Association of Vegetable Intake with Urinary 6-Sulfatoxymelatonin Level

Epidemiologic evidence has suggested that a high consumption of vegetables is associated with reduced risks of cancer and cardiovascular disease (1). The hypothesis that the antioxidant activity of vegetables contributes to their beneficial effects has been advanced (2). Several mechanisms, including vitamins, carotenoids, and flavonoids, are probably involved in these processes. Beyond these recognized constituents of vegetables, certain other substances may confer additional health benefits. In recent years, melatonin has been discovered in edible plants. To date, only a small number of plants have been tested for melatonin (3), but the results of the studies investigating the melatonin contents in plants suggest that melatonin is present in a wide number of plants (4), including commonly consumed vegetables, such as cabbages and white radish sprouts, which are with relatively high amounts of melatonin (3). Normally, melatonin production occurs during the dark phase of the day. The distinct diurnal variation has suggested the possibility of a regulatory function of melatonin in day/night-dependent physiologic processes, such as sleep (5). Recently, laboratory studies have also shown the antiproliferative, antioxidative, and immunostimulatory actions of melatonin (6). The protective effects of vegetable intake against cancer and cardiovascular disease may be partially attributable to an increase in melatonin intake resulting from the consumption of vegetables. As a first step to test this hypothesis, we evaluated whether vegetable intake is associated with an increased urinary melatonin in generally healthy women.

Study subjects were recruited among participants of a breast cancer screening at a general hospital in Gifu, Japan, between June and December, 2000. The hospital has been conducting mass screening campaign for breast cancer since the early 1980s. Municipal letters of invitation to the screening are mailed to women residing in its surroundings. A total of 432 women who were free of breast cancer participated in the present study (response rate was 68.5%). Informed consent was obtained from each woman. The study was approved by the institutional review board.

Each woman responded to a self-administered questionnaire asking basic demographic characteristics, diet, exercise, smoking and drinking habits, and medical and reproductive histories. A nurse epidemiologist visited participants and collected first-void morning urines on the next morning. The blood and urine samples were frozen and stored at −80°C until assayed.

Intakes of vegetables and other foods and nutrients were estimated using a 169-item semiquantitative food-frequency questionnaire. Detailed information on the questionnaire, including its validity and reproducibility, has been described elsewhere (7, 8). For example, the Spearman correlation coefficients between this questionnaire and 12 daily diet records kept over a 1-year period for total energy, fat, protein, carbohydrates, and vegetables were 0.53, 0.52, 0.63, 0.53, and 0.41, respectively.

Urinary 6-sulfatoxymelatonin (aMT6-s) was measured radioimmunologically using kits purchased from IBL Laboratories (Hamburg, Germany). The sensitivity and the intra-assay and interassay coefficients of variation were 1.7 ng/mL, 9.8%, and 15.3%, respectively. The value of assay sensitivity was assigned for a woman (n = 1) who had an undetectable level. To adjust for variation in the diluteness of urine, urinary aMT6-s levels were expressed as urine aMT6-s/urine creatinine.

We excluded 31 women from analyses because of incomplete or unreliable responses to the dietary questionnaire (criteria shown in ref. 7). Women who reported having cancer (n = 6) or heart disease (n = 10) were excluded. As exogenous estrogen may suppress melatonin level (9), women using hormone replacement therapy (n = 6) or contraceptive pills (n = 1) were excluded from the study. Because it was not possible to obtain information on the use of diuretics and β-blockers, which can affect urinary aMT6-s levels (10), we further excluded 55 women with the diagnosis of hypertension. The urine samples from 34 women were insufficient for the measurement. Hence, the remaining 289 women were the focus of this report.

Urinary aMT6-s level was transformed into logarithmic values for statistical analysis. Intakes of vegetables and other foods and nutrients were adjusted for total energy after log transformation by using the residual method proposed by Willett (11). The relationship between vegetable intake and urinary aMT6-s level was assessed by linear regression models. Geometric means of urinary aMT6-s levels according to the quartile of vegetable intake were provided using analysis of covariance models. Potential confounders, such as age, body mass index, alcohol intake, menopausal status, and day length (the number of hours of daylight between dawn and dusk), of the day before urine collection were included into models as covariates.

The characteristics of the study subjects were shown in Table 1.

Table 2 shows the geometric means of urinary aMT6-s according to the quartile of vegetable intake. The mean urinary aMT6-s was 15.9% higher in women with the highest quartile of vegetable intake than it was in those with the lowest quartile of intake after controlling for age, total energy, body mass index, smoking status, alcohol intake, menopausal status, and day length. There was a significant trend between urinary aMT6-s and intake of vegetables after controlling for the covariates. A similar tendency was observed for the association of aMT6-s with green and yellow vegetables as well as other vegetables, although the association with other vegetables was of borderline significance.

Table 3 shows the association of fruit intake with urinary aMT6-s. Fruit intake was positively associated with urinary aMT6-s but this association was not statistically significant. There was no significant association of urinary aMT6 with total energy and other food groups and nutrients, such as grains, potatoes, meats, fishes, dairy foods, protein, fat, and carbohydrate. Carotene, dietary fiber, and vitamins, which are abundant in vegetables, were marginally significantly associated with urinary aMT6-s (data not shown).

As exposure to light at night may affect the urinary aMT6-s levels (12), an additional survey was conducted 3 years after to obtain information on sleeping habits around the time when the urine sampling had been conducted. Out of the 289 women, 204 responded to the second survey. Further adjustment for the frequency of being awake around 1:00 and 2:00 a.m. (the approximate time of the melatonin peak) did not substantially alter the results; the mean urinary aMT6-s was 20.7% higher in women with the highest quartile of vegetable intake than it was in those with the lowest quartile of intake after controlling for the covariates (P_trend = 0.05).

When we analyzed data separately for premenopausal and postmenopausal women, the association of vegetable intake with urinary aMT6-s level was weak in postmenopausal women; mean aMT6-s levels in the lowest and the highest quartiles of vegetable intake was 32.8 and 37.3 ng/mg creatinine, respectively, in postmenopausal women. The corresponding values for premenopausal women were 31.9 and 41.0 ng/mg creatinine, respectively.

To our knowledge, this is the first report on vegetable intake and melatonin levels in humans. Vegetable intake was moderately but significantly associated with urinary aMT6-s level. We speculate that this association may be explained by the melatonin contained in vegetables. Hattori et al. (3) reported that feeding chicks a diet containing plant products rich in melatonin increased blood melatonin levels.

Urinary specimens were not collected over a long period (12 or 24 hours). However, the validity of the use of the first-void morning urine has been previously reported. Urinary aMT6-s level in morning urine is strongly correlated with total nocturnal plasma melatonin output and peal nocturnal melatonin value (13). Sufficient reproducibility of measurement of aMT6-s in morning urine over 5 years (intraclass correlation coefficient = 0.58; ref. 14) has been also reported.

The food-frequency questionnaire, like all methods of dietary assessment, is subject to measurement error. Our questionnaire was designed to measure an individual's relative intakes of foods and nutrients rather than absolute values. The data presented for vegetables may have been overestimated because vegetable intake estimated from the questionnaire was 46% higher than that estimated from the 12 daily diet records. Our questionnaire included 13 items for specific vegetables, such as tomato, pumpkin, spinach, Japanese radish, cabbage, carrot, etc. Besides these food items, we took into account some dishes that include vegetables as ingredients, which may have yielded higher intake compared with the diet records. However, it is likely that this measurement error was unrelated to urinary aMT6-s levels and led to an underestimation of the true associations.

As data for melatonin concentration are available only for a few plants, it is not possible to estimate the dietary melatonin intake from vegetables or from the entire diet. Hattori et al. (3) measured the melatonin concentrations in plants commonly consumed as vegetables among the Japanese. In their study, melatonin concentrations ranged from 24.6 pg/g tissue for cucumber to 657.2 pg/g tissue for white radish sprouts. Serum melatonin levels in normal humans are very low during most of the day but increase significantly to a mean of 80 pg/mL (range, 0-200) between 2:00 and 4:00 a.m., and remain elevated during the normal hours of sleep, falling sharply to daytime values around 9:00 a.m. (15). Even 10 μg of melatonin infusion raises serum melatonin concentration ∼40 times more at 5 minutes after administration (from 12 ± 5 to 487 ± 377 pg/mL; ref. 16). However, consumption of 400 g of white radish sprouts should provide only 0.3 μg of melatonin. We cannot rule out the possibility that melatonin in vegetables may not be sufficient to affect blood melatonin or urinary aMT6-s levels. In such a case, vegetable intake may be merely a correlate of certain factors, which are associated with urinary aMT6-s.

Cagnacci et al. (17) suggested that the effects of melatonin on some biological functions, such as hypothermic response, are reduced in aged women, which may partially explain the observed weak association of urinary aMT6-s with vegetable intake in postmenopausal women.

Thus far, melatonin has been identified in some grains, nuts, and fruits, but we did not observe significant associations of urinary aMT6-s with intakes of these food groups. In these food groups, the number of foods that contain melatonin may be very limited or the concentration of melatonin may vary greatly in different foods belonging to the same food group. Folate deficiency decreases melatonin secretion in rats (18). Although folate intake itself was nonsignificantly associated with urinary aMT6-s levels (P = 0.10), it is possible that the association of vegetable intake with urinary aMT6-s may be attributable to melatonin together with folate in vegetables. Fruits are also rich in folate. Like vegetables, high intake of fruits has been associated with reduced risks of cancer and cardiovascular disease (1). Our data did not deny the possibility that melatonin contained in fruits may be implicated in these associations.

Clearly, more extensive studies to determine the melatonin concentrations in a wider variety of vegetables, as well as other foods, will be necessary before our results are thoroughly understood. In addition, evidences for the beneficial effects of melatonin on cancer and other diseases is mainly based on laboratory data and must be confirmed in epidemiologic studies. Nonetheless, our findings might stimulate studies investigating the role of dietary melatonin in health.