Cancer increased after a reduction of infections in the ®rst half of this century in Italy: Etiologic and preventive implications Giuseppe Mastrangelo, Emanuela Fadda & Giovanni Milan Institute of Occupational Medicine, University of Padua, Italy Accepted in revised form 10 August 1998 Abstract. Two rate ratios indicating the disappear- ance of infections and the growth of tumours, re- spectively, were simultaneously plotted against the calendar years of occurrence in a period during which mortality rates were reasonably comparable to inci- dences. The transformation used gave upward trend time variations for infectious diseases, providing strong evidence that in Italy during the ®rst half of this century variations in infectious diseases preceded variations in cancer. While some bacteria and viruses are known to be cancer agents, sparse studies indicate that a host's immune response to infection may de- stroy cancer cells. With a decreasing mortality from infectious illnesses, there may have been a reduction in the activation of immunological mechanisms against transformed cells in early phases of carcino- genesis. If cancer growth is a consequence of a lower exposure to chronic sublethal doses of microbial agents, bacterial derivates could be potentially useful in cancer chemoprevention. Key words: Biological response modi®ers, Chemoprevention, Ecological studies, Endotoxin, Epidemiology, Infections, Neoplasms Abbreviations: ISTAT  national institute of statistics (Italian abbreviation); lnPR  logarithm of the pro- duction ratio of cigarettes (see PR); lnRR  logarithm of the rate ratio (see RR); PR  production ratio: ratio between cigarette production in the index year and that in 1895; RR  rate ratio: (1) for infectious diseases, ratio between the age world standardized mortality rate in 1895 and that in the index year; (2) for tumours, ratio between the age world standardized mortality rate in the index year and that in 1895; TNF  tumour necrosis factor Introduction In 1891, following his original observation of cancer remission after erysipelas, Coley used bacterial lysates containing endotoxin for cancer therapy [1]. The treatment, which induces haemorrhagic necrosis in tumours, was abandoned with the advent of chemo- therapy and radiotherapy and with the progress in surgery. The phenomenon observed by Coley re- ceived new attention in the 1980s, when Old and collaborators demonstrated that haemorrhagic ne- crosis was not induced by endotoxins per se, but depended on the tumour necrosis factor (TNF), a host factor produced by macrophages [2]. This phe- nomenon probably also involves a mixed response from Th1 and Th2 (the two maturation pathways of T cells), when the in¯amed tissue becomes extremely sensitive to TNF, and the release of cytokines causes necrosis, ®rst involving the microvasculature and la- ter the entire tissue [3]. Another mechanism, by which TNF acts as an additional non-toxic macrophage- activating factor, depends solely on a response from Th1 cells, which produce a set of signals that or- chestrate attacks on unfamiliar cells, thus protecting the body against bacteria and tumour cells [4]. Fur- thermore, an inverse relationship between recent in- fectious diseases and the occurrence of cancer has been found in several epidemiologic studies, and summarised by other authors [5, 6]. Several common viruses are known to cause speci®c malignancies; the major virus±malignancy associa- tions include hepatitis virus B and C and hepatocel- lular carcinoma; human lymphotropic virus-type 1 and adult T-cell leukemia-lymphoma; Epstein±Barr virus and endemic Burkitt lymphoma, nasopharyn- geal carcinoma, and Hodgkin's disease; and human papilloma virus and cervical cancer [7]. Regarding bacteria, attention has currently focused on non-spe- ci®c mechanisms underlying carcinogenesis rather than on speci®c organisms; examples are: the induc- tion of chronic in¯ammation (Helicobacter pylori and gastric cancer) and the production of mutagenic compounds by bacterial metabolism (resident colonic ¯ora and colon cancer) [8]. It was recently suggested that one-third of the world's cancer burden has an infectious origin [9]. Criteria now used to determine the causality of an association are: strength of the association and its European Journal of Epidemiology 14: 749±754, 1998. Ó 1998 Kluwer Academic Publishers. Printed in the Netherlands.