidate pathways involved in protection against various external threats. These include folic acid metabolism, immune function, xenobiotic metabolism, DNA repair and oxidative stress. The research into genetic predisposition is challenged by the fact that the expected effect sizes for common, low-penetrance markers are small requiring large sample sizes.

Environmental risk factors
1. Ionizing radiation
Two forms of ionizing radiation exposure were discussed: singular relatively high exposures and chronic low dose exposures. These included:
- X-ray examinations during pregnancy, i.e. prenatal exposures of the mother through diagnostic x-ray procedures. The Oxford survey of childhood cancers revealed a clear exposure-risk relationship, whereas data on postnatal X-ray exposures were inconclusive.
- Knowledge from studies of the Japanese atomic bomb survivors.
- Proximity to nuclear power plants: the current German study (KiKK-Study) compared with British and other European studies. The results of the KiKK-Study lead to reanalyzes of the national data in the UK, but these could not support the German results. There is no obvious explanation for the differences. Only with regard to site location was it obvious that almost all UK nuclear power plants are in coastal locations, which is not the case in Germany.
- Indoor radon exposure. Studies so far show only inconsistent results. Many studies were poor, because the radon exposure was not measured individually. In a current study from Denmark, a model was developed and validated to calculate the radon levels in residences of children. In this study the cumulative radon exposure (intensity x time) was associated with increased risk for ALL.

2. Non-ionizing radiation
Epidemiological studies on non-ionizing radiation in the low frequency range (50/60 Hz) have consistently shown an increased leukemia risk from magnetic field strengths above 0.3 -0.4 µT. There is no biological explanation or support from animal studies for such findings. Recent powerful epidemiological studies involving high frequency electromagnetic fields caused e.g. by high power radio and TV transmitter, do not indicate a causal relationship.

3. Chemicals
Pesticides and other chemicals have been investigated as risk factors for several decades, with relatively small observed risk factors (OR <2). However, exposure assessment has been rare, and recent validation studies have shown that previous exposure assessments based on self reports were in part heavily over-estimated. Thus, misclassification (false positive) has to be taken into account and, so far, only solvents have been confirmed as a risk factor.

4. Life style
The data showing a relation between elevated risk and elevated birth weight are relatively consistent. The basic hypothesis is that an excess of growth factors may result in an amplification of proliferative processes in the hematopoietic system. In this connection, the role of folic acid, diet and the age of the mother has been investigated without definite results to date. The influence of the social status was included in some studies, but the results were inconsistent.

5. Immune system
Current complex evidence supports the hypothesis that the second (or last) postnatal damaging event, that leads to manifestation of leukemia, may be related to a deregulation of the immune system. The clear incidence peak of the disease in children aged 2-5 years in industrialized countries can be related to the common ALL (aberrant B-lymphocyte precursor cells). It is assumed that insufficient or delayed activation of the immune system (isolation, too little social contact, too much hygiene) might be a risk factor. A current meta-analysis concerning day care studies supports this hypothesis. The results show a consistent and in part considerable reduction in risk for children participating in day care programs. Participation in day care is used as an indicator for first social contacts and the associated activation of the immune system.

Conclusions and suggestions for future work
Leukemia in children is a multi-factorial disease and none of the environmental risk factors mentioned above prove to have major explanatory power. The observed risk factors were small, in general less than 2. In all areas similar problems and uncertainties were discussed: Essential, but difficult to provide retrospectively, is an improved exposure assessment to prevent misclassification. Confounders seem to have little influence on the results. It became evident that new research paths are needed as there is no simple explanation for the etiology of childhood leukemia. Genetic studies on associations between a specific gene and childhood leukemia need sufficient power to detect small effect sizes. Individual studies might not be able to overcome the foreseen challenges. The need for larger sample sizes and for pooling genetic, environmental and behavioral data has led to the initiation of the Childhood Leukemia International Consortium. It was stated at the workshop that concerted efforts must be made to increase our knowledge in the genetic basis of hematopoietic cancers, the role of the immune system and the interactions between genes and environmental factors.

A summary of all presentations is available on the ICNIRP website (www.icnirp.org). The workshop proceedings are published in Radiation Protection Dosimetry 132 (2); December 2008. View the online publication.