INCREASING INCIDENCE OF CHILDHOOD LEUKEMIA IN NORTHWEST ITALY, 1975–98 Corrado MAGNANI 1,2 , * Paola DALMASSO 1,3 , Guido PASTORE 1 , Benedetto TERRACINI 1 , Marco MARTUZZI 4 , Maria Luisa MOSSO 1 and Franco MERLETTI 1 1 Childhood Cancer Registry of Piedmont, Regional Center for Cancer Epidemiology and Prevention (CPO Piemonte), Torino, Italy 2 Medical Statistics Unit, Department of Medical Sciences, University of Eastern Piedmont at Novara, Novara, Italy 3 Medical Statistics Unit, Department of Public Health and Microbiology, University of Torino, Torino, Italy 4 World Health Organization, European Center for Environment and Health, Rome, Italy Although some childhood cancer registries reported in- creasing incidence, the evidence and magnitude of time trends in the incidence of childhood leukemia are debated and the scientific evidence is conflicting. Only limited data have so far been supplied from Southern European countries. We present an analysis of the incidence trend of childhood leukemia in Piedmont (NW Italy) in 1975–98, based on data from the population-based childhood cancer registry. The Childhood Cancer Registry of Piedmont has been recording cases of childhood neoplasms since 1967. Procedures have been uniform and are based on an active search for cases and relevant information. Only cases with confirmed residence in Piedmont at diagnosis are included. Eight hundred cases of leukemia (622 acute lymphoblastic [ALL], 133 acute nonlym- phoblastic [AnLL], 45 other and unspecified) were recorded in the period 1975–98 considered in our study. Incidence trends were analyzed using piecewise regression and Poisson regression, based on annual incidence rates. As results from the 2 analyses were similar, only the former were reported. In the age group 1– 4 years, a statistically significant annual 2.6% increase in incidence rate of ALL (adjusted by age and gender; 95% confidence interval [CI] 1.13– 4.13) was esti- mated. There was no evidence of increase in other age groups. During 1980 –98, a statistically significant 4.4% annual increase (95% CI 1.86 – 6.90) was seen for pre-B-All in the age group 1– 4 years. An increase was also seen for T-ALL that was not statistically significant. Sensitivity analyses were con- ducted, with no relevant differences from the main results. Our data suggest an increasing trend in ALL incidence for children between the ages of 1 and 4 years. These results are unlikely to be explained by changes in quality of data or exhaustiveness in reporting in the study period. The results were not changed in the sensitivity analyses we conducted. Possible causes to be investigated include environmental fac- tors, changes in family size and parental age, socioeconomic conditions and geographical distribution of cases. © 2003 Wiley-Liss, Inc. Key words: childhood leukemia; ALL; AnLL; epidemiology; inci- dence; trends; Italy; cancer registration Temporal trends in disease incidence can reflect changes in the distribution of causal factors but also as changes in diagnostic or registration practices or just chance. Although malignancies in children are a small proportion of the total cancer burden, they are of special interest because the latency is short and time trends are therefore supposed to reflect the variation of causal factors more closely. A number of papers were published in the 1990s on temporal variation in the incidence of childhood neoplasms: most studies supported the evidence for an increasing trend for leukemia, al- though there was wide variation among countries. 1–13 An increase in acute lymphoblastic leukemia (ALL) was reported in the United Kingdom, 4,5 whereas early observations in the United States 8 were later attributed to changes in classification. 11 Two recent studies reported on leukemia incidence in Northern England. McNally et al. 12 described an increasing incidence of ALL, attributable to the pre-B cell subtype, in Northwest England. A study in Yorkshire by Feltbower et al. 13 also showed an increasing trend for pre-B cell ALL, albeit not reaching statistical significance. Incidence rates of childhood leukemia in the Piedmont Region (NW Italy), where a population-based Childhood Cancer Registry (CCRP) has been operating since the sixties, are among the highest in the world. 14 –16 In this article we report on time trends of the incidence of leukemia among children (aged 0 –14 years) resident in Piedmont at diagnosis. The present results provide, as far as we know, the first population-based detailed information on childhood leukemia incidence trends in Southern Europe. MATERIAL AND METHODS Since 1967, the CCRP has been recording cancer incidence in children (aged 0 –14 years). Cases are actively collected through perusal of medical records and other relevant files by trained personnel. The registration is extended to all pediatric wards in the region and to the other wards, both in Piedmont and elsewhere, where a child affected by a malignancy may be admitted. Active registration in hospital wards is supplemented through the search for relevant cases in the files of hospital admissions, reimburse- ment for cancer treatment carried out in other countries, cancer deaths roster and also the data base of the Italian Association of Pediatric Hematology and Oncology (AIEOP). 17 Diagnoses are coded using the ICD-O classification and further grouped using the International Classification for Childhood Cancer (ICCC) crite- ria. 18 Cases are included in the CCRP only after positive confir- mation of residence in Piedmont at diagnosis. For each registered child we check the residence at diagnosis (as reported in the clinical record) at the corresponding Town Office. All Italian residents have to be listed in the roster of a town, and the infor- mation is quickly updated in case of moving or change of status. Additional information is searched for the occasional nonmatching cases until the town of residence at diagnosis is identified. The procedures adopted for data collection and coding were uniform throughout the period of activity of CCRP. Further infor- mation is available elsewhere. 16 The present analyses were limited to the period 1975–98 in order to exclude the effect of improve- ment in diagnostic methods that took place in the early seventies. White blood cell (WBC) count at diagnosis and immunologic typing were systematically abstracted from the clinical records Grant sponsor: the Regional Administration of Piedmont; Grant sponsor: the Italian Association of Cancer Research; Grant sponsor: the Ministry of University and Research (MUIR). *Correspondence to: Childhood Cancer Registry of Piedmont, V. San- tena 7, 10126 Torino, Italy. Fax: +39-011-6336960, E-mail corrado.magnani@cpo.it Received 30 July 2002; Revised 13 January 2003; Accepted 17 January 2003 DOI 10.1002/ijc.11112 Int. J. Cancer: 105, 552–557 (2003) © 2003 Wiley-Liss, Inc. Publication of the International Union Against Cancer