An investigation of clustering of cases of leukaemia in Jamaican children

ABSTRACT

The cause of leukaemia is not yet known. Recently attention has been drawn to its epidemiological aspects and in particular to clustering of cases, i.e. association of cases in regard to time of occurrence and locality. If this phenomenon can be demonstrated, it suggests an environmental factor, virus infection being one possibility. The present study is an attempt to apply mathematical methods of analysis to 58 cases of childhood leukaemia presenting in Kingston, Jamaica, from January 1, 1958, to December 31, 1966. Of these 58 cases, 37 were domicile in the parish of Kingston and St. Andrew and 21 were from the rest of Jamaica. Owing to geographical factors it is felt that whereas the case finding rate for the Kingston and St. Andrew cases is high, the number of cases from the Jamaican countryside which are included in this series can represent only a fraction of the expected total. The sex incidence in the series was 30 boys to 28 girls. There were 35 cases of acute lymphatic leukaemia, 8 acute stem cell, 10 acute myeloid, 4 chronic myeloid and one case of erythro-leukaemia. The peak age incidence was in the fifth year. Acute lymphatic leukaemia was mostly found in those under six years, acute myloid in the over sixes and in the under twos. The incidence of recorded cases of leukaemia has risen over the period studied. In 1964 there was a distict peak of 15 cases (an epidemic year). Some degree of seasonal incidence was shown in a reduction in cases in the last three months of the year, the six months April to September having 37 cases as against 21 for the other half year. Finally the incidence of pairing in Kingston cases was - Leukemia pairs - distance 1 Km, 1 Km, 2 Km, time 1 year, 6 months, 6 months, respectively, expected pairs 12.6, 6.5, 17.9 respectively, observed pairs 16, 7, 22, respectively. The degree of clustering is consistently greater than that calculated on a chance basis. However the increase of observed pairs over expected pairs does not achieve statistical significance. A larger series might be needed to demonstrate clustering in a conclusive fashion (AU)