br after inclusion of the mobile phone penetra tion rate
+ 4.5% (-2.1%, 102.4%) after inclusion of the mobile phone penetra-tion rate into the models, which is also unlikely to have been the result of mobile phone use.
This research aimed to further assess the likelihood of mobile phone use being an important putative cause for the observed increased in-cidence of glioblastoma multiforme and malignant and benign neo-plasms in the temporal lobe from the 1980s. A slightly higher impact of + 37.6% was observed for all malignant neoplasms in the temporal lobe compared to the previous analysis (+35%) (De Vocht, 2016), because a later update of the national cancer registry database was used in this study and some additional covariates were included to calculate the counterfactuals. Assessment of specific cancer subtypes in the temporal lobe in-dicated that the excess incidence was mainly found for GBM, with si-milar trends observed in the frontal lobe and cerebellum. Increased incidence rates of GBM have previously been reported elsewhere, but not everywhere, as well (Ostrom et al., 2014). The increased rates of specific Tofacitinib cancer subtypes in excess of the counterfactuals corre-spond to the spatial and temporal patterns that would be expected if exposure to RF from mobile phones were an important putative factor (Cardis et al., 2008; Morgan et al., 2016), and mobile phone use has been associated with higher mutant type p53 gene expression in the peripheral zone of the glioblastoma (Akhavan-Sigari et al., 2014) and
Analyses of all malignant neoplasms and Glioblastoma multiforme (GBM) in the temporal lobe by age group.
Annual Newly Total cases Cumulative Causal 95% Credible Bayesian tail-area Inclusion mobile 95% CI Bayesian tail-area
diagnosed cases (1985-2014) impact (%) interval probability phones
Fig. 1. Actual (solid line) and modelled counterfactual (dotted line) plus 95% Bayesian Credible Intervals (shaded area) 1985–2014 annual newly diagnosed cases of glioblastoma multiforme in the temporal lobe for different age groups. Note that 75 + and 85 + age groups also included in 65 + group.
with decreased survival of GBM patients (Carlberg and Hardell, 2014; Akhavan-Sigari et al., 2014). However, age group-specific analyses in-dicate that the excess relative impacts increased with age over 65 years and were primarily found in the very old (75/85 + years of age) for whom it is unlikely that mobile phone use had been an important causal factor. In addition, an effect of inclusion of mobile phone penetration rate was also observed in the young (< 24 years of age) for whom mobile phone use is also an unlikely causal factor.
Combining these results indicates that the likelihood of mobile phone usage being an important putative factor for the observed in-creases in primarily GBM since the 1980s is small, and that some other factor may be the cause and would have occurred in over a comparable time period. Analyses from other regions came to the same conclusion as this study (Kim et al., 2015; Sato et al., 2016; Zada et al., 2012), while previous analyses of English national GBM age-standardized rates for 5-year groups similarly showed an increase from 1995–99 to 2011–15 for most age groups, but with a steep increase in the rates for those 65 years of age and above, increasing with age (Philips et al., 2018). However, the authors of the latter study did not sufficiently discuss the implications of this age distribution for the interpretation of their results, while also because of the analytical method used they could not make comparisons with expected, counterfactual, trends in incidence rates.
In addition, these analyses provide no evidence for an association between mobile phone use and increased risk of benign neoplasms, including meningioma and acoustic neuroma specifically; temporal trends did not differ substantially from their counterfactuals. For me-ningioma these results were in agreement with data from Interphone and other case-control studies (Carlberg and Hardell, 2015; Interphone Study Group, 2010; Schuz et al., 2006), but for acoustic neuroma in-creased risks were observed for highest exposure levels (Hardell et al., 2013; Interphone Study Group, 2011). The latter may be because acoustic neuroma is a slow growing tumour, and the 10-year lag in our analyses may not have been sufficient; although there was also no evidence of an increase for the 15-year lagged sensitivity analysis.
If exposure to radiofrequency radiation from mobile phones is not
the driving factor behind the observed excesses in primarily GBM rates, it is interesting to speculate what this other factor could be. Several candidates have been mentioned (Miranda-Filho et al., 2017), including some of environmental origin (Philips et al., 2018), but alternatively it has been suggested that the observed increase in the incidence of GBM is the result of improved diagnosis techniques, especially at older age (Kim et al., 2015; Zada et al., 2012; Davis et al., 1990; Greig et al., 1990), and improvements in classifications of gliomas, including the increased use of molecular markers (Ludwig and Kornblum, 2017). The current analyses seem to support the latter interpretation, at least in so far as a main driving factor is considered, and this is further strength-ened by the patterns observed for newly diagnosed cases of GBM in overlapping lesions or of an unspecified nature. This would however, not explain the observed excess in newly diagnosed cases of malignant neoplasms (but not GBM alone) in the temporal lobe in the under-24 age group. This could be a chance finding, but it has also been high-lighted that secondary glioblastoma progress from low-grade diffuse or anaplastic astrocytoma, rather than de novo glioblastomas in elderly patients, make up the vast majority of glioblastomas in younger patients (Philips et al., 2018), or may occur as a result of different prenatal or early-life exposure(s) of which RF from other sources or extremely-low electromagnetic fields also cannot be completely excluded (Mortazavi et al., 2017). It has been argued that the observed increase in GBM cannot solely be ascribed to improvements in diagnostic techniques because it affects specific areas in the brain only (Philips et al., 2018), which may indeed imply that exposure to RF from mobile phones cannot be excluded completely as a contributing factor, especially since the brain regions identified as showing an excess compared to their counterfactuals are those that absorb 81–86% of all mobile phone ra-diation (Cardis et al., 2008; Morgan et al., 2016). However, the region-specific analyses in this study were suggestive of comparable patterns compared to the counterfactuals in the occipital and cerebral lobes as well, suggesting effects may be less strong, but not exclusive to the aforementioned anatomic regions. Most likely the observed trends are the result of a combination of different factors, environmental and/or other (Miranda-Filho et al., 2017; Ostrom et al., 2014), and which may