Study design and methods
INTERPHONE study group (2010)
4.5 Ecological studies
These studies investigate the occurrence of disease at population level in relation to the prevalence of (a proxy for) exposure in the population. They may analyze for instance the pattern of tumour occurrence over time (either by incidence or by mortality) in geographic entities such as countries, to identify any trends and to see whether these could be explained e.g. by trends in possession or use of mobile phones. Individual data on mobile phone use are not used in these studies. Such studies will inherently be limited by the poor level of insight into trends and patterns of mobile phone use, and hence of actual exposure, particularly for specific age, sex and other population group definitions.
It should be noted that for many countries substantial and wide-spread mobile phone use is relatively recent (Figure 2).
In most Western-European countries approximately half of the population had a mobile phone subscription in the year 2000. In the Nordic countries (Norway, Sweden, Finland and Denmark) the increase started earlier, but was caught up by the other countries around the century mark. By 2005 most people in the countries presented (except France and the USA) owned a mobile phone, but the extent of use is much less certain.
Sato et al. (2010)112 Acoustic neuroma Intensity of phone use and
laterality vs. tumour location and size
Original 51%
≥29 - ≤70 y old Cardis et al. (2011)114 Glioma, meningioma Intensity of phone use; based on
calculated RF energy
Subsample of INTERPHONE Australia, Canada, France, Israel, New Zealand Cases: glioma 42% meningioma 56% Controls: for glioma 36%, for meningioma 40% 30-59 y old Larjavaara et al. (2011)113Glioma Case-specular*
Based on calculated RF exposure
Subsample of INTERPHONE Denmark, Finland, Germany, Italy, Norway, Sweden, Southeast England 63% 18-59 y old * Simulated case
Figure 2 Number of mobile phone subscriptions for some European countries and the USA. Panels B and C show the same data as panel A, but separated for Nordic and other countries. Data from ITU (http://www.itu.int/ITU-D/ict/statistics/explorer/ index.html).
Taking into account that the latency period of brain tumours is likely more than 10 years (see Chapter 6), it is thus possible that any trends in tumour occurrence related to mobile phone use may not yet be visible in most countries, with an exception perhaps for the Nordic countries, since use started earlier there.
In analyzing ecological studies, it has to be realized that trends in mortality can also be influenced by the introduction of more effective treatments and that trends in incidence can be affected by changes in diagnostic techniques.
Ecological studies identified in this search were performed in various countries and totalled 21 publications.115-121,58,122-125,125-133 A summary of the publications
is presented in Table 4.6.
Table 4.6 Ecological studies
.
Reference Tumour type, Endpoint Exposure assessment Time period Country Counsell et al. (1996)126 Brain tumours Trend, not in relation to phone
possession or use
1989-1990 Scotland Howitz et al. (2000)116 Acoustic neuroma Trend, not in relation to phone
possession or use
1977-1995 Denmark Gurney & Kadan-Lottick
(2001)127
Brain tumours Trend, not in relation to phone possession or use
1975-1997 USA, 11 states
Cook et al. (2003)115 Head and neck tumours Trend, in relation to phone possession and exposure (from location of tumour)
1986-1998 New Zealand Inskip et al. (2003)117 Ocular melanoma Right- vs. left sided tumours
(assuming predominantly right sided phone use): trends & contrast pre/post 1995
1974-1998 USA, 5 states & 4 metropolitan areas Hardell et al. (2003)58 Brain tumours, acoustic
neuroma
Trend, not in relation to phone possession or use
1960-1998 Sweden Lönn et al. (2004)119 Primary brain tumours Trend, in relation to phone
subscriptions
1996-1998 Denmark, Finland, Norway and Sweden Muscat et al. (2006)120 Neuronal brain cancers Contrast pre/post 1985, in relation
to phone subscriptions
1973-2002 USA, 5 states & 4 metropolitan areas Nelson et al. (2006)128 Acoustic neuroma Trend, in relation to phone
subscriptions
1979-2001 England & Wales
Röösli et al. (2007)121 Brain tumour Trend, in relation to predicted phone use based on subscriptions
1969-2002 Switzerland Deltour et al. (2009)122 Glioma, meningioma Trends, in relation to general
mobile phone use pre/post mid 1990s
1974-2003 Denmark, Finland, Norway & Sweden Inskip et al. (2010)123 Brain cancer Trends, in relation to phone
subscriptions
1997-2006 USA, 5 states & 4 metropolitan areas (10% USA population) Lehrer et al. (2010)124 Primary brain tumours Relation with subscriptions in
2007; comparison of 19 states
2007 USA, 19 states Johansen et al. (2002)118 Ocular melanoma Trend, in relation to phone
subscriptions
1943-1996 Denmark
Czerninski et al. (2011)129 Parotid gland tumour Trends, in relation to increase in phone use
1970-2000 Israel
De Vocht et al. (2011)130 Brain tumours Trends, in relation to phone
subscriptions
1998-2007 England
Only those investigations using both outcome and exposure are assessed and the most recent investigation of the same data is discussed. These publications are identified in bold type face in Table 4.6.
Two Letters to the Editor were identified and are listed in Annex E. Results from the data extraction are presented in Annex F.
The Committee deemed a quality evaluation of the ecological studies not meaningful.
Kohler et al. (2011)131 Brain tumours Trends, not in relation to phone possession or use
1975-2007 USA, 46 population based cancer (93% USA population) Larjavaara et al. (2011)132 Vestibular Schwannoma
(acoustic neuroma)
Trends incl. birth cohorts, not in relation to phone possession or use
1987-2007 Denmark, Finland, Norway & Sweden
Deltour et al. (2012)135 Glioma Trends, in relation to general
mobile phone use
1974-2008 Denmark, Finland, Norway & Sweden
Little et al. (2012)133 Glioma Trends, in relation to results from
INTERPHONE and Hardell studies
1997-2008 USA 12 SEER regions
5
Chapter