Decreasing incidence of all histological subtypes of oesophagus cancer in Tuscany, Italy.

Abstract

In a paper recently published in this journal, Levi et Ž . al. 2001 analysed the rise in incidence of oesophageal adenocarcinomas observed in the Registry of Vaud, Switzerland. During the long peŽ . riod examined 1976 1998 there was an overall decrease in the incidence of oesophagus cancer due to the balance between the decrease in squamous Ž cell carcinomas about 66% of all oesophagus can. cers in Vaud in 1995 1998 , an increase for adenoŽ . carcinomas about 22% and a decrease for cases Ž . with other and unspecified morphology about 12% . The contrasting trends between squamous and adenocarcinomas may be due, according to the authors, to the different trends of morphology-specific risk factors. The evidenced increase in the rates for adenocarcinomas of the oesophagus agreed with similar results in other but not all 10 European Ž . countries recently analysed Botterweck et al., 2000 . We want to contribute with further descriptive Ž . data from the Tuscany Cancer Registry RTT . The RTT is a population-based cancer registry active in the provinces of Florence and Prato, central Italy Ž . since 1984 Buiatti et al., 1997 . The age-adjusted incidence rates for oesophagus cancer in the RTT area around the 1990s were 3.7 per 100 000 among Ž . males and 0.9 among females Parkin et al., 1997 . Ž These rates were on the Italian average range . 1.0 9.7 among males and 0.4 1.5 among females Ž but lower than those from Vaud 9.5 for males and . 2.0 for females, respectively and from most other Ž . European countries Parkin et al., 1997 . Ž During 1985 1997, 613 oesophagus cancers 427, . 69.7% males and 186, 30.3% females were diagŽ nosed among the resident population about . 1 180 000 at the 1991 census . Incidence rates from Ž 1985 to 1997 directly standardized to the European . Ž population , percentage change PC, as the percentage difference of the mean of the rates of the last 2 . years minus that of the first 2 years considered and the expected annual percentage change in a logŽ . linear model EAPC were computed with the Ž . SEER Stat software seerstat@ims.nci.nih.gov . The overall incidence trend was downward with a PC of Ž 32.4%, similar in both sexes PC 35.4% among . males and 31.4% among females . There was an Ž overall statistically significant EAPC of 3.5% 95% . CI 6.3 0.7 also significant among males Ž . EAPC 4.3, 95% CI 7.9 0.5 but not among Ž . females EAPC 3.0, 95% CI 6.1 0.2 probably because of the relatively small sample. Ž . Following Levi et al. 2001 , we analysed the following morphological categories: squamous cell carŽ . cinoma ICD-O M 8050 8082 , adenocarcinoma Ž . 8140 8573 and other and unspecified cancers. In our data set there were 299 squamous cell carcinoŽ . Ž . mas 48.8% , 71 adenocarcinomas 11.6% and 243 Ž . 39.6% other and unspecified cases. The proportion of microscopic confirmations was rather low although it had improved over time, being only 53% in 1985 and reaching 70% in 1997. The time trends were similar and decreasing for all the three histological categories. In fact, squamous cell carcinomas decreased with a rate of 2.6% every Ž . year 95% CI 5.5 0.5 , adenocarcinomas at Ž . 3.8% 95% CI 10.3 3.1 and others 5.5% Ž . 95% CI 10.1 0.7 . It is unlikely that completely different results would be expected if the proportion of microscopic verifications had been higher. In fact, it would mean to hypothesize that in the unspecified group there would be hidden not only an increasing number of adenocarcinomas but also that this increasing trend would be so strong as to modify the decreasing trend evidenced for the histologically verified cases. Also, another possible source of bias, the misclassification of adenocarcinomas to other bothering cancer sites, may be excluded. In fact, both cardia and hypopharynx cancers had the same decreasing trend as Ž oesophagus. With regard to cardia cancers about 36 . cases diagnosed every year , during the same period Ž there was a yearly 3.9% trend 95% CI 9.3 . 1.9 that was also present for adenocarcinomas