Association among childhood leukemia, growth and population grouping: study of data from São Paulo municipality

• Grant number: 05/00857-2
• Support Opportunities: Scholarships abroad - Research
• Effective date (Start): October 10, 2005
• Effective date (End): March 09, 2006
• Field of knowledge: Health Sciences - Collective Health - Epidemiology
• Principal Investigator: Karina de Cássia Braga Ribeiro
• Grantee: Karina de Cássia Braga Ribeiro
• Host Investigator: Patricia a Buffler
• Host Institution: Hospital A C Camargo. Fundação Antonio Prudente (FAP). São Paulo , SP, Brazil
• Research place: University of California, Berkeley (UC Berkeley), United States

Abstract

Leukemias are the most frequent childhood neoplasia worldwide, corresponding to approximately one-third of all malignancies in this age group. Acute lymphocytic leukemia is the most common sub-type. Most leukemia cases still have an unknown etiology; however, several environmental and genetic risk factors have already been described. At the end of the 1980’s, KINLEN suggested that the excess of leukemia cases noted near nuclear power plants in Seascale, United Kingdom, could be attributed to a rare response to a subclinical infection; transmission was facilitated by contact between many individuals. According to KINLEN, a large group of individuals (many with urban backgrounds and, therefore, exposed to a variety of infections) moving to a sparsely populated area could facilitate this transmission. On the other hand, in areas with high a demographic density of children could be less susceptible, due to inherited immunity. Intense exposures would be leukemogenic, while weak and moderate ones could confer immunity. KINLEN’s hypothesis has been supported in many other studies, where isolated populations were exposed to population mixing or other sources of infectious agents. Similar results were obtained in other countries and it seems that among children with ALL, effects are more pronounced in those aged 4 or less, living in rural areas, reflecting low demographic density and relative isolation. Some authors had used demographic density as a predictor of high cancer incidence rates The city of São Paulo has more than 10 million inhabitants, including 2.5 million children under the age of 15 years. More than 700,000 people compose the rural population (FOUNDATION SEADE 2004). In the whole city, there is a large variation in the demographic density (0.42 hab/km2, Marsilac and 242.86 hab/km2, Bela Vista) as well as in the rates of population growth (measured from 1996 to 2000) that ranged from -2.68 (Pinheiros) to 7.78 (Anhanguera). From 1969 to 1998, pediatric leukemias incidence rates (0-14 years) ranged from 3.5 to 6.9 cases/100.000 inhabitants among male and from 2.5 to 8.6 cases /100.000 inhabitants among female. The observed incidence rates in the period of 1997-1998, as much for boys (8,24/100.000) as for girls (6,16/100.000), are among the highest worldwide and, curiously, these rates are quite similar to the rates reported in the Hispanic population of Los Angeles, California, United States.OBJECTIVE: To evaluate the association between childhood leukemia incidence and population mixing in the city of São Paulo.METHODS: Study design: case-control studyStudy population: The cases will be children with leukemia, who are under 15 years of age, registered in the São Paulo City Population-Based Cancer Registry (RCBP-SP) during the period between 1997 and 2001. For each case, a control will be matched by age and sex. The control is a child (< 15 years) with a solid tumor, chosen randomly from the same RCBP-SP database. Cases and controls will be classified according to their home address regarding: Social Exclusion Index, Poverty Index, Social Inequality Index (CAMPOS et al. 2004), population growth rate (0-10.0%, 10.1-20.0%, > 20.0%), demographic density, and housing situation (urban or rural area). The rates of population growth and the demographic density of each one of the 96 administrative districts of the city of São Paulo will be those from Demographic Censuses of 1991 and 1996. Descriptive statistics will be used to describe quantitative and categorical variables. Conditional logistic regression models (including crude and adjusted odds ratios, with respective 95% confidence intervals) will be built in order to assess the role of population mixing as a risk factor for childhood leukemia. All possible interactions will be tested. Goodness-of–fit will be evaluated through a Hosmer and Lemeshow test. (AU)