Research articles

By Dr. Giovanna Tranfo , Dr. Daniela Pigini , Mr. Vincenzo Brugaletta , Dr. Giancarlo Burriesci , Dr. Rosaria Falsaperla , Dr. Paolo Rossi , Dr. Floriana Sacco , Dr. Renata Sisto
Corresponding Author Dr. Giovanna Tranfo
INAIL - Monteporzio Catone Research Center, - Italy 00040
Submitting Author Dr. Giovanna Tranfo
Other Authors Dr. Daniela Pigini
ISPESL - Occupational Hygiene, Via Fontana Candida 1, Monteporzio Catone (RM) - Italy 00040

Mr. Vincenzo Brugaletta
INAIL - Monteporzio Catone Research Center, Via di Fontana Candida 1, Monteporzio Catone (RM) - Italy 00040

Dr. Giancarlo Burriesci
INAIL - Monteporzio Catone Research Center, Via di Fontana Candida 1, Monteporzio Catone (RM) - Italy 00040

Dr. Rosaria Falsaperla
INAIL - Monteporzio Catone Research Center, Via di Fontana Candida 1, Monteporzio Catone (RM) - Italy 00040

Dr. Paolo Rossi
INAIL - Monteporzio Catone Research Center, Via di Fontana Candida 1, Monteporzio Catone (RM) - Italy 00040

Dr. Floriana Sacco
INAIL - Monteporzio Catone Research Center, Via di Fontana Candida 1, Monteporzio Catone (RM) - Italy 00040

Dr. Renata Sisto
INAIL - Monteporzio Catone Research Center, Via di Fontana Candida 1, Monteporzio Catone (RM) - Italy 00040


Melatonin, Cortisol, Urine, Saliva, GSM, Mobile Phones, Exposure, Volunteers, non ionizing radiations

Tranfo G, Pigini D, Brugaletta V, Burriesci G, Falsaperla R, Rossi P, et al. Measures of melatonin and cortisol variations in volunteers exposed to GSM cellular phones in a double blind experiment. WebmedCentral ENVIRONMENTAL MEDICINE 2010;1(9):WMC00634
doi: 10.9754/journal.wmc.2010.00634
Submitted on: 15 Sep 2010 01:17:27 PM GMT
Published on: 15 Sep 2010 05:42:14 PM GMT


Melatonin and cortisol are two hormones produced according to a circadian rhythm. Cortisol is secreted in pulsatile manner, while melatonin blood concentration is influenced by the daylight; both are considered as stress markers. Animal studies suggest that exposure to ELF magnetic fields and radiofrequencies may influence the secretion of melatonin and cortisol, with possible health implications. To investigate the effect to GSM 900 MHz mobile phones, twenty-six volunteers were studied during three different days but at the same time of the day: no exposure, 30 minutes of exposure to telephone A (on) and 30 minutes exposure to telephone B (modified with connection of GSM signal to an internal load, to get a sham exposure). Urine and saliva samples were collected: 6-Sulphatoxymelatonin , the urine metabolite, showed a wide individual and seasonal variability. The differences among no exposure, exposure A and B were not statistically significant for cortisol and melatonin levels determined in saliva ,(correlated to the blood levels), that therefore appear not influenced by 30 minutes exposure to mobile phones . No statistically significant effect due to cell phone exposure was found in the data both for melatonin and cortisol saliva concentration.


Melatonin is an endogenous hormone which is synthesized in the pineal gland and secreted into the blood: it is well known as a key substance in promoting the falling into night sleep by humans, and has a fundamental role as photoneuroendocrine transducer of information on day length [1, 2]. Blood melatonin concentration in adult humans subjects is extremely low during daytime and increases rapidly at 11 – 12 p.m.up to as much as ten or twenty times daytime values. The high level is maintained till the early morning and then it decreases back to the baseline concentration. Melatonin is metabolized primarily in the liver via a cytocrome P450-mediated hydroxylation into the 6-hydroxymelatonin that is then conjugated with sulphate to form 6-sulphatoxymelatonin and eliminated in the urine [3]. 6-Sulphatoxymelatonin acrophase is reached approximately 2 hours after the plasma melatonin acrophase [4]. Cortisol is the primary glucocorticoid produced in the zone file of the adrenal cortex and secreted in pulsatile manner under the control of ACTH. In physiological conditions cortisol is produced according to a circadian rhythm, characterized by mass production in the early hours of the morning. Free cortisol, that is the amount not bonded to proteins, is released in saliva, whose cortisolo level is correlated to the plasma concentration [5, 6]. The stress is able increase the levels of cortisol, and after some day of loss of sleep [7]. Melatonin also appears to be and effective anthihypertensive agent in animals [8]. On the basis of experimental studies that showed reduction in melatonin concentration in animals exposed to ELF fields and of epidemiological studies in humans suggesting an association between ELF field exposure and cancer, it has been hypothesized over the years that exposure to ELF may alter the pineal gland function reducing the nocturnal increase in melatonin biosynthesis and release[9], but also contradictory results ware published [10, 11]. There is evidence that melatonin is able to neutralize free radicals, and this activity has been linked to the control of tumour production and growth: several studies have been carried out in order to confirm this hypothesis but results are not definitive [12, 13].
The rapid growth of mobile telecommunications and the use of mobile phones working at the frequency of 900 and 1800 MHz increased the public concern on the possible adverse health effects of radiofrequencies and the research needs have been addressed on mobile telephones and human reactions to RF exposure. Healthy individuals are to be exposed to RF fields in controlled laboratory studies to determine whether they are sensitive to these fields [14].
The effect of the exposure to the radiation emitted by GSM mobile telephones has been investigated by measuring serum melatonin in animals [15, 16] and serum melatonin and its urinary metabolite in humans [17, 18], and serum cortisol [19, 20, 21] but results show slight or not significant differences between exposure and sham conditions. The present experiment, conducted in double blind, was performed on volunteers to further investigate the effect of exposure to GSM 900 MHz mobile phones, by measuring salivary melatonin and cortisol during exposure, known to be correlated with the blood level [22], and urinary metabolite 6-sulphatoxymelatonin after the exposure was finished.

Materials and Methodology

In this experiment 26 healthy volunteers participated in the study, 11 females and 15 males. They were asked to fill in a questionnaire in order to collect personal information. Age was between 23 and 51 years. 5 subjects were smokers. Exclusion criteria were: night or shift work, usual exposure to electromagnetic fields, recent long distance flights, regular medical therapy, use of melatonin as dietary supplement.
Exposure parameters
The exposure system consisted of two Motorola Timeport 8900 dual-band (900 and 1800 MHz) GSM mobile phones, called A and B. Telephones were commercially identical, but B was modified by connecting the output of GSM power signal to an internal load instead than into antenna. This implied that, when the telephone was on, the internal circuitry was regularly active, but RF power delivered in space was negligible, around 30 dB lower than phone A. Such condition identified a true SHAM exposure with respect to the GSM signal; nevertheless the weak magnetic fields produced by circuitry and battery currents were identical in A and B. Both telephones were set at 900 MHz at the maximum power (about 2 W). The maximum local SAR value produced by telephone A was 0.5 ± 1 W/kg.
The document [23] recommends both the use of a SHAM exposure and the study blinding. In this experiment nor the volunteers nor the technicians who tested the biological samples were aware of which telephone was emitting the radiofrequency (A or B). Volunteers were sitting and the telephones were attached to a plastic helmet that could be comfortably worn for the duration of exposure.
Phone exposure protocol
Volunteers were examined during three days within two weeks:
- day1: no exposure.
- day 2 and day 3: 30 min exposure to telephone A or B.
Day 2 and 3 were never consecutive, and the sequence A and B was randomized as much as possible (the randomization in exposure presentation is needed to avoid systematic effect due to the fact that the subject could differently react to the first or the second stimulus).
15 volunteers, 5 females and 10 males (morning group) were exposed for 30 minutes from 11:00 to 11:30, while the remaining 11, 6 females and 5 males (evening group) were exposed for 30 minutes from 18:00 to 18:30.
Sampling protocol
All volunteers were asked to collect 4 saliva samples starting 1 hour before the beginning of the exposure (at 10:00 for the morning group and at 17:00 for the evening group), and then at intervals of 1 hour (at 11:00, 12:00 and 13:00 for the morning group and 18:00, 19:00 and 20:00 for the evening group). In addition, all volunteers provided a urine sample at 19:00 of the exposure days and another one at 7:00 of the following day (same hours for both groups). In the day of no exposure (day1) volunteers provided saliva and urine samples following the same protocol of the exposure days. Urine samples were collected in sterile polypropylene containers and stored at 4°C until the next day, and then frozen at -20° C until assay. Volunteers were also asked not to collect saliva samples within 15 minutes after brushing teeth or use of dental floss, nor within 30 minutes after eating or drinking or gum chewing, and to avoid bananas, coffee, alcohol and sport in the sampling days. Saliva samples were collected using the collection devices Salicap® Set from IBL International (Hamburg, Germany) consisting of 1 mL polypropylene tubes and polypropylene straws. Saliva samples were stored at 4°C until the day following the collection and then frozen at -20° until assay.
Urine and saliva assay
The urine samples have been tested for their concentration of 6-sulphatoxymelatonin (6-OHMS) by means of the Melatonin Sulfate Elisa assay provided by IBL International (Hamburg, Germany), and for their creatinine concentration by means of a colorimetric test kit (PKL® Poker Italia s.r.l., Genova Italy). Results have been expressed as ng/mg of creatinine in order to normalize results with respect to the urine dilution.. The saliva samples have been tested for their concentration of melatonin and cortisol. For melatonin the Direct Saliva Melatonin ELISA test kit provided by Bühlmann Laboratories AG (Schönenbuch,CH) was used and the results are expressed in pg/mL. Cortisol ELISA test kit (IBL-Hamburg) was used to determine cortisol on the saliva samples and the result is expressed as ng/mL. The performance characteristics of both tests are reported in Illustration 1. Assays were performed blind to the exposure status.

Results and discussion

Evening exposure – saliva results
The evening exposure experiment, was designed in order to gather melatonin levels in a period of biosynthetic activity of the pineal gland, when melatonin secretion begins. The saliva melatonin concentrations at the four sampling times (exposure starting at time 1) are reported in Illustration 2 for 11 volunteers. Both median values are reported in Illustration 3. The saliva test demonstrated that even during the late afternoon the melatonin secretion is so low: that a suppression effect cannot be demonstrated in this experiment. For this reason and as volunteer recruiting is very difficult in these hours. we decided to continue the experiment in the morning hours.
The values of the mean and median concentrations of salivary cortisol for 11 volunteers in the experiment of exposure of late afternoon show no significant difference between control and exposure, between control and sham, and between exposure and sham (Illustration  4 and 5).
Morning exposure – saliva results
The morning exposure experiment was performed on 15 additional volunteers. The saliva melatonin concentration at the four sampling times are reported in Illustration 2. The median values are also reported (Illustration 6). The values at time 0 are higher than in the evening experiment as the level is still decreasing from the night peak, and during all the experiment time values stay higher than in the evening, but again no suppression effect seems to be evidenced for the exposure compared to sham.
The results of the experiment of exposure in the morning, the period in which the level of cortisol decreases, are reported in Illustration 4 and the median values are shown graphically (Illustration 7). Also in this case no effects of suppression correlated to exposure are evidenced.
Urine analysis
The urine samples for the melatonin metabolite assay were collected at the same time for all volunteers, both participating in the morning and in the evening experiment. The main objective of this test was to assess the eligibility of the subject for the study, that is to show a detectable metabolite excretion and ratio between the metabolite concentration in the morning samples with respect to the night one always higher than 1. The values reported in Illustration 8 show that there is a large individual variability for the metabolite concentration in the urine both at 7 p.m. and at 7 a.m. and more in the morning-to-night ratio, even if normalized for the creatinine concentration of the samples. Moreover a seasonal variability is known to exist [1, 2]. The 6-OHMS concentration reported in order of sampling date, for two volunteers who collected seven urine samples at 7 the period from March to November in the same conditions than the “no exposure day” of the study is shown (Illustration 9).
However, other authors found a statistically significant lower amounts of 6-OHMS normalized for the urine creatinine concentration excreted in the “pre-bed-time” samples from subject exposed to mobile phones compared to subject exposed to sham [12]: in our case the mean values of 6-OHMS in the 7 p.m. urine samples for the exposure experiment are not different from the sham and the control median values reported in Illustration 8, even if morning and evening exposure experiments are separated (Illustration 10).
Statistical analysis
Statistical analysis was performed on a personal computer using R software ((C) R Foundation, [23].
The complete experimental design consisted of two dependent variables, melatonin and cortisol saliva concentration, as function of the following factors:
- exposure to mobile phone (3 levels: no exposure “baseline”, exposure to phone A, exposure to phone B)
- time ( 4 levels: T0 1 hour before the exposure time, T1 corresponding to the beginning of the exposure , T2 and T3, one and two hours after the exposure respectively)
- order (2 levels corresponding to the presentation order of the A or B exposure)
- sex (2 levels)
To take into account the within-subject correlation, Subject was entered as a random–effect factor. Data were analyzed by means of a multiway ANOVA to identify the statistical significance of each source of variation, both main and two ways interaction terms were considered. A mixed-effect linear regression model was also used for generalization purpose. Data from evening and morning exposure were separately analyzed .
In Illustration 11 (top) and (bottom) the endogenous saliva melatonin in pg/ml and the saliva cortisol concentration in ng/ml are shown for each subject for each dosage period in the three different exposure conditions and divided with respect to the factor order: BA  represents the sampling order in which the phone B is firstly presented to the volunteer whilst AB is the opposite case. The most interesting period is clearly the epoch T2 corresponding to about 30 minutes after the exposure. The epochs T0 and T1 are interesting for the inter subject variability evaluation.
The main effect of the factors sex and order are negligible, so these two variables were considered only into the two ways interaction terms, the interaction between exposure condition and time, (exp*Ti) and the interaction between order and exposure (exp*order).
The final considered model was:
(ConcTi ~ Ti+exp+(Ti*exp)+(exp*order), random = ~1|subject)               (1)
where, as above mentioned, the term subject is considered as a random factor and ConcTi means the saliva concentration of endogenous melatonin or cortisol.
The factor of interest, exp, was not significant both as main effect and as through the interaction term (Ti*exp) both in the case of melatonin and cortisol concentration.
A post hoc pairwise test with a Bonferroni correction criterion for molteplicity was performed with the aim at testing the difference between the three levels of the factor exp, in particular the exposure to phone A and phone B. No significant result was found in particular in comparing the two phones A and B both in the case of melatonin as in the case of cortisol saliva concentration. From the point of view of the within subject variation, just two subjects show a significant variation in the endogenous melatonin concentration in the first epoch after the exposure. They are subject v6 who shows an increase with respect to his baseline concentration when is exposed to phone A and the subject v7 who experienced an increase with respect to its baseline condition when was exposed to phone B.
In the case of the two volunteers v6 and v7, who experienced a large increase in melatonin concentration after the exposure to one of the mobile phones, this circumstance occurred in correspondence to the second exposure. This result could be attributed to the fact that the subject should be more relaxed in correspondence to the second exposure because the experimental context is well known.
In this first part of the analysis no statistically significant difference was found both in the melatonin as in the cortisol concentration levels in the different exposure conditions.
In particular, in the first part of the statistical analysis the absence of a statistically significant difference in the exposure to phone A and to phone B was verified. On the other hand, the described experimental design was focused on testing a difference between the exposure to a real phone and a “sham” phone.
The two phones, active and “sham” differ for the 900 MHz emissions: the sham phone dissipates the power in this frequency range on a resistive load. A better characterization of the emission in the ELF range should be performed, however we could hypothesize that the ELF emission of the sham and real phone are presumably comparable. So the second part of the analysis is focused on looking at differences between mobile phone exposure (real or sham) and absence of exposure “baseline”.
The factor exposure was transformed in the two levels factor, cell_exp (“baseline” and exposure to A or B phone considered as the same source of exposure). The analysis by a multiway ANOVA was repeated as previously described in the first part of the statistical analysis.
A statistically significant difference between the condition “baseline” and “phone exposure” was not found both in the case of cortisol and in melatonin concentration levels.
The same statistical analysis was performed on data coming from the morning experiment. No statistically significant difference was found between the three levels of the exposure factor as regards the melatonin saliva concentration.
In the case of cortisol, the multiway ANOVA corresponding to the model of eq. (1) gives a statistically significant result as regards the main effect of the variables Ti (p = 0.00058) and exp (p = 0.01240) whilst the exposure condition does not discriminate data through the interaction term (Ti*exp). A post hoc pairwise test with a Bonferroni correction criterion for molteplicity was performed with the aim at testing the difference between the three levels of the factor exp finding no significant result in particular in comparing the two phones A and B. Transforming the three levels factor, exp, into the two levels factor, cell_exp, as previously described, a statistically significant effect was found in comparing the condition “baseline” and “phone exposure”(p = 0.04). With the aim at testing if this difference could be attributed to some effect due to electromagnetic field, the same analysis was separately performed on the two time epochs before and after the cell phone exposure. Due to causality, if the difference between “baseline” and “phone exposure” condition was due to the radiation it should be present only after the mobile phone exposure. A statistically significant difference between “baseline” and “phone exposure” was not found limiting the data comparison to the two epochs after the mobile phone exposure. The conclusion of this study is that some other systematic effect should be responsible of the difference between the “baseline” and the exposure condition.


No statistically significant difference in endogenous melatonin concentration was found among the three different exposure conditions: GSM 900 MHz real mobile phone, “sham” RF phone, and “baseline” concentration. This difference was investigated both during the morning phase of the cyrcadian rhythm and during the evening phase. The study of these two phases, morning and evening, was chosen at the aim of evaluating a possible exposure effect during a time period of normal using conditions of a mobile phone.The aim of reproducing a normal use condition was followed in spite of the fact that the saliva melatonin concentration is known to be very low during the day with respect to the level during the night. The evening phase showed the advantage that the cyrcadian rhythm curve (in this time interval) is quite flat and consequently a possible increasing or decreasing trend, if present, could be more easily detected than during the morning, when the melatonin concentration is rapidly falling down. The hypothesis of melatonin concentration decreasing as a consequence of the exposure to mobile phone is not supported by this experiment. A weak increasing trend in melatonin concentration was observed in some subjects immediately after the exposure to both type of cell phones, real or sham. This trend was not statistically significant on the entire data set and was usually observed in correspondence to the second presented exposure. This aspect, if not casual, could be explained as a relaxation effect in the subject due to the better experience of the experimental condition, or suggest a potential effect of ELF magnetic fields produced by circuitry and buttery currents, that are present in both real and sham exposure. Indication for future experiments is that a larger sample of subjects is needed, and much care must be put in the randomization of exposure condition presentation. The potential effect of ELF magnetic fields produced by circuitry and buttery currents should be also investigated, provided a detailed exposure assessment to such fields. The metabolite urinary concentration is however not suitable as a biomarker for future experiments, as possible variations are masked by the individual and seasonal variability.
On the other side, the saliva test demonstrated that during the day the melatonin secretion is very low: therefore an eventual suppression effect could be discriminated only repeating the experiment during the night time. The concentration of salivary cortisol is unsuitable as an indicator of exposure to mobile phone for short periods and low intensity infact secretion of salivary cortisol was highly variable over time in all three conditions (control, exposure and sham,). Determination of salivary cortisol seems therefore not be influenced by exposure to mobile phones.

Source(s) of Funding

The present paper is one of the products of a larger research project financed by the Italian Ministry of Health in 2004 regarding the possible effects of mobile phones on human health, in particular about neurobehavioural effects. All the authors were researcher of the Italian Institute for Occupational Prevention and Safety (ISPESL), the leader institution in this project, that involved the cooperation of other private and public Italian reserch institutions.
ISPESL has been suppressed on May 31st 2010 and incorporated into INAIL (Italian Workers' Compensation Authority )

Competing Interests



This article has been downloaded from WebmedCentral. With our unique author driven post publication peer review, contents posted on this web portal do not undergo any prepublication peer or editorial review. It is completely the responsibility of the authors to ensure not only scientific and ethical standards of the manuscript but also its grammatical accuracy. Authors must ensure that they obtain all the necessary permissions before submitting any information that requires obtaining a consent or approval from a third party. Authors should also ensure not to submit any information which they do not have the copyright of or of which they have transferred the copyrights to a third party.
Contents on WebmedCentral are purely for biomedical researchers and scientists. They are not meant to cater to the needs of an individual patient. The web portal or any content(s) therein is neither designed to support, nor replace, the relationship that exists between a patient/site visitor and his/her physician. Your use of the WebmedCentral site and its contents is entirely at your own risk. We do not take any responsibility for any harm that you may suffer or inflict on a third person by following the contents of this website.

0 comments posted so far

Please use this functionality to flag objectionable, inappropriate, inaccurate, and offensive content to WebmedCentral Team and the authors.


Author Comments
0 comments posted so far


What is article Popularity?

Article popularity is calculated by considering the scores: age of the article
Popularity = (P - 1) / (T + 2)^1.5
P : points is the sum of individual scores, which includes article Views, Downloads, Reviews, Comments and their weightage

Scores   Weightage
Views Points X 1
Download Points X 2
Comment Points X 5
Review Points X 10
Points= sum(Views Points + Download Points + Comment Points + Review Points)
T : time since submission in hours.
P is subtracted by 1 to negate submitter's vote.
Age factor is (time since submission in hours plus two) to the power of 1.5.factor.

How Article Quality Works?

For each article Authors/Readers, Reviewers and WMC Editors can review/rate the articles. These ratings are used to determine Feedback Scores.

In most cases, article receive ratings in the range of 0 to 10. We calculate average of all the ratings and consider it as article quality.

Quality=Average(Authors/Readers Ratings + Reviewers Ratings + WMC Editor Ratings)