March 20, 2017

expert reaction to diabetes and rising temperatures

Researchers published in Open Diabetes Research & Care claim that the diabetes incidence rate in the USA and prevalence of glucose intolerance worldwide increase with higher outdoor temperature.

Prof. Ashley Grossman, Professor of Endocrinology at the University of Oxford and member of the Society for Endocrinology, said:

“Obesity with its bedfellow type 2 diabetes is a worldwide scourge, and traditionally we could only advise people to eat less and exercise more. However, the discovery of ‘brown fat’ and the ‘browning’ of white fat means that there may be other ways to decrease its incidence.

“There is some rather encouraging evidence that cooling the body, even by a few degrees, may improve or reduce diabetes. Now, Blauw and colleagues from Holland have speculated that if this is true, then there should be a correlation between environmental temperature and the prevalence of diabetes, and they claim that examining environmental temperature data from US states, and some countries, there is indeed a (small) but significant negative correlation. They also feel this supports the speculation that global warming may play a part in the ever-increasing prevalence of type 2 diabetes.

“Interestingly, this appears to hold even when body weight – obesity – is accounted for. This encourages the ‘keep cool’ theory of decreasing diabetes and possibly obesity, although of course one can never be certain that correlation means causation. And there are still questions about the study; we know that Americans in cold climates tend to stay in centrally-heated homes, while in hotter states there is ubiquitous air-conditioning, so such a correlation with the outside temperature is possibly unexpected and a little puzzling.

“Nevertheless, it is just a little more suggestive evidence that living in a cool environment may be useful to increase insulin sensitivity and ward off diabetes. Together with work indicating that adequate sleep can also help avoid obesity and diabetes, maybe we should all aim have a good night’s sleep in a cool bedroom with the windows open to the night breeze.”

Dr Simon Cork, Research Fellow at the Department of Investigative Medicine, Imperial College London, said:

“To be honest, I’m shocked that this has been published.  It’s publications like this that reduce public trust in the scientific process.

“The authors have correlated global increases in temperature with the increasing incidence of type 2 diabetes, and concluded that increasing temperatures may reduce brown adipose tissue activity which therefore leads to diabetes. Unfortunately the authors do not provide any evidence to back this up, and is a clear case of correlation, not causation.

“A lot has changed since temperatures began to increase. I can’t begin to comprehend how they came to the conclusion that this might be linked to reduced brown adipose tissue activity. Their own data doesn’t even support this idea since cold US states have the same diabetes rate as warm ones (compare Alaska with Iowa for example). Also adults have varying amounts of brown adipose tissue, with some adults having very little to none, and again they haven’t taken this into account – not that that is the most glaring omission in this study.

“They do not correct for any other factors (for example, the increasing consumption of sugar or decreasing activity levels over the same time frame), nor do they provide any evidence that a reduction in brown adipose tissue is driving the increase in diabetes.

“The way the study has been conducted means the conclusions drawn are not appropriate.”

Dr Louise Brown, Senior Statistician at the MRC Clinical Trials Unit at UCL, said:

“The authors have used meta-analytic methods to combine the estimates of association between increase in annual average temperature and measures of poor glucose control such as annual incidence of diabetes, fasting glucose levels or glucose intolerance. They have performed these meta-regressions across states within the USA and more widely across global countries. They have shown association but they are a long way from demonstrating any kind of causality.

“They do not quote any measures of heterogeneity across their meta-analyses nor do they attempt to explain why some states demonstrate a negative association whilst others a positive association. If their hypothesis is that average annual temperature is regulating the incidence of diabetes via the production of brown adipose tissue (BAT) then it is difficult to understand why this association would be reversed in some states, particularly as there is little indication of a pattern of temperatures across those states.

“They dismiss the possibility that there are other important residual confounding effects that might explain this association. I feel this is naïve given that activity levels are known to differ by both temperature and cultural differences across states as well as countries. There is no mention of other curveball effects on human behaviour in relation to temperature, e.g. many of the hotter US states will have air conditioning – could that be having an effect – I think not. If they believe that this association is causal, then they should be calling for a study that is designed to measure longitudinal data on BAT, ambient temperature, glucose levels and diabetes diagnosis within individuals followed over time, with appropriate adjustment for other confounding factors.

“It is not clear to me whether this published analysis was part of a much larger analysis looking at many factors associated with the rising incidence of diabetes. Did they look at lots of metrics and are only publishing the one that demonstrated statistical significance? Are they able to confirm that they performed these analyses according to a pre-specified statistical analysis plan with this as their only hypothesis?

“Overall, the uncovering of this association is interesting but I do not feel that is of great help in our fight against the increasing global incidence of diabetes, unless they are suggesting that we all move to colder climates. Nevertheless, the authors do acknowledge that their findings merely demonstrate association rather than causation and if they have stumbled across a useful pointer that leads to appropriate metabolic research on the role of BAT in the development of diabetes then great, but their claims are too strong at this stage.”

Prof Sir David Spiegelhalter, Winton Professor for the Public Understanding of Risk at the University of Cambridge, said:

“Even if these estimates were true, it would mean a 2 degrees rise in average temperature was associated with an increased incidence of diabetes of 0.7%. In those circumstances I don’t think this would be the biggest concern.”

* ‘Diabetes incidence and glucose intolerance prevalence increase with higher outdoor temperature’ by Lisanne L Blauw et al. will be published in BMJ Open Diabetes Research & Care on Monday 20 March 2017.

Declared interests

None declared