The answer to the question of whether cellphones increase the risk of brain, head and neck tumors is truly a matter of whom you ask.
An analysis published Tuesday of data from 23 epidemiological studies found no connection between cellphone use and the development of cancerous or benign tumors. But when eight of the studies that were conducted with the most scientific rigor were analyzed, cellphone users were shown to have a 10% to 30% increased risk of tumors compared with people who rarely or never used the phones. The risk was highest among those who had used cellphones for 10 years or more.
"The other group of 15 studies were not as high-quality," said study coauthor Joel M. Moskowitz, director of the UC Berkeley Center for Family and Community Health. "They either found no association or a negative association or a protective effect -- which I don't think anyone would have predicted."
The main message of the analysis, published in the Journal of Clinical Oncology, is that studies should be conducted so that findings are harder to refute, he said.
In recent years, concerns have arisen that the radio-frequency energy emitted by cellphones may be high enough to cause tumors and other health problems. But the risks are hotly debated.
"I went into this really dubious that anything was going on," Moskowitz said. "Overall, you find no difference. But when you start teasing the studies apart and doing these subgroup analyses, you do find there is reason to be concerned."
All of the studies were case control studies, which means researchers interviewed people on their past use of cellphones. Some of the people, referred to as controls, had no history of brain tumors; others, known as cases, had been diagnosed with brain tumors. The studies encompassed 37,916 people.
Eight of the studies were singled out as more reliable because the researchers were not told which people had tumors and because the studies were not supported with mobile phone industry funding. However, seven of those eight studies were conducted by a single researcher, Dr. Lennart Hardell, an oncologist in Sweden.
Some of the less-stringent studies were part of the Interphone project coordinated by the World Health Organization's International Agency for Research on Cancer. Interphone is funded in part by the Mobile Manufacturers Forum and the Global System for Mobile Communication Assn.
"Hardell had the higher-quality studies in which he was blinded to the cases and controls," Moskowitz said. "Presumably, he would have less opportunity to bias the results. But was it better methodology? Or was it something about Sweden? More people there live in rural communities, and maybe they were exposed to higher levels of energy."
More radio-frequency energy is typically needed to operate cellphones in rural areas.
Interphone study investigators are not influenced by the funding source, said Michael Milligan, secretary-general of the Mobile Manufacturers Forum, based in Hong Kong.
"The Mobile Manufacturers Forum has provided part funding for the Interphone study that has complemented other public sources of funding -- such as the European Commission and other national bodies," Milligan said in an e-mail. "In providing funding, we have done so on terms that guarantee Interphone's complete scientific independence."
The bulk of all scientific studies, Milligan said, reveals no increased risk of head and neck tumors. However, conclusions should await higher-quality studies that follow diverse groups of people, both phone users and nonusers, over a long period of time, said Dr. Seung-Kwon Myung, lead author of the meta-analysis, from the National Cancer Center in Goyang, South Korea.
"Larger, prospective, cohort studies, independently conducted from the mobile industry, are required to confirm the relationship between mobile phone use and tumor risk," he said.