Shift and tilt of the bony patella in total knee replacement

Resurfacing the patella or not in total knee arthroplasty is generally still a quite controversial topic. In that context we felt a necessity to get more informations about the natural history of a non resurfaced patella in the prosthetic surrounding than only in comparison with a replaced patella. By the fact that we haven't been replacing patellae apart from special cases since up to five years, we have available a patient selection on which this question can be studied. Thirty primary total knee replacements with documented 2 year's follow up have been investigated by calculation of the radiological patellar shift and tilt relative to the natural groove on preoperative X-rays and to the prosthetic groove in the postoperative evolution. These results were matched with the range of motion obtained after two years and with persisting pain. Only mobile polyethylene inlays had been used. With regard to the patella no difference was found for one single complete tibial tray or two separately implanted unicompartimental tibial trays. As a rule an adaptation of the bony patellar contour to the prosthetic groove was observed with an increased density of the bony contact area with or without a fibrous interstitial layer. Based on their clinical follow up examination seven patients had to be adjoined to a "residual problem group", as they presented persisting pain (3) or a poor range of motion of 90/0/0 degrees of flexion/extension or less. Only in this group we found cases with a postoperative shift over 5mm (2) and a postoperative tilt over 15 degrees(3). Inside this group, also with a significant difference from the group without residual problems, there was furthermore a correlation between persisting pain and an important preoperative tilt and between a diminished range of motion and a considerable preoperative shift. It has to be pointed out that even the cases without long lasting severe malposition before the arthroplasty but with unsatisfactory realignement referring to the patellar tilt (7 degrees-15 degrees) have been found pain free at the last follow up. One patient with a severe secondary increase of disalignement (16 mm/139 degrees) due to an aseptic tibial loosening represents a particular case. After corrective revision of the tibial implants this patient also got again a complete pain relief and a good range of motion without any surgical measure to the patella and in spite of an evident residual tilt. Thus we have to conclude that it is justified to take important pains with the realignement to obtain a residual patellar shift of less than 5 mm. Apart from usual operative steps during the implantation to favour a good patellar tracking, e.g. a high quadriceps release or a medially transposed refixation of the detached tibial tubercle can get necessary. On the other hand a residual tilt up to 15 degrees is well tolerated, as the unresurfaced patella presents an astonishing ability of adaptation to the prosthetic groove by remodelling of the bony contour and structure. Nevertheless in the included cases of severe preoperative malposition, in which we didn't replace the patella because of a good intraoperative bone quality, we had to put up with an increased incidence of late problems even with normalized patellar tracking. In such cases an extended indication to a patellar resurfacing may be considered.