Asthma mortality appeared to increase two-fold in the UK from the mid- 1970s to the early 1990s, but there is evidence of inaccuracy in asthma death certification and so a region-wide investigation was undertaken to assess whether this recorded statistical trend might have been partly or wholly artefactual. A total of 35 respiratory physicians, distributed in panels of three, systemically reviewed the hospital and general practice records of 210 subjects with physician-diagnosed asthma who died in 1991 and 1992. The death certificates indicated that asthma was considered to be the primary cause of death in 103 (group 1), a contributory cause in 70 (group 2) and not relevant in 37 (group 3). There was agreement within the panels that 43% of the subjects had probably never suffered from asthma. Discordance between the panels and the certifying physicians over the correct death certification category was high for group 1 (45% for those aged <65 yrs, 75% for those aged > or =65 yrs and 64% overall) and group 2 (67%), but much less for group 3 (22%). This study concludes that asthma death certification provides a markedly inaccurate picture of asthma mortality, particularly in elderly subjects. Thus, it is speculated that if the magnitude of this source of inaccuracy has increased over the last two decades, the apparent recent increase in asthma mortality may be largely artefactual.