Article
Etiologic research on cancers: potential niches for ecologic studies
Search Medline for
Authors
-
Andreas Stang
- Universitätsklinikum Essen, Essen, Deutschland
| Published: | August 27, 2015 |
|---|
Outline
Text
Introduction: Many previous descriptive studies on cancer incidence focused on country-wide incidence estimates and found substantial international variation. A low within-country variation and a large between country variation in cancer incidence may indicate that ecologic factors are involved in the etiology of the disease. The aim of this study is to explore the within and between country variation of cancer incidence to motivate high-quality ecologic
studies.
Methods: We used Cancer Incidence in Five Continents, Volume X, to select population-based cancer registry data. We extracted age-standardized (world standard population) incidence rate estimates from 135 regions for the 10 most frequent invasive cancers for each sex for non-Hispanic white populations. We fitted a weighted multilevel Poisson regression model with a random country effect for each cancer and sex separately. To account for different precisions in rates, rates were weighted with the inverse of the squared standard error. We estimated intraclass correlation coefficients (ICC) and 95% confidence intervals (95%CI).
Results: Patterns of variation within and between countries differed by cancer and sex. The four cancer sites with the highest ICC among men included prostate cancer (0.96, 95%CI: 0.92-0.99), skin melanoma (0.78, 0.64-0.93), stomach cancer (0.77, 0.64-0.91) and lung cancer (0.77, 0.64-0.89). Among women, these sites included lung cancer (0.84, 0.73-0.95), breast cancer (0.80, 0.69-0.91), skin melanoma (0.78, 0.63-0.92) and cervical cancer (0.71, 0.55-0.87). The two most prominent sex differences for ICC occurred for head & neck cancer (men: 0.64, 0.45-0.83, women: 0.18 (0.07-0.30)) and breast cancer (men: 0.04 (0.01-0.07), women: 0.80).
Discussion: ICCs varied by cancer site and sex indicating that ecologic studies may be more or less promising. For
example, the ICC for head & neck cancers was quite high for men but quite low for women indicating that ecologic studies on these cancers may be more fruitful among men than women. Pancreas cancer, uterine cancer, ovarian cancer and leukemia show very low ICC and are not priority candidates for ecologic studies. Future ecologic studies should account for a time lag between exposure prevalence data and cancer incidence data and restrict analyses to high-quality data sets only.
