Tumori, 90: 181-185, 2004
LUNG CANCER MORTALITY IN A DISTRICT OF LA SPEZIA (ITALY) EXPOSED
TO AIR POLLUTION FROM INDUSTRIAL PLANTS
Stefano Parodi1, Roberta Baldi2, Claudia Benco3, Michela Franchini4, Elsa Garrone5, Marina Vercelli5,6,
Floriana Pensa2, Riccardo Puntoni7, and Vincenzo Fontana7
1Epidemiology and Biostatistics Section, Scientific Directorate, G Gaslini Children’s Hospital, Genoa; 2Unit of Public Health, Local Health
Unit 5 (ASL 5), La Spezia; 3Environmental Toxicology Section, Environmental Protection Agency of the Ligurian Region (ARPAL),
La Spezia; 4Department of Epidemiology and Health Service Research, National Research Council (CNR), Pisa; 5Tumor Registry Section,
National Cancer Research Institute, Genoa; 6Department of Oncology, Biology and Genetics, University of Genoa; 7Department of
Environmental Epidemiology, National Cancer Research Institute, Genoa, Italy
Aims and background: In the last decades, many epidemiologi- was found in the two exposed areas among males, after ex-
cal studies have implicated outdoor environmental carcino- cluding rural and semi-rural zones from the analyses
gens in the onset of lung cancer. The present investigation (RR = 1.03 and RR = 0.77). In contrast, a risk excess was ob-
evaluated lung cancer mortality in two areas of the Province served for females in both exposed areas, which remained
of La Spezia (Northern Italy) exposed to environmental pollu- elevated and statistically significant (P <0.05) after restric-
tion emitted by a coal-fired power station and other industrial tion to urban/semi-urban municipalities and after controlling
sources, including a waste incinerator. for deprivation factors (RR = 1.54 and RR = 2.14, respective-
Methods: In the two exposed areas, lung cancer mortality risk ly). Bayesian mapping confirmed the rural/urban gradient
for the 1988-1996 calendar period was evaluated using the and the risk excess observed in females near the industrial
whole Province population as referent. The corresponding rel- sites.
ative risks (RR) were estimated after controlling for age struc- Conclusions: The risk observed among females is consistent
ture, urban/rural gradient and deprivation factors (occupation, with pollution measurements and with other epidemiologic
education, home ownership, housing conditions and family findings, whereas a strong confounding from occupational ex-
structure) by a Poisson regression modeling. The geographic posures and smoking habit could account for the lack of an
pattern of risk for the whole province was evaluated via the excess risk in males. However, the ecologic nature of this in-
Besag, York and Mollié (BYM) bayesian model. vestigation prevented drawing a causal inference. The pollu-
Results: Persons living in urban areas showed the highest tion-related risk observed in the female gender is an impor-
rates in both sexes. No statistically significant risk excess tant clue that deserves further epidemiologic attention.
Key words: Bayesian analysis, ecological monitoring, environmental pollution, lung cancer.
Correspondence to: Dr Stefano Parodi, Epidemiology and Biostatistics Section, Scientific Directorate, G Gaslini Children’s Hospital,
Largo G Gaslini 5, 16145 Genoa, Italy. Tel +39-010-5636301; fax +39-010-3776590; e-mail firstname.lastname@example.org
Received July 7, 2003; accepted November 14, 2003.