Obesity trends in the U.S. have a little connection with sugar intake

I would like to remind to the readers the outdated information presented by Dr Lustig all over again. In the video (16:05) I noticed how he again showed in 2013 the trends in obesity prevalence in the US up to year 2003-4 as the present trend:

I do not doubt the data and their value - in the past. But the more recent trends show that the obesity has been rather leveled off among adults:

The article also says that the same trend is suggested for children, too. Similar trend was observed in the UK.

Here I would like to remind you, that the sugar consumption in the U.S. has been decreasing since 2000 while you can see that the obesity kept rising.  

And, surprise surprise, Dr Lustig, after those years of presenting to you that obesity rose as the sugar consumption rose, he finally admitted in a video at 27:04 (2014) that these two variables do not quite match. 

The yellow shade refers to obesity, the blue line to refined sugar, red line to HFCS and the green line to fruit juice. These are the corporate data, which he refused in my argumentation in the past as unreliable, yet he keeps showing them again when it suits him. The sugar and obesity not only do not correlate perfectly, they hardly correlate at all, especially when other sources of energy are added to the equation:

Just have a look at the time point above year 2000. Sugar intake went down, while fat intake shot up, perfectly matching the trend of rising obesity. Moreover, as the sugar consumption has been increasing over the decades, the fat went along this sugar. Quite interesting, right? 

The seven or 22 countries study - what is the issue?

At 17:16 minute of the video Dr Lustig criticized Ancel Keys that he cherry-picked 7 countries instead of all 22 countries included in the study when examining the link between the mortality due to the coronary heart disease and the consumption of either sugar or fat :

While I do not know exactly why this was done (I am open to suggestions) I have no problem in seeing the trend on the right diagram, too. The correlation simply IS there. And it is NOT A MINOR correlation as Dr Lustig announced. The correlation follows the same curved line, only the points are more widely distributed. So what is the fuss about this? To me, the line looks like almost the best fitted line across the 22 country points.

Straight after, Dr Lustig points at outliers that were not included in the publication at all:

I again do not know the reason why these were not included. The reasons could be various: either a poor validity of the dietary assessment, high energy expenditure in the tribe, distinct physiology as a result of isolated evolution of the tribe, on which our general physiology does not quite apply... anything could have contributed and become a reason that these distinct societies were not included at the end. Ancel Keys was an experienced and endorsed scientist and researcher. Although I cannot comment on the allegations about a possible background lobby by the industry, his professional expertise was undeniable.

In fact, Dr Lustig said that these few indigenous tribes would blow a hole into Ancel's hypothesis. He further said that these tribes only consume fat, have "no carbohydrate at all" in their diet (really?) and suffer the lowest incidence of heart diseases. Maasais also consumed honey (sugar); the diet composition of Rendille tribe is hard to find, but they are nomadic, which means they also eat what they find, which is not always meat or fat; the Tokelau tribe obtains most of their energy from the coconut oil whereas Inuits get a lot of omega-3 fats from their cold-sea animals based diet.

Indeed, Ancel Keys believed that "different kinds of dietary fat have different effects on health", according to Wikipedia. This may well apply on the intake of saturated fat as on the poly-unsaturated fat, or different length of the saturated fatty acids, since coconut oil, albeit mainly saturated fat, does not seem to have the same effect as the saturated fat of animal origin. Even butter (milk fat) has a different fatty acids composition than lard, especially when the cattle is freely grazing instead of being fed with grains. Ancel may not have known all this in the past, but he was certainly on the right track. 

Going back to the diagrams, having outliers in the data is nothing unusual and statisticians can handle that. Outliers are commonly excluded when they disturb otherwise obvious trend in the data and often examined separately. Therefore there is nothing like blowing a hole into the Ancel Keys hypothesis as Dr Lustig tried to imply.

Overall, it is apparent that the positive trend was there even when including all 22 countries and I believe it would have a high statistical significance and a correlation coefficient. I have not read the study myself and I welcome any relevant comments and explanations from others.