An inverse association between tumour size and overdiagnosis may explain the results by Bucchi et al.

Sir,

In a recent paper, Bucchi report that the risk of axillary lymph node metastases for screening detected cancers is increasing more rapidly than for clinically detected cancers even after adjusting for other risk factors. They argue that overdiagnosis may not explain their results. I believe that overdiagnosis actually may explain their results.

Further, Bucchi et al claim there is ‘no published data support on an inverse association between overdiagnosis and tumour size’. Actually, there are a lot of recent papers on the association between mammography screening and overdiagnosis of small tumours (Olsen and Gotzsche, 2001); (Zahl ). Most recently, Joensuu reported an extremely high level of length time bias in screening detected cancers in the Finnish mammography screening programme.