The Results of Acetabular Impaction Grafting in 129 Primary Cemented Total Hip Arthroplasties

The Journal of Arthroplasty 28 (2013) 1394–1400

Contents lists available at ScienceDirect

The Journal of Arthroplasty journal homepage: www.arthroplastyjournal.org

The Results of Acetabular Impaction Grafting in 129 Primary Cemented Total Hip Arthroplasties Matthew J. Wilson, FRCS (Orth) a, Sarah L. Whitehouse, PhD b, Jonathan R. Howell, MSc, FRCS (Orth) a, Matthew J.W. Hubble, FRCS (Orth) a, A. John Timperley, FRCS, DPhil (Oxon) a, Graham A. Gie, FRCSEd (Orth) a a b

The Princess Elizabeth Orthopaedic Centre, Royal Devon and Exeter Hospital, Exeter, Devon UK Institute of Health and Biomedical Innovation, Queensland University of Technology, The Prince Charles Hospital, Chermside, Brisbane, Queensland, Australia

a r t i c l e

i n f o

Article history: Received 8 June 2012 Accepted 19 September 2012 Keywords: impaction grafting socket acetabulum survival cement

a b s t r a c t Between 1995 and 2003, 129 cemented primary THAs were performed using full acetabular impaction grafting to reconstruct acetabular deficiencies. These were classified as cavitary in 74 and segmental in 55 hips. Eighty-one patients were reviewed at mean 9.1 (6.2–14.3) years post-operatively. There were seven acetabular component revisions due to aseptic loosening, and a further 11 cases that had migrated N 5 mm or tilted N5° on radiological review — ten of which reported no symptoms. Kaplan–Meier analysis of revisions for aseptic loosening demonstrates 100% survival at nine years for cavitary defects compared to 82.6% for segmental defects. Our results suggest that the medium-term survival of this technique is excellent when used for purely cavitary defects but less predictable when used with large rim meshes in segmental defects. Crown Copyright © 2013 Published by Elsevier Inc. All rights reserved.

Acetabular impaction grafting for the reconstruction of acetabular defects in total hip arthroplasty has the potential to recreate anatomy whilst also allowing the restoration of bone stock. The incorporation of impacted, morcellised bone graft has been demonstrated in histological studies [1]. The use of bone graft in the reconstruction of acetabular defects has evolved over the last four decades. In 1984, Slooff from Nijmegen published a paper on the modern concept of acetabular impaction grafting (AIG) for protrusio acetabuli using bone chips and hemispherical impactors [2]. Subsequently there have been numerous papers reporting good outcomes using AIG, mainly in revision surgery [3–12]. There are only a few studies which report the results of impaction grafting when used exclusively in primary hip arthroplasty [13–15]. Of these, none clearly differentiate ‘full’ impaction grafting, where cement is in contact with graft over 100% of the interface, from ‘simple’ impaction of cysts and isolated, cavitary, medial defects. In addition no papers report the difference in results when this technique is used in the reconstruction of cavitary defects and segmental defects that have been reconstructed using a rim mesh. In this paper we have studied our results of ‘full’ AIG when used in primary total hip arthroplasty,

The Conflict of Interest statement associated with this article can be found at http:// dx.doi.org/10.1016/j.arth.2012.09.019. Reprint requests: Matthew J. Wilson, FRCS (Orth), The Princess Elizabeth Orthopaedic Centre, Royal Devon and Exeter Hospital, Barrack Road, Exeter, Devon EX2 5DW, UK.

with particular emphasis on the results of AIG in cavitary and segmental defects. Patients and Methods Between August 1995 and August 2003, all patients who underwent primary total hip arthroplasty using AIG were identified. All of the initial clinical and operative data were collected prospectively and no patient was lost to follow-up. Two hundred and three patients had undergone 217 primary total hip arthroplasties which had been coded as requiring the use of impaction grafting. Only cases requiring full, circumferential impaction grafting (defined as those cases in which the acetabular component is in contact with impacted graft over 100% of its interface) were included. This left a total of 129 THAs in 117 patients (89 female) with a mean age of 68.3 (22.9–100.1) years. There was no difference in the mean age at surgery between those with segmental and cavitary defects, 68.9 (SD 13.7) years compared to 66.2 (SD 13.6) respectively (P = 0.27 t-test). Classification of Acetabular Defects Acetabular defects were assessed using the American Academy of Orthopaedic Surgeons (AAOS) nomenclature, according to D'Antonio [16], of segmental, cavitary and combined deficiencies. Fifty hips were classified as Type 1, purely segmental defects, 74 hips as Type 2 cavitary defects and five as Type 3, combined segmental–cavitary defects. For the purpose of the study, Type 1 and Type 3 defects were

0883-5403/2808-0032$36.00/0 – see front matter. Crown Copyright © 2013 Published by Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.arth.2012.09.019

M.J. Wilson et al. / The Journal of Arthroplasty 28 (2013) 1394–1400

1395

Table 1 Primary Pre-Operative Diagnosis, Mean Age and Acetabular Defect. Type of Defect Age (Range)

Segmental (Type 1)

Cavitary (Type 2)

Combined (Type 3)

Total

69.0

–

1

–

1

74.8 (70.3–79.2)

2

–

–

2

55.2 (29.9–86.7) 69.4 (45.9–87.1)

16 14

1 5

1 3

18 22

85.5 (70.5–100.1) 81.7 72.8 (57.6–85.8)

– – 15

3 1 14

– – –

3 1 29

43.0 (31.2–54.7)

–

2

–

2

68.9 (22.9–99.2)

–

42

–

42

70.4 (57.3–85.8)

3 50

5 74

1 5

9 129

Diagnosis Ankylosing Spondylitis Drug induced arthropathy Dysplasia Idiopathic avascular necrosis Paget's disease Previous fracture Primary osteoarthritis Post-traumatic osteoarthritis Primary protrusio acetabuli Rheumatoid TOTAL

analysed together as 55 segmental defects that required reconstruction using a rim mesh. Of the 55 hips with segmental defects, 54 involved the supero-lateral acetabular wall and one case involved the anterior wall. The remaining 74 (57%) hips had pure cavitary defects with a fully supportive acetabular rim. Seven of the cavitary defects had a medial mesh inserted for reinforcement of a thin but intact medial wall. The primary diagnosis, mean age at surgery and classification of acetabular defects are presented in Table 1.

Operative Technique The majority of cases (96%) were performed by consultants or senior hip fellows. All patients received pre-operative systemic antibiotics and antibiotic loaded cement. The posterior approach was used in 126 hips and the direct lateral (transgluteal) in three. Defects were reconstructed using the techniques described by Schreurs et al [17]. The methods used in preparation of the bone graft varied; in 43 (33%) cases the bone chips for impaction were prepared by hand, using rongeurs; in 53 (41%) a commercially

Fig. 1. Pre-operative radiograph of cavitary acetabular defect.

Fig. 2. Immediate post-operative radiograph of cavitary defect demonstrating good socket positioning.

available bone mill was used and in 25 (19%) a combination of milled and hand-prepared bone chips was used. In eight cases the method of graft preparation was not recorded. Following full acetabular exposure, the defect was assessed for containment. In cavitary defects, the acetabulum was prepared using standard acetabular reamers and burrs to reveal bleeding subchondral bone. In very sclerotic surfaces, a 2.5 mm drill was used to perforate the bone, increasing vascularity at the grafting surface. Bone graft was impacted, in layers, using hemispherical acetabular impactors (Stryker Orthopedics, Mahwah, NJ) placed at the level of the transverse acetabular ligament. The size of final impactor was chosen to fill the mouth of the defect but allowing at least 5 mm thickness of graft in the socket. Multiple impactions were made ensuring that the final graft surfaces were tightly packed and had the feel of cortical bone. A cemented polyethylene, component was trialled and the graft washed through a slotted mesh to protect its structure and then dried using gauzes soaked in dilute hydrogen peroxide. At this point low viscosity cement was mixed and then pressurized into the socket prior to socket implantation.

Fig. 3. Thirteen year follow-up radiograph of impaction grafting of cavitary defect showing good incorporation of bone stock with no signs of loosening.

1396

M.J. Wilson et al. / The Journal of Arthroplasty 28 (2013) 1394–1400

Fig. 4. Pre-operative radiograph of segmental acetabular defect.

Fig. 6. Nine year follow-up radiograph of segmental defect showing good incorporation of bone stock and recreation of anatomy.

Segmental defects were first contained using a flexible stainless steel mesh (Stryker Orthopedics, Mahwah, NJ) that was cut to size and attached to the acetabular rim using a minimum of three 3.5 mm cortical screws. Once a stable containment had been achieved, the impaction proceeded as described above. During impaction, care was taken not to place too much graft medially, avoiding a lateralisation of the final socket. Once a tight impaction had been achieved, a cemented polyethylene, flanged component was trialled; the graft was washed through a slotted mesh to protect its structure and then dried using gauzes soaked in dilute hydrogen peroxide. At this point low viscosity cement was mixed and then pressurized into the socket prior to socket implantation. In 91 cases (71%) the use of the autologous femoral head provided sufficient bone graft. In eight cases (6%), the native femoral head was deemed to be of inadequate quality for use in impaction grafting and pure allograft was used. In 28 (22%) cases the native femoral head was deemed to be of inadequate quantity for reconstruction and allograft was used in combination with the autograft. In two cases the type of graft used was not recorded.

A cemented, all-polyethylene acetabular component was used in all cases; 66 Exeter concentric cups, 39 Contemporary cups (Stryker Orthopedics, Mahwah, NJ) and 24 Ogee cups (DePuy, Warsaw, IN) in combination with a cemented Exeter femoral component (Stryker Orthopedics, Mahwah, NJ), the majority (85%) of femoral components having heads of 26 mm diameter. Post-operatively, all patients were allowed to partially bear weight for a period of six weeks at which point check radiographs were taken before allowing the patient to progress to full weight-bearing. Examples of pre- and post-operative radiographs for both cavitary and segmental defects are shown in Figs. 1–6. Clinical Evaluation Clinical evaluation was by the grading system of Merle D'Aubigné and Postel as modified by Charnley [18] (surgeon-derived) and by both the Harris [19] (in part, patient-derived) and Oxford [20] (patient-derived) hip scores, using the transformed 0–48, worst to best score. Data were recorded pre-operatively and at regular intervals post-operatively. Where patients were unable to attend clinic for follow-up due to frailty or distance then Harris and Oxford Hip scores were obtained using postal questionnaires or telephone interview. Radiological Evaluation Plain radiographs taken pre-operatively and at every routine follow-up were digitised and scaled using digital templating software (Orthoview, Jacksonville, FL). In all cases an attempt was made to measure the thickness of the impacted graft layer at three points corresponding to the three DeLee and Charnley zones [21]. In cases of acetabular protrusio, migration of the acetabular wall medial to Kohler's line was measured on the pre-operative radiograph and graded according to Sotelo-Garza and Charnley [22] (Table 2). In Table 2 Grading of Protrusio Acetabuli.

Fig. 5. Immediate post-operative radiograph of segmental defect showing mesh containment and impaction grafting.

Grade 1 (0–5 mm) Grade 2 (6–14 mm) Grade 3 (N15 mm) TOTAL

15 27 7 49

Primary diagnosis: 42 primary protrusio, 3 RA, 2 Paget's, 1 dysplasia and 1 failed fracture.

M.J. Wilson et al. / The Journal of Arthroplasty 28 (2013) 1394–1400

analysis was performed using SPSS for Windows version 18.0 (SPSS Inc, Chicago, IL).

Table 3 Grading of Dysplasia. Crowe I II III IV Total

1397

Eftekhar 8 3 3 2 16

A B C D

3 10 4 1 18

Two cases could not be graded according to Crowe due to inadequate radiographs.

cases of dysplasia, the grading systems of Crowe [23] and Eftekhar [24] were used (Table 3). Radiological assessment of the acetabular component was performed as described by Sutherland [25] and migration of the acetabular component was measured, both horizontally and vertically using methods described by Nunn [26]. We recorded radiographic failure which has traditionally been defined as migration of the acetabular component by more than 5 mm in either a horizontal or vertical direction or a change in acetabular inclination of more than 5° [27,28]. The size of the rim mesh used was estimated by counting the small squares of the mesh in both directions as seen on a cross table lateral radiograph. The product of these values results in a relative area that can be used subsequently to compare meshes. By dividing the approximated mesh area by the number of screws used to secure the mesh, a ‘mesh area per screw’ figure was calculated. Endpoints for survival analysis were defined as revision for aseptic loosening, revision for any reason and as radiographic failure of the acetabular component. Clinical scores are presented as means and standard deviation (SD) due to sample size and score recommendations [20]. Kaplan–Meier survivorship analysis was carried out, with survival rates and 95% confidence intervals calculated and survival curves constructed. Survival rates are reported at nine years where there are 40 cases remaining at risk [29]. After confirmation of proportionality, Cox proportional hazards regression analysis [30] (forward conditional method) was used to determine if age at operation, gender, primary diagnosis, surgeon grade, graft type or type of defect (cavitary or segmental) significantly influenced revision for any reason. Statistical

Results Using a Kaplan–Meier analysis (Fig. 7), survivorship of all acetabular components at nine years with revision for aseptic loosening as the endpoint was 93.2% (95% CI 87.7 to 98.7). With revision of the acetabulum for any reason as the endpoint, survivorship is 88.1% (95% CI 81.0 to 95.2) and including cases of migration (designated as radiographic failures) that were not revised overall survivorship is 82.2% (95% CI 73.2 to 91.2%). It should be noted that ten of the eleven cases that had migrated were asymptomatic (vide infra). Cox regression analysis of various factors in surgical technique demonstrated that age at operation, gender, initial diagnosis, grade of operating surgeon and type of bone graft used (P = 0.13, P = 0.72, P = 0.89, P = 0.81, P = 0.78 respectively) did not influence revision. However the type of defect was a significant factor in the model (P = 0.032) with a hazard ratio of 4.3 (95% CI 1.1 to 16.2) for a segmental compared to cavitary defect. Separate study of the survivorship of cavitary and segmental defects (Fig. 8) demonstrates 100% survivorship of those hips with cavitary defects at nine years, with revision for aseptic loosening as the endpoint; although one hip was revised for aseptic loosening at 9.3 years post-operatively and one hip was symptomatic but not fit for revision. However the survivorship for hips with segmental defects reconstructed using a rim mesh was significantly worse at 82.6% (95% CI 69.3 to 95.9%) (Log rank test P = 0.008), although it should be noted that the number of patients remaining at risk at this stage was only 18. No patient was lost to follow-up and the fate of every implant is known. Eighty-one patients were alive and available for review at a mean of 9.1 (6.2–14.3) years post-operatively. The outcome of the Oxford Hip Score, Harris hip score and Charnley hip scores, improved significantly after surgery for both segmental and cavitary groups (P b 0.001; paired t-test). The Oxford hip score (0–48 worst to best scale) improved from 16.4 (SD 9.4; range 1–36) to 36.1 (SD 10.0; range 14–48) and the mean Harris Hip Score improved from 44.8 (SD 16.1; range 12–86) to 80.0 (SD 16.4; range 45–100). The pre-

Fig. 7. Kaplan–Meier survivorship curves of all hips.

1398

M.J. Wilson et al. / The Journal of Arthroplasty 28 (2013) 1394–1400

Fig. 8. Kaplan–Meier analysis of cavitary and segmental defects with revision as the endpoint.

operative and review scores using the Charnley modification of the Merle d'Aubigné and Postel system are shown in Table 4. The thickness of the graft layer in the three Charnley zones, as measured on the immediate post-operative radiograph of 127 patients is shown in Table 5. Four cases had lucent lines of less than 2 mm; three cases had lucent lines in zone 1 only and one case had a lucent line in zones 1 and 2. All of these lines were present on the initial post-operative radiograph and none progressed. There was no relationship between failure and graft thickness in the three Charnley zones (P = 0.086, P = 0.395 and P = 0.582 respectively). At latest follow-up, 12 of the 129 sockets (9%) had been revised; seven for aseptic loosening with symptomatic radiological failure, two for sepsis, two for recurrent dislocation and one well-fixed socket was revised, using a cement-in-cement technique, to a constrained liner, at the time of femoral revision for peri-prosthetic fracture. There are a further 11 cases who have failed radiographically according to the criteria mentioned above. The details of the 18 cases that were seen radiographically to have migrated N5 mm or tilted N5° are shown in Table 6. It is unclear what proportion of these sockets that undergo initial migration will remain stable over the longer term. Within the group designated “radiological failures” there are certainly some hips that have stabilised and should not be designated as clinical failures. Analysis of the 55 segmental cases, in which a reinforcement mesh was used, reveals that those cases which failed through aseptic loosening had significantly larger meshes with a mean, comparative area of 110.8 squares compared to 88.1 squares in those cases which did not fail (P = 0.02, Mann–Whitney U). Analysis of the ‘mesh area

Table 4 Pre-Operative and Latest Clinical Scores Using the Charnley Modification of the Merle d'Aubigné and Postel Scoring System. Charnley Category Pre-operative A B C Latest follow-up A B C

Number

Pain

Function

Movement

24 58 39

1.9 2.0 1.6

2.5 2.1 1.6

3.1 2.9 2.3

22 42 18

5.7 5.5 5.6

5.2 4.9 2.7

5.1 5.2 4.5

per screw’ demonstrated no significant difference between those hips that failed and those that survived (P = 0.968, Mann–Whitney U). Discussion The results of this study demonstrate the medium-term survival of acetabular impaction grafting (AIG) in the management of acetabular deficiencies. In cases involving cavitary defects, the intact acetabular rim permits firm graft impaction and excellent stability and this is reflected in the impressive results of AIG in cavitary defects in this study. Similar results are demonstrated in other reports of AIG in contained defects. Plominski and Kwiatkowski reviewed 42 cases of primary AIG for protrusio with 100% survival at seven years [15]. Rosenberg et al reviewed 36 primary cemented hip arthroplasties in 31 patients with rheumatoid arthritis and protrusio acetabuli and, even in this complex group of patients, good results were achieved, with 90% survival at a mean of 12 years following AIG [31]. The results in our study for segmental defects requiring reconstruction are less predictable, consistent with work by other authors. In a review of 71 cases, 70% of which had defects that were classified as AAOS Grade III or IV, Van Haaren et al reported an overall survival of 72% at a mean of 7.2 years following revision using AIG [32]. The authors concluded that insufficient support allowed micro-movement of the acetabular component and graft failure. Poor results in large, combined defects were also reported by Buttaro et al, with 90.8% survival at 36 months [6]. They also advised against the use of AIG in large defects. In our series all defects were reconstructed using a stainless steel rim mesh, which is flexible allowing it to be contoured to the acetabular rim. Our results demonstrate 81.6% survival for defects reconstructed using a rim mesh, with increasing mesh size shown to be an independent risk factor for failure. In order to improve mesh stability it has been our usual practice to place screws at approximately 1 cm intervals along the rim. One would expect that,

Table 5 Median Graft Layer Thickness (mm) in 3 Charnley Zones. Interquartile Range in Parentheses. Charnley Zone 1 2 3

Cavitary Defect (n = 73)

Segmental and Combined Defects (n = 54)

7.0 (5) 9.0 (5) 7.0 (4)

9.0 (4) 8.0 (5) 4.0 (5)

P Value (Mann–Whitney U Test) b0.001 0.031 b0.001

M.J. Wilson et al. / The Journal of Arthroplasty 28 (2013) 1394–1400 Table 6 Details of Those Cases Which Were Designated to Have Failed by Radiographic Criteria. Time Since Type of Defect Index (Segmental/ Surgery Cavitary/ Pre-Operative Age at (Years) Combined) Diagnosis Operation Sex 71.2

F

C

Protrusio

9.3

47.0

F

S

DDH

8.4

47.1 45.9 58.1 68.1

F M F M

S Cm S S

DDH IAN DDH OA

8.1 0.2 4.0 10.5

44.6 75.6

F F

C S

Protrusio DDH

9.3 2.5

74.7 72.4

F F

S S

OA IAN

5.6 7.6

64.5

F

S

IAN

7.9

46.9 81.4

F F

S S

DDH IAN

6.0 7.9

68.3 99.2

F F

S C

OA Protrusio

7.1 1.8

70.5

M

C

Paget's

13.3

48.5

F

S

DDH

11.6

61.1

M

S

OA

5.1

Outcome Asymptomatic and under review Asymptomatic and under review Revised Revised Revised Asymptomatic and under review Revised Asymptomatic prior to death Revised to cage Asymptomatic and under review Asymptomatic and under review Revised Asymptomatic and under review Revised Asymptomatic prior to death Symptomatic but not fit for revision surgery Asymptomatic and under review Asymptomatic prior to death

DDH = developmental dysplasia of the hip; IAN = idiopathic avascular necrosis; OA = osteoarthritis.

for a given size of mesh, those cases secured with fewer screws would be less stable and at a greater risk of failure than those meshes secured with a greater number of screws. Our study has not shown any such correlation suggesting that failure is not directly related to mesh fixation. Indeed, in many cases the mode of failure is seen to be cleavage within the graft and subsequent migration of the implant. One of the problems with screw placement along the rim of large defects is that the optimal position of the screw often passes across the defect to engage the medial cortex. This makes it difficult to perform solid graft impaction through a ‘portcullis’ of screws. For this reason, screw placement is often a compromise of placement for fixation and placement to allow good impaction. This may go some way to explaining why defects reconstructed with larger meshes have a higher failure rate. Since this study, we have modified our technique in a way that may address this problem. For the largest defects, we have started using porous foam metal wedges to fill segmental defects and provide constraint for the graft [33]. If a mesh is to be used we have modified the technique to gain initial fixation with a minimum of three screws: one at the apex, one anterior and one at the posterior border. The position of these screws does not interfere with graft placement and, for all but the largest of rim meshes, provides adequate stability for impaction. Following cup cementation, further screws can then be placed in the optimal area for fixation even if this means drilling through the graft bed. It has also been shown [34] that graft stability is improved with the use of larger chips in the region of 0.5–1 cm 3 and that commercially available bone mills produce chips that are too small, less than 3.5 mm on average [35]. In our study graft was prepared using a variety of methods. There was no correlation between the type of graft used and failure although numbers involved in each group were small. It is now our routine practice to prepare chips by hand, using

1399

rongeurs, to create bone chips of an adequate size. In addition, it is now a requirement of the Human Tissue Authority that fresh-frozen allograft chips are washed prior to impaction with the aim of reducing the risk of disease transmission. It has been shown that rinsing the graft improves graft stability, probably by allowing tighter impaction, with no adverse affects on graft incorporation [36,37]. One further change to our surgical technique has involved the use of larger polyethylene acetabular components that more readily fill the mouth of the defect. The mean outer diameter of components used in this study was quite small (46 mm) and the most commonly used size was 44 mm (range 40–58). Whether or not improved mesh stability, better graft impaction and larger acetabular components improve the outcome of AIG in segmental defects remains to be seen. Acknowledgments The authors wish to gratefully acknowledge Mrs R Sculpher, Mrs S Wraight and Mr L Collett, researchers of the Exeter Hip Unit, Princess Elizabeth Orthopaedic Centre, Exeter for their contribution to the ongoing data collection. References 1. van der Donk S, Buma P, Slooff TJ, et al. Incorporation of morselized bone grafts: a study of 24 acetabular biopsy specimens. Clin Orthop Relat Res 2002;396:131. 2. Slooff TJ, Huiskes R, van Horn J, et al. Bone grafting in total hip replacement for acetabular protrusion. Acta Orthop Scand 1984;55(6):593. 3. Comba F, Buttaro M, Pusso R, et al. Acetabular reconstruction with impacted bone allografts and cemented acetabular components: a 2- to 13-year follow-up study of 142 aseptic revisions. J Bone Joint Surg Br 2006;88(7):865. 4. Comba F, Buttaro M, Pusso R, et al. Acetabular revision surgery with impacted bone allografts and cemented cups in patients younger than 55 years. Int Orthop 2009;33(3):611. 5. Atroshi I, Ornstein E, Franzen H, et al. Quality of life after hip revision with impaction bone grafting on a par with that 4 years after primary cemented arthroplasty. Acta Orthop Scand 2004;75(6):677. 6. Buttaro MA, Comba F, Pusso R, et al. Acetabular revision with metal mesh, impaction bone grafting, and a cemented cup. Clin Orthop Relat Res 2008;466(10): 2482. 7. Issack PS, Beksac B, Helfet DL, et al. Reconstruction of the failed acetabular component using cemented shells and impaction grafting in revision hip arthroplasty. Am J Orthop (Belle Mead NJ) 2008;37(10):510. 8. Ochs BG, Schmid U, Rieth J, et al. Acetabular bone reconstruction in revision arthroplasty: a comparison of freeze-dried, irradiated and chemically-treated allograft vitalised with autologous marrow versus frozen non-irradiated allograft. J Bone Joint Surg Br 2008;90(9):1164. 9. Emms NW, Buckley SC, Stockley I, et al. Mid- to long-term results of irradiated allograft in acetabular reconstruction: a follow-up report. J Bone Joint Surg Br 2009;91(11):1419. 10. Iwase T, Masui T, Torii Y, et al. Impaction bone grafting for acetabular reconstruction: mean 5.5-year results in Japanese patients. Arch Orthop Trauma Surg 2010;130(4):433. 11. Mehendale S, Learmonth ID, Smith EJ, et al. Use of irradiated bone graft for impaction grafting in acetabular revision surgery: a review of fifty consecutive cases. Hip Int 2009;19(2):114. 12. Schreurs BW, Luttjeboer J, Thien TM, et al. Acetabular revision with impacted morselized cancellous bone graft and a cemented cup in patients with rheumatoid arthritis. A concise follow-up, at eight to nineteen years, of a previous report. J Bone Joint Surg Am 2009;91(3):646. 13. Figueras Coll G, Salazar Fernandez de Erenchu J, Roca Burniol J. Results of acetabular wiremesh and autograft in protrusio acetabuli. Hip Int 2008;18(1):23. 14. de Kam DC, Gardeniers JW, Hendriks JC, et al. Cemented polyethylene cups in patients younger than 40 years. Clin Orthop Relat Res 2009;467(7):1753. 15. Plominski J, Kwiatkowski K. Cemented primary total arthroplasty for acetabular protrusion in patients with rheumatoid arthritis. Ortop Traumatol Rehabil 2008;10(1):26. 16. D'Antonio JA, Capello WN, Borden LS, et al. Classification and management of acetabular abnormalities in total hip arthroplasty. Clin Orthop Relat Res 1989;243: 126. 17. Schreurs BW, Gardeniers JW, Slooff TJ. Acetabular reconstruction with bone impaction grafting: 20 years of experience. Instr Course Lect 2001;50:221. 18. Charnley J. Numerical grading of clinical results. In: Charnley J, editor. Low friction arthroplasty of the hip — theory and practice. New York: Springer-Verlag; 1979. p. 20. 19. Harris WH. Traumatic arthritis of the hip after dislocation and acetabular fractures: treatment by mold arthroplasty. An end-result study using a new method of result evaluation. J Bone Joint Surg Am 1969;51(4):737. 20. Murray DW, Fitzpatrick R, Rogers K, et al. The use of the Oxford hip and knee scores. J Bone Joint Surg Br 2007;89(8):1010.

1400

M.J. Wilson et al. / The Journal of Arthroplasty 28 (2013) 1394–1400

21. DeLee JG, Charnley J. Radiological demarcation of cemented sockets in total hip replacement. Clin Orthop Relat Res 1976;121:20. 22. Sotelo-Garza A, Charnley J. The results of Charnley arthroplasty of hip performed for protrusio acetabuli. Clin Orthop Relat Res 1978;132:12. 23. Crowe JF, Mani VJ, Ranawat CS. Total hip replacement in congenital dislocation and dysplasia of the hip. J Bone Joint Surg Am 1979;61(1):15. 24. Eftekhar NS. Principles of total hip arthroplasty. C. V. Mosby; 1978. 25. Sutherland CJ, Wilde AH, Borden LS, et al. A ten-year follow-up of one hundred consecutive Muller curved-stem total hip-replacement arthroplasties. J Bone Joint Surg Am 1982;64(7):970. 26. Nunn D, Freeman MA, Hill PF, et al. The measurement of migration of the acetabular component of hip prostheses. J Bone Joint Surg Br 1989;71(4):629. 27. Schreurs BW, Bolder SB, Gardeniers JW, et al. Acetabular revision with impacted morsellised cancellous bone grafting and a cemented cup. A 15- to 20-year followup. J Bone Joint Surg Br 2004;86(4):492. 28. Garcia-Cimbrelo E, Cruz-Pardos A, Garcia-Rey E, et al. The survival and fate of acetabular reconstruction with impaction grafting for large defects. Clin Orthop Relat Res 2010;468(12):3304. 29. Lettin AW, Ware HS, Morris RW. Survivorship analysis and confidence intervals. An assessment with reference to the Stanmore total knee replacement. J Bone Joint Surg Br 1991;73(5):729.

30. Cox DR. Regression models and life-tables. J R Stat Soc Ser B Stat Methodol 1972;34(2):187. 31. Rosenberg WW, Schreurs BW, de Waal Malefijt MC, et al. Impacted morsellized bone grafting and cemented primary total hip arthroplasty for acetabular protrusion in patients with rheumatoid arthritis: an 8- to 18-year follow-up study of 36 hips. Acta Orthop Scand 2000;71(2):143. 32. van Haaren EH, Heyligers IC, Alexander FG, et al. High rate of failure of impaction grafting in large acetabular defects. J Bone Joint Surg Br 2007;89(3):296. 33. Ling RSM, Lee AJC, Gie GA, et al. The Exeter hip — 40 years of innovation in total hip arthroplasty. Exeter: Exeter Hip Publishing; 2010. p. 369. 34. Arts JJ, Verdonschot N, Buma P, et al. Larger bone graft size and washing of bone grafts prior to impaction enhances the initial stability of cemented cups: experiments using a synthetic acetabular model. Acta Orthop 2006;77(2):227. 35. Bolder SB, Schreurs BW, Verdonschot N, et al. Particle size of bone graft and method of impaction affect initial stability of cemented cups: human cadaveric and synthetic pelvic specimen studies. Acta Orthop Scand 2003;74(6):652. 36. Dunlop DG, Brewster NT, Madabhushi SP, et al. Techniques to improve the shear strength of impacted bone graft: the effect of particle size and washing of the graft. J Bone Joint Surg Am 2003;85-A(4):639. 37. van der Donk S, Weernink T, Buma P, et al. Rinsing morselized allografts improves bone and tissue ingrowth. Clin Orthop Relat Res 2003;408:302.