The use of a non-biological, bridging, antiprotrusio cage in complex revision hip arthroplasty and periacetabular reconstructive oncologic surgery. Is still today a valid option?: A mid/long-term survival and complications’ analysis

Revisions of total hip arthroplasty (rTHA) and periacetabular resections due to primary or secondary bone malignancies often pose surgeons to face the same problem, a major acetabular bone loss [1‐3]. This is usually combined to poor quality of the surrounding pelvic bones becoming a true surgical challenge, particularly when bone loss results in compromise of acetabular column support. Several methods of acetabular reconstruction along with various biologic and non-biologic materials to supplement the periacetabular bone stock have already been described in literature, each one with its strengths and weaknesses [4, 5]. However, what finally drives the surgeon through the choice among different types of reconstruction techniques is the degree of the bone loss itself [2, 6]. Nowadays the customized triflange 3D printed acetabular components [7] or other custom-made prostheses [8, 9] are becoming more and more popular. However, Burch–Schneider-like antiprotrusio cages, combined or not with massive allograft or morselized grafts and/or cement, still remain a useful and relatively low-cost method in case of severe acetabular/periacetabular bone loss otherwise not bridgeable with simple cementless hemispherical trabecular metal cup [10‐13]. Considering the fact that those cages do not rely upon a biological fixation but are simply mechanically fixed to the pelvis, there is certainly a risk for cage breakage and subsequent failure over time. This is even more likely in juvenile high-demand patients who face against periacetabular malignant bone tumors [13]. Furthermore, this is a complex surgery that often requires prolonged surgical time and large surgical exposures in patients with comorbidities related to their age or treated preoperatively with radio- and/or chemotherapy, with all the consequences that it carries in terms of intra- and post-operative complications. Indeed, due to the intra-pelvic extension of the malignancy or its closeness to the mayor iliac vessels as well as the presence of a cup protrusio or a huge intra-pelvic pseudotumor, surgical approaches could range from the classical postero-lateral approach to more invasive ones, such as extended Smith–Pethersen ilio-femoral approach or Enneking’s modified ilio-femoral approach. Those approaches are often more invasive for the surrounding soft tissues and they can lead to a higher short-term post-operative instability/dislocation rate irrespective of the type or the position of the prosthetic component used [10, 14].

Therefore, the aim of our study was to evaluate the clinical outcome, and estimate both the risk of cages’ failure and complications in a population of patients in which we implanted Burch–Schneider-like antiprotrusio cage due to failure of THA or periacetabular bone loss after resection of primary and secondary bone malignancies. Moreover, we investigated whether the surgical approach as well as the presence of a proximal femur replacement or the reason for implanting the antiprotrusio cage (revision THA versus oncologic surgery) could result a risk factor enhancing the chance of dislocation and/or infection and therefore predict the outcome of such complex patients.

One of the aims of the study was to investigate whether this non-biological type of acetabular reconstruction could keep its mechanical strength with time at mid/long-term follow-up, independently of the nature of acetabular/periacetabular bone loss. Taking into account the competing risk of death, in our series, the risk of failure was 30% and 10% at 10 years for the rTHA and oncologic group, respectively, with no differences between the two groups.

The first long-term cages’ survival rate was reported by Berry and Müller [23] in which they found 76% survival rate in 32 hips at 5 years of follow-up. More recently, in the review published by Aprato et al. [24], the survival rate of those particular type of reinforcement rings among 13 different articles ranges from 72 to 100% at a medium follow-up of 5.6 years.

Rowell et al. [25] were the first to report the treatment of 47 acetabular bone metastasis with the use of PPR link system between 2006 and 2017. They found a surprisingly high survivorship of 91% free from all-cause revision or reoperation, probably due to the short survival expectancy of metastatic patients.

Focusing on structural/mechanical failure of the cage itself (aseptic loosening/fixation failure/cage breakage), Hsu et al. [14] described a total of six (19.4%) failures with associated component migration at the latest follow-up (10 years), reporting that three implants had broken iliac screws but no cage breakages occurred. Among our series, we reported only four cases (5.5%) of structural/mechanical failure that required a re-intervention equally distributed between the two groups.

Udomkiat et al. [26] retrospectively reviewed a series of 18 Burch–Schneider cages and found a 6-year survivorship for mechanical failure of 63.6% including both cages revised for aseptic loosening and those that had radiographic loosening but had not been revised.

In line with the previously reported literature [11, 23, 25, 27‐31], we also observed good functional results with a median modified Harris hip score of 61 and no significative differences among the two groups. Nevertheless, we reported a high incidence of Trendelenburg gait (24%) among both groups, but we found no association in between this particular type of complication and the post-operative dislocation rate (F = 0.532). Differently from our report, Udomkiat et al. found that 38.8% (7 out of 18 patients) of their cages’ series had a dislocation. They described that the dislocation rate was related to muscle weakness (p = 0.028).

Among our study population, we reported a high incidence of complications. Almost half of our patients had at least one complication (42%). Taking into account the competing risk of death, the risk of getting complication was 53% and 30% at 10 years for the rTHA and oncologic group, respectively, with no differences among the groups. The most common complication was dislocation (primary dislocation in 24.7% of patients) followed by infection (13.7% of patients).

In our study, we observed the surgical approach/exposure along with the presence of a complementary proximal femur replacement as the two major independent risk factors for dislocation. Indeed, the odd ratio of dislocation resulted three times higher (p = 0.02, risk ratio = 3.2) in those patients who underwent an extended ileo-femoral approach and two times higher (p = 0.04, risk ratio = 2.1) for those who also had a proximal femur replacement. This can be related on the fact that this surgical approach was most frequently used in oncologic or challenging cases, in the presence of huge endopelvic pseudotumor in contact with iliac vessels or endopelvic cup protrusion associated with pelvic discontinuity, in which a postero-lateral approach could not be the first choice. Moreover, all the patients (100%) of the rTHA group who got a dislocation underwent also a proximal femur replacement and, actually, every patient of the oncologic group who had a proximal femur replacement (four patients) got a dislocation as well, independently of the surgical approach used. Based on the fact that during the implantation of a proximal femur replacement prostheses both the gluteus muscles and the psoas muscle have to be detached, it also remarks the importance that soft tissues play in the stability of a revision THA. It means that both factors, an extensive invasive surgical approach and the presence of a proximal femur replacement, could justify the reason why we had that high rates of dislocation in both groups (dislocation rate: 35% in group A and 30.3% in group B).

A lower dislocation rate was reported in the periacetabular metastatic series of patients by Rowell et al. [25] who described four events in 46 patients (9%). A slightly higher rate of dislocation (17%) was described by Mark Clayer [12] in a series of 29 patients undergoing implantation of an anti-protrusio cage for metastatic pelvic disease. Such low dislocation rates could be related to the short survival of metastatic patients. Indeed, in the latter report, the median length of patient survival was only 12 months (3 days–100 months) after the procedure.

Regarding the second-most common complication, we found that implanting an anti-protrusio-cage due to failure of a previous THA or revision THA was an independent risk factor in predicting post-operative infections. Indeed, the infection rate among the two groups was, respectively, 2.5% and 24.2% in the oncologic group and revision THA group and the estimated odds ratio of infections was more than two times higher (p = 0.04, risk ratio = 2.6) for the latter group. Despite no previous study has compared those two different categories of patients, single literature reports are in line with our findings concerning the oncologic group (percentage of infection ranging from 2 to 8%) [12, 25, 27] but the infection rates among the revision THA are reported to be lower than what we have found (percentage of infection ranging from 3 to 10%) [14, 30, 31]. We believe that higher infection rate in our revision THR series can be related to previous multiple surgeries, with an average of 2.9 procedures before the index operation.

This study has several limitations. We studied two non-homogeneous groups of patients, but our initial purpose was to evaluate the durability of a non-biologic implant in bypassing a gross acetabular/periacetabular bone defect independently of the cause of the bone loss itself. Furthermore, the two groups have in common a poor host bone healing potential, either because of the scarce quality of bone that sustained many previous surgeries [32] or because the bone lesion (e.g. metastasis) is not expected to heal [33] or because the periacetabular bone resection (e.g. primary bone tumors) has been bypassed with a massive allograft [34] that has no capacity to support the bone growth over an ingrowth material (e.g. trabecular metal).

Second, the sample size of our study is not big enough to give a proper consistency at the statistical analysis performed, therefore the interpretation of data cannot be taken as milestone.

Finally, despite we performed X-rays at every out-patient control, we have only recorded those cases that required a subsequent revision surgery due to cage loosening/migration/structural breakage. This could therefore lead to an underestimation of intrinsic cages’ mechanical failures.

As far as we know, there are no previously reported studied in literature comparing those two groups of patients with the purpose to evaluate the potentialities and weaknesses of this particular bridging acetabular cage implant, independently of the nature causing the acetabular/periacebular bone defect as well as to investigate whether or not there could exist some difference in complication rates among the two different groups and if those could be related to some specific risk factors.

Burch–Schneider-like antiprotrusio cages are useful implants when facing against different types of gross acetabular/periacetabular bone losses, with an overall high long-term survival rate. However, complication rate is high, mostly dislocations and infections. Whenever possible, we suggest both to spare the proximal femur along with its muscular insertions in rTHA, and to use a less-invasive postero-lateral approach to reduce dislocation rates.