Smoking and Drinking Habits among the JPHC Study Participants at Baseline Survey

Smoking and drinking habits at baseline survey among 110,896 male and female residents aged 40-69 are reported as part of an ongoing prospective population-based cohort in 11 geographically diverse health centers in Japan. The age-adjusted proportion of current and former smokers was 54.8% and 21.8% in males and 8.3% and 2.1 % in females, respectively. Mean age at initiation of smoking in males and females was 20.6 and 27.8 years of age, respectively. In males, the age-adjusted proportion of those who drink almost daily was 49.1 % and that of those who drink almost never was 19.7%, while in females, it was 5.9% and 71.6%, respectively. When compared by health center, the proportion of male current smokers was lower in Miyako and Ishikawa, both located in Okinawa Prefecture, while in females the proportion was higher in urban areas, such as Katsushika and Suita. The proportion of those who drink almost daily in males was higher in Yokote, Kashiwazaki, Katsushika and Suita, and lower in Ishikawa and Miyako, but in Ishikawa and Miyako, the proportion of those who drink at social events was higher. In females, the proportion of those who drink almost daily was higher in urban areas. There was substantial variation in the types of alcohol beverages consumed by males. In contrast, alcohol consumption in females comprised mainly beer. J Epidemiol, 2001 ; 11 (Suppl) : S44-S56.


Study population:
The study population, the study design, and other related details have been reported elsewhere in this journal .Table 1 shows the sex and age distribution of the study population .The study population in this analysis was 110 ,896 male and female respondents in Cohort I and II combined .In two districts, Katsushika and Suital, only adults aged from 40 to 50 years 1 Cancer Information and Epidemiology Division, National Cancer Center Research Institute .
were asked to participate, limiting the generalizability for these two districts.1.In all public health centers, the proportion was less than one percent except for females in Kashiwazaki.The subjects with an "unknown" smoking history were excluded from further tabulation. Drinking: The questionnaires used in Cohort I and II differed to a larger extent with respect to the questions on alcohol.Basically, in both Cohort I and II, the frequency of alcohol intake was asked first, followed by the average amount of daily alcohol intake by type of alcohol.In Cohort I, current frequency of any alcohol intake was asked first using 6 categories; "almost never," "1 -3 times per month ," "1-2 times per week," "3-4 times per week," "5-6 times per week" and "daily."For those who consumed alcohol at least once a week, the average amount of intake per day of 5 types of alcohol beverages (Japanese Sake, Shochu/Awamori (spirits), beer, whiskey, and others) was queried.In Cohort II, current drinking status was asked using 3 categories: "No", "Stopped already" and "Yes."For those who responded "Stopped already" or "Yes," the frequency of regular alcohol intake was asked using 4 categories: "1-3 times per month," "1-2 times per week," "3-4 times per week," "almost daily".In addition, the average amount of daily intake was asked of all current and former drinkers for 8 types of alcohol beverages (Japanese Sake , Shochu (spirits), Awamori (spirits), beer (large bottle), beer (medium bottle) , beer (small bottle), wine, and whiskey/brandy/vodka) .In addition, the frequency of drinking at social events was asked as the average number of days per month, and average amount of intake per day was asked for 8 types of alcohol beverages in the same way as usual drinking.Those who did not respond completely on these alcoholrelated questions were defined as "unknown" for drinking habits.The proportion of unknown cases for drinking habits is also shown in Table 1.The proportions tended to be higher in the areas of Cohort II than in Cohort I , especially for males in Miyako, Arikawa and Kashiwazaki.The unknown cases were excluded from further tabulation.
In order to compare the frequency of alcohol intake among the 11 public health centers simultaneously, the category "5-6 times per week" and "daily" were combined as "almost daily" in Cohort I and those who responded "No" or "Stopped already" were combined as "almost never" in Cohort II.For these comparisons, data from drinking at social events were ignored in Cohort II, because subjects were not asked these questions in Cohort I.In order to compare the frequency of intake by type of alcohol beverages, the percent by type of alcohol beverage was calculated only for those who drink any type of alcohol at least 1-2 times per week or more.The data from drinking at social events was not used in these calculations.
Average ethanol intake per week was estimated by multiplying the frequency of intake, by the amount of ethanol for a specific type of alcohol beverage.Answers of "1-3 times per month," "1-2 times per week," "3-4 times per week," "almost daily" were set to be 0.5, 1.5, 3.5, 6.0 times per week, respectively.Amount of ethanol in "Japanese Sake," "Shochu/Awamori" (spirits), "beer," "wine" and "whiskey/brandy/vodka" was set to be 23g/180m1, 36g/180m1, 23g/760ml, 3g/lglass and 10g/lglass, respectively.Average amount of ethanol intake was calculated for those who drink at least 1-2 times per week or more.The analysis was conducted for Cohort I and Cohort II separately, because data from drinking at social events was available only for Cohort II.

Statistical Analysis:
All calculations were based on sex and 10-year age group .For the comparison of 11 Public Health Centers , age adjusted proportions were calculated by two age groups (40-49 and 50-59) using the 1985 Japanese population as the standard population.Ratios of these proportions were computed using the proportion for the 11 heath centers combined as the denominator.

Smoking:
Table 2 shows the proportions of current and former smokers by sex, age and public health center.The age-adjusted proportion of current and former smokers was 54.8% and 21.8% in males and 8.3% and 2.1% in females, respectively, for the 11 public health centers combined, In both males and females, the proportion of current smokers was lower in older age group.The proportion of former smokers was higher in the older age group in males, but there was no remarkable difference in females.
In males, the proportion of current smokers was lower in Miyako and Ishikawa; both located in Okinawa Prefecture.The highest proportion of current smokers was observed in Kasama, but the differences among the health centers were not remarkable except for Miyako and Ishikawa.The proportion of former smokers was low in Ninohe, but the proportion of current smokers was high for this public health center.The proportion of former smokers tended to be slightly lower in areas of Cohort II than Cohort I, which may reflect a systematic bias as a result of using a different questionnaire.
In females, there was more variability in the proportion of current smokers by public health center.The proportion of current smokers was higher in urban areas, such as Katsushika and Suita.In the rural areas of Kashiwazaki, Miyako , Arikawa, Yokote and Ninohe, the proportion was low .In Saku, Ishikawa, Kasama and Tosayamada, which are a mixture of rural and suburban areas, the proportion was relatively high .As with males, the proportion of former smokers tended to be slightly lower in areas of Cohort II than Cohort I .
Table 3 shows the mean age at initiation of smoking and average number of cigarettes smoked per day among current smokers.Approximately, 1.0% of male and 3 .5% of female smokers did not provide proper information on the above items and they were excluded from the analysis .
The average age at initiation of smoking in males and females was 20.6 and 27.8 years of age , respectively.In males, there was little variation on the mean age at initiation of smoking by cohort age group , while in females the average age at initiation of smoking was higher for the older age group .There was no remarkable difference in the average age at initiation of smoking in males across public health centers , although it was slightly higher in Miyako .In females, however, the mean age at starting to smoke was lower in the two urban areas of Katsushika and Suita.
The average number of cigarettes smoked per day tended to decrease in the older age group in males , while there was little difference in females.Across public health centers , the average number smoked in males was higher in Suita and lower in Ninohe and Yokote .In females, the average number was higher in Arikawa and Miyako and lower in Nihohe and Saku , In both males and females, the average number of cigarette Drinking: Table 4a shows the distribution of alcohol consumption in males by age and health center.Overall, the age-adjusted proportion of those who drink almost daily was 49.1% and of those who drink almost none was 19.7% for males.In the older age group, the proportion of those who almost never drink was higher than of those who drink occasionally, while that of those who drink almost daily did not change remarkably across age groups.
For males, the proportion of those who drink almost daily was higher in Yokote, Kashiwazaki, Katsushika and Suita, and lower in Ishikawa and Miyako.The proportion of those who drink almost never was higher in Ishikawa and Arikawa and lower in Miyako, Kashiwazaki, and Yokote.In Ishikawa and Miyako, the proportion of those who drink occasionally was higher, while it was lower in Yokote and Kashiwazaki.
Table 4b shows the distribution of alcohol consumption in females by age and public health center.Overall, the ageadjusted proportion of those who drink almost daily was 5.9%, while that of those who drink almost never was 71.6%.In the older age group, the proportion of those who drink almost never was higher, while that of those who drink occasionally was lower and that of those who drink almost daily was lower than the younger ages.
In females, the proportion of those who drink almost daily was higher in Katsushika and Suita, and lower in Kashiwazaki, Ishikawa, Arikawa, Nihohe and Miyako.In contrast, the proportion of those who almost never drink was higher in Ishikawa and Miyako and lower in Katsushika and Suita.The proportion of those who drink occasionally was higher in Katsushika, Suita, and Tosayamada.

*
Study population consisted of only those aged 40 or 50 exactly in Katsushika and Suital ** Age adjusted proportion by Japanese standard population (1985) using the data for age group 40-49 and 50-59 *** Ratio of the age-adjusted proportions setting that for 11 heath centers combined to be the denominator Table 5a shows alcohol consumption by beverage type in males who drink any type of alcohol at least 1-2 times per week.People in the older age group tended to drink Japanese traditional alcohol beverages such as Japanese Sake, Shochu or Awamori more frequently, while those in the younger age group tended to drink westernized beverages such as beer, whiskey, brandy or vodka more frequently.
Alcohol consumption varied by public health center in males but not in females.In males, beer was the most popular type of alcohol beverage in Suita, Katsushika, Arikawa, Kasama, Ishikawa, while Japanese Sake was the most popular in Kashiwazaki, Yokote, and Saku.In Tosayamada, both beer and Japanese Sake were consumed frequently, while in Ninohe, both were infrequently consumed.
In Miyako, Awamori was preferred to beer, while in Ishikawa, Awamori was consumed as frequently as beer.In both Miyako and Ishikawa, Japanese Sake was seldom consumed.In Arikawa, those in younger age groups preferred beer, while those in older age groups preferred Shochu.Whiskey, brandy and vodka were more frequently consumed in Suita.Table 5b shows the proportion of those who drink at least once a week in females according to the types of alcohol beverages by age and health center.In females, beer was the most popular alcohol beverage in all health centers.The proportion of those who drink beer at least once a week was higher in Suita, Tosayamada, Miyako, Katsushika and Ishikawa and lower in Saku.Japanese Sake ranked second in Saku, Tosayamada, Kasama, Nihohe, Kashiwazaki and Yokote.In Ishikawa and Arikawa, whiskey/brandy/vodka were second, while in Miyako, Awamori was second.
Table 6a shows the estimated amount of ethanol intake among those who drink at least once a week in the areas of Cohort 1, by sex, age and health center.The age-adjusted average amount of ethanol intake was 301g/week in males and proportion of current smokers in males did not vary remarkably by area.In females, however, the proportion of current smokers in Hokkaido was higher than 20%, which was 2 times higher than the other geographic areas.Since this study did not cover Hokkaido, this would be one explanation for the discrepancies in estimates.

Drinking:
The National Nutrition Survey also reports on a frequency of alcohol consumption of 3 days or more per week and 1 "go" (180m1 of Japanese Sake or equivalent amount of alcohol) or more per day 2).In Table 8  Second, the amount of ethanol intake may be underreported * Study population consisted of only those aged 40 or 50 exactly in Katsushika and Suital ** Age adjusted proportion by Japanese standard population (1985) using the data forage group 40 -49 and 50-59 *** Ratio of the age -adjusted proportions setting that for 11 heath centers combined to be the denominator in Cohort I, because social drinking was not specifically queried as in Cohort II.This systematic underreporting would be of particular concern in geographic areas where social drinking constitutes a large proportion of ethanol intake as may exist in Ishikawa.For most previous studies, however, drinking at social events has seldom been asked.There is a possibility that people do not clearly differentiate between regular drinking or drinking at social events and might report the frequency of both types of drinking combined.This may result in the amount of intake being overestimated in Cohort II.
Because the responses in regular drinking might be affected by the different structures of the questionnaires , comparison of alcohol intake in Cohort I and II was limited to the frequency of alcohol intake.Estimated amount of alcohol intake was compared separately.This problem should be further investigated by other dietary survey approaches, such as dietary records for selected subjects .Third, ethanol intake may be overestimated in Miyako and Ishikawa, because of possible confusion whether alcohol intake should be reported by the volume of spirits with or with -  In females, the proportion of current smokers appeared to be highly dependent on the extent of urbanization.There was a five-fold difference on the proportion of current smokers between Katsushika (20%) and Miyako (4%).These findings are consistent with the geographical variation of lung cancer mortality, which shows higher rates in urban areas.Again, the exception is Okinawa Prefecture, where lung cancer mortality in females is also the highest in Japan while the smoking frequency is low.

Drinking:
In males, the proportion of those who drink daily was higher in Yokote and Kashiwazaki, both located in areas where production of Japanese Sake is prospering.In urban areas, such as Katsushika and Suita, the proportion of daily drinkers is higher as well.In Ishikawa and Miyako, people tend to drink less frequently, such as 1-2 times per week.This may be because in Yokote and Kashiwazaki, people prefer to drink at home with dinner, while in Ishikawa and Miyako, people prefer to drink outside of the home.The average number of days of social drinking among those who drink at least once per week in males was 6.2 in Miyako, while it was 2.8 in Kashiwazaki.
Since the amount of alcohol intake is usually higher when drinking together, the total amount of alcohol intake was higher in Miyako.In terms of total alcohol intake, for Miyako 33.5% was from drinking at social events, while for Kashiwazaki, the figure was only 18.6%.However, the estimated amount of alcohol intake may be overestimated in Miyako and Ishikawa.
In females, the proportion of daily drinkers was highly dependent on the extent of urbanization, similar to smoking habits.In Katsushika and Suita, 12% to 15% of women were daily drinkers, while in Kashiwazaki, Ishikawa, Arikawa and Nihohe, it was less than 3%.In Tosayamada, the proportion of daily drinkers combined with the occasional drinkers was relatively higher than the other remaining areas.This may be due to the fact that the local social norm in Tosayamada has not been so strict against women drinking.
There was a great variation on the types of alcohol beverages frequently consumed in males, while there was little variation in females.In Yokote, Saku, and Kashiwazaki, which are located in a relatively cold climate and also famous for the production of Japanese Sake, people like to drink Japanese Sake, especially among the older age group.In Katsushika and Suita, both belonging to urban areas, or Ishikawa, Tosayamada, Arikawa and Miyako, which are located in warm or moderate climates, beer was the most popular alcoholic beverage.In Miyako, where the production of Awamori, a traditional distilled liquor made from rice, is prosperous, people like to drink Awamori most, while in Ishikawa, people like to drink both Awamori and beer.Shochu, which is also a type of a tradition- , the proportion of habitual drinkers observed in the National Nutrition Survey in 1990-93 is shown alongside the estimates from our study.During 1990-93, the proportion declined slightly in males, but remained fairly constant in females.Compared to the average proportion of drinkers in 1990-93 observed in the National Nutrition Survey, Cohort I showed no difference in males and a lower proportion Smoking and Drinking Habits in JPHC Baseline S -51 No" to question B and responded to at least one of the following items: age at initiation of smoking, average number of cigarettes smoked per day, and age at cessation of smoking."Current smoker" was defined those who responded "Yes" to question B. All other combinations of responses including missing responses were defined as "unknown".The proportion of unknown cases for smoking habits is shown in Table

Table 1 .
Sex and age distribution of study subjects and proportion of unknown cases on smoking and drinking habits.
* Study population consisted of only those aged 40 or 50 exactly in Katsushika and Suital

Table 3 .
Mean age at initiation of smoking and average number of cigarettes smoked perday by sex, age and health center.

Table 4b .
Distribution for frequency of alcohol intake by age and health center, female.
tion was slightly higher in males except for the age group of 60-69 years.In females, the estimated proportion of current smokers from our study was slightly lower for all age groups, compared to the National Nutrition Survey, while the estimate of former smokers was slightly higher except for the age group of 60-69 years.According to the Japan Tobacco Survey, the

Table 5a .
Proportion of people who drink at least once per week according to kind of alcohol beverage by age and heath center , male.Study population consisted of only those aged 40 or 50 exactly in Katsushika and Suital ** Age adjusted proportion by Japanese standard population (1985) using the data forage group 40-49 and 50-59 *** Ratio of the age-adjusted proportions setting that for 11 heath centers combined to be the denominator *

Table 5b .
Proportion of people who drink at least once per week according to kind of alcohol beverage by age and heath center .female.

Table 6a .
Average amount of alcohol intake (ethanol g/week) by sex , age and health center in Cohort T .Study population consisted of only those aged 40 or 50 exactly in Katsushika ** Age adjusted proportion by Japanese standard population (1985) using the data for age group 40-49 and 50-59 *** Ratio of the age-adjusted proportions setting that for 5 heath centers combined to be the denominator *

Table 6b .
Average amount of alcohol intake (ethanol g/week) by sex, age and health center in health centers.In contrast to the lower smoking rates, lung can-

Table 6c .
Proportion of alcohol intake (ethanol g/week) derived from drinking at social events by sex, age and health center in Cohort II.Ratio of the age-adjusted proportions setting that for 5 heath centers combined to be the denominator

Table 7 .
Proportion of current and former smoker from other nation-wide surveys.
* Age adjusted proportion by Japanese standard population (1985) using the data forage group 40-49 and 50-59 ** Ratio of the age-adjusted proportions setting that for average proporion of National Nutrition Survey from 1990 to 1993 to be the denominator

Table 8 .
Pronortion of habitual drinker* from other nation-wide surveys.Habitual drinker was defined as those who drink alcohol 3 days or more per week and I go or more.perday ** Age adjusted proportion by Japanese standard population (1985) using the data for age group 40-49 and 50-59 *** Ratio of the age-adjusted proportions setting that for average proporion of National Nutrition Survey from 1990 to 1993 to be the denominator *cer mortality in the Okinawa Prefecture is the highest in Japan in both males and females.These findings seem to be contradictory, since cigarette smoking is the most influential risk factor for lung cancer.In terms of the production and sales of cigarettes, however, Okinawa Prefecture has a special history, different from other prefectures in Japan.During 1945-72, while al distilled liquor, but mainly made from wheat, sweet potatoes