Refutation of dirty electricity hypothesis in obesity: epistemological arguments and trans-disciplinary study using an instrumental variable

To the editors,

We read with interest a paper by Dr. Milham recently published in this journal (Milham, 2013). In this paper, Dr. Milham observes that 8 of 10 countries in the world with the highest Body Mass Index (BMI) in 2008 were small islands in the Pacific (n = 7) and the Caribbean (n = 1) and were also the countries with the highest fasting plasma glucose (FPG) and prevalence of diabetes prevalence. These facts derived from an internet search were abruptly followed by a conclusion that these conditions were caused by the fact these islands obtain electricity from diesel generators that allegedly exposed the inhabitants to “dirty electricity” (high frequency voltage transients (HFVT)).

To place this study into perspective, it is important to emphasize that there is at present no scientifically valid peer-reviewed data indicating that HFVT are associated with increased risks of diabetes and other conditions cited by Dr. Milham (asthma, ADHD, cancer). In citing previous work, Dr. Milham did not include a systematic review in which all available evidence was reviewed by de Vocht (2010). The review concluded that the few studies that were done, mostly by the same researchers, were all subject to methodological problems large enough to prohibit valid assessment of any biological activity of HFVT. In addition to the problems in the design of the studies and the interpretation of the results, it is also very unlikely that any single exposure would cause such a wide range of health effects, further indicating that HFVT most likely do not cause most, if not all of the reported health effects.

Regardless of any prior evidence of health risks of HFVT, this ecological study conducted by Dr. Milham is fatally flawed and is a vivid illustration of incorrect application of ecological analyses in epidemiology. Even if the presence of diesel generators correlates with average BMI, fasting plasma glucose (FPG) and diabetes, any causal arguments arising from this observation must account for confounding and temporality. More likely, high BMI, FPG and prevalence of type 2 diabetes is caused by lifestyle factors (Diabetes UK, 2013), while on small islands genetic factors may also play a role. Obvious suspected risk factors, exhaustively reported in literature, are diet and insufficient activity. Indeed, the World Health Organization described that one of the barriers to an effective response to obesity on many Pacific Island countries is that, culturally, large physical size is considered a mark of beauty and social status; furthermore imported high-fat energy-dense food consumption has been increasing on these islands (WHO, 2002). Another important barrier is reduced physical activity because of the use of modern technology further aggravated by the high rate of violence and crime that reduces the opportunity for outdoor activity (WHO, 2002). These factors indicate that the position of the small islands in the ranking of BMI, FPG and diabetes can satisfactory be explained by recent trends in known risk factors and that it is very unlikely HFVT have an important role to play, if any. Dr. Milham further also fails to address what came first: increase in outcomes or introduction of diesel generators to the islands. Everybody knows that cause must precede effect so the failure to appreciate this by Dr. Milham is baffling.

Obviously, established risk causative factors, although more likely than the one proposed by Dr. Milham, are subject to the same problems of interpretation from ecological cross-sectional data. Therefore, we conducted a similar study to the one by Dr. Milham to investigate the hypothesis that if dirty electricity in Oceania causes obesity in humans, it should have the same effect on Rattus norvegicus (common brown rat). To test the hypothesis, we compared the size and weight of R. norvegicus in Oceania to the rest of the world (especially The Netherlands, where levels of “dirty electricity” are much lower (Milham, 2013)). To minimize residual confounding we stratified the results by wild and laboratory-bred R. norvegicus. We used the methodology of Dr. Milham: internet search. The instrumental variable used in this study was the size of R. norvegicus which has similar exposures to humans, including most notably to “dirty electricity” but without the socio-cultural confounders. Moreover, we hypothesized that in lab-bred R. norvegicus exposure to “dirty electricity” will be similar to all rats in a geographical area because of their controlled environment and proven exposure due to presence of electrical devices in labs. The R. norvegicus populations were found to have a tail length ranging 17–23 cm, a head/body length ranging 20–28 cm, and a weight ranging 275–575 grams (ARKIVE, 2013; Burnie, 2001). No evidence was found of R. norvegicus being larger in Oceania than in the rest of the world (Burnie, 2001; redOrbit, 2013). There was no evidence of laboratory rats being more overweight in laboratories in Oceania than elsewhere in the world. We conclude that dirty electricity does not cause obesity in R. norvegicus, and therefore it does not have corresponding effect in humans.

Although we never conducted this study and it was obviously meant as a parody on the study by Dr. Milham, the proposed experiment conducted at an individual level would provide a sound scientific basis for the evaluation of Dr. Miham’s ideas. We therefore encourage Dr. Milham to visit Oceania and perform such informative experiments in laboratory rodents. In the meantime, we hope that readers of the work by Dr. Milham will put on hold plans to construct alternative sources of power such as nuclear power plants or coal-fired power stations on small islands in Oceania and the Caribbean.

Acknowledgements

The authors report no conflicts of interest.