Postprandial Glycemic Response to Whole Fruit versus Blended Fruit

April 24, 2023 6:47 AM

TLDR: smoothies ok We compared the postprandial glycemic response in 20 young, healthy college students (12 female, 8 male) after consuming whole fruit vs. blended fruit. The fruit included gala apple, with the seeds removed, and blackberries. We used a repeated measures two-way ANOVA with fruit treatment as the within-subject variable, sex as the between-subjects factor, and glucose maximum, glucose incremental area under the curve (iAUC), and 60 min glucose as dependent variables. Glucose maximum and glucose iAUC were significantly lower (p < 0.05) in blended fruit compared to whole fruit and 60 min glucose was marginally significantly lower (p = 0.057) in blended fruit compared to whole fruit. [Lisa T. Crummett and Riley J. Grosso in Nutrients].

Article discussed at length in episode 257 of The Proof With Simon Hill. Alas, ‘our results should not be extrapolated to commercial fruit smoothies, which typically use apple juice, sorbet, or ice cream as the base, rather than water.’

posted by bq (21 comments total)

Man, if only peer-reviewed research came with recipes.
posted by Halloween Jack at 7:17 AM on April 24, 2023 [1 favorite]

I greatly enjoyed the most recent Maintenance Phase episode about sugar which goes into the confusing and contradictory evidence about glycemic responses to different kinds of sugar. That being said you couldn't catch me drinking a smoothie.
posted by dis_integration at 7:19 AM on April 24, 2023 [1 favorite]

Well I use a soy milk maker to make almond/date milk and then i put a cup of that in the blender with maybe 1/2-3/4 cup of frozen berries and a tsp of clover honey and a dash of salt and voop.
posted by seanmpuckett at 7:23 AM on April 24, 2023

I did not realize there was any sort of controversy here. Everything I’ve read just suggests eating the whole fruit because you’ll get the skin, and of course to know what you’re getting in the smoothie if you buy it.
posted by Tell Me No Lies at 7:25 AM on April 24, 2023

every science article/post/etc should come with a couple of resources like the hierarchy of evidence (for which this would be very low) and an explanation of the difference between statistical significance and effect size (for which this seems fairly small even when it was significant)

on a more fun note, we purchased a blender from an estate recently and I've made smoothies galore since. it's so exciting and fun and they taste so good and are so cheap. I've always personally had an issue with fully chewing things before swallowing whole and now in my continuing pursuit of being a dystopian cyborg and thanks to this blender I'm slowly integrating machine chewing into my diet

what a wonderful future we live in
posted by paimapi at 7:27 AM on April 24, 2023 [26 favorites]

I have wondered about this. Saliva contains enzymes that break down some parts of food. Chewing mixed with saliva must have a somewhat different effect versus gulping down puréed food. But I guess it’s not a big deal.
posted by SoberHighland at 7:44 AM on April 24, 2023

Man, if only peer-reviewed research came with recipes.

Nutrients is an MDPI journal, and the extent to which peer review plays a role in publication there is debated.
posted by grouse at 7:54 AM on April 24, 2023 [9 favorites]

Oh, the other usual warning about smoothies is that they contain a lot *more* fruit. A peach smoothie for example will usually contain a lot more than one peach. For the study the amount of fruit consumed was the same for the whole fruit and the prepared smoothies.
posted by Tell Me No Lies at 7:58 AM on April 24, 2023 [4 favorites]

Sample size of 20 young people tells you everything you need to know about whether the finding is actionable
posted by armoir from antproof case at 8:35 AM on April 24, 2023 [6 favorites]

Did I read here that ice cream is low-glycemic?
posted by grobstein at 8:42 AM on April 24, 2023 [1 favorite]

Sample size of 20 young people tells you everything you need to know about whether the finding is actionable

Especially when the only demographic information is gender and the fact that they're college students.

Which I guess is an improvement on the old standard of male and college student.
posted by Tell Me No Lies at 9:14 AM on April 24, 2023

On first read I mistook "postprandial" for "postcoital" and boy I had a very different idea of when these healthy young people were drinking their smoothies — and some questions about how this study got past the review board.
posted by heyitsgogi at 9:17 AM on April 24, 2023 [4 favorites]

Sample size of 20 young people tells you everything you need to know about whether the finding is actionable

I have no comment on this particular study, but in general the above statement is incorrect.

Frequentist statistics methodology says that if you're showing a statistically significant positive result (rejecting the null hypothesis) with a small sample size, that's good enough. It's only when you're claiming inconclusive results (accepting the null hypothesis) that a large sample size is necessary to justify your conclusion.

Of course, there are more general problems with frequentist hypothesis testing methodology, such as people not giving enough emphasis to effect size or hacking p-values or only publishing "positive" results. But sample size by itself doesn't say anything. A sample size of 20 which is statistically significant and has a large effect size is more meaningful than a sample size of 1 million with a small effect size.
posted by splitpeasoup at 12:26 PM on April 24, 2023 [1 favorite]

Especially when the only demographic information is gender and the fact that they're college students.

At least the article's title references the sample limitations for once: "Postprandial Glycemic Response to Whole Fruit versus Blended Fruit in Healthy, Young Adults".
posted by The_Vegetables at 2:21 PM on April 24, 2023

Sample size of 20 young people tells you everything you need to know about whether the finding is actionable

It's actionable. Every group of 20 young people can pool their avocado money for a year and afford one blender.
posted by srboisvert at 3:30 PM on April 24, 2023 [3 favorites]

COI: I am an empirical Bayesian, and there is no one less Bayesian than an empirical Bayesian. That said, these guys would have their results shrunk into nothingness by anything short of a frequentist approach, and even that didn’t help.

Frequentist statistics methodology says that if you're showing a statistically significant positive result (rejecting the null hypothesis) with a small sample size, that's good enough.

Cool, cool, but that’s not what they did. A p-value of 0.057 is not, by commonly accepted (albeit arbitrary) standards, significant. It’s not “marginally significant” any more than it is “marginally insignificant”. They could have powered their experiment properly and they chose not to, which is why this is in an MDPI journal, where peer review is even less rigorous than at CNS.
posted by apathy at 6:44 PM on April 24, 2023 [4 favorites]

I did not realize there was any sort of controversy here. Everything I’ve read just suggests eating the whole fruit because you’ll get the skin, and of course to know what you’re getting in the smoothie if you buy it.

Read it again. These results suggested that consuming blended fruit results in LOWER blood glucose (60 minutes later and integrated over time) compared with consuming whole fruit. That's the opposite of what I would have expected—namely, blended fruit being to some extent "predigested" and therefore processed and absorbed more quickly.
posted by The Tensor at 12:36 AM on April 25, 2023 [2 favorites]

Fruits and their eaters co-evolved for fruit sugars to be highly bioavailable. Not surprising that the digestion puts and takes for something for which neither the fruit nor the gut evolved to optimize (blenders!) could net out or even tend slightly to indigestibility.

Obviously a higher level of processing is going to push the equilibrium. I'd guess the a handful of blackberries will yield lower glucose levels than a spoonful of blackberry jam made from the same quantity of blackberries even if no sugar is added to the jam.
posted by MattD at 5:54 AM on April 25, 2023

I'm a bit surprised nothing here so far has addressed the theoretical mechanism for blending fruit reducing glycemic response. TLDR - the blending releases magic smoke inside the seeds that delays or inhibits sugar absorption, and this effect is absent in fruits without seeds, also replicated in other studies besides this.

-- grinding the seeds in blackberries during the blending process may have released additional fiber, polyphenols, fats, and protein that would not otherwise be released during mastication or normal digestive processes when the fruit is consumed in whole form

-- polyphenols from fruit have been shown to inhibit glucose transport (via inhibition of SGLT1 and GLUT2 transporters) from the intestine into/out of intestinal Caco-2 cells

-- the addition of fiber and protein, from the blackberry seeds to chyme, may increase chyme viscosity, decrease gastric emptying rate, and decrease the rate and degree of glucose absorption.

-- Haber et al. found that maximum plasma glucose and glucose area under the curve were not significantly different after consuming whole apples, blended apples, and apple juice

--- Redfern et al. reported that blended mango did not have a significantly different glycemic index compared to whole mango in healthy subjects, but blended “mixed” fruit that contained mango, banana, passion fruit, pineapple, kiwi, and raspberries had a significantly lower glycemic index compared to whole mixed fruit
posted by xdvesper at 6:40 AM on April 25, 2023 [4 favorites]

It's worth noting the distinction between "blended" and "juiced" fruit
posted by MengerSponge at 2:04 AM on April 26, 2023

It's worth noting the distinction between "blended" and "juiced" fruit
posted by MengerSponge at 5:04 AM on April 26 [+] [!]

Blended fruits wear fedoras and raincoats and hide behind pillars, juiced fruit is wasted and falling down stairs.

Don't even get me started on stone(d) fruit.
posted by yes I said yes I will Yes at 3:43 PM on April 26, 2023