Poster 312: Graft harvest for ACL reconstruction from the contralateral knee does not increase patellofemoral arthritis rates versus a normal knee

OBJECTIVES: Previous outcome studies after anterior cruciate ligament (ACL) reconstruction have suggested a link between patella tendon (PT) graft harvest and increased incidence of patellofemoral (PF) degenerative changes. However, with the graft harvest on the same knee as the ACL reconstruction, it is difficult to specifically elucidate whether this association is directly from the graft harvest itself as opposed to other factors. Knowledge of the incidence of PF osteoarthritis (OA) after graft harvest from the contralateral knee may provide insight into this relationship. The purpose of this study was to compare PF OA rates between contralateral versus ipsilateral graft harvest for ACL reconstruction. METHODS: A database of long-term follow up of ACL reconstruction patients done by a single surgeon was queried to find surgeries done between minimum 15 years and maximum 25 years ago. Patients were included if they had long term radiographs within this window for evaluation of PF OA. Exclusion criteria included revision ACL reconstruction, bilateral ACL involvement, subsequent graft retear, contralateral ACL tear, PF chondral wear seen at the time of surgery, preoperative or subsequent surgeries to either knee, or the presence of preoperative PF OA. Once patients met inclusion criteria, they were separated into groups based on ipsilateral or contralateral PT graft harvest. Standard Merchant’s radiographs were taken by the same technicians with the assistance of a Merchant’s board. Radiographs were graded by the physical therapist that saw the patient back for their long term research visit, each individually trained to grade the images based on the set criteria. Radiographs were graded as none, mild, moderate, or severe PF OA based on joint space narrowing of none, <50%, >50% but <100%, and complete, respectively. Patellofemoral OA rates were then statistically compared utilizing chi-squared analysis or Fisher’s exact test. These criteria produced an ipsilateral group with 95 patients and a contralateral group with 98 patients. Mean ages at time of surgery, height, weight, and sex distribution were not statistically different between groups. The ipsilateral group had a slightly, but statistically significantly, longer mean follow up interval (20.0 vs. 17.4 years, p<.001). RESULTS: In the ipsilateral graft group, PF OA of any grade was present in 20.0% of the involved ACL reconstructed knees and 9.5% of the uninvolved normal knees, which was a statistically significant difference (p=.041, OR 2.4). In the contralateral graft group, PF OA was present in 10.2% of the ACL reconstructed knees and 13.3% of the contralateral graft donor knees (p=.506). Patellofemoral OA of any grade was present in 14.7% of all knees in the ipsilateral group and 11.7% of all knees in the contralateral group (p=.348). When comparing the uninvolved normal knee from the ipsilateral group to the contralateral graft donor knee of the contralateral group, there was no statistical difference in PF OA rates of any grade (9.5% vs. 13.3% p=.407). The ACL reconstructed knee in the ipsilateral group exhibited PF OA in 20.0% and 10.2% in the contralateral group (p=.057). With regard to moderate or severe PF OA in either knee, the rates were not statistically significant between the ipsilateral and contralateral groups. When comparing the ipsilateral group normal knee to the contralateral group donor knee, the rates of moderate or severe PF OA were 1.1% and 2.0% respectively, which was not statistically significantly different. CONCLUSIONS: Previous studies suggest that patella tendon grafts are associated with higher rates of PF OA after ACL reconstruction. Our study reveals that PF OA rates are not statistically increased when a patella tendon graft alone is taken from the knee, when compared to a normal control knee. However, when ACL reconstruction and PT graft are combined in the same knee, there is a statistically significant increase in PF OA versus the uninvolved normal knee, with the ACL reconstructed knee with ipsilateral graft being 2.4 times more likely to experience PF OA than the uninvolved knee. Furthermore, harvesting the ACL graft from the contralateral knee trends toward decreased PF OA rates in the ACL reconstructed knee, though this is not statistically significant. We conclude that the PT graft itself does not increase PF OA rates and that contralateral PT graft harvest may provide advantages to ipsilateral graft harvest with regard to PF OA.