Lloyd Morgan's Column
[Return to column index]
A Comparison Between Two Interphone Studies
Risk of Parotid Gland Tumors from Cellphone Use
Introduction
The Interphone Study is a 13-country case-control study on the risk of brain
and parotid gland tumors as a result of cellphone use. In order to roll-up the
results from all 13 countries, each of the individual country studies are
required to use the Interphone Protocol.
To date (March 2008) there have been two Interphone studies on the risk of
parotid gland tumors (a salivary gland near the ear) from cellphone use. The
two studies, using the same Interphone Protocol, reported diametrically
different results. How could this be? The quick answer is that one study had
far more tumors (cases) than the other. The longer answer is that one study
restricted their study design to the highest levels of exposure, while the other
study made no such effort.
What did each of these studies find?
The most recent study, an Israeli study by Sadetzki et al. reported, "[The]
analysis [was] restricted to ... conditions that may yield higher levels of
exposure (e.g., heavy use in rural areas) showed consistently elevated risks.
For ipsilateral[1] use, the odds
ratios in the highest category of cumulative number of calls and call time ...
were 1.58 (95% confidence interval: 1.11, 2.24) and 1.49 (95% confidence
interval: 1.05, 2.13), respectively." Additionally, "A positive dose-response
trend was found for these measurements." In other words, this study found a
significant[2] 58% increased risk
of parotid gland tumors for the total number of cellphone calls and a
significant 49% increased risk for the total number of cellphone hours. Perhaps
of greater importance, they found the risk increases as the number of calls or
the number of hours increases.
The earlier Swedish-Danish study, by Lönn et al., led by the prestigious
Karolinska Institute, reported, "For regular[3] mobile phone use, regardless of duration, the risk estimates for
malignant and benign tumors were 0.7 (95% confidence interval: 0.4, 1.3) and 0.9
(95% confidence interval: 0.5, 1.5), respectively. Similar results were found
for more than 10 years' duration of mobile phone use." In other words the study,
in contrast to the Israeli study, reported no risk regardless of duration
of use even when the duration exceeded 10 years. As we will see, though
Lönn et al. reported no risk, they did find a risk.
Let's examine the Israeli study and compare it to the Swedish-Danish study.
1 - Ipsilateral means the tumor is on the same
side of the head where the cellphone was held.
2 - Significant (as well as non-significant) is
short had for statistically significant.
3 - "Regular" use is defined as use of a
cellphone for at least once a week for 6 months or more (with tumor diagnosis at
least a year prior to the cut-off date of the study). If "regular" smoking and
risk of lung cancer replaces "regular" cellphone use and risk of brain tumors
and/or parotid gland tumors, would a risk of lung cancers be reported?
Comparison of Studies
There are three reasons why the Sadetzki et al. study found a risk of parotid
gland tumors where the Lönn et al. study did not find a risk. First,
Israeli cellphone users are exceptionally heavy users (both in call time and in
the number of calls). The result was a substantially greater number of tumors in
the Sadetzki et al. study compared to the Lönn et al. study. Second, the
Israeli study was "restricted to ... conditions that may yield higher levels of
exposure" while the Lönn et al. study made no such effort. Third, the
Lönn et al. study did not present any ipsilateral data for cumulative call
hours or cumulative number of calls, much less for the highest category of time
and number of calls as did Sadetzki et al.
Why is "heavy" cellphone use important when investigating tumor risk from
cellphone use? Heavy cellphone use increases the total exposure.
Why is it important to compare rural cellphone use to urban cellphone use?
The quick answer is that the power radiating from a cellphone is higher in rural
areas than in urban areas. Simply stated, rural users are farther from a cell
tower (base station, mask) than urban users. As a result the power of cellphone
must he higher.
Why is it important to report ipsilateral cellphone use for the highest
category of cumulative use time and cumulative number of calls? Because
ipsilateral use (contralateral has no exposure) combined with the highest
category of cumulative use time and/or number of calls results in the highest
level of exposure. The Lönn et al. study made no such effort.
Below we present Table 1. It is a side-by-side comparison of the two studies.
Findings from Table 1
- Number of tumors
With a much smaller eligible population, Israel, for every parameter, had
2.7-fold more tumors. This difference goes a long way towards explaining why
the studies have such diametrically different results.
- Eligible study populations
The eligible population in Israel (~3.2 million) is less than half the
eligible population in Sweden and Denmark (6.7 million). Given the larger
population in Sweden and Denmark we would expect there would be more tumors.
But, it is just the opposite.
Why is this? It suggests the prevalence of cellphone users was greater and
the amount of exposure was larger in Israel than in Denmark and Sweden. The
Lönn et al. study could have, but did not, resolve their dearth of tumors
by using a longer eligibility range (first parameter in Table 1).
- Ten or more years of cellphone use
In an examination of cellphone use for 10 years or more there were 1.8-fold
more tumors in Israel. The higher 1.4-fold risk reported in the
Swedish-Danish study is meaningless because the confidence level is quite
low (51%), a direct result of having too few tumors.
Table 1, Parameter Comparison
Israeli Study |
Swedish-Danish Study |
Israeli OR[4] (CI) |
Swedish-Danish OR (CI) |
Tumor Ratio I: S-D |
Jan. 2000-Dec. 2003 |
Sept. 1, 2000-Aug. 31, 2002 |
NA |
NA |
NA |
Malignant, 59 Benign, 402 |
Malignant, 60 Benign 112 |
NA |
NA |
2.7 |
Malignant, 4 Benign, 12 |
Malignant, 2 Benign, 7 |
0.93 (0.44-1.98) Benign |
1.4 (0.5-1.39) Benign |
1.8 |
Malignant, 6 Benign, 75 (>18,997 calls) |
Malignant, 6 Benign, 21 (>7,350 calls) |
1.09 (0.74-1.62) Benign |
1.0 (0.5-2.1) Benign |
3.0 |
Benign, 148 (>5 years, >266.3 hours) |
Benign, 22 (>450 hours) |
1.44[5] (0.99-2.09) Benign |
1.0 (0.5-2.1) Benign |
6.7 |
10 (Malignant and Benign # not reported) |
Malignant, 1 Benign, 6 |
1.97 (0.81-4.85) Malignant & Benign |
2.6 (0.9-7.9) Benign |
1.4 |
104 (>5,497 calls, Malignant and Benign # not reported) |
Not Reported |
1.49 (1.03-2.10) |
NA |
98 (>266.3 hours, Malignant and Benign # not reported) |
1.45 (1.00-2.10) |
41 (Malignant and Benign # not reported) |
1.39 (0.40-4.84) |
41 (>18,997 calls, Malignant and Benign # not reported) |
1.81 (1.04-3.14)[6] |
38 (>1,035 hours, Malignant and Benign # not reported) |
1.96 (1.11-3.44) |
4 - OR is Odds Ratio is the Relative Risk of
cases (people with tumors) compared to controls. For example, if an OR = 2.2,
there relative 2.2-fold risk for cases compared to controls.
5 - Bold indicates an Odds Ratio is
near-significant: >90% confidence equivalent to p<0.10.
6 - Bold italics indicates an Odds Ratio is
statistically significant, >95% confidence (p<0.05).
- Tumor on the same side of head where cellphone was held
When the tumor reported is on the same as where the cellphone was held
(ipsilateral use), there were 1.4-fold more tumors in Israel. Interestingly,
the Swedish-Danish study listed in a table within their study, a
near-significant[7] 2.6-fold
risk for 10 years or more of ipsilateral cellphone use. Yet, other than
listing this result in a table, the elevated risk was not reported within
the text of the paper.
- Cumulative number of calls
When we look at the cumulative number of calls there were 3.0-fold more
tumors in Israel compared to Sweden and Denmark Further, for the highest
category of cumulative number of calls there were 2.6-fold more calls in
Israel compared to Sweden and Denmark.
- Cumulative number of hours
We see in the Israeli study, for cumulative hours of use, 6.7-fold more
tumors and 2.7-fold more cumulative hours.
- Rural use compared to urban use
Lastly, the Swedish-Danish study did not present any rural/urban data.
With every parameter, except two, the risk was higher in the Israeli study:
Regular cellphone use for >10 years (statistically meaningless), and
ipsilateral use for >10 years. The latter finding was not reported in
the text of the Swedish-Danish paper.
Table 2, compares all significant and near-significant risks for ipsilateral
cellphone use. Why is this important? The highest level of exposure is on the
side where the cellphone is held. There is no exposure on the opposite of the
head (contralateral).
The Swedish-Danish study had only one near-significant finding. The
equivalent risk in the Israeli study is nearly the same as the Swedish-Danish
study's risk (2.53 and 2.6 respectively).
The results from Table 2 are not surprising. We know that ipsilateral use of
a cellphone is the highest exposure, compared to the opposite side
(contralateral) where there in no exposure. Indeed, if risk is reported without
regard to the side of the head (which is the usual case in all studies), the
result will be an underestimation of risk. For example, the 1.96-fold risk in
Table 1 is an underestimation. In other words, the risk is greater than
1.96-fold.
Swedish-Danish Study's Reported Risk
Even with six differences between the Israeli study and the Swedish-Danish
study, each strengthening the Israeli study, the Swedish-Danish study found a
risk, but it was not discussed. For >10 years since first regular ipsilateral
cellphone use a table reported a 2.6 fold near-significant risk (p=0.078, or 92%
confidence). Why didn't the authors state, "There is a near-significant elevated
risk for >10 years of ipsilateral use?"
Table 2, All Reported Ipsilateral Results: Significant and
Near-Significant
Israeli |
Swedish-Danish |
OR (CI) |
Tumors[8] |
OR (CI) |
Tumors[9] |
2.17[10] (0.91-5.18) |
10 |
Not Reported |
2.53 (0.97-6.53) |
10 |
2.6[11] (0.9-7.9) |
8 |
0.71[12] (0.48-1.04) |
50 |
Not Reported |
1.49[13] (1.03-2.14) |
104 |
1.45 (1.00-2.10) |
98 |
1.42 (0.94-2.15) |
70 |
0.46[11] (0.20-1.07) |
13 |
1.48 (0.99-2.0) |
70 |
How Benign are "Benign" Tumors
In both studies, the reported risk was only for "benign" tumors. Not to worry
you say? After all, benign means harmless.
"Benign" tumors - are not benign!
For the researchers, patients, and other readers of science papers, "benign"
creates an expectation that there is nothing to be concerned about.
Let me describe the experience of having a "benign" parotid gland tumor.
First, it must be surgical removed. If not removed it continues to grow until a
facial nerve is damaged. The result is partial or complete paralysis of one side
of the face. Though the surgery is relative simple (one to two nights in the
hospital), some surgeons will leave part of the parotid gland in place for fear
of cutting a facial nerve (with the possibility of tumor re-growth). A common
side effect is a loss of facial feeling. In most cases the feeling returns after
a couple of years. Is this harmless (AKA "benign")?
Interestingly, if the "benign" parotid gland tumor is larger than a certain
size, it is arbitrarily called malignant.
"Benign" tumors, just like "malignant" tumors, are the result of genetic
mutations. These mutations cause the tumor cells to grow without limit (unless
treated). It would be far better to refer to these tumors as "non-malignant", or
even, "low-grade cancer." The Central Brain Tumor Registry of the United States
will soon publish its 2007-2008 Primary Brain Tumors In The United States,
Statistical Report. This edition compared to previous editions, changes the
term "benign" to "non-malignant."[14]
The Swedish-Danish study states, "Benign parotid gland tumors are not
reported to the cancer registries in any of the countries." This was also true
in the United States for "benign" brain tumors until the US Congress
unanimously passed a law mandating data collection of "benign" brain
tumors. Similar legislation for "benign" brain[15], parotid gland, and facial nerve tumors is necessary
in all countries, in order to learn, delayed by a latency time, if the incidence
of these tumors increases as a result of cellphone use.
8 - Benign and malignant tumors combined.
9 - Only benign tumors.
10 - Bold indicates near-significant (p<0.10).
11 - Since first regular use.
12 - Anomalous finding: Ten % of
near-significant results will show a protective effect due to chance.
13 - Bold-italics indicates statistical
significance (p<0.05)
14 - For a copy of this Report go to www.cbtrus.org.
15 - Acoustic neuromas and meningiomas are
"benign" brain tumors exposed to cellphone radiation.
Summation
The Israeli study, using the same Interphone Protocol as the Swedish-Danish
study was designed to emphasize the subjects with the highest exposure (rural
users and ipsilateral cellphone use.
Why wasn't a similar design used in the Swedish-Danish study? There are two
answers to this question. First, there was no effort made to design an approach
that would emphasize highest cellphone exposures (rural users and ipsilateral
use). Second, there were too few tumors. If there had been a design emphasis for
the highest levels of exposure, then it would have been clear that there would
be too few tumors. The solution? Extend the diagnoses eligibility time in order
to capture a sufficient number of tumors.
The rhetorical question must be asked. What was the reason for choosing an
approach that, as reported, found no risk?
[Return to column index]
|