Exposure to non-ionizing electromagnetic radiation from mobile telephony and the association with psychiatric symptoms

The aim of this study was to investigate the association between exposure to non-ionizing electromagnetic radiation from mobile phone base stations and psychiatric symptoms. In a crosssectional study in Salvador, Bahia State, Brazil, 440 individuals were interviewed. Psychiatric complaints and diagnoses were the dependent variables and distance from the individual’s residence to the base station was considered the main independent variable. Hierarchical logistic regression analysis was conducted to assess confounding. An association was observed between psychiatric symptoms and residential proximity to the base station and different forms of mobile phone use (making calls with weak signal coverage, keeping the mobile phone close to the body, having two or more chips, and never turning off the phone while sleeping), and with the use of other electronic devices. The study concluded that exposure to electromagnetic radiation from mobile phone base stations and other electronic devices was associated with psychiatric symptoms, independently of gender, schooling, and smoking status. The adoption of precautionary measures to reduce such exposure is recommended. Nonionizing Radiation; Radiation Exposure; Cell Phones; Psychic Symptoms Resumo O objetivo desse estudo foi investigar a associação entre exposição a radiações eletromagnéticas não ionizantes da estação radiobase de telefonia celular e sintomas à saúde. Em um estudo transversal realizado em Salvador, Bahia, Brasil, entrevistaram-se 440 indivíduos. Queixas e diagnósticos psiquiátricos constituíram as variáveis dependentes e a distância do domicílio para estação radiobase foi considerada a variável independente principal. Realizou-se análise de regressão logística hierarquizada para avaliação de confundimento e efeito. Observou-se associação entre sintomas psiquiátricos e residir próximo à estação radiobase e formas de uso do telefone celular (sinal de cobertura fraco, perto do corpo, dois ou mais chips e nunca desligar o celular quando dorme) e com uso de outros eletroeletrônicos. Concluiu-se que a exposição à radiação eletromagnética não ionizante de telefonia celular e a outros eletroeletrônicos foi associada aos sintomas psiquiátricos independente do sexo, escolaridade e tabagismo. Recomenda-se a adoção de medidas precaucionárias no sentido de se reduzir este tipo de exposição. Radiação Não Ionizante; Exposição à Radiação; Telefones Celulares; Sintomas Psíquicos http://dx.doi.org/10.1590/0102-311X00104114 ARTIGO ARTICLE 2110 Cad. Saúde Pública, Rio de Janeiro, 31(10):2110-2126, out, 2015 SINTOMAS PSIQUIÁTRICOS E EXPOSIÇÃO À RADIAÇÃO DA TELEFONIA CELULAR 2111 Cad. Saúde Pública, Rio de Janeiro, 31(10):2110-2126, out, 2015 Introdução O crescimento da telefonia celular e, portanto, do número de estações radiobase, que estabelecem a comunicação com os aparelhos celulares, aumentou os benefícios para o estilo atual de vida, uma vez que tem facilitado a comunicação, oferecendo conforto, dando oportunidade de manter-se conectado com localidades próximas e distantes, e possibilitando o uso da internet para diversos fins. No entanto, a telefonia celular tem gerado preocupações sobre os possíveis efeitos à saúde das populações expostas às radiações eletromagnéticas não ionizantes. Trata-se de radiações caracterizadas por seu comprimento de onda, frequência e energia irradiada, consideradas como não transportadoras de energia suficiente para alterar o estado físico de um átomo 1. Para que a comunicação da telefonia celular ocorra, o sistema é subdividido em células. Cada célula possui uma estação radiobase capaz de enviar sinais com potência em toda sua extensão. Cada estação radiobase pode atender a diversos aparelhos de telefones ao mesmo tempo, designando para cada um, uma pequena faixa de frequências 2. A radiação eletromagnética não ionizante é absorvida pela pele e por níveis mais profundos do corpo, dissipando-se repetidamente com profundidade, podendo causar um aumento de temperatura não percebido pelos sensores térmicos naturais, localizados superficialmente. O aquecimento gerado internamente depende do tempo de exposição, da intensidade do campo e da espessura do tecido, não podendo às vezes ser compensado pelo organismo, ocasionando efeitos biológicos 3. A profundidade de penetração das ondas ao redor de frequência de 900MHz, usada na telefonia celular, em tecidos com alto conteúdo de água, como o muscular, é de 3cm. As ondas de 2.400MHz dos fornos de micro-ondas penetram cerca de 1,7cm. Em tecidos com baixo teor de água, como o ósseo, esses valores são respectivamente, 17,7cm e 11,2cm 4. Parece provável que as raízes da maioria dos transtornos psiquiátricos residam em alguma combinação de fatores genéticos e ambientais (biológico ou psicossocial) 5. Uma das observações mais documentadas em estudos epidemiológicos é a maior prevalência de transtornos de ansiedade e depressão em mulheres que em homens 6. A associação entre esses efeitos psiquiátricos e a exposição às radiações eletromagnéticas não ionizantes de telefonia celular tem sido investigada. Santini et al. 7 indicaram sintomas mais significativos por influência da distância de até 300m da estação radiobase de telefonia celular: irritabilidade, depressão, perda da memória, tontura, baixa da libido, dor de cabeça, perturbação do sono, desconforto (200m); cansaço (300m). Destaca-se que sete sintomas foram mais significativos nas mulheres: náuseas, perda de apetite, distúrbios visuais, tendência depressiva, dor de cabeça, insônia, sensação de desconforto. No caso dos homens, a diminuição na libido foi a queixa mais referida. Navarro et al. 8 e Bortkiewicz et al. 9 indicaram que moradores próximos a estação radiobase relataram problemas circulatórios, distúrbios do sono, irritabilidade, depressão, visão turva e dificuldades de concentração. Já Abdel-Rassoul et al. 10 sugeriram uma relação entre residir próximo à estação radiobase e problemas neurocomportamentais, tais como, tendência depressiva, tremores, tonturas, cefaleia, distúrbios do sono, distúrbios visuais, dentre outros. Da mesma forma, Oberfeld et al. 11 relataram que as pessoas que viviam perto das estações radiobase referiam mais sintomas de irritabilidade, fadiga, cefaleia, náuseas, perda de memória, distúrbio visual, tontura e problemas cardiovasculares, quanto maior o seu nível de exposição às micro-ondas. Augner et al. 12 estudaram a exposição à radiofrequência de estações radiobase em três grupos de pessoas. Cada grupo foi submetido a um campo de 900MHz, com tempos de exposição variáveis. Houve aumento significativo, da menor para a maior intensidade de exposição, para a secreção salivar de cortisol e alfa-amilase (proteínas de estresse metabólico agudo); esse efeito não foi visto para a secreção de imunoglobulina A. Assim, concluiu-se que exposição às radiofrequências de estações radiobase de telefonia celular de intensidade menor que a estabelecida pela diretriz da International Comission on NonIonizing Radiation Protection (ICNIRP) pode causar estresse fisiológico. Esta diretriz estabelece limites que consideram apenas os efeitos agudos, altos níveis de radiações eletromagnéticas não ionizantes e de curta duração (efeitos térmicos), desconsiderando, portanto, os efeitos crônicos de baixos níveis e de longa duração (efeitos não térmicos) 13. Lakimenko et al. 14 reproduziram os efeitos atérmicos das radiações emitidas por estações radiobase em células, utilizando radiofrequências de baixa intensidade, por um longo período de exposição, demonstrando aumento da desnaturação proteica de diversas proteínas citoplasmáticas, aumento da formação de espécies reativas de oxigênio, aumento de Ca2+ intracelular, dano ao DNA e inibição da reparação do DNA, alterações que podem gerar distúrbios metabólicos. O estudo concluiu que é equivocado relacionar os danos causados por essa radiação apenas ao fator térmico. Outros efeitos decorrentes da Silva DF et al. 2112 Cad. Saúde Pública, Rio de Janeiro, 31(10):2110-2126, out, 2015 exposição às radiações eletromagnéticas não ionizantes de telefonia celular, tais como neoplasias (ovário, mama, pulmão), distúrbio do sono, cefaleia, infertilidade, dentre outros são relatados na literatura 15,16,17. Entretanto, há estudos que não relatam 18 efeitos à saúde em populações expostas a radiações eletromagnéticas não ionizantes. Blettner et al. 19 não encontraram associação entre morar próximo de uma estação radiobase e aumento da incidência de câncer, concluindo que as emissões de radiações eletromagnéticas não ionizantes não estavam relacionadas com efeitos adversos para a saúde. Saravi 20 afirma que os dados não sugerem que as radiações eletromagnéticas não ionizantes de estações radiobase de telefonia celular apresentem riscos para saúde, embora considere que mesmo que os resultados existentes sejam conflitantes, parece clara a necessidade de realização de novos estudos desta e de outras fontes eletromagnéticas como as de rádio e de televisão. Verifica-se, portanto, que a relação entre exposição às radiações eletromagnéticas não ionizantes e aumento da incidência de agravos à saúde humana é um assunto controverso, necessitando de maior investigação por meio dos estudos epidemiológicos. Neste sentido, em resposta à preocupação pública e governamental, a Organização Mundial da Saúde (OMS) desenvolveu um projeto em 1996 para avaliar a evidência científica de possíveis efeitos adversos para a saúde, relacionados com as radiações eletromagnéticas não ionizantes. Em 2011, a OMS se posicionou quando a International Agency for Research on Cancer (IARC) 21 classificou a exposição às radiofrequências como pertencente ao grupo 2B, ou seja, categoria que classifica os agentes em possível carcinogênico. Diante do exposto, o objetivo do presente estudo foi investigar a associação entre exposição à radiação eletromagnética não ionizante de estações radiobase de telefonia celular e sintomas psiquiátricos.

The aim of this study was to investigate the association between exposure to non-ionizing electromagnetic radiation from mobile phone base stations and psychiatric symptoms.In a crosssectional study in Salvador, Bahia State, Brazil, 440 individuals were interviewed.Psychiatric complaints and diagnoses were the dependent variables and distance from the individual's residence to the base station was considered the main independent variable.Hierarchical logistic regression analysis was conducted to assess confounding.An association was observed between psychiatric symptoms and residential proximity to the base station and different forms of mobile phone use (making calls with weak signal coverage, keeping the mobile phone close to the body, having two or more chips, and never turning off the phone while sleeping), and with the use of other electronic devices.The study concluded that exposure to electromagnetic radiation from mobile phone base stations and other electronic devices was associated with psychiatric symptoms, independently of gender, schooling, and smoking status.The adoption of precautionary measures to reduce such exposure is recommended.

Introduction
The growth in mobile telephony and thus in the number of mobile phone base stations, which establish communications with mobile phones, has increasingly benefited contemporary lifestyle by facilitating communications, offering ease and comfort, providing the opportunity to remain connected to nearby and distant locations, and allowing Internet use for various purposes.However, mobile telephony has raised concerns over the possible health effects for populations exposed to non-ionizing electromagnetic radiation.Such radiation is characterized by its wavelength, frequency, and irradiated energy, and is considered not to carry sufficient energy to alter an atom's physical state 1 .
In order for mobile telephony communication to occur, the system is subdivided into cells.Each cell has an mobile phone base stations capable of sending power signals throughout its range.Each mobile phone base stations can serve several telephones at the same time, assigning each of them a narrow range of frequencies 2 .
Non-ionizing electromagnetic radiation is absorbed by the skin and by deeper levels of the body, dissipating repeatedly in depth, potentially causing a temperature increase not perceived by the body's natural thermal sensors (located superficially).The heat generated internally depends on exposure time, field intensity, and tissue thickness, and sometimes cannot be offset by the body, thus resulting in biological effects 3 .The depth of penetration of waves around 900MHz frequency used in mobile telephony, in tissues with high water content, like muscle, is 3cm.2,400MHz waves from microwave ovens penetrate some 1.7cm.In tissues with low water content, like bone, the depths are 17.7cm and 11.2cm, respectively 4 .
Most psychiatric disorders appear to originate from some combination of genetic and environmental (biological or psychosocial) factors 5 .One of the most extensively documented observations in epidemiological studies is the increased prevalence of anxiety and depression in women when compared to men 6 .The association between exposure to non-ionizing electromagnetic radiation/mobile telephony and these psychiatric effects has been investigated.Santini et al. 7 indicated more significant symptoms within a radius of 300m from the mobile phone base stations: irritability, depression, memory loss, dizziness, decreased libido, headache, sleep disorders, malaise (200m); and tiredness (300m).Seven symptoms were more significant in women: nausea, loss of appetite, visual disorders, depressive tendency, headache, insomnia, and malaise.Decreased libido was the most common complaint in men.According to Navarro et al. 8 and Bortkiewicz et al. 9 , individuals living close to the mobile phone base stations reported circulatory problems, sleep disorders, irritability, depression, blurred vision, and difficulties in concentrating.Meanwhile, Abdel-Rassoul et al. 10 suggested a relationship between living near a mobile phone bases station and neurobehavioral problems like depressive tendency, tremors, dizziness, headache, sleep disorders, and visual disorders, among others.Likewise, Oberfeld et al. 11 found that persons living close to base stations reported more symptoms of irritability, fatigue, headache, nausea, memory loss, visual disturbances, dizziness, and cardiovascular problems, directly proportional to their exposure to microwaves.
Augner et al. 12 studied exposure to mobile phone base stations radiofrequency in three groups of individuals.Each group was exposed to a 900MHz field, with variable exposure times.As exposure intensity increased, there was a significant increase in salivary secretion of cortisol and alpha-amylase (acute metabolic stress proteins); the same effect was not seen in the secretion of immunoglobulin A. The authors thus concluded that exposure to mobile phone base stations radiofrequencies at intensities lower than the guidelines of the International Commission on Non-Ionizing Radiation Protection (ICNIRP) can cause physiological stress.The guidelines set limits that only consider the acute effects, namely from high non-ionizing electromagnetic radiation levels and short duration (thermal effects), thus overlooking the chronic effects of low levels with long duration (non-thermal effects) 13 .
Lakimenko et al. 14 reproduced in cells the non-thermal effects of mobile phone base stations radiation using low-intensity radiofrequencies for long exposure times, demonstrating increased denaturation of various cytoplasmic proteins, increased formation of reactive oxygen species, increased intracellular Ca²+, DNA damage, and inhibition of DNA repair, alterations that can lead to metabolic disturbances.The study concluded that it is a mistake to explain the damage caused by this radiation, based exclusively on the thermal factor.Other effects of exposure to non-ionizing electromagnetic radiation/mobile telephony such as neoplasms (ovary, breast, lung), sleep disorders, headache, infertility, and others have been reported in the literature 15,16,17 .
However, some studies 18 have failed to identify negative health effects in populations exposed to non-ionizing electromagnetic radiation.Blettner et al. 19 found no association between living close to mobile phone base station and in-Cad.Saúde Pública, Rio de Janeiro, 31(10):1-16, out, 2015 creased cancer incidence, concluding that nonionizing electromagnetic radiation emissions were not related to adverse health outcomes.According to Saravi 20 , the data do not suggest that non-ionizing electromagnetic radiation/mobile phone base stations pose health risks; even so, they find that even though the results are conflicting, there appears to be a clear need for new studies on this and other electromagnetic sources such as radio and television.
Thus, the relationship between exposure to non-ionizing electromagnetic radiation and increased incidence of human health problems is a controversial topic, requiring further in-depth investigation in epidemiological studies.In this sense, in response to public and government concern, the World Health Organization developed a project in 1996 to assess the scientific evidence for possible adverse health effects related to non-ionizing electromagnetic radiation.In 2011, the World Health Organization (WHO) took a stance when the International Agency for Research on Cancer (IARC) 21 classified exposure to radiofrequencies as belonging to group 2B, the category that classifies agents as possible carcinogens.Given the above, the current study aimed to investigate the association between exposure to non-ionizing electromagnetic radiation from mobile phone base stations and psychiatric symptoms.

Methodology
This was a cross-sectional epidemiological study in two neighborhoods in Salvador, Bahia State, Brazil.By mapping the mobile phone base stations in this city 22 , a study area was defined based on the existence of at least one mobile phone base stations surrounded by urban occupation up to and beyond a 300m radius, as proposed by Santini et al. 7 .Since they met the established criteria, two adjoining neighborhoods in Salvador were selected for the study: individuals living in the neighborhoods of Bonfim and Monte Serrat, up to versus greater than 300m from mobile phone base stations, respectively.The mobile phone base stations reported in the study had been licensed for installation since 2004, operating with GSM transmission antennas, at a frequency of 1,800MHz, shared by three more operators in 2004, 2006, and 2008.Sample size was calculated (n = 474) using the following parameters: α = 5%; frequency of the health problem in the unexposed = 5%: prevalence ratio = 2; study power = 80%, and exposed/ unexposed ratio 1:1.However, due to difficulty in access and the exclusion of interviews held in households outside the established geographic limits, a total of 440 interviews were considered valid for analysis.
The following exclusion criteria were used for households: exclusively commercial buildings or institutions; households in which only individuals less than 18 years of age were home at the time of the intended interview; and homes exposed to non-ionizing electromagnetic radiation from other base stations.
A systematic random sample was taken.The first household was picked, after which every other household was selected.After the household was picked a resident 18 years or older was selected randomly for a face-to-face interview, conducted by trained interviewers.In addition to the instrument produced specifically for this study, a questionnaire was applied, previously validated for Brazil 23 , for screening symptoms of depression, namely the Center for Epidemiological Studies Depression Scale (CES-D).The scale is widely used to facilitate diagnosis in nonpsychiatric outpatient clinics, and especially in research.However, the scale's sensitivity and specificity vary in the literature.Risk of depression was defined as a CES-D score greater than or equal to 16 23 .
The dependent variables were evaluated by means of interviewees' self-reporting, including complaints of irritability, anxiety, decreased libido, palpitation, depressive tendency, and physician diagnoses of depression and anxiety and depressive symptoms (CES-D ≥ 16).The principal independent variable was exposure to nonionizing electromagnetic radiation from mobile phone base stations.Based on existing knowledge on exposure to non-ionizing electromagnetic radiation/mobile telephony and health effects, the following secondary variables were selected: exposure to non-ionizing electromagnetic radiation from mobile phones (years of use, number of mobile phones, number of chips, use with weak signal coverage, duration of calls, keeping the device close to the body, and never turning off the phone), electronic devices (microwave oven, landline telephone, router, devices kept in the bedroom), age, sex, income, schooling, alcohol consumption, and smoking.
After descriptive analyses, a hierarchical logistic regression model was used.We initially verified any association between independent variables using the chi-square test.Thus, the covariates age, alcohol consumption, income, years of mobile phone use, use of microwave oven, duration of mobile phone calls, and use of landline telephone were excluded because they showed strong associations between each other (p ≤ 0.05).We then structured hierarchical models, Cad.Saúde Pública, Rio de Janeiro, 31(10):1-16, out, 2015 preselecting variables whose association with the dependent variable showed p ≤ 0.25.The strategy for entering variables into the model was hierarchical, as proposed by Greenland 24 and Fuchs et al. 25 , maintaining at each level the variables with p ≤ 0.10 26 .
The first level included macro-social and lifestyle variables: sex, schooling (primary/secondary versus university/graduate school) and smoking.The second level included exposure to electronic devices: presence of devices in the bedroom (mobile phones, videogames, notebook, alarm clock/radio, computer); and router use in the home.The third level included exposure to non-ionizing electromagnetic radiation resulting from subjects' behavior in mobile phone use, that is, whether the individual: made mobile phone calls with weak signal coverage; used more than one mobile phone or more than one chip; turned the mobile phone off when asleep; kept the mobile phone close to the body.For inclusion in the fourth level, pertaining to exposure to non-ionizing electromagnetic radiation/mobile phone base stations, three models were constructed based on distance from the households and to the mobile phone base stations: 0-100; 101 to 200m; and > 200m from the mobile phone base stations; ≤ 300m versus > 300m from the mobile phone base stations; and ≤ 400 versus > 400m from the mobile phone base stations.These models were constructed because the antennas installed in the mobile phone base stations are tilted towards the ground in relation to the tower.Thus, at a given distance from the tower, the electromagnetic waves reach the ground with maximum intensity.Before and after this distance, the waves display significantly lower intensities 27 .In addition to distance, time residing in the vicinity (≤ 10 versus > 10 years) and time spent at home (8 to 16 versus 16.1 to 24 hours/day) were also incorporated at this level.The data were analyzed in Stata 10 (StataCorp LP, Colege Station, USA).The Hosmer-Lemeshow statistic 28

Results
The final study population consisted of 440 individuals, with a mean age of 48.4 years (range 18 to 87).Most were females (59.3%), and brown was the most common self-reported skin color (51.4%).More than half (60.2%) had finished secondary school and 39.8% had university or graduate degrees.One-fourth earned between one and two minimum wages, 9.1% smoked, and 40.7% consumed alcoholic beverages.
Most subjects had lived in the vicinity for more than ten years, with an average of 21.3 years, and 61.8% stayed at home between 8 and 16 hours a day.Some 40% lived at a distance of up to 300m from the mobile phone base stations, while the rest lived more than 300 meters away.In relation to mobile phone exposure, 90.9% were users, 31.3% had used mobile phones for more than ten years, had at least two chips (56.5%), were used to carrying their mobile phones close to the body (67.4%), never turned their mobile phones off while sleeping (82.5%), made calls totaling more than 30 minutes a day (28.6%), and used their mobile phone even with weak signal coverage (89.2%).
As for electronic equipment, 50.8% had a landline phone with a wireless receiver, kept electronic devices in the bedroom (80%), used a router at home (52.5%), and used a microwave oven (65.7%).Distribution according to these different variables was similar in the two neighborhoods (≤ 300m and > 300m from the mobile phone base stations) (Table 1).
For variables pertaining to exposure to nonionizing electromagnetic radiation from mobile phone use, an association was observed between having more than one mobile phone and depressive symptoms (CES-D ≥ 16) (OR = 1.63; 90%CI: 1.03-2.56)(Table 3).Using two or more chips and never turning off the mobile phone while sleep-    ing was associated with anxiety (Table 4).Keeping the mobile phone close to the body was associated with palpitation (Table 5).Using the mobile phone with weak signal coverage was associated with irritability (Table 5) and diagnosis of anxiety (Table 3).As for covariates related to exposure to electronic equipment, "keeping devices in the bedroom" was associated with irritability (Table 5) and diagnosis of anxiety (Table 3), while palpitation was significantly less common among individuals with a router at home (Table 5).An association was observed between female gender and eight psychiatric symptoms (irritability, palpitation, anxiety, depressive tendency, decreased libido, diagnosis of anxiety and depression, and depressive symptoms, CES-D ≥ 16) (Tables, 3, 4, and 5).

Discussion
The relationship between exposure to nonionizing electromagnetic radiation/mobile telephony and health effects has raised concern in the scientific community, as well as among policymakers, specifically related to populations around mobile phone base stations and mobile phone users.
The theme is highly relevant for public health and appears in the literature with conflicting results.This calls for adoption of the precautionary principle, a decision made when scientific information is insufficient, inconclusive, or uncertain and there are indications that the possible effects for the environment, human and animal health, or plant protection may be potentially hazardous and inconsistent with the chosen level of protection 29 .
The current study showed an association between exposure to non-ionizing electromagnetic radiation/mobile phone base stations for individuals living between 100 and 200m from the mobile phone base stations and diagnoses of anxiety and depression; those living up to 300m from the mobile phone base stations showed a significantly lower association with depressive symptoms (CES-D ≥ 16); and living up to 400m from the mobile phone base stations was asso-   ciated with depressive tendency and depressive symptoms (CES-D ≥ 16).Staying at home from 16.1 to 24 hours a day was associated with decreased libido in individuals living 100 to 200m from the mobile phone base stations.These findings are similar to those of other studies published in the literature.A French study 7 focused on 530 individuals living within a radius of 300m from the mobile phone base stations, and the significant symptoms according to distance from the source were: irritability, depression, memory loss, dizziness, decreased libido (100m); headache, sleep disturbance, malaise (200m), and tiredness (300m).Navarro et al. 8 indicated an association between exposure and depressive symptoms in the population 150m from the mobile phone base stations.Likewise, Bortkievicz et al. 9 , Abdel-Rassoul et al. 10 , and Hutter et al. 30 found more neurobehavioral effects in populations living near the mobile phone base stations when compared to those living more than 300m or 400m from the mobile phone base stations.Eger et al. 16 found that five years or more after installation and operation of the mobile phone base stations, the risk of developing cancer tripled for residents within less than 400m from the source, the main beam of which touched the ground at a horizontal distance of 350m.Wolf & Wolf 15 also indicated a fourfold increase in can-cer incidence in residents within 350m from the mobile phone base stations.
Antennas in mobile phone base stations are installed such that radiation emission is maximal when perpendicular to the antenna 27 .In relation to the tower, the antennas are tilted towards the ground, such that at a given distance from the tower the waves reach the ground with maximum intensity.Before and after this distance, the waves display significantly lower intensities.The terrain's topography, tower height, antennas' tilt, irradiated energy, and distance from the tower to the households are also preponderant factors in determining the region with the highest incidence of waves emitted by the mobile phone base stations, since the waves bounce off natural obstacles and buildings.Thus, households located between 100 to 200m from the tower show higher risk when compared to those located nearer (0 to 100m) and farther (> 200m) from the mobile phone base stations 27,31 .
One problem with exposure to non-ionizing electromagnetic radiation/mobile phone base stations has been the distance from the source to the households.For example, at 150m from a mobile phone base stations, typical values for an electric field (0.5 to 2V/m) 32 are considered low, but they represent continuous exposure.High radiation levels have been found 16   where the principal beam it reaches the ground, and in combination with local reflection, from that point onward; the intensity of the radiation decreases with the square of the distance from the point of the antenna feed.In addition, the principal radiation lobe is in the direction determined by the tilt's angle, while there are also side lobes arranged at other angles, thereby scattering the radiation.Hardell et al. 33 found an association between mobile phone use and risk of brain tumors.According to the authors, the results indicate that risk of exposure to non-ionizing electromagnetic radiation from mobile phones is lower in areas at a short distance from the mobile phone base stations, due to adaptive power control (APC).This corroborates the findings presented in other studies, by Lönn et al. 34 and Hillert et al. 35 , who demonstrated that APC in cellular phones is regulated by the distance between the mobile phone base stations.Thus, in areas with a long distance between mobile phone base stations (generally rural), the output power level is higher when compared to urban areas, where the distance between mobile phone base stations is shorter.

in areas around
In this sense, a possible explanation for the current study's findings, for example OR > 1 when evaluating psychiatric symptoms at distances between 100 and 200m, may be the exposure to non-ionizing electromagnetic radiation/mobile phone base stations.At a distance of 300m, the risk is expected to decrease, since there is less exposure to non-ionizing electromagnetic radiation/mobile phone base stations .This can be explained by the automatic control of gain or power irradiated by mobile phone devices, which decrease as they approach the mobile phone base stations, in agreement with some studies 33,34,35 .The risk increases again after 400m, possibly due to exposure to non-ionizing electromagnetic radiation from mobile phones which intensifies when the devices are farther from the mobile phone base stations (weak signal coverage).Automatic gain adjustment results in higher power and thus greater exposure of the user's head, resulting in higher OR, due to exposure to non-ionizing electromagnetic radiation from the mobile phone, not from the mobile phone base stations.
One way to estimate protective spacing of homes in neighborhoods around mobile phone base stations is to use the Friis equation to calculate exclusion zones, or areas in which the estimated level of the electric field may exceed the guidelines, considering not only the effects of high levels and short exposure times, like the ICNIRP standards, but also those that consider the effects of low levels and long exposure times.This procedure shows the region with reduction of possible risks of exposure from non-ionizing electromagnetic radiation/mobile phone base stations for the health of individuals living or working outside this region 36 .
For variables related to exposure to nonionizing electromagnetic radiation from mobile phones, having more than one mobile phone device was associated with depressive symptoms (CES-D ≥ 16), and using two or more chips and never turning the mobile phone off when asleep were associated with anxiety.Using the mobile phone with weak signal coverage was associated with diagnosis of anxiety and irritability.Coureau et al. 37 found a statistically significant association between using the mobile phone for more than two hours a day and brain tumors, concluding that the data reinforce the results of previous studies on intensive use of mobile phones and such tumors.Martin et al. 38 conducted measurements of the spectra of non-ionizing electromagnetic radiations in São José dos Campos and Taubaté, São Paulo State, Brazil, indicating that the sources with the heaviest pollution in frequencies ranging from 1MHz to 9.4GHz were mobile phones and VHF.Kuster & Kuhn 39 indicated that mobile phones are the strongest source of the brain's exposure to electromagnetic fields, and that exposure is influenced by the user's behavior and choice of the device.Important factors include the device's design and the use of earphones (which reduces exposure by ten times).
Other studies 40,41 have associated excessive and intensive use of wireless communication technologies with psychiatric effects.
From the physical point of view, when comparing the results of exposure to non-ionizing electromagnetic radiation from mobile phone base stations with that of mobile phones themselves, the individual's mobile phone behavior can certainly favor greater exposure.Using two or more chips, never turning the device off, keeping it close to the body, and using it in places with weak signal coverage can lead to greater exposure to non-ionizing electromagnetic radiation, when compared to exposure to radiation from mobile phone base stations.
The amount of energy one receives when talking for six seconds on a mobile phone 42 is equivalent to being exposed for 24 hours to a mobile phone base station at 100m, and when one holds a mobile phone device very close to the ear, the power level received is greater than when holding the device farther away.However, when signal coverage is poor, individuals tend to hold the device as close as possible, without realizing that this behavior increases the absorption of radiation.Using a mobile phone with weak signal coverage, a habit reported by most of our interviewees, is worrisome, since the reduced number of available channels due to overcrowding of users involves a higher field level and thus higher power in the device, which increases automatically in the attempt to locate another mobile phone base stations to keep the call going.Several studies 7,8,17 have indicated an association between exposure to non-ionizing electromagnetic radiation/ mobile phone base stations and health effects, only considering exposure to mobile phone base stations and overlooking exposure to non-ionizing electromagnetic radiation resulting from mobile phones use and thus APC.
Electronic devices in the bedroom (mobile phone, notebook, router, TV, etc.) can increase the risk of exposure to non-ionizing electromagnetic radiation, and in this study they were associated with irritability and diagnosis of anxiety.Palpitations were significantly less frequent in individuals that used a router in the home.For a distance up to 5m from the source (router), a low electric field is found (0.1 to 0.2V/m) 32 .
In this sense, despite the lack of a convincing explanation as to the biological plausibility of psychiatric effects from exposure to non-ionizing electromagnetic radiation, these findings should serve as a warning, especially due to the intensive use of mobile phone devices, where the electric field levels are much higher (10 to 150 /m, close to the head) 32 when compared to those from mobile phone base stations, which are located far from the individual's body.In addition, the results on exposure to non-ionizing electromagnetic radiation and depressive symptoms (CES-D ≥ 16) reinforce the findings, since CES-D has been validated in Brazil 23 .Nevertheless, the current study's findings should be viewed with caution, since this was a cross-sectional study, subject to reverse causality bias, given that information on exposure and outcome was obtained simultaneously.For example, do individuals that use mobile phones more tend to become more anxious, or does anxiety lead to greater use of the mobile phone?This reasoning can also be applied to other associations described above.
Despite adjustment, one cannot entirely rule out confounding.An example is possible confounding in the association between decreased libido and staying longer at home.Although the majority (54.1%) of the individuals were working, 7.5% were studying and 7.5% were unemployed (Table 1), and it is important to note that because they spend more time at home, they may have some clinical illness that causes limitations, or they may be on some medication, both of which are factors that can alter libido.It is important to identify the degree of difficulty in generalizing the findings to other populations, even within the city of Salvador.The two neighborhoods were not selected randomly, but as a convenience sample, since they displayed the necessary logistic and methodological conditions for conducting the research.Notwithstanding the important formal requisites of statistics, one can raise the hypothesis that the study's findings may be valid for similar populations, especially from the socioeconomic point of view.Important well-known associations, for example between smoking and lung cancer or between ionizing radiation and leukemia, were found in specific populations, but are valid for all human beings, due to a species issue, while respecting the various possibilities of effect modification.In addition, studies on distances from mobile phone base stations and especially the exposure to non-ionizing electromagnetic radiation from mobile phone devices should be conducted in the future, since the effects of non--ionizing electromagnetic radiation depend on the exposure field's characteristics.

Conclusions
Exposure to non-ionizing electromagnetic radiation related to living between 100 and 200m from a mobile phone base stations and individual behavior in the use of mobile phones and other electronic devices were associated with some psychiatric symptoms, after adjusting for various other potential confounding variables, especially gender.Despite the study's limitations and considering the controversies on the issue, described for the first time in Brazil based on a household survey and using a random sample, we recommend the adoption of precautionary measures to reduce the absorption of radiation from this type of exposure, especially in young people, namely: reduce the time per day spent on mobile phone calls, avoid using the mobile phone with weak signal coverage, refrain from keeping the mobile phone close to the body, use earphones or pop phones, avoid mobile phone use by children, whenever possible use the hardwire landline phone, and avoid residing or working within 200m of mobile phone base stations.The results can help improve public policies, in the sense of informing decision-making on risk evaluation, management, and communication aimed at individual health promotion.

Contributors
MPBS100-200: distance between 100 and 200m from the households and the mobile phone base stations; MPBS300: distance of 300m from the households and the mobile phone base stations; MPBS400: distance of 400m from the households and the mobile phone base stations.*Contextual variables;** Selected variables from level I and variables related to exposure to non-ionizing electromagnetic radiation from electronic devices; *** Selected variables from levels I and II and variables related to non-ionizing electromagnetic radiation from mobile phones; # Selected variables from levels I, II, and III and variables related to exposure to non-ionizing electromagnetic radiation/mobile phone base stations.
was used to test the model's goodness-of-fit.The research project complies with Resolution n. 196/96 of the Brazilian National Health Council and was approved by the Ethics Research Committee of the Gonçalo Moniz Research Center, Oswaldo Cruz Foundation (CPqGM/ Fiocruz), filed under protocol n. 358/2011 and final ruling n. 245/2011.All subjects signed a Free and Informed Consent Form.

Table 1
Distribution of individuals according to macro-social and lifestyle variables and exposure to non-ionizing electromagnetic radiation/mobile telephony.Salvador, Bahia State, Brazil, 2014.

Table 2 Prevalence
ratios (PR) for exposure to non-ionizing electromagnetic radiation/mobile telephony and psychiatric symptoms.Salvador, Bahia State, Brazil, 2014.95%CI: 95% confidence interval; CES-D: Centre For Epidemiological Studies Depression Scale.

Table 3
Hierarchical logistic regression model for variables associated with diagnoses of anxiety and depression and depressive symptoms (Center for EpidemiologicalStudies Depression Scale -CES-D ≥ 16) and exposure to non-ionizing electromagnetic radiation/mobile telephony, odds ratios (OR), and 90% confidence intervals (90%CI).Salvador, Bahia State, Brazil, 2014.

Table 3 (
continued) MPBS100-200: distance between 100 and 200m from the households and the mobile phone base stations; MPBS300: distance of 300m from the households and the mobile phone base stations; MPBS400: distance of 400m from the households and the mobile phone base stations.* Contextual variables; ** Selected variables from level I and variables related to exposure to non-ionizing electromagnetic radiation from electronic devices; *** Selected variables from levels I and II and variables related to non-ionizing electromagnetic radiation from mobile phones; # Selected variables from levels I, II, and III and variables related to exposure to non-ionizing electromagnetic radiation/mobile phone base stations.

Table 4
Hierarchical logistic regression model for variables associated with anxiety, depressive tendency, and decreased libido and exposure to non-ionizing electromagnetic radiation/mobile telephony, odds ratios (OR), and 90% confidence intervals (90%CI).Salvador, Bahia State, Brazil, 2014.
and the mobile phone base stations; MPBS400: distance of 400m from the households and the mobile phone base stations.*Contextualvariables;**Selected variables from level I and variables related to exposure to non-ionizing electromagnetic radiation from electronic devices; *** Selected variables from levels I and II and variables related to non-ionizing electromagnetic radiation from mobile phones; # Selected variables from levels I, II, and III and variables related to exposure to non-ionizing electromagnetic radiation/mobile phone base stations.Cad.Saúde Pública, Rio de Janeiro, 31(10):1-16, out, 2015

Table 5
Hierarchical logistic regression model for variables associated with irritability and palpitation and exposure to non-ionizing electromagnetic radiation/mobile telephony, odds ratios (OR), and confidence intervals (90%CI).Salvador, Bahia State, Brazil, 2014.