Epidemiology of childhood leukemia in New Zealand: Studies of infectious hypotheses

https://doi.org/10.1016/j.bcmd.2008.10.008Get rights and content

Abstract

The etiology of childhood leukemia remains an enigma despite decades of research. Hypotheses of an infectious etiology have been around for a long time, and in the last 20 years there have been two main theoretical contenders. One of these involves the possibility of a specific infectious agent having a causative role, and animal leukemia viruses would be analogous to this. Another theory relates to the possible involvement of unusual patterns of infections in infancy and how they might relate to aberrant immune responses. The first of these is easier to test.

In New Zealand, since the early 1990s we have embarked on a program of research on the epidemiology of childhood leukemia. One of the goals has been to test hypotheses about the role of infection in causation. A variety of study designs have been employed, including descriptive, clustering, case-control and contributions to pooled international analyses.

Some of the more interesting findings include: there has been a marked increase in the incidence of acute lymphoblastic leukemia among young children in New Zealand since the mid 1960s, poorer families are at greater risk, and there is no clear support for hypotheses of an infectious cause from the New Zealand data. However because of our small total population (4 million people) we cannot produce clear results on our own. Hence our current international collaborations, for example in CLIC (the Childhood Leukemia International Consortium) represent an important step forward. As countries work together across international boundaries we have a renewed hope that the causes of the childhood leukemias will be unlocked in the foreseeable future.

Introduction

As in other western countries, leukemia is the commonest cancer among people aged from 0–14 in New Zealand. In 2004, there were 43 leukemia diagnoses among 124 new registrations for cancer in children [1]. To researchers in larger countries, these are small numbers; however in the international context, incidence rates for acute lymphoblastic leukemia among children in New Zealand are similar to those observed in countries like Australia, the USA and England and Wales. The National Cancer Registry in New Zealand has been operating since 1948, and there is a strong tradition of quality epidemiological research. A varied contingent of studies on the epidemiology of childhood leukemias has been conducted here, several of which were designed specifically to investigate theories of an infective cause, particularly for acute lymphoblastic leukemia.

Section snippets

New Zealand research: background, time trends and prevailing hypotheses

In the early 1990s, we set up and conducted a national case-control study to study possible causes of leukemia and other cancers in children. One of several primary hypotheses was to assess the possible role of infections and vaccinations on risk of childhood acute lymphoblastic leukemia (ALL) [2]. In a descriptive study of time trends and ethnic differences, we found that there had been a large statistically significant increase in the incidence of leukemia among children aged 0–4 in New

Population mixing study

Kinlen's hypothesis was tested in a New Zealand ecological study of population mixing in rural areas. The population mixing occurred in the context of forestry developments, which led to the establishment of new towns [7]. The data from this study did not support Kinlen's hypothesis, as there was no change in the age-adjusted rate ratios for childhood leukemia during or after the period of population influx that was associated with the forestry developments. A subsequent review of seventeen

Study of spatial clustering

The method of Cuzick and Edwards [11] was used for a New Zealand study of spatial clustering of childhood leukemias and lymphomas [8]. This used a national dataset to look for evidence of clustering by birth address. There was no statistically significant clustering for any tumour group overall, including childhood ALL, although there was evidence of statistically significant clustering for a subgroup of ALL based on age (ages 10–14). This may have been real or a chance occurrence, and was for

Case-control study of possible causes

The New Zealand case-control study of childhood leukemias and other cancers was conducted to assess multiple possible causal factors [2]. These included birth characteristics, infections and vaccinations, chemicals and drugs, smoking, alcohol, pesticides, electromagnetic fields, X-rays, vitamins and minerals and parental occupations. The two major hypotheses relevant to infection and immunity were: (1) that children infected with one or more specific viruses (in utero or in early childhood) are

Results of the case-control study

Most of the case-control study findings for demographic variables, infections and vaccination exposures have been reported previously [2] and are summarized below. Previously unpublished detail on the serological results is given here in Table 1. For demographic data, children whose mothers were married had a lower risk of ALL, with an odds ratio (OR) of 0.5, and a 95% confidence interval (CI) of 0.3 to 0.8. There was a suggestion of an increase in risk of ALL for those whose parents'

Discussion

The various current hypotheses of an infective cause for childhood leukemia have yet to be proven. No specific virus has been shown conclusively to have a role [14], and the hypotheses of Greaves and Kinlen have not found consistent support in the studies that have sought to test them. In New Zealand the conditions are right for conducting epidemiological studies of high quality, because of very good datasets, population-based controls, and good participation rates. For rare diseases like the

Acknowledgment

This paper is based on a presentation at a Focused Workshop on “Immunity and Infection in the Etiology of Leukemia: a Workshop” sponsored by The Leukemia & Lymphoma Society and held in Sausalito, CA, Nov. 14–16, 2007.

References (14)

There are more references available in the full text version of this article.

Cited by (0)

View full text