Background: Many treatment options are available for the revision of large acetabular defects. Debate continues as to which technique is most effective.
Methods: A systematic review was performed according to the Meta-analysis of Observational Studies in Epidemiology (MOOSE) guidelines to evaluate the effectiveness of interventions for large acetabular defects. Quality assessment was performed next with use of 8 items of the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) checklist for reports of observational studies. Large acetabular defects were defined as American Academy of Orthopaedic Surgeons (AAOS) type III or IV or Paprosky type 3A or 3B. Outcomes included re-revision, radiographic loosening, complications, and clinical outcomes.
Results: We found 7 different treatment options for large acetabular defects in 20 included studies: antiprotrusio cage (8 studies), Trabecular Metal (Zimmer) augment and shell (4 studies), bone impaction grafting with a metal mesh (2 studies), hemispherical implant with hook and flanges (2 studies), Trabecular Metal augment or structural allograft with cup (2 studies), cup-cage reconstruction (1 study), and custom-made triflange component (1 study).
Conclusions: Trabecular Metal augments and shells gave the most promising results in terms of the re-revision rate and radiographic loosening. Reconstruction with an antiprotrusio cage was the most frequently reported technique, with good results in a physically low-demand elderly population. Bone impaction grafting seems not appropriate for pelvic discontinuity and prone to failure in patients with Paprosky type-3B defects. In those cases, a custom-made triflange implant or a cup-cage reconstruction might be the best alternative, but few reports of sufficient quality are available yet.
Level of Evidence: Therapeutic Level IV. See Instructions for Authors for a complete description of levels of evidence.
Many different treatment options are available for acetabular revision, including (jumbo) noncemented hemispherical cups, structural allografts, bone impaction grafting, antiprotrusio cages, Trabecular Metal (Zimmer) augments and shells, cup-cage constructs, oblong cups, and custom-made triflange components1-6. Preoperative planning is essential to choose the appropriate implant, and therefore one needs to objectively define the nature of the defect to assess remaining acetabular bone stock and bone quality. Two widely used classification systems that provide detailed anatomical information for defect-specific preoperative planning are the American Academy of Orthopaedic Surgeons (AAOS) system7 and the system of Paprosky et al.8. In general, the larger the defect, the more challenging the acetabular revision. It is important to choose the appropriate strategy to treat these acetabular defects. Many studies evaluating different treatment options are available. However, most are small case series evaluating treatment methods that have been used for the treatment of various types of acetabular defects.
The objective of the present systematic review is to assess the effectiveness of revision options for the treatment of objectively classified large acetabular defects on the basis of re-revision rates, radiographic loosening, complications, and clinical outcomes.
Materials and Methods
A literature search was performed according to the Meta-analysis of Observational Studies in Epidemiology (MOOSE) guidelines9. Studies were identified in PubMed, MEDLINE, and Embase from 2000 to March 2014. The search strategy is shown in Table I. Two investigators independently screened the titles, abstracts, and full texts according to predetermined inclusion and exclusion criteria (Table II). Discrepancies were settled by consensus.
Studies in which the AAOS system7 or the Paprosky system8 was used to objectively define the acetabular defect were included. AAOS types III and IV and Paprosky types 3A and 3B were rated as large acetabular defects. These classification systems are widely used and accepted. AAOS type-III defects are characterized by a combination of segmental bone loss and cavitary deficiency. Type-IV defects are similar to pelvic discontinuity and are characterized by complete separation between the superior and inferior aspects of the acetabulum7. Paprosky type-3A acetabular defects are characterized by moderate-to-severe destruction of the acetabular walls and posterior column, rendering these structures nonsupportive, but the Kohler line remains intact, thus preventing substantial medial displacement of the component. If the acetabulum is considered as a circular structure represented by a clock face, then the bone loss involves the superior rim of the acetabulum from 10 o’clock to 2 o’clock. Paprosky type-3B defects are similar to type-3A defects, but the rim defect involves the region from 9 o’clock to 5 o’clock8. When both classification systems were used, we registered the Paprosky system as it is a more quantitative system.
Data extraction from the included full texts was performed by the primary investigator and was checked by the senior investigator. The data were collected on a prespecified data-extraction form and included authors, publication year, journal, study design, sample size, mean age, primary diagnosis, reason for revision, number of previous revisions, duration of follow-up, treatment method, co-interventions, classification, and the method that was used to determine the classification. Outcome measures were determined as the number of revisions for any reason, the number of implants with radiographic loosening based on the definition of radiographic loosening used in the article (including those that were revised because of loosening), the dislocation rate, complications, and clinical outcomes as determined with use of objective hip scores.
Two investigators independently evaluated the quality of the included full texts. Quality assessment was performed with use of 8 items of the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE)10 checklist for reports of observational studies that we deemed relevant for case series (Table III). Each item was scored as well described (+), partly described (+/−), or poorly/not described (−). If an item contained sub-items, the scores were added. The final score was rounded off downward (e.g., an item that consisted of 1 well-described [+] and 1 partly-described [+/−] sub-item was scored as partly described [+/−]).
In cases of disagreement, consensus was sought between the 2 investigators. Articles were included if ≥75% of items were well described (+). Two partly-described items (+/−) counted as 1 well-described item (+). Finally, if studies (partially) used the same patient data, the studies with newer or more extensive patient data were included.
The results of the included studies are presented according to the available treatment option and in order of frequency. Data were pooled according to treatment option: the total number of hips that were operatively treated as well as total number of hips per outcome (re-revision rate, radiographic loosening, and reoperation for any reason). Percentages were calculated per treatment option by dividing the outcome data by the total number of hips that were treated.
The detailed flow of the search and selection process is shown in Figure 1. A total of 38 articles reporting on 33 studies were eligible for the quality assessment (Table IV)11-48. The basic characteristics of the 20 included studies11-15,19,22,23,25,27,29,30,32,33,37,39,43,45,46,48 are summarized in Table V. All studies were case series with Level-IV evidence49,50. Seven studies were prospectively performed11,13,15,23,43,46,48. In most studies, the mean age of participating patients was between 60 and 70 years; in 1 study, a mean age of 82 years was reported23. The main indication for index revision surgery was aseptic loosening.
The outcome measures are summarized in Table VI. We found 7 different treatment options for large acetabular defects: antiprotrusio cage14,23,27,29,33,37,46,48, Trabecular Metal augment and shell19,22,32,45, bone impaction grafting with metal mesh15,24, hemispherical implant with hook and flanges12,30, Trabecular Metal augment13 or structural allograft with cup39, cup-cage reconstruction11, and a custom-made triflange component43. In all but 2 studies14,33, only large acetabular defects (Paprosky types 3A and 3B and AAOS types III and IV) were operatively treated. We included all data from these 2 studies as they mainly (>90%) involved revisions for large acetabular defects. In 1 study, pelvic discontinuity (which coincides with AAOS type IV) was used as a classification11. In most studies the classification of the defect was based on intraoperative findings11,13-15,23,24,27,29,32,33,36,45,46,48, in 3 studies it was based on preoperative radiographs22,39,43, and in 3 studies the classification system was mentioned but the method was not reported12,19,30. All studies included postoperative hip scores. In the studies in which preoperative hip scores were reported12,14,19,22-24,29,30,32,33,36,39,45,46, the scores improved postoperatively.
The antiprotrusio cage was the most widely used method in the included studies (8 studies, 315 hips). The Burch-Schneider cage was used in 4 studies23,27,37,48, the Kerboull reinforcement device was used in 3 studies29,33,46, and the Richards contour cage was used in 1 study14. Remaining defects were filled with cement in 1 study23, whereas morselized bone allograft was used in other studies. In all studies, cemented cups or liners were used in the antiprotrusio device.
The device was revised in 11 (3.5%) of the pooled 315 hips. One more revision probably should be added as Okano et al.33 excluded 1 hip because of infection and removal of the components 1 month postoperatively. Radiographic loosening was present in 22 hips (7.0%), and a total of 18 fractures of the device or screws were reported. Not all studies counted implant and/or screw breakage as radiographic loosening. The definition of radiographic loosening was well described in all but 2 studies. Bostrom et al.14 did describe the kind of radiographic loosening that was present (i.e., breakage or migration) but did not quantify the migration. Jones et al.27 did not quantify radiographic loosening but did report a mean vertical migration of 2.99 mm (range, 8.79 mm caudal to 4.05 mm cranial) and a mean horizontal migration of 3.43 mm (range, 7.98 mm medial to 4.19 mm lateral). As most studies quantified radiographic loosening as a migration of the implant of >3 to 5 mm, radiographic loosening was underreported in that study.
A total of 27 hips (8.6%) underwent reoperation for any reason, and 13 hips dislocated. Other complications included 9 infections, 9 hematomas, and 4 (partial) neurological deficits (2 neurapraxias of the sciatic nerve, 1 peroneal nerve palsy, and 1 transient sciatic nerve irritation).
A Trabecular Metal augment and shell was used in 4 studies (125 hips). Morselized bone allograft was used in 3 of these studies19,22,32, and the augment was cemented into the shell in 1 study45. Liners were cemented in some cases.
Two hips (1.6%) were revised, and 3 hips (2.4%) showed radiographic loosening. Flecher et al.22 did not provide a definition of radiographic loosening; however, they reported that no mechanical failure, screw breakage, loosening, or migration was noticed during the time of the study. In 1 study19, a patient with radiographic loosening was on the waiting list for revision with a custom-made implant; therefore, the number of hips undergoing revision is to be expected to increase to 3.
Dislocation occurred in 10 hips. Nineteen hips (15.2%) underwent reoperation for any reason, with 6 of them needing a liner revision. Lingaraj et al.32 implanted a liner with an elevated rim or a constrained liner in most patients. A total of 11 other complications were reported, including 5 infections and 3 nerve palsies. Weeden and Schmidt45 only reported the most common complication (dislocation), which may have led to underreporting of the complication rate.
The use of bone impaction grafting with mesh and a cemented cup was reported in 2 studies15,25 but was in fact the second most common technique (204 hips). No patients with pelvic discontinuity were managed with this technique. Garcia-Cimbrelo et al.25 found that hips with Paprosky type-3B defects had a higher risk of failure compared with those with Paprosky type-3A defects.
Fifteen hips (7.4%) underwent reoperation for any reason, and 14 hips (6.9%) had an acetabular revision. Eighteen hips (8.8%) had signs of radiographic loosening; 5 of these hips were not revised because they were only mildly symptomatic. Dislocation occurred in 5 hips. Two other complications, both deep infections, were reported.
An oblong implant with hooks and flanges was used in 2 studies (79 hips). Babis et al.12 only used this technique for Paprosky type-3A defects.
Nineteen hips (24%) were revised, and 4 patients were waiting for re-revision for radiographic loosening12. A total of 22 hips (27.8%) had radiographic loosening. Kim et al.30 reported radiographic loosening in all 3 hips with pelvic discontinuity. Dislocation occurred in 2 hips. Other complications included 1 broken hook and side plate, 4 infections, and 1 nerve palsy.
Two studies (47 hips) involved the use of either a structural allograft39 or a trabecular augment13 to provide stability for the acetabular component. The structural allografts were only used for Paprosky type-3A defects39.
Six hips (12.8%) were revised. Eleven hips (23.4%) had radiographic loosening. One hip dislocated. Only 1 other complication, a nerve palsy, was reported.
The study involving a Trabecular Metal augment13 demonstrated a lower revision rate but a similar rate of radiographic loosening and also had a shorter follow-up period (5 years compared with 10 years).
In 1 article (26 hips), a cup-cage construct was used for the treatment of pelvic discontinuity11. In that study, 2 hips (7.7%) needed revision and 5 (19.2%) had radiographic loosening. Other complications included 2 dislocations, 1 infection, and 1 nerve palsy.
Custom-Made Triflange Component43
Only 1 study (57 hips) evaluated the use of a custom-made component43.
Three hips (5.3%) were revised, 2 because of deep infection. Nine hips (15.8%) had aseptic loosening, but only 1 of them was revised.
Twelve hips dislocated. Twenty-four hips (42.1%) underwent reoperation for any reason; of those, 10 had a liner revision because of instability. Other complications included 2 infections and 2 nerve palsies.
The aim of this systematic review was to assess the effectiveness of available interventions for large acetabular defects. We found 7 treatment options for large acetabular defects as reported in 20 case series.
We are not the first to provide an overview of the different options for acetabular revision reconstruction1,2,5, and others have also limited the search to large acetabular defects3,4,6. Only 1 of those studies4 was a systematic review of the literature; that review, which included 50 studies, evaluated treatment options for large acetabular defects. However, the search in that study4, which was based on different treatment options, was different from our search, which was based on large acetabular defects. Those authors reported on studies investigating treatment options for large but also smaller defect sizes, which may have introduced a bias. Moreover, the use of bone impaction grafting for large defects was not reported.
In the present study, we only included articles that mainly (>90%) reported on revisions for large acetabular defects (Paprosky types 3A and 3B and AAOS types III and IV) in order not to have the outcome data biased by the results of treatment of smaller defects.
The best results for large acetabular defect reconstruction in terms of the rates of re-revision and radiographic loosening were reported for procedures involving a Trabecular Metal augment and shell as described in 4 studies19,22,32,45. However, a high dislocation rate resulted in many liner revisions. In our opinion, the use of a constrained liner might reduce the dislocation rate, although it might also increase the risk of aseptic loosening.
The antiprotrusio cage was the most widely reported technique and was described in 8 studies14,23,27,29,33,37,46,48. The re-revision rate was only slightly higher and the duration of follow-up in the included studies was longer in comparison with those in the studies on the Trabecular Metal augment and shell (Table V). The rate of radiographic loosening was relatively high. However, many hips did not need revision, probably because of satisfactory clinical results in a physically low-demand elderly population (mean age, >65 years). This finding suggests that this technique may be reliable for the treatment of large defects in elderly patients. In younger, more physically demanding patients, implant breakage resulting from a lack of stability and biological fixation may result in poor clinical outcomes.
In the present study, bone impaction grafting with mesh was found to be associated with acceptable results15,24. In cases of failure, the same technique might be used as bone impaction grafting results in at least partial restoration of bone stock. The effective restoration of bone stock is the main advantage of bone impaction grafting. However, this technique is not appropriate for pelvic discontinuity, and inferior results have been reported for Paprosky type-3B defects15,24.
The present study included only 1 study involving a cup-cage solution11 and 1 study involving a custom-made triflange component43. Both studies demonstrated acceptable revision rates, and both techniques were limited to defects with pelvic discontinuity. The higher rates of radiographic loosening associated with both techniques might be explained by the pelvic discontinuity.
Unsatisfactory results were reported when a hemispherical cup with hooks and flanges12,30 was used and when an allograft39 or a Trabecular Metal augment13 with a cup was used. Trabecular Metal augments may be favorable compared with structural allografts, which are only used for Paprosky type-3A defects and may fail because of resorption.
The present study had a few limitations. First, we limited our search to the 2 most commonly used defect-classification systems (Paprosky and AAOS) to provide uniformity, but studies investigating large acetabular defects classified with use of other qualification systems might have been missed as a result. Second, all of the included studies were case series. As far as we know, no instruments are available to sufficiently assess the risk of bias of case series. Therefore, we used the STROBE checklist10, which we adjusted for case series to assess the methodological quality. Third, there was no uniform definition of radiographic loosening in the different studies. Therefore, we reported the definition of radiographic loosening used in each study. Also, demographic factors such as comorbidities and body mass index that may correlate with patient outcomes were inconsistently and poorly reported in the articles. As a consequence, we were not able to comment on the possible influence of these factors on implant choice. Finally, we did not analyze the effect of reconstruction options on clinical outcome scores as different outcome scores were used and some reports did not provide preoperative scores11,13,15,27,43,48.
In conclusion, Trabecular Metal augments and shells to reconstruct large acetabular defects may be considered the technique with the most promising results, whereas the use of antiprotrusio cages is the most frequently reported technique. Antiprotrusio cages may be a valuable option for elderly, less physically demanding patients. Restoration of bone stock is the ultimate goal of bone impaction grafting, but this technique has inferior results for Paprosky type-3B defects, especially those associated with pelvic discontinuity. For large, type-3B defects, custom-made implants or cup-cage reconstructions might work, but few studies are available. In order to make recommendations with regard to the most effective intervention for large acetabular bone defects, prospective controlled studies would be most helpful.
Investigation performed at the Department of Orthopaedic Surgery, Sint Maartenskliniek, Nijmegen, the Netherlands
Disclosure: The authors indicated that no external funding was received for any aspect of this work. On the Disclosure of Potential Conflicts of Interest forms, which are provided with the online version of the article, one or more of the authors checked “yes” to indicate that the author had a relevant financial relationship in the biomedical arena outside the submitted work.
- Copyright © 2016 by The Journal of Bone and Joint Surgery, Incorporated