European Journal of Clinical Nutrition (2015), 1–7 © 2015 Macmillan Publishers Limited All rights reserved 0954-3007/15 www.nature.com/ejcn

ORIGINAL ARTICLE

Fruit and vegetable consumption and hypertriglyceridemia: Korean National Health and Nutrition Examination Surveys (KNHANES) 2007–2009 C Yuan1,2, H-J Lee3, HJ Shin1,4, MJ Stampfer1,2,5 and E Cho5,6,7 BACKGROUND: Limited research has been conducted on the association between intake of fruits and vegetables and hypertriglyceridemia, especially in Asian populations. This study aimed to investigate the association between total fruit and vegetable intake, as well as subgroups of fruit and vegetable intake, with hypertriglyceridemia among Korean adults. METHODS: We conducted a cross-sectional study of 7934 adults aged 19–64 years from the fourth Korean Health and Nutrition Examination Survey. Fruit and vegetable intake was estimated from a food frequency questionnaire. Subgroups of fruits and vegetables included citrus, non-citrus and carotene-rich fruits and cruciferous, green leafy and carotene-rich vegetables. Hypertriglyceridemia (plasma triglyceride ⩾ 150 mg/dl) was diagnosed using a blood sample drawn after 12+ hours of fasting. RESULTS: There were 2001 (25.2%) cases of hypertriglyceridemia among the participants. Total fruit intake was significantly inversely associated with the prevalence of hypertriglyceridemia; the multivariate odds ratios (95% confidence intervals) of hypertriglyceridemia across increasing quintiles were 1.00 (ref), 0.76 (0.62, 0.92), 0.72 (0.58, 0.90), 0.68 (0.54, 0.85) and 0.64 (0.49, 0.82; Ptrend = 0.001) after controlling for survey year, body mass index, waist circumference, smoking, alcohol drinking, physical activity, education and income. Similar inverse associations were found for all fruit subgroups. However, we found no significant association between intakes of total or subgroups of vegetable and hypertriglyceridemia; the odds ratio for top vs bottom quintile was 1.00 (0.81–1.24) for total vegetable intake. CONCLUSIONS: Our findings support a potential beneficial role of fruit consumption to reduce blood triglyceride levels in Asian populations. European Journal of Clinical Nutrition advance online publication, 27 May 2015; doi:10.1038/ejcn.2015.77

INTRODUCTION Hypertriglyceridemia (elevated plasma triglyceride levels) is a strong risk factor for cardiovascular disease1 and is frequently accompanied by other lipid abnormalities, obesity and the metabolic syndrome, which are also associated with cardiovascular disease.2 Hypertriglyceridemia is highly prevalent across countries: for example, 31% in the US in 2008 3 and 33% in Korea in 2007,4 remaining as a major health burden worldwide. Fruit and vegetable intake has been regarded as an important factor for prevention of cardiovascular disease5 and hypertension,6 on the basis of observational studies across different countries.5,7–10 Some studies have also investigated the association between fruit and vegetable intake and hypertriglyceridemia. A Mediterranean-style dietary pattern (higher total fruit, vegetables, nuts and whole grains) was associated with decreased triglyceride levels.3,11 In addition, higher consumption of fruit or vegetable was associated with lower likelihood of hypertriglyceridemia in Latin America.12 Dietary fiber, mostly from fruit and vegetable, also showed an inverse association with serum triglyceride levels.13 It was also noted that beneficial effects of fruits and vegetables could not be simply replaced by dietary supplements, such as purified flavonoids.14 Studies in Asia have

linked fish oil, carbohydrates, dietary pattern and red grapefruit with hypertriglyceridemia or the metabolic syndrome.15–18 However, limited research on fruit and vegetable has been conducted in Asian countries.5 The prevalence of hypertriglyceridemia has gradually increased in Korea.4 Koreans are now consuming more ‘western’ food because of marked lifestyle changes by westernization.19 However, some Korean people still follow the traditional diet that includes relatively higher fruit and vegetable consumption.20 In this study, we aimed to investigate the association between total fruit and total vegetable consumption, including fruit and vegetable subgroups, and hypertriglyceridemia in a nationally representative sample of the Korean population. METHODS Study population We examined data from the Korean National Health and Nutrition Examination Surveys (KNHANES), a national cross-sectional survey of the Korean population, conducted in 2007–2009. Study details have been described previously.21 Participants were selected using a stratified, multistage sampling design based on geographical area, sex and age. The study included a health behavior interview, health examination and dietary

1 Department of Nutrition, Harvard T.H. Chan School of Public Health, Boston, MA, USA; 2Department of Epidemiology, Harvard T.H. Chan School of Public Health, Boston, MA, USA; 3Department of Food and Nutrition, Eulji University, Daejeon, Korea; 4Division of Cardiology, Department of Medicine, Baylor University Medical Center, Dallas, TX, USA; 5 Channing Division of Network Medicine, Brigham and Women's Hospitals, Harvard T.H. Chan School of Public Health, Boston, MA, USA; 6Department of Dermatology, Warren Alpert Medical School of Brown University, Providence, RI, USA and 7Department of Epidemiology, Brown School of Public Health, Providence, RI, USA. Correspondence: Dr E Cho, Channing Division of Network Medicine, Brigham and Women's Hospital, Harvard T.H. Chan School of Public Health, 181 Longwood Avenue, Boston, MA 02115, USA. E-mail: [email protected] Received 23 February 2014; revised 29 January 2015; accepted 17 March 2015

Fruit, vegetable intake and hypertriglyceridemia C Yuan et al

2 survey including a 24-h recall, dietary practice questionnaire and food frequency questionnaire (FFQ). The study was approved by the Institutional Review Board of the Korea Centers for Disease Control and Prevention (KCDC), and informed consent was obtained from all participants. The fourth survey included all seasons to avoid seasonal bias on dietary intake. Data used in our analysis included a total of 4594, 9744 and 10 533 individuals from the 2007, 2008 and 2009 KNHANES, respectively. Excluding subjects who were under 19 and over 65 years old (n = 10 537), pregnant (n = 123), with incomplete dietary survey data (n = 1924), with extreme energy intake of o500 or 46000 kcal/day (n = 103), with missing data on total fruit and vegetable intake frequency (n = 57), body mass index (BMI; n = 869), waist circumference (n = 15) or serum triglyceride (n = 259), or who did not fast at least 12 h before the blood draw (n = 3041) we included a total of 7934 adult participants aged 19–64 years in the final analysis.

Dietary assessment The frequency of fruit and vegetable consumption was estimated from the FFQ for the previous year’s intake. The responses in the FFQ included: almost never, 6–11 times/year, 1 time/month, 2–3 time/month, 1 time/ week, 2–3 times/week, 4–6 times/week, 1 time/day, 2 times/day and 3 times/day. There were 11 fruit items included in the FFQ: tangerines, persimmons, pears, watermelons, oriental melons, strawberries, grapes, peaches, apples, bananas and oranges/other citrus fruits. A total of 11 vegetable items were included in the FFQ: Korean cabbages, radishes, radish leaves, tomatoes, cabbages, soy bean sprouts, spinach, cucumbers, carrots, peppers and pumpkins. The total frequency of intake was calculated for each fruit and vegetable by summing their daily frequencies. Subgroups of fruits and vegetables included citrus fruits, non-citrus fruits, carotene-rich fruits, cruciferous vegetables, green leafy vegetables and carotene-rich vegetables. Individual item included in each subgroup is listed in Table 1. Intake of total calorie, carbohydrate, fat, protein and fiber was calculated using 24-h recall data.

Documentation of hypertriglyceridemia Participants had their blood samples drawn during their health examination after at least 12 h of fasting. The serum level of triglycerides was measured enzymatically using a Hitachi automatic analyzer 7600 (Hitachi, Tokyo, Japan). Hypertriglyceridemia was defined as fasting triglyceride levels ⩾ 150 mg/dl according to the American Diabetic Association and the National Cholesterol Education Program Adult Treatment Panel III guidelines.22

Assessment of other covariates Information on covariates, including age, gender, alcohol drinking, smoking, education, annual income, physical activity and survey year, were collected from health behavior interviews and dietary assessments. Height, weight and waist circumference were measured and recorded by the medical team. BMI was calculated for each participant as weight (kg)/height squared (m2).

Statistical analysis Survey analysis was used in studying the association between fruit and vegetable intake and hypertriglyceridemia. Because of the complex sampling design of the KNHANES study, sampling weights were used.

Table 1.

General subject characteristics were presented as means or as prevalence after age and sex standardization to the distribution of the Korean population. We also conducted logistic regressions to investigate the association between fruit and vegetable intake and hypertriglyceridemia. Fruit and vegetable intake frequencies were divided into quintiles. Hypertriglyceridemia was dichotomized at the cutoff point of 150 ml/dl of fasting triglyceride (yes vs no). For total vegetable and fruit intake and vegetable and fruit subgroups, we conducted analyses adjusted for age (continuous) and sex and then further controlled for potential confounding or risk factors of hypertriglyceridemia: survey year (2007/2008/2009), BMI (continuous, kg/m2), smoking (never smoker, o1 pack/day, 1–2 packs/day and ⩾ 2 packs/ day), alcohol drinking (continuous), physical activity (no exercise or sometimes walk, regularly walk, regular moderate-level activity and regular vigorous-level activity), education (elementary school, middle school, high school, college or higher degree), family income (quartiles) and waist circumference (continuous). Utilizing the 24-h recall data, we additionally adjusted for intakes total calorie, fiber and carbohydrate, fat and protein intake (original and energy adjusted residual) one at a time in a sensitivity analysis. We further conducted analyses mutually adjusting for fruits and vegetables and combining fruits and vegetables. Tests for linear trends across categories of fruit or vegetable intake were conducted by treating the median of each category as a continuous variable. In addition, we treated fruit or vegetable intake as continuous variables and estimated multivariate-adjusted odds ratios (ORs) for each five times/week increase in fruit or vegetable intake frequency. We also assessed effect modification by gender, age (omedian age, ⩾ median age), BMI ( o23 kg/m2, ⩾ 23 kg/ m2) and smoking status (never, current or ever smoker). All analyses were conducted using SAS software, version 9.2 (SAS Institute, Cary, NC, USA).

RESULTS There were 2001 (25.2%) cases of hypertriglyceridemia among the included participants of KNHANES in 2007–2009. The age and sexstandardized characteristics of the study participants for the total population and by quintiles of vegetable and fruit intake are presented in Table 2. The average frequencies of total fruit intake and total vegetable intake were 10 times/week and 32 times/ week, respectively. Fruit and vegetable intake was positively correlated (Pearson's correlation coefficient = 0.25, P o0.001). Participants with higher vegetable intake tended to be older, male, non-smokers and more engaged in vigorous exercise and tended to have higher BMI, alcohol intake, fruit intake and household income. Individuals with higher fruits intake were more likely to be younger, female, non-smokers and tended to have lower alcohol intake and higher vegetable intake, education, physical activity and household income. Total fruit intake was significantly inversely associated with the prevalence of hypertriglyceridemia in both age- and genderadjusted and multivariate analyses (Table 3 and Supplementary Figure 1). The multivariate-adjusted ORs (95% confidence intervals) of hypertriglyceridemia across increasing quintiles of total fruit intake were 1.00 (ref), 0.76 (0.62, 0.92), 0.72 (0.58, 0.90), 0.68 (0.54, 0.85) and 0.64 (0.49, 0.82), respectively (Ptrend = 0.001). The multivariate-adjusted OR (95% CI) for each five times/week increase in fruit intake frequency was 0.95 (0.89–1.00). Similar

Food items in fruit and vegetable subgroups

Subgroup Vegetable subgroup Cruciferous vegetables Green leafy vegetables Carotene-rich vegetables Fruit subgroup Citrus fruits Non-citrus fruits Carotene-rich fruits

European Journal of Clinical Nutrition (2015), 1 – 7

Food items Korean cabbages, cabbages, radish and radish leaves Spinach, radish leaves Carrots, spinach, pumpkins and tomatoes Tangerines and oranges/other citrus fruits Persimmon, watermelon, strawberry, grape, pear, oriental melon, peach, apple and banana Tangerines, oranges/other citrus fruits, watermelon and persimmon

© 2015 Macmillan Publishers Limited

Fruit, vegetable intake and hypertriglyceridemia C Yuan et al

3 Table 2. Age- and sex-standardized characteristics of subjects according to quintiles of vegetable or fruit intake in Korean National Health and Nutrition Examination Surveys Total (n = 7934)

Variables

Mean Age (mean ± s.d.; years) BMI (mean ± s.d.; kg/m2) Waist circumference (mean ± s.d.; cm) Alcohol intake (mean ± s.d.; servings/week) Total vegetable intake (mean ± s.d.; times/week) Total fruit intake (mean ± s.d.; times/week) Daily calorie intake (mean ± s.d.; kcal/day)c Daily carbohydrate intake (mean ± s.d.; g/day) Daily fat intake (mean ± s.d.; g/day) Daily protein intake (mean ± s.d.; g/day) Daily fiber intake (mean ± s.d.; g/day) Percent Female percentage (weighted) Smoking status percentage (weighted) Never smoke o 1 pack/day 1–2 pack/day ⩾ 2 pack/day Regular physical activity percentage (weighted) No exercise or sometimes walk Regularly walk Regular moderate-level activity Regular vigorous-level activity Education level percentage (weighted) Elementary school Middle school High school College or higher Income percentage (weighted) Quartile 1 Quartile 2 Quartile 3 Quartile 4

Vegetable intake quintilesa

Fruit intake quintilesb

Q1 (n = 1586)

Q3 (n = 1586)

Q5 (n = 1587)

Q1 (n = 1587)

Q3 (n = 1589)

Q5 (n = 1588)

39 ± 0.02 24 ± 0.1 81 ± 0.2 6 ± 0.2 32 ± 0.2 10 ± 0.2 1966 ± 13 309 ± 2 42 ± 0.5 72 ± 0.6 7.4 ± 0.1

39 ± 0.04 23 ± 0.1 80 ± 0.3 6 ± 0.3 16 ± 0.2 8 ± 0.3 1868 ± 25 296 ± 4 40 ± 0.8 67 ± 1.1 6.7 ± 0.1

39 ± 0.05 23 ± 0.1 80 ± 0.3 7 ± 0.4 31 ± 0.1 10 ± 0.2 1969 ± 27 309 ± 5 42 ± 1.0 71 ± 1.1 7.4 ± 0.2

39 ± 0.05 24 ± 0.1 81 ± 0.3 7 ± 0.4 51 ± 0.4 13 ± 0.5 2053 ± 26 324 ± 4 42 ± 0.8 78 ± 1.4 8.2 ± 0.2

39 ± 0.05 23 ± 0.1 81 ± 0.3 8 ± 0.4 27 ± 0.4 3 ± 0.04 1852 ± 25 287 ± 4 38 ± 0.9 65 ± 1.1 6.3 ± 0.1

39 ± 0.04 23 ± 0.1 80 ± 0.3 6 ± 0.4 32 ± 0.4 8 ± 0.03 1990 ± 23 317 ± 4 42 ± 0.9 73 ± 1.0 7.5 ± 0.1

39 ± 0.06 24 ± 0.1 81 ± 0.3 5 ± 0.4 38 ± 0.6 22 ± 0.43 2066 ± 31 327 ± 4 44 ± 1.2 77 ± 1.4 8.3+0.3

51

54

53

49

38

51

67

54 19 15 11

53 18 15 13

54 20 15 11

56 18 16 10

50 18 16 15

53 22 15 11

57 19 15 8

43 30 7 18

46 29 7 15

44 29 8 18

41 30 7 21

47 28 8 16

42 29 8 19

42 29 7 20

11 9 33 47

12 9 30 49

10 10 34 45

10 9 34 47

16 13 32 39

9 10 35 45

8 7 31 54

23 25 25 25

25 25 24 23

24 23 23 28

21 24 28 26

35 25 21 17

23 28 22 25

19 20 28 31

Abbreviations: BMI, body mass index; Q, quintile. aCharacteristics that are significantly different across vegetable intake quintiles are age, alcohol intake, sex, household income (Po0.05) and BMI, fruit intake, regular physical activity, protein intake, fiber intake, carbohydrate intake and calorie intake (Po0.001). b Characteristics that are significantly different across fruit intake quintiles are age, regular physical activity (P o0.05) and alcohol intake, vegetable intake, sex, smoking status, education level and household income, protein intake, fiber intake, fat intake, carbohydrate intake and calorie intake (Po 0.001). cCalorie, carbohydrates, protein, fat and fiber intake were estimated in the 24-h recall.

Table 3.

Age- and gender-adjusted and multivariate-adjusted odds ratios and 95% confidence intervals for intakes of vegetables and fruits and hypertriglyceridemia in Korean National Health and Nutrition Examination Surveys Variables

Quintile 1

Quintile 2

Quintile 3

Quintile 4

Quintile 5

P for trend

Five times/week

Total vegetables Median (frequency/week) Participants, n Cases, n Age- and gender-adjusted Multivariate adjusteda

16.8 1586 380 Ref Ref

25.8 1587 376 0.88 (0.71–1.08) 0.82 (0.66–1.02)

31.0 1586 403 0.90 (0.73–1.10) 0.88 (0.70–1.09)

36.8 1588 415 0.91 (0.73–1.13) 0.89 (0.72–1.10)

48.0 1587 427 1.07 (0.87–1.32) 1.00 (0.81–1.24)

0.36 0.68

1.01 (0.99, 1.04) 1.01 (0.98, 1.03)

Total fruits Median (frequency/week) Participants, n Cases, n Age- and gender-adjusted Multivariate adjusteda

2.7 1587 550 Ref Ref

5.5 1586 444 0.76 (0.63–0.93) 0.76 (0.62–0.92)

8.0 1589 378 0.70 (0.57–0.86) 0.72 (0.58–0.90)

11.6 1584 339 0.67 (0.54–0.83) 0.68 (0.54–0.85)

19.5 1588 290 0.60 (0.47–0.76) 0.64 (0.49–0.82)

o0.0001 0.001

0.92 (0.87–0.98) 0.95 (0.89–1.00)

a

The Multivariate model was adjusted for age (continuous, years), gender (male/female), survey year (2007/2008/2009), BMI (continuous, kg/m2), smoking (never smoke, o1 pack/day, 1–2 packs/day and ⩾ 2 packs/day), alcohol drinking (continuous, times/week), physical activity (no exercise or sometimes walk, regularly walk, regular moderate-level activity and regular vigorous-level activity), education (elementary school, middle school, high school, college or higher degree), income (quartiles) and waist circumference (continuous, cm).

inverse associations were found for the intake of citrus fruits, non-citrus fruits and carotene-rich fruits (Table 4). For the association between vegetable intake and hypertriglyceridemia, we found no significant association in age- and © 2015 Macmillan Publishers Limited

gender-adjusted analyses (Table 3). Further adjustment for survey year, BMI, waist circumference, smoking, alcohol drinking, physical activity, education and income did not materially change the results. The multivariate-adjusted OR for top quintile (median European Journal of Clinical Nutrition (2015), 1 – 7

Fruit, vegetable intake and hypertriglyceridemia C Yuan et al

4 Table 4.

Multivariate odds ratios and 95% confidence intervals of intakes of fruit and vegetable subgroups and hypertriglyceridemia in Korean National Health and Nutrition Examination Surveys Variable

Q1

Q2

Q3

Q4

Q5

Green leafy vegetable Median frequency/week Participants, n Cases, n Multivariate adjusteda

0.4 1502 360 Ref

0.9 1202 285 0.99 (0.78–1.25)

1.6 2123 532 1.06 (0.87–1.30)

2.9 1806 470 1.00 (0.80–1.26)

6.0 1300 354 1.13 (0.88–1.45)

Cruciferous vegetable Median frequency/week Participants, n Cases, n Multivariate adjusteda

9.6 1562 368 Ref

17.6 1610 364 0.77 (0.62–0.96)

23.3 1535 388 0.84 (0.67–1.05)

27.9 1619 408 0.76 (0.61–0.95)

35.3 1606 472 1.04 (0.83–1.30)

Carotene-rich vegetable Median frequency/week Participants, n Cases, n Multivariate adjusteda

24.6 1602 394 Ref

27.4 1570 348 0.75 (0.60–0.93)

29.9 1587 406 0.79 (0.63–0.98)

32.2 1570 439 0.84 (0.67–1.06)

37.0 1601 411 0.85 (0.67–1.08)

Citrus fruit Median frequency/week Participants, n Cases, n Multivariate adjusteda

0.31 1581 564 Ref

1.00 1807 460 0.76 (0.63–0.93)

2.30 1381 332 0.80 (0.63–1.00)

2.93 1784 395 0.75 (0.60–0.93)

6.00 1380 249 0.72 (0.56–0.92)

Non-citrus fruit Median frequency/week Participants, n Cases, n Multivariate adjusteda

1.84 1584 536 Ref

3.84 1589 453 0.84 (0.69–1.02)

5.70 1552 376 0.67 (0.54–0.84)

8.56 1620 343 0.70 (0.56–0.89)

15.3 1581 292 0.62 (0.48–0.81)

Carotene-rich fruit Median frequency/week Participants, n Cases, n Multivariate adjusteda

0.90 1596 560 Ref

2.13 1569 405 0.70 (0.57–0.86)

3.30 1572 381 0.74 (0.60–0.92)

5.10 1623 344 0.66 (0.52–0.83)

8.28 1570 309 0.69 (0.54–0.87)

P for trend

Five times/week

0.36

1.09 (0.96–1.23)

0.73

1.01 (0.97–1.04)

0.51

1.01 (0.91–1.13)

0.05

0.90 (0.76–1.07)

0.0004

0.94 (0.86–1.02)

0.008

0.93 (0.82–1.05)

a

The multivariate model was adjusted for age (continuous, years), gender (male/female), survey year (2007/2008/2009), BMI (continuous, kg/m2), smoking (never smoke, o1 pack/day, 1–2 packs/day and ⩾ 2 packs/day), alcohol drinking (continuous, times/week), physical activity (no exercise or sometimes walk, regularly walk, regular moderate-level activity and regular vigorous-level activity), education (elementary school, middle school, high school, college or higher degree), income (quartiles) and waist circumference (continuous, cm).

intake: 48.0 times/week) vs bottom quintile (median intake: 16.8 times/week) of total vegetable intake was 1.00 (0.81–1.24; Table 3 and Supplementary Figure 1). For specific types of vegetables including green leafy vegetables, cruciferous vegetables and carotene-rich vegetables, we also found no significant associations (Table 4). After further adjusting for total calorie, fiber, carbohydrate, fat or protein intake one at a time, the multivariate ORs for the intake of fruits or vegetable and hypertriglyceridemia remained essentially unchanged (data not shown). After including both total vegetable and total fruit intakes simultaneously in a multivariate model, we found similar associations as the primary analysis (data not shown). We also further adjusted for total fruit intake for individual vegetable groups and adjusted for total vegetable intake for individual fruit groups, respectively; the associations remained similar (data not shown). When combining total vegetable and fruit intakes together, no significant association with hypertriglyceridemia was found (data not shown). When we conducted the analyses stratified by gender (female, male), age group (⩽43, 443 years old), BMI category (o 23 kg/m2, European Journal of Clinical Nutrition (2015), 1 – 7

⩾ 23 kg/m2) and smoking status (never, current/ever), no significant interactions were found (all P for interactions 40.25; Table 5). DISCUSSION We found that high intake of fruit was associated with reduced prevalence of hypertriglyceridemia in a nationally representative sample of the Korean population. Moreover, the intake of subgroups of fruits (citrus, non-citrus and carotene-rich fruits) was each inversely associated with hypertriglyceridemia. However, no significant association with hypertriglyceridemia was found for the consumption of either total vegetables or any of the subgroups. In addition to the beneficial effect of fruits on coronary heart disease,23 stroke24 and hypertension25 found in previous studies, our study adds to the literature that higher consumption of fruit may have a protective role against elevated blood triglyceride levels in Asian populations. Previous evidence on the association or effect of fruit or vegetable and hypertriglyceridemia has come mainly from © 2015 Macmillan Publishers Limited

© 2015 Macmillan Publishers Limited

215 Ref

109 Ref

271 Ref

191 Ref

271 Ref

Age443 Cases, n OR (95% CI)

BMI o23 Cases, n OR (95% CI)

BMI ⩾ 23 Cases, n OR (95% CI)

Never smoker Cases, n OR (95% CI)

Ever/current smoker Cases, n OR (95% CI)

282 0.85 (0.61–1.18)

185 0.76 (0.55–1.05)

282 0.80 (0.61–1.05)

94 0.88 (0.59–1.31)

226 0.79 (0.59–1.06)

150 0.90 (0.65–1.24)

206 0.92 (0.67–1.27)

170 0.69 (0.51–0.94)

Q2

314 0.87 (0.64–1.20)

198 0.86 (0.64–1.17)

314 0.96 (0.75–1.24)

89 0.78 (0.52–1.18)

263 0.88 (0.65–1.18)

140 1.00 (0.72–1.39)

218 1.01 (0.74–1.38)

185 0.73 (0.56–0.95)

Q3

Q4

321 1.05 (0.77–1.43)

181 0.71 (0.52–0.97)

321 1.02 (0.77–1.35)

94 0.66 (0.46–0.95)

278 0.96 (0.71–1.32)

137 0.89 (0.65–1.21)

247 1.05 (0.79–1.41)

168 0.71 (0.52–0.98)

Total vegetables

337 1.04 (0.76–1.43)

199 0.91 (0.66–1.26)

337 1.02 (0.80–1.31)

90 0.96 (0.64–1.46)

261 0.91 (0.67–1.24)

166 1.22 (0.90–1.65)

249 1.16 (0.86–1.56)

178 0.86 (0.63–1.16)

Q5

Abbreviations: BMI, body mass index; CI, confidence interval; OR, odds ratio; Q, quintile. *P for interaction.

165 Ref

182 Ref

198 Ref

Q1

Age ⩽ 43 Cases, n OR (95% CI)

OR (95% CI)

Male Cases, n

Female Cases, n OR (95% CI)

Variable (times/week)

0.30

0.31

0.61

0.50

P*

331 Ref

218 Ref

421 Ref

129 Ref

350 Ref

200 Ref

350 Ref

200 Ref

Q1

256 0.76 (0.58–1.01)

188 0.70 (0.51–0.97)

328 0.70 (0.55–0.90)

116 0.88 (0.62–1.26)

288 0.84 (0.65–1.08)

156 0.74 (0.55–1.00)

277 0.75 (0.58–0.98)

167 0.77 (0.57–1.04)

Q2

203 0.70 (0.52–0.95)

174 0.69 (0.50–0.95)

286 0.65 (0.50–0.84)

92 0.86 (0.57–1.30)

228 0.76 (0.56–1.04)

150 0.73 (0.54–1.00)

210 0.74 (0.56–0.99)

168 0.74 (0.53–1.03)

Q3

Q4

143 0.59 (0.42–0.82)

196 0.74 (0.55–1.01)

260 0.65 (0.50–0.86)

79 0.69 (0.44–1.07)

201 0.73 (0.54–0.99)

138 0.67 (0.48–0.92)

152 0.61 (0.44–0.84)

187 0.81 (0.58–1.12)

Total fruits

111 0.53 (0.38–0.75)

178 0.70 (0.50–0.99)

230 0.63 (0.46–0.85)

60 0.57 (0.35–0.94)

176 0.73 (0.54–1.00)

114 0.56 (0.39–0.82)

113 0.56 (0.39–0.83)

177 0.75 (0.54–1.04)

Q5

0.41

0.35

0.80

0.59

P*

Table 5. Multivariate odds ratios (ORs) and 95% confidence intervals for intakes of vegetables and fruits and hypertriglyceridemia by risk factors of hypertriglyceridemia in Korean National Health and Nutrition Examination Surveys

Fruit, vegetable intake and hypertriglyceridemia C Yuan et al

5

European Journal of Clinical Nutrition (2015), 1 – 7

Fruit, vegetable intake and hypertriglyceridemia C Yuan et al

6 western countries. Several trials looking at the effect of dietary patterns that included a component of higher fruit or vegetable intake on serum triglyceride levels found that a Mediterraneanstyle dietary pattern (higher intakes of fruit, vegetable, nuts and whole grains) was associated with decreased triglyceride levels,3,11 whereas the DASH (Dietary Approaches to Stop Hypertension) diet, which emphasizes fruits and vegetables and low-fat dairy products, did not affect serum triglyceride levels.26 One crosssectional study among Brazilian adults found that a dietary intake of ⩾ 3 servings of fruit or ⩾ 4 servings of vegetable per day was associated with reduced serum triglyceride levels.12 Our study was one of the few studies that explored the association between fruit and vegetable intake and hypertriglyceridemia in an Asian population. One study investigated the association between fruit and dairy dietary patterns and their individual components and the metabolic syndrome among 406 Korean adults and found that subjects with higher intake of fruit and dairy foods had a 61% decreased likelihood of hypertriglyceridemia.15 A randomized trial in 57 hyperlipidemic patients in Israel also found that a diet supplemented with fresh red or blond grapefruit had beneficial effects on serum lipid levels, especially serum triglycerides.17 Our findings in the Korean population are consistent with these studies demonstrating strong dose response relationships, providing another piece of evidence that higher total fruit, as well as individual subgroups of fruit intake are strongly associated with lower prevalence of hypertriglyceridemia. The mechanisms whereby fruit intake may exert effects on hypertriglyceridemia are not fully understood. First, phytochemicals are important bioactive compounds found in fruits and include flavonoids and phenolic acids and so on. The antioxidant activity and anti-inflammatory potential of those compounds could reduce systemic inflammation through cellular signaling processes,27 which have been shown to protect atherosclerosis and cardiovascular disease.5,28,29 In addition, fruits have low dietary glycemic load and low energy density,30 which have been suggested to have a beneficial role in lowering serum triglycerides levels and increasing HDL-C levels over time, compared with high glycemic load diets high in refined carbohydrates and sugar.31 Recently, several short-term controlled feeding studies found that dietary fructose significantly increases postprandial triglyceride levels;31,32 thus, there may be concerns raised with regard to the fructose content in fruits. This dietary setting is different from that of our study because fructose was added to processed food, and the level of dietary fructose used in those studies was much higher than the intake from fruits in the general population. We found no association of vegetable intake with hypertriglyceridemia, which is inconsistent with previous studies that reported an inverse association with serum triglyceride levels.11,12 One possible explanation for the null association in our study is that Koreans usually eat steamed, boiled and pickled vegetable mixed with salt and soy sauce,15 whereas people in Western countries have a higher intake of raw vegetables. Although some fat-soluble nutrients from vegetables can be better absorbed after cooking with oil,33 studies have shown that cooking may reduce the bioavailability of certain nutrients, destroy digestive enzymes and alter the structure and digestibility of food.34–37 The duration and temperature during cooking may result in differential loss of nutrients, affecting the beneficial effect of vegetables.36 To support this possible explanation, previous studies found stronger inverse associations for coronary heart disease incidence38 and total mortality39 for the intake of raw vegetables compared with cooked vegetables; another study also found that high intake of baked vegetables was not associated with coronary heart disease incidence.40 There are several strengths in our study. To the best of our knowledge, this study was the first national-level study to investigate the association of total fruit and total vegetable intake European Journal of Clinical Nutrition (2015), 1 – 7

with hypertriglyceridemia in Korea. We also evaluated different subgroups including citrus fruits, non-citrus fruits, cruciferous vegetables, green leafy vegetables and carotene-rich fruits and vegetables. Our study also provided evidence regarding an inverse association of fruit consumption with hypertriglyceridemia in Korean populations, although future prospective studies are needed to confirm the relationship. The FFQ used in our study could represent long-term intake from the participants. Our study also had limitations. First, the study design was crosssectional rather than prospective. Thus, we could not provide evidence of temporal relationship between fruit intake and hypertriglyceridemia. Second, the FFQ asked frequency of intake only and did not collect portion size information. Therefore, we could not calculate the amount of fruit and vegetable intake. However, dietary recommendation is often made based on serving or frequency of intake instead of exact amount of intake. Previous study also showed that frequency of intake is more important than portion size to distinguish between high and low consumption of fruits and vegetables.41 Thus, our findings could still provide valuable information in establishing dietary recommendations. Third, although our analysis controlled for major risk factors for hypertriglyceridemia, we could not rule out the possibility of residual confounding. Also, we had caloric intake information from 24-h recall, which might not be sufficient to adjust for confounding by caloric intake. Unmeasured confounding could have occurred from additional factors related to dietary selection. For example, participants who eat more fruit and vegetable may also adopt a healthier lifestyle including a better diet, which would bias the relationship;42 however, we did adjust for several important lifestyle factors, including physical activity level, smoking status and alcohol drinking, and this had little effect on the effect estimates. In conclusion, our study supports a potential beneficial role of fruit consumption to reduce blood triglyceride levels, especially among Asian populations. Hypertriglyceridemia is prevalent and is often accompanied with other lipid abnormalities, obesity and the metabolic syndrome, which are all associated with cardiovascular disease.43 Most recently, the Global Burden of Disease Study 2010 identified diets low in fruit as one of the top five risk factors for global burden diseases, especially cardiovascular disease,44 and revealed a low intake of fruit in many regions around the world, in particular several East Asian countries. Thus, more research studies are warranted to investigate the specific effect of fruit intake across different countries. As different preparation and preservation methods may affect the nutrient level and health benefit of vegetable across countries, further studies may provide more information on the cooking method of vegetables and could investigate the effect of fresh vegetables, as well as cooked vegetables. CONFLICT OF INTEREST The authors declare no conflict of interest.

ACKNOWLEDGEMENTS We thank Dr Frank Sacks for his advice on the manuscript.

REFERENCES 1 Brunzell JD. Hypertriglyceridemia. N Engl J Med 2007; 357: 1009–1017. 2 Hodis HN, Mack WJ, Krauss RM, Alaupovic P. Pathophysiology of triglyceride-rich lipoproteins in atherothrombosis: clinical aspects. Clin Cardiol 1999; 22: II15–II20. 3 Miller M, Stone NJ, Ballantyne C, Bittner V, Criqui MH, Ginsberg HN et al. Triglycerides and cardiovascular disease: a scientific statement from the american heart association. Circulation 2011; 123: 2292–2333. 4 Lim S, Shin H, Song JH, Kwak SH, Kang SM, Won Yoon J et al. Increasing prevalence of metabolic syndrome in Korea: the Korean national health and nutrition examination survey for 1998-2007. Diabetes Care 2011; 34: 1323–1328.

© 2015 Macmillan Publishers Limited

Fruit, vegetable intake and hypertriglyceridemia C Yuan et al

7 5 Dauchet L, Amouyel P, Dallongeville J. Fruits, vegetables and coronary heart disease. Nat Rev Cardiol 2009; 6: 599–608. 6 Appel LJ, Moore TJ, Obarzanek E, Vollmer WM, Svetkey LP, Sacks FM et al. A clinical trial of the effects of dietary patterns on blood pressure. DASH collaborative research group. N Engl J Med 1997; 336: 1117–1124. 7 Boeing H, Bechthold A, Bub A, Ellinger S, Haller D, Kroke A et al. Critical review: vegetables and fruit in the prevention of chronic diseases. Eur J Nutr 2012; 51: 637–663. 8 Patel A, Barzi F, Jamrozik K, Lam TH, Ueshima H, Whitlock G et al. Serum triglycerides as a risk factor for cardiovascular diseases in the Asia-Pacific region. Circulation 2004; 110: 2678–2686. 9 Joshipura KJ, Hung HC, Li TY, Hu FB, Rimm EB, Stampfer MJ et al. Intakes of fruits, vegetables and carbohydrate and the risk of CVD. Public Health Nutr 2009; 12: 115–121. 10 Salehi-Abargouei A, Maghsoudi Z, Shirani F, Azadbakht L. Effects of dietary approaches to stop hypertension (DASH)-style diet on fatal or nonfatal cardiovascular diseases–incidence: a systematic review and meta-analysis on observational prospective studies. Nutrition 2013; 29: 611–618. 11 Rumawas ME, Meigs JB, Dwyer JT, McKeown NM, Jacques PF. Mediterranean-style dietary pattern, reduced risk of metabolic syndrome traits, and incidence in the Framingham Offspring Cohort. Am J Clin Nutr 2009; 90: 1608–1614. 12 Takahashi MM, de Oliveira EP, Moreto F, Portero-McLellan KC, Burini RC. Association of dyslipidemia with intakes of fruit and vegetables and the body fat content of adults clinically selected for a lifestyle modification program. Arch Latinoam Nutr 2010; 60: 148–154. 13 Ylonen K, Saloranta C, Kronberg-Kippila C, Groop L, Aro A, Virtanen SM. Associations of dietary fiber with glucose metabolism in nondiabetic relatives of subjects with type 2 diabetes: the Botnia Dietary Study. Diabetes Care 2003; 26: 1979–1985. 14 Egert S, Rimbach G. Which sources of flavonoids: complex diets or dietary supplements? Adv Nutr 2011; 2: 8–14. 15 Hong S, Song Y, Lee KH, Lee HS, Lee M, Jee SH et al. A fruit and dairy dietary pattern is associated with a reduced risk of metabolic syndrome. Metabolism 2012; 61: 883–890. 16 Sartika RA. Effect of trans fatty acids intake on blood lipid profile of workers in East Kalimantan, Indonesia. Malays J Nutr 2011; 17: 119–127. 17 Gorinstein S, Caspi A, Libman I, Lerner HT, Huang D, Leontowicz H et al. Red grapefruit positively influences serum triglyceride level in patients suffering from coronary atherosclerosis: studies in vitro and in humans. J Agric Food Chem 2006; 54: 1887–1892. 18 Baik I, Abbott RD, Curb JD, Shin C. Intake of fish and n-3 fatty acids and future risk of metabolic syndrome. J Am Diet Assoc 2010; 110: 1018–1026. 19 Pingali P. Westernization of Asian diets and the transformation of food systems: implications for research and policy. Food Policy 2007; 32: 281–298. 20 Godfray HCJ, Crute IR, Haddad L, Lawrence D, Muir JF, Nisbett N et al. The future of the global food system. Philos Trans R Soc Lond B Biol Sci 2010; 365: 2769–2777. 21 Park HS, Oh SW, Cho SI, Choi WH, Kim YS. The metabolic syndrome and associated lifestyle factors among South Korean adults. Int J Epidemiol 2004; 33: 328–336. 22 Anon AD. Standards of medical care in diabetes 2008. Diabetes Care 2008; 31: S12–S54. 23 Dauchet L, Amouyel P, Hercberg S, Dallongeville J. Fruit, and vegetable consumption and risk of coronary heart disease: a meta-analysis of cohort studies. J Nutr 2006; 136: 2588–2593.

24 He FJ, Nowson CA, MacGregor GA. Fruit and vegetable consumption and stroke: meta-analysis of cohort studies. Lancet 2006; 367: 320–326. 25 Appel LJ, Champagne CM, Harsha DW, Cooper LS, Obarzanek E, Elmer PJ et al. Effects of comprehensive lifestyle modification on blood pressure control: main results of the PREMIER clinical trial. JAMA 2003; 289: 2083–2093. 26 Obarzanek E, Sacks FM, Vollmer WM, Bray GA, Miller ER 3rd, Lin PH et al. Effects on blood lipids of a blood pressure-lowering diet: the Dietary Approaches to Stop Hypertension (DASH) trial. Am J Clin Nutr 2001; 74: 80–89. 27 Han X, Shen T, Lou H. Dietary polyphenols and their biological significance. Int J Mol Sci 2007; 8: 950–988. 28 Liu RH. Health benefits of fruit and vegetables are from additive and synergistic combinations of phytochemicals. Am J Clin Nutr 2003; 78: 517S–520S. 29 Diaz MN, Frei B, Vita JA, Keaney JF. Antioxidants and atherosclerotic heart disease. N Engl J Med 1997; 337: 408–416. 30 Bazzano LA, Serdula MK, Liu S. Dietary intake of fruits and vegetables and risk of cardiovascular disease. Curr Atheroscler Rep 2003; 5: 492–499. 31 Schaefer EJ, Gleason JA, Dansinger ML. Dietary fructose and glucose differentially affect lipid and glucose homeostasis. J Nutr 2009; 139: 1257S–1262S. 32 Rizkalla SW. Health implications of fructose consumption: a review of recent data. Nutr Metab 2010; 7: 82. 33 Livny O, Reifen R, Levy I, Madar Z, Faulks R, Southon S et al. Beta-carotene bioavailability from differently processed carrot meals in human ileostomy volunteers. Eur J Nutr 2003; 42: 338–345. 34 Moreno DA, Lopez-Berenguer C, Garcia-Viguera C. Effects of stir-fry cooking with different edible oils on the phytochemical composition of broccoli. J Food Sci 2007; 72: S064–S068. 35 Lyimo MH, Nyagwegwe S, Mnkeni AP. Investigations on the effect of traditional food processing, preservation and storage methods on vegetable nutrients: a case study in Tanzania. Plant Foods Hum Nutr 1991; 41: 53–57. 36 Yuan GF, Sun B, Yuan J, Wang QM. Effects of different cooking methods on health-promoting compounds of broccoli. J Zhejiang Univ Sci B 2009; 10: 580–588. 37 Link LB, Potter JD. Raw versus cooked vegetables and cancer risk. Cancer Epidemiol Biomarkers Prev 2004; 13: 1422–1435. 38 Oude Griep LM, Geleijnse JM, Kromhout D, Ocké MC, Verschuren WMM. Raw and processed fruit and vegetable consumption and 10-year coronary heart disease incidence in a population-based cohort study in The Netherlands. PLoS One 2010; 5: e13609. 39 Leenders M, Sluijs I, Ros MM, Boshuizen HC, Siersema PD, Ferrari P et al. Fruit and vegetable consumption and mortality: European prospective investigation into cancer and nutrition. Am J Epidemiol 2013; 178: 590–602. 40 Dauchet L, Ferrieres J, Arveiler D, Yarnell JW, Gey F, Ducimetiere P et al. Frequency of fruit and vegetable consumption and coronary heart disease in France and Northern Ireland: the PRIME study. Br J Nutr 2004; 92: 963–972. 41 Ashfield-Watt PA, Welch AA, Day NE, Bingham SA. Is 'five-a-day' an effective way of increasing fruit and vegetable intakes? Public Health Nutr 2004; 7: 257–261. 42 Joshipura KJ, Ascherio A, Manson JE, Stampfer MJ, Rimm EB, Speizer FE et al. Fruit and vegetable intake in relation to risk of ischemic stroke. JAMA 1999; 282: 1233–1239. 43 Hodis HN, Mack WJ, Krauss RM, Alaupovic P. Pathophysiology of triglyceride-rich lipoproteins in atherothrombosis: clinical aspects. Clin Cardiol 1999; 22: 15–20. 44 Lim SS, Vos T, Flaxman AD, Danaei G, Shibuya K, Adair-Rohani H et al. A comparative risk assessment of burden of disease and injury attributable to 67 risk factors and risk factor clusters in 21 regions, 1990–2010: a systematic analysis for the Global Burden of Disease Study 2010. Lancet 2012; 380: 2224–2260.

Supplementary Information accompanies this paper on European Journal of Clinical Nutrition website (http://www.nature.com/ejcn)

© 2015 Macmillan Publishers Limited

European Journal of Clinical Nutrition (2015), 1 – 7