Megavoltage linac photon outputs are determined using the IAEA TRS-398 and TRS-277 codes of practice and the results compared. It is found that for two types of thimble chambers the contributions to absorbed dose to water due to the two codes of practice are comparable and it is suggested that observed differences are mainly due to the different standards used to determine ND,W and NK calibration coefficients at a standards laboratory. A method is proposed to determine a priori the expected difference between absorbed dose to water from the two codes of practice. The contribution to the difference that depends on the respective codes is expressed by a code dependent factor fC. It is shown that for Wellhöfer FC65-G and NE 2571 ionisation chambers this factor is largely independent of beam quality. To calculate the expected shift in absorbed dose to water when moving from TRS-277 to TRS-398, an fC. value which is representative of the range of beam qualities in the clinic is calculated and multiplied by the ratio ND,W/NK.