buffy, do you want to consider submitting out paper to the Lancet. It’s as good as this paper.
poikilotherm said:
mollwollfumble said:
> a percentage of 50–55% energy from carbohydrate was associated with the lowest risk of mortality.
They need to distinguish between proteins and fats. Also, are they taking into account the enormous difference between pure carbohydrates and carbohydrate-rich foods. I can see that I’ll have to read the technical article to make sense of it.
they did, you should read the actual study instead of my cut’n‘past brief findings section.
Reading actual study now.
> We studied 15 428 adults aged 45–64 years, in four US communities, who completed a dietary questionnaire at enrolment in the Atherosclerosis Risk in Communities (ARIC) study between 1987 and 1989, and who did not report extreme caloric intake. During a median follow-up of 25 years there were 6283 deaths in the ARIC cohort.
Good.
> After multivariable adjustment
Good.
> Low carbohydrate dietary patterns favouring animal-derived protein and fat sources, from sources such as lamb, beef, pork, and chicken, were associated with higher mortality, whereas those that favoured plant-derived protein and fat intake, from sources such as vegetables, nuts, peanut butter, and whole-grain breads, were associated with lower mortality.
Should that be the real conclusion of the study?
> The meta-analysis of large cohort studies in North America and Europe has suggested increased mortality associated with low carbohydrate intake. Conversely, recently published multinational and Asian studies have reported increased mortality in association with high carbohydrate intake.
Perhaps that should be the real conclusion. Perhaps that means that the methodology is stuffed.
> Combining these data with those from other North American, European, Asian and multinational cohorts. We identified seven studies in addition to the index cohort (432 179 participants, 40 181 deaths).
Well in that case, perhaps the present study (because 6 << 40) has only negligible significance. Reading further, no, the same conclusions from the present study and from the other studies combined would be important confirmation of both the present study and the methodology.
> Participants completed an interview that included a 66-item semi-quantitative food frequency questionnaire. We used the Harvard Nutrient Database to derive nutrient intakes from the FFQ responses.
So far so good.
> We analysed the covariates of age, sex, race, education, smoking, exercise, energy intake (kcal), and diabetes.
So far so good.
> total energy from carbohydrate
As I my reply to the OP, there’s still no mention of whether the carbohydrate was from whole foods or in purified form.
> We created actuarial estimates of the age-specific probabilities of death according to each category of carbohydrate
“Category” is the wrong word here because it’s not different types of carbohydrate. “Percentile range” is more accurate, 6 classes.
> We chose a reference group of 50–55% … Mean carbohydrate intake was 48·9%.
The smallest percentile range, is that because it’s what most people do naturally? Looks like it to me.
> We used restricted cubic splines
Like buffy and me
> Participants with the lowest carbohydrate range … high body-mass index … smoking
Wait on, is this really a paper about the dangers of obesity and smoking, ie. are these treated as independent variables or not? You see, that can be read both ways: one would be directly caused by carbohydrate intake and the other would be indirect via smoking or obesity (or other). Both approaches are important and the paper should analyse it both ways. Do they?
> Prevalence of hypertension was similar across carbohydrate quantiles
That’s good to know.
I wish they’d written down number of deaths in Table 1. That was too important to be omitted, considering we’re supposed to be talking about mortality.