Scientific journals claim that correspondence sections are for post-publication peer review. We compared the conditions for
submission and the bibliometrics of letters-to-editors published in leading medical journals in 2002 and 2007 using journal-derived
information and data from PubMed and Journal Citation Reports. The median time limit for letter submissions decreased from
6 to 3.5 weeks, the median word limit from 400 to 350. The median number of letters per published article was near one in
both years. Only about half of the letters were followed by an author reply in either year. Electronic response systems were
available for four journals in 2007.