There have been allegations in the media and in the courts that cell phones and other types of hand-held transceivers are a cause of cancer. There have also been numerous public objections to the siting of TV, radio and cell phone transmission facilities because of a fear of cancer induction. A recent publication in Radiation Research by Repacholi et al. (147, 631-640, 1997) which suggests that exposure to radiofrequency (RF) radiation may increase lymphoma incidence in mice has contributed to this controversy. The goal of this review is to provide biomedical researchers a brief overview of the existing RF radiation-cancer studies. This article begins with a brief review of the physics and technology of cell phones. It then reviews the existing epidemiological studies of RF radiation, identifying gaps in our knowledge. Finally, the review discusses the cytogenetics literature on RF radiation and the whole-animal RF-radiation carcinogenesis studies. The epidemiological evidence for an association between RF radiation and cancer is found to be weak and inconsistent, the laboratory studies generally do not suggest that cell phone RF radiation has genotoxic or epigenetic activity, and a cell phone RF radiation-cancer connection is found to be physically implausible. Overall, the existing evidence for a causal relationship between RF radiation from cell phones and cancer is found to be weak to nonexistent.

As part of a health and safety assessment of ultrawideband sources, it was useful to determine stimulation thresholds for an electrically excitable tissue down into the low nanosecond range. Stimulation thresholds were measured using gastrocnemius muscles isolated from 16 frogs (Rana sp.). Single pulses were delivered with a pair of surface electrodes, and muscle twitch was measured with an isotonic transducer. Pulse durations of 100, 10, and 1 ms; 100, 10 and 1 s; and 100 and 1 ns were used. Tissue voltage and current strength-duration (S-D) curves on log-log plots had a classic appearance, with thresholds for ultrashort pulses being linear. For a pulse of 1 ns, the mean threshold voltage in the muscle was 4.5 kV and the mean threshold peak current was 35 A. When delivered by direct contact, a single ultrawideband pulse of 1 ns could reliably produce a biological effect, stimulation of an electrically excitable tissue. The observation that the S-D curves extended downward to 1 ns in a linear manner suggested that classical ion channel mechanisms regulated excitation and that other processes, such as electroporation, did not occur. Although a single nanosecond pulse delivered by direct contact can elicit a biological response, such a stimulus in air is unlikely to produce an effect.

Since the reported alterations of permeability of the blood-brain barrier by microwave radiation have implications for safety considerations in man, studies were conducted to replicate some of the initial investigations. No transfer of parenterally-administered fluorescein across the blood-brain barrier of rats after 30 min of 1.2-GHz radiation at power densities from 2--75 mW/cm2 was noted. Increased fluorescein uptake was seen only when the rats were made hyperthermic in a warm-air environment. Similarly, no increase of brain uptake of 14C-mannitol using the Oldendorf dual isotope technique was seen as a result of exposure to pulsed 1.3-GHz radiation at peak power densities up to 20 mW/cm2, or in the continuous wave mode from 0.1--50 mW/cm2. An attempt to alter the permeability of the blood-brain barrier for serotonin with microwave radiation was unsuccessful. From these studies it would appear that the brain must be made hyperthermic for changes in permeability of the barrier induced by microwave radiation to occur.

Experimental studies that sought teratologic effects or developmental abnormalities from exposure to radiofrequency electromagnetic fields (RFEMF) in the range 3 kHz-300 GHz are critically reviewed for their possible consequences on human health. Those studies were conducted on beetles, birds, rodents, and nonhuman primates. Collectively, those experimental studies indicate that teratologic effects can occur only from exposure levels that cause biologically detrimental increases in body temperature. No reliable experimental evidence was found for nonthermal teratologic effects; rodents, mouse fetuses, and perinatal mice are more susceptible to such effects than rats. The primary confirmed effect in rats at high RFEMF levels was initial weight deficits in fetuses and neonates that decreased with infant growth. More generally from findings with pregnant mammals, exposures at RFEMF levels far higher than those permitted under the IEEE human exposure guidelines are necessary to reach or exceed cited experimental thresholds for maternal temperature increases. Some results indicated that the levels necessary to cause such effects in pregnant mammals could exceed those lethal to the dams. In a behavioral study of squirrel monkeys, no effects were observed on usual dam-offspring interactions or EEGs, but unexpected deaths of a number of offspring had occurred. However, this finding was not confirmed in a study solely on infant death using a larger number of subjects for greater statistical validity. Also reviewed were epidemiologic studies of various human populations considered to have been chronically exposed to environmental levels of RFEMF. Early studies on the incidence of congenital anomalies yielded no credible evidence that chronic exposure of pregnant women or of fathers exposed to RFEMF from nearby sources at levels below those guidelines would cause any anomalies in their offspring. The findings of studies on pregnancy outcomes of female physiotherapists occupationally exposed while treating patients with RFEMF were mixed, but taken collectively, the findings were negative. Bioelectromagnetics Supplement 6:S174-S186, 2003.

The purpose of this study was to determine if chronic, low-level exposure of mice prone to mammary tumors to 435 MHz radiofrequency (RF) radiation promotes an earlier onset, a faster growth rate or a greater total incidence of mammary tumors than in sham-exposed controls. Two hundred female C3H/HeJ mice were exposed for 21 months (22 h/day, 7 days/week) to a horizontally polarized 435 MHz pulse-wave (1.0 micros pulse width, 1.0 kHz pulse rate) RF radiation environment with an incident power density of 1.0 mW/cm2 (SAR = 0.32 W/kg). An additional 200 mice were sham-exposed. Animals that died spontaneously, became moribund or were euthanized after 21 months of exposure were completely necropsied; tissues were subjected to histopathological examinations. Concerning mammary carcinomas, there were no significant differences between the two groups with respect to latency to tumor onset, tumor growth rate and overall tumor incidence. Histopathological examination revealed no significant differences in numbers of malignant, metastatic or benign neoplasms between groups. Survival probability was estimated by the Kaplan-Meier method; no significant difference between groups was noted (Cox's test). Under the conditions of this long-term study, low-level exposure of mice prone to mammary tumors to 435 MHz RF radiation did not affect the incidence of mammary tumors, tumor growth rate, latency to tumor onset or animal longevity when compared to sham-exposed controls.