Article publication date
March 2026
ARPANSA review date
May 2026
Summary
This commentary by Belenki et al. from the German Federal Office for Radiation Protection (BfS) provides an in-depth criticism of the World Health Organisation (WHO) commissioned systematic review of radiofrequency electromagnetic field (RF‑EMF) exposure and cancer in laboratory animals by Mevissen et al. (2025). It includes a detailed critique and re-analysis, an overview of methodological and analytical flaws, and a comprehensive list of identified errors.
The commentary’s main critique is that the systematic review diverged from their pre-published protocol and did not perform meta-analyses. Mevissen et al., (2025) argues that the methods and results of individual studies are too different and can’t be combined in meta-analyses. Instead, Mevissen et al., (2025) further argues that any positive outcome can be used to conclude an effect of RF-EMF on cancer, despite this method meaning much of the available evidence is ignored. This meant that the majority of conclusions of the animal systematic review were based on the results of National Toxicology Program (NTP) study which is one of the largest studies on RF-EMF and cancer in rats and mice. However, the NTP study has been heavily critiqued and many of the flaws were not considered by Mevissen et al. (2025). Belenki et al. (2026) corrected this by conducting meta-analyses and certainty of evidence evaluations for sufficiently similar studies following the originally published protocol. These meta-analyses resulted in odds ratios (OR) for both malignant glioma (OR 3.3, 95% confidence interval (CI) 0.46-23.84) and malignant cardiac schwannoma (OR 9.11, 95% CI 1.51-54.84) with very wide confidence intervals. This means that there is a lot of imprecision in the data and the true risk is hard to determine.
GRADE and OHAT provide guidance on how to interpret confidence intervals when conducting systematic reviews. Both guidance documents recommend that, when confidence intervals are as wide as observed in these results, the certainty of evidence (CoE) should be downgraded twice. Consequently, Belenki et al. downgraded CoE from high certainty to low. This is a substantially different conclusion to that reported by Mevissen et al. (2025) who claimed a high CoE.
Mevissen et al. (2026), responded to the commentary made by Belenki et al. (2026), contending that rats of different strains could not be compared to each other in a meta-analysis. It was further argued that, if there is a positive outcome in one study and a negative outcome in another study, only positive outcomes should be considered when examining rare cancers. Mevissen et al. (2026) concedes many of the other errors pointed out by Belenki et al. (2026).
Published in
Environment International
Link to
ARPANSA commentary
The paper by Mevissen et al. (2025) has now been critiqued by a number of national organisations including the Swiss expert group on electromagnetic fields and non-ionising radiation, BfS and ARPANSA. The general assessment is that the methods used by Mevissen et al. (2025) were flawed because they resulted in an assessment that did not consider the evidence as a whole and instead only took into account studies showing positive associations.
Since the publication of Mevissen et al. (2025) two large studies, simultaneously but independently conducted in Korea (Kim et al., 2026) and Japan (Imaida et al., 2026), investigated whether long‑term exposure to RF-EMF can cause tumours or genetic damage in rats. The findings of these studies were remarkably consistent with each other as there were no statistically significant changes in tumour incidence in either study, in contrast to the results reported in NTP study. Any future meta-analysis on this topic would be significantly influenced by the Korean and Japanese studies, improving the CoE and providing further evidence that RF-EMF does not cause cancer in laboratory animals.
The WHO commissioned systematic review process aimed to assess the possible implications of RF-EMF exposure on human health. The most significant evidence on the impact of RF-EMF on human health comes from studies on humans,not animals. The WHO commissioned systematic reviews looking at observational studies in humans (Karipidis et al., 2024, Karipidis et al., 2025) did not find an association between RF-EMF and any cancer type, including brain cancer.
The meta-analyses conducted by Belenki et al. (2026) has completed the work that the WHO commissioned animal systematic review set out to do in their protocol. Their syntheses of the evidence combined with the newly published results from the Korean and Japanese studies do not the support the association between RF-EMF exposure and cancer in animals.
