Intensity modulated radiotherapy (IMRT) has become a treatment of choice in many oncological institutions. Small fields or beamlets with sizes of 1 to 5 cm2 are now routinely used in IMRT delivery. Therefore small ionization chambers (IC) with sensitive volumes 0.1 cm3 are generally used for dose verification of an IMRT treatment. The measurement conditions during verification may be quite different from reference conditions normally encountered in clinical beam calibration, so dosimetry of these narrow photon beams pertains to the so-called non-reference conditions for beam calibration. This work aims at estimating the error made when measuring the organ at risk's (OAR) absolute dose by a micro ion chamber (microIC) in a typical IMRT treatment. The dose error comes from the assumption that the dosimetric parameters determining the absolute dose are the same as for the reference conditions. We have selected two clinical cases, treated by IMRT, for our dose error evaluations. Detailed geometrical simulation of the microIC and the dose verification set-up was performed. The Monte Carlo (MC) simulation allows us to calculate the dose measured by the chamber as a dose averaged over the air cavity within the ion-chamber active volume (D(air)). The absorbed dose to water (D(water)) is derived as the dose deposited inside the same volume, in the same geometrical position, filled and surrounded by water in the absence of the ion chamber. Therefore, the D(water)/D(air) dose ratio is the MC estimator of the total correction factor needed to convert the absorbed dose in air into the absorbed dose in water. The dose ratio was calculated for the microIC located at the isocentre within the OARs for both clinical cases. The clinical impact of the calculated dose error was found to be negligible for the studied IMRT treatments.