Lloyd Morgan's Column

[Return to column index]

Commentary, Second Japanese Interphone Study¹

As we will see, this Interphone study discussed here has two methodology improvements, not previously seen in published Interphone brain tumor studies. The two improvements are:

1. Looking at three categories of Specific Absorption Rate (SAR)² exposure within the brain tumor.
   
   However, one table (Table 2), for various indices of cellphone use, reports results for glioma,³ meningioma, ⁴ and pituitary adenoma.⁵ This table considered all tumors as exposed. It is not clear why tumors of the pituitary gland were investigated, for as the study states, "[P]ituitary adenomas occur [at the base of the skull], where RF exposure is negligible..." As a result, this study, while introducing two improvements not previous seen in Interphone studies, makes only a partial improvement in one.
    
2. Control subjects eligible for participation were not told their participation would be for a cellphone study.

First Improvement

A cellphone's radiation plume penetrates a small volume of the brain. Thus, tumors outside the plume are unexposed tumors. Treating all brain tumors, as in Table 2, as exposed, results in an underestimation of the risk of brain tumors.

This second Japanese Interphone study examines the power absorbed within gliomas and meningiomas (SAR). Previously, no Interphone brain tumor study examined the power absorbed within the brain tumor. This study presents results for three categories of power absorption within the brain. Though hampered by design flaws that underestimate the risk, a 5.84-fold risk of glioma was found [see Results, Risk of Glioma (Brain Cancer)].

Determining the power absorbed at the site of a brain tumor is essential. Yet, even though this study does look at tumors within the radiation plume (Tables 3, and 4), because of other design flaws, it is almost incapable of finding a risk.

Five of these six design flaws independently underestimate the risk of brain tumors. Three of the flaws are resolved in this study. For a thorough examination of these design flaws see Interphone Studies to Date, An Examination of Poor Study Design, Resulting in an Underestimation of the Risk of Brain Tumors.

Second Improvement

When potential controls are asked to participate in a cellphone study, it is reasonable to assume that potential controls using a cellphone are more likely to agree to participate compared to potential controls that are not using a cellphone. The result is a problem called selection bias. Selection bias will underestimate the risk of brain tumors. Not telling potential controls, recruited to participate in this study, it is a cellphone study, eliminates selection bias.

¹ - Takebayashi et al. Mobile phone use, exposure to radiofrequency electromagnetic field, and brain tumour: a case-control study. British Journal of Cancer (2008) 98, 652 - 659.
² - SAR is the amount of energy absorbed is tissue, in this case brain tissue, from cellphone radiation. SAR is measured in Watts per kilogram of tissue.
³ - A glioma is a tumor of the glial cells in the brain. Colloquially it is referred to as brain cancer.
⁴ - A meningioma is a tumor of the meninges that surround the brain and spinal cord.
⁵ - The pituitary (master) gland is located at the skull base.

Six Interphone Protocol Design Flaws and Limitations of this Study

Flaw 1: Selection Bias

This flaw is resolved in the third improvement.

Flaw 2: Inclusion of Tumors Outside the Cellphone's Radiation Plume

This flaw is resolved in the first and second improvement.

Flaw 3: Latency Time and Definition of Regular User

This flaw results in an underestimation of risk and is not resolved in this study (see The limitations of this study below). It is a major contributor to the underestimation of risk. Thus, this flaw alone, is why it is nearly impossible to find a risk of brain tumors.

Latency Time: Latency time is the time from exposure to diagnosis. Data from ionizing radiation, reports brain tumor latency time is between 20 and 40 years (similar to smoking and lung cancer). With only 2 cases (2.4% of all cases) at 10+ years, there is virtually no possibility of finding a risk for 10+ years of use.

Of particular interest is that the number of cases (8) in the first Takebayashi study was 2.4% of all cases, identical to the this Takebayashi study.

Definition of Regular User: The Internet Protocol defines regular cellphone use as use for at least once a week, for 6 months or more with diagnosis at least a year previous. Given the rapid adoption of cellphone everywhere, the vast percentage of cellphone users have used a cellphone for far too short of time for any reasonable brain tumor latency time.

Flaw 4: Children and Young Adults Are Not Included in the Interphone studies

Children, due to their rapid cell growth, are far more at risk to a tumorgenic exposure than adults. The range of age in this study was 30 to 59. Not having younger subjects in the study results in an underestimation of risk.

Flaw 4 results in an underestimation of risk and is not resolved in this study.

Flaw 5: Cellphone's Radiated Power

This flaw is resolve by the inclusion of various SAR metrics.

Flaw 6: Number of Cases Included in a Study

This study has far too few cases, considering the latency time for a tumor to be diagnosed. In total it has 83 glioma cases and 128 meningioma cases. The result: only 2 glioma cases and only 4 meningioma cases had used a cellphone for 10+ years. This study required far more cases (at least 15 times more glioma cases) in order to have meaningful results for a 10+ year brain tumor latency time.

Results: Risk of Glioma (AKA Brain Cancer)

At the beginning of the Discussion section this study states, "No consistent increase [in risk] was observed in the overall risk of glioma or meningioma among mobile phone users ..." The authors are using "consistent" to hide the important results of this study. There are three important results: one is alarming, two raise concerns.

Alarm

For the highest compared to the lowest Cumulative maxSAR-hours (> 10, or > 1,000 hours of use) there was a 5.84-fold risk of glioma. The text comments that the finding was not significant (< 95% confidence). It is outrageous that this finding was described a non-significant without further qualification! The nearly 6-fold risk of brain cancer has a confidence level of 94.9% (7 cases).

Concern

For the "mid-high" category of mean maxSAR⁶, the risk of glioma (15 cases) was 2.98-fold and was a tad lower than 95% confidence. For mean maxSAR with values between 0.001 and 0.01 Watts per kilogram the risk of glioma was 2.30-fold and the confidence was 91% (17 cases). These findings are a reason enough to be concerned about using a cellphone.

Conclusion

While this study made two improvements, the existing flaws, particularly the short latency time, made it nearly impossible to find a risk. Because the study did find risks, it should result in a world-wide set of public health actions to substantially reduce the absorbed cellphone radiation. For example, the consistent use of a wired headset and not allowing children to use cellphones.

The 5.84-fold glioma risk found for > 1,000 hours (2 cases) of use it confirms the Hardell team's findings of a 1.3-fold risk, for < 1,000 hours (355 cases), a 1.8-fold risk for 1,001 to 2,000 hour (26 cases), and a 3.7-fold risk for > 2,000 hours (21 cases)⁷. However, this study found a much higher risk than the Hardell study. Suggesting, because the > 1,000 hours was > 1,000 hours at the highest value for mean maxSAR, that higher SAR values create higher risk.

Bottom line: in spite of the BBC's headline "Mobiles 'not brain cancer risk'", the risk of brain tumors from cellphone use exists.

⁶ - Mean maxSAR is the average maximum for all the phones used by a subject.
⁷ - Hardell et al, Pooled analysis of two case-control studies on use of cellular and cordless telephones and the risk for malignant brain tumours diagnosed in 1997-2003. Int Arch Occup Environ Health. 2006 Sep;79(8):630-9. Epub 2006 Mar 16.