The aim of this study was the production of the first Italian survey of radiation dose in computed tomography (CT) prior to the widespread adoption of multislice CT, in order to have a reference point to facilitate later investigation of dose exposure changes brought by this new CT modality. The collected dose data were compared with diagnostic reference levels (DRLs). The agreement between experimental dose evaluation and Monte Carlo (MC) simulations was investigated. The survey was carried out in 29 Italian hospitals, covered 48 CT scanners and 232 examinations. The dose-length product (DLP) and effective dose (E) values were estimated based on MC simulations for seven clinical protocols using the CT-Dose program. Statistical analysis showed a significant difference (p<0.01) in the DLP between the two methods, with MC values being greater than the experimental ones. For E, the MC values were greater in routine head (8.2%), cervical spine (2.7%) and lumbar spine (2.9%) studies. The weighted CT dose index, the DLP and E were always below the DRLs set by the European Community. This dose survey gives a good but incomplete picture of the Italian CT dose situation and may be useful as a reference baseline for defining clinical multislice protocols in the near future.