Knee PJI accompanied with bone loss is a challenging problem in knee joint revision. Several studies have reported the outcomes of revision knees with bone defects reconstructed by bone allografts. However, limited literature focuses on the same issue in knee PJI cases. Using allogenous bone graft for bone defect reconstruction remains a controversial issue, especially for knee PJI revision. In this current study, we compared the outcomes of knee PJI with bone defect reconstruction by structural allografts to outcomes of general cases in knee PJI without using structural allografts. There was no significant difference in the relapse rate of infection and implant survival rate between the groups. We believe using structural allogenous bone graft in the second stage of knee PJI reconstruction can be safe and feasible.
Two-stage treatment for knee PJI
PJI is one of the most common complications that leads to the revision of TKA(2), and two-stage total knee revision is considered as the proper treatment for PJI with a success rate of around 85%-90%(3, 17, 18). Bongers et al.(19) reported a 20% re-infection rate at five years in 113 PJI knees with 2-stage revision. We performed a 2-stage revision protocol for all knee PJI patients from our database including knees with bone defects. In our results, both groups presented with around 90% of 8-year prosthesis survival rate and approximately 80% of 8-year infection relapse-free survival rate. The outcome of 2-stage treatment for knee PJI was in line with previous studies.
Bone defect management
Bone defects reconstruction is challenging in knee revision arthroplasty(20). When it comes to large bone defects, metal augment is feasible in managing AORI type II bone defects, and structural allograft is the treatment option for AORI type IIB to type III(5, 7). Hockman et al. (21) compared the efficacy of using metal augment to allograft for bone loss treatments in knee revision. The result revealed a 59% failure rate of using metal augment alone, and 48% of the knees required additional structural allograft. It was believed that metal augment reduced the contact surface between the host bone and the implant, making it relatively more unstable. In addition, augment could not manage bone defects over 20 mm depth, and there were risks of fretting and erosion, as well as limitations on bone stock restoration (6, 22). Richards et al.(23) also declared that patients treated with allograft demonstrated better clinical outcomes and lower complication rates than those treated with metal augment.
Metal cones were another treatment choice for knee revision with large bone defects as AORI type III. A systematic review had reported a lower loosening rate of the porous metal cones than structural allograft in revision knees (24). However, the re-infection rate revealed no significant difference between these two methods. In our experiences, it was necessary to remove additional host bones when applying metal cones in particular cases. Besides, metal cones have a limited choice in sizes and are inapplicable to knees with small bone sizes. Furthermore, metal cones may cause an additional financial burden to patients comparing to allograft, which is more cost-effective. All these factors need to be considered when managing large bone defects indicated for both structural allograft and cones systems.
Structural allogenous bone graft
Structural allograft is known for its capability of incorporating the host bone and stress protection. Literature had reported a prosthesis survival rate between 76% and 93% at a 5-year follow-up, showing the capability of structural allograft in treating knee revision cases with bone loss (9, 10, 25). For mid-term to long-term survivorship, Chun et al.(26) reported the results of 27 patients undergoing revision TKA with severe bone defect using a fresh-frozen femoral head allograft with a minimum of 8-year follow up. They demonstrated an improved Hospital for Special Surgery knee score from 46 to 83, and 26 out of 27 patients presented no complications. Engh et al. reported a 91% survivorship at 10 years for femoral head allograft in tibial defects in 46 patients receiving revision TKA(27). Clatworthy et al. presented a prosthesis survival rate of 92% at 5 years and 72% at 10 years in 52 patients with uncontained bone defect constructed with structural allograft in knee revisions (28). In this current study, cases enrolled in the study group were all knee PJI cases with bone defects. We demonstrated a 90.9% 8-year implant survival rate in structural allograft reconstruction knees, indicating a similar outcome compared with the previous literatus though our cases were all PJI revision knees.
Despite the reliable results of structural allograft in revision knees, some complications were still with concerns. The potential risk of disease transmission remains an unresolved problem of using bone allografts(29). Several studies had reported a higher infection rate of using allograft bone, which led to revision failure. Franke et al. reported a study in which 30 patients were treated with allografts for revision TKA, and the infection rate was 10%(10). Bauman et al. reviewed structural allograft for TKA reconstruction with an infection rate of 7.1% (9). Backstein et al. used structural allograft in revision TKA and reported a re-operation rate of 4.9% for secondary infection (30). However, Wang et al. conducted a case series study with revision knee reconstructed with femoral head allograft showed no recurrent infection among PJI cases(31). In this current study, one patient in the study group sustained recurrent infection, while the other 11 patients presented free of infection relapse. The results showed an infection rate of 8.3%, and a 100% 5-year infection-free survival rate and an 80% 8-year infection-free survival rate. Though the infection rate was similar to previous articles, it was acceptable owing to cases in this study were all PJI cases. We supposed the application of structural allograft to knee PJI cases did not increase the infection rate compared to general cases. The outcome can be attributed to the following reasons. First, the bone bank in our institute is under strict regulations which are supervised by the government, and all of the bone grafts were harvested by experienced surgeons (15). Wu et al. had reported the result of using bone allografts from our bone bank for surgery with a relatively low infection rate of 1.2% (15). Second, all allografts used in this study were femoral heads harvested from the patients undergoing total hip arthroplasty in our institute. Based on our previous study, administering prophylaxis antibiotics before surgery resulted in their presence in fresh-frozen femoral heads, which exhibited inhibitory effects against bacteria in vitro after two weeks of deep-frozen storage (32).
For other complications, Backstein et al. reported three cases of resorption and one case of nonunion in 58 patients (30). Franke et al. reported one case of nonunion to the graft and host bone in 30 cases of revision knee using allograft bone (10). Bauman et al. reported two cases of nonunion and three cases of post-operative fracture in 79 cases of revision knee treated by structural bone allograft (9). In our study, no graft resorption, nonunion or fracture was reported during our radiographic follow-up concerning structural allograft. One patient suffered from tibial component breakage two and half years after the reconstruction surgery. The structural allograft was applied to the tibia bone defect without nonunion or resorption by x-ray follow-up.
By comparing our study group results to the control group (PJI cases without allografts) and previous articles on revision knees managed with structural allograft, the results revealed the capability of allograft to resolve bone defects in knee PJI cases. The outcome showed no additional infection relapse rate and a satisfying prosthesis survival rate. This study validated the efficacy of using allograft in solving PJI with bone loss in the two-stage revision procedure.
Limitations
The study has several limitations. First, this was a retrospective study with a relatively small sample size. Second, the size of the control group's bone defects was inconsistent with that of the study group. Some of the cases in the control group might be simple revision cases that presented with minimal bone defects. Generally, the bone defect would be more extensive in the study group. The infection rate in complex revisions knees, which require structural bone reconstruction, might be higher than simple revision or small bone defect cases. However, we presented a different point of view according to our results. Third, we did not compare the functional outcome between the study and control groups, which is also essential for comparison. Nevertheless, the overall KSS was good in our study group after the revisions. At last, the implants were not matched between cases in the study group and control group. The different implants might affect the survival rate and complication rate.