One-sided push towards the link between drinking soda and diabetes

I continue watching that 2013 video of Dr Lustig. At 31:17 minute he talks about a study, which analyzed data of cohort from the EPIC - InterAct study performed in Europe. This particular analysis was published in a journal Diabetologia.

The study worked out the association between drinking sweet beverages (i.e. juices and nectars, sugar-sweetened soft drinks and artificially sweetened soft drinks and the incidence of type 2 diabetes by following the cohort over time.  

The study says: For the present analysis we used information about: smoking status. alcohol intake, educational level, and an ordered four-category index of physical activity. They also adjusted the results for BMI and total energy intake.

These were the potential confounders that could have messed-up with the association and providing false positive or negative results, so they adjusted for them, too. Do you see fats in there anywhere? Or some other dietary factors, except the alcohol? Liver is a major organ for processing not only sugar/fructose, but also fats and synthesizing cholesterol. Too much fat from fructose has been claimed to lead to insulin resistance. How about too much fat from the diet that the liver has to process as well? Nothing like this has been discussed here or elsewhere in similar videos. 

Dr Lustig happily presented the diagram:

and saying that: 

"Adjust for calories and BMI, you just look at sugar sweetened beverages against a risk for diabetes, every sugar sweetened beverages you consume increases your risk of diabetes by 29% over time. This is in doses commonly used, this is human data, but this is correlational data, because it is a snapshot over time. "

Then he offers results of their study, claiming to prove causation data. I will discuss that in the next article. 

For now I would like to say, that Dr Lustig has used this source to support his anti-sugar agenda, despite the evidence is incomplete and flawed. It correlated one kind of food/drink and controlled for too few confounders for it to be taken seriously. Why? 

Because this is an example of basic fallacy in the epidemiology research. People consuming soft drinks also tend to have generally unhealthy diet, especially rich in fat, which was not controlled for in this “study”. The scientists who analyzed the selected data did not look for a real correlation between the diabetes and any possible dietary variable, they were looking whether they can find the correlation pointing at sugar. And they found it. Surprised?

The sugary drinks could only serve as a marker of wrong diet here, not a hint for a causal relationship. As a proof can be considered the lack of association with the fruit juices, which is often the case of health conscious individuals, who also have a generally healthier lifestyle.

The only value I would put in it is that it adjusted for BMI, making the risk of diabetes independent of BMI. However, other studies found different results, so one study like this does not prove anything. It is a whole body of evidence that is important and that takes time and a lot more studies.