Nutrition, Metabolism & Cardiovascular Diseases (2015) xx, 1e6
Available online at www.sciencedirect.com
Nutrition, Metabolism & Cardiovascular Diseases journal homepage: www.elsevier.com/locate/nmcd
Dietary intake of carotenoids and risk of type 2 diabetes I. Sluijs a,*, E. Cadier a, J.W.J. Beulens a, D.L. van der A b, A.M.W. Spijkerman b, Y.T. van der Schouw a a b
Julius Center for Health Sciences and Primary Care, University Medical Center Utrecht, Utrecht, The Netherlands Center for Nutrition, Prevention and Health Services, National Institute for Public Health and the Environment (RIVM), Bilthoven, The Netherlands
Received 4 June 2014; received in revised form 28 November 2014; accepted 22 December 2014 Available online - - -
KEYWORDS Carotenoids; Cohort study; Diet; Type 2 diabetes; b-carotene; a-carotene
Abstract Background and aims: Carotenoids may reduce diabetes risk, due to their antioxidant properties. However, the association between dietary carotenoids intake and type 2 diabetes risk is still unclear. Therefore, the objective of this study was to examine whether higher dietary carotenoid intakes associate with reduced type 2 diabetes risk. Methods and results: Data from 37,846 participants of the European Prospective Investigation into Cancer and Nutrition- Netherlands study were analyzed. Dietary intakes of b-carotene, a-carotene, b-cryptoxanthin, lycopene, lutein & zeaxanthin and the sum of these carotenoids were assessed using a validated food frequency questionnaire. Incident type 2 diabetes was mainly selfreported, and verified against general practitioner information. Mean SD total carotenoid intake was 10 4 mg/day. During a mean SD follow-up of 10 2years, 915 incident cases of type 2 diabetes were ascertained. After adjustment for age, sex, diabetes risk factors, dietary intake, waist circumference and BMI, higher b-carotene intakes associated inversely with diabetes risk [Hazard Ratio quartile 4 versus quartile 1 (HRQ4): 0.78 (95%CI:0.64,0.95), P-linear trend 0.01]. For a-carotene, a borderline significant reduced risk was observed, with a HRQ4 of 0.85 (95%CI:0.70,1.03), and P-linear trend 0.05. b-cryptoxanthin, lycopene, lutein & zeaxanthin, and the sum of all carotenoids did not associate with diabetes risk. Conclusions: This study shows that diets high in b-carotene and a-carotene are associated with reduced type 2 diabetes in generally healthy men and women. ª 2015 Elsevier B.V. All rights reserved.
Introduction Carotenoids are plant pigments that provide the yellow, orange and red pigments in fruits and vegetables [1]. Smaller quantities of carotenoids are also present in bread, eggs and oils [2]. Carotenoids have antioxidant functions, and may reduce type 2 diabetes risk by reducing oxidative stress, which plays an important role in the development of diabetes [3e5]. However, evidence from prospective
* Corresponding author. Julius Center for Health Sciences and Primary Care, University Medical Center Utrecht, PO Box 85500, 3508GA Utrecht, The Netherlands. Tel.: þ31887550465; fax: þ31887568099. E-mail address: [email protected] (I. Sluijs).
studies on the associations of dietary carotenoids and diabetes risk is scarce and inconsistent. So far, b-carotene has been studied most often, with two observational studies reporting an inverse association of dietary b-carotene with diabetes [6,7] and another reporting no association among male smokers [8]. In addition, studies that addressed serum or plasma b-carotene values in association with diabetes risk showed inconsistent results [6,9e11]. Randomized controlled trials with b-carotene supplementation consistently showed no effect on diabetes among women with a history of cardiovascular disease [12], generally healthy men [13] and among male smokers [14]. For dietary lycopene, three cohort studies showed no association with diabetes
http://dx.doi.org/10.1016/j.numecd.2014.12.008 0939-4753/ª 2015 Elsevier B.V. All rights reserved.
Please cite this article in press as: Sluijs I, et al., Dietary intake of carotenoids and risk of type 2 diabetes, Nutrition, Metabolism & Cardiovascular Diseases (2015), http://dx.doi.org/10.1016/j.numecd.2014.12.008
2
[7,8,15]. Very little research has been performed on other individual dietary carotenoids such as a-carotene, b-cryptoxanthin, lutein, zeaxanthin and the total of carotenoids, and therefore no conclusions can yet be drawn on such intakes [7,8]. Inconsistencies in results between studies performed thus far could be due to differences in study design, range of carotenoids intake, or population characteristics such as age and sex. Moreover, previous research suggests that smoking increases oxidative stress [16], and that smokers may need a higher intake of carotenoids in comparison with non-smokers to reduce their diabetes risk [9]. One study indeed found reduced diabetes risk with higher serum b-carotene and total carotenoid values in nonsmokers only [9], and a study among male smokers reported no associations of dietary [8] or supplementary [14] b-carotene with diabetes risk. However, two other studies did not confirm such an interaction with dietary carotenoids [7] or supplementary b-carotene [12]. In the European Prospective Investigation into Cancer and Nutrition- Netherlands (EPIC-NL) cohort, we investigated associations of dietary intake of six carotenoids and incidence of type 2 diabetes. In addition, we explored whether smoking modified these associations. Methods Study population EPIC-NL consists of the two Dutch contributions to EPIC, Prospect-EPIC and MORGEN-EPIC [17]. Both cohorts were set up simultaneously in 1993e1997 and were merged in 2007 according to standardized processes into one large Dutch EPIC cohort [17]. Prospect-EPIC includes 17,357 women aged 49e70 years at baseline, participating in the national breast cancer screening program, and living in the city of Utrecht and its surroundings [18]. The MORGEN-EPIC study includes 22,654 men and women aged 21e64 years selected from random samples of the Dutch population in three cities in the Netherlands (Amsterdam, Doetinchem, and Maastricht) [19]. All participants provided written informed consent before study inclusion. The study complied with the Declaration of Helsinki and was approved by local ethical committees. After exclusion of prevalent diabetes cases (n Z 615), participants without consent to linkage with disease registries (n Z 1027), participants with abnormal high or low energy intake (n Z 400) and missing data on carotenoid intake (n Z 218), 37,846 participants were included in the analysis. Dietary intake Participants completed a self-administered validated food frequency questionnaire (FFQ), containing questions about the usual consumption frequency of 178 foods during the year preceding enrollment. The FFQ was administrated once at baseline. The Dutch food composition table was
I. Sluijs et al.
used to calculate energy and nutrient intakes [20]. All nutrient values were adjusted for total energy intake by the regression residual method. Nutritional intakes of carotenoids examined were b-carotene, a-carotene, b-cryptoxanthin, lycopene, lutein plus zeaxanthin, and the sum of these carotenoids. The FFQ has been validated against twelve 24-h recalls [21]. Spearman correlations for carotenoids ranged from 0.16 for b-carotene to 0.31 for alpha-carotene for men, and from 0.17 for b-cryptoxanhin to 0.32 for lycopene for women (unpublished data). Energy reporting Basal metabolic rate (BMR) was estimated using the Schofield equations. Participants with an energy intake vs. BMR of 2.40 were classified as energy over-reporters, according to the Goldberg cut-offs [22]. Energy misreporters were defined as energy under-plus over-reporters. The remaining was defined as normal energy-reporters. Type 2 diabetes Ascertainment and verification of diabetes has been described elsewhere [23]. In summary, occurrence of diabetes during follow-up was obtained from self-report in two follow-up questionnaires and from linkage with hospital discharge diagnoses registries. Follow-up was complete until 1 January 2006. In the Prospect study, occurrence of diabetes was also obtained from a urinary glucose strip test for detection of glucosuria. Cases notified by any of these methods were verified against general practitioner or pharmacist information. Only cases confirmed by either the general practitioner or pharmacist were included in the analyses. Other variables At baseline, participants completed a general questionnaire containing questions on demographics, presence of chronic diseases, and chronic disease risk factors. Physical activity, assessed by validated questionnaire, was categorized using the Cambridge Physical Activity Score [24]. Blood pressure was measured twice in a supine position on the right arm using a Boso Oscillomat (Bosch & Son) (Prospect) or on the left arm using a random zero Sphygmomano-meter (MORGEN) from which the mean was taken. Hypertension was defined as: diastolic BP > 90 mm Hg, and/or systolic BP > 140 mm Hg, and/or self-report. Smoking was categorized into never, past and current. Body weight and waist circumference were measured, and BMI was calculated as weight divided by height squared (kg/m2). Data analysis Baseline characteristics were presented according to quartiles of carotenoid intake. Pearson correlations were
Please cite this article in press as: Sluijs I, et al., Dietary intake of carotenoids and risk of type 2 diabetes, Nutrition, Metabolism & Cardiovascular Diseases (2015), http://dx.doi.org/10.1016/j.numecd.2014.12.008
Risk of type 2 diabetes
3
calculated to determine correlations of carotenoids with other dietary intakes. Cox proportional hazards regression analysis was used to determine hazard ratios (HR) and 95% confidence intervals (95%CI) for the associations of carotenoid (in quartiles) and diabetes risk. P-values for linear trend were estimated by using the median carotenoid intake per quartile as a continuous variable in the regression model. In model 1, we adjusted for age (continuous) and sex. In model 2, educational status (low/middle/high), physical activity (inactive/moderately-inactive/moderately-active/ active) and systolic blood pressure (continuous) were added. In model 3, dietary intakes of fiber and vitamin E (continuous) were added. In model 4, we added waist circumference and BMI (continuous). Interactions with smoking were investigated by means of a likelihood ratio test. Former smokers and never smokers were combined into “non-smokers”. A series of sensitivity analyses were performed by excluding participants diagnosed with diabetes within the first two years of follow-up (n Z 361), excluding those with baseline chronic diseases (coronary heart disease, stroke, hyperlipidemia and/or hypertension; n Z 20,742), excluding energy misreporters (n Z 9894), and by adding vitamin C as additional confounder to the multivariable model. The analyses were performed using SPSS version 20.0. The significance level was set two sided at 0.05. Results The study population had a mean SD age of 49.1 11.9 years. The majority of participants was female (74%), and
31% were smokers (Table 1). Mean SD carotenoid intake was 10 4 mg/day, with b-carotene and lycopene as the largest contributors. Correlations between individual carotenoids ranged from 0.02 for a-carotene and lycopene to 0.87 for a-carotene and b-carotene. Correlations between carotenoids and fiber ranged from 0.06 for lycopene to 0.51 for b-cryptoxanthin. Correlations between carotenoids and vitamin C ranged from 0.24 for lutein & zeaxanthin to 0.88 for b-cryptoxanthin. Mean age and intakes of vitamin C, vitamin E and fiber increased over the quartiles of total carotenoid intake, whereas male sex, waist circumference, physical inactivity, low education, current smoking, diastolic blood pressure and percentage with hypertension decreased (Table 1). During a mean SD follow-up of 10 2 years, 915 incident cases of diabetes were identified. After adjustment for confounders (model 4), an inverse association was found for b-carotene comparing the highest vs. the lowest quartile. The corresponding HR (HRQ4) was 0.78 (95%CI:0.64,0.95), and P-for linear trend 0.01. For a-carotene, a borderline significant inverse association was found, with a HRQ4 of 0.85 (95%CI:0.70,1.03), and P-for linear trend of 0.05. b-cryptoxanthin, lycopene, lutein & zeaxanthin and the total of all carotenoids did not associate with diabetes in multivariable analysis (Table 2). Smoking did not modify associations between carotenoid intake and diabetes (P-values for interaction > 0.49). Exclusion of participants who developed diabetes within the first 2 years after the start of the study did not materially change our findings (HRQ4 b-carotene: 0.78 95% CI:0.64,0.98; a-carotene: 0.86 95%CI:0.70,1.06). After exclusion of participants diagnosed with chronic disease at baseline, all estimates remained essentially the same
Table 1 Baseline characteristics according to quartiles of total dietary carotenoid intake.a Quartiles of total carotenoid intake
Participants, n Total carotenoids, mg [median (p25, p75)] b-carotene, mg [median (p25, p75)] a-carotene, mg [median (p25, p75)] b-cryptoxanthin, mg [median (p25, p75)] Lycopene, mg [median (p25, p75)] Lutein & zeaxanthin, mg [median (p25, p75)] Energy, kcal Vitamin C, mg Vitamin E, mg Fiber, g Alcohol, g [median (p25, p75)] Age, y [median (p25, p75)] Male sex BMI, kg/m2 Waist circumference, cm Physically (moderately) inactive Low education Current smoker Systolic blood pressure, mmHg Diastolic blood pressure, mmHg Hypertension a
Q1 (low)
Q2
Q3
Q4 (high)
9461 6.2 (5.1, 6.9) 1.4 (1.1, 1.7) 0.3 (0.2, 0.4) 0.08 (0.05, 0.13) 1.4 (0.8, 2.2) 2.4 (1.8, 3.0) 2080 651 82 33 12 3 21 5 4.65 (0.54, 15.5) 51 (41, 57) 3509 (36.1) 25.7 4.0 87 12 3433 (36.3) 6927 (61.3) 3447 (36.4) 127 19 79 11 3535 (37.4)
9462 8.6 (8.1, 9.1) 2.0 (1.7, 2.4) 0.4 (0.3, 0.5) 0.11 (0.07, 0.17) 2.5 (1.5, 3.5) 3.3 (2.7, 4.0) 2084 611 101 36 12 3 23 4 4.56 (0.82, 13.9) 51 (41, 57) 2623 (27.7) 25.6 3.9 85 11 3033 (32.0) 5415 (57.2) 2840 (30.0) 126 19 78 11 3443 (36.4)
9462 10.7 (10.2, 11.3) 2.6 (2.1, 3.0) 0.5 (0.3, 0.8) 0.13 (0.08, 0.20) 3.2 (2.1, 4.4) 4.1 (3.3, 4.9) 2063 587 116 41 12 3 24 4 4.66 (0.89, 13.7) 51 (42, 58) 2114 (22.3) 25.6 3.9 85 11 2902 (30.6) 5328 (56.3) 2678 (28.3) 126 19 78 11 3442 (36.4)
9461 14.2 (13.0, 16.1) 3.5 (2.8, 4.3) 0.7 (0.4, 1.1) 0.17 (0.10, 0.24) 4.4 (2.8, 6.4) 5.3 (4.0, 6.7) 1984 576 138 50 13 3 26 5 4.65 (0.75, 14.4) 52 (43, 58) 1438 (15.2) 25.7 4.0 84 11 2800 (29.6) 5056 (53.5) 2569 (27.2) 126 19 77 11 3521 (37.2)
Values are mean SD or n (percentage), unless stated otherwise; all nutrients are adjusted for total energy intake.
Please cite this article in press as: Sluijs I, et al., Dietary intake of carotenoids and risk of type 2 diabetes, Nutrition, Metabolism & Cardiovascular Diseases (2015), http://dx.doi.org/10.1016/j.numecd.2014.12.008
4
I. Sluijs et al.
Table 2 Associations between quartiles of dietary carotenoid intake and risk of type 2 diabetes.a Quartiles of intake
Total carotenoids, median intake, mg/d Cases/at risk, n Model 1b Model 2c Model 3d Model 4e b-carotene, median intake, mg/d Cases/at risk, n Model 1b Model 2c Model 3d Model 4e a-carotene, median intake, mg/d Cases/at risk, n Model 1b Model 2c Model 3d Model 4e b-cryptoxanthin, median intake, mg/d Cases/at risk, n Model 1b Model 2c Model 3d Model 4e Lycopene, median intake, mg/d Cases/at risk, n Model 1b Model 2c Model 3d Model 4e Lutein & Zeaxanthin, median intake, mg/d Cases/at risk, n Model 1b Model 2c Model 3d Model 4e
P-linear trend
Q1 (low)
Q2
Q3
Q4 (high)
6.2 253/9461 1 1 1 1 1.3 223/9461 1 1 1 1 0.2 260/9460 1 1 1 1 0.05 235/9460 1 1 1 1 0.9 312/9460 1 1 1 1 2.1 198/9460 1 1 1 1
8.6 225/9462 0.93 (0.78,1.11) 0.93 (0.78,1.12) 0.92 (0.77,1.10) 0.91 (0.76,1.10) 1.9 220/9462 0.90 (0.75,1.09) 0.96 (0.79,1.15) 0.93 (0.77,1.13) 0.93 (0.77,1.13) 0.4 206/9462 0.89 (0.74,1.07) 0.94 (0.79,1.13) 0.93 (0.77,1.12) 0.99 (0.82,1.19) 0.10 210/9462 0.88 (0.73, 1.06) 0.95 (0.79, 1.15) 0.93 (0.77, 1.13) 0.93 (0.77, 1.12) 2.6 266/9462 0.88 (0.74, 1.04) 0.94 (0.79, 1.11) 0.92 (0.78, 1.09) 0.90 (0.76, 1.07) 3.1 200/9462 1.01 (0.83, 1.23) 1.02 (0.84, 1.25) 1.02 (0.83, 1.24) 1.02 (0.83, 1.24)
10.7 209/9462 0.88 (0.73,1.05) 0.89 (0.74,1.07) 0.86 (0.71,1.04) 0.84 (0.69,1.02) 2.6 220/9462 0.84 (0.70,1.01) 0.89 (0.74,1.08) 0.86 (0.71,1.04) 0.86 (0.70,1.04) 0.6 217/9464 0.77 (0.64,0.93)* 0.84 (0.69,1.01) 0.81 (0.67,0.98)* 0.85 (0.70,1.02) 0.15 224/9464 0.82 (0.68, 0.99)* 0.93 (0.77,1.12) 0.91 (0.74,1.10) 0.96 (0.79, e 1.17) 3.3 188/9464 0.86 (0.72, 1.04) 0.96 (0.80, 1.15) 0.93 (0.78, 1.12) 0.97 (0.81, 1.16) 4.1 222/9464 1.01 (0.83, 1.23) 1.00 (0.82, 1.22) 0.98 (0.80, 1.20) 0.89 (0.73, 1.09)
14.2 228/9461 0.97 (0.80,1.16) 0.99 (0.82,1.19) 0.92 (0.76,1.13) 0.85 (0.69,1.04) 3.7 252/9461 0.85 (0.70, 1.02) 0.91 (0.76,1.10) 0.85 (0.69,1.04) 0.78 (0.64,0.95)* 0.9 232/9460 0.80 (0.67,0.96)* 0.87 (0.73,1.05) 0.84 (0.69,1.01) 0.85 (0.70,1.03) 0.24 246/9460 0.83 (0.69, 1.00)* 0.99 (0.82, 1.19) 0.96 (0.78, 1.18) 0.96 (0.78, 1.19) 5.3 149/9460 0.79 (0.65, 0.96)* 0.94 (0.78, 1.15) 0.92 (0.75, 1.12) 0.97 (0.79, 1.18) 5.9 295/9460 1.11 (0.92, 1.34) 1.08 (0.89, 1.30) 1.04 (0.85, 1.27) 0.91 (0.74, 1.11)
0.19 0.81 0.38 0.09
0.09 0.32 0.10 0.01
0.02 0.13 0.06 0.05
0.07 0.97 0.77 0.89
0.01 0.55 0.36 0.73
0.23 0.42 0.72 0.22
*p 0.05 for difference from quartile 1. a Values are hazard ratio’s (95%CI) derived from cox regression; all carotenoid intakes were adjusted for total energy intake. b Adjusted for age and sex. c Adjusted for model 1 þ educational status, physical activity and systolic blood pressure. d Adjusted for model 2 þ fiber intake and vitamin E intake. e Adjusted for model 3 þ BMI and waist circumference.
(HRQ4 b-carotene: 0.72 95%CI:0.46,1.12; a-carotene: 0.74 95%CI:0.49,1.11). Excluding energy misreporters did not materially change our findings either (HRQ4 b-carotene: 0.84 95%CI:0.65,1.09; a-carotene: 0.79 95%CI:0.61,1.02) In addition, after adding vitamin C to model 4, associations did not materially change (HRQ4 b-carotene: 0.78 95% CI:0.63,0.95; a-carotene: 0.85 95%CI:0.70,1.03).
Discussion In this prospective study among 37,846 men and women followed-up for a mean of 10 years, higher dietary intakes of b-carotene and a-carotene were associated with a reduced diabetes risk, whereas dietary intakes of other individual carotenoids were not associated with diabetes risk. The total of individual carotenoids tended to be
associated with reduced diabetes, similar to a-carotene and b-carotene, but this was not statistically significant. The associations were not modified by smoking status. Strengths of this study include the large size of the population, long follow-up time and prospective design. Additionally, in this study only cases of diabetes that were confirmed to be diagnosed with diabetes by their general practitioner were included in the analyses. This study has several limitations. First of all, participants with a higher carotenoid intake may have an overall healthier lifestyle than participants with a lower intake of carotenoids. Although we did adjust for several lifestyle factors such as physical activity and fiber intake, we cannot exclude that residual confounding related to a healthier lifestyle may still exist. Second, although our study design allows us to draw conclusions on associations of dietary carotenoids and diabetes, it does not allow us to make causal
Please cite this article in press as: Sluijs I, et al., Dietary intake of carotenoids and risk of type 2 diabetes, Nutrition, Metabolism & Cardiovascular Diseases (2015), http://dx.doi.org/10.1016/j.numecd.2014.12.008
Risk of type 2 diabetes
inferences. Third, presence of type 2 diabetes goes often undetected, and may be preclinical up to 9e12 years [25]. Undetected diabetes cases may have been misclassified as non-diabetic individuals, resulting in attenuation of associations. Fourth, we used a FFQ to measure carotenoid intake, which in general is known to be prone to imprecision and reporting bias [26]. Correlation coefficients between the FFQ and repeated 24 h recalls for measurement of carotenoid intake were moderate. Random variation in reporting intake of carotenoid-containing foods is expected to underlie the moderate correlation coefficients. This is supported by the finding that excluding energy misreporters did not materially change our findings. Such random variation causes bias toward the null and can only have attenuated our associations. Still, we need to be careful with drawing conclusions from our study. Finally, although the food composition table used in our study provides detailed information on carotenoid content of foods, 2% (for beta-carotene) to 29% (for zeaxanthin) of fruits and vegetables listed in the food composition table were not assigned carotenoid values. We therefore cannot exclude that we underestimated associations due to such missing information. The inverse association of dietary b-carotene intake and diabetes is consistent with two smaller scale observational studies [6,7]: One study among 846 Swedish men, found a 29% reduced diabetes risk in the highest quartile of bcarotene intake (>1.9 mg/d) compared with the lowest, during 7 years follow-up [6]. Another study, including 4304 adults, found a borderline significant 26% reduced diabetes risk in the highest (>2.1 mg/d) vs. lowest quartile of b-carotene intake [7]. We found a 22% reduced risk in the highest b-carotene intake quartile, which is only slightly lower than the estimates in those studies. In contrast, a cohort study including 29,000 smokers, with similar b-carotene and higher a-carotene intake as in our study, did neither find associations with b-carotene nor with a-carotene [8]. In contrast with our study, the study population consisted of men only and had a higher mean age. Another study also did not find associations of dietary a-carotene and diabetes [7], but a-carotene intakes in that study were much lower compared with our study. Studies that investigated b-carotene in serum or plasma have shown inconsistent results, with some reporting inverse [6,9] and others reporting no associations [10,11] with diabetes risk. Serum or plasma a-carotene did not associate with diabetes risk in previous studies [9,11]. Although serum and plasma carotenoid values reflect nutritional status, they do not directly relate to intake. In fact, serum and plasma carotenoid values are a reflection of many more factors than intake, including absorption and metabolism [27]. Therefore, the comparability of studies examining serum or plasma and intake values is limited. Several randomized controlled trials with b-carotene supplementation have been conducted [12e14], with doses of supplementation ranging from 20 mg/d to 50 mg every other day. None of the trials found significant effects on diabetes [12e14]. The discrepancy with the observational studies may be due to a higher intake of
5
b-carotene in the trials as a consequence of supplementation. The ATBC trial was the first to show increased mortality with b-carotene supplementation [28], later confirmed in a large meta-analysis [29]. This thus suggests that b-carotene may even become harmful at higher doses. Differences in bioavailability between natural and synthetic forms of b-carotene may also underlie the differences between the intervention and observational studies [30]. However, residual confounding should also be considered. We found no association of b-cryptoxanthin and lutein plus zeaxanthin with diabetes, whereas another prospective study did suggest inverse associations for those carotenoids [7]. That study completed their food database with analyzed carotenoid values from national foods, which may explain the difference with our study. Our finding of no association for dietary lycopene with diabetes is consistent with previous research [7,8,15]. In our study we only found associations for some individual carotenoids. For the total of individual carotenoids we found a tendency towards a reduced diabetes risk, similar to a-carotene and b-carotene, but this was not statistically significant. This is likely due to the lack of association for the other individual carotenoids such as lycopene and b-cryptoxanthin. One observational study with 4304 adults did find an inverse association between total carotenoids and diabetes [7]. We found no interactions with smoking, which is consistent with previously conducted studies on dietary carotenoids and diabetes [7,12]. Carotenoids are known to have antioxidant functions, which may underlie the observed inverse associations with diabetes. It has been suggested that antioxidants such as carotenoids might be effective in reducing diabetes by reducing oxidative stress [4]. Of all carotenoids addressed in this study, b-carotene is known to be a strong antioxidant [1], which may explain the observed association with diabetes in our study. We also found an inverse association with a-carotene, which could possibly be explained by the high correlation between b-carotene and a-carotene. Lycopene also has high antioxidant properties [1]. Even though lycopene was a major contributor to total carotenoid intake, we did not find an association of lycopene with diabetes risk. It is unclear why lycopene shows no association with diabetes in the present study. Relative validity for lycopene was higher than for b-carotene, the bioavailability for lycopene and b-carotene are comparable [1], and the range of intake for lycopene was wider than for b-carotene. Therefore this unlikely explains why we were unable to find an association of lycopene with diabetes. The higher completeness of the food database for bcarotene values of foods (92% for vegetables; 98% for fruits) than for lycopene values of foods (79% for both vegetables and fruits) may partly explain these differences. In conclusion, this study shows that diets high in bcarotene and a-carotene are associated with reduced type 2 diabetes in generally healthy men and women. The association does not differ by smoking status.
Please cite this article in press as: Sluijs I, et al., Dietary intake of carotenoids and risk of type 2 diabetes, Nutrition, Metabolism & Cardiovascular Diseases (2015), http://dx.doi.org/10.1016/j.numecd.2014.12.008
6
I. Sluijs et al.
Conflict of interest None of the authors declared a conflict of interest. Acknowledgments The EPIC-NL study was funded by ‘European Commission: Public Health and Consumer Protection Directorate 1993e2004; Research Directorate-General 2005’; Dutch Ministry of Public Health, Welfare and Sports (VWS), Netherlands Cancer Registry (NKR), Dutch Prevention Funds, Dutch ZON (Zorg Onderzoek Nederland), World Cancer Research Fund (WCRF) (The Netherlands).
[14]
[15]
[16] [17]
[18]
References [19] [1] Perera C, Yen G. Functional properties of carotenoids in human health. Int J Food Prop 2007 Apr 27;10:201e30. [2] Rao AV, Rao LG. Carotenoids and human health. Pharmacol Res 2007 Mar;55(3):207e16. [3] Ceriello A, Motz E. Is oxidative stress the pathogenic mechanism underlying insulin resistance, diabetes, and cardiovascular disease? the common soil hypothesis revisited. Arterioscler Thromb Vasc Biol 2004 May;24(5):816e23. [4] Johansen JS, Harris AK, Rychly DJ, Ergul A. Oxidative stress and the use of antioxidants in diabetes: linking basic science to clinical practice. Cardiovasc Diabetol 2005;4(1):5. [5] Psaltopoulou T, Ilias I, Alevizaki M. The role of diet and lifestyle in primary, secondary, and tertiary diabetes prevention: a review of meta-analyses. Rev Diabet Stud 2010;7(1):26e35. [6] Arnlov J, Zethelius B, Riserus U, Basu S, Berne C, Vessby B, et al. Serum and dietary beta-carotene and alpha-tocopherol and incidence of type 2 diabetes mellitus in a community-based study of Swedish men: report from the Uppsala Longitudinal Study of Adult Men (ULSAM) study. Diabetologia 2009 Jan;52(1):97e105. [7] Montonen J, Knekt P, Jarvinen R, Reunanen A. Dietary antioxidant intake and risk of type 2 diabetes. Diabetes Care 2004 Feb;27(2): 362e6. [8] Kataja-Tuomola MK, Kontto JP, Mannisto S, Albanes D, Virtamo J. Intake of antioxidants and risk of type 2 diabetes in a cohort of male smokers. Eur J Clin Nutr 2011 May;65(5):590e7. [9] Hozawa A, Jacobs DRJ, Steffes MW, Gross MD, Steffen LM, Lee DH. Associations of serum carotenoid concentrations with the development of diabetes and with insulin concentration: interaction with smoking: the Coronary Artery Risk Development in Young Adults (CARDIA) Study. Am J Epidemiol 2006 May 15;163(10): 929e37. [10] Reunanen A, Knekt P, Aaran RK, Aromaa A. Serum antioxidants and risk of non-insulin dependent diabetes mellitus. Eur J Clin Nutr 1998 Feb;52(2):89e93. [11] Wang L, Liu S, Pradhan AD, Manson JE, Buring JE, Gaziano JM, et al. Plasma lycopene, other carotenoids, and the risk of type 2 diabetes in women. Am J Epidemiol 2006 Sep 15;164(6):576e85. [12] Song Y, Cook NR, Albert CM, Van Denburgh M, Manson JE. Effects of vitamins C and E and beta-carotene on the risk of type 2 diabetes in women at high risk of cardiovascular disease: a randomized controlled trial. Am J Clin Nutr 2009 Aug;90(2):429e37. [13] Liu S, Ajani U, Chae C, Hennekens C, Buring JE, Manson JE. Longterm beta-carotene supplementation and risk of type 2 diabetes
[20] [21]
[22]
[23]
[24]
[25]
[26]
[27] [28]
[29]
[30]
mellitus: a randomized controlled trial. J Am Med Assoc 1999 Sep 15;282(11):1073e5. Kataja-Tuomola M, Sundell JR, Mannisto S, Virtanen MJ, Kontto J, Albanes D, et al. Effect of alpha-tocopherol and beta-carotene supplementation on the incidence of type 2 diabetes. Diabetologia 2008 Jan;51(1):47e53. Wang L, Liu S, Manson JE, Gaziano JM, Buring JE, Sesso HD. The consumption of lycopene and tomato-based food products is not associated with the risk of type 2 diabetes in women. J Nutr 2006 Mar;136(3):620e5. Isik B, Ceylan A, Isik R. Oxidative stress in smokers and nonsmokers. Inhal Toxicol 2007 Jul;19(9):767e9. Beulens JWJ, Monninkhof EM, Verschuren WMM, van der Schouw YT, Smit J, Ocke MC, et al. Cohort profile: the EPIC-NL study. Int J Epidemiol 2010 Oct;39(5):1170e8. Boker LK, van Noord PA, van der Schouw YT, Koot NV, Bueno de Mesquita HB, Riboli E, et al. Prospect-EPIC Utrecht: study design and characteristics of the cohort population. European prospective investigation into cancer and nutrition. Eur J Epidemiol 2001; 17(11):1047e53. Verschuren WMM, Blokstra A, Picavet HSJ, Smit HA. Cohort profile: the doetinchem cohort study. Int J Epidemiol 2008 Dec;37(6): 1236e41. Dutch food composition table. The Hague: NEVO Foundation; 2006. Ocke MC, Bueno-de-Mesquita HB, Goddijn HE, Jansen A, Pols MA, van Staveren WA, et al. The Dutch EPIC food frequency questionnaire. I. Description of the questionnaire, and relative validity and reproducibility for food groups. Int J Epidemiol 1997;26(Suppl. 1): S37e48. Goldberg GR, Black AE, Jebb SA, Cole TJ, Murgatroyd PR, Coward WA, et al. Critical evaluation of energy intake data using fundamental principles of energy physiology: 1. Derivation of cutoff limits to identify under-recording. Eur J Clin Nutr 1991 Dec; 45(12):569e81. Sluijs I, van der A DL, Beulens JWJ, Spijkerman AMW, Ros MM, Grobbee DE, et al. Ascertainment and verification of diabetes in the EPIC-NL study. Neth J Med 2010 Aug;68(1):333e9. Wareham NJ, Jakes RW, Rennie KL, Schuit J, Mitchell J, Hennings S, et al. Validity and repeatability of a simple index derived from the short physical activity questionnaire used in the European Prospective Investigation into Cancer and Nutrition (EPIC) study. Public Health Nutr 2003 Jun;6(4):407e13. Harris MI, Klein R, Welborn TA, Knuiman MW. Onset of NIDDM occurs at least 4e7 yr before clinical diagnosis. Diabetes Care 1992 Jul;15(7):815e9. Margetts BM, Pietinen P. European Prospective Investigation into Cancer and Nutrition: validity studies on dietary assessment methods. Int J Epidemiol 1997;26(Suppl. 1):S1e5. Potischman N. Biologic and methodologic issues for nutritional biomarkers. J Nutr 2003 Mar;133(Suppl. 3):875Se80S. The effect of vitamin E and beta carotene on the incidence of lung cancer and other cancers in male smokers. The alpha-tocopherol, beta carotene cancer prevention study group. N Engl J Med 1994 Apr 14;330(15):1029e35. Vivekananthan DP, Penn MS, Sapp SK, Hsu A, Topol EJ. Use of antioxidant vitamins for the prevention of cardiovascular disease: meta-analysis of randomised trials. Lancet 2003 Jun 14;361(9374): 2017e23. Ben-Amotz A, Levy Y. Bioavailability of a natural isomer mixture compared with synthetic all-trans beta-carotene in human serum. Am J Clin Nutr 1996 May;63(5):729e34.
Please cite this article in press as: Sluijs I, et al., Dietary intake of carotenoids and risk of type 2 diabetes, Nutrition, Metabolism & Cardiovascular Diseases (2015), http://dx.doi.org/10.1016/j.numecd.2014.12.008
Available online at www.sciencedirect.com
Nutrition, Metabolism & Cardiovascular Diseases journal homepage: www.elsevier.com/locate/nmcd
Dietary intake of carotenoids and risk of type 2 diabetes I. Sluijs a,*, E. Cadier a, J.W.J. Beulens a, D.L. van der A b, A.M.W. Spijkerman b, Y.T. van der Schouw a a b
Julius Center for Health Sciences and Primary Care, University Medical Center Utrecht, Utrecht, The Netherlands Center for Nutrition, Prevention and Health Services, National Institute for Public Health and the Environment (RIVM), Bilthoven, The Netherlands
Received 4 June 2014; received in revised form 28 November 2014; accepted 22 December 2014 Available online - - -
KEYWORDS Carotenoids; Cohort study; Diet; Type 2 diabetes; b-carotene; a-carotene
Abstract Background and aims: Carotenoids may reduce diabetes risk, due to their antioxidant properties. However, the association between dietary carotenoids intake and type 2 diabetes risk is still unclear. Therefore, the objective of this study was to examine whether higher dietary carotenoid intakes associate with reduced type 2 diabetes risk. Methods and results: Data from 37,846 participants of the European Prospective Investigation into Cancer and Nutrition- Netherlands study were analyzed. Dietary intakes of b-carotene, a-carotene, b-cryptoxanthin, lycopene, lutein & zeaxanthin and the sum of these carotenoids were assessed using a validated food frequency questionnaire. Incident type 2 diabetes was mainly selfreported, and verified against general practitioner information. Mean SD total carotenoid intake was 10 4 mg/day. During a mean SD follow-up of 10 2years, 915 incident cases of type 2 diabetes were ascertained. After adjustment for age, sex, diabetes risk factors, dietary intake, waist circumference and BMI, higher b-carotene intakes associated inversely with diabetes risk [Hazard Ratio quartile 4 versus quartile 1 (HRQ4): 0.78 (95%CI:0.64,0.95), P-linear trend 0.01]. For a-carotene, a borderline significant reduced risk was observed, with a HRQ4 of 0.85 (95%CI:0.70,1.03), and P-linear trend 0.05. b-cryptoxanthin, lycopene, lutein & zeaxanthin, and the sum of all carotenoids did not associate with diabetes risk. Conclusions: This study shows that diets high in b-carotene and a-carotene are associated with reduced type 2 diabetes in generally healthy men and women. ª 2015 Elsevier B.V. All rights reserved.
Introduction Carotenoids are plant pigments that provide the yellow, orange and red pigments in fruits and vegetables [1]. Smaller quantities of carotenoids are also present in bread, eggs and oils [2]. Carotenoids have antioxidant functions, and may reduce type 2 diabetes risk by reducing oxidative stress, which plays an important role in the development of diabetes [3e5]. However, evidence from prospective
* Corresponding author. Julius Center for Health Sciences and Primary Care, University Medical Center Utrecht, PO Box 85500, 3508GA Utrecht, The Netherlands. Tel.: þ31887550465; fax: þ31887568099. E-mail address: [email protected] (I. Sluijs).
studies on the associations of dietary carotenoids and diabetes risk is scarce and inconsistent. So far, b-carotene has been studied most often, with two observational studies reporting an inverse association of dietary b-carotene with diabetes [6,7] and another reporting no association among male smokers [8]. In addition, studies that addressed serum or plasma b-carotene values in association with diabetes risk showed inconsistent results [6,9e11]. Randomized controlled trials with b-carotene supplementation consistently showed no effect on diabetes among women with a history of cardiovascular disease [12], generally healthy men [13] and among male smokers [14]. For dietary lycopene, three cohort studies showed no association with diabetes
http://dx.doi.org/10.1016/j.numecd.2014.12.008 0939-4753/ª 2015 Elsevier B.V. All rights reserved.
Please cite this article in press as: Sluijs I, et al., Dietary intake of carotenoids and risk of type 2 diabetes, Nutrition, Metabolism & Cardiovascular Diseases (2015), http://dx.doi.org/10.1016/j.numecd.2014.12.008
2
[7,8,15]. Very little research has been performed on other individual dietary carotenoids such as a-carotene, b-cryptoxanthin, lutein, zeaxanthin and the total of carotenoids, and therefore no conclusions can yet be drawn on such intakes [7,8]. Inconsistencies in results between studies performed thus far could be due to differences in study design, range of carotenoids intake, or population characteristics such as age and sex. Moreover, previous research suggests that smoking increases oxidative stress [16], and that smokers may need a higher intake of carotenoids in comparison with non-smokers to reduce their diabetes risk [9]. One study indeed found reduced diabetes risk with higher serum b-carotene and total carotenoid values in nonsmokers only [9], and a study among male smokers reported no associations of dietary [8] or supplementary [14] b-carotene with diabetes risk. However, two other studies did not confirm such an interaction with dietary carotenoids [7] or supplementary b-carotene [12]. In the European Prospective Investigation into Cancer and Nutrition- Netherlands (EPIC-NL) cohort, we investigated associations of dietary intake of six carotenoids and incidence of type 2 diabetes. In addition, we explored whether smoking modified these associations. Methods Study population EPIC-NL consists of the two Dutch contributions to EPIC, Prospect-EPIC and MORGEN-EPIC [17]. Both cohorts were set up simultaneously in 1993e1997 and were merged in 2007 according to standardized processes into one large Dutch EPIC cohort [17]. Prospect-EPIC includes 17,357 women aged 49e70 years at baseline, participating in the national breast cancer screening program, and living in the city of Utrecht and its surroundings [18]. The MORGEN-EPIC study includes 22,654 men and women aged 21e64 years selected from random samples of the Dutch population in three cities in the Netherlands (Amsterdam, Doetinchem, and Maastricht) [19]. All participants provided written informed consent before study inclusion. The study complied with the Declaration of Helsinki and was approved by local ethical committees. After exclusion of prevalent diabetes cases (n Z 615), participants without consent to linkage with disease registries (n Z 1027), participants with abnormal high or low energy intake (n Z 400) and missing data on carotenoid intake (n Z 218), 37,846 participants were included in the analysis. Dietary intake Participants completed a self-administered validated food frequency questionnaire (FFQ), containing questions about the usual consumption frequency of 178 foods during the year preceding enrollment. The FFQ was administrated once at baseline. The Dutch food composition table was
I. Sluijs et al.
used to calculate energy and nutrient intakes [20]. All nutrient values were adjusted for total energy intake by the regression residual method. Nutritional intakes of carotenoids examined were b-carotene, a-carotene, b-cryptoxanthin, lycopene, lutein plus zeaxanthin, and the sum of these carotenoids. The FFQ has been validated against twelve 24-h recalls [21]. Spearman correlations for carotenoids ranged from 0.16 for b-carotene to 0.31 for alpha-carotene for men, and from 0.17 for b-cryptoxanhin to 0.32 for lycopene for women (unpublished data). Energy reporting Basal metabolic rate (BMR) was estimated using the Schofield equations. Participants with an energy intake vs. BMR of 2.40 were classified as energy over-reporters, according to the Goldberg cut-offs [22]. Energy misreporters were defined as energy under-plus over-reporters. The remaining was defined as normal energy-reporters. Type 2 diabetes Ascertainment and verification of diabetes has been described elsewhere [23]. In summary, occurrence of diabetes during follow-up was obtained from self-report in two follow-up questionnaires and from linkage with hospital discharge diagnoses registries. Follow-up was complete until 1 January 2006. In the Prospect study, occurrence of diabetes was also obtained from a urinary glucose strip test for detection of glucosuria. Cases notified by any of these methods were verified against general practitioner or pharmacist information. Only cases confirmed by either the general practitioner or pharmacist were included in the analyses. Other variables At baseline, participants completed a general questionnaire containing questions on demographics, presence of chronic diseases, and chronic disease risk factors. Physical activity, assessed by validated questionnaire, was categorized using the Cambridge Physical Activity Score [24]. Blood pressure was measured twice in a supine position on the right arm using a Boso Oscillomat (Bosch & Son) (Prospect) or on the left arm using a random zero Sphygmomano-meter (MORGEN) from which the mean was taken. Hypertension was defined as: diastolic BP > 90 mm Hg, and/or systolic BP > 140 mm Hg, and/or self-report. Smoking was categorized into never, past and current. Body weight and waist circumference were measured, and BMI was calculated as weight divided by height squared (kg/m2). Data analysis Baseline characteristics were presented according to quartiles of carotenoid intake. Pearson correlations were
Please cite this article in press as: Sluijs I, et al., Dietary intake of carotenoids and risk of type 2 diabetes, Nutrition, Metabolism & Cardiovascular Diseases (2015), http://dx.doi.org/10.1016/j.numecd.2014.12.008
Risk of type 2 diabetes
3
calculated to determine correlations of carotenoids with other dietary intakes. Cox proportional hazards regression analysis was used to determine hazard ratios (HR) and 95% confidence intervals (95%CI) for the associations of carotenoid (in quartiles) and diabetes risk. P-values for linear trend were estimated by using the median carotenoid intake per quartile as a continuous variable in the regression model. In model 1, we adjusted for age (continuous) and sex. In model 2, educational status (low/middle/high), physical activity (inactive/moderately-inactive/moderately-active/ active) and systolic blood pressure (continuous) were added. In model 3, dietary intakes of fiber and vitamin E (continuous) were added. In model 4, we added waist circumference and BMI (continuous). Interactions with smoking were investigated by means of a likelihood ratio test. Former smokers and never smokers were combined into “non-smokers”. A series of sensitivity analyses were performed by excluding participants diagnosed with diabetes within the first two years of follow-up (n Z 361), excluding those with baseline chronic diseases (coronary heart disease, stroke, hyperlipidemia and/or hypertension; n Z 20,742), excluding energy misreporters (n Z 9894), and by adding vitamin C as additional confounder to the multivariable model. The analyses were performed using SPSS version 20.0. The significance level was set two sided at 0.05. Results The study population had a mean SD age of 49.1 11.9 years. The majority of participants was female (74%), and
31% were smokers (Table 1). Mean SD carotenoid intake was 10 4 mg/day, with b-carotene and lycopene as the largest contributors. Correlations between individual carotenoids ranged from 0.02 for a-carotene and lycopene to 0.87 for a-carotene and b-carotene. Correlations between carotenoids and fiber ranged from 0.06 for lycopene to 0.51 for b-cryptoxanthin. Correlations between carotenoids and vitamin C ranged from 0.24 for lutein & zeaxanthin to 0.88 for b-cryptoxanthin. Mean age and intakes of vitamin C, vitamin E and fiber increased over the quartiles of total carotenoid intake, whereas male sex, waist circumference, physical inactivity, low education, current smoking, diastolic blood pressure and percentage with hypertension decreased (Table 1). During a mean SD follow-up of 10 2 years, 915 incident cases of diabetes were identified. After adjustment for confounders (model 4), an inverse association was found for b-carotene comparing the highest vs. the lowest quartile. The corresponding HR (HRQ4) was 0.78 (95%CI:0.64,0.95), and P-for linear trend 0.01. For a-carotene, a borderline significant inverse association was found, with a HRQ4 of 0.85 (95%CI:0.70,1.03), and P-for linear trend of 0.05. b-cryptoxanthin, lycopene, lutein & zeaxanthin and the total of all carotenoids did not associate with diabetes in multivariable analysis (Table 2). Smoking did not modify associations between carotenoid intake and diabetes (P-values for interaction > 0.49). Exclusion of participants who developed diabetes within the first 2 years after the start of the study did not materially change our findings (HRQ4 b-carotene: 0.78 95% CI:0.64,0.98; a-carotene: 0.86 95%CI:0.70,1.06). After exclusion of participants diagnosed with chronic disease at baseline, all estimates remained essentially the same
Table 1 Baseline characteristics according to quartiles of total dietary carotenoid intake.a Quartiles of total carotenoid intake
Participants, n Total carotenoids, mg [median (p25, p75)] b-carotene, mg [median (p25, p75)] a-carotene, mg [median (p25, p75)] b-cryptoxanthin, mg [median (p25, p75)] Lycopene, mg [median (p25, p75)] Lutein & zeaxanthin, mg [median (p25, p75)] Energy, kcal Vitamin C, mg Vitamin E, mg Fiber, g Alcohol, g [median (p25, p75)] Age, y [median (p25, p75)] Male sex BMI, kg/m2 Waist circumference, cm Physically (moderately) inactive Low education Current smoker Systolic blood pressure, mmHg Diastolic blood pressure, mmHg Hypertension a
Q1 (low)
Q2
Q3
Q4 (high)
9461 6.2 (5.1, 6.9) 1.4 (1.1, 1.7) 0.3 (0.2, 0.4) 0.08 (0.05, 0.13) 1.4 (0.8, 2.2) 2.4 (1.8, 3.0) 2080 651 82 33 12 3 21 5 4.65 (0.54, 15.5) 51 (41, 57) 3509 (36.1) 25.7 4.0 87 12 3433 (36.3) 6927 (61.3) 3447 (36.4) 127 19 79 11 3535 (37.4)
9462 8.6 (8.1, 9.1) 2.0 (1.7, 2.4) 0.4 (0.3, 0.5) 0.11 (0.07, 0.17) 2.5 (1.5, 3.5) 3.3 (2.7, 4.0) 2084 611 101 36 12 3 23 4 4.56 (0.82, 13.9) 51 (41, 57) 2623 (27.7) 25.6 3.9 85 11 3033 (32.0) 5415 (57.2) 2840 (30.0) 126 19 78 11 3443 (36.4)
9462 10.7 (10.2, 11.3) 2.6 (2.1, 3.0) 0.5 (0.3, 0.8) 0.13 (0.08, 0.20) 3.2 (2.1, 4.4) 4.1 (3.3, 4.9) 2063 587 116 41 12 3 24 4 4.66 (0.89, 13.7) 51 (42, 58) 2114 (22.3) 25.6 3.9 85 11 2902 (30.6) 5328 (56.3) 2678 (28.3) 126 19 78 11 3442 (36.4)
9461 14.2 (13.0, 16.1) 3.5 (2.8, 4.3) 0.7 (0.4, 1.1) 0.17 (0.10, 0.24) 4.4 (2.8, 6.4) 5.3 (4.0, 6.7) 1984 576 138 50 13 3 26 5 4.65 (0.75, 14.4) 52 (43, 58) 1438 (15.2) 25.7 4.0 84 11 2800 (29.6) 5056 (53.5) 2569 (27.2) 126 19 77 11 3521 (37.2)
Values are mean SD or n (percentage), unless stated otherwise; all nutrients are adjusted for total energy intake.
Please cite this article in press as: Sluijs I, et al., Dietary intake of carotenoids and risk of type 2 diabetes, Nutrition, Metabolism & Cardiovascular Diseases (2015), http://dx.doi.org/10.1016/j.numecd.2014.12.008
4
I. Sluijs et al.
Table 2 Associations between quartiles of dietary carotenoid intake and risk of type 2 diabetes.a Quartiles of intake
Total carotenoids, median intake, mg/d Cases/at risk, n Model 1b Model 2c Model 3d Model 4e b-carotene, median intake, mg/d Cases/at risk, n Model 1b Model 2c Model 3d Model 4e a-carotene, median intake, mg/d Cases/at risk, n Model 1b Model 2c Model 3d Model 4e b-cryptoxanthin, median intake, mg/d Cases/at risk, n Model 1b Model 2c Model 3d Model 4e Lycopene, median intake, mg/d Cases/at risk, n Model 1b Model 2c Model 3d Model 4e Lutein & Zeaxanthin, median intake, mg/d Cases/at risk, n Model 1b Model 2c Model 3d Model 4e
P-linear trend
Q1 (low)
Q2
Q3
Q4 (high)
6.2 253/9461 1 1 1 1 1.3 223/9461 1 1 1 1 0.2 260/9460 1 1 1 1 0.05 235/9460 1 1 1 1 0.9 312/9460 1 1 1 1 2.1 198/9460 1 1 1 1
8.6 225/9462 0.93 (0.78,1.11) 0.93 (0.78,1.12) 0.92 (0.77,1.10) 0.91 (0.76,1.10) 1.9 220/9462 0.90 (0.75,1.09) 0.96 (0.79,1.15) 0.93 (0.77,1.13) 0.93 (0.77,1.13) 0.4 206/9462 0.89 (0.74,1.07) 0.94 (0.79,1.13) 0.93 (0.77,1.12) 0.99 (0.82,1.19) 0.10 210/9462 0.88 (0.73, 1.06) 0.95 (0.79, 1.15) 0.93 (0.77, 1.13) 0.93 (0.77, 1.12) 2.6 266/9462 0.88 (0.74, 1.04) 0.94 (0.79, 1.11) 0.92 (0.78, 1.09) 0.90 (0.76, 1.07) 3.1 200/9462 1.01 (0.83, 1.23) 1.02 (0.84, 1.25) 1.02 (0.83, 1.24) 1.02 (0.83, 1.24)
10.7 209/9462 0.88 (0.73,1.05) 0.89 (0.74,1.07) 0.86 (0.71,1.04) 0.84 (0.69,1.02) 2.6 220/9462 0.84 (0.70,1.01) 0.89 (0.74,1.08) 0.86 (0.71,1.04) 0.86 (0.70,1.04) 0.6 217/9464 0.77 (0.64,0.93)* 0.84 (0.69,1.01) 0.81 (0.67,0.98)* 0.85 (0.70,1.02) 0.15 224/9464 0.82 (0.68, 0.99)* 0.93 (0.77,1.12) 0.91 (0.74,1.10) 0.96 (0.79, e 1.17) 3.3 188/9464 0.86 (0.72, 1.04) 0.96 (0.80, 1.15) 0.93 (0.78, 1.12) 0.97 (0.81, 1.16) 4.1 222/9464 1.01 (0.83, 1.23) 1.00 (0.82, 1.22) 0.98 (0.80, 1.20) 0.89 (0.73, 1.09)
14.2 228/9461 0.97 (0.80,1.16) 0.99 (0.82,1.19) 0.92 (0.76,1.13) 0.85 (0.69,1.04) 3.7 252/9461 0.85 (0.70, 1.02) 0.91 (0.76,1.10) 0.85 (0.69,1.04) 0.78 (0.64,0.95)* 0.9 232/9460 0.80 (0.67,0.96)* 0.87 (0.73,1.05) 0.84 (0.69,1.01) 0.85 (0.70,1.03) 0.24 246/9460 0.83 (0.69, 1.00)* 0.99 (0.82, 1.19) 0.96 (0.78, 1.18) 0.96 (0.78, 1.19) 5.3 149/9460 0.79 (0.65, 0.96)* 0.94 (0.78, 1.15) 0.92 (0.75, 1.12) 0.97 (0.79, 1.18) 5.9 295/9460 1.11 (0.92, 1.34) 1.08 (0.89, 1.30) 1.04 (0.85, 1.27) 0.91 (0.74, 1.11)
0.19 0.81 0.38 0.09
0.09 0.32 0.10 0.01
0.02 0.13 0.06 0.05
0.07 0.97 0.77 0.89
0.01 0.55 0.36 0.73
0.23 0.42 0.72 0.22
*p 0.05 for difference from quartile 1. a Values are hazard ratio’s (95%CI) derived from cox regression; all carotenoid intakes were adjusted for total energy intake. b Adjusted for age and sex. c Adjusted for model 1 þ educational status, physical activity and systolic blood pressure. d Adjusted for model 2 þ fiber intake and vitamin E intake. e Adjusted for model 3 þ BMI and waist circumference.
(HRQ4 b-carotene: 0.72 95%CI:0.46,1.12; a-carotene: 0.74 95%CI:0.49,1.11). Excluding energy misreporters did not materially change our findings either (HRQ4 b-carotene: 0.84 95%CI:0.65,1.09; a-carotene: 0.79 95%CI:0.61,1.02) In addition, after adding vitamin C to model 4, associations did not materially change (HRQ4 b-carotene: 0.78 95% CI:0.63,0.95; a-carotene: 0.85 95%CI:0.70,1.03).
Discussion In this prospective study among 37,846 men and women followed-up for a mean of 10 years, higher dietary intakes of b-carotene and a-carotene were associated with a reduced diabetes risk, whereas dietary intakes of other individual carotenoids were not associated with diabetes risk. The total of individual carotenoids tended to be
associated with reduced diabetes, similar to a-carotene and b-carotene, but this was not statistically significant. The associations were not modified by smoking status. Strengths of this study include the large size of the population, long follow-up time and prospective design. Additionally, in this study only cases of diabetes that were confirmed to be diagnosed with diabetes by their general practitioner were included in the analyses. This study has several limitations. First of all, participants with a higher carotenoid intake may have an overall healthier lifestyle than participants with a lower intake of carotenoids. Although we did adjust for several lifestyle factors such as physical activity and fiber intake, we cannot exclude that residual confounding related to a healthier lifestyle may still exist. Second, although our study design allows us to draw conclusions on associations of dietary carotenoids and diabetes, it does not allow us to make causal
Please cite this article in press as: Sluijs I, et al., Dietary intake of carotenoids and risk of type 2 diabetes, Nutrition, Metabolism & Cardiovascular Diseases (2015), http://dx.doi.org/10.1016/j.numecd.2014.12.008
Risk of type 2 diabetes
inferences. Third, presence of type 2 diabetes goes often undetected, and may be preclinical up to 9e12 years [25]. Undetected diabetes cases may have been misclassified as non-diabetic individuals, resulting in attenuation of associations. Fourth, we used a FFQ to measure carotenoid intake, which in general is known to be prone to imprecision and reporting bias [26]. Correlation coefficients between the FFQ and repeated 24 h recalls for measurement of carotenoid intake were moderate. Random variation in reporting intake of carotenoid-containing foods is expected to underlie the moderate correlation coefficients. This is supported by the finding that excluding energy misreporters did not materially change our findings. Such random variation causes bias toward the null and can only have attenuated our associations. Still, we need to be careful with drawing conclusions from our study. Finally, although the food composition table used in our study provides detailed information on carotenoid content of foods, 2% (for beta-carotene) to 29% (for zeaxanthin) of fruits and vegetables listed in the food composition table were not assigned carotenoid values. We therefore cannot exclude that we underestimated associations due to such missing information. The inverse association of dietary b-carotene intake and diabetes is consistent with two smaller scale observational studies [6,7]: One study among 846 Swedish men, found a 29% reduced diabetes risk in the highest quartile of bcarotene intake (>1.9 mg/d) compared with the lowest, during 7 years follow-up [6]. Another study, including 4304 adults, found a borderline significant 26% reduced diabetes risk in the highest (>2.1 mg/d) vs. lowest quartile of b-carotene intake [7]. We found a 22% reduced risk in the highest b-carotene intake quartile, which is only slightly lower than the estimates in those studies. In contrast, a cohort study including 29,000 smokers, with similar b-carotene and higher a-carotene intake as in our study, did neither find associations with b-carotene nor with a-carotene [8]. In contrast with our study, the study population consisted of men only and had a higher mean age. Another study also did not find associations of dietary a-carotene and diabetes [7], but a-carotene intakes in that study were much lower compared with our study. Studies that investigated b-carotene in serum or plasma have shown inconsistent results, with some reporting inverse [6,9] and others reporting no associations [10,11] with diabetes risk. Serum or plasma a-carotene did not associate with diabetes risk in previous studies [9,11]. Although serum and plasma carotenoid values reflect nutritional status, they do not directly relate to intake. In fact, serum and plasma carotenoid values are a reflection of many more factors than intake, including absorption and metabolism [27]. Therefore, the comparability of studies examining serum or plasma and intake values is limited. Several randomized controlled trials with b-carotene supplementation have been conducted [12e14], with doses of supplementation ranging from 20 mg/d to 50 mg every other day. None of the trials found significant effects on diabetes [12e14]. The discrepancy with the observational studies may be due to a higher intake of
5
b-carotene in the trials as a consequence of supplementation. The ATBC trial was the first to show increased mortality with b-carotene supplementation [28], later confirmed in a large meta-analysis [29]. This thus suggests that b-carotene may even become harmful at higher doses. Differences in bioavailability between natural and synthetic forms of b-carotene may also underlie the differences between the intervention and observational studies [30]. However, residual confounding should also be considered. We found no association of b-cryptoxanthin and lutein plus zeaxanthin with diabetes, whereas another prospective study did suggest inverse associations for those carotenoids [7]. That study completed their food database with analyzed carotenoid values from national foods, which may explain the difference with our study. Our finding of no association for dietary lycopene with diabetes is consistent with previous research [7,8,15]. In our study we only found associations for some individual carotenoids. For the total of individual carotenoids we found a tendency towards a reduced diabetes risk, similar to a-carotene and b-carotene, but this was not statistically significant. This is likely due to the lack of association for the other individual carotenoids such as lycopene and b-cryptoxanthin. One observational study with 4304 adults did find an inverse association between total carotenoids and diabetes [7]. We found no interactions with smoking, which is consistent with previously conducted studies on dietary carotenoids and diabetes [7,12]. Carotenoids are known to have antioxidant functions, which may underlie the observed inverse associations with diabetes. It has been suggested that antioxidants such as carotenoids might be effective in reducing diabetes by reducing oxidative stress [4]. Of all carotenoids addressed in this study, b-carotene is known to be a strong antioxidant [1], which may explain the observed association with diabetes in our study. We also found an inverse association with a-carotene, which could possibly be explained by the high correlation between b-carotene and a-carotene. Lycopene also has high antioxidant properties [1]. Even though lycopene was a major contributor to total carotenoid intake, we did not find an association of lycopene with diabetes risk. It is unclear why lycopene shows no association with diabetes in the present study. Relative validity for lycopene was higher than for b-carotene, the bioavailability for lycopene and b-carotene are comparable [1], and the range of intake for lycopene was wider than for b-carotene. Therefore this unlikely explains why we were unable to find an association of lycopene with diabetes. The higher completeness of the food database for bcarotene values of foods (92% for vegetables; 98% for fruits) than for lycopene values of foods (79% for both vegetables and fruits) may partly explain these differences. In conclusion, this study shows that diets high in bcarotene and a-carotene are associated with reduced type 2 diabetes in generally healthy men and women. The association does not differ by smoking status.
Please cite this article in press as: Sluijs I, et al., Dietary intake of carotenoids and risk of type 2 diabetes, Nutrition, Metabolism & Cardiovascular Diseases (2015), http://dx.doi.org/10.1016/j.numecd.2014.12.008
6
I. Sluijs et al.
Conflict of interest None of the authors declared a conflict of interest. Acknowledgments The EPIC-NL study was funded by ‘European Commission: Public Health and Consumer Protection Directorate 1993e2004; Research Directorate-General 2005’; Dutch Ministry of Public Health, Welfare and Sports (VWS), Netherlands Cancer Registry (NKR), Dutch Prevention Funds, Dutch ZON (Zorg Onderzoek Nederland), World Cancer Research Fund (WCRF) (The Netherlands).
[14]
[15]
[16] [17]
[18]
References [19] [1] Perera C, Yen G. Functional properties of carotenoids in human health. Int J Food Prop 2007 Apr 27;10:201e30. [2] Rao AV, Rao LG. Carotenoids and human health. Pharmacol Res 2007 Mar;55(3):207e16. [3] Ceriello A, Motz E. Is oxidative stress the pathogenic mechanism underlying insulin resistance, diabetes, and cardiovascular disease? the common soil hypothesis revisited. Arterioscler Thromb Vasc Biol 2004 May;24(5):816e23. [4] Johansen JS, Harris AK, Rychly DJ, Ergul A. Oxidative stress and the use of antioxidants in diabetes: linking basic science to clinical practice. Cardiovasc Diabetol 2005;4(1):5. [5] Psaltopoulou T, Ilias I, Alevizaki M. The role of diet and lifestyle in primary, secondary, and tertiary diabetes prevention: a review of meta-analyses. Rev Diabet Stud 2010;7(1):26e35. [6] Arnlov J, Zethelius B, Riserus U, Basu S, Berne C, Vessby B, et al. Serum and dietary beta-carotene and alpha-tocopherol and incidence of type 2 diabetes mellitus in a community-based study of Swedish men: report from the Uppsala Longitudinal Study of Adult Men (ULSAM) study. Diabetologia 2009 Jan;52(1):97e105. [7] Montonen J, Knekt P, Jarvinen R, Reunanen A. Dietary antioxidant intake and risk of type 2 diabetes. Diabetes Care 2004 Feb;27(2): 362e6. [8] Kataja-Tuomola MK, Kontto JP, Mannisto S, Albanes D, Virtamo J. Intake of antioxidants and risk of type 2 diabetes in a cohort of male smokers. Eur J Clin Nutr 2011 May;65(5):590e7. [9] Hozawa A, Jacobs DRJ, Steffes MW, Gross MD, Steffen LM, Lee DH. Associations of serum carotenoid concentrations with the development of diabetes and with insulin concentration: interaction with smoking: the Coronary Artery Risk Development in Young Adults (CARDIA) Study. Am J Epidemiol 2006 May 15;163(10): 929e37. [10] Reunanen A, Knekt P, Aaran RK, Aromaa A. Serum antioxidants and risk of non-insulin dependent diabetes mellitus. Eur J Clin Nutr 1998 Feb;52(2):89e93. [11] Wang L, Liu S, Pradhan AD, Manson JE, Buring JE, Gaziano JM, et al. Plasma lycopene, other carotenoids, and the risk of type 2 diabetes in women. Am J Epidemiol 2006 Sep 15;164(6):576e85. [12] Song Y, Cook NR, Albert CM, Van Denburgh M, Manson JE. Effects of vitamins C and E and beta-carotene on the risk of type 2 diabetes in women at high risk of cardiovascular disease: a randomized controlled trial. Am J Clin Nutr 2009 Aug;90(2):429e37. [13] Liu S, Ajani U, Chae C, Hennekens C, Buring JE, Manson JE. Longterm beta-carotene supplementation and risk of type 2 diabetes
[20] [21]
[22]
[23]
[24]
[25]
[26]
[27] [28]
[29]
[30]
mellitus: a randomized controlled trial. J Am Med Assoc 1999 Sep 15;282(11):1073e5. Kataja-Tuomola M, Sundell JR, Mannisto S, Virtanen MJ, Kontto J, Albanes D, et al. Effect of alpha-tocopherol and beta-carotene supplementation on the incidence of type 2 diabetes. Diabetologia 2008 Jan;51(1):47e53. Wang L, Liu S, Manson JE, Gaziano JM, Buring JE, Sesso HD. The consumption of lycopene and tomato-based food products is not associated with the risk of type 2 diabetes in women. J Nutr 2006 Mar;136(3):620e5. Isik B, Ceylan A, Isik R. Oxidative stress in smokers and nonsmokers. Inhal Toxicol 2007 Jul;19(9):767e9. Beulens JWJ, Monninkhof EM, Verschuren WMM, van der Schouw YT, Smit J, Ocke MC, et al. Cohort profile: the EPIC-NL study. Int J Epidemiol 2010 Oct;39(5):1170e8. Boker LK, van Noord PA, van der Schouw YT, Koot NV, Bueno de Mesquita HB, Riboli E, et al. Prospect-EPIC Utrecht: study design and characteristics of the cohort population. European prospective investigation into cancer and nutrition. Eur J Epidemiol 2001; 17(11):1047e53. Verschuren WMM, Blokstra A, Picavet HSJ, Smit HA. Cohort profile: the doetinchem cohort study. Int J Epidemiol 2008 Dec;37(6): 1236e41. Dutch food composition table. The Hague: NEVO Foundation; 2006. Ocke MC, Bueno-de-Mesquita HB, Goddijn HE, Jansen A, Pols MA, van Staveren WA, et al. The Dutch EPIC food frequency questionnaire. I. Description of the questionnaire, and relative validity and reproducibility for food groups. Int J Epidemiol 1997;26(Suppl. 1): S37e48. Goldberg GR, Black AE, Jebb SA, Cole TJ, Murgatroyd PR, Coward WA, et al. Critical evaluation of energy intake data using fundamental principles of energy physiology: 1. Derivation of cutoff limits to identify under-recording. Eur J Clin Nutr 1991 Dec; 45(12):569e81. Sluijs I, van der A DL, Beulens JWJ, Spijkerman AMW, Ros MM, Grobbee DE, et al. Ascertainment and verification of diabetes in the EPIC-NL study. Neth J Med 2010 Aug;68(1):333e9. Wareham NJ, Jakes RW, Rennie KL, Schuit J, Mitchell J, Hennings S, et al. Validity and repeatability of a simple index derived from the short physical activity questionnaire used in the European Prospective Investigation into Cancer and Nutrition (EPIC) study. Public Health Nutr 2003 Jun;6(4):407e13. Harris MI, Klein R, Welborn TA, Knuiman MW. Onset of NIDDM occurs at least 4e7 yr before clinical diagnosis. Diabetes Care 1992 Jul;15(7):815e9. Margetts BM, Pietinen P. European Prospective Investigation into Cancer and Nutrition: validity studies on dietary assessment methods. Int J Epidemiol 1997;26(Suppl. 1):S1e5. Potischman N. Biologic and methodologic issues for nutritional biomarkers. J Nutr 2003 Mar;133(Suppl. 3):875Se80S. The effect of vitamin E and beta carotene on the incidence of lung cancer and other cancers in male smokers. The alpha-tocopherol, beta carotene cancer prevention study group. N Engl J Med 1994 Apr 14;330(15):1029e35. Vivekananthan DP, Penn MS, Sapp SK, Hsu A, Topol EJ. Use of antioxidant vitamins for the prevention of cardiovascular disease: meta-analysis of randomised trials. Lancet 2003 Jun 14;361(9374): 2017e23. Ben-Amotz A, Levy Y. Bioavailability of a natural isomer mixture compared with synthetic all-trans beta-carotene in human serum. Am J Clin Nutr 1996 May;63(5):729e34.
Please cite this article in press as: Sluijs I, et al., Dietary intake of carotenoids and risk of type 2 diabetes, Nutrition, Metabolism & Cardiovascular Diseases (2015), http://dx.doi.org/10.1016/j.numecd.2014.12.008
