Abstract
Aim
To assess the associations of total dietary fibre and fibre from different food sources (ie, cereal, fruit and vegetables) with the risk of diabetes.
Materials and methods
The Melbourne Collaborative Cohort Study enrolled 41 513 participants aged 40 to 69 years from 1990 to 1994. The first and second follow‐ups were conducted in 1994 to 1998 and 2003 to 2007, respectively. Self‐reported diabetes incidence was recorded at both follow‐ups. We analysed data from 39 185 participants, with a mean follow‐up of 13.8 years. The relationships between dietary fibre intake (total, fruit, vegetable and cereal fibre) and the incidence of diabetes were assessed using modified Poisson regression, adjusted for dietary, lifestyle, obesity, socioeconomic and other possible confounders. Fibre intake was categorized into quintiles.
Results
At total of 1989 incident cases were identified over both follow‐up surveys. Total fibre intake was not associated with diabetes risk. Higher intake of cereal fibre (P for trend = 0.003), but not fruit (P for trend = 0.3) and vegetable fibre (P for trend = 0.5), was protective against diabetes. For cereal fibre, quintile 5 versus quintile 1 showed a 25% reduction in diabetes risk (incidence risk ratio [IRR] 0.75, 95% confidence interval [CI] 0.63‐0.88). For fruit fibre, only quintile 2 versus quintile 1 showed a 16% risk reduction (IRR 0.84, 95% CI 0.73‐0.96). Adjustment for body mass index (BMI) and waist‐to‐hip ratio eliminated the association and mediation analysis showed that BMI mediated 36% of the relationship between fibre and diabetes.
Conclusion
Intake of cereal fibre and, to a lesser extent, fruit fibre, may reduce the risk of diabetes, while total fibre showed no association. Our data suggest that specific recommendations regarding dietary fibre intake may be needed to prevent diabetes.
Keywords: cohort study, dietary intervention, population study, type 2 diabetes
1. INTRODUCTION
Worldwide, half a billion people live with diabetes, with type 2 diabetes accounting for 90% of the disease burden. 1 According to self‐reported data, an estimated 1.2 million Australians (4.9% of the total population) had diabetes in 2017 to 2018. 2 The global prevalence and incidence of type 2 diabetes are expected to increase further due to high rates of obesity. Global obesity rates increased from 3.2% in 1975 to 10.8% in 2014 for men and from 6.4% to 14.9% for women. 3 Prevention strategies focusing on lifestyle modification, including dietary interventions, are crucial to reduce the burden. 3
A high dietary fibre intake (>25 g/d in women and > 38 g/d in men) has been linked to a 20% to 30% lower incidence of type 2 diabetes. 4 Previous studies assessing the effect of dietary fibre on diabetes have reported that total fibre reduced diabetes incidence 5 , 6 ; others report that cereal fibre, but not total fibre, had a protective role. 6 , 7 It is less widely understood whether high intakes of whole grains and insoluble cereal fibre, which are generally non‐viscous and hence do not affect postprandial glucose, are the principal drivers of these benefits. 4 Fibre from different sources varies in solubility and effect on digestion. Fruit and vegetables contain mostly soluble fibre, while cereal grains contain primarily insoluble fibre. 8 Soluble fibre generally prolongs transit through the human digestive system, delaying stomach emptying and slowing down glucose absorption. Insoluble fibre generally increases the faecal bulk and excretion of bile acids. 8 Previous studies have primarily focused on how fibre might affect glycaemic control in people with type 2 diabetes, 9 , 10 with less focus on how it might affect diabetes risk. 11 , 12
The mechanisms by which fibre could reduce the risk of diabetes are attributed mainly to its effect on the gut microbiome and weight control. 13 , 14 , 15 , 16 Studies have also shown that the concentrations of hormones regulating gastric emptying and central satiety, such as cholecystokinin and glucagon‐like peptide YY (GLP‐YY), are elevated after consuming a high‐fibre diet compared to a low‐fibre diet. 17 Higher fibre intakes are also reported to reduce weight, waist circumference, and visceral adiposity. 18 , 19
A previous analysis using Melbourne Collaborative Cohort Study (MCCS) data found no association between cereal or total fibre intake and type 2 diabetes incidence. 12 With growing evidence for the effects of cereal fibre on the gut microbiome and the prevention of diabetes, we were able to update the MCCS analysis with more incident cases of type 2 diabetes, a longer follow‐up period, and a more robust data analysis method, including assessment of mediation by body mass index (BMI). In this study, we examined associations of total fibre and fibre from different food sources with the incidence of type 2 diabetes using data from the MCCS.
2. MATERIALS AND METHODS
2.1. The Melbourne Collaborative Cohort Study
The MCCS is a prospective cohort of 41 513 Melbourne residents recruited between 1990 and 1994. Recruitment of participants, follow‐up, and the data collection process have been described in detail elsewhere. 20 The electoral roll was used to recruit participants, and a direct approach was additionally made through clubs, churches, and ethnicity‐specific media. At the outset, interviewer‐administered questionnaires were used to collect sociodemographic and nutritional data. Between 1995 and 1998 (Wave 1) and 2003 and 2007 (Wave 2), follow‐up surveys were conducted. As shown in Figure 1, at baseline, 39 185 participants were included; 34 444 were included in the first wave of follow‐up and 25 888 in the second wave of follow‐up. The average follow‐up period was 13.8 years.
FIGURE 1.
Flow diagram showing the recruitment and follow‐up of the cohort. MCCS, Melbourne Collaborative Cohort Study
During the first follow‐up, data were obtained using either a mailed self‐administered questionnaire or an interview via telephone. In the second follow‐up visit, self‐administered questionnaires were used to collect data, and anthropometric measures except height were repeated. 20
2.2. Dietary assessment
At baseline, dietary consumption data were collected using a 121‐item self‐administered Food Frequency Questionnaire (FFQ) 21 and this was validated with 24‐hour recall dietary data. 22 Each food item was given a sex‐specific average portion size, and certain fruits' frequency was adjusted to account for the season when the FFQ was completed. Mean daily nutrient intakes were calculated by multiplying the daily frequency of each food item by the nutritional composition and portion. Most nutrient composition data came from the Australian food composition tables. 23 Data were collected from applicable UK (folate and vitamin), 24 Royal Melbourne Institute of Technology (fatty acids), 25 and US Department of Agriculture (Carotenoids) 26 sources when Australian tables were unavailable.
Data from the FFQ were used to calculate the Alternative Healthy Eating Index‐2010 (AHEI‐2010), based on scientific evidence for the relationship between food and health. 27 A higher score indicates a higher intake of fruits, vegetables, whole grains, nuts, legumes, omega‐3, and polyunsaturated fatty acids. It also indicates a moderate alcohol intake and a lower intake of sugar‐sweetened beverages, fruit juice, red and processed meat, trans‐fat, and sodium. The 11 dietary component scores, ranging from 0 to 10 depending on adherence to recommended intakes, are summed to give total AHEI‐2010 scores ranging from 0 to 110.
2.3. Non‐dietary assessments
At baseline, interviewer‐administered questionnaires were used to collect information on age, sex, country of origin, smoking, alcohol intake, and physical activity. Participants were categorized into three groups based on their region of origin: (a) Australia/New Zealand/other; (b) Northern European (mainly British); and (c) Southern European (Greek and Italian). Deciles of the Socioeconomic Indices for Areas (SEIFA) Index of Relative Socioeconomic Disadvantage based on postcode at baseline were used to indicate socioeconomic standing. SEIFA deciles were recoded into quintiles, with the first quintile being the most disadvantaged and the fifth quintile the most prosperous. History of comorbid health problems was also assessed for all the participants. Participants were also asked about their family history of diabetes. A standardized questionnaire was used to assess how much time participants spent walking or taking part in moderate‐intensity and vigorous physical activity. These data were combined to give an overall score that weighted time spent doing vigorous activity twice that of less vigorous activity. The score ranged from 0 to 16 and was divided into four categories: 0, >0 to 4, >4 to 6, and >6, for this analysis.
Height, weight, and waist and hip circumferences were all measured using standard procedures, and the BMI (in kg per meter squared) and waist‐to‐hip ratio (WHR) were calculated.
2.4. Outcome measurement
A self‐administered questionnaire mailed to individuals 4 years after baseline was used to identify incident cases of diabetes. The participants were asked, “Has a doctor ever told you that you have diabetes?” Those who answered yes were asked the year of their diagnosis. Except for those who indicated a diagnosis date before baseline, who were excluded, 76% of participants with a self‐reported incident case had their diagnosis confirmed by doctors nominated by them. The remaining participants were also considered to have type 2 diabetes because of their age. At the second follow‐up, similar questions were used to identify additional incident cases of diabetes. However, no additional confirmation was sought; all cases were considered to be type 2 diabetes mellitus.
2.5. Data analysis
Dietary fibre intakes (g/d) were categorized into quintiles to minimize the effect of outliers and provide visualization of the association between fibre and diabetes. Sex‐specific cut‐off points were used to account for differing distributions of fibre intake between men and women but a single mean fibre intake for each quintile has been presented for simplicity. The lowest quintile of intake was used as the reference category. Similarly, energy intakes were also categorized into quintiles.
Normality distribution was checked for continuous variables using the Kolmogrov‐Smirnov test. Chi‐squared tests for categorical variables and analysis of variance tests for continuous variables were used to assess associations between baseline characteristics and fibre intake quintiles. At the first and second follow‐ups, the cumulative incidences of diabetes were compared across predictor categories. A generalized estimating equation Poisson regression model 28 with robust error variance 29 was used to investigate the associations of baseline dietary fibre intakes with the incidence of diabetes after adjusting for possible confounders. Survival analysis is not suitable for this dataset because the outcome ascertainment was done at three timepoints only, hence the time and censorship information was not measured accurately for the outcome variable, that is, diabetes.
The associations of total and different fibre types were assessed separately. The selection of possible predictor variables was made using a priori univariate analysis and based on previous scientific literature.
Three graduated models were used to calculate the incidence risk ratio (IRR) and 95% confidence intervals (CIs). Age, sex, SEIFA (quintiles [Q]1–5), smoking status (never, former and current), drinking status (never, former and current), family history of diabetes, physical activity level, AHEI‐2010 quintile, quintile of energy intake, and region of origin were all adjusted for in Model 1. Model 2 was fitted using Model 1 variables plus BMI, while Model 3 was fitted using Model 2 variables plus WHR. Trends across quintiles were calculated by assigning a median fibre intake to each person in that quintile, and P for trend was reported. stata version 16 was used for all statistical analyses.
We also modelled the association between the consumption of cereal fibre and type 2 diabetes using BMI as a mediator. 30 , 31 The natural indirect effect (NIE) is only attributable to mediation, whereas the controlled direct effect (CDE) is the average impact of an increase in cereal fibre consumption. The total effect is the product of the CDE and NIE. 32 The proportion mediated was calculated using the following formula (CDE × [NIE – 1])/([CDE × NIE] − 1). 33
3. RESULTS
At baseline, participants in higher total fibre intake groups (Q4 and Q5) were likely to be older, socioeconomically less disadvantaged (SEIFA Q5), former smokers and physically active, and to have high AHEI‐2010, normal WHR, normal BMI, high energy intake and fewer comorbidities than participants in Q1. All variables except comorbidity and family history of diabetes showed a significant association (P < 0.01) with total fibre quintiles based on chi‐squared test (Table 1).
TABLE 1.
Descriptive analysis of variables by total fibre quintile
| Fibre Q1 | Fibre Q2 | Fibre Q3 | Fibre Q4 | Fibre Q5 | Total | Test statistic | |
|---|---|---|---|---|---|---|---|
| Total fibre, g/d a | 16.6 (3.5) | 24 (1.6) | 29.2 (1.6) | 35.5 (2.1) | 50.4 (14.6) | 31.1 (13.4) | P < 0.001 |
| Cereal fibre, g/d a | 5.9 (2.6) | 8.6 (3.0) | 10.6 (3.6) | 12.9 (4.5) | 17.2 (8.5) | 11.0 (6.2) | P < 0.001 |
| Fruit fibre, g/d a | 3.3 (2.1) | 5.5 (2.8) | 7.1 (3.4) | 9.1 (4.1) | 14.4 (8.8) | 7.86 (6.2) | P < 0.001 |
| Vegetable fibre, g/d a | 2.8 (1.5) | 4.0 (1.8) | 4.9 (2.1) | 5.8 (2.5) | 8.4 (5.6) | 5.20 (3.6) | P < 0.001 |
| Age |
Chi‐squared = 73.10; df = 8 P < 0.001 |
||||||
| <50 years | 2693 (34.36) | 2677 (34.16) | 2552 (32.56) | 2407 (30.71) | 2340 (29.86) | 12 669 (32.33) | |
| 50‐59 years | 2523 (32.19) | 2571 (32.81) | 2574 (32.84) | 2548 (32.51) | 2577 (32.88) | 12 793 (32.65) | |
| ≥60 years | 2621 (33.44) | 2589 (33.04) | 2711 (34.59) | 2882 (36.77) | 2920 (37.26) | 13 723 (35.02) | |
| Sex |
Chi‐squared = 305.67; df = 4 P < 0.0001 |
||||||
| Male | 2866 (36.57) | 2870 (36.62) | 3055 (38.98) | 3214 (41.01) | 3786 (48.31) | 15 791 (40.30) | |
| Female | 4971 (63.43) | 4967 (63.38) | 4782 (61.02) | 4623 (58.99) | 4051 (51.69) | 23 394 (59.70) | |
| SEIFA quintile |
Chi‐squared = 390.96; df = 16 P < 0.001 |
||||||
| Q1 | 1744 (22.25) | 1473 (18.80) | 1292 (16.49) | 1270 (16.21) | 1297 (16.55) | 7076 (18.06) | |
| Q2 | 1893 (24.15) | 1663 (21.22) | 1608 (20.52) | 1520 (19.40) | 1465 (18.69) | 8149 (20.80) | |
| Q3 | 1292 (16.49) | 1248 (15.92) | 1284 (16.38) | 1149 (14.66) | 1198 (15.29) | 6171 (15.75) | |
| Q4 | 1260 (16.08) | 1435 (18.31) | 1472 (18.78) | 1538 (19.62) | 1549 (19.77) | 7254 (18.51) | |
| Q5 | 1648 (21.03) | 2018 (25.75) | 2181 (27.83) | 2360 (30.11) | 2328 (29.71) | 10 535 (26.89) | |
| Region of origin |
Chi‐squared = 287.95; df = 8 P < 0.001 |
||||||
| AUS/NZ/other | 5001 (63.81) | 5476 (69.87) | 5720 (72.99) | 5732 (73.14) | 5402 (68.93) | 27 331 (69.75) | |
| Northern Europe | 481 (6.14) | 469 (5.98) | 508 (6.48) | 545 (6.95) | 525 (6.70) | 2528 (6.45) | |
| Southern Europe | 2355 (30.05) | 1892 (24.14) | 1609 (20.53) | 1560 (19.91) | 1910 (24.37) | 9326 (23.80) | |
| Smoking status |
Chi‐squared = 526.39; df = 8 P < 0.001 |
||||||
| Never smoked | 4193 (53.50) | 4442 (56.68) | 4605 (58.76) | 4815 (61.44) | 4659 (59.45) | 22 714 (57.97) | |
| Current smoker | 1378 (17.58) | 944 (12.05) | 791 (10.09) | 599 (7.64) | 632 (8.06) | 4344 (11.09) | |
| Former smoker | 2266 (28.91) | 2451 (31.27) | 2441 (31.15) | 2423 (30.92) | 2546 (32.49) | 12 127 (30.95) | |
| Alcohol drinking status |
Chi‐squared = 39.29; df = 8 P < 0.001 |
||||||
| Lifetime abstainers | 2371 (30.25) | 2212 (28.23) | 2140 (27.31) | 2092 (26.69) | 2326 (29.68) | 11 141 (28.43) | |
| Ex‐drinker | 810 (10.34) | 804 (10.26) | 812 (10.36) | 827 (10.55) | 846 (10.79) | 4099 (10.46) | |
| Current drinker | 4656 (59.41) | 4821 (61.52) | 4885 (62.33) | 4918 (62.75) | 4665 (59.53) | 23 945 (61.11) | |
| Physical activity score |
Chi‐squared = 727.38; df = 12 P < 0.001 |
||||||
| 0 | 2255 (28.77) | 1912 (24.40) | 1617 (20.63) | 1454 (18.55) | 1332 (17.00) | 8570 (21.87) | |
| >0 and <4 | 1731 (22.09) | 1603 (20.45) | 1614 (20.59) | 1557 (19.87) | 1380 (17.61) | 7885 (20.12) | |
| ≥4 and <6 | 2631 (33.57) | 2792 (35.63) | 2760 (35.22) | 2834 (36.16) | 2912 (37.16) | 13 929 (35.55) | |
| ≥6 | 1220 (15.57) | 1530 (19.52) | 1846 (23.55) | 1992 (25.42) | 2213 (28.24) | 8801 (22.46) | |
| WHR |
Chi‐squared = 67.26; df = 4 P < 0.001 |
||||||
| Normal | 4835 (61.69) | 5110 (65.20) | 5189 (66.21) | 5302 (67.65) | 5073 (67.65) | 25 509 (67.65) | |
| High | 3002 (38.31) | 2727 (38.31) | 2648 (33.79) | 2535 (32.35) | 2764 (35.27) | 13 676 (34.90) | |
| BMI |
Chi‐squared = 234.02; df = 8 P < 0.001 |
||||||
| < 25.0 kg/m2 | 2553 (32.58) | 2743 (35.00) | 2927 (37.35) | 3154 (40.24) | 3084 (39.35) | 14 461 (36.90) | |
| 25.0‐29.9 kg/m2 | 3355 (42.81) | 3440 (43.89) | 3394 (43.31) | 3347 (42.71) | 3378 (43.10) | 16 914 (43.16) | |
| ≥30.0 kg/m2 | 1929 (24.61) | 1654 (21.11) | 1516 (19.34) | 1336 (19.34) | 1375 (17.54) | 7810 (19.93) | |
| AHEI‐2010 quintile |
Chi‐squared = 4500; df = 16 P < 0.001 |
||||||
| Q1 | 2996 (38.23) | 1964 (25.05) | 1454 (18.55) | 962 (12.28) | 618 (7.89) | 7994 (20.04) | |
| Q2 | 2144 (27.36) | 1864 (23.78) | 1682 (21.46) | 1497 (19.10) | 1309 (16.70) | 8496 (21.68) | |
| Q3 | 1385 (17.67) | 1496 (19.09) | 1521 (19.41) | 1574 (20.08) | 1558 (19.88) | 7534 (19.23) | |
| Q4 | 869 (11.09) | 1396 (17.81) | 1578 (20.14) | 1748 (22.30) | 1896 (24.19) | 7487 (19.11) | |
| Q5 | 443 (5.65) | 1117 (14.25) | 1602 (20.44) | 2056 (26.23) | 2456 (31.34) | 7674 (19.58) | |
| Family history of diabetes |
Chi‐squared = 8.24; df = 4 P = 0.083 |
||||||
| No | 6383 (81.45) | 6482 (82.71) | 6435 (82.11) | 6509 (83.05) | 6436 (82.12) | 32 245 (82.29) | |
| Yes | 1454 (18.55) | 1355 (17.29) | 1402 (17.89) | 1328 (16.95) | 1401 (17.88) | 6940 (17.71) | |
| Comorbidity |
Chi‐squared = 1.58; df = 4 P = 0.81 |
||||||
| No | 3501 (44.67) | 3490 (44.53) | 3537 (45.13) | 3540 (45.17) | 3555 (45.36) | 17 623 (44.97) | |
| Yes | 4336 (55.33) | 4347 (55.47) | 4300 (54.87) | 4297 (54.83) | 4282 (54.64) | 21 562 (55.03) | |
| Energy intake (kJ/d) |
Chi‐squared = 22 000; df = 16 P < 0.001 |
||||||
| Q1 | 4634 (59.13) | 2029 (25.89) | 866 (11.05) | 255 (3.25) | 53 (0.68) | 7837 (20.00) | |
| Q2 | 1887 (24.08) | 2515 (32.09) | 1975 (25.20) | 1119 (14.28) | 341 (4.35) | 7837 (20.00) | |
| Q3 | 849 (10.83) | 1854 (23.66) | 2251 (28.72) | 1993 (25.43) | 890 (11.36) | 7837 (20.00) | |
| Q4 | 390 (4.98) | 1041 (13.28) | 1823 (23.26) | 2496 (31.86) | 2087 (26.63) | 7837 (20.00) | |
| Q5 | 77 (0.98) | 398 (5.08) | 922 (11.76) | 1974 (25.19) | 4466 (56.99) | 7837 (20.00) | |
Abbreviations: AHEI‐2010, Alternative Healthy Eating Index‐2010, AUS/NZ, Australia/New Zealand; BMI, body mass index; SEIFA, Socioeconomic Indices for Areas; WHR, waist to hip ratio.
Mean (SD).
During the first wave, 740 cases of diabetes were reported and 1249 cases in the second wave, giving a total of 1989 incident cases of self‐reported diabetes. The incidence of diabetes increased with age in both waves. In both Wave 1 and Wave 2, a higher incidence was observed among female participants, socioeconomically disadvantaged participants, current smokers, alcohol abstainers, participants with high BMI, with high WHR, in the low AHEI quintile, with family history of diabetes, with comorbidity, in the lowest fibre intake quintile, and those from southern Europe. All variables showed a significant association (P < 0.05) with incidence of type 2 diabetes in both Wave 1 and Wave 2, except energy intake quintiles in Wave 2 (P = 0.056; Table 2).
TABLE 2.
Incidence of diabetes in Waves 1 and 2 by possible predictor variable
| Wave 1 | Wave 2 | |||
|---|---|---|---|---|
| N (%) | P value | N (%) | P value | |
| Age | ||||
| <50 years | 122 (1.12) |
<0.001 |
276 (3.44) |
<0.001 |
| 50‐59 years | 263 (2.38) | 499 (6.44) | ||
| ≥60 years | 355 (3.04) | 447 (7.15) | ||
| Sex | ||||
| Male | 376 (2.81) | <0.001 | 591 (6.86) | <0.001 |
| Female | 364 (1.80) | 658 (4.78) | ||
| SEIFA quintile | ||||
| Q1 | 214 (3.63) |
<0.001 |
282 (8.40) |
<0.001 |
| Q2 | 188 (2.75) | 289 (7.22) | ||
| Q3 | 111 (2.08) | 189 (5.62) | ||
| Q4 | 98 (1.55) | 221 (4.89) | ||
| Q5 | 129 (1.4) | 268 (3.76) | ||
| Region of origin | ||||
| AUS/NZ/Other | 344 (1.45) |
<0.001 |
734 (4.47) |
<0.001 |
| Northern Europe | 41 (1.88) | 61 (4.03) | ||
| Southern Europe | 355 (4.60) | 454 (10.23) | ||
| Smoking status | ||||
| Never | 379 (1.92) |
<0.001 |
699 (5.14) |
<0.001 |
| Current smoker | 92 (2.68) | 142 (6.97) | ||
| Former smoker | 269 (2.60) | 408 (6.06) | ||
| Alcohol drinking | ||||
| Lifetime abstainer | 276 (2.92) |
<0.001 |
417 (7.01) |
<0.001 |
| Ex‐drinker | 91 (2.56) | 150 (6.43) | ||
| Current drinker | 373 (1.81) | 682 (4.84) | ||
| Physical activity | ||||
| 0 | 226 (3.15) |
<0.001 |
356 (7.67) |
<0.001 |
| >0 and <4 | 171 (254) | 273 (5.89) | ||
| ≥4 and <6 | 243 (2.03) | 430 (5.58) | ||
| ≥6 | 100 (1.29) | 190 (3.53) | ||
| WHR | ||||
| Normal | 238 (1.04) | P = 0.001 | 503 (3.27) | <0.001 |
| High | 502 (4.40) | 746 (10.65) | ||
| BMI | ||||
| < 25.0 kg/m2 | 63 (0.5) |
<0.001 |
133 (1.51) |
<0.001 |
| 25.0‐29.9 kg/m2 | 297 (2.06) | 541 (5.66) | ||
| ≥ 30.0 kg/m2 | 380 (5.84) | 575 (14.46) | ||
| AHEI‐2010 quintiles | ||||
| Q1 | 175 (2.62) |
<0.001 |
313 (7.37) |
<0.001 |
| Q2 | 189 (2.62) | 284 (6.18) | ||
| Q3 | 160 (2.47) | 242 (5.58) | ||
| Q4 | 122 (1.88) | 232 (5.26) | ||
| Q5 | 94 (1.40) | 168 (3.64) | ||
| Family history of diabetes | ||||
| No | 493 (1.79) | <0.001 | 865 (4.66) | <0.001 |
| Yes | 247 (4.14) | 384 (10.04) | ||
| Comorbidity | ||||
| No | 191 (1.24) | <0.001 | 403 (3.77) | <0.001 |
| Yes | 549 (3.01) | 846 (7.25) | ||
| Fibre | ||||
| Q1 | 181 (2.78) |
=0.001 |
270 (6.78) |
<0.001 |
| Q2 | 129 (1.93) | 263 (5.83) | ||
| Q3 | 131 (1.94) | 269 (5.86) | ||
| Q4 | 131 (1.92) | 222 (4.71) | ||
| Q5 | 168 (2.47) | 225 (4.93) | ||
| Energy intake | P = 0.002 | P = 0.056 | ||
| Q1 | 167 (2.52) | 256 (6.19) | ||
| Q2 | 131 (1.95) | 263 (5.87) | ||
| Q3 | 134 (1.97) | 229 (4.98) | ||
| Q4 | 127 (1.89) | 237 (5.12) | ||
| Q5 | 181 (2.69) | 264 (5.83) | ||
Abbreviations: AHEI‐2010, Alternative Healthy Eating Index‐2010; AUS/NZ, Australia/New Zealand; BMI, body mass index; SEIFA, Socioeconomic Indices for Areas; WHR, waist to hip ratio.
Table 3 presents the associations of quintiles of fibre intake (total, cereal, fruit, vegetables) with the incidence of diabetes. In Model 1, no association was found across total fibre quintiles (P for trend = 0.7). For cereal fibre, Q3 showed an 18% reduction of risk of diabetes (IRR 0.82, 95% CI 0.71‐0.94), Q4 showed a 22% reduction of risk of diabetes (IRR 0.78, 95% CI 0.67‐0.91) and Q5 showed 25% reduction of risk of diabetes (IRR 0.75, 95% CI 0.63‐0.88; P for trend = 0.003). For fruit fibre (P for trend = 0.018), only the second quintile (Q2) showed a 16% reduction in risk of diabetes (IRR 0.84, 95% CI 0.73‐0.96), whereas the other quintiles (Q3‐Q5) showed no association. No association was found across vegetable fibre quintiles.
TABLE 3.
Association of quintiles of total and different types of fibre with diabetes after controlling for confounder
| Model 1 a | Model 2: Model 1 + BMI b | Model 3: Model 2 + WHR c | |||||
|---|---|---|---|---|---|---|---|
| Category | Adjusted IRR (with 95% CI) | P‐value | Adjusted IRR (with 95% CI) | P value | Adjusted IRR (with 95% CI) | P value | |
| Total fibre | Q1 | Reference | Reference | Reference | |||
| Q2 | 0.91 (0.79‐1.05) | 0.22 | 0.91 (0.79‐1.05) | 0.22 | 0.92 (0.81‐1.05) | 0.24 | |
| Q3 | 1.00 (0.86‐1.17) | 0.95 | 1.01(0.87‐1.18) | 0.86 | 1.03 (0.90‐1.18) | 0.65 | |
| Q4 | 0.89 (0.74‐1.07) | 0.21 | 0.92 (0.71‐1.11) | 0.40 | 0.96 (0.83‐1.10) | 0.57 | |
| Q5 | 0.94 (0.76‐1.16) | 0.56 | 1.00 (0.81‐1.23) | 0.99 | 1.08 (0.94‐1.25) | 0.26 | |
| P for trend | 0.74 | 0.33 | 0.18 | ||||
| Cereal fibre | Q1 | Reference | Reference | Reference | |||
| Q2 | 0.89 (0.78‐1.02) | 0.09 | 0.91(0.80‐1.03) | 0.14 | 0.91 (0.80‐1.04) | 0.16 | |
| Q3 | 0.82 (0.71‐0.94) | 0.006* | 0.87 (0.75‐0.99) | 0.04 | 0.88 (0.77‐1.01) | 0.07 | |
| Q4 | 0.78 (0.67‐0.91) | 0.002* | 0.85 (0.73‐0.99) | 0.038 | 0.87 (0.75‐1.00) | 0.06 | |
| Q5 | 0.75 (0.63‐0.88) | 0.001* | 0.85 (0.72‐1.00) | 0.05 | 0.87 (0.74‐1.03) | 0.11 | |
| P for trend | 0.003* | 0.31 | 0.55 | ||||
| Fruit fibre | Q1 | Reference | Reference | Reference | |||
| Q2 | 0.84 (0.73‐0.96) | 0.014* | 0.84 (0.73‐0.97) | 0.02* | 0.85 (0.74‐1.02) | 0.07 | |
| Q3 | 0.91 (0.79‐1.04) | 0.18 | 0.93 (0.81‐1.07) | 0.30 | 0.94 (0.82‐1.08) | 0.38 | |
| Q4 | 0.94 (0.81‐1.09) | 0.43 | 0.95 (0.82‐1.09) | 0.46 | 0.97 (0.84‐1.12) | 0.67 | |
| Q5 | 0.87 (0.75‐1.02) | 0.09 | 0.88 (0.75‐1.03) | 0.10 | 0.89 (0.77‐1.04) | 0.15 | |
| P for trend | 0.34 | 0.68 | 0.90 | ||||
| Vegetable fibre | Q1 | Reference | Reference | Reference | |||
| Q2 | 1.07 (0.96‐1.22) | 0.27 | 1.09 (0.96‐1.24) | 0.19 | 1.09 (0.96‐1.24) | 0.18 | |
| Q3 | 0.97 (0.85‐1.12) | 0.71 | 0.97 (0.85‐1.12) | 0.72 | 0.96 (0.84‐1.10) | 0.59 | |
| Q4 | 1.09 (0.94‐1.26) | 0.24 | 1.08 (0.94‐1.24) | 0.29 | 1.07 (0.93‐1.23) | 0.36 | |
| Q5 | 1.06 (0.91‐1.24) | 0.44 | 1.05 (0.90‐1.22) | 0.51 | 1.05 (0.90‐1.22) | 0.54 | |
| P for trend | 0.54 | 0.45 | 0.45 | ||||
Abbreviations: BMI, body mass index; CI, confidence interval; IRR, incidence risk ratio; SEIFA, Socioeconomic Indices for Areas; WHR, waist to hip ratio. * p < 0.05.
Adjusted for age, sex, SEIFA, smoking status, drinking status, family history of diabetes and physical activity level at baseline, country of birth, quintiles of energy intake, comorbidity at baseline, Alternative healthy eating index quintiles.
Adjusted for all in 1 plus BMI.
Adjusted for all in 2 plus WHR.
In Model 2 and Model 3, we additionally adjusted for BMI and WHR, respectively. In both models, any association of cereal and fruit fibre with the incidence of diabetes disappeared.
We further conducted a mediation analysis using BMI and found that 36% of the effect of cereal fibre intake on incidence of type 2 diabetes was mediated by BMI (Figure 2).
FIGURE 2.
Mediation analyses of the association cereal fibre with risk of type 2 diabetes. BMI, body mass index; CI, confidence interval
4. DISCUSSION
In this study, we found a 25% reduction in the incidence of type 2 diabetes with higher cereal fibre intake (Q5 vs. Q1). We also found a smaller reduction in the incidence of type 2 diabetes with fruit fibre. However, total fibre and vegetable fibre intake were not associated with diabetes risk. After additional adjustment for measures of obesity, the associations with cereal and fruit fibre disappeared, suggesting BMI is probably along the causal pathway. Mediation analysis revealed that a significant proportion of the effect of cereal fibre intake was mediated by BMI.
Previous studies have assessed the effect of intake of whole grains on the incidence of type 2 diabetes and cardiovascular disease and hypothesized that the effect might be attributable to the fibre content. 34 , 35 In this study, we assessed the separate associations of total dietary fibre, and fibre from different sources (cereal, fruit and vegetables) with diabetes incidence. Intake of total dietary fibre was found to have no association with the incidence of diabetes. This finding is consistent with some 36 , 37 but not all previous studies. 38 , 39 The latter studies showed that a higher intake of total dietary fibre had an inverse association with the risk of developing type 2 diabetes. 38 , 39 The inconsistency may be attributable to differences in study design and population. Alessa et al conducted their study in women only. 38 The study by Lindstrom et al 39 used an interventional study design. As reported in previous studies, gender plays a role in the metabolism of dietary fibre. 40 , 41 Haro et al noted that there is a significant difference in gut microbiota composition in men and women, which may explain the difference in risk of metabolic diseases among men and women. 40
In our study, cereal fibre intake was associated with reduced diabetes risk. This finding is consistent with findings of previous research that reported an inverse association between the incidence of diabetes and cereal fibre intake. 4 , 11 , 42 In contrast, results from the ARIC study, a US study of 16 000 individuals followed up for 30 years, showed no association between cereal fibre and diabetes incidence. 36 The difference in the populations and dietary assessment may explain the inconsistent finding. The FFQ used in the current study included 121 items, whereas the ARIC study used an FFQ with only 61 items. Moreover, the ARIC study included a significant number of African‐American participants. African‐American people reportedly consume significantly lower amounts of dietary fibre compared to those of other races/ethnicities, 43 which may explain the lack of an association of cereal fibre with diabetes in the ARIC study.
Our study showed a weak association between fruit fibre intake and diabetes, with the risk being lower in Q2 than in Q1. Other similar studies found no association for fruit fibre. 37 , 43 In addition, meta‐analyses from Wang et al and the European Prospective Investigation into Cancer and Nutrition (EPIC)–InterAct consortium study also reported no association between fruit fibre intake and diabetes risk. 44 , 45 Our finding contrasts with the study conducted in Taiwan that reported an inverse association between fruit fibre intake and diabetes. 46 The inconsistency may be attributable to the study follow‐up time difference. The follow‐up time of the Weng et al study was 4.6 years, in contrast to our study, for which the mean follow‐up time was almost 14 years. In addition, most of the studies included in both meta‐analyses have a follow‐up time of >8 years. 44 , 45
In our study, vegetable fibre intake was found to have no association with the incidence of diabetes. This result is consistent with the different studies, including the EPIC–InterAct consortium study. 39 , 42 , 43 , 44 However, several studies have reported inverse associations between vegetable fibre and diabetes risk. 45 , 47 , 48 The difference in the study populations may be an explanation for the inconsistency. Barclay et al conducted the study in the older Australian population (age > 49 years), 47 which would result in more incident cases of diabetes compared to our study, where nearly one‐third of the participants were aged <50 years. Hopping et al 48 reported a significant association among men. In our study, nearly 60% of the participants were female. In addition, the effect of fibre on metabolic disease risk differs among men and women due to differences in the composition of their gut microbiomes, with a more positive effect of fibre being observed in women. 40 , 41
An association between cereal fibre and the incidence of diabetes may be mediated via a change in the gut microbiome. 10 A healthy gut microbiome contributes to the health of the immune system. By contrast, abnormal microbiomes (dysbiosis) are associated with inflammation, poor immune function, and metabolic problems. Dysbiosis results from an unhealthy diet, use of antibiotics, and other causes. 49 Dietary fibre intake is associated with increased gut microbial diversity, abundance, and stability. 50 , 51 A recent cross‐sectional study also reported that higher microbiome α diversity, along with more butyrate‐producing gut bacteria, was associated with a lower prevalence of type 2 diabetes and lower insulin resistance among individuals without diabetes. 52 , 53 Other studies have also highlighted that a high cereal fibre intake increases the number of butyrate‐producing bacterial colonies, 52 , 54 which might be one mechanism for the association with diabetes risk.
Food sources of soluble fibre are mainly fruit and vegetables, whereas insoluble fibre is mainly found in cereals and grains. Generally, insoluble fibres increase faecal volume and bile acid excretion, whereas the transit time through the human digestive system is generally prolonged by soluble fibres, delaying stomach emptying. 8 Insoluble fibres, such as cellulose, increase the relative abundance of gut bacteria, while soluble fibre mainly stimulates the growth of gut bacteria responsible for producing short‐chain fatty acids. Furthermore, feeding soluble fibre led to a higher concentration of short‐chain fatty acids and microbial diversity. 55 , 56
Although there is evidence from prospective studies which assessed the association of dietary fibre intake with diabetes incidence, no study directly assessed the effect of fibre on the incidence of diabetes. However, some studies examined the effect of different fibre supplements on metabolic markers using randomized clinical trials. 46 , 57 , 58 , 59 All of them reported an inverse relationship between fibre intake and metabolic markers. However, the OptiFiT study only showed a trendwise association due to lack of power. 57 However, the above trial studies 57 , 58 mainly included individuals at high risk of developing diabetes. They also highlighted the need for further studies with large sample sizes.
The findings of this study can still be applicable to the current Australian population. We have used baseline dietary intake measures for an average follow‐up of 13.8 years for diabetes outcome. Although participants might change their dietary habits after finishing the FFQ, the literature suggests dietary intake is quite stable over time. 60 , 61 In addition, average total fibre intake was 31 g in the MCCS at baseline, and the most up‐to‐date Australian data from 2011 to 2012 62 show that average fibre intake for adults is 25 g/d, suggesting that people are still eating enough to see benefit.
Strengths of this study include the use of a large population‐based prospective cohort with a long follow‐up period. All the results were reported after adjustment for many possible confounders. Anthropometric data were measured rather than self‐reported, which is common in large cohort studies. In addition, this study indicated a causal pathway via BMI with mediation analysis.
Our study also has some limitations. We used self‐reported dietary data from an FFQ, which is known to measure intake with considerable error. Diabetes was self‐reported, albeit the participant's nominated doctor validated the diagnosis at the first follow‐up. Given the age of the research participants, we considered that all incident cases were type 2 diabetes. The numbers within each country‐of‐birth stratum were modest, and Greek‐born participants, who had the most significant incidence of diabetes at the first follow‐up, had lower attendance at the second follow‐up. 20
In conclusion, higher intakes of cereal fibre and, to a lesser extent, fruit fibre, may reduce the risk of type 2 diabetes. However, further adjustment for BMI and WHR eliminated the association. In mediation analysis, BMI mediated 36% of the relationship between cereal fibre and type 2 diabetes. These results suggest that more specific recommendations regarding fibre intake may be helpful, along with other lifestyle modifications to prevent diabetes. In addition, the results imply the need for further studies, including clinical trials investigating the effects of different fibre types on type 2 diabetes risk.
AUTHOR CONTRIBUTIONS
Robel Hussen Kabthymer, Md Nazmul Karim, Allison M. Hodge and Barbora de Courten planned the analysis, Robel Hussen Kabthymer and Md Nazmul Karim conducted the statistical analysis, All authors interpreted the results. Robel Hussen Kabthymer wrote the first draft of the manuscript. All authors revised the manuscript. Barbora de Courten had primary responsibility for final content. All authors read and approved the final manuscript.
FUNDING INFORMATION
The MCCS cohort recruitment was funded by VicHealth and Cancer Council Victoria. The MCCS was further augmented by Australian National Health and Medical Research Council grants 209057, 396414 and 1074383 and by infrastructure provided by Cancer Council Victoria.
CONFLICTS OF INTEREST
No conflict of interest to declare.
PEER REVIEW
The peer review history for this article is available at https://www.webofscience.com/api/gateway/wos/peer‐review/10.1111/dom.15054.
ETHICS STATEMENT
This study was conducted according to the guidelines laid down in the Declaration of Helsinki, and all procedures involving research participants were approved by the Cancer Council Victoria Human Research Ethics Committee. Written informed consent was obtained from all participants. The study received approval from the Monash University Human Research Ethics Committee.
ACKNOWLEDGMENT
Open access publishing facilitated by Monash University, as part of the Wiley ‐ Monash University agreement via the Council of Australian University Librarians.
Kabthymer RH, Karim MN, Hodge AM, de Courten B. High cereal fibre but not total fibre is associated with a lower risk of type 2 diabetes: Evidence from the Melbourne Collaborative Cohort Study. Diabetes Obes Metab. 2023;25(7):1911‐1921. doi: 10.1111/dom.15054
Allison M. Hodge and Barbora de Courten are senior authors.
DATA AVAILABILITY STATEMENT
Details on how to access data for the Melbourne Collaborative Cohort Study are available at: https://www.cancervic.org.au/research/epidemiology/pedigree.
REFERENCES
- 1. International Diabetes Foundation . IDF Diabetes Atlas. 10th ed. Belgium Brussels: IDF; 2021. [Google Scholar]
- 2. Australian Institute of Health and Welfare . Diabetes: Australian Facts. Canberra, Australia: AIHW; 2022. [Google Scholar]
- 3. NCD Risk Factor Collaboration (NCD‐RisC) . Trends in adult body‐mass index in 200 countries from 1975 to 2014: a pooled analysis of 1698 population‐based measurement studies with 19.2 million participants. Lancet. 2016;387(10026):1377‐1396. doi: 10.1016/S0140-6736(16)30054-X [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Weickert MO, Pfeiffer AFH. Impact of dietary fibre consumption on insulin resistance and the prevention of type 2 diabetes. J Nutr. 2018;148(1):7‐12. doi: 10.1093/jn/nxx008 [DOI] [PubMed] [Google Scholar]
- 5. Vitale M, Giacco R, Laiola M, et al. Acute and chronic improvement in postprandial glucose metabolism by a diet resembling the traditional Mediterranean dietary pattern: can SCFAs play a role? Clin Nutr. 2021;40(2):428‐437. doi: 10.1016/j.clnu.2020.05.025 [DOI] [PubMed] [Google Scholar]
- 6. Haro C, Montes‐Borrego M, Rangel‐Zúñiga OA, et al. Two healthy diets modulate gut microbial community improving insulin sensitivity in a human obese population. J Clin Endocrinol Metab. 2016;101:233‐242. doi: 10.1210/jc.2015-3351 [DOI] [PubMed] [Google Scholar]
- 7. Hald S, Schioldan AG, Moore ME, et al. Effects of arabinoxylan and resistant starch on intestinal microbiota and short‐chain fatty acids in subjects with metabolic syndrome: a randomized crossover study. Plos One. 2016;11(7):e0159223. doi: 10.1371/journal.pone.0159223 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Mudgil D, Barak S. Composition, properties and health benefits of indigestible carbohydrate polymers as dietary fibre: a review. Int J Biol Macromol. 2013;61:1‐6. doi: 10.1016/j.ijbiomac.2013.06.044 [DOI] [PubMed] [Google Scholar]
- 9. Zhao L, Zhang F, Ding X, et al. Gut bacteria selectively promoted by dietary fibres alleviate type 2 diabetes. Science. 2018;359:1151‐1156. doi: 10.1126/science.aao5774 [DOI] [PubMed] [Google Scholar]
- 10. Jenkins DJ, Kendall CW, McKeown‐Eyssen G, et al. Effect of a low‐glycemic index or a high‐cereal fibre diet on type 2 diabetes: a randomized trial. JAMA. 2008;300(23):2742‐2753. doi: 10.1001/jama.2008.808 [DOI] [PubMed] [Google Scholar]
- 11. Davison KM, Temple NJ. Cereal fibre, fruit fibre, and type 2 diabetes: explaining the paradox. J Diabetes Complications. 2018;32:240‐245. doi: 10.1016/j.jdiacomp.2017.11.002 [DOI] [PubMed] [Google Scholar]
- 12. Allison M, Hodge DRE, O'dea K, Graham G. Glycemic index and dietary fibre and the risk of type 2 diabetes. Diabetes Care. 2004;27(11):2701‐2706. doi: 10.2337/diacare.27.11.2701 [DOI] [PubMed] [Google Scholar]
- 13. Flint HJ, Duncan SH, Scott KP, Louis P. Interactions and competition within the microbial community of the human colon: links between diet and health. Environ Microbiol. 2007;9(5):1101‐1111. doi: 10.1111/j.1462-2920.2007.01281.x [DOI] [PubMed] [Google Scholar]
- 14. Louis P, Scott KP, Duncan SH, Flint HJ. Understanding the effects of diet on bacterial metabolism in the large intestine. J Appl Microbiol. 2007;102(5):1197‐1208. doi: 10.1111/j.1365-2672.2007.03322.x [DOI] [PubMed] [Google Scholar]
- 15. Hills RD, Pontefract BA, Mishcon HR, Black CA, Sutton SC, Theberge CR. Gut microbiome: profound implications for diet and disease. Nutrients. 2019;11(7):1‐40. doi: 10.3390/nu11071613 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Wannamethee SG, Whincup PH, Thomas MC, Sattar N. Associations between dietary fibre and inflammation, hepatic function, and risk of type 2 diabetes in older men: potential mechanisms for the benefits of fibre on diabetes risk. Diabetes Care. 2009;32(10):1823‐1825. doi: 10.2337/dc09-0477 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Massimino SP, McBurney MI, Field CJ, et al. Fermentable dietary fibre increases GLP‐1 secretion and improves glucose homeostasis despite increased intestinal glucose transport capacity in healthy dogs. J Nutr. 1998;128:1786‐1793. doi: 10.1093/jn/128.10.1786 [DOI] [PubMed] [Google Scholar]
- 18. Du H, Vimaleswaran KS, Angquist L, Hansen RD, et al. Dietary fibre and subsequent changes in body weight and waist circumference in European men and women. Am J Clin Nutrit. 2010;91(2):329‐336. doi: 10.3945/ajcn.2009.28191 [DOI] [PubMed] [Google Scholar]
- 19. Davis JN, Alexander KE, Ventura EE, Toledo‐Corral CM, Goran MI. Inverse relation between dietary fibre intake and visceral adiposity in overweight Latino youth. Am J Clin Nutrit. 2009;90(5):1160‐1166. doi: 10.3945/ajcn.2009.28133 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Milne RL, Fletcher AS, MacInnis RJ, et al. Cohort profile: the Melbourne collaborative cohort study (health 2020). Int J Epidemiol. 2017;46(6):1757‐i. doi: 10.1093/ije/dyx085 [DOI] [PubMed] [Google Scholar]
- 21. Ireland P, Jolley D, Giles G, et al. Development of the Melbourne FFQ: a food frequency questionnaire for use in an Australian prospective study involving an ethnically diverse cohort. Asia Pac J Clin Nutr. 1994;3(1):19‐31. [PubMed] [Google Scholar]
- 22. Hodge AM, Simpson JA, Fridman M, et al. Evaluation of an FFQ for assessment of antioxidant intake using plasma biomarkers in an ethnically diverse population. Public Health Nutr. 2009;12(12):2438‐2447. doi: 10.1017/S1368980009005539 [DOI] [PubMed] [Google Scholar]
- 23. Lewis J, Milligan G, Hunt A. NUTTAB95 Nutrient Data Table for Use in Australia. Canberra, Australia: Australian Government Publishing Service; 1995. [Google Scholar]
- 24. Holland B, Welch AA, Unwin ID, et al. McCance and Widdowson's the Composition of Foods. 5th ed. Cambridge, UK: Royal Society of Chemistry; 1993. [Google Scholar]
- 25. Fatty Aacid Compositional Database [Internet]. Xyris Software. 2001. http://www.xyris.com.au/fatty_acids/default.htm
- 26. Holden JM, Eldridge AL, Beecher GR, et al. Carotenoid content of u.s. foods: an update of the database. J Food Compos Anal. 1999;12(3):169‐196. doi: 10.1006/jfca.1999.0827 [DOI] [Google Scholar]
- 27. Chiuve SE, Fung TT, Rimm EB, et al. Alternative dietary indices both strongly predict risk of chronic disease. J Nutr. 2012;142(6):1009‐1018. doi: 10.3945/jn.111.157222 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. McNutt LA, Wu C, Xue X, Hafner JP. Estimating the relative risk in cohort studies and clinical trials of common outcomes. Am J Epidemiol. 2003;157(10):940‐943. doi: 10.1093/aje/kwg074 [DOI] [PubMed] [Google Scholar]
- 29. Zou G. A modified Poisson regression approach to prospective studies with binary data. Am J Epidemiol. 2004;159(7):702‐706. doi: 10.1093/aje/kwh090 [DOI] [PubMed] [Google Scholar]
- 30. Emsley R, Liu H. PARAMED: Stata module to perform causal mediation analysis using parametric regression models. 2013. https://econpapers.repec.org/software/bocbocode/s457581.htm
- 31. Valeri L, Vanderweele TJ. Mediation analysis allowing for exposure‐mediator interactions and causal interpretation: theoretical assumptions and implementation with SAS and SPSS macros. Psychol Methods. 2013;18:137‐150. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Ikram MA, VanderWeele TJ. A proposed clinical and biological interpretation of mediated interaction. Eur J Epidemiol. 2015;30:1115‐1118. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33. VanderWeele TJ. Explanation in causal inference: developments in mediation and interaction. Int J Epidemiol. 2016;45(6):1904‐1908. doi: 10.1093/ije/dyw277 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34. De Munter JS, Hu FB, Spiegelman D, Franz M, van Dam RM. Whole grain, bran, and germ intake and risk of type 2 diabetes: a prospective cohort study and systematic review. PLoS Med. 2006;4(8):1385‐1395. doi: 10.1371/journal.pmed.0040261 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. McKeown NM, Meigs JB, Liu S, Wilson PW, Jacques PF. Whole‐grain intake is favorably associated with metabolic risk factors for type 2 diabetes and cardiovascular disease in the Framingham offspring study. Am J Clin Nutr. 2002;76(2):390‐398. doi: 10.1093/ajcn/76.2.390 [DOI] [PubMed] [Google Scholar]
- 36. Stevens J, Ahn K, Juhaeri HD, Steffan L, Couper D. Dietary fibre intake and glycemic index and incidence of diabetes in African‐American and white adults: the ARIC study. Diabetes Care. 2002;25(10):1715‐1721. doi: 10.2337/diacare.25.10.1715 [DOI] [PubMed] [Google Scholar]
- 37. Meyer KA, Kushi LH, Jacobs DR, Slavin J, Sellers TA, Folsom AR. Carbohydrates, dietary fibre, and incident type 2 diabetes in older women. Am J Clin Nutr. 2000;71(4):921‐930. doi: 10.1093/ajcn/71.4.921 [DOI] [PubMed] [Google Scholar]
- 38. AlEssa HB, Bhupathiraju SN, Malik VS, et al. Carbohydrate quality and quantity and risk of type 2 diabetes in US women. Am J Clin Nutr. 2015;102(6):1543‐1553. doi: 10.3945/ajcn.115.116558 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39. Lindstrom J, Peltonen M, Eriksson JG, et al. High‐fibre, low‐fat diet predicts long‐term weight loss and decreased type2 diabetes risk: the Finnish diabetes prevention study. Diabetologia. 2006;49(5):912‐920. doi: 10.1007/s00125-006-0198-3 [DOI] [PubMed] [Google Scholar]
- 40. Haro C, Rangel‐Zúñiga OA, Alcalá‐Díaz JF, et al. Intestinal microbiota is influenced by gender and body mass index. PLoS One. 2016;11(5):1‐16. doi: 10.1371/journal.pone.0154090 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41. Benítez‐Páez A, Hess AL, Krautbauer S, et al. Sex, food, and the gut microbiota: disparate response to caloric restriction diet with fibre supplementation in women and men. Mol Nutr Food Res. 2021;65(8):2000996. doi: 10.1002/mnfr.202000996 [DOI] [PubMed] [Google Scholar]
- 42. Montonen J, Knekt P, Järvinen R, Aromaa A, Reunanen A. Whole‐grain and fibre intake and the incidence of type 2 diabetes. Am J Clin Nutr. 2003;77(3):622‐629. doi: 10.1093/ajcn/77.3.622 [DOI] [PubMed] [Google Scholar]
- 43. Schulze MB, Schulz M, Heidemann C, Schienkiewitz A, Hoffmann K, Boeing H. Fibre and magnesium intake and incidence of type 2 diabetes: a prospective study and meta‐analysis. AMA Arch Intern Med. 2007;167(9):956‐965. doi: 10.1001/archinte.167.9.956 [DOI] [PubMed] [Google Scholar]
- 44. InterAct Consortium . Dietary fibre and incidence of type 2 diabetes in eight European countries: the EPIC‐InterAct study and a meta‐analysis of prospective studies. Diabetologia. 2015;58(7):1394‐1408. doi: 10.1007/s00125-015-3585-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 45. Wang PY, Fang JC, Gao ZH, Zhang C, Xie SY. Higher intake of fruits, vegetables or their fibre reduces the risk of type 2 diabetes: a meta‐analysis. J Diabetes Investig. 2016;7(1):56‐69. doi: 10.1111/jdi.12376 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 46. Weng LC, Lee NJ, Yeh WT, Ho LT, Pan WH. Lower intake of magnesium and dietary fibre increases the incidence of type 2 diabetes in Taiwanese. J Formos Med Assoc. 2012;111(11):651‐659. doi: 10.1016/j.jfma.2012.07.038 [DOI] [PubMed] [Google Scholar]
- 47. Barclay AW, Flood VM, Rochtchina E, Mitchell P, Brand‐Miller JC. Glycemic index, dietary fibre, and risk of type 2 diabetes in a cohort of older Australians. Diabetes Care. 2007;30(11):2811‐2813. doi: 10.2337/dc07-0784 [DOI] [PubMed] [Google Scholar]
- 48. Hopping BN, Erber E, Grandinetti A, Verheus M, Kolonel LN, Maskarinec G. Dietary fibre, magnesium, and glycemic load alter risk of type 2 diabetes in a multiethnic cohort in Hawaii. J Nutr. 2010;140(1):68‐74. doi: 10.3945/jn.109.112441 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 49. Hawrelak JA, Myers SP. The causes of intestinal dysbiosis: a review. Altern Med Rev. 2004;9(2):180‐197. [PubMed] [Google Scholar]
- 50. Martínez I, Lattimer JM, Hubach KL, et al. Gut microbiome composition is linked to whole grain‐induced immunological improvements. ISME J. 2013;7(2):269‐280. doi: 10.1038/ismej.2012.104 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 51. Tap J, Furet JP, Bensaada M, et al. Gut microbiota richness promotes its stability upon increased dietary fibre intake in healthy adults. Environ Microbiol. 2015;17(12):4954‐4964. doi: 10.1111/1462-2920.13006 [DOI] [PubMed] [Google Scholar]
- 52. Iversen KN, Dicksved J, Zoki C, et al. Effects of high fibre Rye, compared to refined wheat, on gut microbiota composition, plasma short chain fatty acids, and implications for weight loss and metabolic risk factors (the RyeWeight study). Nutrients. 2022;14(8):1669. doi: 10.3390/nu14081669 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 53. Naderpoor N, Mousa A, Gomez‐Arango LF, Barrett HL, Dekker Nitert M, de Courten B. Faecal microbiota are related to insulin sensitivity and secretion in overweight or obese adults. Journal of. Clin Med. 2019;8(4):452. doi: 10.3390/jcm8040452 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 54. Jefferson A, Adolphus K. The effects of intact cereal grain Fibres, including wheat bran on the gut microbiota composition of healthy adults: a systematic review. Front Neurol. 2019;6:33. doi: 10.3389/fnut.2019.00033 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 55. Holscher HD. Dietary fibre and prebiotics and the gastrointestinal microbiota. Gut Microbes. 2017;8(2):172‐184. doi: 10.1080/19490976.2017.1290756 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 56. So D, Whelan K, Rossi M, et al. Dietary fibre intervention on gut microbiota composition in healthy adults: a systematic review and meta‐analysis. Am J Clin Nutr. 2018;170(6):965‐983. doi: 10.1093/ajcn/nqy041 [DOI] [PubMed] [Google Scholar]
- 57. Honsek C, Kabisch S, Kemper M, et al. Fibre supplementation for the prevention of type 2 diabetes and improvement of glucose metabolism: the randomised controlled optimal fibre trial (OptiFiT). Diabetologia. 2018;61(6):1295‐1305. doi: 10.1007/s00125-018-4582-6 [DOI] [PubMed] [Google Scholar]
- 58. Weickert MO, Roden M, Isken F, et al. Effects of supplemented isoenergetic diets differing in cereal fibre and protein content on insulin sensitivity in overweight humans. Am J Clin Nutr. 2011;94(2):459‐471. doi: 10.3945/ajcn.110.004374 [DOI] [PubMed] [Google Scholar]
- 59. Gibb RD, McRorie JW Jr, Russell DA, et al. Psyllium fibre improves glycemic control proportional to loss of glycemic control: a meta‐analysis of data in euglycemic subjects, patients at risk of type 2 diabetes mellitus, and patients being treated for type 2 diabetes mellitus. Am J Clin Nutr. 2015;102(6):1604‐1614. doi: 10.3945/ajcn.115.106989 [DOI] [PubMed] [Google Scholar]
- 60. Davis JA, Mohebbi M, Collier F, et al. The role of diet quality and dietary patterns in predicting muscle mass and function in men over a 15‐year period. Osteoporos Int. 2021;32(11):2193‐2203. doi: 10.1007/s00198-021-06012-3 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 61. Baldwin JN, Forder PM, Haslam RL, et al. Change in diet quality over 12 years in the 1946‐1951 cohort of the Australian longitudinal study on Women's health. Nutrients. 2020;12(1):147. doi: 10.3390/nu12010147 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 62. Australian Bureau of Statistics . Australian Health Survey: Nutrition First Results‐Foods and Nutrients [Internet]. Canberra: ABS. 2011. https://www.abs.gov.au/statistics/health/health‐conditions‐and‐risks/australian‐health‐survey‐nutrition‐first‐results‐foods‐and‐nutrients/latest‐release
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
Details on how to access data for the Melbourne Collaborative Cohort Study are available at: https://www.cancervic.org.au/research/epidemiology/pedigree.