Content uploaded by Myrtill Simko
Author content
All content in this area was uploaded by Myrtill Simko on Sep 17, 2019
Content may be subject to copyright.
Available via license: CC BY 4.0
Content may be subject to copyright.
International Journal of
Environmental Research
and Public Health
Review
5G Wireless Communication and Health E↵ects—A
Pragmatic Review Based on Available Studies
Regarding 6 to 100 GHz
Myrtill Simkó* and Mats-Olof Mattsson
SciProof International AB, Vaktpoststigen 4, 83132 Östersund, Sweden;
mats-olof.mattsson@sciproof-international.se
*Correspondence: myrtill.simko@sciproof-international.se
Received: 19 August 2019; Accepted: 11 September 2019; Published: 13 September 2019
!"#!$%&'(!
!"#$%&'
Abstract:
The introduction of the fifth generation (5G) of wireless communication will increase
the number of high-frequency-powered base stations and other devices. The question is if such
higher frequencies (in this review, 6–100 GHz, millimeter waves, MMW) can have a health impact.
This review analyzed 94 relevant publications performing
in vivo
or
in vitro
investigations. Each
study was characterized for: study type (
in vivo
,
in vitro
), biological material (species, cell type, etc.),
biological endpoint, exposure (frequency, exposure duration, power density), results, and certain
quality criteria. Eighty percent of the
in vivo
studies showed responses to exposure, while 58% of the
in vitro
studies demonstrated e↵ects. The responses a↵ected all biological endpoints studied. There
was no consistent relationship between power density, exposure duration, or frequency, and exposure
e↵ects. The available studies do not provide adequate and sufficient information for a meaningful
safety assessment, or for the question about non-thermal e↵ects. There is a need for research regarding
local heat developments on small surfaces, e.g., skin or the eye, and on any environmental impact.
Our quality analysis shows that for future studies to be useful for safety assessment, design and
implementation need to be significantly improved.
Keywords: radiofrequency electromagnetic fields; MMW; in vivo; in vitro
1. Introduction
Recent decades have experienced an unparalleled development of technologies that are categorized
as information and communication technologies (ICT), which include wireless communication used
for mobile telephony (MP) and e.g., Wi-Fi by using electromagnetic fields (EMF). The first generation
of handheld mobile phones were available for individual, private, customers in a few countries in
the late 1980’s. Subsequently, the second (2G), third (3G), and fourth (4G, LTE) generations increased
their penetration rates in the society in a dramatic way, so that today there are more devices than
inhabitants of the Earth. In addition, Wi-Fi and other forms of wireless data transfer have become
ubiquitous, and are globally available. At present we are starting to introduce the next generation, 5G,
of mobile networks. Importantly, 5G is not a new technology, but an evolution of already existing G1
to G4 technologies.
With the upcoming deployment of 5G mobile networks, significantly faster mobile broadband
speeds and increasingly extensive mobile data usage will be ensured. This is made possible by the use
of additional higher frequency bands. 5G is intended to be the intersection of communications, from
virtual reality to autonomous vehicles to the industrial Internet and smart cities. In addition, 5G is
considered the base technology for the Internet of Things (IoT), where machines communicate with
machines (M2M communication). At the same time, a change in the exposure to electromagnetic fields
(EMF) of humans and the environment is expected (see, for example [1,2]).
Int. J. Environ. Res. Public Health 2019,16, 3406; doi:10.3390/ijerph16183406 www.mdpi.com/journal/ijerph
Int. J. Environ. Res. Public Health 2019,16, 3406 2 of 23
The 5G networks will work with within several di↵erent frequency bands (Table 1), of which
the lower frequencies are being proposed for the first phase of the 5G networks. Several of these
frequencies (principally below 1 GHz; Ultra-high frequencies, UHF) have actually been or are presently
used for earlier mobile communication generations. Furthermore, much higher radio frequencies
(RF) are also planned to be used at later stages of technology evolutions. The new bands are well
above the UHF ranges, having wavelengths in the centimeter (3–30 GHz) or the millimeter ranges
(30–300 GHz; millimeter waves, MMW). These latter bands have traditionally been used for radars
and microwave links.
Table 1. Subdivision of the 5G frequency spectrum.
Frequency Range Use Comments
<1 GHz Net coverage, IoT Already partly used for earlier MP generations, longer
range coverage, less costly infrastructure
1–6 GHz Net coverage, IoT,
capacity for data transfer
More spectrum available, shorter range and reduced
performance compared to higher frequencies
>6 GHz Capacity for very high
data transfer
Short range, allows high speed data transfer and short
latency times
The introduction of wireless communication devices that operate in the high frequency parts
of the electromagnetic spectrum has attracted considerable amounts of studies that focus on health
concerns. These studies encompass studies on humans (epidemiology as well as experimental studies),
on animals, and on
in vitro
systems. Summaries and conclusions from such studies are regularly
published by both national and international committees containing relevant experts (see e.g., [
3
–
5
].
The conclusions from these agencies and committees are that low level RF exposure does not cause
symptoms (“Idiopathic Environmental Intolerance attributed to Electromagnetic Fields”, IEI-EMF), but
that a “nocebo” e↵ect (expectation of a negative outcome) can be at hand. Some studies suggest that RF
exposure can cause cancer, and thus the International Agency for Research on Cancer classified RF EMF
as a “possibly carcinogenic to humans” (Group 2B) [
3
]. In a recent recommendation of a periodically
working Advisory Group for IARC “to ensure that the Monographs evaluations reflect the current state
of scientific evidence relevant to carcinogenicity” the group recommended radiofrequency exposure
(among others) for re-evaluation “with high priority” [
6
]. There is further no scientific support for
that e↵ects on other health parameters occur at exposure levels that are below exposure guideline
levels, even though some research groups have published non-carcinogen related findings after RF
exposure at such levels (see [
4
,
5
]). Environmental aspects of this technological development are much
less investigated.
Frequencies in the MMW range are used in applications such as radar, and for some medical uses.
Occupational exposure to radars have been investigated in some epidemiological studies, and the
overall conclusion is that this exposure does not constitute a health hazard for the exposed personnel [
7
].
This is due to that exposures for all practical purposes are below the guideline levels and thus not
causing tissue heating. However, further studies are considered necessary concerning the possible
cancer risk in exposed workers. Medical use of MMW has been recently reviewed [
8
,
9
] suggesting
a possibility for certain therapeutic applications, although the action mechanisms are unclear.
The 5G networks and the associated IoT will greatly increase the number of wireless devices
compared to the present situation, necessitating a high density of infrastructure. Thus, a much higher
mobile data volume per geographic area is to be created. Consequently, it is necessary to build
a higher network density because the higher frequencies have shorter ranges. The question that arises,
is whether using the higher frequencies can cause health e↵ects?
Exposure limits for both the general public and occupational exposure are available and
recommended by the WHO in most countries, based on recommendations from ICNIRP [
10
] or
IEEE [
11
] guidelines. These limits, which have considerable safety factors included, are set so that
Int. J. Environ. Res. Public Health 2019,16, 3406 3 of 23
exposure will not cause thermal damage to the biological material (thermal e↵ects). Thus, for 10 GHz
to 300 GHz, 10 W/m
2
is recommended as the basic restriction (no thermal e↵ects), with reference values
for 400 MHz to 2 GHz (2–10 W/m
2
) and >2 GHz (10 W/m
2
). It should be pointed out that the present
ICNIRP guidelines [
10
] are currently being revised, and new versions are to be expected in the near
future. In addition, ICNIRP proposes two categories of recommendations: (1) the basic restriction
values based on proven biological e↵ects from the exposure and (2) the reference levels given for the
purpose of comparison with physical value measurements. ICNIRP guidelines present no reference
values above 10 GHz, only considering the basic restriction values. This is due to that only surface
heating occurs since the penetration depth is so small at these frequencies. Therefore any calculations
of the Specific Absorption Rate (SAR) values, that take larger volumes into consideration, are not
reasonable to perform.
The SAR is the measure of the absorption of electromagnetic fields in a material and is expressed
as power per mass/volume (W/kg), where the penetration depth of the electromagnetic fields depends
on the wavelength of the radiation and the type of matter. The penetration depth of MMW is very
shallow, hence the exposed surface area and not the volume is considered. The appropriate exposure
metric for MMW is therefore the power density, power per area (W/m2).
It is of course too early to forecast the actual exposures to 5G networks. However, the antennas
planned for 5G will have narrow antenna beams with direct alignment [
12
] to the receiving device.
This could possibly significantly reduce environmental exposure compared to the present exposure
situation. However, it is also argued that the addition of a very high number of 5G network components
will increase the total EMF exposure in the environment, and that higher exposures to the higher
frequencies can lead to adverse health e↵ects.
Therefore, the question arises, what do we know so far about the e↵ects on biological structures
and on health due to exposure to the higher frequency bands (in this review we consider 6–100 GHz,
since lower frequencies have been extensively investigated due to their use in already existing wireless
communication networks)? Do so-called “non-thermal” e↵ects (e↵ects that occur below the thermal
e↵ect threshold) occur, that can lead to health e↵ects? Is there relevant health-oriented research
using the 5G technology relevant frequencies? Is there relevant research that can make a significant
contribution to improving the risk assessment of exposure to the general population? Answers to these
questions are necessary for a rapid and safe implementation of a technology with great potential.
2. Materials and Methods
This review takes into account scientific studies that used frequencies from 6 GHz to 100 GHz as the
source of exposure. The review is based on available data in the field of public literature, papers written
in English until the end of 2018 (PubMed database: www.ncbi.nlm.nih.gov/pubmed), EMF-Portal
(www.emf-portal.org), and other relevant literature such as documents from ICNIRP, SCENIHR, WHO,
IARC, IEEE, etc.). In addition, more refined research was conducted when necessary from sources that
were not included in the above-mentioned databases (relevant abstracts from conferences, abstract
books, and archives of journals). The resulting studies were examined for technical and scientific data
and presented in the supplementary Table S1.
As a pragmatic approach, we interpreted the results as a “response” when the authors themselves
reported the result as an “e↵ect/response” based on a statistical analysis and the p-value <0.05.
Next we defined necessary criteria for study quality, both from a biomedical and physical point of
view (see [
13
]). The results of the studies were (if possible) analysed for correlations with study quality
according to the correlation approach done by Simk
ó
et al. [
14
]. The studies were analysed with reference
to a minimum of criteria in terms of experimental design and implementation. The following criteria
were considered: were the experiments performed in the presence of an appropriate sham/exposure
control, temperature control, positive control, were the samples blinded, and was a comprehensive
dosimetry presented.
Int. J. Environ. Res. Public Health 2019,16, 3406 4 of 23
The study is divided into a descriptive part, which covers the description of all selected studies,
their exposure conditions, frequency ranges (6 GHz to 100 GHz), dose levels, etc., as well as the
biological results, presented in a Master-Table (Table S1). Review articles were not considered.
The outcomes of the studies were furthermore analyzed and discussed according to frequency domains,
and power density and exposure duration. If appropriate, we include an evidence-based interpretative
part regarding risk from exposures according to the criteria of SCHEER [15].
3. Results
In the following, health-related published scientific papers dealing with frequencies from 6 GHz
to 100 GHz (using the term MMW for all the frequencies) are described in detail. It should be noted
that there are no epidemiological studies dealing with wireless communication for this frequency
range, thus, this review will cover studies performed in vivo and in vitro.
Thermal biological e↵ects of radiofrequency electromagnetic fields occur when the SAR values
exceed a certain limit, namely 4 W/kg (general population exposure limit: SAR 0.08 W/kg), which
causes a tissue heating of 1
C. However, in the literature, biological e↵ects below 4 W/kg SAR
values have been described. Since such e↵ects are considered to be not due to warming, they are
termed non-thermal e↵ects. In the present review, in some individual studies, the authors interpreted
thermal e↵ects as “no e↵ect”. Those ones and studies without response/e↵ect of MMW exposure were
considered as “no response/e↵ect” in our present analysis.
3.1. Grouping of Selected Parameters
For analysis, 94 publications were identified and selected from the accessible databases (
in vivo
and
in vitro
)[
16
–
109
]. It should be noted that the total number of individual examinations is larger
than the number of publications, since some authors investigated several physical and/or biological
conditions in the same publication.
Various biological endpoints have been identified, which are referred to as “response” or e↵ects
when appropriate. Since the list of these endpoints is relatively long, we have not mentioned them
in detail, but summarized them in groups: Physiological, neurological, histological changes, or in
in vitro
studies gene or protein expression, cytotoxic e↵ects, genotoxic changes, and also temperature-
related reactions.
For a detailed analysis, a “Master-table” (Table S1) was prepared in which all parameters considered
in the studies were included. The table contains the following information: frequency,
in vivo
or
in vitro
study (the latter distinguishes between primary cells and cell lines), power density, exposure
duration, biological endpoints, and response. Some studies lack information on individual parameters.
For example, a publication had to be excluded completely because there was no information about the
frequency. In nine studies the power density data were absent and in seven studies the calculated SAR
values were provided instead of the power density. In ten studies, the exposure time was not given.
The 45
in vivo
studies were mainly conducted on mammals (mouse, rat, rabbit) and a few on
humans. In some studies, bacteria, fungi, and other living material were also used for the experiments.
80% of all in vivo studies showed exposure-related reactions.
Primary cells (n =24) or cell lines (n =29) were used in the 53
in vitro
studies, with approximately
70% of the primary cell studies and 40% of the cell line investigations showing exposure-related
responses (Table 2).
All identified studies were analyzed as a function of frequency. For this purpose, frequency
domains (groups) have been created (Figure 1) to analyze and illustrate the results. The frequency
groups from 30 to 60 GHz were grouped in ten-GHz increments (up to 30, 30.1–40, 40.1–50, 50.1–60 GHz).
The frequency range 60–65 GHz was extra analyzed as in this group a larger number of publications
was identified (in comparison to the other groups). Due to the low number of publications above
65.0 GHz, data was merged into the groups of “65.1–90” and “above 90 GHz”. As shown in Figure 1,
the majority of studies show a frequency-independent response after MMW exposure.
Int. J. Environ. Res. Public Health 2019,16, 3406 5 of 23
Table 2. Overview of the total number of publications examinations.
All Publications (94) No Response Response All
In vivo 10 35 45
In vitro 22 31
53
Primary cells 6 18
Cell lines 16 13
Int. J. Environ. Res. Public Health 2019, 16, x FOR PEER REVIEW 5 of 23
Table 2. Overview of the total number of publications examinations.
All Publications (94) No Response Response All
In vivo 10 35 45
In vitro 22 31
53 Primary cells 6 18
Cell lines 16 13
All identified studies were analyzed as a function of frequency. For this purpose, frequency
domains (groups) have been created (Figure 1) to analyze and illustrate the results. The frequency
groups from 30 to 60 GHz were grouped in ten-GHz increments (up to 30, 30.1–40, 40.1–50, 50.1–60
GHz). The frequency range 60–65 GHz was extra analyzed as in this group a larger number of
publications was identified (in comparison to the other groups). Due to the low number of
publications above 65.0 GHz, data was merged into the groups of “65.1–90” and “above 90 GHz”. As
shown in Figure 1, the majority of studies show a frequency-independent response after MMW
exposure.
Figure 1. The number of publications as a function of frequency domains. The black line represents
the total number of publications, and bars represent the in vivo (dark blue) and in vitro (light blue)
studies with biological responses.
3.1.1. Frequency Ranges
All data regarding the individual papers are found in Table S1.
Up to 30 GHz
The first group “up to 30 GHz” was introduced since some of the 5G frequencies fall within this
frequency range. Unfortunately, there are only two publications in this group, both showing
responses to the MMW exposure. A study that was conducted on bacteria and fungi showed an
increase in cell growth [58]. The other in vitro study was performed on fibroblasts (25 GHz, 0.80
mW/cm2, 20 min), with genotoxic effects observed at high SAR levels (20 W/kg) [24]. A graphical
presentation of the outcomes is presented in Figure 1 for this and all other frequency domains.
Frequency Group 30.1–40 GHz
As shown in Figure 1, responses were detected in approximately 95% of the 19 studies. In all in
vivo studies responses were described after exposure [25,27,36,37,55,56,78,79,87,91,103,104].
Endpoints ranged from recorded footpad edema, which is a frequent endpoint for the measurement
of inflammatory responses, to morphological changes, changes in skin temperature, blood pressure,
heart rate, body temperature, neuronal electrical activity, and EEG analyses. Protein expression
Figure 1.
The number of publications as a function of frequency domains. The black line represents the
total number of publications, and bars represent the
in vivo
(dark blue) and
in vitro
(light blue) studies
with biological responses.
3.1.1. Frequency Ranges
All data regarding the individual papers are found in Table S1.
Up to 30 GHz
The first group “up to 30 GHz” was introduced since some of the 5G frequencies fall within this
frequency range. Unfortunately, there are only two publications in this group, both showing responses
to the MMW exposure. A study that was conducted on bacteria and fungi showed an increase in cell
growth [
58
]. The other
in vitro
study was performed on fibroblasts (25 GHz, 0.80 mW/cm
2
, 20 min),
with genotoxic e↵ects observed at high SAR levels (20 W/kg) [
24
]. A graphical presentation of the
outcomes is presented in Figure 1for this and all other frequency domains.
Frequency Group 30.1–40 GHz
As shown in Figure 1, responses were detected in approximately 95% of the 19 studies.
In all
in vivo
studies responses were described after exposure [
25
,
27
,
36
,
37
,
55
,
56
,
78
,
79
,
87
,
91
,
103
,
104
].
Endpoints ranged from recorded footpad edema, which is a frequent endpoint for the measurement
of inflammatory responses, to morphological changes, changes in skin temperature, blood pressure,
heart rate, body temperature, neuronal electrical activity, and EEG analyses. Protein expression
studies, oxidative stress marker measurements, histological investigations, and induction of cell death
(apoptosis) were performed. Only one study used lower power densities (0.01 mW/cm
2
, 0.1 mW/cm
2
;
SAR: 0.15, 1.5 W/kg; 20 min, 40 min) to study inflammatory responses [
27
]. The authors determined the
frequency-dependent anti-inflammatory e↵ect as a function of power density and exposure duration
and did not rule out temperature-related e↵ects. The power densities of the other
in vivo
studies were
extremely high (10, 75, 500–5000 mW/cm
2
), so the induced e↵ects were likely temperature dependent.
Eight
in vitro
studies were performed [
18
,
20
,
47
,
91
,
97
,
99
,
101
,
102
] of which seven reported
responses. In one study [
99
], human blood cells (ex vivo) were exposed to MMW for 5, 15 and
Int. J. Environ. Res. Public Health 2019,16, 3406 6 of 23
30 min (32.9–39.6 GHz, 10 mW/cm
2
). The activation of the cells was examined in the presence or
absence of bacteria. It was shown that in the presence of bacterial activation and after 15 min of
exposure, the cells were activated to release free radicals. These results were similar to the heated
samples (positive controls), so a temperature e↵ect is plausible. The induction of di↵erentiation of
bone marrow cells in to neuronal phenotype cells was also demonstrated (36.11 GHz, 10 mW/cm
2
,
3
⇥
10 min every 2 h for 24 h) [
97
]. In two studies, temperature-related reactions were described at
the protein level [
18
,
91
]. When the cell cultures were cooled during exposure to prevent the induced
temperature increase, no responses were detected.
In three publications, a research group described cell cycle changes, induction of cell death and
activation of di↵erentiation processes in primary cells (rat bone cells and mesenchymal stem cells)
after exposure to 30–40 GHz (4 mW/cm
2
, di↵erent exposure durations) [
47
,
101
,
102
]. Unfortunately,
the minimum quality criteria were not fulfilled in any of the three studies, mainly because there were
no temperature controls.
Frequency Group 40.1–50 GHz
In the 40.1–50 GHz frequency group, 26 studies were identified, 13
in vivo
[
16
,
17
,
26
,
48
,
49
,
51
,
53
,
65
,
69
,
74
,
80
,
84
,
98
] and 13
in vitro
[
29
–
31
,
62
,
64
,
86
,
89
,
92
,
93
,
100
,
105
,
107
] with nine studies showing
responses. A large number of studies have tested cell biology endpoints such as cell proliferation,
gene or protein expression, and changes in oxidative stress. In addition, immunological, neurological,
morphological and genotoxic e↵ects were investigated. The power densities used vary enormously,
from 0.02 to 450 mW/cm2, and one publication gave no information.
In healthy volunteers, a double-blind study was performed to investigate the e↵ects of MMW on
experimentally induced cold pain (42.25 GHz, <17.2 mW/cm
2
, 30 min) [
74
]. The authors found no
di↵erence from the placebo e↵ect. This study was a repeat of a previous study with volunteers and the
results of the older study could not be confirmed. The other four
in vivo
studies with no detectable
e↵ects were investigating genotoxic e↵ects or oxidative stress [17,48,49,98].
Five
in vivo
publications addressed the e↵ects of MMW on the immune system of mice or rats,
finding activation of the immune system at both the cellular and molecular levels (41.95 or 42.2 GHz,
19.5 µW/cm2, 0, 1, 31.5 mW/cm2, 20 min or intermittently over 3 days) [26,48,51,53,84].
MMW exposure of frog isolated nerve cells, (41.34 GHz, 0.02, 0.1, 0.5, 2.6 mW/cm
2
, 10–23 min)
lead to a reduction of the action potential frequency. Interestingly, the e↵ects at higher power density
(2.6 mW/cm2) were similar to conventional heating [49].
One study detected an increase in the motility of human spermatozoa after 15 min of exposure
(42.25 GHz, 0.03 mW/cm
2
)[
100
]. Additional
in vitro
tests have identified the formation of free radicals,
the activation of calcium-dependent potassium ion channels (around 42 GHz, 100, 150, 240
µ
W/cm
2
,
20–40 min) as well as changes at the cell membrane in exposed cells [29,30,100].
No responses on cell biological endpoints (cell cycle changes, cell death, heat shock proteins) were
detected in four additional in vitro studies.
Frequency Group 50.1–60 GHz
We identified 16 studies in the frequency group 50.1-60 GHz (six
in vivo
, ten
in vitro
) and 60% of
the studies showed responses to MMW exposures [21,23,35,38,43,46,59,61,72,77,81,83,85,94,109].
In five of the
in vivo
studies very di↵erent responses were shown. In a study on healthy volunteers,
the authors wanted to find out whether the human skin at a so-called acupuncture point has di↵erent
dielectric properties during exposure to MMW. They found that these properties change during
exposure to 50–61 GHz from the surrounding skin [23].
A pilot study on mice (60 GHz, 0.5 mW/cm
2
, lifelong exposure for 30 min/5 days a week) showed
that MMW exposure a↵ects cancer-induced cells and increases in motor activity of healthy mice [61].
In rats, the influence of 54 GHz, 150 mW/cm
2
, on an area of approximately 2 cm
2
on the head
was examined [
81
]. This transcranial electromagnetic brain stimulation induced pain prevention and
Int. J. Environ. Res. Public Health 2019,16, 3406 7 of 23
prevented the conditioned avoidance response to a pain stimulus in 50% of the animals. However,
no changes were detected when serotonin inhibitors were previously administered. Therefore,
the authors concluded that transcranial electromagnetic brain stimulation promotes the synthesis of
serotonin, a transmitter that changes the animals’ pain threshold.
The e↵ects of MMW were also tested (60 GHz, 475 mW/cm
2
, 1.898 mW/cm
2
, 6, 30 min) on rabbit
eyes, describing acute thermal injuries of various types [
38
]. The authors also pointed out that the
higher temperature just below the eye surface could induce injury.
Neurological investigations were performed on leeches (60 GHz, 1 min, 1, 2, 4 mW/cm
2
)[
77
]
and electrophysiological studies were performed on frog oocytes (60 GHz, up to 5 min) [
85
]. In both
experimental systems e↵ects were described, which were induced by the temperature rise.
Cell biological and morphological changes after exposure to 0.7–1.0
µ
W/cm
2
(intermittent) were
reported in three
in vitro
studies [
72
,
83
,
94
], with two publications providing no information regarding
power density or exposure duration. At the level of protein analysis and total genome analysis no
changes were identified in four in vitro studies [35,46,59,109].
Frequency Group 60.1–65 GHz
The number of studies in the 60.1–65 GHz frequency group is 27. Of these, twelve reported e↵ects
from exposure to MMW, and no responses were found in 15 studies.
The
in vivo
studies investigated di↵erent topics [
23
,
27
,
44
,
52
,
67
,
68
,
70
,
71
,
73
,
75
,
76
]. Thus, two
studies examined the e↵ects on tumor development in mice injected with tumor cells [
52
,
70
]. In one
of the studies it was reported that exposure to 61.22 GHz, 13.3 mW/cm
2
, inhibited the growth of
melanoma cells (exposure 15 days after tumor cell injection, 15 min/day) [70].
Other publications from one research group investigated the potential of MMW for pain relief and
the associated biological mechanisms of action [
67
,
71
,
73
,
75
,
76
]. Several of the studies were performed
on mice skin exposed to 61.22 GHz for 15 min. The most commonly used power density was 15 mW/cm
2
.
Another study addressed the dose issue with no effect below 1.5 mW/cm
2
.Theauthors’conclusionis
that MMW can lower the hypoalgesia threshold, which is likely mediated by the release of opioids.
The e↵ects of 61.22 GHz exposure of mice were examined also with respect to the immune
system [
52
]. The animals were exposed on three consecutive days for 30 min per day. The exposure
caused peak SAR values of 885 W/kg on the nose of the animals where the exposure took place.
The power density was 31 mW/cm
2
and the measured temperature rise reached 1
C. It was found that
MMW modulates the e↵ects of the cancer drug cyclophosamide. In particular, the T-cell system of the
immune system was activated and various other immune system relevant parameters a↵ected.
The similar exposure condition was used in a study on gastrointestinal function, however no
e↵ects were identified [68].
A single exposure for eight hours (61 GHz, 10 mW/cm
2
), or five times four hours, did not
cause eye damage to rabbits and rhesus monkeys [
44
]. It should be emphasized that several of
the mentioned studies come from the same laboratory, and all criteria for the study quality are
met. However, the authors were able to replicate their own findings on pain relief whereas other
laboratories have not replicated this work. In the
in vitro
studies, various biological endpoints were
examined [28,32–34,42,45,50,59,60,66,83,88,94,95,108].
In one study, neurons of snails (Lymnea) were exposed at 60.22–62.22 GHz and no non-thermal
responses on the ion currents were identified [28].
In a series of investigations with nerve cell-relevant cell lines, the dopamine transmission properties,
stress, pain and membrane protein expression were investigated (60.4 GHz, 10 mW/cm
2
, 24 h) and no
responses were detected [32–34,59,60,108].
The same exposure setup has also been used in studies examining di↵erent stress response related
genes (0.14–20 mW/cm
2
)[
59
]. No e↵ects were found at the gene expression level. Interestingly,
the overall genome impact was influenced when the exposure (60.4 GHz, 20 mW/cm
2
, 3 h) of the
primary human keratinocytes was combined with 2-deoxyglucose, a glucose-6- phosphatase inhibitor.
Int. J. Environ. Res. Public Health 2019,16, 3406 8 of 23
This co-exposure caused a change in the amount of six di↵erent transcription factors, the e↵ect di↵ering
from the e↵ect of 2-deoxyglucose alone and 60.4 GHz alone (both factors alone induced no changes).
Other studies also examined human keratinocytes and astrocytoma glial cells after exposure to
60 GHz (0.54, 1 and 5.4 mW/cm
2
)[
60
,
108
]. Various parameters such as cell survival, intracellular
protein homeostasis, and stress-sensitive gene expression were investigated. Also, in these studies,
no e↵ects were observed. In contrast, in one publication, the elevation of an inflammatory marker
(IL1-
) was observed in human keratinocytes after exposure (61.2 GHz, 29 mW/cm
2
, 15, 30 min), while
other inflammatory markers (chemotaxis, adhesion and proliferation) have remained unchanged [
95
].
Another type of study was performed on rat brain cortical slices [
66
]. The brain slices were
exposed to a field of 60.125 GHz (1
µ
W/cm
2
) for 1 min, and then specific electrophysiological parameters
were measured. In many slices, transient responses on membrane characteristics and action potential
amplitude and duration were observed. The exposure caused a temperature rise of the medium
(of 3
C) in which the sections were stored. Interestingly, a chronically induced Ca
2+
blockade did not
a↵ect the MMW response.
Frequency Group 65.1–90 GHz
The studies in the frequency group of 65.1 to 90 GHz were performed both
in vivo
and
in vitro
in
a total of 14 articles (four
in vivo
and 11
in vitro
investigations). The studies vary widely, based on
di↵erent hypotheses, biological endpoints, power densities, and exposure durations. In addition, some
studies have used biological materials to identify physical properties such as dielectric properties and
skin reflection coefficient. The latter studies are discussed in Section 4.2.
Four
in vivo
studies reported responses after MMW exposure. One study examined the dose
of eye damage (especially damage to the corneal epithelium) [
40
]. The dose was calculated as DD
50
(based on the results for which the probability of eye damage was 50%). The experiments were carried
out on rats with an exposure of 75 GHz, the DD50 value being 143 mW/cm2.
Other
in vivo
studies were performed on rats and mice as well as on insects [
27
,
42
,
57
]. The study
on mice used di↵erent frequencies of 37.5 to 70 GHz, with power densities of 0.01 and 0.3 mW/cm
2
for
20 to 40 min. A single whole-body exposure of the animals reduced both the footpad edema and local
hyperthermia on average by 20% at the frequencies of 42.2, 51.8, and 65 GHz. Other frequencies had
no influence.
The study on insects (Chironomidae) focused on DNA e↵ects of giant chromosomes of the salivary
glands of the animals with di↵erent frequencies (64.1–69.1, 67.2, 68.2 GHz) [
42
]. All frequencies, using
power densities <6 mW/cm
2
, caused a reduction in the size of a particular area of the chromosome.
This in turn led to the expression of certain secretory proteins of the salivary gland.
Di↵erent aspects were studied in the
in vitro
studies [
18
,
28
,
39
,
50
,
64
,
72
,
83
,
89
,
94
,
106
], where nerve
cell function was investigated in three studies. Two studies used nerve cells from the snail Lymnea that
were exposed at 75 GHz for a few minutes at very high SAR levels (up to 4200 W/kg, power density
was not reported) [
28
,
39
]. The authors observed thermal e↵ects on the ion currents and the firing rate
of the action potentials. Another study also described thermal e↵ects on transmembrane currents and
ionic conductivity of the cell membrane. Again, the exposure was at very high SAR levels (2000 W/kg),
and the authors emphasized the temperature dependence of the reaction.
Broadband frequencies (52–78 GHz) have been used in several publications, mainly investigating
the e↵ects on cell growth and cell morphology as well as the ultrastructure of di↵erent cell
lines
[50,72,83,94]
. The values for the power densities were not given consistently but appear to
have been very low (not higher than 1
µ
W/cm
2
). The results indicated the inhibition of cell growth,
accompanied by changes in cell morphology.
Another group of studies used hamster fibroblasts, BHK cells, and exposed the cells at 65 to
75 GHz, with the power density reaching 450 mW/cm
2
[
18
,
64
,
89
]. The authors noted the inhibition of
protein synthesis and cell proliferation as well as cell death at higher power densities. In a study using
human dermal fibroblasts and human glioblastoma cells, no e↵ects at the protein level (proliferation
Int. J. Environ. Res. Public Health 2019,16, 3406 9 of 23
or cytotoxicity markers) were detected (70 GHz and higher, in 1 GHz increments; 3, 70 or 94 h) [
106
].
Power densities varied across frequencies, ranging from 1.27
µ
W/cm
2
in the lower frequency range to
0.38 µW/cm2at higher frequencies.
The
in vitro
studies in this group are similar to the
in vivo
studies in their diversity. The majority
of studies in which responses were reported are thermal-e↵ects due to MMW exposure. In three
studies, responses at low power densities were described, but all results were from the same laboratory,
and were not replicated by others. Moreover, the quality of these studies is questionable, as the quality
criteria were not met.
Frequency Group 90.1–100 GHz
Eight out of eleven studies in the 90.1–100 GHz frequency group are
in vitro
studies [
22
,
41
,
57
,
82
,
90
,
96
,
106
]. The three
in vivo
investigations addressed a variety of issues including acute e↵ects on
muscle contraction, skin-reflection properties (which are more of a dose-related than health-related
issue), and skin cancer [
19
,
54
,
57
]. The rat skin cancer study (one to two weekly, short-term exposures
at 94 GHz, 1 W/kg; DMBA-initiated animals) did not show any positive outcome [
54
]. Another study
examined the muscle contraction of mice and described some responses [
19
]. Again, 94 GHz was used,
but power density or SAR values were not reported.
Seven of the eight
in vitro
studies showed responses after MMW exposure. In some studies,
primary neurons were used to study the cytoskeleton (94 GHz, 31 mW/cm
2
)[
82
] or specific
electrophysiological parameters (90–160 GHz) [
22
]. In the latter study it was found that the observed
responses were more likely due to interactions with the cell culture medium than with the cells,
although the mechanisms of action were not clear. Other studies identified responses on the DNA
integrity (100 GHz and higher) [
41
] or described changes in intracellular signaling pathways (94 GHz,
90–160 GHz) using di↵erent cell types [
57
,
96
]. The exposure time ranged from minutes to 24 h for
partially unknown exposure values. In one study no cytotoxic influence at power density levels of
a few µW/cm2was detected in either normal or in tumor cells.
3.1.2. Power Densities
All identified studies were analyzed as a function of the used power densities. The studies
were grouped depending on the power density as follows: below 1; 1.1–10; 10.1 to 50; 50.1–100, and
100.1 mW/cm
2
or higher. Studies that do not provide information on power density or SAR values
are not displayed in these groups. As shown in Figure 2, the vast majority of studies show responses
regardless of the power density used.
Int. J. Environ. Res. Public Health 2019, 16, x FOR PEER REVIEW 10 of 23
Figure 2. The number of publications as a function of power density. The black line represent the
total number of publications, and bars represent the in vivo (dark blue) and in vitro (light blue)
studies with biological responses.
3.1.2. Power Densities
All identified studies were analyzed as a function of the used power densities. The studies were
grouped depending on the power density as follows: below 1; 1.1–10; 10.1 to 50; 50.1–100, and 100.1
mW/cm2 or higher. Studies that do not provide information on power density or SAR values are not
displayed in these groups. As shown in Figure 2, the vast majority of studies show responses
regardless of the power density used
3.1.3. Exposure Duration
Exposure duration of the studies was also grouped for data analysis (Figure 3). The time groups
were selected as seconds to 10 min; 10–30 min; 30–60 min; over 60 min-days and
alternately/intermittently. The groups were selected so that the used exposure times and the number
of studies are meaningfully summarized. Here, too, it becomes clear that the majority of all studies
show responses regardless of the exposure time. Interestingly, longer exposure times (over 60
min—days) seemingly lead to fewer reactions than in the other groups.
Figure 3. The number of publications as a function of exposure duration. The black line represent the
total number of publications, and bars represent the in vivo (dark blue) and in vitro (light blue)
studies with biological responses.
Figure 2.
The number of publications as a function of power density. The black line represent the total
number of publications, and bars represent the
in vivo
(dark blue) and
in vitro
(light blue) studies with
biological responses.
Int. J. Environ. Res. Public Health 2019,16, 3406 10 of 23
3.1.3. Exposure Duration
Exposure duration of the studies was also grouped for data analysis (Figure 3). The time
groups were selected as seconds to 10 min; 10–30 min; 30–60 min; over 60 min-days and alternately/
intermittently. The groups were selected so that the used exposure times and the number of studies are
meaningfully summarized. Here, too, it becomes clear that the majority of all studies show responses
regardless of the exposure time. Interestingly, longer exposure times (over 60 min—days) seemingly
lead to fewer reactions than in the other groups.
Int. J. Environ. Res. Public Health 2019, 16, x FOR PEER REVIEW 10 of 23
Figure 2. The number of publications as a function of power density. The black line represent the
total number of publications, and bars represent the in vivo (dark blue) and in vitro (light blue)
studies with biological responses.
3.1.2. Power Densities
All identified studies were analyzed as a function of the used power densities. The studies were
grouped depending on the power density as follows: below 1; 1.1–10; 10.1 to 50; 50.1–100, and 100.1
mW/cm2 or higher. Studies that do not provide information on power density or SAR values are not
displayed in these groups. As shown in Figure 2, the vast majority of studies show responses
regardless of the power density used
3.1.3. Exposure Duration
Exposure duration of the studies was also grouped for data analysis (Figure 3). The time groups
were selected as seconds to 10 min; 10–30 min; 30–60 min; over 60 min-days and
alternately/intermittently. The groups were selected so that the used exposure times and the number
of studies are meaningfully summarized. Here, too, it becomes clear that the majority of all studies
show responses regardless of the exposure time. Interestingly, longer exposure times (over 60
min—days) seemingly lead to fewer reactions than in the other groups.
Figure 3. The number of publications as a function of exposure duration. The black line represent the
total number of publications, and bars represent the in vivo (dark blue) and in vitro (light blue)
studies with biological responses.
Figure 3.
The number of publications as a function of exposure duration. The black line represent the
total number of publications, and bars represent the
in vivo
(dark blue) and
in vitro
(light blue) studies
with biological responses.
3.2. Studies without Responses
Table 3shows the number of studies in which no responses were detected after or during MMW
exposure. As “no response” also such investigations were referred to, which were considered by the
authors themselves as such. This means that in some cases the observed e↵ects were described as
temperature-related and not as a non-thermal MMW e↵ect.
Table 3. Studies without responses.
Frequency (GHz) No Response
In Vivo In Vitro
Up to 30 0 0
0.1–40 0 2
40.1–50 6 4
50.1–60 1 5
60.1–65 2 10
65.1–90 0 6
90.1–100 1 1
Few
in vivo
studies have shown no response at all. Noticeable is the frequency group 40.1–50 GHz,
in which 6 studies were identified. These studies investigated immunosuppression, genotoxic e↵ects,
changes in pain sensitivity, and changes in enzyme activity. One study was carried out on bacteria
and fungi.
There are a variety of
in vitro
studies in which no responses were detected. Interestingly, studies
on protein or gene expression levels often failed to detect any changes after MMW exposure. This
could be due to the fact that in
in vitro
studies the possibility of non-thermal e↵ects were specifically
investigated, where cooling was used to counteract the temperature increase.
Int. J. Environ. Res. Public Health 2019,16, 3406 11 of 23
3.3. Quality Analysis
We analyzed the quality of the selected studies according to specific criteria [
14
]. The studies were
categorized by the presence of sham/control, dosimetry, positive control, temperature control, and
whether the study was blinded. The presence of these five criteria while performing an MMW study is
the minimum requirement for qualifying as a study with sufficient technical quality.
Of the 45
in vivo
studies, 78% (35) demonstrated biological responses after exposure to MMW.
Of all studies, 73% were performed with sham/controls, 76% employed appropriate dosimetry, 44% used
positive control, and 67% were done under temperature control conditions (Figure 4). Unfortunately,
only 16% of the studies were performed according to protocols that ensured blinding and only three
publications were identified that met all five criteria [
26
,
51
,
53
]. If the blinding criterion was excluded,
13 studies could be identified that met the remaining four criteria. Considering three criteria only,
namely sham, dosimetry, and temperature control, 40% (20 papers) were identified. Thus, the quality
of the in vivo studies is unsatisfactory.
Int. J. Environ. Res. Public Health 2019, 16, x FOR PEER REVIEW 12 of 23
Figure 4. The quality of all publications: The number of in vivo (top) and in vitro (bottom)
experiments (blue: no reaction, red: reaction) using the listed quality features (y-axis). The spider
web shows the percentage of the quality characteristics in all examinations.
These results show that the number of examinations and the quality criteria are insufficient for
a statistical analysis. It should be stressed that this quality analysis covers all publications dealing
with the responses/effects of exposure to 6 to 100 GHz MMW, irrespective of the endpoints tested.
To perform a correlation analysis, a larger number of comparable studies (e.g., identical endpoints in
a frequency group) would be required.
4. Discussion
The first relevant observation during the analysis of the studies is that in most publications the
aim of the investigations has been to determine the effects of MMW exposure for medical purposes.
This means that the exposure devices used primarily come from medical applications (therapy or
diagnostics). Very few publications dealt with health-related issues after MMW exposure in general,
or with the specific topic of 5G. Therefore, the 94 publications are very heterogeneous.
We divided the frequency bands into seven ranges and placed the studies in the relevant
groups. All available information on physical and experimental parameters was collected, but the
exact number of experiments in each study was not taken into account. (One publication can contain
more than one experiment.) Therefore, it is the provided numbers of studies/publications that
constitute the data set, not the exact numbers of experiments performed, which is significantly higher.
This report does not provide a statistical analysis of the correlation between the exposure
conditions and the results, which was our original ambition. In the correlation study according to
Simkó et al. [14] a frequency group was selected, with only one group of biological endpoints
considered. About one hundred, exclusively in vitro, studies were identified and broken down into
individual experiments in that paper. In this way, the number of experiments was sufficient to
perform a correlation analysis. In the present review, the spread of biological endpoints in the
individual frequency groups and the models used (in vivo and in vitro) is large and the number of
studies is very low. Therefore, it was not possible to group the studies by specific endpoints and
perform a statistical analysis.
Figure 4.
The quality of all publications: The number of
in vivo
(top) and
in vitro
(bottom) experiments
(blue: no reaction, red: reaction) using the listed quality features (y-axis). The spider web shows the
percentage of the quality characteristics in all examinations.
Out of the 53
in vitro
studies, 31 showed biological responses. Only in 13 studies (42%) were three
of the five quality criteria satisfied, namely the presence of sham/control, dosimetry, and temperature
control (Figure 4). Positive controls were used in 47% and only one study was performed with blinded
protocol (2%).
These results show that the number of examinations and the quality criteria are insufficient for
a statistical analysis. It should be stressed that this quality analysis covers all publications dealing
with the responses/e↵ects of exposure to 6 to 100 GHz MMW, irrespective of the endpoints tested.
To perform a correlation analysis, a larger number of comparable studies (e.g., identical endpoints in
a frequency group) would be required.
Int. J. Environ. Res. Public Health 2019,16, 3406 12 of 23
4. Discussion
The first relevant observation during the analysis of the studies is that in most publications the
aim of the investigations has been to determine the e↵ects of MMW exposure for medical purposes.
This means that the exposure devices used primarily come from medical applications (therapy or
diagnostics). Very few publications dealt with health-related issues after MMW exposure in general,
or with the specific topic of 5G. Therefore, the 94 publications are very heterogeneous.
We divided the frequency bands into seven ranges and placed the studies in the relevant groups.
All available information on physical and experimental parameters was collected, but the exact number
of experiments in each study was not taken into account. (One publication can contain more than one
experiment.) Therefore, it is the provided numbers of studies/publications that constitute the data set,
not the exact numbers of experiments performed, which is significantly higher.
This report does not provide a statistical analysis of the correlation between the exposure conditions
and the results, which was our original ambition. In the correlation study according to Simk
ó
et al. [
14
]
a frequency group was selected, with only one group of biological endpoints considered. About one
hundred, exclusively
in vitro
, studies were identified and broken down into individual experiments in
that paper. In this way, the number of experiments was sufficient to perform a correlation analysis.
In the present review, the spread of biological endpoints in the individual frequency groups and the
models used (
in vivo
and
in vitro
) is large and the number of studies is very low. Therefore, it was not
possible to group the studies by specific endpoints and perform a statistical analysis.
Interestingly, more than half of the studies (53 publications) were conducted in the frequency
bands 40.1–50 and 60.1–65 GHz (with di↵erent models and endpoints). One possible reason for this
is that medical use of MMW has a long tradition in Eastern Europe. These applications use specific
frequencies that fall in these two frequency groups. The studies were conducted with the aim of
testing specific e↵ects with medical relevance. In these two frequency groups, the “with responses”
percentage was generally lower than in the other frequency bands (see Figure 1), where a majority of
studies showed responses to exposure.
With regard to the power densities used, about half of the studies were carried out in the range up
to 10 mW/cm
2
(Figure 2). This value is ten times higher than the current ICNIRP exposure guideline [
10
]
for the general population. Based on available data, there is no indication that higher power densities
cause more frequent responses, since the percentage of responses in all groups is already at 70%
(Figure 2). One exception from this high response rate is the group 50.1–100 mW/cm
2
, where the
proportion of studies with reactions is slightly lower (54%). However, the total number of examinations
(11) is relatively small in this group.
The results of some of the studies may suggest that exposure to power densities at or below the
guideline recommendations induce biological e↵ects. There are, however, some arguments against it.
One of these is the apparent heterogeneity of the study design and the outcomes studied. There are very
few (if any) independent replication studies that confirm the reported results. It is also noteworthy that
there is no trend towards a classic dose-response pattern where stronger or more frequent e↵ects would
be caused by higher exposure levels. Since the studies with conditions promoting tissue warming
show no greater e↵ect than below the guideline values (1 mW/cm
2
), this would either mean that the
same interactions are present at all power densities tested, or that experimental artifacts unknown to
the scientists are present.
The most important physical experimental parameter is the temperature during exposure, therefore,
the temperature must be consistently controlled. The need for stringent temperature control is not an
insignificant or trivial matter and has been neglected or at least undervalued in many studies. Although
some authors report that they performed specific temperature measurements during the experiments,
this does not necessarily mean that this represents the actual temperature in the biological material.
Measurements can be made, for example, in the surrounding medium but not in the exposed tissue or
in the cell. It also has to be considered that the “bulk” heating (from outside to inside with a certain
time course) can di↵er from a heating that occurs at a rather limited point (“hot spot”). In addition,
Int. J. Environ. Res. Public Health 2019,16, 3406 13 of 23
the intensity of a short burst can be lost if the measurements are based on average exposure times.
Such errors and problems are possible factors that have contributed to the questionable interpretation
of “non-thermal e↵ects” in some studies.
E↵ects after MMW exposure were shown at all exposure times with no clear time dependency.
The data presented shows one exception, namely in the group “>60 min to days”, where fewer reactions
were detected (Figure 3). It has to be taken into account that 27 examinations were carried out in
this group, 23 of which were
in vitro
studies.
In vitro
experiments can be carried out under cooling,
therefore the results can be di↵erent (see further below).
Two research groups together provide 30 of the 94 publications in the data set, and could thus
possibly have a large impact on the analysis of the outcomes. One group presented at least 21
publications (42.25 and 61.82 GHz; 10 to 30 mW/cm
2
; with di↵erent exposure durations), with a variety
of
in vivo
and
in vitro
studies, which mostly reported responses to exposure. The other group mainly
studied gene and protein expressions (60 GHz; 5.4 to 20 mW/cm
2
; exposure durations from minutes to
days) and found mainly no responses. Studies from both groups adhered well to the quality criteria in
our analysis.
4.1. Temperature Controls in In Vitro Studies
In vivo
studies that are performed within or directly on the living organism have shown both
thermal and purportedly non-thermal e↵ects after or during MMW exposure.
In vitro
studies are
carried out on cells and most experimental parameters can be accurately set and observed. Cell cultures
can thus be very carefully controlled, e.g., an induced temperature increase can be counter-cooled.
Many
in vitro
studies considered in this review were performed using cooling of the cell culture
vessels and the authors did not detect any non-thermal e↵ects in these studies. In
in vivo
studies
counter-cooling is not possible, thus it is very difficult to di↵erentiate between thermal and non-thermal
reactions. Therefore,
in vivo
and
in vitro
studies regarding the induced e↵ects cannot be directly
compared. An accurate dosimetry could solve this problem.
4.2. Dosimetry
It is important to know what the exposure of the MMW will be due to the expected introduction
of a large number of 5G wireless communication devices. Given the novelty of the technology, it is
currently unlikely that a large number of relevant exposure assessment studies will be available.
However, an example from a recent study [
110
] shows that a “typical” office environment with wireless
communication transmitters (5.50 GHz) leads to power densities well below the exposure guideline
limits. Thus, the maximum power density was measured at 0.89 µW/cm2.
Partly (n =25) the experimental studies on biological and health e↵ects of MMW exposure are at
or below the ICNIRP exposure guidelines. The power densities were often chosen so that the exposure
caused no or very moderate tissue warming (<1
C), namely in the range of 1 to 10 mW/cm
2
. Since the
penetration into the tissue of these frequencies are on the order of millimeters and below, it is important
to study biological e↵ects directly or indirectly related to skin and eyes exposure. As mentioned
previously, the number of available studies in the 6–100 GHz frequency range is relatively low, which
is in contrast to the number of studies for lower radio frequencies. Similarly, the number of tissue
dosimetry studies (especially for the skin) is very limited. However, such studies are very relevant
because they show how certain exposure parameters can influence the energy input and thus the
thermal behavior of the skin.
Currently, both the ICNIRP guidelines and the IEEE standards are being revised to replace the
SAR values with power density above 6 GHz. However, it has already been recognized that there is
a reactive near field close to the transmitter (around the antennas). Here, the energy is not radiated, but
the energy envelopes the antennas. The question is whether these “reactive near fields” are important
for the energy delivery to a human body near the transmitter? If this is not the case, it is sufficient to
comply with the existing exposure limits based on free space power density measurements. On the
Int. J. Environ. Res. Public Health 2019,16, 3406 14 of 23
other hand, a strong reactive near field would considerably complicate the exposure situation [
111
].
Therefore, for dosimetry modeling of distances (from the antenna) below the wavelength of the MMW
(mm), temperature measurements should rather be performed in suitable phantoms rather than direct
measurements of the power densities in the free space [111].
The question is how reliably the power density (in free space) can be extrapolated to possible
temperature increases in human tissue? For example, Neufeld et al. [
112
] found that 10 GHz “bursts”
(considered “safe” by ICNIRP and IEEE) can cause temperature increases of >1
C if the burst duration
is long enough. It was also discussed whether the average values of the power densities for the safety
assessment are the right ones. In addition, the temperature increase by the MMW also depends on the
size of the area. Thus, the factors such as the amplitude of the burst, the “averaging area” and the
“averaging time” for the dosimetry would have to be considered.
Foster et al. [
113
] reviewed and modelled data on MMW-induced temperature increases in human
skin. The model takes into account the frequencies of 3–100 GHz and smaller skin areas with the
diameter of 1–2 cm. Available data on exposures lasting more than a few minutes, as well as areas
of skin larger than 2 cm in diameter, were limited and made modeling difficult, but consistent with
existing data. This means that this model, after appropriate evaluation for dosimetry, could use smaller
areas of the skin. The authors also commented on the exposure guidelines for frequencies from 3 to
300 GHz in a separate article [
114
]. Based on “thermal modeling,” the authors considered the current
guidelines to be conservative in terms of protection against temperature increases in the tissue. They
also pointed out that the averaging time and average area provisions require further refinement and
that the e↵ects of short high intensity bursts may not be protected by the guidelines.
Zhadobov et al. [
115
] addressed the problem of accurate temperature measurement in
in vitro
MMW studies. They found that the type of thermal probe (thermocouples are better than fiber optic
probes) and the size of the probe (smaller probes are more accurate) are relevant. In addition, they were
able to show that the initial temperature rise during exposure is rapid (within seconds until a plateau
is reached) and that the cells absorb very small amounts of energy, since most of the energy is already
absorbed in the cell culture medium. Nevertheless, the authors have calculated that the exposure of
58.4 GHz with 10 mW/cm
2
leads to SAR values of more than 100 W/kg in a cell monolayer. This value
is a fraction of the SAR values of the fluid surrounding the cells.
Several studies focused on the distribution of power density and the change in skin temperature
as a result of exposure to MMW in the 6 to 100 GHz frequency range. The studies are experimental
and/or modeling studies using previously published data. Alekseev et al. [
116
,
117
] investigated the
absorption of the skin of mice and humans at frequencies between 30 and 82 GHz (10 mW/cm
2
). They
found that in both species absorption into both the epidermis and the dermis occurs with a concomitant
loss of power density in the deeper regions. An extended study from the same group [
118
] on human
forearm skin showed that both temperature increase and SAR values depend on frequency (in the
interval of 25 to 75 GHz; 25, 73.3 and 128 mW/cm2).
Frequency dependence for temperature increases was also observed in a modeling study with
human facial skin [
119
]. Pulsed MMWs were used (6–100 GHz, 100 mW/cm
2
, 200–10,000 ms pulse
length) and the skin temperatures were modeled as the function of both pulse length and frequency.
Peak skin temperature increased as a function of frequency up to 20 GHz, while above 20 GHz it proved
to be dependent on “absorption hotspots”. In deeper regions (>2 mm), the temperature increases were
very low and highest around 10 GHz.
In addition, certain skin constituents have been shown to a↵ect energy absorption. It has been
shown that the presence of sweat glands [
120
,
121
] and also capillaries in the dermis can cause locally
elevated SAR levels [
122
]. The latter study showed that SAR levels in vessels could be up to 30 times
higher than in the surrounding skin, depending on the diameter of the vessels.
Both [
23
] and [
123
] have reported that the dielectric properties of di↵erent areas of the skin di↵er.
The first study found that so-called acupuncture points in healthy volunteers show di↵erent dielectric
Int. J. Environ. Res. Public Health 2019,16, 3406 15 of 23
properties when exposed to MMW (50–75 GHz, 14 mW/cm
2
), while the second study even found
di↵erences between the epidermis and dermis (0–110 GHz).
These studies suggest that both the frequency and the specific condition and composition of the
skin are relevant for tissue dosimetry. However, too few and very di↵erent studies are available to give
a conclusive picture on dosimetry of 5G-relevant MMW exposures.
4.3. ICNIRP and other Exposure Recommendations
The guidelines for exposure limits for radiofrequency electromagnetic fields from 3 to 300 GHz in
many countries are based on the recommendations of the International Commission on Non-Ionizing
Radiation Protection (ICNIRP) [
10
]. However, there are also other organizations dealing with limit
values such as the Institute of Electrical and Electronics Engineers, IEEE [
11
] or the US Federal
Communications Commission, FCC [124].
The guidelines contain basic exposure limits that are indicated as SAR or power density. The limits
for a given frequency di↵er only slightly, if at all, between the di↵erent guidelines. However,
an important di↵erence between the guidelines concerns frequency, as the SAR basic restriction values
change to power density. This frequency (range) is currently set by ICNIRP at 10 GHz, while IEEE
and FCC see this between 3–6 GHz. The current revision of these guidelines aims to harmonize
these frequencies.
The exposure limits specified in the guidelines should protect against warming of tissue above
1
C. The reason is that the perceived dangers of MMW energy are associated with excessive heating,
called thermal e↵ects. However, it must be considered that the guidelines mean a temperature increase
of 1
C relative to the starting temperature, regardless of the starting temperature. Elevations in
temperature may cause pain in the skin when moderately increased, whereas at temperatures of
43–44 C it may even induce burns [124,125].
At present, only thermal e↵ects due to high-frequency electromagnetic fields are recognized as
e↵ects. This means that e↵ects have a thermal component even if it is obviously not due to tissue that
has been damaged by excessive heating. On the other hand, it has been suggested that the MMW
exposure may also cause non-thermal e↵ects. So far, however, no recognized expert committee has
supported such an assertion.
4.4. Knowledge Gaps and Research Recommendations
Exposure of humans can occur through 5G devices with frequencies above 6 GHz, and may be
primarily on the skin and, to a lesser extent, on the eyes. This is due to the very low penetration depth
of this MMW. Therefore, it is important to investigate whether there are any health-related e↵ects on
the skin and/or e↵ects associated with the skin. These include acute skin damage from tissue heating
(burns), but possibly also less acute e↵ects (such as inflammation, tumor development, etc.). Such
e↵ects could appear after prolonged and repeated heating of superficial structures (the skin). This
would mean that thermal e↵ects occur that are not due to acute but to chronic damage.
It may also be that local exposure causes energy deposition in the dermis of the skin, which may
be so great as to a↵ect nerve endings and peripheral blood vessels through warming mechanisms. Such
scenarios were proposed by Ziskin [
9
] based on a series of studies by his group. These studies typically
used exposures around 60 GHz at a power density of 10 mW/cm
2
on the skin in the sternum area to
produce systemic e↵ects. The aim was to treat certain diseases and complaints. The idea was that the
treatment induces the release of the body’s own opioids and additionally stimulates the peripheral
nerves. The stimulation would depend on a local thermal e↵ect, which, due to the frequencies, induces
locally high SAR values, even at low power densities, thus warming the tissue.
Due to the contradictory information from various lines of evidence that cannot be scientifically
explained, and given the large gaps in knowledge regarding the health impact of MMW in the
6–100 GHz frequency range at relevant power densities for 5G, research is needed at many levels. It is
important to define exact frequency ranges and power densities for possible research projects. There is
Int. J. Environ. Res. Public Health 2019,16, 3406 16 of 23
an urgent need for research in the areas of dosimetry,
in vivo
dose-response studies and the question of
non-thermal e↵ects. It is therefore recommended that the following knowledge gaps should be closed
by appropriate research (the list of research recommendations is not prioritized):
•
Exact dosimetry with consideration of the skin for relevant frequency ranges, including the
consideration of short intense pulses (bursts)
•Studies on inflammatory reactions starting from the skin and the associated tissues
•In vivo
studies on the influence of a possible tissue temperature increase (e.g., nude mouse or
hairless mouse model)
•In vivo dose-response studies of heat development
•Use of in vitro models (3D models) of the skin for molecular and cellular endpoints
•Clarification of the question about non-thermal e↵ects (in vitro)
There are also questions about the environmental impact, with potential consequences for human
health. Since many MMW devices will be installed in the environment, the impact of MMW on
insects, plants, bacteria, and fungi is relevant to investigate. Particularly relevant is the question of
temperature increase in very small organisms, as the depth of penetration of the MMW could warm
the whole organism.
An unrealistic scenario, however, is that MMW exposures at realistic power densities could cause
systemic body warming in humans. Any local heat exposure would be dissipated by the body’s
normal heat regulation system. This is mainly due to convection caused by blood flow adjacent to the
superficial skin areas where the actual exposure takes place.
In summary, it should be noted that there are knowledge gaps with respect to local heat
developments on small living surfaces, e.g., on the skin or on the eye, which can lead to specific health
e↵ects. In addition, the question of any possibility of non-thermal e↵ects needs to be answered.
5. Conclusions
Since the ranges up to 30 GHz and over 90 GHz are sparingly represented, this review mainly
covers studies done in the frequency range from 30.1 to 65 GHz.
In summary, the majority of studies with MMW exposures show biological responses. From
this observation, however, no in-depth conclusions can be drawn regarding the biological and health
e↵ects of MMW exposures in the 6–100 GHz frequency range. The studies are very di↵erent and the
total number of studies is surprisingly low. The reactions occur both
in vivo
and
in vitro
and a↵ect all
biological endpoints studied.
There does not seem to be a consistent relationship between intensity (power density), exposure
time, or frequency, and the e↵ects of exposure. On the contrary, and strikingly, higher power densities
do not cause more frequent responses, since the percentage of responses in most frequency groups is
already at 70%. Some authors refer to their study results as having “non-thermal” causes, but few have
applied appropriate temperature controls. The question therefore remains whether warming is the
main cause of any observed MMW e↵ects?
In order to evaluate and summarize the 6–100 GHz data in this review, we draw the following
conclusions:
•
Regarding the health e↵ects of MMW in the 6–100 GHz frequency range at power densities not
exceeding the exposure guidelines the studies provide no clear evidence, due to contradictory
information from the in vivo and in vitro investigations.
•
Regarding the possibility of “non-thermal” e↵ects, the available studies provide no clear
explanation of any mode of action of observed e↵ects.
•
Regarding the quality of the presented studies, too few studies fulfill the minimal quality criteria
to allow any further conclusions.
Int. J. Environ. Res. Public Health 2019,16, 3406 17 of 23
Supplementary Materials:
The following are available online at http://www.mdpi.com/1660-4601/16/18/3406/s1,
Table S1: Master-table of the selected (
in vivo
and
in vitro
) studies and the extracted physical, biological, and
quality parameters.
Author Contributions:
M.S. and M.-O.M. have contributed equally to conceptualization, structuring, data
collection and analysis, interpretation of data, and all aspects of writing of the manuscript.
Funding:
This research was funded by Deutsche Telekom Technik GmbH, Bonn, Germany, PO number 4806344812.
Conflicts of Interest:
The authors declare no conflict of interest. The funders had no role in the design of the
study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to
publish the results.
References
1.
Ch
á
vez-Santiago, R.; Szydełko, M.; Kliks, A.; Foukalas, F.; Haddad, Y.; Nolan, K.E.; Kelly, M.Y.; Masonta, M.T.;
Balasingham, I. 5G: The Convergence of Wireless Communications. Wirel. Pers. Commun.
2015
,83, 1617–1642.
[CrossRef][PubMed]
2.
Chih-Lin, I.; Han, S.; Xu, Z.; Sun, Q.; Pan, Z. 5G: Rethink mobile communications for 2020+.Philos. Trans. R.
Soc. A Math. Phys. Eng. Sci. 2016,374, 20140432. [CrossRef][PubMed]
3.
International Agency for Research on Cancer. Non-Ionizing Radiation, Part. 2: Radiofrequency Electromagnetic
Fields; International Agency for Research on Cancer: Lyon, France, 2013; Volume 102, pp. 1–460.
4.
Scientific Committee on Emerging and Newly Identified Health Risks (SCENIHR). Opinion on: Potential
health e↵ects of exposure to electromagnetic fields. Bioelectromagnetics
2015
,36, 480–484. [CrossRef][PubMed]
5.
SSM’s Scientific Council on Electromagnetic Fields. Recent Research on EMF and Health Risk: Eleventh report
from SSM’S Scientific Council on Electromagnetic Fields; Swedish Radiation Safety Authority: Stockholm,
Sweden, 2016.
6.
Marques, M.M.; Berrington de Gonzalez, A.; Beland, F.A.; Browne, P.; Demers, P.A.; Lachenmeier, D.W.;
Bahadori, T.; Barupal, D.K.; Belpoggi, F.; Comba, P.; et al. Advisory Group recommendations on priorities for
the IARC Monographs. Lancet Oncol. 2019,20, 763–764. [CrossRef]
7.
WHO. Available online: https://www.who.int/peh-emf/publications/facts/fs226/en/(accessed on
8 August 2019).
8.
Mattsson, M.O.; Zeni, O.; Simk
ó
, M. Is there a Biological Basis for Therapeutic Applications of Millimetre
Waves and THz Waves? J. Infrared Millim. Terahertz Waves 2018,39, 863–878. [CrossRef]
9.
Ziskin, M.C. Millimeter waves: Acoustic and electromagnetic. Bioelectromagnetics
2013
,34, 3–14. [CrossRef]
[PubMed]
10.
International Commission on Non-Ionizing Radiation Protection (ICNIRP). Guidelines for limiting exposure
to time-varying electric, magnetic, and electromagnetic fields (Up to 300 GHz). Health Phys.
1998
,74, 494–522.
11.
IEEE Standards Coordinating Committee. IEEE Standard for Safety Levels with Respect to Human Exposure to
Radio Frequency Electromagnetic Fields, 3kHz to 300 GHz; Technical report No. C95.1-2005; IEEE: Piscataway,
NJ, USA, 2006; pp. 1–1991.
12.
IEEE. Available online: https://spectrum.ieee.org/video/telecom/wireless/everything-you-need-to-know-
about-5g (accessed on 8 August 2019).
13.
Zeni, O.; Scarf
ì
, M.R. Experimental requirements for
in vitro
studies aimed to evaluate the biological e↵ects of
radiofrequency radiation. In Microwave Materials Characterization; InTech: Rijeka, Croatia, 2012; pp. 121–138.
14.
Simk
ó
, M.; Remondini, D.; Zeni, O.; Scarfi, M.R. Quality Matters: Systematic Analysis of Endpoints Related
to ‘Cellular Life’ in Vitro Data of Radiofrequency Electromagnetic Field Exposure. Int. J. Environ. Res.
Public Health 2016,13, 701. [CrossRef]
15.
SCHEER–Scientific Committee on Health Environmental and Emerging Risks. Memorandum on Weight of
Evidence and Uncertainties; European Comission: Brussels, Belgium, 2018.
16.
Alekseev, S.I.; Gordiienko, O.V.; Radzievsky, A.A.; Ziskin, M.C. Millimeter wave e↵ects on electrical responses
of the sural nerve in vivo. Bioelectromagnetics 2010,31, 180–190. [CrossRef]
17.
Alekseev, S.I.; Radzievsky, A.A.; Szabo, I.; Ziskin, M.C. Local heating of human skin by millimeter waves:
E↵ect of blood flow. Bioelectromagnetics 2005,26, 489–501. [CrossRef]
Int. J. Environ. Res. Public Health 2019,16, 3406 18 of 23
18.
Bush, L.G.; Hill, D.W.; Riazi, A.; Stensaas, L.J.; Partlow, L.M.; Gandhi, O.P. E↵ects of millimeter-wave
radiation on monolayer cell cultures. III. A search for frequency-specific athermal biological e↵ects on
protein synthesis. Bioelectromagnetics 1981,2, 151–159. [CrossRef][PubMed]
19.
Chatterjee, I.; Yoon, J.; Wiese, R. Millimeter wave bioe↵ects at 94 GHz on skeletal muscle contraction.
In Proceedings of the 2013 IEEE Topical Conference on Biomedical Wireless Technologies, Networks, and
Sensing Systems, Austin, TX, USA, 20–23 January 2013; pp. 67–69.
20.
Chen, Q.; Zeng, Q.L.; Lu, D.Q.; Chiang, H. Millimeter Wave Exposure Reverses TPA Suppression of Gap
Junction Intercellular Communication in HaCaT Human Keratinocytes. Bioelectromagnetics
2004
,25, 1–4.
[CrossRef][PubMed]
21.
D’Agostino, S.; Della Monica, C.; Palizzi, E.; Di Pietrantonio, F.; Benetti, M.; Cannat
à
, D.; Cavagnaro, M.;
Sardari, D.; Stano, P.; Ramundo-Orlando, A. Extremely High Frequency Electromagnetic Fields Facilitate
Electrical Signal Propagation by Increasing Transmembrane Potassium E✏ux in an Artificial Axon Model.
Sci. Rep. 2018,8, 9299. [CrossRef][PubMed]
22.
Deghoyan, A.; Heqimyan, A.; Nikoghosyan, A.; Dadasyan, E.; Ayrapetyan, S. Cell bathing medium as
a target for non thermal e↵ect of millimeter waves. Electromagn. Biol. Med.
2012
,31, 132–142. [CrossRef]
[PubMed]
23.
Egot-Lemaire, S.J.-P.; Ziskin, M.C. Dielectric properties of human skin at an acupuncture point in the
50–75 GHz frequency range: A pilot study. Bioelectromagnetics 2011,32, 360–366. [CrossRef][PubMed]
24.
Franchini, V.; Franchini, V.; Regalbuto, E.; De Amicis, A.; De Sanctis, S.; Di Cristofaro, S.; Coluzzi, E.;
Marinaccio, J.; Sgura, A.; Ceccuzzi, S.; et al. Genotoxic E↵ects in Human Fibroblasts Exposed to Microwave
Radiation. Health Phys. 2018,115, 126–139. [CrossRef][PubMed]
25.
Frei, M.R.; Ryan, K.L.; Berger, R.E.; Jauchem, J.R. Sustained 35-GHz radiofrequency irradiation induces
circulatory failure. Shock 1995,4, 289–293. [CrossRef]
26.
Gapeyev, A.B.; Kulagina, T.P.; Aripovsky, A.V.; Chemeris, N.K. The role of fatty acids in anti-inflammatory
e↵ects of low-intensity extremely high-frequency electromagnetic radiation. Bioelectromagnetics
2011
,32,
388–395. [CrossRef]
27.
Gapeyev, A.B.; Gapeyev, A.B.; Mikhailik, E.N.; Mikhailik, E.N.; Chemeris, N.K. Anti-inflammatory e↵ects
of low-intensity extremely high-frequency electromagnetic radiation: Frequency and power dependence.
Bioelectromagnetics 2008,29, 197–206. [CrossRef]
28.
Alekseev, S.I.; Ziskin, M.C. E↵ects of millimeter waves on ionic currents of Lymnaea neurons.
Bioelectromagnetics 1999,20, 24–33. [CrossRef]
29.
Gapeyev, A.B.; Safronova, V.G.; Chemeris, N.K.; Fesenko, E.E. Inhibition of the production of reactive oxygen
species in mouse peritoneal neutrophils by millimeter wave radiation in the near and far field zones of the
radiator. Bioelectrochem. Bioenerg. 1997,43, 217–220. [CrossRef]
30.
Geletyuk, V.; Kazachenko, V. Dual e↵ects of microwaves on single Ca 2
+
-activated K
+
channels in cultured
kidney cells Vero. FEBS Lett. 1995,359, 85–88. [CrossRef]
31.
Grundier, W.; Keilmann, F. Nonthermal E↵ects of Millimeter Microwaves on Yeast Growth. Z. Naturforsch. C
1978,33, 15–22. [CrossRef]
32.
Haas, A.J.; Le Page, Y.; Zhadobov, M.; Boriskin, A.; Sauleau, R.; Le Dr
é
an, Y. Impact of 60-GHz millimeter
waves on stress and pain-related protein expression in di↵erentiating neuron-like cells. Bioelectromagnetics
2016,37, 444–454. [CrossRef][PubMed]
33.
Haas, A.J.; Le Page, Y.; Zhadobov, M.; Sauleau, R.; Le Dr
é
an, Y.; Saligaut, C. E↵ect of acute millimeter wave
exposure on dopamine metabolism of NGF-treated PC12 cells. J. Radiat. Res.
2017
,58, 439–445. [CrossRef]
[PubMed]
34.
Haas, A.J.; Le Page, Y.; Zhadobov, M.; Sauleau, R.; Le Dr
é
an, Y. E↵ects of 60-GHz millimeter waves on neurite
outgrowth in PC12 cells using high-content screening. Neurosci. Lett.
2016
,618, 58–65. [CrossRef][PubMed]
35.
Habauzit, D.; Le Qu
é
ment, C.; Zhadobov, M.; Martin, C.; Aubry, M.; Sauleau, R.; Le Dr
é
an, Y. Transcriptome
analysis reveals the contribution of thermal and the specific e↵ects in cellular response to millimeter wave
exposure. PLoS ONE 2014,9, e109435. [CrossRef]
36.
Ivanov, V.B.; Subbotina, T.I.; Khadartsev, A.A.; Yashin, M.A.; Yashin, A.A. Exposure to low-intensive
superhigh frequency electromagnetic field as a factor of carcinogenesis in experimental animals. Bull. Exp.
Biol. Med. 2005,139, 241–244. [CrossRef]
Int. J. Environ. Res. Public Health 2019,16, 3406 19 of 23
37.
Jauchem, J.R.; Ryan, K.L.; Walters, T.J. Pathophysiological alterations induced by sustained 35-GHz
radio-frequency energy heating. J. Basic Clin. Physiol. Pharmacol. 2016,27, 79–89. [CrossRef]
38.
Kojima, M.; Hanazawa, M.; Yamashiro, Y.; Sasaki, H.; Watanabe, S.; Taki, M.; Suzuki, Y.; Hirata, A.;
Kamimura, Y.; Sasaki, K. Acute ocular injuries caused by 60-GHZ millimeter-wave exposure. Heal. Phys.
2009,97, 212–218. [CrossRef]
39.
Alekseev, S.I.; Ziskin, M.C.; Kochetkova, N.V.; Bolshakov, M.A. Millimeter waves thermally alter the firing
rate of the Lymnaea pacemaker neuron. Bioelectromagnetics 1997,18, 89–98. [CrossRef]
40.
Kojima, M.; Kojima, M.; Suzuki, Y.; Sasaki, K.; Taki, M.; Wake, K.; Watanabe, S.; Mizuno, M.; Tasaki, T.;
Sasaki, H. Ocular E↵ects of Exposure to 40, 75, and 95 GHz Millimeter Waves. J. Infrared Millim. Terahertz Waves
2018,39, 912–925. [CrossRef]
41.
Korenstein-Ilan, A.; Barbul, A.; Hasin, P.; Eliran, A.; Gover, A.; Korenstein, R. Terahertz Radiation Increases
Genomic Instability in Human Lymphocytes. Radiat. Res. 2008,170, 224–234. [CrossRef]
42.
Koschnitzke, C.; Kremer, F.; Santo, L.; Quick, P.; Poglitsch, A. A Non-Thermal E↵ect of Millimeter Wave
Radiation on the Puffing of Giant Chromosomes. Zeitschrift für Naturforsch C 1983,38, 883–886. [CrossRef]
43.
Koyama, S.; Narita, E.; Shimizu, Y.; Suzuki, Y.; Shiina, T.; Taki, M.; Shinohara, M.; Miyakoshi, J. E↵ects of
long-term exposure to 60 GHz millimeter-wavelength radiation on the genotoxicity and heat shock protein
(HSP) expression of cells derived from human eye. Int. J. Environ. Res. Public Health
2016
,13, 802. [CrossRef]
44.
Kues, H.A.; Anna, S.A.D.; Osiander, R.; Green, W.R.; Monahan, J.C. Absence of Ocular E↵ects After Either
Single or Repeated Exposure to 10 mW/cm
2
from a 60 GHz CW Source. Bioelectromagnetics
1999
,473, 463–473.
[CrossRef]
45.
Le Qu
é
ment, C.; Nicolas Nicolaz, C.; Zhadobov, M.; Desmots, F.; Sauleau, R.; Aubry, M.; Michel, D.;
Le Dr
é
an, Y. Whole-genome expression analysis in primary human keratinocyte cell cultures exposed to
60 GHz radiation. Bioelectromagnetics 2012,33, 147–158.
46.
LeQu
é
ment, C.; Nicolaz, C.N.; Habauzit, D.; Zhadobov, M.; Sauleau, R.; Le Dr
é
an, Y. Impact of 60-GHz
millimeter waves and corresponding heat e↵ect on endoplasmic reticulum stress sensor gene expression.
Bioelectromagnetics 2014,35, 444–451. [CrossRef]
47.
Li, X.; Ye, H.; Yu, F.; Cai, L.; Li, H.; Chen, J.; Wu, M.; Chen, W.; Lin, R.; Li, Z.; et al. Millimeter wave treatment
promotes chondrocyte proliferation via G 1/S cell cycle transition. Int. J. Mol. Med.
2012
,29, 823–831.
[CrossRef]
48. Logani, M.K.; Agelan, A.; Ziskin, M.C. E↵ect of millimeter wave radiation on catalase activity. Electromagn.
Biol. Med. 2002,21, 303–308. [CrossRef]
49.
Logani, M.K.; Anga, A.; Szabo, I.; Agelan, A.; Irizarry, A.R.; Ziskin, M.C. E↵ect of millimeter waves on
cyclophosphamide induced suppression of the immune system. Bioelectromagnetics
2002
,23, 614–621.
[CrossRef]
50.
Alekseev, S.I.; Ziskin, M.C. Millimeter microwave e↵ect on ion transport across lipid bilayer membranes.
Bioelectromagnetics 1995,16, 124–131. [CrossRef]
51.
Makar, V.; Logani, M.; Szabo, I.; Zlskin, M. E↵ect of Millimeter Waves on Cyclophosphamide Induced
Suppression of T Cell Functions. Bioelectromagnetics 2003,24, 356–365. [CrossRef]
52.
Makar, V.R.; Logani, M.K.; Bhanushali, A.; Alekseev, S.I.; Ziskin, M.C. E↵ect of cyclophosphamide and
61.22 GHz millimeter waves on T-cell, B-cell, and macrophage functions. Bioelectromagnetics
2006
,27, 458–466.
[CrossRef]
53.
Makar, V.R.; Logani, M.K.; Bhanushali, A.; Kataoka, M.; Ziskin, M.C. E↵ect of millimeter waves on natural
killer cell activation. Bioelectromagnetics 2005,26, 10–19. [CrossRef]
54.
Mason, P.A.; Walters, T.J.; DiGiovanni, J.; Beason, C.W.; Jauchem, J.R.; Dick, E.J., Jr.; Mahajan, K.; Dusch, S.J.;
Shields, B.A.; Merritt, J.H.; et al. Lack of e↵ect of 94 GHz radio frequency radiation exposure in an animal
model of skin carcinogenesis in the radio frequency radiation (RFR) band is mutagenic, as either a promoter
or co-promoter in some animal models of carcinogenesis. Recent develop. Carcinogenesis
2001
,22, 1701–1708.
[CrossRef]
55.
Millenbaugh, N.J.; Kiel, J.L.; Ryan, K.L.; Blystone, R.V.; Kalns, J.E.; Brott, B.J.; Cerna, C.Z.; Lawrence, W.S.;
Soza, L.L.; Mason, P.A. Comparison of blood pressure and thermal responses in rats exposed to millimeter
wave energy or environmental heat. Shock 2006,25, 625–632. [CrossRef]
Int. J. Environ. Res. Public Health 2019,16, 3406 20 of 23
56.
Millenbaugh, N.J.; Roth, C.; Sypniewska, R.; Chan, V.; Eggers, J.S.; Kiel, J.L.; Blystone, R.V.; Mason, P.A. Gene
expression changes in the skin of rats induced by prolonged 35 GHz millimeter-wave exposure. Radiat. Res.
2008,169, 288–300. [CrossRef]
57.
Narinyan, L.; Ayrapetyan, S. Cyclic AMP-dependent signaling system is a primary metabolic target for
non-thermal e↵ect of microwaves on heart muscle hydration. Electromagn. Biol. Med.
2017
,36, 182–191.
[CrossRef]
58.
Nguyen, T.H.; Pham, V.T.; Nguyen, S.H.; Baulin, V.; Croft, R.J.; Phillips, B.; Crawford, R.J.; Ivanova, E.P.
The bioe↵ects resulting from prokaryotic cells and yeast being exposed to an 18 GHz electromagnetic field.
PLoS ONE 2016,11, e0158135. [CrossRef]
59.
Nicolaz, C.N.; Zhadobov, M.; Desmots, F.; Ansart, A.; Sauleau, R.; Thouroude, D.; Michel, D.; Le Drean, Y.
Study of narrow band millimeter-wave potential interactions with endoplasmic reticulum stress sensor
genes. Bioelectromagnetics 2009,30, 365–373. [CrossRef]
60.
Nicolas Nicolaz, C.; Zhadobov, M.; Desmots, F.; Sauleau, R.; Thouroude, D.; Michel, D.; Le Drean, Y. Absence
of direct e↵ect of low-power millimeter-wave radiation at 60.4 GHz on endoplasmic reticulum stress. Cell Biol.
Toxicol. 2009,25, 471–478. [CrossRef]
61.
Bellossi, A.; Dubost, G.; Moulinoux, J.P.; Himdi, M.; Ruelloux, M.; Rocher, C. Biological E↵ects of
Millimeter-Wave Irradiation on Mice—Preliminary Results. IEEE Trans. Microw. Theory Tech.
2000
,
48, 2104–2110.
62.
Pakhomov, A.G.; Prol, H.K.; Mathur, S.P.; Akyel, Y.; Campbell, C.B.G.B.G. Role of field intensity in the
biological e↵ectiveness of millimeter waves at a resonance frequency. Bioelectrochem. Bioenerg.
1997
,43,
27–33. [CrossRef]
63.
Pakhomov, A.G.; Prol, H.K.; Mathur, S.P.; Akyel, Y.; Campbell, C.B. Search for frequency-specific e↵ects of
millimeter-wave radiation on isolated nerve function. Bioelectromagnetics 1997,18, 324–334. [CrossRef]
64.
Partlow, L.M.; Bush, L.G.; Stensaas, L.J.; Hill, D.W.; Riazi, A.; Gandhi, O.P. E↵ects of millimeter-wave
radiation on monolayer cell cultures. I. Design and validation of a novel exposure system. Bioelectromagnetics
1981,2, 123–140. [CrossRef]
65.
Partyla, T.; Hacker, H.; Edinger, H.; Leutzow, B.; Lange, J.; Usichenko, T. Remote E↵ects of Electromagnetic
Millimeter Waves on Experimentally Induced Cold Pain: A Double-Blinded Crossover Investigation in
Healthy Volunteers. Anesth. Analg. 2017,124, 980–985. [CrossRef]
66.
Pikov, V.; Arakaki, X.; Harrington, M.; Fraser, S.E.; Siegel, P.H. Modulation of neuronal activity and plasma
membrane properties with low-power millimeter waves in organotypic cortical slices. J. Neural Eng.
2010
,7,
045003. [CrossRef]
67.
Radzievsky, A.; Gordiienko, O.; Cowan, A.; Alekseev, S.; Ziskin, M. Millimeter-Wave-Induced Hypoalgesia
in Mice: Dependence on Type of Experimental Pain. IEEE Trans. Plasma Sci.
2004
,32, 1634–1643. [CrossRef]
68.
Radzievsky, A.A.; Cowan, A.; Byrd, C.; Radzievsky, A.A.; Ziskin, M.C. Single millimeter wave treatment
does not impair gastrointestinal transit in mice. Life Sci. 2002,71, 1763–1770. [CrossRef]
69.
Radzievsky, A.A.; Gordiienko, O.V.; Alekseev, S.; Szabo, I.; Cowan, A.; Ziskin, M.C. Electromagnetic
millimeter wave induced hypoalgesia: Frequency dependence and involvement of endogenous opioids.
Bioelectromagnetics 2008,29, 284–295. [CrossRef]
70.
Radzievsky, A.; Gordiienko, O.; Szabo, I.; Alekseev, S.I.; Ziskin, M.C. Millimeter wave-induced suppression
of B16 F10 melanoma growth in mice: Involvement of endogenous opioids. Bioelectromagnetics
2004
,25,
466–473. [CrossRef]
71.
Radzievsky, A.A.; Rojavin, M.A.; Cowan, A.; Alekseev, S.I.; Radzievsky, A.A.; Ziskin, M.C. Peripheral neural
system involvement in hypoalgesic e↵ect of electromagnetic millimeter waves. Life Sci.
2001
,62, 1143–1151.
[CrossRef]
72.
Beneduci, A.; Chidichimo, G.; De Rose, R.; Filippelli, L.; Straface, S.V.; Venuta, S. Frequency and irradiation
time-dependant antiproliferative e↵ect of low-power millimeter waves on RPMI 7932 human melanoma cell
line. Anticancer Res. 2005,25, 1023–1028.
73.
Radzievsky, A.A.; Rojavin, M.A.; Cowan, A.; Alekseev, S.I.; Ziskin, M.C. Hypoalgesic e↵ect of millimeter
waves in mice: Dependence on the site of exposure. Life Sci. 2000,66, 2101–2111. [CrossRef]
74.
Radzievsky, A.A.; Rojavin, M.A.; Cowan, A.; Ziskin, M.C. Suppression of pain sensation caused by millimeter
waves: A double-blinded, cross-over, prospective human volunteer study. Anesth. Analg.
1999
,88, 836–840.
[CrossRef]
Int. J. Environ. Res. Public Health 2019,16, 3406 21 of 23
75.
Rojavin, M.A.; Cowan, A.; Radzievsky, A.A.; Ziskin, M.C. Antipruritic e↵ect of millimeter waves in mice:
Evidence for opioid involvement. Life Sci. 1998,63, 251–257. [CrossRef]
76.
Rojavin, M.A.; Radzievsky, A.A.; Cowan, A.; Ziskin, M.C. Pain relief caused by millimeter waves in mice:
Results of cold water tail flick tests. Int. J. Radiat. Biol. 2000,76, 575–579.
77.
Romanenko, S.; Siegel, P.H.; Wagenaar, D.A.; Pikov, V. E↵ects of millimeter wave irradiation and equivalent
thermal heating on the activity of individual neurons in the leech ganglion. J. Neurophysiol.
2014
,112,
2423–2431. [CrossRef]
78.
Ryan, K.L.; Frei, M.R.; Berger, R.E.; Jauchem, J.R. Does nitric oxide mediate circulatory failure induced by
35-GHz microwave heating? Shock 1996,6, 71–76. [CrossRef]
79.
Ryan, K.L.; Frei, M.R.; Jauchem, J.R. Circulatory failure induced by 35 GHz microwave heating: E↵ects of
chronic nitric oxide synthesis inhibition. Shock 1997,7, 70–76. [CrossRef]
80.
Safronova, V.G.; Gabdoulkhakova, A.G.; Santalov, B.F. Immunomodulating Action of Low Intensity Millimeter
Waves on Primed Neutrophils. Bioelectromagnetics 2002,23, 599–606. [CrossRef]
81.
Samoilov, V.O.; Shadrin, E.B.; Filippova, E.B.; Katsnelson, Y.; Backho↵, H.; Eventov, M. The e↵ect of
transcranial electromagnetic brain stimulation on the acquisition of the conditioned response in rats.
Biophysics 2015,60, 303–308. [CrossRef]
82.
Samsonov, A.; Popov, S.V. The e↵ect of a 94 GHz electromagnetic field on neuronal microtubules.
Bioelectromagnetics 2013,34, 133–144. [CrossRef]
83.
Beneduci, A.; Chidichimo, G.; Tripepi, S.; Perrotta, E. Transmission electron microscopy study of the e↵ects
produced by wide-band low-power millimeter waves on MCF-7 human breast cancer cells in culture.
Anticancer Res. 2005,25, 1009–1013.
84.
Shanin, S.N.; Rybakina, E.G.; Novikova, N.N.; Kozinets, I.A.; Rogers, V.J.; Korneva, E.A. Natural killer cell
cytotoxic activity and c-Fos protein synthesis in rat hypothalamic cells after painful electric stimulation of
the hind limbs and EHF irradiation of the skin. Med. Sci Monit. 2005,11, BR309–BR315.
85.
Shapiro, M.G.; Priest, M.F.; Siegel, P.H.; Bezanilla, F. Thermal Mechanisms of Millimeter Wave Stimulation of
Excitable Cells. Biophys. J. 2013,104, 2622–2628. [CrossRef]
86.
Shckorbatov, Y.G.; Grigoryeva, N.N.; Shakhbazov, V.G.; Grabina, V.A.; Bogoslavsky, A.M. Microwave
irradiation influences on the state of human cell nuclei. Bioelectromagnetics 1998,19, 414–419. [CrossRef]
87.
Sivachenko, I.B.; Medvedev, D.S.; Molodtsova, I.D.; Panteleev, S.S.; Sokolov, A.Y.; Lyubashina, O.A. E↵ects of
Millimeter-Wave Electromagnetic Radiation on the Experimental Model of Migraine. Bull. Exp. Biol. Med.
2016,160, 425–428. [CrossRef]
88.
Soubere Mahamoud, Y.; Aite, M.; Martin, C.; Zhadobov, M.; Sauleau, R.; Le Dr
é
an, Y.; Habauzit, D. Additive
E↵ects of Millimeter Waves and 2-Deoxyglucose Co-Exposure on the Human Keratinocyte Transcriptome.
PLoS ONE 2016,11, e0160810. [CrossRef]
89.
Stensaas, L.J.; Partlow, L.M.; Bush, L.G.; Iversen, P.L.; Hill, D.W.; Hagmann, M.J.; Gandhi, O.P. E↵ects of
millimeter-wave radiation on monolayer cell cultures. II. Scanning and transmission electron microscopy.
Bioelectromagnetics 1981,2, 141–150. [CrossRef][PubMed]
90.
Sun, S.; Titushkin, I.; Varner, J.; Cho, M. Millimeter Wave-induced Modulation of Calcium Dynamics in
an Engineered Skin Co-culture Model: Role of Secreted ATP on Calcium Spiking. J. Radiat. Res.
2012
,53,
159–167. [CrossRef][PubMed]
91.
Sypniewska, R.K.; Millenbaugh, N.J.; Kiel, J.L.; Blystone, R.V.; Ringham, H.N.; Mason, P.A.; Witzmann, F.A.
Protein changes in macrophages induced by plasma from rats exposed to 35 GHz millimeter waves.
Bioelectromagnetics 2010,31, 656–663. [CrossRef][PubMed]
92.
Szabo, I.; Kappelmayer, J.; Alekseev, S.I.; Ziskin, M.C. Millimeter wave induced reversible externalization
of phosphatidylserine molecules in cells exposed
in vitro
.Bioelectromagnetics
2006
,27, 233–244. [CrossRef]
[PubMed]
93.
Szabo, I.; Manning, M.R.; Radzlevsky, A.A.; Wetzel, M.A.; Rogers, T.J.; Ziskin, M.C. Low Power Millimeter
Wave Irradiation Exerts No Harmful E↵ect on Human Keratinocytes In Vitro. Bioelectromagnetics
2003
,24,
165–173. [CrossRef]
94.
Beneduci, A.; Chidichimo, G.; Tripepi, S.; Perrotta, E.; Cufone, F. Antiproliferative e↵ect of millimeter radiation
on human erythromyeloid leukemia cell line K562 in culture: Ultrastructural and metabolic-induced changes.
Bioelectrochemistry 2007,70, 214–220. [CrossRef]
Int. J. Environ. Res. Public Health 2019,16, 3406 22 of 23
95.
Szabo, I.; Rojavin, M.A.; Rogers, T.J.; Ziskin, M.C. Reactions of keratinocytes to
in vitro
millimeter wave
exposure. Bioelectromagnetics 2001,22, 358–364. [CrossRef]
96.
Titushkin, I.A.; Rao, V.S.; Pickard, W.F.; Moros, E.G.; Shafirstein, G.; Cho, M.R. Altered Calcium Dynamics
Mediates P19-Derived Neuron-Like Cell Responses to Millimeter-Wave Radiation. Radiat. Res.
2009
,172,
725–763. [CrossRef]
97.
Tong, Y.; Yang, Z.; Yang, D.; Chu, H.; Qu, M.; Liu, G.; Wu, Y.; Liu, S. Millimeter-wave exposure promotes
the di↵erentiation of bone marrow stromal cells into cells with a neural phenotype. J. Huazhong Univ. Sci.
Technolog. Med. Sci. 2009,29, 409–412. [CrossRef]
98.
Logani, M.K.; Bhanushali, A.; Ziskin, M.C.; Prihoda, T.J. Micronuclei in peripheral blood and bone marrow
cells of mice exposed to 42 GHz electromagnetic millimeter waves. Radiat. Res.
2004
,161, 341–345. [CrossRef]
99.
Vlasova, I.I.; Mikhalchik, E.V.; Gusev, A.A.; Balabushevich, N.G.; Gusev, S.A.; Kazarinov, K.D.
Extremely high-frequency electromagnetic radiation enhances neutrophil response to particulate agonists.
Bioelectromagnetics 2018,39, 144–155. [CrossRef][PubMed]
100.
Volkova, N.A.; Pavlovich, E.V.; Gapon, A.A.; Nikolov, O.T. E↵ects of millimeter-wave electromagnetic
exposure on the morphology and function of human cryopreserved spermatozoa. Bull. Exp. Biol. Med.
2014
,
157, 574–576. [CrossRef][PubMed]
101.
Wu, G.; Sferra, T.; Chen, X.; Chen, Y.; Wu, M.; Xu, H.; Peng, J.; Liu, X. Millimeter wave treatment inhibits
the mitochondrion-dependent apoptosis pathway in chondrocytes. Mol. Med. Rep.
2011
,4, 1001–1006.
[PubMed]
102.
Wu, G.W.; Liu, X.X.; Wu, M.X.; Zhao, J.Y.; Chen, W.L.; Lin, R.H.; Lin, J.M. Experimental study of millimeter
wave-induced di↵erentiation of bone marrow mesenchymal stem cells into chondrocytes. Int. J. Mol. Med.
2009,23, 461–467. [PubMed]
103.
Xia, L.; Luo, Q.-L.; Lin, H.-D.; Zhang, J.-L.; Guo, H.; He, C.-Q. The e↵ect of di↵erent treatment time of
millimeter wave on chondrocyte apoptosis, caspase-3, caspase-8, and MMP-13 expression in rabbit surgically
induced model of knee osteoarthritis. Rheumatol. Int. 2012,32, 2847–2856. [CrossRef][PubMed]
104.
Xie, T.; Pei, J.; Cui, Y.; Zhang, J.; Qi, H.; Chen, S.; Qiao, D. EEG changes as heat stress reactions in rats
irradiated by high intensity 35 GHZ millimeter waves. Health Phys. 2011,100, 632–640. [CrossRef]
105.
Beneduci, A. Evaluation of the Potential In Vitro Antiproliferative E↵ects of Millimeter Waves at Some
Therapeutic Frequencies on RPMI 7932 Human Skin Malignant Melanoma Cells. Cell Biochem. Biophys.
2009
,
55, 25–32. [CrossRef][PubMed]
106.
Yaekashiwa, N.; Otsuki, S.; Hayashi, S.; Kawase, K. Investigation of the non-thermal e↵ects of exposing cells
to 70-300 GHz irradiation using a widely tunable source. J. Radiat. Res. 2018,59, 116–121. [CrossRef]
107.
Yu, G.; Coln, E.A.; Schoenbach, K.H.; Gellerman, M.; Fox, P.; Rec, L.; Beebe, S.J.; Liu, S. A study on biological
e↵ects of low-intensity millimeter waves. IEEE Trans. Plasma Sci. 2002,30, 1489–1496.
108.
Zhadobov, M.; Desmots, F.; Thouroude, D.; Michel, D.; Drean, Y. Evaluation of the potential biological
e↵ects of the 60-GHz millimeter waves upon human cells. IEEE Trans. Antennas Propag.
2009
,57, 2949–2956.
[CrossRef]
109.
Zhadobov, M.; Sauleau, R.; Le Coq, L.; Debure, L.; Thouroude, D.; Michel, D.; Le Dr
é
an, Y.
Low-power millimeter wave radiations do not alter stress-sensitive gene expression of chaperone proteins.
Bioelectromagnetics 2007,28, 188–196. [CrossRef][PubMed]
110.
Aminzadeh, R.; Thielens, A.; Bamba, A.; Kone, L.; Gaillot, D.P.; Lienard, M.; Martens, L.; Joseph, W. On-body
calibration and measurements using personal radiofrequency exposimeters in indoor di↵use and specular
environments. Bioelectromagnetics 2016,37, 298–309. [CrossRef][PubMed]
111.
Colombi, D.; Thors, B.; Tornevik, C.; Balzano, Q. RF Energy Absorption by Biological Tissues in Close
Proximity to Millimeter-Wave 5G Wireless Equipment. IEEE Access 2018,6, 4974–4981. [CrossRef]
112.
Neufeld, E.; Carrasco, E.; Murbach, M.; Balzano, Q.; Christ, A.; Kuster, N. Theoretical and numerical
assessment of maximally allowable power-density averaging area for conservative electromagnetic exposure
assessment above 6 GHz. Bioelectromagnetics 2018,39, 617–630. [CrossRef][PubMed]
113.
Foster, K.R.; Ziskin, M.C.; Balzano, Q. Thermal modeling for the next generation of radiofrequency exposure
limits: Commentary. Health Phys. 2017,113, 41–53. [CrossRef]
114.
Foster, K.R.; Ziskin, M.C.; Balzano, Q. Thermal Response of Human Skin to Microwave Energy: A Critical
Review. Health Phys. 2016,111, 528–541. [CrossRef]
Int. J. Environ. Res. Public Health 2019,16, 3406 23 of 23
115.
Zhadobov, M.; Alekseev, S.I.; Sauleau, R.; Le Page, Y.; Le Drean, Y.; Fesenko, E.E. Microscale temperature
and SAR measurements in cell monolayer models exposed to millimeter waves. Bioelectromagnetics
2017
,38,
11–21. [CrossRef]
116.
Alekseev, S.I.; Gordiienko, O.V.; Ziskin, M.C. Reflection and penetration depth of millimeter waves in murine
skin. Bioelectromagnetics 2008,29, 340–344. [CrossRef]
117.
Alekseev, S.I.; Radzievsky, A.A.; Logani, M.K.; Ziskin, M.C. Millimeter wave dosimetry of human skin.
Bioelectromagnetics 2008,29, 65–70. [CrossRef]
118.
Alekseev, S.I.; Ziskin, M.C.; Fesenko, E.E. Frequency dependence of heating of human skin exposed to
millimeter waves. Biophysics 2012,57, 90–93. [CrossRef]
119.
Laakso, I.; Morimoto, R.; Heinonen, J.; Jokela, K.; Hirata, A. Human exposure to pulsed fields in the frequency
range from 6 to 100 GHz. Phys. Med. Biol. 2017,62, 6980–6992. [CrossRef]
120.
Feldman, Y.; Puzenko, A.; Ben Ishai, P.; Cadu↵, A.; Agranat, A.J. Human skin as arrays of helical antennas in
the millimeter and submillimeter wave range. In Proceedings of the 2008 33rd International Conference on
Infrared, Millimeter and Terahertz Waves, Pasadena, CA, USA, 27 March 2008; pp. 1–2.
121.
Shafirstein, G.; Moros, E.G. Modelling millimetre wave propagation and absorption in a high resolution skin
model: The e↵ect of sweat glands. Phys. Med. Biol. 2011,56, 1329. [CrossRef]
122.
Alekseev, S.I.; Ziskin, M.C. Enhanced absorption of millimeter wave energy in murine subcutaneous blood
vessels. Bioelectromagnetics 2011,32, 423–433. [CrossRef]
123.
Sasaki, K.; Wake, K.; Watanabe, S. Measurement of the dielectric properties of the epidermis and dermis at
frequencies from 0.5 GHz to 110 GHz. Phys. Med. Biol. 2014,59, 4739–4747. [CrossRef][PubMed]
124.
Federal Communications Commission (FCC). Available online: https://www.fcc.gov/general/radio-frequency-
safety-0 (accessed on 8 August 2019).
125.
Foster, K.R.; Morrissey, J.J. Thermal aspects of exposure to radiofrequency energy: Report of a workshop.
Int. J. Hyperth. 2011,27, 307–319. [CrossRef]
©
2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).