Fruit and diabetes - a cocktail of results

Do you remember how much fresh fruit you ate last year?  Or, ok, in the last three months?  Or even last week?  What about differentiating between fruits; strawberries vs cantaloupe vs blueberries?  And how many??

We ask because a new study published in the British Medical Journal ("Fruit consumption and risk of type 2 diabetes: results from three prospective longitudinal cohort studies," Muraki et al.), that's getting a lot of news play, reports that eating a serving of blueberries at least three times a week protects against type 2 diabetes (t2d).  Blueberries are best, but grapes and raisins are second best and apples and pears are third.  Cantaloupe, on the other hand, seems to be a risk factor, as does drinking your fruit as juice rather than eating it whole. 

The study looked at food questionnaire data from a ton of people; "66 105 women from the Nurses’ Health Study (1984-2008), 85 104 women from the Nurses’ Health Study II (1991-2009), and 36 173 men from the Health Professionals Follow-up Study (1986-2008)."  So the data must be robust--yes?

Food questionnaires are a well-established tool for eliciting dietary information, and they are used all the time.  But they are also notoriously unreliable, for reasons that are easy to understand; it's really hard, first, to think in terms of standardized portion sizes and, second, to remember how often you eat a food, particularly if it's seasonal.  And there's the subtle possibility of responder bias--knowing what the study is looking for.

An anecdote, for what it's worth.  My mother has been filling out food questionnaires as a subject in the Nurses' Health Study for decades.  I remember her reaction to the first one she was asked to complete because I was a graduate student in public health at the time, learning in epidemiology classes about state-of-the-art tools like diet surveys, so the fact that they might be seriously flawed was an eye opener.  Suffice it to say, she did a lot of guessing. It's not good science to extrapolate from a single case, I know, but the following, from the paper, suggests it might actually be valid in this instance.

The food frequency questionnaires were validated against diet records among 173 participants in the Nurses’ Health Study in 1980 and 127 participants in the Health Professionals Follow-up Study in 1986. Corrected correlation coefficients between food frequency questionnaire and diet record assessments of individual fruit consumption were 0.80 for apples, 0.79 for bananas, and 0.74 for oranges in women, and 0.67 for total whole fruits, 0.76 for fruit juice, 0.95 for bananas, 0.84 for grapefruit, 0.76 for oranges, 0.70 for apples and pears, 0.59 for raisins and grapes, and 0.38 for strawberries in men.

So portion size and frequency are hard to remember, and that's a problem.  But maybe there's something else influencing the results of this study.  We're betting it's a lot easier to remember when you consume something every day, like a glass of orange juice, or banana on your cereal, but harder to remember when you have something that is more likely to be only seasonally available, like plums, or apricots.  This could explain why there seems to be such variation in the accuracy with which people remember different fruits and juice.

But probably more significantly, we'd bet that people who are already at lower risk of t2d because they exercise, or watch their diets, are thinner and so forth, are the same people who include more fruits and vegetables in their diets.  So, blueberry consumption may be indicative of a low-risk lifestyle rather than nutritional components that protect against t2d.  And this is what was found.

In all three cohorts, total whole fruit consumption was positively correlated with age, physical activity, multivitamin use, total energy intake, fruit juice consumption, and the modified alternate health eating index score, and was inversely associated with body mass index and current smoking. Whole fruit consumption was associated with an increased probability of using post-menopausal hormones in the Nurses’ Health Study and with a reduced probability of using oral contraceptives in the Nurses’ Health Study II.

The investigators corrected for these correlations -- that is, essentially asked the question, "In the lower risk group in the study, are those who eat X fruit at even lower risk?".  These adjustments weakened the associations, which isn't surprising, and in fact adjusting for other things like gestational diabetes or cancer also attenuated the associations.  But they still found that blueberries, raisins, grapes, pears and apples were correlated with lower risk, and cantaloupe with higher.  Methodological questions aside, could blueberries really be protective against type 2 diabetes, but not cantaloupe?

The idea is that fruits are rich in fiber, antioxidants and phytochemicals, all of which are presumed to be protective.  But, some are sugary, which might instead be a risk factor.  Results of studies of the role of specific fruits in risk of t2d have varied, with different fruits sometimes implicated in risk and sometimes with protection. "In eight previous prospective studies, the association between total fruit consumption and risk of type 2 diabetes was examined, and the results were mixed."

And, the study found that association of risk with glycemic load was inconsistent, and varied by cohort; "...a significant, inverse association was found in the Nurses’ Health Study, but not in the other two cohorts."  Indeed, this wasn't the only outcome that varied by cohort.    

In the Nurses’ Health Study II and Health Professionals Follow-up Study, banana consumption was associated with a lower risk of type 2 diabetes, whereas in the Nurses’ Health Study a non-significant positive association was found. The association for strawberry consumption was significantly positive in the Health Professionals Follow-up Study but was non-significant and inverse in the Nurses’ Health Study.

The authors conclude about their findings that most, but not all, "were quite consistent among three cohorts." We would suggest, however, that despite the large cohort sizes, methodological issues, particularly recall issues with the food questionnaires, are significant enough that you shouldn't make your decisions about which fruit to put on your cereal based on these results.

The issues are mainly examples of data (recall) error, and confounding:  so many factors go together in this society, bombarded by health-advice 'news' (and 'scientific' reports like the blueberry study), that it is very difficult to disentangle them.  Usually, then each individual factor has low effect on its own.

You can make your own judgement about whether these kinds of study are worth reading, publishing, or funding.  You might, for example, classify them with GWAS and other 'Big Data' studies, and say that when many different factors are at play, with individually minor effects, and people know what the latest advice is and adjust their behavior (often subtly) accordingly, that we're just asking to be confounded when we ask about the wisdom of exposure to individual factors.

Unless, of course, you're from New England and you're partial to blueberries.  In that case: go ahead and dose up!  Otherwise, read something more worthwhile than massive, inconclusive, weak-factor studies.