Fruit, vegetables and vascular risk: reading the 200-gram association
Prospective-study data linked greater produce intake with lower coronary and stroke risk, without proving causation.

Illustrative food image, not a photograph of the study intervention.
What dose-response means here
A 2017 systematic review examined how fruit and vegetable intake related to cardiovascular disease, cancer and mortality in prospective studies. The review included 95 studies reported in 142 publications, with searches extending through September 2016. Its dose-response analysis expressed several results per additional 200 grams a day of fruits and vegetables combined. That is an analytical increment, not a randomized daily prescription. People were not assigned by this review to eat a specified extra amount. Consequently, its estimates describe associations across prospective observations and should not be presented as a guaranteed effect of changing an individual's intake by exactly 200 grams.
Aune D, Giovannucci E, Boffetta P, Fadnes LT et al.. International journal of epidemiology, 2017. PMID 28338764 · DOI
Coronary heart disease and stroke
For coronary heart disease, the summary relative risk per additional 200 grams a day of combined fruit and vegetable intake was 0.92, with a 95% confidence interval of 0.90 to 0.94, based on 15 studies. For stroke, the corresponding relative risk was 0.84, with a confidence interval of 0.76 to 0.92, based on 10 studies. These figures correspond to relative associations of 8% and 16% lower risk, respectively. They are not absolute percentage-point reductions in a person's chance of an event. The baseline risk needed to translate them into an absolute difference is not supplied by these pooled estimates.
Aune D, Giovannucci E, Boffetta P, Fadnes LT et al.. International journal of epidemiology, 2017. PMID 28338764 · DOI
Why the two outcomes differ
The coronary heart disease analysis reported an I-squared statistic of 0%, whereas the stroke analysis reported 73%. Those values describe different degrees of between-study heterogeneity in these analyses. The lower stroke relative risk therefore came with substantially more variation across the underlying studies, as well as a wider confidence interval. For overall cardiovascular disease, the pooled relative risk per 200 grams a day was 0.92, with a confidence interval of 0.90 to 0.95 and an I-squared value of 31%, based on 13 studies. Reading these details together is more informative than selecting only the largest-looking relative association from the abstract.
Aune D, Giovannucci E, Boffetta P, Fadnes LT et al.. International journal of epidemiology, 2017. PMID 28338764 · DOI
The upper range is not a promise
The authors observed reductions in risk up to 800 grams a day for the cardiovascular and mortality outcomes; cancer differed, with reductions observed up to 600 grams a day. Similar associations were reported for fruit and vegetables considered separately. These findings describe the shape of the pooled observational data, not a requirement that everyone reach an identical intake or a demonstration that each additional portion has the same effect. The abstract explicitly made its global attributable-death estimates conditional on the associations being causal. That qualification is important: the review supports an association with higher produce intake, but its design does not itself establish causation.
Aune D, Giovannucci E, Boffetta P, Fadnes LT et al.. International journal of epidemiology, 2017. PMID 28338764 · DOI