Vegetables and fruit and human cancer: Update of an Italian study

A substantial amount of epidemiological evidence indicates that vegetables and fruit are associated with reduced risk of several cancer sites (Willett and Trichopoulos, 1996; World Cancer Research Fund, 1997), although the issue is open to discussion (Botterweck et al., 1998). The relation between frequency of consumption of vegetables and fruit and cancer risk was analyzed using data from a series of case-control studies conducted in northern Italy between 1983 and 1990 (Negri et al., 1991). For digestive tract neoplasms, the relative risks ranged between 0.2 and 0.5 for the highest compared with the lowest tertile of vegetable intake. Protective effects were also observed for hormone-related epithelial neoplasms. Fruit was related to reduced relative risks for cancers of the upper digestive tract, stomach and urinary tract (Negri et al., 1991). For digestive tract cancers, population attributable risks for low intake (i.e., <2 vs. ≥4 portions per day) of vegetables and fruit ranged between 15 and 40% (Tavani and La Vecchia, 1995). Given the widespread interest in the issue (Nelson, 1996; Potter and Steinmetz, 1996), we have updated our findings, including data collected up to 1997. Thus, the relation between frequency of consumption of green vegetables and fruit and cancer risk has been analyzed using data from an integrated series of case-control studies conducted in greater Milan and the province of Pordenone, northern Italy, between 1983 and 1997. For breast and colorectal cancers, only data collected until 1991 were included. The overall dataset included the following incident, histologically confirmed cancers, below age 75 years: oral cavity and pharynx, 524; esophagus, 410; stomach, 745; colon, 955; rectum, 625; liver, 435; gallbladder, 65; pancreas, 402; larynx, 388; breast, 3,412; endometrium, 750; ovary, 971; soft tissue sarcomas, 217; prostate, 127; bladder, 431; kidney, 190; thyroid, 428; Hodgkin's disease, 201; non-Hodgkin's lymphomas, 529; myelomas, 185; and a total of 10,058 controls admitted to hospital for acute, non-neoplastic conditions, unrelated to long-term dietary modifications (30% traumas, 16% non-traumatic orthopaedic conditions, 29% acute surgical diseases, 25% other miscellaneous). All interviews were conducted in hospital, and, on the average, refusal rate of eligible subjects (cases and controls) was below 5%. All questionnaires comprised a basic structured section, including sociodemographic factors and general characteristics and habits. Patients were asked to indicate the frequency of consumption per week of selected indicator foods, including between 14 and 37 items (D'Avanzo et al., 1997). The questionnaires also included a summary question on total vegetable and fruit consumption, thus making possible a combination of data from various studies. Multivariate odds ratios (OR) and corresponding 95% confidence intervals (CI) for subsequent tertiles of vegetable and fruit consumption were derived after allowance for age (quinquennia), sex, calendar year at interview, area of residence (Milan/Pordenone/other northern regions/other regions), education (<7, 7–11, ≥12 years), smoking (never, ex, current ≤15, ≥15 cigarettes per day) and alcohol consumption (0, <2, 2–4, >4 drinks per day) (Breslow and Day, 1980). Further allowance for body mass index did not appreciably modify any of the ORs. The best approximate tertile cutoffs were 7 and 10 portions per week for vegetables and 7 and 14 portions for fruit. For vegetables (Table I), a consistent pattern of protection was observed for all epithelial cancers, with ORs in the highest tertile of intake between 0.3 and 0.7 for digestive sites, 0.4 for larynx, between 0.5 and 0.8 for breast and female genital tract, around 0.2 for cancers of the prostate and urinary tract and 0.8 for the thyroid. In contrast, no appreciable association was observed for non-epithelial neoplasms, the ORs ranging between 0.7 and 1.1 for the highest tertile of consumption for lymphomas and myeloma. None of these estimates for non-epithelial neoplasms was significant. Likewise, for soft tissue sarcomas, which were not included in the previous report (Negri et al., 1991), the ORs were 1.1 for the intermediate and 1.3 for the highest vegetable consumption tertile. With reference to fruit (Table II), strong protection was observed for cancers of the upper digestive tract (oral cavity and pharynx, OR = 0.5 for the highest tertile), esophagus (OR = 0.4), stomach (OR = 0.4), as well as for colon (OR = 0.6), rectum (OR = 0.8), gallbladder (OR = 0.4), pancreas (OR = 0.7) and larynx (OR = 0.5). Significant inverse trends in risk were also observed for cancers of the prostate (OR = 0.5), bladder (OR = 0.5) and kidney (OR = 0.6). The associations were inconsistent and showed no appreciable protection for other epithelial neoplasms, including breast, female genital tract and thyroid, as well as for lymphomas and myeloma. For soft tissue sarcomas, the ORs were 1.1 and 1.0 for the intermediate and high fruit consumption tertiles. When subjects in the highest tertile of both vegetable and fruit consumption were compared with those in the lowest tertile of both variables, the ORs were 0.2 (95% CI = 0.1–0.3) for esophagus, 0.3 (95% CI = 0.2–0.4) for stomach, 0.5 (95% CI = 0.3–0.6) for colon, 0.3 (95% CI = 0.2–0.4) for larynx, 0.1 (95% CI = 0.0–0.3) for prostate, 0.1 (95% CI = 0.1–0.2) for bladder and 0.2 (95% CI = 0.1–0.4) for kidney cancer. For other neoplasms, no appreciable joint effect of fruit and vegetable consumption was observed. Our findings confirm, on the basis of a substantially larger dataset, the observation that in this population frequent green vegetable intake is associated with a markedly reduced risk of several common neoplasms and that fruit intake has a favorable effect, too, especially on cancers of the digestive and urinary tracts (Negri et al., 1991). The consistency of the findings in subsequent time periods (1983–1990 and 1991–1997) provides further support for the existence of a real association. However, it has been suggested (Botterweck et al., 1998) that the inverse association of vegetable and fruit consumption is less clear in cohort than in case-control studies, at least for gastric cancer. Even in the absence of a univocal and clear biological explanation (Potter and Steinmetz, 1996), these findings nonetheless may have an immediate and substantial impact on a preventive and public health level (La Vecchia and Tavani, 1998). This work was conducted with the contribution of the Italian Association for Cancer Research. M.S. is a recipient of a fellowship awarded by the Zambon Group, Spain. Carlo La Vecchia* , Liliane Chatenoud*, Silvia Franceschi , Maria Soler*, Fabio Parazzini* ?, Eva Negri*