One of the main reasons to read papers is to learn things. Preferably new things. This is one of those papers where you can pick up a few ideas for analyzing or looking at data or measuring something that you may not have initially thought of. This was certainly the case for me while I read this paper.
There are two effects of additional filtration: reduced patient entrance exposure (from removal of low-energy x-rays) and increased tube loading (from increased technique to compensate for the radiation removed by the additional filtration).
The change in mAs for varying amounts of added filtration and different projections was examined under three conditions: constant patient entrance exposure, constant patient exit exposure and constant film optical density. If I was doing the same study, I probably only would have thought to look at entrance exposure (since it is easiest to measure) or CR exposure index (we use digital systems here).
Maintaining constant entrance exposure required the highest increase in mAs for a given amount of filtration, while maintaining constant film density required the lowest increase. To avoid increasing the amount of image noise with added filtration, I think the ideal parameter to maintain would be constant optical density (or receptor exposure for digital systems). Since maintaining constant optical density is more labour intensive to measure, constant exit exposure is probably and easier parameter to work with (and already measured by phototimer systems). Both methods resulted in similar increases in mAs.
Increased technique also affects image quality due to focal spot blooming when the tube current is increased. However this is normally a very small effect, which was confirmed by the authors. There is also the potential for increased motion blur when exposure time must be increased, but with the constant exit exposure/optical density methods, the increase in exposure time was small enough so that motion was not a problem.
Even with 4.0 mm added filtration, the increase in mAs should be well within the range of x-ray units to accomodate by increasing mA and/or time without adversely affecting image quality.
Overall I thought this was a well done paper that clearly showed that the impact of added filtration was minimal and easily accomodated by most x-ray systems.