Several nutrition-specific programmes promoting the daily intake of fruits and vegetables (FV) have been developed around the world. Catchy slogans such as the NHS' “5-a-day” or the CDC's “Fruits & Veggies-More Matters” invade the supermarket shelves. Used as a marketing strategy by the food industry or as a health awareness campaign, these slogans aim to tackle inadequate nutrition, a risk factor responsible for 16.0 million (1.0%) disability adjusted life years and 1.7 million (2.8%) deaths worldwide. (FAO/WHO, 2004) However, despite the popularity recently gained by these programmes, there are still some controversies around their scientific support and their effectiveness for improving population's health. Aiming to solve these controversies, the present essay focuses on 1) evaluating the robustness of the current recommendations of fruits and vegetables daily intake (FVDI) by critically appraising the existing evidence; and 2) analysing the impact and limitations of on-going programmes to promote the consumption of FV in the European and American population.
“5-a-day”, “≥400 grams/day”, “10 portions/day” … what does the evidence say?
The WHO recommends eating ≥400 g per day of FV, not counting potatoes and other starchy tubers such as cassava. (FAO/WHO, 2004) This recommendation is shared by Europe and the USA, who adopted it for their campaign of “5-a-day” and “Fruits & Veggies-More Matters”, respectively. (CDC, 2016, NHS, 2016) However, a recently published systematic review reported that the mere consumption of 200g of FV per day might reduce by 16% the risk of heart disease, 18% the risk of strokes and 13% the risk of cardiovascular diseases; while the major risk reduction was observed with intakes up to 800g/day, which implies doubling the current recommendation. (Aune et al., 2017) Therefore… should the WHO re-evaluate the FVDI recommendation based on these new findings? To be able to answer this question it is necessary to analyse the evidence that initially led the WHO to adopt the cut-off point of 400g/day.
To start with, it is known that carrying out a research study of great scientific quality is complex; and it is even more complex in a field like Nutrition, in which mandatory changes on dietary habits reduce participant's compliance -increasing the proportion of loss to follow-up and therefore, compromising the validity of the results.(Burns et al., 2011) For this reason, there is a limited number of good quality randomised control trials (RCTs) that study the association between FVDI and health outcomes. Only two systematic reviews were published in the Cochrane Library: one evaluates the effect on FVDI in children and the other one analyses the association between FVDI and cardiovascular diseases (CVDs) in adults. (Wolfenden et al., 2012, Hartley et al., 2013) However, both reviews conclude that the limited number of available RCTs, the heterogeneity of the published studies and the short period of intervention hampers the achievement of conclusive results. Notwithstanding, both studies show a weak tendency toward the improvement of some health outcomes with a FVDI. Mytton et al (2014) published another systematic review containing 8 RCTs. (Mytton et al., 2014) They evaluated the association between FVDI and changes on body weight, finding that a mean difference of 133g of FVDI between both study arms was linked to a significant difference in the reduction of body weight (Mean reduction: 0.68kg ;[95% CI: 0.15;1.20]). Yet, when evaluating the heterogeneity, the I2 value equalled 83% compromising the validity of this meta-analysis. One last systematic review of RCTs published by Ledoux et at supports the conclusion of Mytton, endorsing the limitations caused by the heterogeneity of the studies to be able to perform a thorough examination of the association between FVDI and body weight. (Ledoux et al., 2011) Hence, the results coming from these systematic reviews are poor and deficient to make a recommendation of adequate quality and strength.
Because of the previously stated, the WHO (and other national agencies) resorted to epidemiological studies to sustain their recommendations. It is known that the susceptibility to residual or unmeasured confounding variables is an intrinsic limitation of observational studies. (Brozek et al., 2011) Therefore, to be able to make a strong recommendation based on them there must be a great homogeneity on the trend and directionality of the results as well as a large effect size, even after adjusting by known covariates. With this in mind, four published systematic reviews of prospective cohort studies evaluated the association between FVDI and non-communicable diseases (NCDs), such as CVDs, cancer, overweight/obesity and type 2 diabetes. A review by Wang et al (2014) meta-analysed 16 prospective cohort and found that a) an increase of one serving of FV a day reduces the risk of “all-cause mortality” (HR:0.95; [95%CI 0.92-098]) up to a threshold of around five servings a day; b) mortality due to cardiovascular causes has an inverse relationship with the number of FVDI (HR:0.96; [95% CI 0.92-0.99]) and c) there is no significant association between FVDI and cancer (HR:0.97; [95% CI 0.90-1.03]). (Wang et al., 2014) Another review evaluated the relationship between FVDI and changes in body weight, concluding that, based on available evidence and without another co-intervention to reduce dietary energy intake, FVDI alone has no effect as treatment or prevention of obesity (Mean difference on body weight: 0.04 [95% CI: −0.10, 0.17]).(Kaiser et al., 2014) Finally, Li et al found that an increase in fruit intake alone reduces the risk of type 2 diabetes (RR: 0.93; [95% CI 0.88 to 0.99]) but found no effect when combining fruit and vegetables together (RR: 0.96, [95% CI 0.86 to 1.07]).(Li et al., 2014)
Moreover, earlier this year, Aune et al published a systematic review that included 95 prospective cohort studies. (Aune et al., 2017) This review aimed to evaluate the effectiveness of FVDI and the appropriate dose to reduce the risk of cancer, stroke, coronary heart disease and “all-cause mortality”. These meta-analyses show that daily consumption of 200g of FV reduces the risk of coronary heart disease, stroke and CVDs by 16%, 18% and 13% respectively, with a stronger reduction occurring with intakes up to 800g/day (24%, 33%, 28%). The results reached by Aune can be a call-to-action for the WHO, NHS, CDC among others, when fighting to tackle NCDs. However, it is important to emphasise that the association between FVDI and the reduction of NCDs were made depending on the validity of three assumptions: 1) a causal relationship between FVDI and the study outcomes; 2) lack of confounding and; 3) that the results can be generalised across population. (Aune et al., 2017)
Then again, the review of the existing literature leaves us with nothing but doubts: can we determine causation between FVDI and a 1 to 5% reduction in the risk of NCDs based on observational studies? Are meta-analyses with up to an 83% of heterogeneity between studies conclusive to support this recommendation? Or can we just attribute this results to the fact that those with an adequate FVDI are also healthier, non-smokers, more physically active, drink less alcohol and have a more balanced nutrition? Or maybe those that have better access to their “5-a-day” are wealthier, more educated and live in “health-friendly” environments? And assuming these covariates were already taken into account in the statistical analyses, aren't there other unmeasured or even unknown covariates that might impact the direction or magnitude of the associations found in these studies?
Based on all the previously stated and with the help of the GRADE approach to judge the quality of the evidence and the strength of the recommendation, (Brozek et al., 2011) we conclude that not only the available evidence has proven to be of low quality and highly heterogeneous; but also, the directness and the effect size found in a diverse range of systematic reviews make the “5-a-day” recommendation weak and lacking clear scientific support.
Tackling the problem: fruit and vegetable for….all?
However, despite the limited evidence presented before, there are still a great number of health programmes promoting the daily intake of FV in Europe and the US: SchoolGruiten (Netherlands), Frugtkvaster (Denmark), Fruitness (Italy), 5 a day (UK), Un fruit pour la récré (France), are just some of them. One can only assume that the argument to keeping those campaigns running might be based on the following hypotheses: 1) the possible biases on the measurement or collection of the data in nutritional epidemiological studies dilutes the magnitude of the effect of FVDI on health, and there is actually a higher impact that the one stated by the evidence; 2) the potential benefit of the recommendation outweighs the risk of not having a recommendation at all, and might as well help to prevent other health-related issues such as micronutrient deficiency (even though no study has been found with micronutrient deficiency as a primary outcome) ; 3) the theoretical support from a biological and physiological perspective is sufficient, despite not being enough efficacy and epidemiological evidence to back it up.
So, if we are considering these hypotheses to be true… What has been the impact of these health programmes on the FVDI in the European and American population? According to the data provided by Eurostat, the statistical office of the European Union, only 1 in 7 people older than 15 years (14.2%) eats at least 5 portions of FV daily, while 1 in 3 (33.3%) does not consume any FV every day. (Eurostat, 2016) And this tendency is increasing: based on a theoretical calculation of the market balance, there has been a drop in the consumption of FV between 2006 and 2010. (De Cicco, 2016) This same trend has been detected in the USA, with a 7% decline in the intake of vegetables and a 6% in the intake of fruit between 2010 and 2015.(PBHF, 2015)
Hence, despite the number of assumptions that had to be made to justify the ongoing campaigns on the promotion of FVDI, the overall consumption in the population continues to decline. Why might this be? When evaluating the current programmes, most of them are focused on improving the availability and accessibility of FV through agreements with retail stores, farm-to-institution type of strategies, agricultural expansion programmes or by ensuring healthy options in workplace or school cafeterias.(CDC, 2011) Therefore, trying to find the logic to these unidimensional interventions we can ironically reduce its explanation to basic economy: increase the offer and reduce the prices, then demand goes up and people will easily reach to their 5-a-day goal. However, the theory of demand and supply does not apply to such a complex field as it is nutrition. This simplistic approach looks at food as a mere mean to nourishment, and alienates it from its cultural and social meaning. (Verstraeten et al., 2014) Thus, intervening only in the availability and accessibility of FV will not guarantee its utilisation but it will lead to a waste of resources to sustain awareness and promotion campaigns without any true impact in the FVDI of the population.
Limited scientific evidence, lack of a holistic approach in ongoing programmes and a continued decline in the amount of FV consumption are three barriers towards improving population's nutrition. But still, we are surrounded by publicity recommending the best products to reach to our desired “5-a-day” goal. So, going back to the initial question: should the WHO re-evaluate the FVDI recommendation?
Yes. We need more studies of high methodological quality to evaluate the true impact of FVDI on health if we want to achieve strong and conclusive recommendations. We need an integral view of the feeding process to design campaigns that encourage the consumption of FV but in an effective, efficient and realistic way.
This essay is not refuting the importance of the daily intake of FV, but it emphasises the normalisation of non-evidence based practices as well as the unjustified investment on public health programmes of limited impact. If our aim is to reduce the disease burden in these countries by improving their nutrition, the next step should be to study cost-effective practices that could lead to an efficient and evidence-based use of the available resources.
...(download the rest of the essay above)