Another potential limitation of our study is the lack of control for median sternotomy. Although we had internal controls to assure that no significant changes in the bony thorax occurred with time, we were unable to control for possible changes that can occur after median sternotomy without LVRS. We know that after median sternotomy there is a reduction in rib cage and spine motion that contributes to a restrictive chest wall process. The reduction in the bony thorax dimensions that we measured, however, correlated with RV and TLC, suggesting that the changes were not attributable to the effect of median sternotomy alone.
There are several strengths of our study design. It is a controlled, prospective study, in which two independent investigators made measurements using two roentgenographic modalities, coupled with a 12-month follow-up period. The seven patients whose measurements served as reference data helped solidify our results by demonstrating that the small improvements in thoracic cage configuration found postoperatively were not merely caused by variations in two roentgenograms taken at different times. To further strengthen the standard chest film findings, CT scan measurements were obtained as well. The CT measurements confirmed the AP diameter reductions measured on plain roentgenograms and strengthened their significance; however, they did not confirm the small transverse diameter changes seen with plain films. This difference presumably reflects the better resolution of CT measurements compared with plain roentgenogram measurements. Unlike plain roentgenograms, CT images are not hampered by superimposition of soft tissues. We repeated the CT measurements 12 months postoperatively and demonstrated stability of the lower thoracic AP diameter reductions seen 3 months after LVRS.

In conclusion, we evaluated by CXR the bony thorax of select COPD patients undergoing bilateral LVRS. Using two techniques, we demonstrated that a change in rib cage configuration occurs after LVRS, especially in the lower AP axis. This improvement in thoracic shape remained stable for a 12-month follow-up period. Although the demonstrated changes are small, they likely occur in all three dimensions, and are probably underestimated with two-dimensional measurements. The strong correlations between the roentgenographic bony thorax changes and physiologic measurements suggest a possible mechanism for improvement in respiratory mechanics after LVRS, and suggest that bony thorax configuration may be an important factor affecting diaphragm pressure generation.