- Email: [email protected]
Dietary intake of flavonoid subclasses and risk of type 2 diabetes in prospective cohort studies: A dose–response meta-analysis
Clinical Nutrition xxx (2018) 1e5
Contents lists available at ScienceDirect
Clinical Nutrition journal homepage: http://www.elsevier.com/locate/clnu
Letter to the Editor
Dietary intake of flavonoid subclasses and risk of type 2 diabetes in prospective cohort studies: A doseeresponse meta-analysis s u m m a r y Keywords: Type 2 diabetes Flavonoids Doseeresponse analysis
Flavonoids intake has been linked to risk of type 2 diabetes theoretically, but the findings were conflicting from observational studies. Results from the recent meta-analysis suggested a moderate favorable effect of total dietary flavonoids consumption on type 2 diabetes risk. However, the relationship, if exists, between total dietary flavonoids consumption and type 2 diabetes risk is unclear, especially further in the association between flavonoid subclasses and risk of type 2 diabetes. Thus, the doseeresponse relationship was assessed by restricted cubic spline model and multivariate random-effect metaregression. Among flavonoid subclasses, protective effects were obtained for intakes of flavonols, flavan3-ols, isoflavones and anthocyanidins in high vs. low analysis and doseeresponse analysis. © 2018 Elsevier Ltd and European Society for Clinical Nutrition and Metabolism. All rights reserved.
Dear editor, We read with great interest the article by Liu et al. [1] entitled “Dietary flavonoids intake and risk of type 2 diabetes: A metaanalysis of prospective cohort studies”. The study suggested that an inverse relation was found of total dietary flavonoids with type 2 diabetes. This is a very important investigation considering the widespread consumption of flavonoids and high incidence of type 2 diabetes. Thus, it is meaningful regarding the exploration of the causes and control of type 2 diabetes. However, we would like to 1) explore the association between flavonoid subclasses and type 2 diabetes risk for the study conducted by Liu et al. [1] only examined the relationship between total flavonoids intake; 2) draw attention to the type of doseeresponse analysis between total flavonoids and flavonoid subclasses consumption with risk of type 2 diabetes, because Liu et al. [1] assumed a linear dose relationship without performing a formal test; 3) reveal a more robust and wide association between flavonoid subclasses consumption and risk of type 2 diabetes because only three studies in USA and one study in Finland performed by Liu et al. [1]. Moreover, categories of flavonoids consumption per day differed between studies, which might complicate the interpretation of the pooled results across study populations with different categories. In this respect, a doseeresponse meta-analysis provides a solution to the problem, from which a summary risk estimate can be derived for a standardized increase and specific exposure values considering flavonoid consumption per day. Studies included in this doseeresponse analysis should provide the following information: the number of cases and participants, and relative risk (RR) estimates with 95% confidence intervals (CIs) for each category of flavonoid subclasses were extracted. We extracted data on the amount of flavonoids/subclasses intake, distributions of cases and person-years (when available), and hazard
ratios with 95% confidence intervals for 3 exposure categories. The median or mean daily take of flavonoids in each category was assigned to the corresponding hazard ratio with the 95%CI for each study. Detailed information of the included publications is shown in Table 1. Two-stage random-effects doseeresponse meta-analysis was performed to examine linear and nonlinear relationships between flavonoids intake and type 2 diabetes. In the first stage, the method reported by Greenland and Orsini [2] (generalized least-squares) was used to calculate study-specific coefficients on the basis of results across categories of flavonoids intake taking into account the correlation within each set of retrieved risk ratios. Nonlinear doseeresponse analysis was modeled by using restricted cubic splines with 3 knots at fixed percentiles [3] (25th, 50th, and 75th) of the distribution. We combined the coefficients that had been estimated within each study by performing random-effects meta-analysis. In linear doseeresponse meta-analysis, the method of Der-Simonian and Laird was used, and in nonlinear dose response meta-analysis, the multivariate extension of the method of moments was used. We calculated an overall P value by testing that the 2 regression coefficients were simultaneously equal to zero. We then calculated a P value for nonlinearity by testing that the coefficient of the second spline was equal to zero. All statistical analyses were performed with STATA version 12 (Stata Corporation, College Station, Texas, USA). All reported probabilities (p values) were two-sided, with p < 0.05 considered statistically significant. 1. Total diet flavonoids Nine studies [4e12] including 12 results investigated the association of total flavonoids with incidence of type 2 diabetes. Comparing the highest with the lowest intake, the dietary intake of flavonoids significantly reduced the type 2 diabetes risk by
https://doi.org/10.1016/j.clnu.2018.08.024 0261-5614/© 2018 Elsevier Ltd and European Society for Clinical Nutrition and Metabolism. All rights reserved.
Please cite this article in press as: Zhou Y, et al., Dietary intake of flavonoid subclasses and risk of type 2 diabetes in prospective cohort studies: A doseeresponse meta-analysis, Clinical Nutrition (2018), https://doi.org/10.1016/j.clnu.2018.08.024
2
First author (year of publication)
Country
Study size
Sex
Measure of intake
Exposure
RR (95%CI)
Adjustment for Covariates
Quality assessment
Zamora-Ros R et al. [12]
European countries
15258/729
F/M
A self-administered questionnaire of 98e266 food items
8
Spain
7447/314
F/M
a 137-item semiquantitative FFQ
0.90 0.73 0.81 1.03 0.98 0.89 0.94 0.67 0.97 0.69 0.98 0.88
Age, sex, and total energy
Tresserra-Rimbau A et al. [10]
Flavonoids Flavan-3-ols Flavonols Flavanone Isoflavones Flavone Anthocyanidin Flavonoids Flavonols Flavanone Flavone Anthocyanidin
6
Grosso G et al. [4]
Poland
5806/456
F/M
a self-administered machinereadable 148-item FFQ
Quansah DY et al. [8]
Korea
24377/603
F/M
3193 food items
Flavonoids Flavonols Flavanone Isoflavones Flavone Anthocyanidin Flavonoids
Flavonoids
0.44 (0.33e0.63) 0.89 (0.63e1.26) 0.75 (0.54e1.03) 0.93 (0.65e1.34) 0.47 (0.33e0.68) 0.68 (0.48e0.98) M:0.75 (0.49e1.15) F:0.97 (0.60e1.57) M:0.67 (0.44e1.01) F:0.88 (0.54e1.42) M:0.79 (0.52e1.19) F:1.07 (0.69e1.66) M:0.96 (0.64e1.45) F:0.91 (0.58e1.42) M:0.95 (0.63e1.42) F:1.35 (0.82e2.24) M:0.83 (0.57e1.23) F:0.67 (0.43e1.04) M:1.03 (0.67e1.56) F:0.83 (0.53e1.28) 0.98 (0.77e1.24)
Age, sex, recruitment center, and intervention group, smoking, BMI, physical activity, dyslipidemia, hypertension, education level, total energy intake, alcohol intake, adherence to the Mediterranean diet, and fasting glucose concentrations at baseline. Age, total energy intake, BMI, physical activity, educational status, smoking status, alcohol consumption, alcohol intake, menopausal status and dietary fiber.
Flavonoids
0.92 (0.78e1.09)
Flavan-3-ols Flavonols Flavanone Isoflavones Flavone Anthocyanidin a dietary history questionnaire of 100 items 131-item semiquantitative FFQ
(0.77e1.04) (0.57e0.93) (0.69e0.95) (0.79e1.34) (0.77e1.25) (0.70e1.14) (0.82e1.08) (0.48e0.93) (0.68e1.39) (0.49e0.97) (0.71e1.35) (0.62e1.24)
Knekt P et al. [6]
Finland
10054/526
F/M
Song Y et al. [9]
USA
38018/1614
F
Wedick NM et al. [11](NHS)
USA
70359/6878
F
a comprehensive FFQ with 118e131 food items
Flavonoids Flavan-3-ols Flavonols Flavanone Flavone Anthocyanidin
0.85 0.87 0.84 1.05 1.07 0.83
(0.79e0.92) (0.81e0.94) (0.78e0.91) (0.97e1.13) (0.99e1.16) (0.77e0.90)
Wedick NM et al. [11](NHS II)
USA
89201/3084
F
a comprehensive FFQ with 118e131 food items
Flavonoids Flavan-3-ols Flavonols Flavanone Flavone Anthocyanidin
0.99 1.01 0.99 1.08 1.02 0.83
(0.89e1.11) (0.91e1.12) (0.89e1.10) (0.97e1.22) (0.91e1.16) (0.73e0.94)
6
Age, BMI, current smoking, education, household income, physical activity, and regular alcohol consumption
7
sex and age
6
age, BMI, total energy intake, smoking, exercise, alcohol use, history of hypertension, history of high cholesterol, family history of diabetes, fiber intake, glycemic load, magnesium, and total fat. Age, BMI, smoking status alcohol intake, multivitamin use, physical activity, a family history of diabetes, postmenopausal status and hormone use, ethnicity, total energy, polyunsaturated:saturated fat ratio, intakes of red meat, fish, whole grains, coffee, high-calorie sodas, and trans fat. Age, BMI, smoking status alcohol intake, multivitamin use, physical activity, a family history of diabetes, postmenopausal status and hormone use, oral contraceptive use, ethnicity, total energy, polyunsaturated:saturated fat ratio, intakes of red meat, fish, whole grains, coffee, high-calorie sodas, and trans fat.
7
8
8
Letter to the Editor / Clinical Nutrition xxx (2018) 1e5
Please cite this article in press as: Zhou Y, et al., Dietary intake of flavonoid subclasses and risk of type 2 diabetes in prospective cohort studies: A doseeresponse meta-analysis, Clinical Nutrition (2018), https://doi.org/10.1016/j.clnu.2018.08.024
Table 1 Characteristics of studies on dietary flavonoids and incidence of type 2 diabetes.
First author (year of publication)
Country
Study size
Sex
Measure of intake
Exposure
RR (95%CI)
Adjustment for Covariates
Quality assessment
Wedick NM et al. [11](HPFS)
USA
41334/2649
M
a comprehensive FFQ with 118e131 food items
USA
35816/3395
F
a 127-item semiquantitative FFQ
0.92 0.88 0.88 1.09 1.07 0.93 0.97
(0.81e1.04) (0.78e0.99) (0.78e1.00) (0.96e1.24) (0.94e1.22) (0.81e1.05) (0.86e1.10)
Jacques PF et al. [5]
USA
2915/308
F/M
the Harvard semiquantitative FFQ
Ding M et al. [13] (NHS)
USA
63115/4519
F
a validated 116-item FFQ
Flavonoids Flavan-3-ols Flavonols Flavanone Flavone Anthocyanidin Isoflavones
0.89 0.90 0.68 1.08 1.02 0.83 0.97
(0.75e1.05) (0.80e1.01) (0.54e0.86) (0.97e1.22) (0.91e1.16) (0.73e0.94) (0.88e1.07)
Age, BMI, smoking status alcohol intake, multivitamin use, physical activity, a family history of diabetes, ethnicity, total energy, polyunsaturated:saturated fat ratio, intakes of red meat, fish, whole grains, coffee, high-calorie sodas, and trans fat. age, energy, education level, BMI, waist:hip ratio, activity level, smoking status, multivitamin use, and hormone therapy. sex and time dependent variables age, cardiovascular disease, current smoker, BMI, vegetable, fruit intake and cumulative mean energy intake.
8
Nettleton JA et al. [7]
Flavonoids Flavan-3-ols Flavonols Flavanone Flavone Anthocyanidin Flavonoids
Ding M et al. [13] NHS II)
USA
79061/3920
F
a validated 116-item FFQ
Isoflavones
0.85 (0.76e0.95)
Ding M et al. [13](HPFS)
USA
21281/742
M
a validated 116-item FFQ
Isoflavones
0.80 (0.62e1.02)
Japan
59791/1114
a 147-item FFQ
Isoflavones
M:0.97 (0.79e1.20) F:0.91 (0.71e1.17)
Korea Finland
7963/225 25514/660
a 1549 food item a validated self-administered 276-item FFQ
Flavone Flavone
0.59 (0.41e0.85) 1.03 (0.91e1.06)
age, race, family history of T2D, baseline disease status, BMI, physical activity, overall dietary pattern, total energy intake, smoking status, menopausal status, and postmenopausal hormone use age, race, family history of T2D, baseline disease status, BMI, physical activity, overall dietary pattern, total energy intake, smoking status, menopausal status, and postmenopausal hormone use age, race, family history of T2D, baseline disease status, BMI, physical activity, overall dietary pattern, total energy intake, smoking status age, study area, BMI, smoking habit, alcohol consumption, family history of diabetes mellitus, leisure time physical activity, history of hypertension, coffee, green tea, magnesium, calcium, vegetable, fiber intake, fish, and total energy intake age, BMI, education, menopausal status age, supplementation,BMI, cigarettes smoked daily, smoking years, blood pressure, total cholesterol, high-density lipoprotein cholesterol, leisure-time physical activity and daily intake of alcohol and energy
7
7
7
7
7
6
6 7
Letter to the Editor / Clinical Nutrition xxx (2018) 1e5
Please cite this article in press as: Zhou Y, et al., Dietary intake of flavonoid subclasses and risk of type 2 diabetes in prospective cohort studies: A doseeresponse meta-analysis, Clinical Nutrition (2018), https://doi.org/10.1016/j.clnu.2018.08.024
Table 1 (continued )
F, Female; M, Male; FFQ, food frequency questionnaire; BMI, body mass index; NHS, Nurses' Health Study; NHS II, Nurses' Health Study II; HPFS, Health Professionals Follow-Up Study.
3
4
Letter to the Editor / Clinical Nutrition xxx (2018) 1e5
11.00% (RR ¼ 0.89; 95%CI: 0.82e0.95) with considerable heterogeneity (Pheterogeneity ¼ 0.01, I2 ¼ 56.40%). Five studies (seven comparisons [4,7,10e12]) were eligible for the doseeresponse analysis. The departure from linear relationship was significant (P < 0.05). The RR (95%CI) of type 2 diabetes was 0.97 (0.94, 0.99), 0.96 (0.92, 0.99), 0.95 (0.91, 0.99), 0.95 (0.91, 1.00) and 0.93 (0.89, 0.97) for total flavonoids intake of 100, 200, 300, 400 and 500 mg/d, respectively (Fig. 1). The risk of type 2 diabetes decreased by 2% (RR ¼ 0.98; 95%CI 0.97, 0.99) for every 100 mg/d increment in total flavonoids intake. In the analysis of flavonoid subclasses with risk of type 2 diabetes, the beneficial effects were only observed in flavonols, flavan-3-ols, anthocyanins and isoflavones not found in flavones and flavanones. Besides, the studies in flavan-3-ols eligible for the doseeresponse analysis were limited.
2. Flavonols The summary risk estimate was 0.87 (95% CI, 0.83e0.91), with little evidence of heterogeneity (I2 ¼ 38.7%, Pheterogeneity ¼ 0.11).
Four studies with six cohorts [4,10e12] reporting flavonols intake were eligible for the doseeresponse analysis. The departure from linear relationship was significant (Pnon-linearity ¼ 0.02). The RR (95%CI) of type 2 diabetes was 0.95(0.91, 0.98), 0.89(0.86, 0.93), 0.86(0.81, 0.90), 0.81 (0.75, 0.88) and 0.83 (0.72, 0.96) for flavonols intake of 10, 20, 30, 50 and 100 mg/d, respectively (Fig. 2). The risk of type 2 diabetes decreased by 3% (RR ¼ 0.97; 95%CI 0.95, 0.99) for every 100 mg/d increment in flavonols intake. 3. Anthocyanins The summary risk estimate was 0.89 (95%CI, 0.84e0.93), with no evidence of heterogeneity (I2 ¼ 28.5%, Pheterogeneity ¼ 0.19). Four studies with six cohorts [4,10e12] reporting anthocyanins intake were eligible for the doseeresponse analysis. The departure from linear relationship was significant (Pnon-linearity ¼ 0.001). The RR (95%CI) of type 2 diabetes was 0.93(0.90, 0.96), 0.90(0.86, 0.94), 0.88(0.85, 0.92), 0.88(0.84, 0.91) and 0.86 (0.82, 0.91) for anthocyanins intake of 5, 8, 13, 20 and 30 mg/d, respectively (Fig. 3). A
Fig. 1. Nonlinear doseeresponse relationship between total flavonoid intake and type 2 diabetes risk in cohort studies assessed by restricted cubic spline model with three knots.
Fig. 3. Nonlinear doseeresponse relationship between anthocyanins intake and type 2 diabetes risk in cohort studies assessed by restricted cubic spline model with three knots.
Fig. 2. Nonlinear doseeresponse relationship between flavonols intake and type 2 diabetes risk in cohort studies assessed by restricted cubic spline model with three knots.
Fig. 4. Nonlinear doseeresponse relationship between isoflavones intake and type 2 diabetes risk in cohort studies assessed by restricted cubic spline model with three knots.
Please cite this article in press as: Zhou Y, et al., Dietary intake of flavonoid subclasses and risk of type 2 diabetes in prospective cohort studies: A doseeresponse meta-analysis, Clinical Nutrition (2018), https://doi.org/10.1016/j.clnu.2018.08.024
Letter to the Editor / Clinical Nutrition xxx (2018) 1e5
2 mg/day increment in intake of anthocyanins was associated with a risk ratio of 0.99 (95% CI: 0.98, 1.00). 4. Isoflavones The summary risk estimate was 0.92 (95% CI, 0.87e0.98), with some evidence of heterogeneity (I2 ¼ 0.00%, Pheterogeneity ¼ 0.53). Three studies with five cohorts [4,12,13] reporting isoflavones intake were eligible for the doseeresponse analysis. The departure from linear relationship was significant (Pnon-linearity ¼ 0.02). The RR (95%CI) of type 2 diabetes was 0.97(0.93, 1.00), 0.94(0.90, 0.99), 0.93(0.89, 0.98), 0.91 (0.86, 0.96) and 0.90 (0.84, 0.96) for isoflavones intake of 0.2, 0.4, 0.6, 1.1 and 1.5 mg/d, respectively (Fig. 4). The risk of type 2 diabetes decreased by 1% (RR ¼ 0.99; 95%CI: 0.98, 1.00) for every 0.1 mg/d increment in isoflavones intake. In the subgroup analyses, the protective effect of total flavonoids, flavonols, flavan-3-ols, anthocyanins and isoflavones were only found in USA and Europe, not in Asia.
[7] Nettleton JA, Harnack LJ, Scrafford CG, Mink PJ, Barraj LM, Jacobs Jr DR. Dietary flavonoids and flavonoid-rich foods are not associated with risk of type 2 diabetes in postmenopausal women. J Nutr 2006;136:3039e45. [8] Quansah DY, Ha K, Jun S, Kim SA, Shin S, Wie GA, et al. Associations of dietary antioxidants and risk of type 2 diabetes: data from the 2007-2012 korea national health and nutrition examination survey. Molecules 2017;22. [9] Song Y, Manson JE, Buring JE, Sesso HD, Liu S. Associations of dietary flavonoids with risk of type 2 diabetes, and markers of insulin resistance and systemic inflammation in women: a prospective study and cross-sectional analysis. J Am Coll Nutr 2005;24:376e84. [10] Tresserra-Rimbau A, Guasch-Ferre M, Salas-Salvado J, Toledo E, Corella D, Castaner O, et al. Intake of total polyphenols and some classes of polyphenols is inversely associated with diabetes in elderly people at high cardiovascular disease risk. J Nutr 2016;146(4):767e77. [11] Wedick NM, Pan A, Cassidy A, Rimm EB, Sampson L, Rosner B, et al. Dietary flavonoid intakes and risk of type 2 diabetes in us men and women. Am J Clin Nutr 2012;95:925e33. [12] Zamora-Ros R, Forouhi NG, Sharp SJ, Gonzalez CA, Buijsse B, Guevara M, et al. The association between dietary flavonoid and lignan intakes and incident type 2 diabetes in european populations: the epic-interact study. Diabetes Care 2013;36:3961e70. [13] Ding M, Pan A, Manson JE, Willett WC, Malik V, Rosner B, et al. Consumption of soy foods and isoflavones and risk of type 2 diabetes: a pooled analysis of three us cohorts. Eur J Clin Nutr 2016;70:1381e7.
Conflicts of interest
Yunping Zhou School of Nursing, Qingdao University, Qingdao, Shandong, China
None.
Tao Wang School of Public Health, Qingdao University, Qingdao, Shandong, China
References [1] Liu YJ, Zhan J, Liu XL, Wang Y, Ji J, He QQ. Dietary flavonoids intake and risk of type 2 diabetes: a meta-analysis of prospective cohort studies. Clin Nutr 2014;33:59e63. [2] Greenland S, Longnecker MP. Methods for trend estimation from summarized dose-response data, with applications to meta-analysis. Am J Epidemiol 1992;135:1301e9. [3] Durrleman S, Simon R. Flexible regression models with cubic splines. Stat Med 1989;8:551e61. [4] Grosso G, Stepaniak U, Micek A, Kozela M, Stefler D, Bobak M, et al. Dietary polyphenol intake and risk of type 2 diabetes in the polish arm of the health, alcohol and psychosocial factors in eastern europe (hapiee) study. Br J Nutr 2017;118:60e8. [5] Jacques PF, Cassidy A, Rogers G, Peterson JJ, Meigs JB, Dwyer JT. Higher dietary flavonol intake is associated with lower incidence of type 2 diabetes. J Nutr 2013;143:1474e80. [6] Knekt P, Kumpulainen J, Jarvinen R, Rissanen H, Heliovaara M, Reunanen A, et al. Flavonoid intake and risk of chronic diseases. Am J Clin Nutr 2002;76: 560e8.
5
Daiqin Song Zibo Center for Disease Control and Prevention, Zibo, Shandong Province, PR China Aimin Wang* School of Nursing, Qingdao University, Qingdao, Shandong, China *
Corresponding author. 38 Dengzhou Road, Shandong Province, Qingdao, 266021, PR China. E-mail address: [email protected] (A. Wang). 30 May 2018
Please cite this article in press as: Zhou Y, et al., Dietary intake of flavonoid subclasses and risk of type 2 diabetes in prospective cohort studies: A doseeresponse meta-analysis, Clinical Nutrition (2018), https://doi.org/10.1016/j.clnu.2018.08.024
Contents lists available at ScienceDirect
Clinical Nutrition journal homepage: http://www.elsevier.com/locate/clnu
Letter to the Editor
Dietary intake of flavonoid subclasses and risk of type 2 diabetes in prospective cohort studies: A doseeresponse meta-analysis s u m m a r y Keywords: Type 2 diabetes Flavonoids Doseeresponse analysis
Flavonoids intake has been linked to risk of type 2 diabetes theoretically, but the findings were conflicting from observational studies. Results from the recent meta-analysis suggested a moderate favorable effect of total dietary flavonoids consumption on type 2 diabetes risk. However, the relationship, if exists, between total dietary flavonoids consumption and type 2 diabetes risk is unclear, especially further in the association between flavonoid subclasses and risk of type 2 diabetes. Thus, the doseeresponse relationship was assessed by restricted cubic spline model and multivariate random-effect metaregression. Among flavonoid subclasses, protective effects were obtained for intakes of flavonols, flavan3-ols, isoflavones and anthocyanidins in high vs. low analysis and doseeresponse analysis. © 2018 Elsevier Ltd and European Society for Clinical Nutrition and Metabolism. All rights reserved.
Dear editor, We read with great interest the article by Liu et al. [1] entitled “Dietary flavonoids intake and risk of type 2 diabetes: A metaanalysis of prospective cohort studies”. The study suggested that an inverse relation was found of total dietary flavonoids with type 2 diabetes. This is a very important investigation considering the widespread consumption of flavonoids and high incidence of type 2 diabetes. Thus, it is meaningful regarding the exploration of the causes and control of type 2 diabetes. However, we would like to 1) explore the association between flavonoid subclasses and type 2 diabetes risk for the study conducted by Liu et al. [1] only examined the relationship between total flavonoids intake; 2) draw attention to the type of doseeresponse analysis between total flavonoids and flavonoid subclasses consumption with risk of type 2 diabetes, because Liu et al. [1] assumed a linear dose relationship without performing a formal test; 3) reveal a more robust and wide association between flavonoid subclasses consumption and risk of type 2 diabetes because only three studies in USA and one study in Finland performed by Liu et al. [1]. Moreover, categories of flavonoids consumption per day differed between studies, which might complicate the interpretation of the pooled results across study populations with different categories. In this respect, a doseeresponse meta-analysis provides a solution to the problem, from which a summary risk estimate can be derived for a standardized increase and specific exposure values considering flavonoid consumption per day. Studies included in this doseeresponse analysis should provide the following information: the number of cases and participants, and relative risk (RR) estimates with 95% confidence intervals (CIs) for each category of flavonoid subclasses were extracted. We extracted data on the amount of flavonoids/subclasses intake, distributions of cases and person-years (when available), and hazard
ratios with 95% confidence intervals for 3 exposure categories. The median or mean daily take of flavonoids in each category was assigned to the corresponding hazard ratio with the 95%CI for each study. Detailed information of the included publications is shown in Table 1. Two-stage random-effects doseeresponse meta-analysis was performed to examine linear and nonlinear relationships between flavonoids intake and type 2 diabetes. In the first stage, the method reported by Greenland and Orsini [2] (generalized least-squares) was used to calculate study-specific coefficients on the basis of results across categories of flavonoids intake taking into account the correlation within each set of retrieved risk ratios. Nonlinear doseeresponse analysis was modeled by using restricted cubic splines with 3 knots at fixed percentiles [3] (25th, 50th, and 75th) of the distribution. We combined the coefficients that had been estimated within each study by performing random-effects meta-analysis. In linear doseeresponse meta-analysis, the method of Der-Simonian and Laird was used, and in nonlinear dose response meta-analysis, the multivariate extension of the method of moments was used. We calculated an overall P value by testing that the 2 regression coefficients were simultaneously equal to zero. We then calculated a P value for nonlinearity by testing that the coefficient of the second spline was equal to zero. All statistical analyses were performed with STATA version 12 (Stata Corporation, College Station, Texas, USA). All reported probabilities (p values) were two-sided, with p < 0.05 considered statistically significant. 1. Total diet flavonoids Nine studies [4e12] including 12 results investigated the association of total flavonoids with incidence of type 2 diabetes. Comparing the highest with the lowest intake, the dietary intake of flavonoids significantly reduced the type 2 diabetes risk by
https://doi.org/10.1016/j.clnu.2018.08.024 0261-5614/© 2018 Elsevier Ltd and European Society for Clinical Nutrition and Metabolism. All rights reserved.
Please cite this article in press as: Zhou Y, et al., Dietary intake of flavonoid subclasses and risk of type 2 diabetes in prospective cohort studies: A doseeresponse meta-analysis, Clinical Nutrition (2018), https://doi.org/10.1016/j.clnu.2018.08.024
2
First author (year of publication)
Country
Study size
Sex
Measure of intake
Exposure
RR (95%CI)
Adjustment for Covariates
Quality assessment
Zamora-Ros R et al. [12]
European countries
15258/729
F/M
A self-administered questionnaire of 98e266 food items
8
Spain
7447/314
F/M
a 137-item semiquantitative FFQ
0.90 0.73 0.81 1.03 0.98 0.89 0.94 0.67 0.97 0.69 0.98 0.88
Age, sex, and total energy
Tresserra-Rimbau A et al. [10]
Flavonoids Flavan-3-ols Flavonols Flavanone Isoflavones Flavone Anthocyanidin Flavonoids Flavonols Flavanone Flavone Anthocyanidin
6
Grosso G et al. [4]
Poland
5806/456
F/M
a self-administered machinereadable 148-item FFQ
Quansah DY et al. [8]
Korea
24377/603
F/M
3193 food items
Flavonoids Flavonols Flavanone Isoflavones Flavone Anthocyanidin Flavonoids
Flavonoids
0.44 (0.33e0.63) 0.89 (0.63e1.26) 0.75 (0.54e1.03) 0.93 (0.65e1.34) 0.47 (0.33e0.68) 0.68 (0.48e0.98) M:0.75 (0.49e1.15) F:0.97 (0.60e1.57) M:0.67 (0.44e1.01) F:0.88 (0.54e1.42) M:0.79 (0.52e1.19) F:1.07 (0.69e1.66) M:0.96 (0.64e1.45) F:0.91 (0.58e1.42) M:0.95 (0.63e1.42) F:1.35 (0.82e2.24) M:0.83 (0.57e1.23) F:0.67 (0.43e1.04) M:1.03 (0.67e1.56) F:0.83 (0.53e1.28) 0.98 (0.77e1.24)
Age, sex, recruitment center, and intervention group, smoking, BMI, physical activity, dyslipidemia, hypertension, education level, total energy intake, alcohol intake, adherence to the Mediterranean diet, and fasting glucose concentrations at baseline. Age, total energy intake, BMI, physical activity, educational status, smoking status, alcohol consumption, alcohol intake, menopausal status and dietary fiber.
Flavonoids
0.92 (0.78e1.09)
Flavan-3-ols Flavonols Flavanone Isoflavones Flavone Anthocyanidin a dietary history questionnaire of 100 items 131-item semiquantitative FFQ
(0.77e1.04) (0.57e0.93) (0.69e0.95) (0.79e1.34) (0.77e1.25) (0.70e1.14) (0.82e1.08) (0.48e0.93) (0.68e1.39) (0.49e0.97) (0.71e1.35) (0.62e1.24)
Knekt P et al. [6]
Finland
10054/526
F/M
Song Y et al. [9]
USA
38018/1614
F
Wedick NM et al. [11](NHS)
USA
70359/6878
F
a comprehensive FFQ with 118e131 food items
Flavonoids Flavan-3-ols Flavonols Flavanone Flavone Anthocyanidin
0.85 0.87 0.84 1.05 1.07 0.83
(0.79e0.92) (0.81e0.94) (0.78e0.91) (0.97e1.13) (0.99e1.16) (0.77e0.90)
Wedick NM et al. [11](NHS II)
USA
89201/3084
F
a comprehensive FFQ with 118e131 food items
Flavonoids Flavan-3-ols Flavonols Flavanone Flavone Anthocyanidin
0.99 1.01 0.99 1.08 1.02 0.83
(0.89e1.11) (0.91e1.12) (0.89e1.10) (0.97e1.22) (0.91e1.16) (0.73e0.94)
6
Age, BMI, current smoking, education, household income, physical activity, and regular alcohol consumption
7
sex and age
6
age, BMI, total energy intake, smoking, exercise, alcohol use, history of hypertension, history of high cholesterol, family history of diabetes, fiber intake, glycemic load, magnesium, and total fat. Age, BMI, smoking status alcohol intake, multivitamin use, physical activity, a family history of diabetes, postmenopausal status and hormone use, ethnicity, total energy, polyunsaturated:saturated fat ratio, intakes of red meat, fish, whole grains, coffee, high-calorie sodas, and trans fat. Age, BMI, smoking status alcohol intake, multivitamin use, physical activity, a family history of diabetes, postmenopausal status and hormone use, oral contraceptive use, ethnicity, total energy, polyunsaturated:saturated fat ratio, intakes of red meat, fish, whole grains, coffee, high-calorie sodas, and trans fat.
7
8
8
Letter to the Editor / Clinical Nutrition xxx (2018) 1e5
Please cite this article in press as: Zhou Y, et al., Dietary intake of flavonoid subclasses and risk of type 2 diabetes in prospective cohort studies: A doseeresponse meta-analysis, Clinical Nutrition (2018), https://doi.org/10.1016/j.clnu.2018.08.024
Table 1 Characteristics of studies on dietary flavonoids and incidence of type 2 diabetes.
First author (year of publication)
Country
Study size
Sex
Measure of intake
Exposure
RR (95%CI)
Adjustment for Covariates
Quality assessment
Wedick NM et al. [11](HPFS)
USA
41334/2649
M
a comprehensive FFQ with 118e131 food items
USA
35816/3395
F
a 127-item semiquantitative FFQ
0.92 0.88 0.88 1.09 1.07 0.93 0.97
(0.81e1.04) (0.78e0.99) (0.78e1.00) (0.96e1.24) (0.94e1.22) (0.81e1.05) (0.86e1.10)
Jacques PF et al. [5]
USA
2915/308
F/M
the Harvard semiquantitative FFQ
Ding M et al. [13] (NHS)
USA
63115/4519
F
a validated 116-item FFQ
Flavonoids Flavan-3-ols Flavonols Flavanone Flavone Anthocyanidin Isoflavones
0.89 0.90 0.68 1.08 1.02 0.83 0.97
(0.75e1.05) (0.80e1.01) (0.54e0.86) (0.97e1.22) (0.91e1.16) (0.73e0.94) (0.88e1.07)
Age, BMI, smoking status alcohol intake, multivitamin use, physical activity, a family history of diabetes, ethnicity, total energy, polyunsaturated:saturated fat ratio, intakes of red meat, fish, whole grains, coffee, high-calorie sodas, and trans fat. age, energy, education level, BMI, waist:hip ratio, activity level, smoking status, multivitamin use, and hormone therapy. sex and time dependent variables age, cardiovascular disease, current smoker, BMI, vegetable, fruit intake and cumulative mean energy intake.
8
Nettleton JA et al. [7]
Flavonoids Flavan-3-ols Flavonols Flavanone Flavone Anthocyanidin Flavonoids
Ding M et al. [13] NHS II)
USA
79061/3920
F
a validated 116-item FFQ
Isoflavones
0.85 (0.76e0.95)
Ding M et al. [13](HPFS)
USA
21281/742
M
a validated 116-item FFQ
Isoflavones
0.80 (0.62e1.02)
Japan
59791/1114
a 147-item FFQ
Isoflavones
M:0.97 (0.79e1.20) F:0.91 (0.71e1.17)
Korea Finland
7963/225 25514/660
a 1549 food item a validated self-administered 276-item FFQ
Flavone Flavone
0.59 (0.41e0.85) 1.03 (0.91e1.06)
age, race, family history of T2D, baseline disease status, BMI, physical activity, overall dietary pattern, total energy intake, smoking status, menopausal status, and postmenopausal hormone use age, race, family history of T2D, baseline disease status, BMI, physical activity, overall dietary pattern, total energy intake, smoking status, menopausal status, and postmenopausal hormone use age, race, family history of T2D, baseline disease status, BMI, physical activity, overall dietary pattern, total energy intake, smoking status age, study area, BMI, smoking habit, alcohol consumption, family history of diabetes mellitus, leisure time physical activity, history of hypertension, coffee, green tea, magnesium, calcium, vegetable, fiber intake, fish, and total energy intake age, BMI, education, menopausal status age, supplementation,BMI, cigarettes smoked daily, smoking years, blood pressure, total cholesterol, high-density lipoprotein cholesterol, leisure-time physical activity and daily intake of alcohol and energy
7
7
7
7
7
6
6 7
Letter to the Editor / Clinical Nutrition xxx (2018) 1e5
Please cite this article in press as: Zhou Y, et al., Dietary intake of flavonoid subclasses and risk of type 2 diabetes in prospective cohort studies: A doseeresponse meta-analysis, Clinical Nutrition (2018), https://doi.org/10.1016/j.clnu.2018.08.024
Table 1 (continued )
F, Female; M, Male; FFQ, food frequency questionnaire; BMI, body mass index; NHS, Nurses' Health Study; NHS II, Nurses' Health Study II; HPFS, Health Professionals Follow-Up Study.
3
4
Letter to the Editor / Clinical Nutrition xxx (2018) 1e5
11.00% (RR ¼ 0.89; 95%CI: 0.82e0.95) with considerable heterogeneity (Pheterogeneity ¼ 0.01, I2 ¼ 56.40%). Five studies (seven comparisons [4,7,10e12]) were eligible for the doseeresponse analysis. The departure from linear relationship was significant (P < 0.05). The RR (95%CI) of type 2 diabetes was 0.97 (0.94, 0.99), 0.96 (0.92, 0.99), 0.95 (0.91, 0.99), 0.95 (0.91, 1.00) and 0.93 (0.89, 0.97) for total flavonoids intake of 100, 200, 300, 400 and 500 mg/d, respectively (Fig. 1). The risk of type 2 diabetes decreased by 2% (RR ¼ 0.98; 95%CI 0.97, 0.99) for every 100 mg/d increment in total flavonoids intake. In the analysis of flavonoid subclasses with risk of type 2 diabetes, the beneficial effects were only observed in flavonols, flavan-3-ols, anthocyanins and isoflavones not found in flavones and flavanones. Besides, the studies in flavan-3-ols eligible for the doseeresponse analysis were limited.
2. Flavonols The summary risk estimate was 0.87 (95% CI, 0.83e0.91), with little evidence of heterogeneity (I2 ¼ 38.7%, Pheterogeneity ¼ 0.11).
Four studies with six cohorts [4,10e12] reporting flavonols intake were eligible for the doseeresponse analysis. The departure from linear relationship was significant (Pnon-linearity ¼ 0.02). The RR (95%CI) of type 2 diabetes was 0.95(0.91, 0.98), 0.89(0.86, 0.93), 0.86(0.81, 0.90), 0.81 (0.75, 0.88) and 0.83 (0.72, 0.96) for flavonols intake of 10, 20, 30, 50 and 100 mg/d, respectively (Fig. 2). The risk of type 2 diabetes decreased by 3% (RR ¼ 0.97; 95%CI 0.95, 0.99) for every 100 mg/d increment in flavonols intake. 3. Anthocyanins The summary risk estimate was 0.89 (95%CI, 0.84e0.93), with no evidence of heterogeneity (I2 ¼ 28.5%, Pheterogeneity ¼ 0.19). Four studies with six cohorts [4,10e12] reporting anthocyanins intake were eligible for the doseeresponse analysis. The departure from linear relationship was significant (Pnon-linearity ¼ 0.001). The RR (95%CI) of type 2 diabetes was 0.93(0.90, 0.96), 0.90(0.86, 0.94), 0.88(0.85, 0.92), 0.88(0.84, 0.91) and 0.86 (0.82, 0.91) for anthocyanins intake of 5, 8, 13, 20 and 30 mg/d, respectively (Fig. 3). A
Fig. 1. Nonlinear doseeresponse relationship between total flavonoid intake and type 2 diabetes risk in cohort studies assessed by restricted cubic spline model with three knots.
Fig. 3. Nonlinear doseeresponse relationship between anthocyanins intake and type 2 diabetes risk in cohort studies assessed by restricted cubic spline model with three knots.
Fig. 2. Nonlinear doseeresponse relationship between flavonols intake and type 2 diabetes risk in cohort studies assessed by restricted cubic spline model with three knots.
Fig. 4. Nonlinear doseeresponse relationship between isoflavones intake and type 2 diabetes risk in cohort studies assessed by restricted cubic spline model with three knots.
Please cite this article in press as: Zhou Y, et al., Dietary intake of flavonoid subclasses and risk of type 2 diabetes in prospective cohort studies: A doseeresponse meta-analysis, Clinical Nutrition (2018), https://doi.org/10.1016/j.clnu.2018.08.024
Letter to the Editor / Clinical Nutrition xxx (2018) 1e5
2 mg/day increment in intake of anthocyanins was associated with a risk ratio of 0.99 (95% CI: 0.98, 1.00). 4. Isoflavones The summary risk estimate was 0.92 (95% CI, 0.87e0.98), with some evidence of heterogeneity (I2 ¼ 0.00%, Pheterogeneity ¼ 0.53). Three studies with five cohorts [4,12,13] reporting isoflavones intake were eligible for the doseeresponse analysis. The departure from linear relationship was significant (Pnon-linearity ¼ 0.02). The RR (95%CI) of type 2 diabetes was 0.97(0.93, 1.00), 0.94(0.90, 0.99), 0.93(0.89, 0.98), 0.91 (0.86, 0.96) and 0.90 (0.84, 0.96) for isoflavones intake of 0.2, 0.4, 0.6, 1.1 and 1.5 mg/d, respectively (Fig. 4). The risk of type 2 diabetes decreased by 1% (RR ¼ 0.99; 95%CI: 0.98, 1.00) for every 0.1 mg/d increment in isoflavones intake. In the subgroup analyses, the protective effect of total flavonoids, flavonols, flavan-3-ols, anthocyanins and isoflavones were only found in USA and Europe, not in Asia.
[7] Nettleton JA, Harnack LJ, Scrafford CG, Mink PJ, Barraj LM, Jacobs Jr DR. Dietary flavonoids and flavonoid-rich foods are not associated with risk of type 2 diabetes in postmenopausal women. J Nutr 2006;136:3039e45. [8] Quansah DY, Ha K, Jun S, Kim SA, Shin S, Wie GA, et al. Associations of dietary antioxidants and risk of type 2 diabetes: data from the 2007-2012 korea national health and nutrition examination survey. Molecules 2017;22. [9] Song Y, Manson JE, Buring JE, Sesso HD, Liu S. Associations of dietary flavonoids with risk of type 2 diabetes, and markers of insulin resistance and systemic inflammation in women: a prospective study and cross-sectional analysis. J Am Coll Nutr 2005;24:376e84. [10] Tresserra-Rimbau A, Guasch-Ferre M, Salas-Salvado J, Toledo E, Corella D, Castaner O, et al. Intake of total polyphenols and some classes of polyphenols is inversely associated with diabetes in elderly people at high cardiovascular disease risk. J Nutr 2016;146(4):767e77. [11] Wedick NM, Pan A, Cassidy A, Rimm EB, Sampson L, Rosner B, et al. Dietary flavonoid intakes and risk of type 2 diabetes in us men and women. Am J Clin Nutr 2012;95:925e33. [12] Zamora-Ros R, Forouhi NG, Sharp SJ, Gonzalez CA, Buijsse B, Guevara M, et al. The association between dietary flavonoid and lignan intakes and incident type 2 diabetes in european populations: the epic-interact study. Diabetes Care 2013;36:3961e70. [13] Ding M, Pan A, Manson JE, Willett WC, Malik V, Rosner B, et al. Consumption of soy foods and isoflavones and risk of type 2 diabetes: a pooled analysis of three us cohorts. Eur J Clin Nutr 2016;70:1381e7.
Conflicts of interest
Yunping Zhou School of Nursing, Qingdao University, Qingdao, Shandong, China
None.
Tao Wang School of Public Health, Qingdao University, Qingdao, Shandong, China
References [1] Liu YJ, Zhan J, Liu XL, Wang Y, Ji J, He QQ. Dietary flavonoids intake and risk of type 2 diabetes: a meta-analysis of prospective cohort studies. Clin Nutr 2014;33:59e63. [2] Greenland S, Longnecker MP. Methods for trend estimation from summarized dose-response data, with applications to meta-analysis. Am J Epidemiol 1992;135:1301e9. [3] Durrleman S, Simon R. Flexible regression models with cubic splines. Stat Med 1989;8:551e61. [4] Grosso G, Stepaniak U, Micek A, Kozela M, Stefler D, Bobak M, et al. Dietary polyphenol intake and risk of type 2 diabetes in the polish arm of the health, alcohol and psychosocial factors in eastern europe (hapiee) study. Br J Nutr 2017;118:60e8. [5] Jacques PF, Cassidy A, Rogers G, Peterson JJ, Meigs JB, Dwyer JT. Higher dietary flavonol intake is associated with lower incidence of type 2 diabetes. J Nutr 2013;143:1474e80. [6] Knekt P, Kumpulainen J, Jarvinen R, Rissanen H, Heliovaara M, Reunanen A, et al. Flavonoid intake and risk of chronic diseases. Am J Clin Nutr 2002;76: 560e8.
5
Daiqin Song Zibo Center for Disease Control and Prevention, Zibo, Shandong Province, PR China Aimin Wang* School of Nursing, Qingdao University, Qingdao, Shandong, China *
Corresponding author. 38 Dengzhou Road, Shandong Province, Qingdao, 266021, PR China. E-mail address: [email protected] (A. Wang). 30 May 2018
Please cite this article in press as: Zhou Y, et al., Dietary intake of flavonoid subclasses and risk of type 2 diabetes in prospective cohort studies: A doseeresponse meta-analysis, Clinical Nutrition (2018), https://doi.org/10.1016/j.clnu.2018.08.024