It may appear that scientific skeptics are always criticizing any newly published scientific article that doesn't fit some imaginary point of view. Personally, I evaluate and critique a lot of "scientific" articles that make the rounds on the pseudoscience websites, such as the recent "canola oil causes Alzheimer's disease" nonsense. Now, a new article has been published that claims that non-ionizing radiation causes miscarriage in a respected journal.
Of course, based on this one article, many news organization and legitimate websites have jumped all over it – "Miscarriage rates triple for women with top radiation exposures" is a typical headline. When I see a new science article get that much play in the press, my skeptical radar goes on full sensitivity mode. I just know there's something wrong with the original science.
On the other hand, maybe the new science article is of high quality and may be indicating to us that there's an issue. Let's take a look at whether non-ionizing radiation causes marriage – or not.
First, a little science
Before we get to the paper itself, I need to explain what is meant by "non-ionizing radiation." It refers to any type of electromagnetic radiation that does not carry enough energy to ionize atoms or molecules. In other words, this type of radiation is incapable of removing an electron from an atom or molecule.
On the other hand, ionizing radiation has a higher frequency and shorter wavelength than non-ionizing radiation. Examples of this type of radiation are gamma rays (which are formed by radioactive substances), X-rays, and some types of UV radiation. Ionizing radiation actually damages the biochemicals in the cell, like DNA.
You are exposed to all types of non-ionizing radiation every day in your life. Visible light, for example, is a type of this radiation – it doesn't damage the molecules in your body if exposed to it, but it can cause thermal injury when the light energy is converted to heat. Microwave energy is not ionizing, yet it can "excite" molecules like water creating heat – this is how it cooks your food.
Sunlight also includes another type of radiation in the form of ultraviolet light. UV light can be technically non-ionizing or ionizing depending on the wavelength. However, as we know UV light causes DNA damage, even the non-ionizing form, which can lead to cancers.
This post (along with the underlying peer-reviewed article) focus on a very specific type of microwave radiation – magnetic field (MF) non-ionization radiation. MFs are generated by power lines, home appliances, wireless networks, mobile phone towers, and wireless devices such as mobile phones. As this type of radiation has become more ubiquitous, especially with the proliferation of these wireless networks and cell phones, humans encounter it more frequently than just a few years ago (see Note 1).
There have been some adverse health claims made about several types of magnetic field radiation. For example, the World Health Organization's International Agency for Research on Cancer (IARC) has claimed (pdf) that cell phones are "possibly carcinogenic to humans (Group 2B), based on an increased risk for glioma, a malignant type of brain cancer, associated with wireless phone use." This classification appeared to be based on a major epidemiological study on the use of cell phones and cancer risk.
Dr. David Gorski thoroughly reviewed the evidence that IARC used to determine if cell phone usage was causally linked to certain types of cancer. Let's just say he didn't agree, based on the evidence:
From all this it would appear that the decision making process in which IARC classed cell phones under Category 2B as “possibly carcinogenic” was flawed. IARC was ill informed on the true state of studies of brain cancer incidence rates, claiming that (6) “most of the analyses examined trends until theearly 2000s only“. It ignored numerous warnings by expert groups who had determined that the case control studies by Hardell et al. were outliers in the literature. It accepted a reincarnation of an old study (13, 17, 18) that had been roundly criticized. It undervalued studies which established the existence of recall bias (14) with the statement that: “the Working Group concluded that the findings could not be dismissed as reflecting bias alone, and that a causal interpretation between mobile phone RF-EMF exposure and glioma is possible”. IARC undervalued the solid cohort studies (15, 16) by Schutz et al. of Denmark. “In this study, reliance on subscription to a mobile phone provider, as a surrogate for mobile phone use, could have resulted in considerable misclassification in exposure assessment.”
Of course, we're reviewing whether non-ionizing radiation causes miscarriage, not cancer, but it's important to note that the IARC classification probably has nothing to do with what we're discussing here (see Note 2).
The claim that non-ionizing radiation causes miscarriage
A paper, published in December 2017 in Scientific Reports (a Nature journal), asserted that pregnant women exposed to high radiation levels from magnetic field non-ionization radiation, such as cell phones, wireless networks, and cell towers, had miscarriage rates that were nearly three times the rate as those exposed to lower levels. The lead author, Dr. De-Kun Li, is a senior research scientist at the Kaiser Permanente Division of Research in Oakland, California.
The researchers tracked 913 pregnant women in the San Francisco area who carried meters that measured their exposure to the MF non-ionizing radiation. After controlling for age, race, education, and smoking, pregnant mothers with the highest exposure levels of the MF radiation were 2.7 times more likely to have a miscarriage compared to women who encountered the lowest levels.
Furthermore, researchers could not determine the emission sources of the radiation. So we are not able to ascertain whether a specific type of radiation may or may not have a stronger effect than others.
I think there are some positive attributes to this study that could lead one to say, "yeah, non-ionizing radiation causes miscarriage." Below are some of those positives:
- The population of women in the study are all members of Kaiser Permanente, a vertically integrated healthcare system. Therefore, a complete and thorough health record can be easily accessed for every one of these 913 women.
- This was a cohort study, one of the more powerful forms of observational epidemiology studies.
- The differences in the incidence of miscarriage seemed to be statistically significant.
- A systematic review found that there may be a link between this type of radiation and spontaneous abortions.
Despite these results, Kaiser Permanente has not issued any kind of warning to their pregnant patients about the potential risks of non-ionizing radiation. On the other hand, the California Department of Public Health issued warnings in December 2017 about keeping cell phones away from the body. I'm not sure if this warning was made coincidentally to this article or as a result.
But I can't stop there. This article has some serious issues.
Concerns about the article
Despite the apparent high quality of the article, along with its publication in a Nature journal, there are a lot of things that bother me about it. In no particular order, here are my concerns:
- The study only utilizes Kaiser Permanente patients in one area of Northern California. This is a select group of individuals who may not represent a larger population. Epidemiological studies should have broader patient enrollment for better statistical accuracy.
- The article partially relies on the IARC claims that cell phones cause cancer. Dr. Gorski debunked that link pretty thoroughly, but I'm sure some readers will say in a WHO vs Dr. Gorski battle, WHO wins. Like I keep saying, science is not dogmatic but is provisional. Science relies on evidence, and if that evidence is of bad quality and easily criticized, then one can only criticize and dismiss the conclusions that may be based on that bad evidence.
- The article also relies upon a rat study, another rodent study which has limited applicability to clinical expectations. And once again, Dr. Gorski thoroughly critiques the research (the results of which have not been published). Gorski writes, "the results reported are almost certainly due to chance and are not indicative of a real biological effect."
- Although this study may indicate correlation, it has not established causation. There are numerous variables that have not been described that may have influenced the results. I could hypothesize that those who are around more non-ionizing radiation are in more stressful environments which could have an effect on pregnancy outcomes. Additionally, this study did not account for infectious diseases, like the flu, which are known to have negative outcomes on pregnancy.
- To establish causation, researchers need to establish biological plausibility, which Dr. Li has not done. What possible pathway could be demonstrated that will provide a link between the radiation and the effect? I'm having a difficult time imagining one. These forms of radiation are of extraordinarily low power (like I said, a cell phone transmits its signal at a 0.1-0.3 watts). Is there any evidence that these microwaves, at that power level, can penetrate to the fetus. And even if it could, what biological effect has ever been established? None, as far as I can tell. It is possible that the microwave energy from cell phones, for example, could heat a cell, but by how much? I'm not a physicist, but it would take over 17 watts of energy over 2 minutes to raise the temperature of a 1 kg fetus by 1ºC. That's 100 times more energy than you'd find in a cell phone planted right next to the fetus.
- As far as I can tell (and please correct me so that I can edit this claim), the meter was only used to measure the patients for a grand total of 24 hours. A 9-month pregnancy would include around 6500 total hours – so, they measured 0.3% of the total pregnancy. This is a troubling issue because we have no indication whether the actual radiation received over time varies enough that the two groups end up being the same.
- The testing device, an EMDEX Lite meter, in this study measured frequencies in the range of 40–1000 Hz. WiFi and cell phones use frequencies from 800 to 2600 Mhz. Those sets of frequencies do not overlap. How does this device measure cell phone radiation if it does not receive in the cell phone's frequency range? I don't know, and the authors of the study did not go into any detail about the device. I had to actually Google the device to get information. Was that an intentional omission?
- There are dozens of radiation sources that occur in the range of the EMDEX meter yet were not accounted for by the authors. Vacuum cleaners, hair dryers, solar magnetic storms, and other common things produce far more magnetic radiation than cell phones – was a difference in vacuum cleaner usage between the groups? The potential error of their measurements is so imaginably large, I'm beginning to wonder whether the peer-reviewers actually thought about these issues.
Summary of non-ionizing radiation causes miscarriage
I went into this review of this study with an open mind. I thought that it was of sufficient quality that it would provide possibly robust evidence to actually support the clickbait headlines I was reading in the press. But after digging into the results, the lack of plausibility to establish correlation, and the very weak methodology, I now think that this study tells us almost nothing. This is not convincing science.
You might subscribe to the precautionary principle fallacy that if you think it might happen, you should avoid it. But this study is so flawed at every level, I think you'd be wasting your time worrying about it. Besides, it is borderline impossible to avoid non-ionizing radiation, but I guess you could go live in the mountains in Montana and get away from it all.
This study could be improved by doing things like finding a better radiation meter and vastly increasing the amount of time the patient wears the meter (although compliance may be a huge issue). Researchers also have to establish a compelling and robust body of evidence that there is a plausible relationship between cell phones and miscarriage.
So, do cell phones and other non-ionizing radiation have anything to do with miscarriage? Not as far as I can tell. Right now we are at the point that the absence of evidence is still evidence of absence.
Notes
- Just to be clear, even though your cell phone transmits microwave radiation, it does not mean it can cook your frozen dinner. Microwave describes the wavelength of the radiation. A typical microwave used for cooking has a power of 1000 watts (give or take a few watts). A typical cell phone pushes about 0.1 to 0.3 watts. So please do not set your iPhone next to your frozen chicken and expect it to do something.
- I have been highly critical of IARC previously. And again. That being said, IARC does a good job with classifying potential carcinogens, although they are not perfect. I do not get their cherry-picking of studies that they seem to use to support a priori conclusions about certain things that may or may not cause cancer. And some of their conclusions go against a strong scientific consensus that says otherwise.
Citations