Do Sweetened Drinks Cause Type II Diabetes?


When I was in my early thirties, I several times visited an island in the Pacific called Nauru. From the medical point of view, it was of the utmost interest because fifty percent of the population has Type II diabetes and it therefore represented the epidemiological shape of things to come.

The Nauruans had become diabetic only recently, when they suddenly (and briefly, as it turned out) became the richest people per capita in the world, thanks to the phosphate rock in which their tiny island was covered. From a life of subsistence on fish and coconuts they went straight to being millionaires. They abandoned their traditional diet and started to eat, on average, 7000 calories per day. Not surprisingly they were enormously fat. They liked sweet drinks and consumed Fanta by the case-load. For those who liked alcohol as well there was Château Yquem. They were unique in the world in being both rich and having a short life expectancy.

The Nauruans were, in a sense, pioneers of the diabetogenic diet.

Type II diabetes is now a threat to public health that dwarfs Ebola virus in scale, but kills slowly and undramatically, rather by stealth than by coups de théâtre. No one ever walked around in spacesuits because there was a Type II diabetic on the ward.

The Nauruans (and the Pima Indians of Alaska) were almost certainly susceptible genetically to the disease, which did not affect them until they adopted their horrible diet. But, to a lesser extent admittedly, what happened to them has happened everywhere, particularly in the U.S. and Britain, where patterns of consumption bear some slight resemblance to those of the Nauruans.

A paper in a recent edition of the British Medical Journal draws attention to the connection, statistical but probably causal as well, between the consumption of sweetened drinks and the prevalence of Type II diabetes. The authors performed a meta-analysis of the results of all the papers on the connection between sweetened drinks and the development of Type II diabetes that in their view were methodologically more or less sound. They tried to distinguish between the effects of the drinks themselves and obesity: obese people drink more sweetened drinks than do lean, and sweetened drinks cause obesity, and obesity causes diabetes, sweet drinks or no.

As is usually the case in such papers, the validity of the conclusions depends importantly on the validity of the statistical methods used, and not one doctor in a thousand is competent to judge them. I am certainly not, but a few things occurred to me, for example that all the papers they reviewed were subject to what is known as recall bias. Patients with diabetes were compared to controls and asked about their past consumption of sweetened drinks. Since the notion that sweetened drinks cause diabetes has entered popular consciousness, it may be that those with diabetes exaggerate their consumption while those without minimize theirs. And the authors use a strange term – “excess diabetes events” – as the endpoint to be measured without explaining what this strange locution means.

After putting the data through the numerical meat-mincer, the authors conclude that sweetened drinks are associated with Type II diabetes, even in the absence of obesity. They come to the conclusion that 1,800,000 “excess diabetes events” are attributable to the consumption of sweetened drinks, and also, in a separate analysis, that artificial-sweetened drinks and fruit juices are not suitable or beneficial substitutes.

I have to admit that I welcomed the authors’ conclusions, but for entirely extra-medical reasons. Sweetened drinks seem to me to be an aesthetic abomination, and their consumption consequent upon a mass outbreak of childishness. I want people to suffer ill-effects from their bad taste, and therefore I believe the conclusions to this paper to be true.