Of the 488 peer-reviewed studies in this collection on radiofrequency exposure, 364 — 74.6% — reported biological effects. Almost all of that research covers frequencies 4G and 5G share; the millimeter-wave band unique to high-band 5G is barely represented. The meaningful difference is not the generation label but the dose: how close the transmitter is, how long it transmits, and how hard it has to work to reach the network.
Based on analysis of 493 peer-reviewed studies
People often ask whether 5G is more dangerous than 4G. This question requires understanding how 5G technology differs from previous generations and what research exists on each.
5G networks operate across multiple frequency bands. Low-band 5G (600-900 MHz) is actually similar to 4G frequencies. Mid-band 5G (2.5-4 GHz) overlaps with existing WiFi. High-band 5G (24-40+ GHz, "millimeter wave") represents the newest frequencies for consumer wireless exposure.
This page compares what research shows about radiation exposure from 5G versus 4G technologies.
4G networks operate roughly between 700 MHz and 2.6 GHz. 5G spans a far broader range: low and mid bands that overlap 4G and WiFi, and high bands extending toward about 100 GHz. Those are spectrum allocations rather than research findings, and they frame the comparison.
Of the 488 peer-reviewed studies indexed here on radiofrequency exposure, 364 — 74.6% — reported measurable biological effects. That literature was built on the lower frequencies, which both generations use. No study here compares 4G against 5G directly.
Higher frequencies penetrate tissue less deeply. Millimeter waves deposit their energy largely in skin and the outer surface of the eye rather than reaching deep organs, which changes which tissues are exposed rather than removing exposure. Lower frequencies penetrate further but carry less energy per photon — still far too little to break chemical bonds directly, which is why every effect discussed here is non-thermal and indirect rather than ionizing.
Tice and colleagues (2002) tested four cellular signal technologies — analog, TDMA, CDMA and GSM formats at 837 and 1909 MHz — on cultured human blood cells. Three or 24 hours of exposure produced no significant DNA damage in leukocytes. Extended exposure at an average specific absorption rate of at least 5.0 W/kg did induce chromosomal damage in lymphocytes. Signal format did not determine the outcome; dose did.
Lai and Singh (1995) found DNA single-strand breaks in rat brain cells rising with dose rate across 0.6 and 1.2 W/kg after two hours at 2450 MHz, and their 1996 study found both single- and double-strand breaks at 1.2 W/kg. Their 1997 work showed melatonin and a spin-trap compound both blocked the effect, implicating free radicals rather than heating.
For almost everyone, the largest single radiofrequency exposure is their own phone against their head, not any tower. That exposure depends on transmit power, which depends on signal quality. A handset with a weak signal increases power to reach the network; a handset with a strong signal backs off. This is why adding antennas can reduce rather than increase personal exposure, and why the generation label on the network is a poor predictor of what any individual absorbs.
No study here establishes that 5G is more harmful than 4G, or less. The comparison has not been run. What the research supports is that radiofrequency exposure at sufficient dose produces non-thermal biological effects, that the doses involved are typically at handset-against-head levels rather than ambient network levels, and that the newest part of 5G has not been meaningfully studied at all.
The steps that reduce exposure are the same on either network: speaker mode or wired headphones for calls, texting over calling, phone out of the bedroom overnight, and WiFi rather than cellular where the cellular signal is weak. None depends on which generation the network runs.
Grisanti G et al. · 1998
Researchers studied bioelectromagnetic effects induced by TACS phones on the auditory system. Modulated electromagnetic fields modified distortion products in about all examined subjects. The method used registration of evoked otoacoustic emissions, both transient and distortion products.
Novoselova ET, Fesenko EE. · 1998
Researchers studied whether weak electromagnetic waves in the centimeter range could affect tumor necrosis factor production in macrophages from mice. Whole-body microwave irradiation of male NMRI mice, exposure of macrophages in vitro, and irradiation of culture medium with 8.15-18 GHz at a power density of 1 microW/cm2 each caused a significant enhancement of tumor necrosis factor production in peritoneal macrophages.
Behari J, Kunjilwar KK, and Pyne S · 1998
Researchers studied the effect of low-level amplitude-modulated radiofrequency radiation on Na+–K+-ATPase activity in the brains of developing male Wistar rats aged 23 days. Rats were exposed to carrier wave frequencies of 147 MHz and its sub-harmonic frequencies 73.5 and 36.75 MHz, amplitude modulated at 16 and 76 Hz, for 30–35 days at 3 hours per day. Na+–K+-ATPase activity increased by around 19–20% in rats exposed with amplitude modulation at 16 Hz, and by around 15–16% in rats exposed with amplitude modulation at 76 Hz, compared to controls. The researchers concluded that the effect on Na+–K+-ATPase activity was insensitive to the carrier wave frequencies under investigation.
Phillips et al. · 1998
Researchers exposed Molt-4 T-lymphoblastoid cells in vitro to pulsed radiofrequency signals at cellular telephone frequencies of 813.5625 MHz (iDEN® signal) and 836.55 MHz (TDMA signal) to detect DNA damage. The alkaline comet assay was used to measure DNA single-strand breaks in exposed cell cultures compared to sham-exposed cultures. At an SAR of 24 μW g−1, exposure to the iDEN® signal for 2 h and 21 h significantly increased DNA damage, while at 2.4 μW g−1 the same signal significantly decreased DNA damage. The researchers concluded that the data indicate a need to study the effects of exposure to RF signals on direct DNA damage and on the rate at which DNA damage is repaired.
Vijayalaxmi, Mohan, N, Meltz, ML, Wittler, MA, · 1997
Stagg RB, Thomas WJ, Jones RA, Adey WR · 1997
Researchers exposed transformed and normal rat glial cells to an 836.55 MHz packet-modulated RF field at power densities of 0.09, 0.9, and 9 mW/cm2, with SARs from 0.15 to 59 muW/g. Primary rat glial cells showed no significant differences in DNA synthesis between sham and RF-exposed conditions for either log-phase or serum-starved cultures. C6 glioma cells exposed at 5.9 muW/g SAR showed small but significant increases in thymidine incorporation in 38% of experiments. Growth curves showed no differences between sham and RF-exposed cultures for either cell type, and doubling times of C6 glioma cells did not differ significantly.
Klug S, Hetscher M, Giles S, Kohlsmann S, Kramer K, · 1997
Researchers exposed 9.5-day-old rat embryos for up to 36 hours to radiofrequency electric and magnetic fields with modulation frequencies of 16, 60, and 120 Hz. The exposure included electric field strengths of 60 and 600 V/m, magnetic inductions of 0.2 and 2.0 microtesla, and specific absorption rates of 0.2, 1.0, and 5.0 W/kg. Neither the electric nor the magnetic fields interfered significantly with normal embryo growth and differentiation when assessed on day 11.5 using a scoring system, photographs, histology, and protein content.
Ivaschuk OI et al. · 1997
Researchers exposed nerve growth factor-treated PC12 rat pheochromocytoma cells to a packet-modulated 836.55 MHz radiofrequency field at power densities of 0.09, 0.9, and 9 mW/cm2 for 20, 40, or 60 minutes, with an intermittent 20-minute on/off regimen. Sham-exposed cells served as controls, and c-fos and c-jun expression was measured by Northern blot analysis of total RNA. C-fos transcript levels did not change after 20 minutes at any intensity, while c-jun transcripts decreased by 38% only after 20 minutes at 9 mW/cm2.
Gos, P, Eicher, B, Kohli, J, Heyer, WD · 1997
Researchers exposed exponentially growing Saccharomyces cerevisiae cells to electromagnetic fields from 41.682 GHz to 41.710 GHz in 2 MHz increments at power densities of 0.5 and 50 microW/cm2. Cell division was assessed by time-lapse photography using two identical chambers that allowed concurrent control and exposure experiments. Control experiments showed the cells were dividing at submaximal rates, ensuring that any increase or decrease in division rate could be observed. No consistently significant differences were found between exposed and unexposed cells across multiple independent experiment series.
Cain CD, Thomas DL, Adey WR · 1997
The study tested whether a TDMA-modulated RF field and the tumor promoter TPA together increase focus formation of transformed cells cocultured with parental C3H/10T1/2 murine fibroblasts. Cell cultures were exposed to an 836.55 MHz TDMA-modulated field at specific absorption rates of 0.15, 1.5, and 15 μW/g in a repeating cycle of 20 minutes on and 20 minutes off continuously for 28 days. RF exposure did not significantly alter TPA-induced focus formation at any of the tested exposure levels across multiple independent experiments.
Malyapa RS et al. · 1997
Researchers exposed mouse C3H 10T1/2 fibroblasts and human glioblastoma U87MG cells to FMCW (835.62 MHz) and CDMA (847.74 MHz) radiation at 0.6 W/kg SAR for up to 24 hours. They used the alkaline comet assay to measure DNA damage. No significant differences were found between exposed cells and sham controls.
Persson BRR, Salford LG, Brun A · 1997
Researchers studied blood-brain barrier permeability in Fischer 344 rats of both sexes exposed to radio frequency electromagnetic fields. The exposure used 915 MHz microwaves delivered as either pulse-modulated or continuous wave signals. The rate of pathological leakage rose from 0.17 in controls to 0.39 across all exposed rats, a highly significant difference. Continuous wave exposure produced a higher rate of pathological rats (0.50 ± 0.07) than pulse-modulated exposure (0.35 ± 0.03), and that difference was also statistically significant.
Rojavin MA, Ziskin MC · 1997
Researchers injected BALB/c mice with ketamine 80 mg/kg or chloral hydrate 450 mg/kg and exposed them to 61.22 GHz millimeter waves at 420 W/kg peak specific absorption rate for 15 minutes. Millimeter wave exposure increased anesthesia duration by about 50% compared to sham exposure, which had no effect.
Mason PA et al. · 1997
Researchers measured extracellular amino acid concentrations in the hypothalamus and caudate nucleus of rats using in vivo microdialysis before and during exposure to RFR. Rats were exposed to 5.02 GHz radiation (10 microseconds pulse width, 1000 pulses/s) for 40 minutes while brain temperatures were maintained at approximately 41.2 and 41.7 degrees C. Aspartic acid, serine, and glycine concentrations increased significantly in both regions during exposure. Glutamic acid and glutamine changes were not statistically significant.
Litovitz et al. · 1997
Researchers exposed L929 cells to 60 Hz AM microwave fields or a 50 Hz burst-modulated DAMPS digital cellular phone field together with 30-100 Hz ELF noise up to 10 microT for 8 h. ODC enhancement decreased exponentially with increasing noise amplitude, with complete inhibition at 2 microT or higher for 60 Hz AM microwaves and 5 microT or higher for the DAMPS signal.
Gapeev AB, Lakushina VS, Chemeris NK, Fesenko EE · 1997
Researchers investigated how low-intensity modulated extremely high frequency electromagnetic radiation affected the respiratory burst of peritoneal neutrophils from NMRI mice. Continuous radiation at 50 microW/cm2 inhibited the reaction, with a maximum effect of about 25% at a carrier frequency of 41.95 GHz. When modulation was applied at 1 Hz, the reaction was activated, while modulation at 0.1, 16, and 50 Hz inhibited it. The effect was observed only when intracellular free calcium concentration was raised and calcium was flowing through the plasma membrane.
Cleary, SF, Cao, G, Liu, LM, Egle, PM, Shelton, KR · 1997
Researchers investigated stress protein induction in HeLa and CHO cells following a 2-hour exposure to radiofrequency or microwave radiation. Cells were exposed under isothermal conditions maintained at 37 ± 0.2 °C. Neither microwave nor RF radiation had a detectable effect on stress protein induction.
Ohmoto Y et al. · 1996
Neshev NN, Kirilova EI · 1996
The authors present a theoretical model describing how pulse-modulated microwaves might influence enzyme conformational oscillations in living organisms. The model predicts that specific pulse-repetition times matching enzyme oscillation periods can produce effects at extremely low power levels. The model indicates possible environmental-health risks from long-term exposure to ambient fields from radar, navigation, and communication systems.
Gapeev AB, Safronova VG, Chemeris NK, Fesenko EE · 1996
Researchers compared horn, dielectric, and channel antennas for matching with biological loads. Electromagnetic radiation at 41.8-42.05 GHz inhibited luminol-dependent chemiluminescence of neutrophils activated by opsonized zymosan. No frequency dependence was found in the far field zone, but a quasi-resonance dependence appeared in the near field zone. The channel antenna provided uniform specific absorption rate distribution and wide-band matching in both near and far field zones.
Detlavs I et al. · 1996
Researchers exposed Wistar rats with full-thickness dermal wounds to RF EMF for 30 min daily during the first 5 days using 53.53 GHz without modulation, 42.19 GHz without modulation, and 42.19 GHz with a 200-MHz frequency modulation band. RF EMF without frequency modulation decreased glycoprotein macromolecules and reduced inflammatory exudation and total collagen accumulation. RF EMF with frequency modulation significantly elevated hexoses and sialic acid concentrations, indicating intensified exudative phenomena, and increased collagenous protein accumulation. The authors concluded that non-thermal EMF affects reparative-proliferative processes in soft tissue wounds.
Balode, Z · 1996
Researchers examined micronuclei in bovine peripheral erythrocytes from female Latvian Brown cows near the Skrunda Radar and from a control area. They counted micronuclei in 2000 erythrocytes per animal at 1000x magnification using a simplified Schiff DNA staining method. The exposed group had an average of 0.6 micronuclei per 1000 erythrocytes versus 0.1 per 1000 in controls, a statistically significant difference (P < 0.01).
Urech, M, Eicher, B, Siegenthaler, J · 1996
Researchers conducted field experiments lasting one to three years alongside laboratory tests, exposing Parmelia tiliacea and Hypogymnia physodes to 2.45 GHz microwaves at 0.2, 5, and 50 mW/cm² with a control. Both species showed a substantially reduced growth rate at the highest exposure of 50 mW/cm². Temperature measurements revealed that microwaves at 2.45 GHz and 50 mW/cm² substantially increased lichen surface temperature and accelerated drying. When the same species were exposed near a 9.5 MHz short-wave broadcast transmitter at up to 235 V/m, no visible effects were detected.
Kubinyi G, Thuroczy G, Bakos J, Boloni E, Sinay H, Szabo LD, · 1996
Pregnant CFLP mice were exposed to 2.45 GHz microwave radiation for 100 minutes each day throughout gestation in an anechoic chamber. Brain aminoacyl-tRNA synthetase activity decreased significantly after continuous-wave exposure but not after amplitude-modulated exposure. Liver enzyme activity increased after both continuous-wave and amplitude-modulated exposure, while body and organ weights were unaffected.
Kolosova LI, Akoev GN, Avelev VD, Riabchikova OV, Babu KS · 1996
Researchers examined the effect of low-intensity millimeter wave electromagnetic radiation on regeneration of the rat sciatic nerve after transection and microsurgical reapproximation. At the twentieth postoperative day, treated animals showed a 32% increase in regeneration distance and a 26% increase in conduction velocity compared to controls.
Further Reading:
For a comprehensive exploration of EMF health effects and practical protection strategies, explore these books by R Blank and Dr. Martin Blank.
