The update of the Danish cohort study of mobile phone use and cancer risk has the serious flaws present in the original publication.1 2 The first flaw in the update is the correct identification of mobile phone subscribers between 1987 and 1995. This resulted in a participation rate of only 58% of identified early subscribers, with the remaining 42% placed in the “non-exposed” category (which, by the authors’ own admission, could well be the heaviest user group in the dataset).

http://www.bmj.com/content/343/bmj.d7899.full?ijkey=znFmXOUrJgbqxWT&keytype=ref#aff-1