A cemented acetabular component with a reinforcement cross provides excellent medium-term fixation in total hip arthroplasty after pelvic irradiation

Materials and methods

The study was undertaken between 1 January 1995 and 31 December 2011. All consecutive operations were performed at two tertiary centres which specialise in joint reconstruction and in the treatment of bone tumours (Cochin Hospital Paris, France and Edouard Herriot Hospital Lyon, France). All patients were operated on by senior surgeons or by fellows with significant experience in joint reconstruction. Patients were included in the study if they had undergone pelvic irradiation with a minimum dose of 30 Gy and THA was undertaken using a cemented acetabular reconstruction and a Kerboull reinforcement cross (CMK, Lépine, Genay, France; CMK, Biomet; Valence, France); were aged > 18 years and had follow-up of >  24 months.

The indications for THA were: symptomatic radiation osteitis causing incapacity (n = 20; 41%), symptomatic avascular necrosis of the femoral head (n = 17; 35%) and pathological fracture of the femoral neck or acetabulum (n = 12; 24%) (Figs 1 and 2). Radiation osteitis was confirmed radiologically before the operation (Fig. 1) or at the time of the acetabular reaming by finding alternating zones of bleeding bone and dense sclerotic bone. A total of 45 patients (49 hips) met these inclusion criteria. There were 28 women (29 hips: 59%) and 17 men (20 hips: 41%) with a mean age of 69 years (19 to 85). Two hips (4%) were American Society of Anesthesiology (ASA) grade 1;¹⁵ 24 (53%) grade 2; 18 (40%) grade 3, and one (2%) grade 4. The median pre-operative Postel Merle d’Aubigné (PMA) score was 9 (Q1-Q3: 7 to 11).¹⁶ The median dose of radiation was 50 Gy (Q1 to Q3: 45 to 60;) with a median time from radiation to THA of 27 months (Q1 to Q3: 10 to 91;). Most patients underwent pelvic radiotherapy for prostate cancer; uterine cancer and metastasis from breast cancer to the femoral neck (n = 8; 18%) (Table II).

Figs. 1a - 1b

Anteroposterior radiographs of a 19-year-old patient showing a) a large area of radiation osteitis in the right iliac wing and acetabulum, and avascular osteonecrosis of the right femoral head 24 months after irradiation with 55 Gy for a sarcoma of the gluteal region, and the same patient b) 42 months after a cemented total hip arthroplasty with a reinforcement cross.

Figs. 2a - 2b

Anteroposterior radiographs of a 52-year-old patient a) 57 months after pelvic irradiation for cancer of the ureter showing osteonecrosis of the acetabulum with a protrusion of the right femoral head and the same patient b) 51 months after right cemented total hip arthroplasty (THA) with a reinforcement cross and 44 months after left THA for osteonecrosis of the femoral head.

The hip was approached through a partial anterior hemi-trochanteric osteotomy in 24 patients (49%), a trochanteric osteotomy in 11 (23%), an anterolateral approach in nine (18%) and a posterolateral approach in five patients (10%). The surgical approach to the hip was at the discretion of the surgeon. The Kerboull reinforcement cross (CMK, Lépine, Genay, France; CMK, Biomet; Valence, France) has a hook which is placed at the inferior margin of the acetabulum (Fig. 3). The superior arm is positioned against the roof of the acetabulum and fixed with between one and four 5 mm screws angled at 45° towards the sacro-iliac joint (Figs 1 and 2). Usually, because no graft was used, two screws were sufficient. The bearing surface used was metal on an all-polyethylene component (MKIII, Howmedica, Herouville Saint-Clair, France; SN, Smith & Nephew, Le Mans, France), among which there were ten (20%) dual mobility components (SATURNE, Amplitude, Valence, France).

Figs. 3a - 3d

Photographs showing the stainless steel Kerboull acetabular reinforcement cross MARC.K size 2 right side (for size 42 mm acetabular components; Lepine, Lyon, France). The posterior branch is more prominent than the right (b) which ensures 10° to 15° of anteversion when positioning the component (d).

There were eight (16%) uncemented femoral components (INTEGRALE, Amplitude, Valence, France), and 41 (84%) cemented components (5 MKIII Howmedica, Herouville Saint-Clair, France; 9 CMK, Smith & Nephew, Le Mans, France; 6 GENERIC, Amplitude, Valence, France; INITIALE, 5 Amplitude, Valence, France; 16 OCEANE, Tornier, Montbonnot Saint-Martin, France). All patients received six weeks of thrombo-embolic prophylaxis with low molecular weight heparin. Post-operative rehabilitation was the same as that for a standard THA. Patients were mobilised fully weight-bearing immediately after the operation unless a trochanteric osteotomy had been performed.

The primary outcome measure was the cumulative incidence of revision for mechanical failure of the acetabular component. The time to revision was the interval between the index operation and revision, if undertaken. Patients who died during follow-up or underwent revision for another reason were treated as competing events; the observations of patients for whom the event of interest, or a competing event, were not seen during follow-up were censored. Revision for mechanical reasons was defined as revision of the acetabular component for loosening, fracture, instability or wear.

Secondary outcomes included the cumulative incidence of revision of the acetabular component for any reason; the Postel Merle d’Aubigné (PMA) score;¹⁶ the radiological evaluation of loosening according to De Lee and Charnley¹⁷ and Grüen et al¹⁸ and patient survival. Patients alive at the time of the study were contacted by the authors by telephone and/or had a clinical review on a voluntary basis.

Results

The mean follow-up was 51 months (17 to 137). A total of 19 patients (42%) died during follow up. Patient survival at 24, 60 and 120 months was 71% (95% CI 0.57 to 0.86), 52% (95% CI 0.37 to 0.71), and 41% (95% CI% 0.26 to 0.65), respectively (Fig. 4). At final follow-up, 18 patients (40%) were alive and free of disease and eight (18%) were alive with disease.

Fig. 4

Kaplan–Meier survivorship curve with confidence intervals.

The median PMA score was 15 (Q1 to Q3 14 to 17) at final follow-up with a median score of 5 for hip mobility (Q1 to Q2 5 to 6), 6 for pain (Q1 to Q2 5 to 6) and 5 for walking (Q1 to Q2 4 to 6). Patients who did not undergo revision showed no sign of migration of the components or of progressive lucency. A stable lucency was found around eight acetabular components (16%) and around 16 femoral components (33%). A total of five hips (10%) showed heterotopic ossification: two (4%) grade 1, one (2%) grade 2 and two (4%) grade 3.

The cumulative probability of revision for mechanical reasons at 12, 24, 60 and 120 months was 0% (0% to 0%), 0% (0% to 0%), 2.9% (0.2% to 13.3%), and 2.9% (0.2% to 13.3%), respectively (Fig. 5). One patient underwent a -revision for acetabular loosening 34 months post-operatively (Fig. 6). This 56 year-old woman had a total radiation dose of 60 Gy for a uterine cancer 12 years before her hip surgery. At final follow-up (38 months) she was doing well with a PMA score of 14 (mobility = 6, pain = 4, walk = 4) and no radiological or clinical signs of loosening.

Fig. 5

Graph showing the cumulative probability of revision of a cemented THA with an acetabular reinforcement cross for mechanical reasons (plain line) and for any reason (dashed line) in patients who have undergone pelvic irradiation.

Fig. 6

Anteroposterior radiograph of a 56-year-old patient showing mechanical loosening of the left acetabular component 34 months after a primary cemented total hip arthroplasty with an acetabular reinforcement cross and significant radiation osteitis of the pelvis

The cumulative probability of revision of the acetabular component for any reason at 12, 24, 60 and 120 months was 2.2% (0.2% to 10.3%), 2.2% (0.2% to 10.3%), 8.1% (2% to 20%), and 20.2% (2.9% to 48.6%), respectively (Fig. 5). Three (6%) patients underwent revision for infection.

A 70-year-old man treated for cancer of the bladder by surgery and irradiation with 50 Gy developed an Escherichia coli (E. coli) periprosthetic joint infection 18 months post-operatively. He developed a radiation-induced rectovesical fistula after the THA and underwent a two-stage revision to allow the fistula to heal before the second stage. At final follow-up (9.4 years) he was doing well with a PMA score of 15 (mobility = 6, pain = 5, walk = 4).

A 71-year-old woman who had undergone surgery for a soft-tissue sarcoma of the buttock with resection of a considerable amount of muscle followed by irradiation with 30 Gy developed a periprosthetic infection with a methicillin sensitive Staphylococcus caprae nine months after THA. She underwent a one-stage revision and was doing poorly at final follow-up (9.5 years) with a PMA score at 6 (mobility = 3, pain = 2, walk = 1).

The last patient was an 81-year-old woman who underwent surgery and irradiation with 68 Gy for uterine cancer. She developed ureteric stenosis with frequent pyelonephritis and a periprosthetic infection with E. coli 35 months after THA. She underwent a one-stage revision after surgery for stenosis of the ureter and was doing well at final follow-up (40 months) with a PMA score of 14 (mobility = 5, pain = 6, walk = 3). There has been no recurrent infection in any of these three patients and none has required further revision.

Discussion

We recommend the use of a cemented reconstruction with the addition of an acetabular reinforcement cross to improve the medium-term results after THA in patients who have received radiation to the pelvis.

At ten years, we report a cumulative probability of revision for mechanical failure of 3%. Massin et al¹⁰ in a retrospective study which included 42 THAs, reported a rate of acetabular failure for mechanical reasons of 52% at six years using a standard cemented THA for an irradiated hip. Cho et al⁸ in a retrospective series of 18 patients, reported a failure rate of 50% at a mean follow-up of 58 months (Table I). We suspect that the reasons for these poor results include both inadequate cement interdigitation at the time of the procedure and a structure of bone which was, or eventually became, unable to withstand the stress around the implant.

Radiation osteitis is characterised by mixed areas of dense sclerotic bone together with zones of focal demineralisation.⁴ Previous in vitro studies have shown that the penetration of cement is compromised when the bone is less porous.^23,24 Therefore, the abnormal bone structure seen in patients after radiation is likely to reduce the interdigitation between cement and bone. Moreover, damage to the microvasculature of the bone and cells leave a bone matrix with a reduced capacity to remodel.²⁵ The resultant sclerotic structure is prone to microfracture.^6,26 Because the bone cannot heal, the forces around the implant are transmitted through an interface with an increasingly incompetent surface; consequently the stress increases, more microfractures are created, and the cycle continues until the implant fails. The effects of irradiation also increase with the passage of time.⁸ Therefore, bone that appears adequate at the time of THA involving bleeding bone without signs of radiation osteitis on imaging studies, may become incompetent after a few years.⁸

The use of the Kerboull reinforcement cross reduces the risk of implant failure because it decreases the stress applied to the peri-acetabular bone, allows fixation over a large surface area and gives elasticity to the reconstruction.^13,14 The inferior hook and the long screws passing through the plate allow transmission of forces to the supra- and infra-acetabular regions where resistance is better.¹⁴ Kawanabe et al¹³ have shown in a finite element analysis that the areas under high stress are the hook and the plate of the Kerboull-type device rather than the outer surface of the cemented acetabular component. With other reinforcement rings the stress is mainly transmitted around the screw holes and to the central part of the outer surface of the acetabular component.¹³ This may explain why we have seen better results in our series than in those which used other types of reinforcement ring.^8,10

Given the poor results of conventional cemented reconstruction, surgeons have looked at uncemented implants as an alternative. Conflicting survival results have been reported.^9,27 The proposed reasons for their poor results are the presence of a paucicellular and sclerotic matrix with decreased properties for osseous integration and the progression of cellular damage over time leading to a debonded interface.⁸ Recently, surgeons have been using trabecular metal implants which have shown better primary stability and bone ingrowth because of their high coefficient of friction and porosity.^11,12,28,29 The long-term survival of such uncemented reconstructions has yet to be seen given the potential for the progression of radiation damage over time. Joglekar et al¹² reported no direct proof of osseous integration five years post-operatively. We also believe that the benefits of uncemented components are limited in patients in whom the mortality is approximately 50% at five years.^8,10

We noted a 6% rate of prosthetic infection. This is unusually high and compares unfavourably with the expected 1% or less after conventional THA.³⁰ In other series it ranges between 0% and 12%^8,10 (Table I). These patients are often immunodeficient and present with modifications of their urinary tract microflora resulting in frequent infections which are difficult to treat.^31-33 Furthermore, the presence of radiation-induced entero-urinary fistulae, an increased rate of wound fistula and lymphatic stasis increase the risk of implant colonisation^10,34,35 These patients should be subject to extra care pre- and post-operatively with repeated mid-stream urine examinations and careful management with prophylactic antibiotics.³⁶ Bacteriological culture of joint fluid obtained at operation may be helpful because septic arthritis in irradiated hips is frequently misdiagnosed as radionecrosis.³⁷

This study has some limitations. First, the size of the series is limited to provide a precise estimate of the rate of revision.^38,39 However, it remains the second largest series published to date. A second limitation is the design of the study: it is a non-comparative study and we cannot rule out that the improved survival of this series over non-reinforced cemented acetabular reconstruction is due to chance alone. However, given the mechanical and biophysical backgrounds, previous reports in the literature and the low mechanical failure rate, we believe that our results are significant. A third limitation is that there is no way of determining the extent of radiation injury at the time of surgery. This may be one explanation for the differences in survival reported in the literature. In the present series, we included only patients who had received a minimum of dose of 30 Gy: radiation osteitis was confirmed either radiologically or at the time of reaming of the acetabular bone. Fourthly, all operations were carried out at two centres for hip reconstruction and bone tumour and it cannot be inferred that similar results would be obtained at other centres. However, the surgical technique is straightforward and can be performed as easily as a primary THA. Lastly, although two different acetabular components were used in the study, there was no difference in survival between the two (log rank test, p = 1). The use of a dual mobility cup is increasingly popular: we would recommend its more frequent use in this group of patients with a limited life expectancy.