Cancer incidence in EPIC-ITALY at first follow-up.

Abstract

A descriptive analysis was performed in order to evaluate the completeness of follow-up and to explore the occurrence of malignancy in the Italian section of the European Prospective Investigation into Cancer and Nutrition (EPIC-Italy) at the first follow-up for cancer incidence. The EPIC-Italy cohort consisted of 47,749 subjects, aged 35-65 years, who voluntarily accepted to participate in the project from 1992-1997. Tabulations of the enrolled subjects are presented by sex, age groups, population at risk and person-years as calculated at the first follow-up in 1998; alive, lost to follow-up and dead subjects were tabulated by sex and centers. Cancer occurrence is described by quality indexes of data collected and the lapse of time between the date of recruitment and date of diagnosis. External comparisons for each center and pooled data were carried out by calculation of standard incidence ratios (SIRs) using the rates of the population-based cancer registries covering the areas of EPIC Italian centers. Similarly, an internal comparison was also performed using as the reference population the EPIC-Italy center with the lowest crude cancer incidence rates. A total of 148,968 person-years (43,568 men and 105,400 women) was calculated as the denominator; the percentages of lost to follow-up were 1.34% for men and 0.9% for women. We found 781 malignancies (216 in males and 565 in females), 17 in situ breast cancers and 8 in situ cervical cancers. In men 65 malignant cases (30.1%) and in women 186 malignancies (32.9%) occurred in the first year following enrollment. The proportion of microscopically verified cancers was 93.1%. In pooled data for men, statistically significant SIRs of less than 1 were calculated for all cancer sites combined (SIR = 0.81), lung (SIR = 0.49) and bladder (SIR = 0.62), whereas statistically significant excesses of observed cases were found for melanoma and cancers of ill-defined sites within respiratory system and intrathoracic organs (ICD-0-2 = C39). In pooled data for women, none of the SIRs were statistically significant. For men, SIRs disaggregated by center showed statistically significant excesses of cases only in Florence for the thyroid (SIR = 5.01). For women, statistically significant excesses of observed cases were computed in Florence and Varese for breast cancer (SIRs, 1.36 and 1.27, respectively), Florence for uterine cancer NOS (SIR = 20.3) and Varese for kidney (SIR 2.24). Internal comparisons showed some evidence of an excess of cases in northern compared to southern centers. In conclusion, after an average follow-up of 3 years, nearly 99% of the enrolled individuals were traced and checked for their vital status. Cancer occurrence in women was not far from that expected in comparison to the local general population, whereas in men significantly fewer cases than expected were observed. This preliminary descriptive analysis will be used as a starting point for monitoring the validity of EPIC-Italy over time.