Tohoku J. exp. Med., 1964, 82, 218-223 Coffee Consumption and Mortality for Prostate Cancer By Eiji Takahashi From the Department of Hygiene, Tohoku University School of Medicine, Sendai (Received for publication, January 28, 1964) National average consumption of coffee per person per year is estimated for 5 years (1955-59) on 20 not-producing countries from data in the Yearbook of International Trade Statistics. Age-adjusted death rate from cancer of prostate and of other urogential and digestive organs is adopted from data of Segi and WHO. Correlation between estimated national average consumption of coffee and mortality for prostate cancer is on the significance level, but not between coffee consumption and cancer mortality of other urogential and digestive organs, though sugar consumption has significant correlation not only with prostate cancer but also with cancer of some other organs. The significance of much coffee drinking to prostate cancer is discussed. There are big differences on the dominant cause of deaths between Western countries and Asian countries. Although in some Asian countriess till infectious diseases are dominant as the cause of death, in Japan now degenerative diseases are the most dominant. However, mortality of some kind of degenerative diseases is very different in Japan from that in Western countries. Mortality from prostate cancer and breast cancer is very low in Japan compared to the Western countries. It is epidemiologically interesting to study the reason for the difference between these countries. Living conditions and customs especially alimentary habits are different be tween these countries, though the origin of the nation is also different. Japanese people customarily take green tea as daily beverages, but not so often coffee or black tea as the Western people do, though the habit is changing little by little and tending to westernize. Long ago the author himself used to feel slight stimulation on the part of prostate after coffee drinking and it was supposed that there should be some effect of coffee on the prostate. Recently, yearly import amount of coffee in not producing countries is found in the Yearbook of International Trade Statistics, United Nations1). 218
Coffee Consumption and Prostate Cancer 219 Estimated coffee consumption and mortality for prostate cancer National average consumption of coffee per person per year is estimated for 5 years, 1955-59, on 20 countries: Austria, Belgium, Canada, Czechoslovakia, Denmark, Finland, France, Germany (Federal Republic), Greece, Hungary, Israel, Italy, Japan, Netherlands, Norway, Portugal, Sweden, Switzerland, United Kingdom and United States of America. Export amount of coffee after manufactur ing in these countries until 1959 was rather small compared to import amount and it seems to be negligible. As shown in Fig. 1, Sweden, Denmark, Norway, Finland, U.S.A., etc. are the countries of the highest national coffee consumption per person. Japan belongs to the lowest. The nation of the United Kindgom rather consumes more tea and less coffee than other nations in Europe. Fig. 1. Estimated national average consumption of coffee and age-adjusted death rate for prostate cancer in 20 countries. Age-adjusted death rate for prostate cancer is calculated for 4 years (1956 59) from Segi's international statistics of cancer mortality for 17 eountries2), excluding Greece, Hungary and Czechoslovakia. As for these 3 countries the age adjusted death rate is substitutionally calculated only for 1959 from the data published in the Annual Epidemiological and Vital Statistics, WHO3). As shown in the right side of Fig. 1, age-adjusted death rate for prostate cancer does not regulary diminish according to the estimated national average consumption of coffee. However, the tendency should be recognized that in northern and western Europe where the coffee consumption is higher the death rate for prostate cancer is also higher, and that in eastern Europe both are lower. The correlation coefficient between the estimated national average coffee consumption and the mortality for prostate cancer is calculated by these 20
220 E. Takahashi countries, though the former does not show exact normal distribution. The correlation diagram is as shown in Fig. 2. The correlation coefficient is on the significance level of P<0.001. Japan is exceptionally low both in coffee consumption and in prostate cancer mortality. The prostate cancer mortality of Greece, Hungary and Czechoslovakia is only for one year. Also when these 4 countries excluded, the correlation coefficient still remains on the significance level of P<0.01. Fig. 2. Correlation diagram between estimated national average consumption of coffee and age-adjusted death rate for prostate cancer by 20 countries. Coffee consumption and prostate hyperplasia Not only prostate cancer but also hyperplasia of prostate should be considered as a disease which might be caused by some stimulation of prostate. Age adjusted death rate for hyperplasia of prostate on the age group of 50 and over in 1959 is computed from WHO report3) on 20 countries above mentioned. Correla tion coefficient between estimated national average consumption of coffee per person per year and corrected death rate from hyperplasia of prostate is not on the significance level, as shown in Fig. 3. Prostate hyperplasia itself should not make a direct cause of death, but uremia accompanied might cause death. Therefore, in such a country where the surgical operation of prostate hyperplasia is common, mortality from prostate hyperplasia might not be so high, even if there are many patients. The United
Coffee Consumption and Prostate Cancer 221 Fig. 3. Correlation diagram between estimated national average consumption of coffee and age-adjusted death rate for hyperplasia of prostate by 20 countries. States, Canada, Belgium, France, etc. might be guessed to belong to such countries. Consumption of coffee and sugar in relation to cancer mortality o f other organs Correlation with estimated national average consumption is tested for cancer mortality of the other organs of urogential and digestive system, by 20 countries (breast, uterus, esophagus, stomach, intestine and rectum) and by 13 countries (ovarium and bladder). No correlation is found between coffee consumption and cancer of those organs except prostate cancer, as given in Table I. Except for coffee, no international statistics of other beverages is found. How ever, estimated national average supply of sugar and syrups is reported in the Production Yearbook of the Food and Agricultural Organization4). Correlation coefficient between sugar supply and prostate cancer mortality is as high as that between coffee and prostate cancer. Although there is a significant correlation between coffee and sugar consumption (r=0.564, P<0.02), partial correlation coefficients are on the significance level of P<0.05 both for coffee and prostate cancer, and for sugar and prostate cancer. Correlation with sugar supply is also significant for breast-, ovarium-, intestine- and rectum-cancer. Reason why so is not clear. Although sugar consumption seems to be a barometer for economic level of nation, economic level itself should not be any disease agent. Rather food than economic level
222 E. Takahashi TABLE I. Correlation Coefficient between Estimated National Average Consumption of Coffee and Sugar and Age-adjusted Death Rate from Cancer of Urogential and Digestive Organs should be more directly related to probability of those cancer invasion. According to the result of analysis by correlationship among food stuff groups in the national average supplies, which has been calculated from the data of 22 countries in the Production Yearbook, national average supplies of sugar and syrups have positive significant correlation with meat and milk, and negative correlation with cereals5). DISCUSSION On the carcinogenic agent of coffee, there are analytical works of Kuratsune and Hueper6) in the National Cancer Institute in Bethesda. Four different commercial brands of roasted coffee were analyzed by them for their content of polycyclic aromatic hydrocarbons. Nine hydrocarbons including benzo pyrene, as shown in Table II, were identified in the darkest brand (D), though the amount of these hydrocarbons was small. In two other brands of moderately dark roasted coffee (A and B), pyrene and flouranthene were identified, but no other hydrocarbons. Another brand of darkly roasted coffee beans (C) did not show any appreciable amount of hydrocarbons. It seems to be possible that there should be some association between large amount coffee drinking and higher incidence of prostate cancer from the point of view of nation's unit, though the reason why the agents in coffee affect prostate by choice is not clear. It seems to be true that prostate cancer correlates not only with coffee consumption but also with sugar consumption. However, notwithstanding mortality for breast, ovarium, intestine and rectum cancer correlate with sugar consumption, they do not correlate with coffee consumption. From this point
Coffee Consumption and Prostate Cancer 223 TABLE U. Polycylic Aromatic Hydrocarbons in Coffee (Đg. per 1,000 gm. coffee) After Kuratsune and Huepere6 of veiw the correlation between coffee consumption and prostate cancer mortality should not be neglected. On the real meaning of coffee to prostate cancer, the roasting procedure should be important. It might be supposed in the period, before long after the World War U, when almost all not-producing countries lacked in coffee, in some countries such as Scandinavian where population had had a habit to consume much coffee, also ill-roasted (coarse) coffee might be consumed. To confirm the actual meaning of the effect of coffee to prostate cancer, still it might be needed some statistical study on individual cases by prostate cancer patient and by the control. However, in Japan it seems to be difficult to do such study, because patients of prostate cancer are small in number and habitual coffee drinkers are rare. References 1) United Nations, Yearbook of International Trade Statistics, 1955-1959. 2) Segi, M. & Kurihara, M., Cancer mortality for selected sites in 24 countries No. 2, 1962. 3) WHO, Ann. epidem Vit. Stat. (1959), 1962, 438. 4) FAO, Production Yearbook, 1960. 5) Takahashi, E., Tohoku J. exp. Med., 1962, 77, 239. 6) Kuratsune, M. & Heuper, W.C., J. Nat. Cancer. Inst., 1960, 24, 463.