Effects of electromagnetic fields emitted by mobile phones (GSM 900 and WCDMA UMTS) on the macrostructure of sleep

Transcription

1 J. Sleep Res. (2011) 20, Mobile phones and sleep doi: /j x Effects of electromagnetic fields emitted by mobile phones (GSM 900 and WCDMA UMTS) on the macrostructure of sleep HEIDI DANKER-HOPFE 1, HANS DORN 1, ACHIM BAHR 2, PETER ANDERER 3 and CORNELIA SAUTER 1 1 Department of Psychiatry and Psychotherapy, Competence Center of Sleep Medicine and Sleep Research, Charité-Universitaetsmedizin Berlin, Campus Benjamin Franklin, Berlin, Germany, 2 IMST GmbH, Kamp-Lintfort, Germany and 3 Department of Psychiatry and Psychotherapy, Medical University of Vienna, Vienna, Austria Accepted in revised form 30 March 2010; received 01 July 2009 SUMMARY In the present double-blind, randomized, sham-controlled cross-over study, possible effects of electromagnetic fields emitted by Global System for Mobile Communications (GSM) 900 and Wideband Code-Division Multiple Access (WCDMA) Universal Mobile Telecommunications System (UMTS) cell-phones on the macrostructure of sleep were investigated in a laboratory environment. An adaptation night, which served as screening night for sleep disorders and as an adjustment night to the laboratory environment, was followed by 9 study nights (separated by a 2-week interval) in which subjects were exposed to three exposure conditions (sham, GSM 900 and WCDMA UMTS). The sample comprised 30 healthy male subjects within the age range years (mean ± standard deviation: 25.3 ± 2.6 years). A cell-phone usage at maximum radio frequency (RF) output power was simulated and the transmitted power was adjusted in order to approach, but not to exceed, the specific absorption rate (SAR) limits of the International Commission on Non-Ionizing Radiation Protection (ICNIRP) guidelines for general public exposure (SAR 10g = 2.0 W kg )1 ). In this study, possible effects of long-term (8 h) continuous RF exposure on the central nervous system were analysed during sleep, because sleep is a state in which many confounding intrinsic and extrinsic factors (e.g. motivation, personality, attitude) are eliminated or controlled. Thirteen of 177 variables characterizing the initiation and maintenance of sleep in the GSM 900 and three in the WCDMA exposure condition differed from the sham condition. The few significant results are not indicative of a negative impact on sleep architecture. From the present results there is no evidence for a sleep-disturbing effect of GSM 900 and WCDMA exposure. keywords GSM,, macrostructure of sleep, RF-EMF exposure, sleep architecture, sleep disturbance, UMTS INTRODUCTION Due to the enormous increase in the number of cell-phone users, concerns about possible health effects have been expressed by a substantial percentage of people. Annual Correspondence: Professor Dr Heidi Danker-Hopfe, Competence Center of Sleep Medicine and Sleep Research, Department of Psychiatry and Psychotherapy, Charité-Universitaetsmedizin Berlin, Campus Benjamin Franklin, Eschenallee 3, Berlin, Germany. Tel.: ; fax: ; heidi.dankerhopfe@charite.de surveys on attitudes regarding concerns about possible health risks of the electromagnetic fields (EMF) of mobile telecommunications document a robust percentage of 8 10% in a representative German population who fear adverse effects [Institute for Applied Social Sciences (INFAS), 2006]. Several studies during the last decade have dealt with the effects of EMF on brain activity during wake and sleep with divergent results (for reviews see Barth et al., 2008; Danker- Hopfe and Dorn, 2005; Sienkiewicz et al., 2005; Valentini et al., 2007). So far, all studies on the effects of the EMF of s on sleep have exposed their subjects to Ó 2010 European Sleep Research Society 73

2 74 H. Danker-Hopfe et al. Global System for Mobile Communications signals (GSM 900 or GSM 1800) (Borbely et al., 1999; Fritzer et al., 2007; Hinrichs et al., 2005; Huber et al., 2000, 2002; Hung et al., 2007; Loughran et al., 2005; Mann and Roschke, 1996; Regel et al., 2007; Wagner et al., 1998, 2000). In some of the studies subjects were exposed for 30 min prior to daytime (Huber et al., 2000; Hung et al., 2007) or prior to night sleep (Huber et al., 2002; Loughran et al., 2005; Regel et al., 2007), whereas in others, subjects were exposed during the entire night sleep (Borbely et al., 1999; Fritzer et al., 2007; Hinrichs et al., 2005; Mann and Roschke, 1996; Wagner et al., 1998, 2000) (see Table 1). The early finding of Mann and Roschke (1996), which showed a reduction in sleep latency and of percentage of rapid eye movement (REM sleep) (% of sleep period time) with exposure to GSM were not confirmed by the same research group in larger groups of subjects with different exposure with lower- or higher-power flux density (Wagner et al., 1998, 2000). The finding of the first Swiss study was a reduced wake after sleep onset time (Borbely et al., 1999), which was not confirmed by the following studies by that research group (Huber et al., 2000, 2002; Regel et al., 2007). The only study in which people of a broader age range, i.e years, took part revealed a reduced REM-sleep latency after 30 min exposure to a (Loughran et al., 2005). Studies so far on the effect of Wideband Code-Division Multiple Access (WCDMA) Universal Mobile Telecommunications System (UMTS) are missing. We present results of both types of s, i.e. GSM 900 and WCDMA (UMTS) on sleep. From a clinical viewpoint, and due to the intensive use of mobile telephony in the general population, it is important to gather and provide research data on the Table 1 Main results of studies on electromagnetic fields (EMF) effects on sleep architecture Changes with exposure n Sex Age (years) Signal type Exposure duration Authors Reduced sleep latency Reduced REM sleep (% SPT) 12 Males GSM 900 (217 Hz) No effect 22 Males GSM 900 (217 Hz) No effect 20 Males GSM 900 (217 Hz) Reduced WASO 24 Males GSM 900 (2, 8, 217, 1736 Hz) base station-like signal No effect 16 Males GSM 900 (2, 8, 217, 1736 Hz) base station-like signal No effect 16 Males GSM 900 (non-modulated continuous wave and pulse-modulated signal) No effect Females GSM 1800 (far-field characteristic, 1736 Hz pulse frequency) base station-like signal Reduced REM sleep latency Pronolonged sleep latency Females GSM 900 (217 Hz) 10 males GSM 900 (2, 8, 217 Hz) No effect 10 Males GSM 900 (2, 8, 217 Hz and 1736 Hz) No effect 15 Males GSM 900 (2, 8, 217 Hz and 1736 Hz) During 8 h bed time Mann and Roschke (1996) During 8 h bed time Wagner et al. (1998) During 8 h bed time Wagner et al. (2000) Alternating 15-min on 15-min off intervals during entire night sleep 30 min prior to a 3-h daytime sleep episode in the morning after 4 h of restricted night sleep 30 min prior to 8 h bed time Whole night 30 min prior to 7 h night sleep Borbely et al. (1999) Huber et al. (2000) Huber et al. (2002) Hinrichs et al. (2005) Loughran et al. (2005) 30 min prior to a 90-min Hung et al. (2007) sleep opportunity in the afternoon (13:30 h) after restriction of night sleep to 6 h 8 h Fritzer et al. (2007) 30 min prior to 8 h bed time Regel et al. (2007) GSM, Global System for Mobile Communications; REM, rapid eye movement; SPT, sleep period time; WASO, wake after sleep onset.

3 Electromagnetic fields of mobile phones and sleep 75 possible effects of electromagnetic fields emitted by mobile phones on human brain function during sleep, not only to concerned patients but also to physicians who see the patients complaining of sleep disturbances due to EMF. METHODS Subjects The sample comprised 30 healthy young male subjects [mean age ± standard deviation (SD): 25.3 ± 2.6, range years). All subjects, who had been recruited by newspaper advertising, underwent a detailed screening programme which began with a structured telephone interview on medical history and sleep habits. In a second step, questionnaires were used to check the exclusion criteria with regard to sleep quality (Pittsburgh Sleep Quality Index; Buysse et al., 1989; score > 5; PSQI-score: sample mean ± SD: 2.4 ± 1.3), excessive daytime sleepiness (Epworth Sleepiness Scale; Johns, 1991; ESS-score > 10; ESS-score: sample mean ± sd: 5.3 ± 2.7), anxiety (Zung Self Rating Anxiety Scale; Zung, 1971; SAS-score > 36, SAS-score: sample mean ± SD: 23.2 ± 2.9) and depression (Zung Self Rating Depression; Zung, 1965; SDS-score > 40, SDS-score: sample mean ± sd: 25.7 ± 3.5). Subjects who passed the questionnaire screening were invited for an extensive medical screening procedure, including a detailed medical interview, physical, psychiatric and neurological examination, electrocardiography (ECG), electroencephalogram (EEG) and blood tests. The subjects did not take any drugs and were non-smokers. Prior to the sleep laboratory nights, subjects were screened for sleep-related breathing disorders and sleep quality by ambulatory devices. Subjects proceeded with treatment nights only if the apnoea hypopnoea index (AHI) was <5 h and the periodic leg movements arousal index (PLMAI) was <10 h. For all subjects in the sample the AHI was <1 and the PLMAI was 0 for 26 subjects, and between 0.6 and 2.7 for the remaining four subjects. Of 46 subjects who underwent a screening adaptation polysomnography, four subjects failed to meet the inclusion criteria, and one declined further participation. During the progress of the study 10 subjects terminated prematurely due to job reasons or loss of interest. The study was conducted in accordance with the Declaration of Helsinki and the protocol was approved by the ethics committee of the Charite -Universitaetsmedizin Berlin. Written informed consent was given by all participants and they were paid for their participation. Protocol Each subject spent 10 nights in the sleep laboratory, separated by a 2-week interval (see Fig. 1). The first night served as an adaptation and screening night. The following 9 nights were treatment nights. During these nine study nights, GSM 900, WCDMA (UMTS 2000 MHz) and sham exposure were applied for three nights each in a randomized double-blind design. The exposure system was developed and installed by IMST GmbH, Kamp-Lintfort, in collaboration with the authors. Exposure The study was carried out in a shielded room to prevent interference to radio services and to minimize external signals. Radio frequency (RF) EMF (broadband 100 khz 3 GHz) were below 0.2 V m )1. Low-frequency fields at 50 Hz (power grid) were 1.0 V m )1 and nt; for other frequencies fields were 0.6 V m )1 and 22 nt. As audio amplifiers and other electronic devices are often disturbed by mobile phone RF signals transmitted in the proximity and would render blindness, a suitable type of computer display was selected, supplementary filter blocks were added and special RF immune devices were constructed. Other equipment, such as portable players, clocks, etc. were not permitted in the examination room. The system allowed double-blind exposure to sham, GSM 900 and WCDMA signals, respectively. A cushioned lightweight head-worn antenna connected to a thin microwave RF coaxial cable was used (see Fig. 2). Field distribution was both simulated and measured in a phantom. The maximum localized specific absorption rate (SAR) value was set to reach but not to exceed the limit of Figure 1. Schematic overview of the study design. The first night served as an adaptation and screening night. During the following nine study nights, Global System for Mobile Communications (GSM) 900, Wideband Code-Division Multiple Access (WCDMA) Universal Mobile Telecommunications System (UMTS) (2000 MHz) and sham exposure were applied for three nights each in a randomized double-blind design. AS: adaptation and screening night; N: night.

4 76 H. Danker-Hopfe et al. Polysomnography Figure 2. Subject with a head-worn antenna, which was designed especially for the present study, and which approximated the spatial field distribution of a common dual-band cell phone. 2Wkg )1 averaged over 10 g (ICNIRP, 1998). This corresponds to the upper limit of exposure while using a cellphone at maximum transmission power. Measurement of residual field strengths and monitoring of the system were performed by IMST GmbH independently from the local study team. The GSM signal consisted of a pulse-modulated 900 MHz carrier. Pulse width was 553 ls at a pulse frequency of 217 Hz. The Quadrature Phase Shift Keying (QPSK) modulated WCDMA signal at 1966 MHz, which was designed for investigations of biological effects, comprised a power control scheme with a repeated 1-min sequence consisting of 45 s toggling between 50% and 100% maximum output power every 0.67 ms and 15 s simulation of power fading with a dynamic range of 30 db (Ndoumbe Mbonjo Mbonjo et al., 2004). The dosimetry of the exposure system was carried out by measurements and calculations. Measurements were performed on a Specific Anthropomorphic Mannequin (SAM) head phantom, as used for compliance testing of mobile phones. Measurements revealed similar SAR distributions as published for mobile phones with integrated antennas for 900 MHz and in the MHz frequency range (Manteuffel et al., 2001). Using a finite-difference time-domain (FDTD) simulation, SAR distributions for both the phantom and the visible human head model were calculated. For the phantom, the measured 10 g averaged SAR result for 900 MHz agrees with the simulated value within 3%. For 1966 MHz, the measured and simulated SAR 10g agree within 22%. The simulated SAR 10g in the visible human head model agrees with measured values to within 20%. A variation of the antenna rotation angle by 15 results in a SAR 10g change below 17%. The increase of the antenna distance by 2 mm with respect to the human head leads to a SAR 10g change of 9% (for a detailed description of the exposure system and dosimetry, see Bahr et al., 2006, 2007). Three nights per exposure condition were recorded in order to reduce the variability between sleep nights. Sleep studies were recorded from 23:00 h to 07:00 h. Subjects were not allowed to fall asleep before Ôlights offõ. EEG signals were recorded from 19 gold electrodes fixed at positions Fp 1,Fp 2,F z,f 3,F 4,F 7, F 8,C z,c 3,C 4,T 3,T 4,P z,p 3,P 4,T 5,T 6,O 1 and O 2, according to the International system. The contralateral mastoid electrodes served as reference. The mental and submental chin electromyogram (EMG) were recorded with two bipolar channels. A horizontal and a vertical electro-occulogram (EOG) were recorded as bipolar derivations. All impedances of head electrodes were kept below 5 kw. Beyond the EEG, the EOG and the chin EMG the screening sleep studies comprised the recording of an ECG, an EMG of both legs (bipolar recording at the anterior musculus of both legs) and thoracic breathing excursions. Furthermore, a light sensor placed at the forehead was used. All signals were recorded digitally using a Varioport-System (Becker Meditec, Karlsruhe, Germany), configured especially for the present study. To prevent artefacts caused by spurious demodulation of the RF signals which could compromise blindness or could lead directly to false positive findings, the RF was kept away from the recorder electronics by several means. The battery-operated recorder was placed into a shielding aluminium cast case. All biosignal inputs were guided through shielding chambers with fifthorder LC low-pass filters. Remote control and marker signals were connected using optical cables. Optoelectronics were located inside the shielding. The effectiveness of the measures was checked by recordings under different exposure conditions and using a melon as a phantom. High-resolution spectra of the recordings showed the utility grid 50 Hz signal and weak disturbances from real low-frequency sources such as computer displays, but no signals resulting from exposure. Special electrode cables for the polysomnographic recording were designed to minimize disturbances of the RF field distribution and or additional RF currents injected at the electrode sites. Cables featured a distributed inductance and were made from thin stainless steel wire helix material protected and insulated by polyolefine tubing. The low-friction coefficient of the latter provided for handling convenience of the material. These special cables were also used to resolve RF interference issues on electronics in the study room. The sample rate for all channels was 200 Hz. All signals were lowpass filtered at a cutoff of 70 Hz. To avoid effects due to inter-rater variability, which would result from different human scorers, computer-assisted scoring (Somnolyzer 24 7) as described by Anderer et al. (2005), according to the instructions of Rechtschaffen and Kales (1968), was performed for all recordings (for a detailed description of the algorithm, see Anderer et al., 2005). The expert controlled analysis provided 177 variables to characterize the initiation and maintenance of sleep (27 variables), changes of sleep stages (25 variables), sleep cycles (six variables), sleep according to quarters of the night (17

5 Electromagnetic fields of mobile phones and sleep 77 variables for each quarter, e.g. duration of sleep stages in min and percentage of total sleep time; wake in min; total sleep time; sleep efficiency; number of awakenings) and sleep according to cycles of the night (17 variables each for the first cycle, the middle cycle and the last cycle). Statistical analysis A possible effect of whole-night exposure was analysed separately for GSM 900 and WCDMA exposure. To obtain robust parameters to characterize sleep under a specific exposure condition, the information was pooled by exposure condition and a median was used for further analyses instead of an arithmetic mean, as the latter can be highly biased by single deviating observations, i.e. outliers. Normal distribution of the data was tested with the Shapiro Wilk test. Depending upon the result the difference between exposure conditions was either tested by t-tests for dependent variables or with WilcoxonÕs matched-pairs signedranks test. The level of significance for two-sided tests was set at P < A Bonferroni-corrected significance level was applied in case of multiple tests (P < ). All statistical analyses were performed with sas version 9.1 for Windows. RESULTS GSM 900 exposure The results for conventional sleep parameters are shown in Table 2. Additional variables which showed significant differences between sham and verum exposure (GSM and WCDMA) are listed in Table 3. Overall, the distribution of 13 of the 177 (7.3%) variables investigated proved to be significantly different between GSM 900 and sham exposure; six referred to REM sleep, four to non-rem (NREM) sleep (Stage 2), two to movement time and one to the number of stage shifts from slow wave sleep to Stage 1, respectively. The duration of Stage 2 sleep during the whole night ()5.4 min; )1.1%) and especially during the third quarter of the night ()3.7 min; )3.8%) both in minutes and as percentage of total sleep time were significantly shorter in the exposure condition (see Tables 2 and 3, Figs 3 and 4). Conversely, the duration of REM sleep [whole night, third quarter of the night, in minutes and percentage of total sleep time (TST)] was increased significantly (see Tables 2 and 3, Figs 3 and 4). Furthermore, the average duration of all REM periods and of REM periods in the middle sleep cycles was longer in the exposure condition (see Table 3, Figs 3 and 4). The number of sleep stage changes from slow wave sleep to Stage 1 ()0.3) was decreased significantly (see Table 3, Fig. 5), and the movement time was increased in the first sleep cycle (0.4 min) and in the first quarter of the night (0.3 min; see Table 3, Fig. 5). After applying a Bonferroni correction for multiple tests, none of the P-values reported in Table 2 met the significance criterion afterwards (P < ). WCDMA exposure Only three variables (1.7%) showed deviating distributions under WCDMA compared to sham exposure. The average duration of the REM periods was increased by 1.3 min, the Table 2 Conventional parameters of the macrostructure of sleep: mean differences [± standard deviation (SD)] between Global System for Mobile Communications (GSM) 900 and sham exposure Variable GSM 900 Sham P (P75) (median) (P25) Sleep period time (SPT) (min) ± ± * )1.0 ± )1.0 )4.3 Total sleep time (TST) (min) ± ± * )1.1 ± )1.5 )10.7 Wake after sleep onset (min) 14.1 ± ± ± )3.8 Stage 1 latency (min) 9.7 ± ± * 0.2 ± )2.2 Sleep onset latency (three epochs S ± ± * 0.4 ± )2.0 or one epoch S2) (min) REM sleep latency (from sleep onset to 88.1 ± ± * )0.2 ± )13.0 first epoch of REM sleep) (min) Sleep efficiency (% TST TIB) 93.1 ± ± * )0.3 ± )0.1 )1.9 Stage 1 (min) 37.3 ± ± * )3.5 ± )3.9 )9.7 Stage 1 (% TST) 8.2 ± ± * )0.7 ± )0.8 )1.9 Stage 2 (min) ± ± * )5.4 ± )1.2 )17.0 Stage 2 (% TST) 48.5 ± ± * )1.1 ± )0.8 )3.0 Slow wave sleep (min) ± ± * 1.5 ± )3.3 Slow wave sleep (% TST) 24.3 ± ± * 0.5 ± )0.6 REM sleep (min) 86.5 ± ± * 6.3 ± )2.8 REM sleep (% TST) 19.0 ± ± * 1.4 ± )0.5, difference; P75, percentile 75; P25, percentile 25; REM, rapid eye movement; TIB, time in bed; positive differences indicate higher values under GSM 900 exposure. *Significance level of paired t-test. Significance level of WilcoxonÕs matched-pairs signed-ranks test.

6 78 H. Danker-Hopfe et al. Table 3 Additional significant parameters of the macrostructure of sleep: mean differences [± standard deviation (SD)] between Global System for Mobile Communications (GSM) 900 and sham exposure, and mean differences between Wideband Code-Division Multiple Access (WCDMA) and sham exposure Variable GSM 900 Sham P (P75) (median) (P25) Stage 2 (third quarter of the night) (min) 63.9 ± ± )3.7 ± )2.9 )6.3 Stage 2 (third quarter of the night) (% TST) 54.3 ± ± )3.8 ± 7.1 )0.4 )2.9 )5.1 Stage shifts from slow wave sleep to stage 1 (no.) 0.4 ± ± )0.3 ± )0.3 REM (third quarter of the night) (min) 26.6 ± ± * 3.5 ± )3.3 REM (third quarter of the night) (%) 22.5 ± ± * 2.8 ± )3.4 REM period (mean duration) (min) 25.3 ± ± * 2.3 ± )0.1 REM period (mean duration middle sleep cycle) 23.6 ± ± * 2.7 ± )1.3 (min) Movement time (first quarter of the night) (min) 1.2 ± ± * 0.3 ± )0.3 Movement time (first sleep cycle) (min) 1.2 ± ± * 0.4 ± )0.3 Variable WCDMA Sham P (P75) (median) (P25) Stage 2 (middle sleep cycle) (min) 52.2 ± ± )2.3 ± )3.2 )8.8 NREM period (mean duration) (min) 82.6 ± ± * )2.5 ± )3.3 )6.7 REM period (mean duration) (min) 24.3 ± ± * 1.3 ± )0.5 SD: standard deviation; : difference; P75: percentile 75; P25: percentile 25; NREM period: stage 1 + stage 2 + stage 3 + stage 4; positive differences indicate higher values under GSM 900 exposure. *Significance level of paired t-test. Significance level of WilcoxonÕs matched-pairs signed-ranks test erences [GSM Sham (min)] P = 0.038* P = 0.005* P = 0.011* P = P = 0.037* P = 0.008* erences [GSM Sham (%)] P = 0.019* P = P = 0.003* P = 0.016* 40 S2_TST S2_3Q REM_TST REM_3Q REM_MC REM_period Figure 3. Boxplot of the distribution of individually calculated differences [Global System for Mobile Communications (GSM) 900 exposure sham]: duration of non-rapid eye movement (NREM2) sleep (min) during the whole night [S2_ total sleep time (TST)] as well as in the third quarter of the night (S2_3Q), duration of REM sleep (min) during the whole night (REM_TST), the middle sleep cycles (REM_MC) as well as in the third quarter of the night (REM_3Q) and mean duration of REM cycles (REM_period). *Significance level of paired t-test; significance level of WilcoxonÕs matched-pairs signedranks test 40 S2P_TST S2P_3Q REMP_TST REMP_3Q Figure 4. Boxplot of the distribution of individually calculated differences [Global System for Mobile Communications (GSM) 900 exposure sham]: duration [% of total sleep time (TST)] of non-rapid eye movement (NREM2) and REM sleep for the whole night (S2P_TST; REMP_TST) as well as in the third quarter of the night (S2_3Q; REMP_3Q). *Significance level of paired t-test; significance level of WilcoxonÕs matched-pairs signed-ranks test. applied. None of the conventional sleep parameters was affected by WCDMA (see Table 4). mean duration of NREM periods was diminished by 2.5 min and the duration of NREM sleep in the middle sleep cycles was decreased by 2.3 min (see Table 3 and Fig. 6). These results are no longer significant when a Bonferroni correction is DISCUSSION The sleep wake cycle, as well as sleep structure, reflect the spontaneous activity of autoregulatory central nervous processes (Anderer et al., 2006). In the present study sleep was

7 Electromagnetic fields of mobile phones and sleep 79 erences [GSM Sham (number)] P = P = 0.025* P = 0.047* SWS_S1 MT_1C MT_1Q Figure 5. Boxplot of the distribution of individually calculated differences [Global System for Mobile Communications (GSM) 900 exposure sham]: number of stage shifts from slow wave sleep (SWS) to non-rapid eye movement (NREM1) sleep (SWS_S1) and movement time (MT) in the first sleep cycle (MT_1C) and in the first quarter of the night (MT_1Q). *Significance level of paired t-test; significance level of WilcoxonÕs matched-pairs signed-ranks test. erences [WCMDA Sham (number/min)] P = 0.032* P = 0.049* P = NREM_period REM_period S2_MC Figure 6. Boxplot of the distribution of individually calculated differences [Wideband Code-Division Multiple Access (WCDMA) exposure ssham]: mean duration (min) of non-rapid eye movement (NREM) periods (NREM_period) and REM period (NREM_period) and duration of NREM2 sleep (min) in the middle sleep cycles (S2_MC). *Significance level of paired t-test; significance level of WilcoxonÕs matched-pairs signed-ranks test. used as a model for investigating the effect of electromagnetic fields from cell-phones on the central nervous system (CNS). The specific advantage of the physiological state of sleep for this kind of research results from the fact that sleep is a reversible state, which is characterized by uncoupling of perception, and where the subject is comparatively insensitive to environmental and internal stimuli. Thus, many factors which have to be considered as confounders in research on EMF effects on the CNS in the waking state are controlled for. Unlike most of the previous research the present exposure was applied for 8 h, from Ôlights offõ to Ôlights onõ at an intensity approaching but not exceeding the upper limit given by the International Commission on non-ionizing radiation protection (ICNIRP) guidelines for general public exposure (International Commission on Non-Ionizing Radiation Protection (ICNIRP)., 1998). Furthermore, to our knowledge, this study examines for the first time the effect of WCDMA exposure on sleep. A further differentiating factor is the number of investigated variables, which in previously published papers ranged from one (Borbely et al., 1999; Hung et al., 2007) to 14 (Fritzer et al., 2007). Most of the studies published on GSM exposure did not find a significant EMF effect on variables characterizing macrostructural aspects of sleep architecture. Isolated effects observed so far in different publications affect sleep-onset latency, REM sleep latency and duration and wake after sleep onset (see Table 1). The sample size of 30 would have been sufficient to test all but one of the standardized effect sizes derived from the significant results listed in Table 1 with P < 0.05 and a power of The only exception is the estimated standardized effect size of 0.35 derived from the Australian study (Loughran et al., 2005). However, in the present study, sleep onset latency, REM sleep latency and wake after sleep onset, which proved to be affected significantly by EMF in some previous studies (see Table 1), did not vary with exposure status (see Table 2). Of the previously relevant variables of sleep structure, only REM sleep duration showed a significant difference between sham and GSM 900 exposure. However, in contrast to the REM suppressive effect of EMF reported by Mann and Roschke (1996), REM sleep duration was increased in the present study. Because (1) the six REM sleep-related variables for which significant differences between sham and verum exposure have been observed in the present study are not independent; (2) these variables are different from those reported to be affected in other studies; (3) REM sleep duration even shows an inverse effect; and (4) two of the variables do not seem to have a physiological meaning (changes in movement time, which is a rare state, refer to a difference of less than 0.5 min), it is assumed that the results are due probably to chance unless replicated in an independent study. It appears that the macrostructural level of sleep analysis under GSM 900 exposure does not lead to effects which can explain sleep complaints that are assigned frequently to EMF exposure in the general population. The statistically significant changes in Stage 2 sleep mirror the results for REM sleep, e.g. the decreases in Stage 2 sleep (whole night and third quarter of the night) parallel the increases in REM sleep. The fact that the increase in REM sleep does not lead to an increase of Stage 1 sleep and or wake after sleep onset underlines that GSM 900 exposure does not result in disturbed sleep. Together with the observation that the number of sleep stage changes from slow wave sleep to Stage 1 sleep is decreased significantly, the results could be interpreted if the data need to be

8 80 H. Danker-Hopfe et al. Table 4 Conventional parameters of the macrostructure of sleep: mean differences (± standard deviation (SD)] between Wideband Code- Division Multiple Access (WCDMA) and sham exposure Variable WCDMA Sham P (P75) (median) (P25) Sleep period time (SPT) (min) ± ± )2.7 ± )1.1 )6.0 Total sleep time (TST) (min) ± ± )2.2 ± )0.7 )8.5 Wake after sleep onset (min) 16.1 ± ± )0.5 ± )0.3 )3.5 Stage 1 latency (min) 10.8 ± ± ± )2.7 Sleep onset latency (three epochs S1 or 11.9 ± ± ± )3.2 one epoch S2) (min) REM sleep latency (from sleep onset 85.1 ± ± * )3.2 ± )2.9 )13.2 to first epoch of REM sleep) (min) Sleep efficiency (% TST TIB) 93.1 ± ± * )0.3 ± )1.8 Stage 1 (min) 39.2 ± ± )1.5 ± )6.2 Stage 1 (% TST) 8.7 ± ± )0.3 ± )1.3 Stage 2 (min) ± ± * )4.0 ± )3.6 )10.7 Stage 2 (% TST) 48.9 ± ± * )0.6 ± )1.7 Slow wave sleep (min) ± ± * )0.2 ± )1.0 )9.0 Slow wave sleep (% TST) 23.9 ± ± * 0.1 ± )2.1 REM sleep (min) 83.7 ± ± * 3.5 ± )3.7 REM sleep (% TST) 18.4 ± ± * 0.8 ± )0.7 : difference; P75: percentile 75; P25: percentile 25; REM: rapid eye movement; TIB: time in bed; positive differences indicate higher values under GSM 900 exposure. *Significance level of paired t-test. Significance level of WilcoxonÕs matched-pairs signed-ranks test. interpreted physiologically at all as pointing towards a sleep-consolidating effect of EMF. From this laboratory study there is no evidence for a sleep-disturbing effect of short-term GSM 900 exposure. This conclusion corresponds with the direction of effects seen in almost all other studies, and have been interpreted formerly as slightly sleep-promoting (Mann and Roschke, 2004). As mentioned above, this is the first study on the effect of WCDMA exposure on sleep. Overall, the number of significant differences (three) is much lower compared to GSM 900 exposure. The increase in the duration of REM periods was 1.3 min, and the mean duration of NREM periods decreased by 2.5 min, which was due mainly to the decrease of Stage 2 sleep spent in the middle NREM sleep cycle ()2.3 min). Overall, the observed differences between sham and WCDMA exposure in the sleep macrostructure cannot explain a subjectively perceived sleep-disturbing effect from this kind of signal. However, none of the P-values reported in Tables 2 and 3 meet the significance criterion after Bonferroni correction for multiple tests. In conclusion, it is noteworthy that neither electromagnetic fields from GSM900 nor from WCDMA UMTS mobile phones influenced the sleep macrostructure in healthy young male subjects. Given the number of subjects who claim that electromagnetic fields of mobile phones are the cause of subjectively perceived sleep disturbances, the present findings are of clinical importance. The sleep architecture of patients who suffer from insomnia is changed, and marked by an increase in sleep latency in wake after sleep onset time and in Stage 1, and by a decrease in total sleep time and or sleep efficiency. None of these alterations were observed in subjects exposed to electromagnetic fields of GSM 900 or WCDMA UMTS mobile phones. DECLARATIONS OF INTEREST All authors have indicated no conflicts of interest. ACKNOWLEDGEMENTS We would like to thank all participants and Blanka Pophof from the BfS for her expert monitoring. The study was funded by the Federal Office for Radiation Protection (BfS) within the German Mobile Telecommunication Research Programme (DMF). REFERENCES Anderer, P., Gruber, G., Parapatics, S. et al. An E-health solution for automatic sleep classification according to Rechtschaffen and Kales: validation study of the Somnolyzer 24 7 utilizing the Siesta database. Neuropsychobiology, 2005, 51: Anderer, P., Saletu, B., Saletu-Zyhlarz, G. et al. Electrophysiological evaluation of sleep. In: T Kinoshita (Ed.) Textbook for the Training Course of the International Pharmaco-EEG Society. Awaji Island of Hyogo Prefecture, Japan, 2006: Bahr, A., Dorn, H. and Bolz, T. Dosimetric assessment of an exposure system for simulating GSM and WCDMA mobile phone usage. Bioelectromagnetics, 2006, 27: Bahr, A., Adami, C., Bolz, T., Rennings, A., Dorn, H. and Ruttiger, L. Exposure setups for laboratory animals and volunteer studies using body-mounted antennas. Radiat. Prot. Dosimetry, 2007, 124:

9 Electromagnetic fields of mobile phones and sleep 81 Barth, A., Winker, R., Ponocny-Seliger, E. et al. A meta-analysis for neurobehavioural effects due to electromagnetic field exposure emitted by GSM mobile phones. Occup. Environ. Med., 2008, 65: Borbely, A. A., Huber, R., Graf, T., Fuchs, B., Gallmann, E. and Achermann, P. Pulsed high-frequency electromagnetic field affects human sleep and sleep electroencephalogram. Neurosci. Lett., 1999, 275: Buysse, D. J., Reynolds, C. F., III, Monk, T. H., Berman, S. R. and Kupfer, D. J. The Pittsburgh Sleep Quality Index: a new instrument for psychiatric practice and research. Psychiatry Res., 1989, 28: Danker-Hopfe, H. and Dorn, H. Biological effects of electromagnetic fields at mobile phone frequencies on sleep: current state of knowledge from laboratory studies. Somnologie, 2005, 9: Fritzer, G., Goder, R., Friege, L. et al. Effects of short- and long-term pulsed radiofrequency electromagnetic fields on night sleep and cognitive functions in healthy subjects. Bioelectromagnetics, 2007, 28: Hinrichs, H., Heinze, H.-J. and Rotte, M. Human sleep under the influence of a GSM 1800 electromagnetic far field. Somnologie, 2005, 9: Huber, R., Graf, T., Cote, K. A. et al. Exposure to pulsed highfrequency electromagnetic field during waking affects human sleep EEG. Neuroreport, 2000, 11: Huber, R., Treyer, V., Borbely, A. A. et al. Electromagnetic fields, such as those from mobile phones, alter regional cerebral blood flow and sleep and waking EEG. J. Sleep Res., 2002, 11: Hung, C. S., Anderson, C., Horne, J. A. and McEvoy, P. Mobile phone Ôtalk-modeÕ signal delays EEG-determined sleep onset. Neurosci. Lett., 2007, 421: Institute for Applied Social Sciences (INFAS). Identifying the general publicõs fears and anxieties with regard to the possible risks of high frequency electromagnetic fields of mobile telecommunications (survey 2003, 2004, 2005, 2006). Complete report [in German] Available at: html (accessed 30 June 2009). International Commission on Non-Ionizing Radiation Protection (ICNIRP). ICNRP Guidelines for limiting exposure to time-varying electric, magnetic, and electromagnetic fields (up to 300 GHz). International Commission on Non-Ionizing Radiation Protection. Health Phys., 1998, 74: Johns, M. W. A new method for measuring daytime sleepiness: the Epworth sleepiness scale. Sleep, 1991, 14: Loughran, S. P., Wood, A. W., Barton, J. M., Croft, R. J., Thompson, B. and Stough, C. The effect of electromagnetic fields emitted by mobile phones on human sleep. Neuroreport, 2005, 16: Mann, K. and Roschke, J. Effects of pulsed high-frequency electromagnetic fields on human sleep. Neuropsychobiology, 1996, 33: Mann, K. and Roschke, J. Sleep under exposure to high-frequency electromagnetic fields. Sleep Med. Rev., 2004, 8: Manteuffel, D., Bahr, A., Waldow, P. and Wolff, I. Numerical analysis of absorption mechanisms for mobile phones with integrated multiband antennas. Proceedings of the AP-S-International Symposium on Antennas and Propagation. IEEE, Boston MA, USA, 2001: 3, Ndoumbe Mbonjo Mbonjo, H., Streckert, J., Bitz, A. et al. Generic UMTS test signal for RF bioelectromagnetic studies. Bioelectromagnetics, 2004, 25: Rechtschaffen, A. and Kales, A. A Manual of Standardized Terminology, Techniques and Scoring System for Sleep Stages of Human Subjects. University of California, Brain Information Service Brain Research Institute, Los Angeles, Regel, S. J., Tinguely, G., Schuderer, J. et al. Pulsed radio-frequency electromagnetic fields: dose-dependent effects on sleep, the sleep EEG and cognitive performance. J. Sleep Res., 2007, 16: Sienkiewicz, Z., Jones, N. and Bottomley, A. Neurobehavioural effects of electromagnetic fields. Bioelectromagnetics, 2005, (Suppl. 7): S Valentini, E., Curcio, G., Moroni, F., Ferrara, M., De Gennaro, L. and Bertini, M. Neurophysiological effects of mobile phone electromagnetic fields on humans: a comprehensive review. Bioelectromagnetics, 2007, 28: Wagner, P., Roschke, J., Mann, K., Hiller, W. and Frank, C. Human sleep under the influence of pulsed radiofrequency electromagnetic fields: a polysomnographic study using standardized conditions. Bioelectromagnetics, 1998, 19: Wagner, P., Roschke, J., Mann, K. et al. Human sleep EEG under the influence of pulsed radio frequency electromagnetic fields. Results from polysomnographies using submaximal high power flux densities. Neuropsychobiology, 2000, 42: Zung, W. W. A Self-Rating Depression Scale. Arch. Gen. Psychiatry, 1965, 12: Zung, W. W. A rating instrument for anxiety disorders. Psychosomatics, 1971, 12: