Attenuation correction may improve the diagnostic accuracy of myocardial perfusion imaging (MPI). However, few studies have dealt with the clinical consequences for reporting. We compared routine reports based on scatter-corrected MPI (MPI-routine) with consensus readings of scatter-corrected (MPI-scatter) and scatter plus attenuation-corrected studies (MPI-attenuation) to investigate the impact of attenuation correction on reporting. One hundred consecutive stable angina patients (including 55 men) were investigated in a 99mTc-sestamibi 2-day gated protocol with scatter and attenuation correction. With MPI-routine, 53 patients had normal perfusion and 47 abnormal perfusion, compared to 62 and 38 with MPI-attenuation, and 54 and 46, respectively, with MPI-scatter. Agreement between MPI-routine and MPI-attenuation with respect to overall diagnosis (normal/abnormal perfusion) was 89% (kappa=0.78) compared to 95% (kappa=0.90) between MPI-routine and MPI-scatter. With MPI-attenuation, the overall routine diagnosis changed in 11 patients, of which ten cases were judged normal after scatter plus attenuation correction. The majority of the "normalised" studies were among patients with apparently single-vessel RCA disease as judged from MPI. Agreement rates with regard to normal, reversible or irreversible defects between MPI-attenuation and MPI-routine for the LAD, LCX and RCA territories were 88%, 97% and 85%, respectively, without significant sex differences. In conclusion, attenuation correction caused a change in diagnosis in approximately 10% of the patients, corresponding to one-fifth of the abnormal studies. In all but one case, the shift was from abnormal to normal, mostly because of a different interpretation in the RCA territory.