Childhood cancer in the Northern Region, 1968-82: incidence in small geographical areas

A comparison of some simple methods to identify geographical areas with excess incidence of a rare disease such as childhood leukaemia

Six statistics are compared in a simulation study for their ability to identify geographical areas with a known excess incidence of a rare disease. The statistics are the standardized incidence ratio, the empirical Bayes method of Clayton and Kaldor, Poisson probability, a statistic based on the 'Breslow T' test (BT) and two statistics based on the 'Potthoff-Whittinghill' test (PW) for extra-Poisson variance. Two alternative processes of clustering are simulated in which high-risk locations could be caused by environmental sources or could be sites of microepidemics of an infectious agent contributing to a rare disease such as childhood leukaemia. The simulation processes use two parameters (proportion of cases found in clusters and mean cluster size) which are varied to embrace a variety of situations. Real and artificial data sets of small area populations are considered. The most extreme of the artificial sets has all areas of equal population size. The other data sets use the small census areas (municipalities) in Finland since these have extremely heterogeneous population size distribution. Subset selection allows examination of this variability. Receiver operator curve methodology is used to compare the efficacy of the statistics in identifying the cluster areas; statistics are compared for the proportion of true high-risk areas identified in the top 1 per cent and 10 per cent of ranked areas. One of the PW statistics performed consistently well under all circumstances, although the results for the BT statistic were marginally better when only the top 1 per cent of ranked areas was considered. The standardized incidence ratio performed consistently worst.

Space-time clustering of acute lymphoblastic leukaemia in parts of the U.K. (1984-1993)

Age-related differences in the incidence and immunological subtypes of acute lymphoblastic leukaemia (ALL) suggest that it may be composed of more than one disease entity, each with different aetiologies. Childhood leukaemia (of which the majority of cases are ALL) has been suspected of having an infectious aetiology, but few studies have systematically examined ALL for clustering by age group. The aim of this study was to examine ALL for evidence of space-time clustering of date and place of diagnosis by age group. Knox space-time analysis was carried out separately for three different age groups: childhood (0-14 years), young adult (15-34 years) and older adults (35-79 years). Data on 968 cases of ALL aged 0-79 years, arising during 1984-1993 in the areas covered by a specialist population based register of leukaemias and lymphomas in parts of the U.K., were used in the analysis. Space-time clustering of diagnoses was limited to children aged 0-14 years. It was more prominent in those diagnosed in the period 1984-1988, than in those diagnosed in 1989-1993. The clustering may indicate an infectious aetiology for childhood ALL, or could be the result of episodic exposures to some environmental hazard.

Adult leukemia and proximity-based surrogates for exposure to Pilgrim plant's nuclear emissions

Possible associations between adult leukemia incidence and proximity-based surrogate measures of potential for exposure to radioactive emissions from the Pilgrim nuclear power plant in Plymouth, Massachusetts, were investigated. Included in this study were 105 nonchronic lymphocytic leukemia cases, diagnosed between 1978 and 1986 at age 13 y or older, that occurred in 22 towns near Pilgrim; population controls numbered 208. Residence within 4 mi (6.4 km) of Pilgrim during "high-emissions" years was related to case-control status (adjusted odds ratio [OR] = 3.88, 95% confidence interval [95% CI] = 0.81-10.64). A high "exposure" score (i.e., a value that accounted for downwind time) was also related to case-control status (OR = 3.46, 95% CI = 1.50-7.96). Some statistically significant dose-response trends were found. Cautious interpretation of associations is warranted in light of the low levels of reported emissions.

Viruses, clusters and clustering of childhood leukaemia: a new perspective?

Clusters of childhood leukaemia have, during a lengthy and controversial history, focussed attention on two alternative putative aetiological agents: infections and localised environmental pollution. In the United Kingdom emphasis is currently placed on the latter because of reports of localised clusters in the vicinity of two nuclear reprocessing plants. Now the most recent studies of spatial clustering in the United Kingdom also support the hypothesis that a substantial population of cases of childhood acute lymphoblastic leukaemia (ALL) arise as a rare host response to certain patterns of exposure to common infectious agents--the aberrant response model. Relevant aspects of the epidemiology of ALL are reviewed from this perspective and the hypothesis shown to be capable of unifying reported associations with different types of risk factor. It is probable that specific agent(s) are involved though none have been identified and these may share many epidemiological characteristics of herpes viruses. The possible relevance of these results to associations with prenatal parental occupational exposures to dusts and ionising radiation is explored.

On statistical methods for analysing the geographical distribution of cancer cases near nuclear installations

There is great public concern, often based on anecdotal reports, about risks from ionising radiation. Recent interest has been directed at an excess of leukaemia cases in the locality of civil nuclear installations at Sellafield and Sizewell, and epidemiologists have a duty to pursue such information vigorously. This paper sets out to show that the epidemiological methods most commonly used can be improved upon. When analysing geographical data it is necessary to consider location. The most obvious quantification of location is ranked distance, though other measures which may be more meaningful in relation to aetiology may be substituted. A test based on distance ranks, the "Poisson maximum test", depends on the maximum of observed relative risk in regions of increasing size, but with significance level adjusted for selection. Applying this test to data from Sellafield and Sizewell shows that the excess of leukaemia incidence observed at Seascale, near Sellafield, is not an artefact due to data selection by region, and that the excess probably results from a genuine, if as yet unidentified cause (there being little evidence of any other locational association once the Seascale cases have been removed). So far as Sizewell is concerned, geographical proximity to the nuclear power station does not seem particularly important.

Epidemiological studies of environmental exposure and specific diseases

This paper includes many different epidemiological approaches to environmental health studies, and examples and discussion in terms of a limited number of exposures and diseases are provided. The methods described are not unfamiliar, but, by presenting a panorama in a linked context, it is intended to be constructive in developing ideas.

The Northern region Children's malignant disease registry 1968-82: incidence and survival

All cases of childhood cancer diagnosed before the 15th birthday in the years 1968-1982 and resident in the Northern Health Authority region have been registered. There were 1171 registrations and only two have been completely lost to follow up. The overall annual incidence of cancer was 107.1 per million children, similar to previously reported figures. There was no significant change in the rate over the 15 year period either for all cancers or individual cancer types. Eighty six percent of registrations had central review of pathological material. There has been a significant move towards centralisation of care over the 15 years and a significant improvement in survival over the three quinquennia for all cases and for most individual types. High white blood cell count at presentation was confirmed as a bad prognostic feature in children with acute lymphoblastic leukaemia (ALL). Children treated for ALL in a peripheral hospital had a significantly worse survival than those referred to a specialist centre. Survival rates were calculated for all of the major types of malignancy. The registry includes four sibling pairs with cancer and one family with three siblings affected. Ten children developed secondary primary tumours.

Spatial patterns in electoral wards with high lymphoma incidence in Yorkshire health region

The possibilities of clustering between those electoral wards which display higher than expected incidences of cases of the lymphomas occurring between 1978 and 1982 are examined. Clusters are defined as being those wards with cases in excess (at a probability of less than 10%) which are geographically adjacent to each other. A separate analysis extends the definition of cluster to include high incidence wards that are adjacent or separated by one other ward. The results indicate that many high incidence lymphoma wards do occur close together and when computer simulations are used to compute expected results, many of the observed results are shown to be highly improbable both in the overall number of clustering wards and in the largest number of wards comprising a 'cluster'.

Epidemiology of the lymphomas in the United Kingdom: recent developments

A discussion of the difficulties associated with classification of the lymphomas, particularly by subtype, precedes the descriptive epidemiology of the disease. Recent data collected in the United Kingdom and Yorkshire are presented to illustrate demographic features and variations in lymphoma incidence, both in relation to certain time periods and geographical distribution. Lymphoma aetiology is addressed in the light of current epidemiological approaches, using as an example data from a Yorkshire based case-control study. Three broad areas of aetiological significance--genetics, viruses and immunity--form the bases for presenting descriptions of known or suspected aetiological factors associated with Hodgkin's disease and non-Hodgkin's lymphoma. The final overview points the way to relevant future research.

Variation in lymphoma incidence within Yorkshire Health Region

The spatial distribution of new cases of lymphoma occurring in Yorkshire between 1978 and 1982 has been studied. Administrative districts were used as the basis for analysis and differences in age standardised incidence rates between districts were determined. Excessive rates for NHL were found to occur in Scarborough, York and Leeds districts. In addition an analysis contrasting broadly urban and rural districts showed an excess of NHL in rural areas, particularly of the follicular subtypes.