Gastric cancer incidence in the Romagna Region of Italy: A spatial and temporal analysis

Abstract

Background: The worldwide decrease in gastric cancer incidence is due to a birth-cohort-dependent decrease in exposure to major risk factors.

Methods: In an area of northern Italy with a historically strong internal geographical gradient in gastric cancer incidence, variations in rates by municipality and age group between 1987 and 2008 were evaluated. The study period was divided into three nonadjacent periods. End points included: age- and sex-standardised incidence rates; incidence rate ratio between age- and sex-standardised incidence rates; smoothed relative risks of gastric cancer incidence, and posterior probabilities of the relative risk being >1.

Results: In 1987-1990, the estimate of posterior probabilities of relative risk being >1 showed a higher incidence in hilly/mountainous areas. Between 1987-1990 and 2005-2008, a uniform decrease of more than 50% was observed (incidence rate ratio: plain, 0.45 (95% confidence interval 0.40-0.51); hill, 0.44 (0.34-0.58); mountain, 0.48 (0.22-1.02)). The decrease in the mountainous area was weak in the middle time period, with an incidence rate ratio of 0.92 (0.46-1.84), and intensified afterwards. The decrease occurred earlier and was more pronounced among younger people. In 2005-2008, gastric cancer risk was uniform across ages and municipalities.

Conclusions: The observed changes in gastric cancer incidence is the epilogue of a birth-cohort-dependent decrease in exposure to major risk factors.

Keywords: Birth cohort; Gastric cancer; Incidence; Spatial epidemiology.