Whole fruit, fruit juice and sugary drinks: not the same for diabetes

Whole fruit is associated with a lower risk of type 2 diabetes, sugary drinks with a clearly higher risk, and fruit juice — despite coming from fruit — behaves more like the drinks than like the fruit [s1][s2]. The physical form of the sugar, not just the amount, appears to matter, which is why "it's natural" is a poor guide to how a fruit product affects metabolic risk.

Whole fruit, and where juice diverges

A 2013 BMJ analysis followed 66,105 women in the Nurses' Health Study, 85,104 in its second cohort and 36,173 men in the Health Professionals Follow-up Study; over 3,464,641 person-years, 12,198 developed type 2 diabetes [s1]. Every three weekly servings of whole fruit carried a pooled hazard ratio of 0.98 — a small reduction — but the effect varied sharply by fruit [s1]. Blueberries came in at 0.74, grapes and raisins at 0.88, and apples and pears at 0.93, while cantaloupe ran the other way at 1.10 [s1].

The heterogeneity was itself statistically significant (P<0.001) and did not line up neatly with a fruit's sugar content [s1]. Bananas and grapefruit both landed at 0.95, while oranges (0.99) and strawberries (1.03) showed essentially no association [s1]. That the individual fruits differ so much is a caution against treating "fruit" as a single food, and against reading too much into the small pooled figure for total fruit.

The contrast that gives the study its point is juice. The same three-servings-a-week increment in fruit juice was associated with a hazard ratio of 1.08 — higher risk, in the opposite direction to the whole fruit it is pressed from [s1]. Juicing strips most of the fibre and delivers sugar in a fast-absorbed liquid, and the authors treated that as the likely explanation for why a "fruit" product tracks with worse outcomes [s1].

Sugary drinks, and the limits on juice

A 2015 BMJ meta-analysis of 17 cohorts, covering 38,253 diabetes cases over more than 10 million person-years, quantified the drinks side [s2]. Each daily serving of sugar-sweetened beverages was associated with an 18 percent higher incidence of type 2 diabetes, falling to 13 percent after adjusting for body weight — meaning part, but not all, of the risk runs through weight gain [s2]. Artificially sweetened beverages showed 8 percent and fruit juice 7 percent after the same adjustment [s2].

Here the authors were unusually candid about their own weak spots, and it is worth repeating rather than smoothing over. For artificially sweetened beverages, they flagged likely publication bias and residual confounding — people already at risk often switch to diet drinks — as covered in what the WHO concluded on artificial sweeteners [s2]. For fruit juice, the association went away in the studies that confirmed diabetes objectively rather than by self-report, so its 7 percent should be read as suggestive, not settled [s2]. Only for sugar-sweetened beverages did they find no evident bias, and they estimated that of 20.9 million projected diabetes cases over a decade in the United States, 1.8 million would be attributable to these drinks — a population attributable fraction of 8.7 percent [s2]. In the United Kingdom, where consumption is lower, the same modelling attributed 79,000 of a projected 2.6 million cases to these drinks, a fraction of 3.6 percent [s2]. Those population figures depend on assuming the association is causal, which the authors flag as an assumption rather than a demonstrated fact [s2].

What to make of it

The reliable signal is the ranking. Whole fruit is favourable or neutral, sugar-sweetened beverages are unfavourable with the firmest evidence of the three, and fruit juice sits closer to the drinks — a useful correction to the marketing instinct that juice is a healthy way to "get your fruit" [s1][s2]. The claims weaken as you move down that list: the whole-fruit and sugary-drink findings are robust, the artificial-sweetener and juice findings are hedged by their own authors, and all of it is observational. The added-sugar thread runs through our coverage of how much added sugar is too much and the debate over sugary-drink taxes. This is informational, not dietary advice.

Sources

• Fruit consumption and risk of type 2 diabetes: three prospective cohort studies — BMJ, 2013
• Consumption of sugar sweetened beverages, artificially sweetened beverages, and fruit juice and incidence of type 2 diabetes — BMJ, 2015