(View plain language abstract)
Many studies have demonstrated a link between body weight and morbidity and mortality risk in women. While it is well recognised that weight gain tends to occur throughout adulthood, the aetiology of this weight gain has been relatively under explored. The 5 fruit and vegetables-a-day message is a central tenet of existing health advice and there is some degree of assumption that adherence to this guideline will prevent the development of many chronic diseases, including obesity. However, relatively few analyses of prospective studies have been conducted to confirm this.
In this project a prospective investigation of the effects of eating patterns relating to fruit and vegetable consumption on body weight change in the UK Women's Cohort Study (UKWCS) was undertaken. The UKWCS is a 10 year prospective investigation of the relationship between diet and health in middle-aged women across the UK. Information on fruit and vegetable consumption was captured using questions at baseline and follow-up that asked the participants to separately report how many servings of fruit and vegetables they consumed in an average week, and also by analysis of a 217-item food frequency questionnaire completed at study baseline. Dietary patterns associated with fruit and vegetable consumption were also derived from the FFQ data. These included common patterns (vegetarianism, meat-eating, fish-eating), the Mediterranean dietary pattern, the WHO Healthy Diet Indicator, and patterns derived using principal components and cluster analysis techniques. Body weight measurements were self-reported at study baseline in 1995-8 and again after follow up in 1999-2004 (average follow-up 4.3 years) in 14,172 women within the cohort. After excluding women with missing data, incident cancer, self-reported diabetes mellitus or an eating disorder, 12,946 women were available for inclusion in linear regression analyses that assessed the association between change in body weight and consumption of fruit and vegetables and related patterns.
The women experienced a small increase in body weight (+1.5 kg SD 4.7) during the follow-up period. Weight was gained at a rate of 357g per year of follow-up which is similar to that observed in other cohorts of health-conscious individuals. During the period of follow-up, the women reported an increase in the number of servings of fruit and vegetables consumed (by 0.71 servings per day SD 1.7), from a baseline intake of 3.1 servings per day. In a model that adjusted for a range of confounders, effects of baseline fruit and vegetables on weight change did not show clear trends. Effect sizes were small and crossed zero indicating non-significance. However, small but clear trends appeared when the change in fruit and vegetable consumption between baseline and follow-up was considered. When analysed as a continuous variable, each portion increase in fruit was associated with 104g per year less weight gain (p<0.001) when compared against women who maintained their fruit intake. When change in vegetable consumption was used as the exposure, similar but slightly reduced effects were seen (reduction of 52g per year per portion increase, ns). Change in fruit and vegetable consumption combined also displayed similar significant protective effects against weight gain.
Total dietary fibre (non-starch polysaccharide) intake assessed at baseline tended to be negatively associated with weight change, although the effect sizes were small and not statistically significant (regression slope per gram fibre per day -0.17, 95% CI -2.4 to 2.0, ns). Fibre was also inversely associated with waist change. The effect size for the highest quintile was -0.05 (95% CI -0.16 to 0.06, ns) compared to the lowest quintile. Overall, the strongest associations concerned dietary fibre derived from cereal foods, both for weight change and waist circumference change. However, we found that fruit and vegetable fibre was also associated with waist change, regression slope=-0.006 (95% CI -0.01 to -0.007) for the continuous analysis (change in waist per gram increment of fruit and vegetable fibre per day).
In the common diet analysis, being a vegetarian or a fish-eater was associated with significantly reduced weight and waist change when compared to meat-eaters. There was an inverse relationship between the Mediterranean diet score and weight gain although the difference between the groups was statistically non-significant and overall adherence to the Mediterranean diet was not strongly related to weight and waist changes. Similarly, the WHO healthy diet indicator, a general marker of healthy eating was not strongly related to changes in weight or waist circumference in this cohort.
Principal components analysis revealed some interesting relationships concerning groupings of foods selected by participants and the impact on weight and waist change. Using this approach to describe dietary patterns provided some evidence to support the hypothesis that high vegetable consumption might help prevent weight gain or waist increase, since women who scored highly on component 1, which was characterised by vegetarian diets and those high in complex carbohydrates tended to gain less weight and women scoring highly on component 4 (fruit) tended to have lower waist circumference changes. Similarly, the cluster analysis found that the vegetarian groups experienced lower weight and waist changes compared to the group described as the higher diversity traditional omnivores who would typically consume lesser amounts of fruit and vegetables.
While causality may not be inferred from these observational data, the results of these analyses indicate that increasing consumption of fruit and vegetables may minimise some of the weight gain generally experienced by middle-aged women.
Women with a high intake of fruit and/or vegetables experience less weight gain in late middle-age than women with a low intake. A high intake of nutrients associated with fruit and vegetables (particularly dietary fibre) protects women from weight gain in middle age. Women who consume a dietary pattern characterised by high fruit and/or vegetables experience less weight gain than women with a low fruit and vegetable pattern.
The past few decades have been characterised by major changes in lifestyle, leading to steady increases in average body weight and measures of obesity within many countries. Within Europe, recent estimates suggest that about half of all men and a third of all women may be classified as either overweight or obese and similar data for the US indicate an even higher prevalence. Being overweight or obese is a well-known risk factor for cardiovascular disease and diabetes, but population studies are also providing growing evidence for a link between body weight and cancer risk. In particular, previous studies have suggested that in women, weight gain during and after the menopause is most strongly linked with the risk of developing cancer. An understanding of the factors that are associated with weight gain at this time will guide the development of dietary and lifestyle advice aimed at prevention of weight gain. By using data from more 12,000 women who have been followed up over 5 years, this study aimed to determine whether certain dietary patterns influence weight change in middle-aged UK women.
The UKWCS is a 10 year follow-up study of the relationship between diet and health in middle-aged women across the UK. Information on fruit and vegetable consumption was captured using questions at study start (baseline) and follow-up that asked the participants to separately report how many servings of fruit and vegetables they consumed in an average week, and also by analysis of a 217-item food frequency questionnaire. Dietary patterns associated with fruit and vegetable consumption were also derived from the FFQ data. These included commonly recognised patterns (vegetarianism, meat-eating, fish-eating), the Mediterranean dietary pattern, the World Health Organisation Healthy Diet Indicator, and patterns derived using advanced statistical techniques (principal components and cluster analysis) which explore how foods are typically consumed together to define a pattern. Body weight measurements were self-reported at study baseline in 1995-8 and again after follow up in 1999-2004 (average follow-up 4.3 years) in 14172 women within the cohort. After excluding women who developed cancer, self-reported diabetes mellitus or an eating disorder, 12946 women were available for inclusion in linear regression analyses that assessed the association between change in body weight and waist circumference and consumption of fruit and vegetables and related patterns.