Percutaneous coronary intervention (PCI) has been associated with excessive radiation exposure, so the present study was designed to investigate the determinants of fluoroscopic time during PCI among characteristics inherent to the patient, procedure and hospital in 388 consecutive patients between November 1996 and March 1999 in 11 hospitals included in the Nationwide Database for Cost Analysis of percutaneous transluminal coronary angioplasty in Japan. Fluoroscopic time, which reflects radiation exposure, was used as the dependent variable in a multiple linear regression analysis. The mean fluoroscopic time was 27±15 min. The negative value of partial correlation of the hospital code (r=-0.374) and acute myocardial infarction (-0.163) indicated that these were explanatory variables that decreased fluoroscopic time. The positive value of the number of intraaortic balloon pump catheters (r=0.144), the number of balloons (r=0.304) and hospital stay (r=0.147) indicated these were explanatory variables that increased fluoroscopic time. The absolute value of r, (r=0.374), of the hospital code was the highest. In the final R² (=0.304) of the model, which is the partial R-square that determined the dependent variable, the hospital code showed the highest value. In conclusion, there was an inter-hospital difference in fluoroscopic time. Although the hospital code was found to be the most powerful determinant of fluoroscopic time, the R² (=0.304) of the model showed so lower value that other hospital characteristics that were not included in the model may influence the fluoroscopic time. (Circ J 2003; 67: 676 - 681)