Effects of Passive Smoking on Serum Levels of Carotenoids and α-Tocopherol

To evaluate the effects of passive smoking identified by urine cotinine on serum carotenoids and alpha-tocopherol, we categorized 124 residents in a rural city of Japan into 4 groups by their urine cotinine/creatinine (Cot/Cr) ratio (u.d. (undetectable); low (0-50 ng/mg); moderate (50-120 ng/mg); high (> 120 ng/mg)) and compared the serum carotenoids and alpha-tocopherol levels among these groups. We identified passive smoking by low Cot/Cr ratio level. After controlling on related factors, men with low Cot/Cr ratio showed significantly lower serum zeaxanthin/lutein levels than men with u.d. Cot/Cr ratio. The difference was still marginally significant after excluding self-reported current smokers from the low Cot/Cr group. It is suggested that low level exposure to tobacco smoke, which has been reported to be equivalent to that for passive smokers, could be associated with decreased serum zeaxanthin/lutein levels in men.

passive smoking, urine cotinine, carotenoids, * -tocopherol Tobacco smoke contains much oxidants and free radicals '). Previous epidemiologic studies demonstrated that active smokers have lower concentration of serum carotenoids and tocopherols than nonsmokers [2][3][4][5]. They suggested that serum carotenoids and tocopherols; so-called antioxidant elements would be consumed by reducing oxidants and free radicals6, 7,8). However, it is not clear whether serum levels of carotenoids and tocopherols are decreased in passive smokers. The urine cotinine is a metabolite of nicotine in tobacco smoke 9). It is well known that the levels of urine cotinine in passive smokers usually indicate intermediate values between nonsmokers and active smokers [10][11][12][13][14][15] In this study we compared serum carotenoids and a-tocopherol levels for 124 subjects by levels of urine cotinine.

MATERIALS AND METHODS
We randomly selected 600 persons from 36,990 residents who were 35 years old and over in a rural city of Japan in May 1992. We sent recruiting mails to them and 192 persons agreed to join our study.
In August 1992, dietitians and public health nurses visited the participants and instructed them to record food consumption for 3 days. The participants recorded gram or the number of the materials of their meals. Three days after the first visit, they visited the participants again and checked each miswriting and supplemented the missing answers by interview. Dietitians calculated nutrients intake for each subject based on Japan Food Composition Tables, 4th edition 11). Vitamin supplements were not included into the calculation.
The participants also provided blood samples when public health nurses visited for the dietary survey. Within 3 hours after the blood collection, the sera were separated and stored in brown tubes intercepting light at -801C for one year until the analyses. The concentrations of serum carotenoids (Qcarotene, cryptoxanthin, zeaxanthin/lutein) and a -tocopherol were measured by high-performance liquid chromatography (HPLC)'). Serum total cholesterol and triglycerides were mea-sured by a commercial laboratory using enzyme methods. Serum y*-GTP was also measured by the laboratory. We also collected their spot urine samples at the same time we collected blood samples. Their urine were stored at -8 0*for two years until the analyses. The urine cotinine were measured by a commercial laboratory using gas chromatography -mass spectrometry 18). The concentration of urine cotinine is shown as the ratio of cotinine to creatinine (Cot/Cr ratio (ng/mg)) in this study. Previous studies indicated that the Cot/Cr levels for passive smokers were below 50 ng/mg 14) or 120 ng/mg 15). We categorized the Cot/Cr ratio into the following four categories: u.d. = undetectable; low = 0-50 ng/mg; moderate = 50-120 ng/mg; and high = >120 ng/mg.
One month after our visit, we conducted a self-administered questionnaire survey for the participants. We collected information on their birthday, height and weight, past history of cancer and smoking status, period from the start of habitual smoking and number of cigarettes smoked per day using the questionnaire. Alcohol consumption per day in the last year and vitamin supplements use was also asked in the questionnaire. We converted alcohol consumption into amount of pure ethanol intake.
After excluding 68 persons with one or more missing values in their background information or with a history of cancer, we analyzed the data from 55 men (mean age = 58.1 years old; SD = 11.2) and 69 women (mean age = 54.4 years old; SD = 10.9). None of the subjects had 40 IU or over levels of * -GTP which is a possible indicator of impaired liver function. Seventy three percent of the men were employees or employers in companies and eighty seven percent of the women were housewives.
In statistical analysis, we calculated Spearman rank correlation coefficients between urine Cot/Cr ratio and serum levels of carotenoids and a -tocopherol. Analysis of covariance was also employed to compare the serum carotenoids and a -tocopherol levels by Cot/Cr category after common log-transformation. Other variables which we included into the model were age (years old), body mass index (BMI (kg/m2)), common logtransformed values of carotene, vitamin C and vitamin E intake, pure ethanol intake (categorized quintile), serum total cholesterol, common log-transformed value of serum triglycerides and vitamin supplements use (4 or more times per week for any kinds of vitamin supplements = 1; others -0). These variables were reported as influential factors on serum carotenoids and * -tocopherol in previous studies 2-5) . We carried out the analysis by sex, since quintile category of ethanol intakes of men were different from that of women, i.e., the amount of ethanol intakes of men were much larger than that of women. All of the analyses were performed by SAS Ver 6.0419).

RESULTS
No women indicated the Cot/Cr ratio of 120 ng/mg or over (Table 1). Men with moderate and high Cot/Cr ratio showed significantly lower mean values of serum /3 -carotene than men with u.d. Cot/Cr ratio. Men with high Cot/Cr ratio and women with moderate Cot/Cr ratio showed significantly lower levels of serum cryptoxanthin than subjects with u.d. Cot/Cr ratio.
Analysis of covariance showed that men with high Cot/Cr ratio had significantly lower serum concentrations of (3carotene, zeaxanthin/lutein and a -tocopherol than men with u.d. Cot/Cr ratio (Table 2). Moreover, men with low Cot/Cr ratio had significantly lower serum concentrations of zeaxanthin/lutein and than men with u.d. Cot/Cr ratio. Serum concentrations of carotenoids and a -tocopherol for women were not statistically different among the groups by Cot/Cr ratio.
Out of 29 men and 62 women with Cot/Cr ratio below 50 ng/mg, 4 men (14%) and one woman (2%) reported to be current smokers in the self-administered questionnaires (Table 3).
Even when we excluded the subjects who reported themselves as current smokers from the group with low Cot/Cr ratio, the group showed marginally lower concentration of serum zeaxanthin/lutein than men with u.d. Cot/Cr ratio (P = 0.06). No such findings were observed in other carotenoids and a -tocopherol.

DISCUSSION
We found possible effects of passive smoking on serum zeaxanthin /lutein in men. However, we couldn't find such effects of passive smoking on other carotenoids and a -tocopherol. Zeaxanthin/lutein is an oxygenated carotenoids. It is reported that it retards the lipid peroxidation more efficiently than *-carotene and is powerful antioxidant than icarotene 20). Our finding suggest that serum zeaxanthin/lutein might be consumed to reduce oxidants faster than other antioxidant vitamins, being consistent with the in vitro study. Tribble et al. reported that female passive smokers indicated lower plasma ascorbic acid concentration than nonsmokers 21). Ascorbic acid is one of water-soluble antioxidant and is known to be consumed by regenerating reduced a-tocopherol. Furthermore, * -tocopherol is known to protect * -carotene from peroxidation8). Ascorbic acid may be consumed faster by reducing oxidants and free radicals than carotenoids and a - tocopherol.
We could not observe any statistically significant differences in the concentrations of serum carotenoids and a -tocopherol by Cot/Cr category in women. We must also consider sex differences in antioxidants metabolism. Ito et al. reported girls aged 7-19 years old had higher serum concentration of * -carotene and cryptoxanthin than same age boys although the differences were not larger than subjects aged 20-69 years  Japan 22). Male nonsmokers most of whom are working in their office are probably exposed more to environmental tobacco smoke than female nonsmokers most of whom are housewives.
We attempted to evaluate the effects of passive smoking on serum carotenoids and tocopherols by using an index of urine Cot/Cr ratio, because self reported exposure to tobacco smoke might be incorrect due to response bias. Previous studies indicated that passive smokers had urine Cot/Cr ratio below 50 ng/mg'") or below 120 ng/mg'5. We used the former criteria to evaluate lower exposure level of tobacco smoke. Comparison of Cot/Cr ratio and the reported number of cigarettes smoked per day indicated that 78-97% of those with low Cot/Cr ratio (0-50 ng/mg) reported themselves as nonsmokers, supporting our identification of passive smokers. Even when we limited the analysis to those who reported themselves as nonsmokers, men with low Cot/Cr ratio had lower levels of serum zeaxanthin/lutein than those not exposed to tobacco smoke (i.e. u.d. Cot/Cr group), although, the difference was not statistically significant (P = 0.06). However, we need more validated criteria or other indicators to identify passive smokers.