Risk Factors for Oral and Pharyngeal Cancer in Shanghai, with Emphasis on Diet1

A population-based case-control study of oral and pharyngeal cancer was conducted in Shanghai, China, from 1988 to 1990, in which 204 (115 male, 89 female) incident cases and 414 (269 male, 145 female) controls were interviewed. Cigarette smoking and alcohol consumption, as well as occupational exposures to asbestos and to petroleum products and the use of kerosene stoves in cooking, were associated with increased risk of oral and pharyngeal cancer. In addition, more cases than controls reported having chronic oral diseases and false teeth. Dietary intakes of 42 major foods and selected salt-preserved or deepfried foods during the past 10 years, ignoring any recent changes, were measured by a strudured quantitative food questionnaire. After adjusting for known etiological factors, risks decreased with increasing intake of fruits, particularly oranges and tangerines, and some vegetables, including dark yellow vegetables and Chinese white radish. Men in the highest tertile of intake of these fruits and vegetables had about 30-50% the risk of those in the lowest tertile, with a less pronounced effed among women. A new finding was an excess risk associated with high consumption of salt-preserved meat and fish. The findings from this study provide further evidence that dietary factors play an important role in the development of oral and pharyngeal cancer. Introdudion Cigarette smoking and alcohol Consumption are the primary determinants of oral and pharyngeal cancer in most areas of the world (1). Risk has been shown to increase with the intensity and duration of both smoking and drinking and to decrease following cessation of these Received 12/18/91. , Supported by a research grant from the Young Scientist Foundation, Shanghai Municipal Bureau of Health. 2 To whom requests for reprints should be addressed, at National Cancer Institute, Executive Plaza North, Room 431, Bethesda, MD 20892. habits. Dietary factors have also been linked to this cancer (1). Earlier in this century, Plummer-Vinson syndrome, caused by deficiency of iron and some other nutrients, was related to the high incidence of oral and pharyngeal cancer among Swedish women (1). Recent epidemiological studies, mostly in Western countries, have indicated that other dietary factors, notably a low intake of certain types of fruits and vegetables, also increase the risk ofthis cancer (2-1 1). In a previous study of oral and pharyngeal cancer in China (Beijing), smoking and alcohol were found to be important determinants, but relationships to dietary factors were not reported (12). To provide further evidence on dietary and other risk factors for this cancer, we analyzed data from a recent population-based case-control study conducted in Shanghai, where the dietary pattern is quite different from that of Western societies. Materials and Methods All residents of urban Shanghai aged 20-75 years who were newly diagnosed with oral or pharyngeal cancer (9th Revision lCD3 141, 143-6, 148-9) during the period January 1, 1988 to February 28, 1990 were eligible for the study and are referred to simply as cases of oral cancer in the subsequent text. Cancers of the salivary gland (lCD 142) and nasopharynx (lCD 147), which are thought to have different causes than oral cancer, were not included in the study. A total of 261 eligible cases were identified from a population-based cancer registry in Shanghai during the study period. This registry, which started operation in 1963, has essentially a complete ascertainment of all cancer cases diagnosed in the urban Shanghai area (13). Of the cases identified, 204 (78.2%) were interviewed, while 57 (21 .8%) were not included because ofdeath (18.0%) and unlocatability (3.8%). Data for pathologic diagnoses were collected for the large majority (92.4%) of the cases. Controls were randomly selected from the general population of the Shanghai urban area. The number of controls in each sexand age(5-year interval) specific stratum was determined in advance according to the sex and age distribution of the incident cases of oral, pharyngeal, laryngeal, and nasal cancers reported to the Shanghai Cancer Registry in 1985 and 1986. The personal registry cards kept in the Shanghai Resident Registry were used to select controls. For all adult residents in urban Shanghai the cards contain name, address, date of birth, sex, and other demographic factors. The cards are kept in a total of 4410 drawers, with approximately equal on October 13, 2017. © 1992 American Association for Cancer Research. cebp.aacrjournals.org Downloaded from 442 Dietary Fadors for Oral Cancer in Shanghai numbers of cards in each drawer, which is numbered by a unique 4-digit code. For each control, a 4-digit random number was generated to define the drawer from which the control was randomly selected. A second potential control was also selected. Of the 414 controls interviewed, 48 (12%) were second controls. The major reasons for using second controls were that the first control died before interviewing (2%) or could not be located (10%). No controls or cases refused to be interviewed. The cases and controls were interviewed in person by trained personnel, usually at the homes of the subjects. A structured questionnaire was used to elicit detailed information on demographic factors, tobacco smoking, alcohol consumption, dietary habits, occupational history and exposures, and previous disease history. Diet was assessed by asking about the “usual frequency and amount of consumption in the previous 10year period, ignoring any recent changes” of 30 fruits and vegetables and 1 1 animal foods, which are the major foods in the diets of Shanghai residents (Appendix). Information on the consumption frequency of salt-preserved and deep-fried foods, as well as supplementary vitamin intake, was also obtained. The monthly amount (in grams) and monthly frequency of consumption were derived for each food item. For seasonal foods, the average monthly intake was obtamed by weighing the length of time the food was in the market in a 1-year interval. Food group intake was calculated as the sum of the monthly frequencies or grams of the food items comprising the food group. Individual nutrient intake was computed using a Chinese food composition table (14) and the amount of food consumed. The Mantel-Haenszel stratified estimation method and unconditional logistic regression were used to adjust for confounders and derive adjusted ORs (15). Factors found to be related to cancer risk in this study were examined for their potential confounding effect on other variables individually and in combination. After adjustment for smoking and education, adjustment for other factors such as age, alcohol consumption, and income did not substantially change the ORs of oral cancer with dietary factors. The strata used for smoking adjustment were as follows: nonsmoking, light (<25 pack-years) and heavy ( 25 pack-years) smoking for males, and ever smoking (yes/no) for females. The strata of education used for adjustment are specified in Table 1. In the diet analysis, food intakes were grouped into categorical vanables according to the tertile distribution among controls because no a priori cut points on threshold effects had been reported. Since the results based on frequency of consumption were similar to those based on total grams consumed, only the latter results are presented in the paper. Tests for trend across the tertiles were performed in logistic regression by assigning scores of 1, 2, on 3 to the first, second, and third tertiles, respectively. All statistical tests were based on two-tailed probability.