Guest guest Posted March 21, 2008 Report Share Posted March 21, 2008 Hi All, Love those legumes. Root vegetables in this study did not include potatoes. The pdf of the below paper is now available. Ute Nöthlings, Matthias B. Schulze, Cornelia Weikert, Heiner Boeing, T. van der Schouw, Bamia, Vasiliki Benetou, Pagona Lagiou, Vittorio Krogh, Joline W. J. Beulens, Petra H. M. Peeters, Jytte Halkjær, Anne Tjønneland, rio Tumino, Salvatore Panico, Giovanna Masala, Francoise Clavel-Chapelon, Blandine de Lauzon, Marie- Boutron-Ruault, Marie-Noël Vercambre, Rudolf Kaaks, Jakob Linseisen, Kim Overvad, Larraitz Arriola, Eva Ardanaz, A. , Marie- Tormo, Sheila Bingham, Kay-Tee Khaw, Tim J. A. Key, Paolo Vineis, Elio Riboli, Pietro Ferrari, Paolo Boffetta, H. Bas Bueno-de-Mesquita, Daphne L. van der A, Göran Berglund, bet Wirfält, Göran Hallmans, Ingegerd Johansson, Eiliv Lund, and Antonia Trichopoulo Intake of Vegetables, Legumes, and Fruit, and Risk for All-Cause, Cardiovascular, and Cancer Mortality in a European Diabetic PopulationJ. Nutr. 2008 138: 775-781. ABSTRACT We examined the associations of intake of vegetables, legumes and fruit with all-cause and cause-specific mortality in a population with prevalent diabetes in Europe. A cohort of 10,449 participants with self-reported diabetes within the European Prospective Investigation into Cancer and Nutrition study was followed for a mean of 9 y. Intakes of vegetables, legumes, and fruit were assessed at baseline between 1992 and 2000 using validated country-specific questionnaires. A total of 1346 deaths occurred. Multivariate relative risks (RR) for all-cause mortality were estimated in regression models and RR for cause-specific mortality were derived in a competing risk model. An increment in intake of total vegetables, legumes, and fruit of 80 g/d was associated with a RR of death from all causes of 0.94 [95% CI 0.90–0.98]. Analyzed separately, vegetables and legumes were associated with a significantly reduced risk, whereas nonsignificant inverse associations for fruit intake were observed. Cardiovascular disease (CVD) mortality and mortality due to non-CVD/non-cancer causes were significantly inversely associated with intake of total vegetables, legumes, and fruit (RR 0.88 [95% CI 0.81–0.95] and 0.90 [0.82–0.99], respectively) but not cancer mortality (1.08 [0.99–1.17]). Intake of vegetables, legumes, and fruit was associated with reduced risks of all-cause and CVD mortality in a diabetic population. The findings support the current state of evidence from general population studies that the protective potential of vegetable and fruit intake is larger for CVD than for cancer and suggest that diabetes patients may benefit from a diet high in vegetables and fruits. Abbreviations used: CVD, cardiovascular disease; EPIC, European Prospective Investigation into Cancer and Nutrition; ICD-10, 10th revision of the International Statistical Classification of Diseases, Injuries and Causes of Death; RR, relative risk. Introduction Dietary recommendations are an integral part in the treatment and management of diabetes mellitus (1), although data on their efficacy are limited (2). The objective of diabetes treatment is to reduce the risk of long-term vascular complications (3), predominantly cardiovascular diseases (CVD).29 Increased risks of CVD (4,5) and a higher mortality rate than in the general population have been consistently found in diabetic populations (5,6). Recently, evidence for a higher risk of cancer at various sites has been reported to occur in diabetic compared with nondiabetic individuals (7–9). Vegetable and fruit intakes have been extensively investigated as risk factors for cancer (10), CVD (11,12), or mortality (13–15) in the general population; however, data on associations between dietary behavior, including vegetables and fruit intake, and disease outcomes in diabetic populations are scarce (2). Therefore, we analyzed data from the European Prospective Investigation into Cancer and Nutrition (EPIC) study to examine the associations between intake of vegetables, legumes, and fruit, and risk for all-cause and cause-specific mortality in a population of diabetic individuals. Subjects, Materials, and Methods ... EPIC is an ongoing multicenter prospective cohort study designed to investigate associations between diet and other lifestyle behaviors with chronic diseases, especially cancer. A detailed description of the study design and methods used can be found elsewhere (16). In brief, between 1992 and 2000, >500,000 study participants were recruited in 23 study centers located in 10 European countries to be followed for cancer incidence and cause-specific mortality. Participants ranging from 35 to 70 y were recruited in study centers in France, Germany, Greece, Italy, The Netherlands, Spain, United Kingdom, Sweden, Denmark, and Norway. Study populations were population-based samples from designated geographic areas (e.g. town or province), with the exceptions of France (members of the health insurance for state school employees), subsamples of the Italian and Spanish cohort (blood donors), Utrecht (The Netherlands), Florence (Italy; women invited for breast cancer screening), and a subsample in Oxford (health-conscious group, vegetarians). Individuals signed informed consent forms, after which self-administered diet and lifestyle questionnaires were mailed to the participants. In the Spanish centers, Ragusa, Naples, and Greece, interviewer-administered questionnaires were used. All participants were invited to a study center for an examination that included anthropometric and blood pressure measurements. Approval for this study was obtained from the ethical review boards of the International Agency for Research on Cancer and from all local institutions where subjects had been recruited for the EPIC study. Study population. The study population for this study was the group of participants in the EPIC cohort reporting a diagnosis of diabetes mellitus at baseline (n = 13,838). After exclusion of diabetic participants missing dietary questionnaire information (n = 144) or date of death (n = 3) or with implausible energy intake (top or bottom 1% of the ratio of energy intake:energy requirement; n = 437), a total of 13,254 participants with self-reported diabetes mellitus were identified. ... 10,449 participants with self-reported diabetes mellitus were identified for this analysis. ... Results Mean age at baseline ranged from 56 to 59 y across quartiles of vegetable, legume, and fruit intake, not showing a trend (Table 1). The percentage of diabetic participants treated with insulin was inversely associated with intake of vegetables, legumes, and fruit, ranging from 32% in the lowest quartile to 16% in the highest. The percentage of diabetics with a self-reported heart attack at baseline was also inversely associated with intake of total vegetables, legumes, and fruit. There was a positive association with self-reported absence of hypertension at baseline and hyperlipidemia with vegetable, legume, and fruit intake. A total of 1346 deaths occurred, 517 due to circulatory diseases, 319 due to cancer, and 323 due to other specified causes. TABLE 1 Baseline characteristics of participants with self-reported diabetes^1.======================================= ---Quartile of fruit, vegetables, and legumes intake^2=======================================Characteristic---1 2 3 4=======================================n 2612 2612 2613 2612Men, % 59 36 36 53Age, y 57±8 59±9 59±9 56±8Age at diabetes diagnosis, y 47±13 49±13 50±12 48±11BMI, kg/m 2 28.3±4.9 28.8±5.3 29.0±5.1 29.1±4.7Waist-to-hip ratio 0.93±0.09 0.90±0.09 0.90±0.09 0.91±0.09Insulin treatment, % Yes 32 20 17 16 No 51 61 67 73 Unknown 18 19 16 11Heart attack, % Yes 9 6 5 4 No 79 87 91 94Cancer, % Yes 5 4 3 1 No 78 86 90 95Hypertension, % Yes 35 45 45 40 No 33 41 49 57Hyperlipidemia, % Yes 27 38 39 42 No 41 50 54 54Smoking status, % Never 34 52 55 51 Former, .10 y 26 21 18 15 Former, #10 y 13 10 10 13 Current, ,20 cigarettes/d 15 10 9 13 Current, $20 cigarettes/d 12 7 7 7 Unknown 0 1 1 1Physical activity, % Inactive 16 11 10 14 Moderately inactive 36 31 29 28 Moderately active 40 51 54 50 Active 6 7 6 8 Missing 2 1 0 0Energy intake, kJ/d 8816±1749 7987±1887 7945±1669 9096±1992======================================= 1 Values are percentages or mean±SD; n=10,449. 2 Differences across quartiles were tested with the chi-square test for categorical variables and with the t test for slope in linear regression models of mean values on intake of fruit, vegetables, and legumes for continuous variables. P <0.01 for all variables. Due to the large sample size, all tests were significant, so all variables differed Intake of total vegetables, legumes, and fruit was inversely associated with risk for all-cause mortality (Table 2). An increment in intake by 80 g/d was associated with a significant risk reduction of 6% ([RR 0.94 [95% CI 0.90–0.98]). The respective RR were 0.95 [0.89–1.00] for men and 0.93 [0.85–1.03] for women. Analyzed separately, intakes of vegetables and legumes were also inversely associated with risk for all-cause mortality (Table 2). Associations for total fruit intake were inverse but did not reach significance in the multivariate model (P = 0.42). In a model including vegetables, legumes, and fruit as separate variables, RR were below 1, with a significant RR for vegetables (data not shown). Additional adjustment for physical activity (inactive, moderately inactive, moderately active, active, missing), self-reported hyperlipidemia, or educational attainment changed RR estimates only marginally (data not shown). Adjustment for actual measured blood pressure, which was available for approximately two-thirds of the diabetic population, did not lead to different conclusions (data not shown). RR were essentially the same when we replaced waist-to-hip ratio with waist-to-height ratio, BMI, or waist circumference in multivariate models (data not shown). No evidence for nonlinear associations was found for any of the food group variables (data not shown). Sex, smoking status, age at diagnosis, and waist-to-hip ratio did not modify the associations between intake of fruits and vegetables and mortality risk (data not shown). Associations among participants 60 y of age (n = 4591) were similar to results for the overall group and reached significance for legumes only (data not shown). To confirm our findings, we calculated a multivariate-adjusted regression model with observed intake of total vegetables, legumes, and fruit, i.e. without regression calibration. For the fully adjusted model, RR [95% CI] were 0.83 [0.71–0.99] for the 2nd quartile compared with the lowest, 0.75 [0.62–0.90] for the 3rd quartile, and 0.78 [0.62–0.97] for the highest quartile (P-trend = 0.046). TABLE 2 RR [95% CI] for intake of vegetables, legumes, and fruit, and all-cause mortality in a diabetic population^1.========================================================= ---Quartiles of food intake---Food group---1 2 3 4---PTrend Continuous exposure^2=========================================================Vegetables, legumes, and fruitMedian intake, g/d 283 390 474 630Cases/person-years 478/ 3,890 315/22,065 326/22,399 227/25,174Sex, age, energy-adjusted RR [95% CI] 1 0.73 [0.62–0.87] 0.79 [0.66–0.95] 0.75 [0.59–0.94] 0.02 0.93 [0.89–0.97]Multivariate adjusted RR [95% CI] 1 0.75 [0.63–0.89] 0.80 [0.66–0.97] 0.76 [0.60–0.96] 0.03 0.94 [0.90–0.98] Vegetables Median intake, g/d 127 164 198 259Cases/person-years 432/24,028 333/22,839 271/22,933 310/23,730Sex, age, energy-adjusted RR [95% CI] 1 0.86 [0.73–1.01] 0.73 [0.60–0.88] 0.81 [0.65–1.01] 0.05 0.88 [0.79–0.99]Multivariate adjusted RR [95% CI] 1 0.91 [0.77–1.07] 0.76 [0.62–0.92] 0.78 [0.63–0.98] 0.03 0.87 [0.77–0.97] LegumesMedian intake, g/d 0 5 17 32Cases/person-years 450/24,442 301/22,581 280/22,768 315/23,737Sex, age, energy-adjusted RR [95% CI] 1 0.91 [0.77–1.08] 0.84 [0.66–1.06] 0.73 [0.56–0.95] 0.02 0.92 [0.85–1.00]Multivariate adjusted RR [95% CI] 1 0.95 [0.80–1.14] 0.85 [0.66–1.08] 0.72 [0.55–0.95] 0.02 0.93 [0.86–1.01] FruitMedian intake, g/d 130 195 262 379Cases/person-years 477/23,360 351/22,098 291/22,860 227/25,210Sex, age, energy-adjusted RR [95% CI] 1 0.80 [0.68–0.94] 0.83 [0.70–0.99] 0.85 [0.68–1.05] 0.14 0.94 [0.89–0.99]Multivariate adjusted RR [95% CI] 1 0.83 [0.71–0.98] 0.88 [0.74–1.05] 0.91 [0.73–1.12] 0.42 0.95 [0.90–1.01]========================================================= 1 n=10,449; all models are stratified on age and study center, and adjusted for sex, smoking status (never, former<10 y, former =/>10 y, current <20 cigarettes, current >/= 20 cigarettes, unknown), self-reported heart attack at baseline (yes, no/unknown), self-reported hypertension at baseline (yes, no/unknown), self-reported cancer at baseline (yes, no/unknown), WHR (continuous), insulin treatment (yes, no/unknown), age at diabetes diagnosis (continuous), energy intake (continuous), alcohol intake (continuous). 2 Total of 80 g/d for vegetables, legumes, and fruit; vegetables; fruit; 20 g/d for legumes. For sensitivity analyses, we restricted our population to those individuals with a diabetes diagnosis at age 40 or older (n = 8408). By doing so, we aimed at excluding all type 1 diabetes patients. RR for intake of total vegetables, legumes, and fruit was 0.95 [95% CI 0.90–1.00] (P = 0.041) for an increment of 80 g/d, which was essentially the same as for the overall group. Associations for participants reporting insulin treatment (n = 2197) or no insulin treatment (n = 6555) were different (P < 0.0001), with a stronger effect in those not treated with insulin (RR 0.90 [95% CI 0.84–0.96]) compared with those treated with insulin (RR 0.96 [95% CI 0.87–1.06]). Analyses of vegetable subgroups showed intake of root vegetables was associated with a significantly decreased RR (0.91 [95% CI 0.84–0.99]) for an increase in intake of 20 g/d (Table 3). Associations for the remaining vegetable subgroups, except mushrooms, were inverse but not significant (Table 3). TABLE 3 RR [95% CI] for intake of subtypes of vegetables, and all-cause mortality in a diabetic population^1.==============================================Vegetable subtype RR [95% CI]==============================================Fruiting vegetables (40 g/d) 0.92 [0.84–1.02]Root vegetables (20 g/d) 0.91 [0.84–0.99]Leafy vegetables (8 g/d) 0.97 [0.92–1.03]Cabbages (8 g/d) 0.99 [0.94–1.04]Mushrooms (1 g/d) 1.01 [0.99–1.03]Garlic/onion vegetables (8 g/d) 0.97 [0.87–1.10]============================================== 1 n=10,449; all models are stratified on age and study center, and adjusted for sex, smoking status (never, former ,10 y, former =/>10 y, current <20 cigarettes/d, current =/>20 cigarettes/d, unknown), self-reported heart attack at baseline (yes, no/unknown), self-reported hypertension at baseline (yes, no/unknown), self-reported cancer at baseline (yes, no/unknown), waist-to-hip ratio (continuous), insulin treatment (yes, no/unknown), age at diabetes diagnosis (continuous), energy intake (continuous), alcohol intake (continuous). The results for cause of death-specific analyses (Table 4) suggested that intake of total vegetables, legumes, and fruit was inversely associated with CVD mortality and mortality due to other causes but not with cancer mortality. Differences in RR for CVD mortality or mortality due to non-CVD/non-cancer causes compared with the RR risk for cancer mortality were significant. Associations for CVD mortality were inverse for all dietary exposure variables. RR were essentially unchanged when participants with prevalent heart attack were excluded from the analysis (data not shown). No significant associations of any dietary exposure variable with cancer mortality was observed. All RR were close to unity. TABLE 4 RR [95% CI] for fruit, vegetables, and legumes intake, and cause-specific mortality in a diabetic population^1.===================================================================== ---Cause of death---P for differencesFood group---CVD^2 Cancer^3 Other causes^4---CVD vs. cancer CVD vs. other Cancer vs. other=====================================================================Vegetables, legumes, and fruit (80 g/d) 0.88 [0.81–0.95] 1.08 [0.99–1.17] 0.90 [0.82–0.99] ,0.01 0.67 ,0.01Vegetables (80 g/d) 0.85 [0.68–1.07] 1.09 [0.87–1.36] 0.72 [0.57–0.91] 0.13 0.32 0.01Legumes (20 g/d) 0.72 [0.60–0.88] 1.09 [0.96–1.24] 1.02 [0.93–1.12] ,0.01 ,0.01 0.41Fruit (80 g/d) 0.90 [0.81–0.99] 1.08 [0.98–1.19] 0.92 [0.82–1.03] ,0.01 0.72 0.04===================================================================== 1 n=10,262; all models are stratified on age and study center, and adjusted for sex, smoking status (never, former <10 y, former =/>10 y, current <20 cigarettes/d, current =/>20 cigarettes/d, unknown), self-reported heart attack at baseline (yes, no/unknown), self-reported cancer at baseline (yes, no/unknown), self-reported hypertension at baseline (yes, no/unknown), WHR (continuous), insulin treatment (yes, no/unknown), age at diabetes diagnosis (continuous), energy intake (continuous), alcohol intake (continuous). 2 ICD-10 codes I00–99; n=517. 3 ICD-10 codes C00–97; n=319. 4 n=323. Discussion In this study, a significant inverse association existed between intake of total vegetables, legumes, and fruit and all-cause mortality in a European diabetic population. Associations were inverse for deaths due to CVD and non-CVD/non-cancer causes, but not for deaths due to cancer. To our knowledge, only 3 prospective studies investigated food intakes as risk factors for disease incidence or mortality in diabetic populations to date (30–32) and only one of those evaluated vegetables, legumes, and fruit (32). In that study, overall mortality and cardiovascular mortality was investigated in the Greek arm of the EPIC study, analyzing data of a subgroup of 1013 participants who took drugs for diabetes mellitus and did not report any comorbidities at enrollment (32). Estimated RR for total mortality and vegetables (1.10 [95% CI 0.80–1.51]), legumes and potatoes (0.85 [95% CI 0.63–1.13]), and fruits and nuts (0.93 [95% CI 0.69–1.26]) were not significant. The associations between vegetables and fruit intake and CVD and cancer have been extensively investigated in the general population and have been found to be inverse for CVD (11,12,33–36). Findings for cancer were less consistent (10,33). Of note, inverse associations seemed to be stronger for vegetables and legumes than for fruit intake in the regression models for all-cause mortality and cause-specific mortality, suggesting that consumption of vegetables and legumes is more beneficial for diabetes patients than consumption of fruit. Our findings seem plausible, because type 2 diabetes often is associated with overweight or obesity. Modest weight loss has been shown to improve insulin sensitivity and reduces the CVD risk factors associated with type 2 diabetes (37–39). A diet high in vegetables, legumes, and fruit might help diabetic patients lose weight (39). However, numerous possible mechanisms explaining the effect of fruits and vegetables on health outcomes have been discussed, e.g. the antioxidative effects of vitamins, minerals, or polyphenols, a homocysteine-reducing effect of folate and B vitamins, or the enhancement of detoxification enzymes (12,40). The question arises whether amounts of vegetables and legumes consumed by diabetic men and women differ from those consumed by the nondiabetic population. A comparison of intakes assessed in the EPIC study showed that vegetable and legume intakes, as well as intake of fruit, were slightly higher in the diabetic than the nondiabetic population (data not shown). However, different studies have reported that the majority of adults with type 2 diabetes did not follow guidelines for fruit and vegetable consumption (41) and healthy foods might be overreported, because dietary energy intake has been shown to be underreported in obese diabetic patients (42). Although this type of misreporting would not be of concern for risk assessments within the diabetic population, absolute intakes as reported in our study should be interpreted cautiously. Our study has several limitations. First is its reliance on self-reports of diabetes. However, in a different setting, a comparison between self-reports and medical records in a group of elderly men in the US has shown that concordance was excellent for diabetes ( = 0.84) and substantial for hypertension ( = 0.70) (43). Other studies in the US and Canada reported similar (44,45). Second, we were unable to distinguish between type 1 and 2 diabetes diagnoses. However, our sensitivity analysis using data of those individuals diagnosed at or after 40 y old showed the same inverse association as in the overall cohort. Third, we were unable to adjust for use of oral hypoglycemic medication and information on use of insulin was not available for all participants. We also lacked information about if and what kind of advice about diet diabetes patients might have received. Those participants reporting higher vegetable and fruit intake might comply better with other dietary and lifestyle recommendations and treatment, which could have affected the course of their diabetes. Fourth, information on prevalent heart attacks, hypertension, and cancer, which we adjusted for in our analysis, were also self-reported at baseline and the percentage of unknown conditions was higher for participants with low intake of fruit, vegetables, and legumes. We therefore cannot rule out residual confounding. However, substitution of self-reported hypertension by blood pressure measurement, which was available for about two-thirds of the diabetic population, did not change our conclusions. Also, after excluding participants with unknown status of prevalent heart attacks, hypertension, or cancer, the association with total vegetables, legumes, and fruit intake was significantly inverse. Finally, as described previously (24), categorical variables based on the predicted dietary intake distribution in the calibration model have to be interpreted with caution, because the predicted variation does not reflect true variation. The most important strength of our study is its sample size. To our knowledge, this is the largest observational cohort of diabetic individuals that investigated associations between lifestyle factors and mortality. Furthermore, the multicentric design with centers spread across Europe enabled the coverage of a large variation in exposures. Also, we controlled for important risk factors in our analyses, including the age at diabetes diagnosis and treatment with insulin. Additionally, the application of a competing risk model allowed the evaluation of several outcomes at the same time without losing statistical power due to exclusion of subjects. In conclusion, our study showed that a diet high in vegetables, legumes, and fruit was associated with a reduced risk of all-cause mortality in a European diabetic population. Vegetables, legumes and fruit seemed to have impact especially on CVD mortality, and not on cancer mortality. Our study lends support to the current state of evidence from general population studies that the protective potential of high vegetable and fruit intake is larger for CVD than for cancer. Furthermore, our study underlines the recommendation for the diabetic population to eat large amounts of vegetables, legumes, and fruit. -- Al Pater, alpater@... Never miss a thing. Make your homepage. Quote Link to comment Share on other sites More sharing options...
Guest guest Posted March 22, 2008 Report Share Posted March 22, 2008 In light of Rodney's 2 most recent posts concerning the anti-cancer properties of compounds in citrus fruits, apples, pears, prunes, tomatoes, red grapes, pomegranates, cruciferous vegetables, carrots, some berries, etc., can anyone make sense of the finding in this study that fruit and vegetable intake had no bearning on cancer mortality?? " Vegetables, legumes and fruit seemed to have impact especially on CVD mortality, and not on cancer mortality. " Thanks, -Dave > > Hi All, > > > Love those legumes. Root vegetables in this study did not include potatoes. The pdf of the below paper is now available. > > > Ute Nöthlings, Matthias B. Schulze, Cornelia Weikert, Heiner Boeing, T. van der Schouw, Bamia, Vasiliki Benetou, Pagona Lagiou, Vittorio Krogh, Joline W. J. Beulens, Petra H. M. Peeters, Jytte Halkjær, Anne Tjønneland, rio Tumino, Salvatore Panico, Giovanna Masala, Francoise Clavel-Chapelon, Blandine de Lauzon, Marie- Boutron-Ruault, Marie-Noël Vercambre, Rudolf Kaaks, Jakob Linseisen, Kim Overvad, Larraitz Arriola, Eva Ardanaz, A. , Marie- Tormo, Sheila Bingham, Kay-Tee Khaw, Tim J. A. Key, Paolo Vineis, Elio Riboli, Pietro Ferrari, Paolo Boffetta, H. Bas Bueno-de-Mesquita, Daphne L. van der A, Göran Berglund, bet Wirfält, Göran Hallmans, Ingegerd Johansson, Eiliv Lund, and Antonia Trichopoulo > Intake of Vegetables, Legumes, and Fruit, and Risk for All-Cause, Cardiovascular, and Cancer Mortality in a European Diabetic Population > J. Nutr. 2008 138: 775-781. > > ABSTRACT > > We examined the associations of intake of vegetables, legumes and fruit with all-cause and cause-specific mortality in a population with prevalent diabetes in Europe. > > A cohort of 10,449 participants with self-reported diabetes within the European Prospective Investigation into Cancer and Nutrition study was followed for a mean of 9 y. Intakes of vegetables, legumes, and fruit were assessed at baseline between 1992 and 2000 using validated country-specific questionnaires. A total of 1346 deaths occurred. Multivariate relative risks (RR) for all-cause mortality were estimated in regression models and RR for cause-specific mortality were derived in a competing risk model. > > An increment in intake of total vegetables, legumes, and fruit of 80 g/d was associated with a RR of death from all causes of 0.94 [95% CI 0.90–0.98]. Analyzed separately, vegetables and legumes were associated with a significantly reduced risk, whereas nonsignificant inverse associations for fruit intake were observed. Cardiovascular disease (CVD) mortality and mortality due to non-CVD/non-cancer causes were significantly inversely associated with intake of total vegetables, legumes, and fruit (RR 0.88 [95% CI 0.81–0.95] and 0.90 [0.82–0.99], respectively) but not cancer mortality (1.08 [0.99–1.17]). Intake of vegetables, legumes, and fruit was associated with reduced risks of all-cause and CVD mortality in a diabetic population. > > The findings support the current state of evidence from general population studies that the protective potential of vegetable and fruit intake is larger for CVD than for cancer and suggest that diabetes patients may benefit from a diet high in vegetables and fruits. > > Abbreviations used: CVD, cardiovascular disease; EPIC, European Prospective Investigation into Cancer and Nutrition; ICD-10, 10th revision of the International Statistical Classification of Diseases, Injuries and Causes of Death; RR, relative risk. > > Introduction > > Dietary recommendations are an integral part in the treatment and management of diabetes mellitus (1), although data on their efficacy are limited (2). The objective of diabetes treatment is to reduce the risk of long-term vascular complications (3), predominantly cardiovascular diseases (CVD).29 Increased risks of CVD (4,5) and a higher mortality rate than in the general population have been consistently found in diabetic populations (5,6). Recently, evidence for a higher risk of cancer at various sites has been reported to occur in diabetic compared with nondiabetic individuals (7–9). > > Vegetable and fruit intakes have been extensively investigated as risk factors for cancer (10), CVD (11,12), or mortality (13–15) in the general population; however, data on associations between dietary behavior, including vegetables and fruit intake, and disease outcomes in diabetic populations are scarce (2). Therefore, we analyzed data from the European Prospective Investigation into Cancer and Nutrition (EPIC) study to examine the associations between intake of vegetables, legumes, and fruit, and risk for all-cause and cause-specific mortality in a population of diabetic individuals. > > Subjects, Materials, and Methods ... EPIC is an ongoing multicenter prospective cohort study designed to investigate associations between diet and other lifestyle behaviors with chronic diseases, especially cancer. A detailed description of the study design and methods used can be found elsewhere (16). In brief, between 1992 and 2000, >500,000 study participants were recruited in 23 study centers located in 10 European countries to be followed for cancer incidence and cause-specific mortality. Participants ranging from 35 to 70 y were recruited in study centers in France, Germany, Greece, Italy, The Netherlands, Spain, United Kingdom, Sweden, Denmark, and Norway. Study populations were population-based samples from designated geographic areas (e.g. town or province), with the exceptions of France (members of the health insurance for state school employees), subsamples of the Italian and Spanish cohort (blood donors), Utrecht (The Netherlands), Florence (Italy; women > invited for breast cancer screening), and a subsample in Oxford (health-conscious group, vegetarians). Individuals signed informed consent forms, after which self-administered diet and lifestyle questionnaires were mailed to the participants. In the Spanish centers, Ragusa, Naples, and Greece, interviewer-administered questionnaires were used. All participants were invited to a study center for an examination that included anthropometric and blood pressure measurements. Approval for this study was obtained from the ethical review boards of the International Agency for Research on Cancer and from all local institutions where subjects had been recruited for the EPIC study. > > Study population. The study population for this study was the group of participants in the EPIC cohort reporting a diagnosis of diabetes mellitus at baseline (n = 13,838). After exclusion of diabetic participants missing dietary questionnaire information (n = 144) or date of death (n = 3) or with implausible energy intake (top or bottom 1% of the ratio of energy intake:energy requirement; n = 437), a total of 13,254 participants with self-reported diabetes mellitus were identified. ... 10,449 participants with self-reported diabetes mellitus were identified for this analysis. > > ... > > Results > > Mean age at baseline ranged from 56 to 59 y across quartiles of vegetable, legume, and fruit intake, not showing a trend (Table 1). The percentage of diabetic participants treated with insulin was inversely associated with intake of vegetables, legumes, and fruit, ranging from 32% in the lowest quartile to 16% in the highest. The percentage of diabetics with a self-reported heart attack at baseline was also inversely associated with intake of total vegetables, legumes, and fruit. There was a positive association with self-reported absence of hypertension at baseline and hyperlipidemia with vegetable, legume, and fruit intake. A total of 1346 deaths occurred, 517 due to circulatory diseases, 319 due to cancer, and 323 due to other specified causes. > > TABLE 1 Baseline characteristics of participants with self-reported diabetes^1. > ======================================= > ---Quartile of fruit, vegetables, and legumes intake^2 > ======================================= > Characteristic---1 2 3 4 > ======================================= > n 2612 2612 2613 2612 > Men, % 59 36 36 53 > Age, y 57±8 59±9 59±9 56±8 > Age at diabetes diagnosis, y 47±13 49±13 50±12 48±11 > BMI, kg/m 2 28.3±4.9 28.8±5.3 29.0±5.1 29.1±4.7 > Waist-to-hip ratio 0.93±0.09 0.90±0.09 0.90±0.09 0.91±0.09 > Insulin treatment, % > Yes 32 20 17 16 > No 51 61 67 73 > Unknown 18 19 16 11 > Heart attack, % > Yes 9 6 5 4 > No 79 87 91 94 > Cancer, % > Yes 5 4 3 1 > No 78 86 90 95 > Hypertension, % > Yes 35 45 45 40 > No 33 41 49 57 > Hyperlipidemia, % > Yes 27 38 39 42 > No 41 50 54 54 > Smoking status, % > Never 34 52 55 51 > Former, .10 y 26 21 18 15 > Former, #10 y 13 10 10 13 > Current, ,20 cigarettes/d 15 10 9 13 > Current, $20 cigarettes/d 12 7 7 7 > Unknown 0 1 1 1 > Physical activity, % > Inactive 16 11 10 14 > Moderately inactive 36 31 29 28 > Moderately active 40 51 54 50 > Active 6 7 6 8 > Missing 2 1 0 0 > Energy intake, kJ/d 8816±1749 7987±1887 7945±1669 9096±1992 > ======================================= > 1 Values are percentages or mean±SD; n=10,449. > 2 Differences across quartiles were tested with the chi-square test for categorical variables and with the t test for slope in linear regression models of mean values on intake of fruit, vegetables, and legumes for continuous variables. P <0.01 for all variables. Due to the large sample size, all tests were significant, so all variables differed > > Intake of total vegetables, legumes, and fruit was inversely associated with risk for all-cause mortality (Table 2). An increment in intake by 80 g/d was associated with a significant risk reduction of 6% ([RR 0.94 [95% CI 0.90–0.98]). The respective RR were 0.95 [0.89–1.00] for men and 0.93 [0.85–1.03] for women. Analyzed separately, intakes of vegetables and legumes were also inversely associated with risk for all-cause mortality (Table 2). Associations for total fruit intake were inverse but did not reach significance in the multivariate model (P = 0.42). In a model including vegetables, legumes, and fruit as separate variables, RR were below 1, with a significant RR for vegetables (data not shown). Additional adjustment for physical activity (inactive, moderately inactive, moderately active, active, missing), self-reported hyperlipidemia, or educational attainment changed RR estimates only marginally (data not shown). Adjustment for actual measured blood pressure, > which was available for approximately two-thirds of the diabetic population, did not lead to different conclusions (data not shown). RR were essentially the same when we replaced waist-to-hip ratio with waist-to-height ratio, BMI, or waist circumference in multivariate models (data not shown). No evidence for nonlinear associations was found for any of the food group variables (data not shown). Sex, smoking status, age at diagnosis, and waist-to-hip ratio did not modify the associations between intake of fruits and vegetables and mortality risk (data not shown). Associations among participants 60 y of age (n = 4591) were similar to results for the overall group and reached significance for legumes only (data not shown). To confirm our findings, we calculated a multivariate-adjusted regression model with observed intake of total vegetables, legumes, and fruit, i.e. without regression calibration. For the fully adjusted model, RR [95% CI] were 0.83 [0.71–0.99] for the 2nd > quartile compared with the lowest, 0.75 [0.62–0.90] for the 3rd quartile, and 0.78 [0.62–0.97] for the highest quartile (P-trend = 0.046). > > TABLE 2 RR [95% CI] for intake of vegetables, legumes, and fruit, and all-cause mortality in a diabetic population^1. > ========================================================= > ---Quartiles of food intake--- > Food group---1 2 3 4---PTrend Continuous exposure^2 > ========================================================= > Vegetables, legumes, and fruit > Median intake, g/d 283 390 474 630 > Cases/person-years 478/ 3,890 315/22,065 326/22,399 227/25,174 > Sex, age, energy-adjusted RR [95% CI] 1 0.73 [0.62–0.87] 0.79 [0.66–0.95] 0.75 [0.59–0.94] 0.02 0.93 [0.89–0.97] > Multivariate adjusted RR [95% CI] 1 0.75 [0.63–0.89] 0.80 [0.66–0.97] 0.76 [0.60–0.96] 0.03 0.94 [0.90–0.98] > > Vegetables > Median intake, g/d 127 164 198 259 > Cases/person-years 432/24,028 333/22,839 271/22,933 310/23,730 > Sex, age, energy-adjusted RR [95% CI] 1 0.86 [0.73–1.01] 0.73 [0.60–0.88] 0.81 [0.65–1.01] 0.05 0.88 [0.79–0.99] > Multivariate adjusted RR [95% CI] 1 0.91 [0.77–1.07] 0.76 [0.62–0.92] 0.78 [0.63–0.98] 0.03 0.87 [0.77–0.97] > > Legumes > Median intake, g/d 0 5 17 32 > Cases/person-years 450/24,442 301/22,581 280/22,768 315/23,737 > Sex, age, energy-adjusted RR [95% CI] 1 0.91 [0.77–1.08] 0.84 [0.66–1.06] 0.73 [0.56–0.95] 0.02 0.92 [0.85–1.00] > Multivariate adjusted RR [95% CI] 1 0.95 [0.80–1.14] 0.85 [0.66–1.08] 0.72 [0.55–0.95] 0.02 0.93 [0.86–1.01] > > Fruit > Median intake, g/d 130 195 262 379 > Cases/person-years 477/23,360 351/22,098 291/22,860 227/25,210 > Sex, age, energy-adjusted RR [95% CI] 1 0.80 [0.68–0.94] 0.83 [0.70–0.99] 0.85 [0.68–1.05] 0.14 0.94 [0.89–0.99] > Multivariate adjusted RR [95% CI] 1 0.83 [0.71–0.98] 0.88 [0.74–1.05] 0.91 [0.73–1.12] 0.42 0.95 [0.90–1.01] > ========================================================= > 1 n=10,449; all models are stratified on age and study center, and adjusted for sex, smoking status (never, former<10 y, former =/>10 y, current <20 cigarettes, current >/= 20 cigarettes, unknown), self-reported heart attack at baseline (yes, no/unknown), self-reported hypertension at baseline (yes, no/unknown), self-reported cancer at baseline (yes, no/unknown), WHR (continuous), insulin treatment (yes, no/unknown), age at diabetes diagnosis (continuous), energy intake (continuous), alcohol intake (continuous). > 2 Total of 80 g/d for vegetables, legumes, and fruit; vegetables; fruit; 20 g/d for legumes. > > For sensitivity analyses, we restricted our population to those individuals with a diabetes diagnosis at age 40 or older (n = 8408). By doing so, we aimed at excluding all type 1 diabetes patients. RR for intake of total vegetables, legumes, and fruit was 0.95 [95% CI 0.90–1.00] (P = 0.041) for an increment of 80 g/d, which was essentially the same as for the overall group. Associations for participants reporting insulin treatment (n = 2197) or no insulin treatment (n = 6555) were different (P < 0.0001), with a stronger effect in those not treated with insulin (RR 0.90 [95% CI 0.84–0.96]) compared with those treated with insulin (RR 0.96 [95% CI 0.87–1.06]). > > Analyses of vegetable subgroups showed intake of root vegetables was associated with a significantly decreased RR (0.91 [95% CI 0.84–0.99]) for an increase in intake of 20 g/d (Table 3). Associations for the remaining vegetable subgroups, except mushrooms, were inverse but not significant (Table 3). > > TABLE 3 RR [95% CI] for intake of subtypes of vegetables, and all-cause mortality in a diabetic population^1. > ============================================== > Vegetable subtype RR [95% CI] > ============================================== > Fruiting vegetables (40 g/d) 0.92 [0.84–1.02] > Root vegetables (20 g/d) 0.91 [0.84–0.99] > Leafy vegetables (8 g/d) 0.97 [0.92–1.03] > Cabbages (8 g/d) 0.99 [0.94–1.04] > Mushrooms (1 g/d) 1.01 [0.99–1.03] > Garlic/onion vegetables (8 g/d) 0.97 [0.87–1.10] > ============================================== > 1 n=10,449; all models are stratified on age and study center, and adjusted for sex, smoking status (never, former ,10 y, former =/>10 y, current <20 cigarettes/d, current =/>20 cigarettes/d, unknown), self-reported heart attack at baseline (yes, no/unknown), self-reported hypertension at baseline (yes, no/unknown), self-reported cancer at baseline (yes, no/unknown), waist-to-hip ratio (continuous), insulin treatment (yes, no/unknown), age at diabetes diagnosis (continuous), energy intake (continuous), alcohol intake (continuous). > > The results for cause of death-specific analyses (Table 4) suggested that intake of total vegetables, legumes, and fruit was inversely associated with CVD mortality and mortality due to other causes but not with cancer mortality. Differences in RR for CVD mortality or mortality due to non-CVD/non-cancer causes compared with the RR risk for cancer mortality were significant. Associations for CVD mortality were inverse for all dietary exposure variables. RR were essentially unchanged when participants with prevalent heart attack were excluded from the analysis (data not shown). No significant associations of any dietary exposure variable with cancer mortality was observed. All RR were close to unity. > > TABLE 4 RR [95% CI] for fruit, vegetables, and legumes intake, and cause-specific mortality in a diabetic population^1. > ===================================================================== > ---Cause of death---P for differences > Food group---CVD^2 Cancer^3 Other causes^4---CVD vs. cancer CVD vs. other Cancer vs. other > ===================================================================== > Vegetables, legumes, and fruit (80 g/d) 0.88 [0.81–0.95] 1.08 [0.99–1.17] 0.90 [0.82–0.99] ,0.01 0.67 ,0.01 > Vegetables (80 g/d) 0.85 [0.68–1.07] 1.09 [0.87–1.36] 0.72 [0.57–0.91] 0.13 0.32 0.01 > Legumes (20 g/d) 0.72 [0.60–0.88] 1.09 [0.96–1.24] 1.02 [0.93–1.12] ,0.01 ,0.01 0.41 > Fruit (80 g/d) 0.90 [0.81–0.99] 1.08 [0.98–1.19] 0.92 [0.82–1.03] ,0.01 0.72 0.04 > ===================================================================== > 1 n=10,262; all models are stratified on age and study center, and adjusted for sex, smoking status (never, former <10 y, former =/>10 y, current <20 cigarettes/d, current =/>20 cigarettes/d, unknown), self-reported heart attack at baseline (yes, no/unknown), self-reported cancer at baseline (yes, no/unknown), self-reported hypertension at baseline (yes, no/unknown), WHR (continuous), insulin treatment (yes, no/unknown), age at diabetes diagnosis (continuous), energy intake (continuous), alcohol intake (continuous). > 2 ICD-10 codes I00–99; n=517. > 3 ICD-10 codes C00–97; n=319. > 4 n=323. > > Discussion > > In this study, a significant inverse association existed between intake of total vegetables, legumes, and fruit and all-cause mortality in a European diabetic population. Associations were inverse for deaths due to CVD and non-CVD/non-cancer causes, but not for deaths due to cancer. > > To our knowledge, only 3 prospective studies investigated food intakes as risk factors for disease incidence or mortality in diabetic populations to date (30–32) and only one of those evaluated vegetables, legumes, and fruit (32). In that study, overall mortality and cardiovascular mortality was investigated in the Greek arm of the EPIC study, analyzing data of a subgroup of 1013 participants who took drugs for diabetes mellitus and did not report any comorbidities at enrollment (32). Estimated RR for total mortality and vegetables (1.10 [95% CI 0.80–1.51]), legumes and potatoes (0.85 [95% CI 0.63–1.13]), and fruits and nuts (0.93 [95% CI 0.69–1.26]) were not significant. The associations between vegetables and fruit intake and CVD and cancer have been extensively investigated in the general population and have been found to be inverse for CVD (11,12,33–36). Findings for cancer were less consistent (10,33). > > Of note, inverse associations seemed to be stronger for vegetables and legumes than for fruit intake in the regression models for all-cause mortality and cause-specific mortality, suggesting that consumption of vegetables and legumes is more beneficial for diabetes patients than consumption of fruit. > > Our findings seem plausible, because type 2 diabetes often is associated with overweight or obesity. Modest weight loss has been shown to improve insulin sensitivity and reduces the CVD risk factors associated with type 2 diabetes (37–39). A diet high in vegetables, legumes, and fruit might help diabetic patients lose weight (39). However, numerous possible mechanisms explaining the effect of fruits and vegetables on health outcomes have been discussed, e.g. the antioxidative effects of vitamins, minerals, or polyphenols, a homocysteine-reducing effect of folate and B vitamins, or the enhancement of detoxification enzymes (12,40). > > The question arises whether amounts of vegetables and legumes consumed by diabetic men and women differ from those consumed by the nondiabetic population. A comparison of intakes assessed in the EPIC study showed that vegetable and legume intakes, as well as intake of fruit, were slightly higher in the diabetic than the nondiabetic population (data not shown). However, different studies have reported that the majority of adults with type 2 diabetes did not follow guidelines for fruit and vegetable consumption (41) and healthy foods might be overreported, because dietary energy intake has been shown to be underreported in obese diabetic patients (42). Although this type of misreporting would not be of concern for risk assessments within the diabetic population, absolute intakes as reported in our study should be interpreted cautiously. > > Our study has several limitations. First is its reliance on self-reports of diabetes. However, in a different setting, a comparison between self-reports and medical records in a group of elderly men in the US has shown that concordance was excellent for diabetes ( = 0.84) and substantial for hypertension ( = 0.70) (43). Other studies in the US and Canada reported similar (44,45). Second, we were unable to distinguish between type 1 and 2 diabetes diagnoses. However, our sensitivity analysis using data of those individuals diagnosed at or after 40 y old showed the same inverse association as in the overall cohort. Third, we were unable to adjust for use of oral hypoglycemic medication and information on use of insulin was not available for all participants. We also lacked information about if and what kind of advice about diet diabetes patients might have received. Those participants reporting higher vegetable and fruit intake might comply better with other dietary and > lifestyle recommendations and treatment, which could have affected the course of their diabetes. Fourth, information on prevalent heart attacks, hypertension, and cancer, which we adjusted for in our analysis, were also self-reported at baseline and the percentage of unknown conditions was higher for participants with low intake of fruit, vegetables, and legumes. We therefore cannot rule out residual confounding. However, substitution of self-reported hypertension by blood pressure measurement, which was available for about two-thirds of the diabetic population, did not change our conclusions. Also, after excluding participants with unknown status of prevalent heart attacks, hypertension, or cancer, the association with total vegetables, legumes, and fruit intake was significantly inverse. Finally, as described previously (24), categorical variables based on the predicted dietary intake distribution in the calibration model have to be interpreted with caution, because the > predicted variation does not reflect true variation. > > The most important strength of our study is its sample size. To our knowledge, this is the largest observational cohort of diabetic individuals that investigated associations between lifestyle factors and mortality. Furthermore, the multicentric design with centers spread across Europe enabled the coverage of a large variation in exposures. Also, we controlled for important risk factors in our analyses, including the age at diabetes diagnosis and treatment with insulin. Additionally, the application of a competing risk model allowed the evaluation of several outcomes at the same time without losing statistical power due to exclusion of subjects. > > In conclusion, our study showed that a diet high in vegetables, legumes, and fruit was associated with a reduced risk of all-cause mortality in a European diabetic population. Vegetables, legumes and fruit seemed to have impact especially on CVD mortality, and not on cancer mortality. Our study lends support to the current state of evidence from general population studies that the protective potential of high vegetable and fruit intake is larger for CVD than for cancer. Furthermore, our study underlines the recommendation for the diabetic population to eat large amounts of vegetables, legumes, and fruit. > > -- Al Pater, alpater@... > > > --------------------------------- > Never miss a thing. Make your homepage. > Quote Link to comment Share on other sites More sharing options...
Recommended Posts
Join the conversation
You are posting as a guest. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.