For assessment of bone remodelling around total hip arthroplasty using dual-emission X-ray absorptiometry (DEXA), a variety of different systems to identify regions of interest (ROI) have been used, making comparisons between stem designs difficult. The Gruen zones are now widely used for this purpose. We present the results of a randomised clinical trial comparing 2 uncemented stem designs with proximal coating, using a modification of the Gruen zones to allow improved representation of the effect of the implant on bone mineral density (BMD) over time. DEXA-data were used in a randomised trial with 2 years follow up, comparing the uncemented Symax(TM) (n=25) and Omnifit(®) (n=24) stems. The effect on BMD was determined using the 'standard' adapted Gruen zones, and a modification which studied an equal length and position for zones 1 and 7 around both stems, assuring that the same regions in terms of cancellous and cortical bone were compared. The 'modified' regions of interest give lower BMD values around the Omnifit(®) than using the 'standard' Gruen zones (3.6 % in zone 7, p<0.05). The difference with the Symax(TM) BMD values, which had been concealed using the standard Gruen zones, became statistically significant in favour of the Symax(TM) implant. This adaptation can detect a statistically significant difference in bone preservation in zone 7 between stems that would otherwise not have been revealed. We recommend the use of 'modified' Gruen zones for more valid comparison of remodelling caused by different implant designs.