One of authors proposed a creep life assessment method for USC boiler pipes that can consider heat-to-heat variations of the creep property of each welded joint, where the creep property of the welded joint is estimated from that of each base metal. In the method, it is assumed that the creep property of each base metal in actual pipes is approximately constant in the thickness direction of the pipes, and test results with small samples cut from base metals at the outer surface of pipes are useful for representing the creep properties of the pipes. In this work, the assumption was examined for five pipes of Grade 91 steel, which had been used for longer than 100,000 h at USC power stations. The microstructure, chemical composition, hardness, void density and remaining creep life were investigated in detail in the thickness direction of the pipes. No difference was observed for these items, except in an area less than about 0.2 mm from the outer surface of the pipes, which means that the assumption in the assessment method is valid except in this area. Therefore, it is suggested that an effective portion of the sample taken from USC boiler pipes to consider heat-to-heat variations of the creep properties of base metals is the material excluding the area less than about 1.0 mm from the outer surface of the pipes.