Archives

  • 2019-07
  • 2019-08
  • 2019-09
  • 2019-10
  • 2019-11
  • 2020-03
  • 2020-07
  • 2020-08
  • br Table describes the results for the analyses of

    2020-07-28


    Table 2 describes the results for the analyses of trend in newly di-agnosed cases of subtypes of Sotrastaurin (AEB071) cancers and benign neoplasms in the temporal lobe. Newly diagnosed cases of primary malignant neoplasms in the temporal lobe have increased faster post-2005 then the coun-terfactual predicted; with an estimated additional increase of 33.1% (95% BCI 9.8%, 54.4%), which corresponds to 1659 cases over the 2006–2014 period. The analysis of specific subtypes provides some evidence that gliomas ‘not otherwise specified’ (NOS) have increased beyond expectation, although the estimate of the causal effect is very imprecise (95% BCI -16.8%, 372.8%). There is stronger evidence that the increase of all malignant neoplasms in the temporal lobe was more specifically an excess in newly diagnosed cases of GBM (+37.6% [95% BCI -6.6%, 77.6%]; p-value 0.05) and anaplastic astrocytoma (+45.2% [95%BCI 8.1%, 79.3%]; p-value 0.01). There was no evidence of other 
    astrocytoma subtypes or of oligodendrogliomas deviating from ex-pected, counterfactual, trends. These analyses further also do not sup-port an association between the introduction of mobile phones and increased incidence of benign neoplasms (-5.2% [-29.1%, 16.9%]), nor for meningioma (-39.1% [-344.4%, 394.4%]) or acoustic neuroma (-7.7% [-38.7%, 20.4%]) specifically. Sensitivity analyses in which the hypothesised lag is changed to 0 years, 5 years or 15 years show that for these alternative lags nearly all observed temporal trends are compar-able to the counterfactuals. The only exceptions are all malignant neoplasms in the temporal lobe, which also shows a + 36% excess when no lag is assumed, but not for 5 and 15 years, and anaplastic oligodendroglioma, for which an excess effect is found for 0 years and a reduction in newly diagnosed cases for a 15-year lag. Neither of these point towards mobile phones being an important risk factor (Online supplementary materials Table S2).
    To assess whether mobile phone usage could be an important pu-tative factor, the national mobile phone penetration rate (assuming a 5-year lag) was included in the models for GBM in the anatomic regions and for subtypes in the temporal lobe with excess incidence post-2005 (Table 3). Inclusion reduced the effect size 61% and 75% for GBM in the frontal and temporal lobes, but not for those in the Cerebellum. Within the temporal lobe, mobile phone penetration rates also reduced the effect size for all malignant neoplasms and glioma (NOS) by about 50%, but had no effect on the observed impact on anaplastic astrocytoma.
    Additional Age-group specific analyses were conducted to assess whether excesses in newly diagnosed cases in the temporal lobe com-pared to the counterfactuals were distributed evenly across age groups or were located in specific age groups, and explore how this impacts on the likelihood of mobile phone usage being an important putative
    Table 2
    Modelled causal effects for specific neoplasm subtypes in the temporal lobe with over 100 cases (1985–2014), assuming a 10-year lag between exposure and measurable effect.
    Temporal lobe cancers Total cases Cumulative Causal impact Cumulative Causal 95% Credible Bayesian tail-area
    Table 3
    Comparison of modelled causal effects with and without the mobile phone penetration rate included in the Bayesian structural timeseries for selected cancer subtypes.
    Original
    Inclusion of mobile phone penetration rate (10-year lag)
    Cumulative Causal impact (%) 95% Credible interval
    Cumulative Causal impact (%) 95% Credible interval
    Frontal lobe
    All malignant
    factor. Results are provided in Table 4 and also shown graphically in Fig. 1. Excess, compared to the counterfactuals, including numbers of newly diagnosed cases of all malignant neoplasms in the temporal lobe and of GBM specifically were primarily observed for those over 45 years of age, which corresponds to the likely ages “first adopters” would have had. Inclusion of mobile phone penetration rates however, only affected the effect size for GBM and not for all malignant neoplasms. Moreover, the effect size of both all malignant neoplasms and GBM in the temporal lobe increases with older age groups from about 30% in those aged 45–64 when diagnosed, to about 45% for the 65 + age group, and to 50–80% and 127–177% for the 75 + and 85 + age groups, respec-tively. These trends correspond to a decreasing likelihood that these were “early adopters” (for example, the 85 + group would have been about 60 + years of age when first relevant exposure occurred). On the other side of the age spectrum, excess incidence of all malignant neo-plasms in the temporal lobe (but not for GBM) was also observed for the 0–24 years of age group (41.0% [-32.%, 84.4%]) which reduced to