Abstract
Introduction
Early ceramic bearing systems in total hip arthoplasty (THA) sought to provide long term wear improvement over traditional metal on polyethylene systems. However, previous designs exhibited fractures of the ceramic acetabular liner, leading to the development of the Implex Hedrocel ceramic bearing THA system where the ceramic liner was supported on a layer of polyethylene intended to transition liner loads to the metal shell, a so-called “sandwich” design. Unfortunately, the device trial was stopped to further enrollment when liner fractures were reported. The current study examines nearly 10-year follow-up on 28 devices implanted by two surgeons at one institution in order to document ceramic bearing system performance over a longer time period.
Methods
Radiographic and patient reported outcomes, in the form of Harris Hip Scores (HHS) and 12-Item Short Form Health Survey (sF-12), were collected.
Results
During the study period two cups were replaced, one at three years and a second at seven years. At the five year follow-up HHS were similar to those reported in the literature for devices with traditional metal-on-polyethylene bearing surfaces and for other sandwich ceramic bearing designs. At the nine year follow-up, the HHS had not changed significantly and SF-12 scores measuring overall physical and mental health were higher than age matched national norms (p<0.001). There were no signs of cup migration, stem subsidence, osteolysis or cup loosening at any time up to the last follow-up in this patient cohort. The 89% survivorship rate and device revisions due to delamination of the liner observed in this group were similar to those reported earlier for this device and for other “sandwich design” ceramic bearing systems.
Discussion
This cohort did not exhibit new failure modes and HHS and SF-12 scores indicated high functionality for the majority of patients. These data suggest that a focus on preventing ceramic liner fracture through design and/or materials improvements may result in a device with long-term functionality.
Introduction
In the United States, the majority of total hip arthroplasty (THA) involves a metal stem, head and acetabular component with a polyethylene liner1. The primary drawback of this bearing surface is the osteolysis associated with wear debris generated by the polyethylene bearing surface2,3,4. The development of highly cross-linked polyethylene (HXPLE) has improved outcomes for total hip arthroplasty, but is still not seen as the ideal bearing surface in young active adults5,6. Low wear rates and longer lasting bearing surfaces are of specific interest in addressing the needs of young patients with a long life expectancy and a high activity level5,6,7. One alternative is the use of ceramic-on-ceramic bearing surfaces.
Ceramic-on-ceramic bearing surfaces in THA have been used in an effort to combat the osteolysis and aseptic loosening associated with the wear debris generated by polyethylene2,3,4,8,9,10. Excellent wear and biologically inert characteristics of ceramic appear to yield an ideal bearing surface, however, the risk of ceramic fracture presents a unique concern4,7,8,10,11. Since the initial use of ceramics in THA in the 1970s, improvements in its mechanical properties, grain structure, purity, and proof testing have significantly improved the quality of modern ceramics7,11,12,13. The improvements in physical characteristics decreased fracture rates4,7,11 but revealed aseptic loosening as another challenge. The mechanism for loosening was thought to be caused by a mismatch in the modulus of elasticity between the ceramic and bone4,8. “Sandwich” designs utilize a ceramic inlay with polyethylene interposed between the ceramic inlay and the metal shell which decreases the stiffness mismatch, thus potentially reducing the rate of fracture and improving bony in-growth4,7.
A number of “sandwich” acetabular component designs have been trialed with varying degrees of success. The first commercialized “sandwich” acetabular design was produced in 1993 and consisted of polyethylene sandwiched between an alumina ceramic inlay and a titanium shell3. Viste et al.3 studied the long term results of this cup and found a fracture rate of 3.3% (5 of 151) at an average of 9.9 years due to failure. Evidence of impingement of the femoral neck on the rim of the ceramic liner was also noted and thought to be involved in the failure3. Iwakiri et al.13 utilized a unique alumina ceramic insert on a polyethylene shell that was directly cemented to bone with no metal backing. At an average of 5.6 years they reported a 5.6% (4 of 72) alumina fracture rate. Notching was again noted on the femoral neck, which suggested that impingement likely played a role in the ceramic failure13. Park et al.7 reviewed 357 hips at an average of 3.9 years with a ceramic liner embedded in polyethylene which then had a Morris taper fit into a titanium shell. They reported an alumina fracture rate of 1.7% (6 of 357) (two alumina heads, four alumina liners). Based on microscopic and gross analysis of the fractured liner at revision, impingement of the femoral stem on the rim of the ceramic was seen as a primary cause of fracture7.
A “sandwich” design with the goal of decreasing the impact of femoral stem impingement on the rim of the acetabular liner was developed. The Hedrocel ceramic bearing cup (Implex, Allendale, New Jersey) consists of a Trabecular Metal tantalum shell, compression-molded polyethylene, and a press fit ceramic inlay (Image 1). The alumina ceramic inlay was recessed so any femoral stem impingement would result in polyethylene contact, as opposed to ceramic contact, thus decreasing the risk of delamination of the alumina component and subsequent dislocation and/or fracture. A multicenter prospective randomized clinical study under the supervision of the United States Food and Drug Administration (FDA) was initiated in 1999 to evaluate this device. In 2003 enrollment in the study was suspended due to failures of the ceramic liner8. Two measures of patient outcomes were used in the Hedrocel Ceramic Bearing Cup trial: the Harris Hip Score (HHS) and the 12-Item Short Form Health Survey (SF-12). The Harris Hip score is a well-known and often utilized measure of hip function that was initially described by William H. Harris in 196914. The SF-12 is a shortened version of the 36-Item Health Survey (SF-36) which generates a mental component score (MCS) and physical component score (PCS). These scores reflect overall health-related quality of life15,16. The SF-12 provides a generic measure of health that is not related to a specific disease or condition. The score is scaled from zero, being the lowest level of health, to 100, representing the highest level of health16. As patients age, the PCS component of the score tends to trend down while the MCS component of the score tends to trend up, thus any score must be compared to an age-matched control to be of any meaning. The SF-12 was utilized to provide a tool by which we can compare the patients with the ceramic total hip against the general population.
Image 1. Cross sectional diagram of Hedrocel Implex Acetabular component (a: porous shell, b: polyethylene, c: ceramic).
The current study examines 5 and 10 year functional outcomes and radiographic appearance of the Hedrocel ceramic bearing cup at our institution. Though this system is part of an investigational device trial and will never be commercialized, many patients received it as their THA and their long-term results are of interest.
Methods
The Implex Hedrocel system is an investigational device implanted in 315 patients as part of a FDA monitored, randomized multicenter prospective trial from 1999 to 2003. The current study examines data collected from one site where a total of 28 experimental devices were implanted in 25 patients by two surgeons (NW, SB) between March 2001 and December 2002 with an average of 9.2 year follow-up. The control group was completed but was not required to be followed when the long term follow up was initiated. Average follow-up after exclusion of patients who died or underwent revision procedure (six patients) is 9.8 years (6.1-10.5). The device studied is a “sandwich” monoblock acetabular component consisting of an alumina ceramic inlay (BIOLOX forte; CeramTec, Stuttgart, Germany) press fit into a direct-compression-molded ultra-high molecular weight polyethylene backed by a porous tantalum shell. The cup and femoral stem were press fit. The femoral stem was a porous-coated cobalt-chromium-alloy implant (Implex ProxiLock design). Patients with clinical indications for a total hip replacement, age between 18 and 75 years, and having a body mass index (BMI) less than 40 were included in the study. Patients with bilateral hips replacements received the same implant on both sides. The protocol was approved by the site's investigational review board. All study participants provided informed consent prior to participation.
The study group included 16 females and 9 males (Table 1). All prostheses were implanted using a lateral approach. The standard limited incision muscle-sparing anterolateral Watson-Jones approach for THA, as described by Kenneth Gustke M.D.17, was used with one exception: the leg, when dislocated, was placed over the anterior edge of the table, not the posterior edge. Briefly, the incision was made anterior and proximal to the greater trochanter and the joint exposed through a small opening created along the intermuscular interval between the gluteus medius and tensor fascia lata. Patients were seen pre-operatively, post-operatively at 2 weeks, 6 weeks, 3 months, 6 months, 1 year and annually thereafter. At each visit HHS and SF-12 data was collected. Radiographs obtained at 2 weeks, 3 months, 6 months, 1 year and at subsequent yearly visits were examined for evidence of osteolysis in the seven femoral Gruen Zones18 and three acetabular zones19. The cup inclination was also measured using a transischial line and the inferior and superior cup margins.
Table 1.
Demographics of patients enrolled
| Demographics | Men | Women |
|---|---|---|
| Number (patients/hips) | 6/9 | 16/16 |
| Average Age (yrs) | 63.2 | 64.6 |
| Average BMI | 30.1 | 29.4 |
Harris hip score and improvement in score (HHS at follow-up -preoperative score) for the study device were compared to those reported in the literature for similar designs using student T-test statistics. Rates of revision and osteolysis were compared using the chi-square test. The post-op SF-12 data was compared to pre-operative data using the T-test statistic and compared to national norms by age group using both the Wilcoxon Signed Rank test and a 95th percentile confidence interval test based on the standard error of measure16.
Results
The average age at the time of surgery was 63 years (standard deviation (SD) 8.9, range 42-75). The average BMI at the time of surgery was 29.68 (SD 4.5, range 2339). The average pre-operative HHS and the SF-12 PCS were indicative of reduced joint function and quality of life (Table 1). Two patients underwent revision (7.1%, 2 of 28) during the course of study: one at 3.2 yrs due to perceived imminent failure (female, 63 yrs, BMI 25 revised to metal on polyethylene) and the other at seven years due to liner displacement (female, 68 yrs, BMI 23, revised to metal on polyethylene). The short-term revision case was excluded from the five year follow-up group, as the revision occurred outside of the five year window (as determined using Chauvenet's criterion), while the longer term case was included in both the 5 and 10 year groups. Two patients died during the course of the study: 5.3 yrs and 7.9 yrs, of unrelated causes. Both patients were retained in the 5 year group and the latter was included in the 10 year group.
At the five year follow-up (4.9 yrs, SD 0.5, range 3.9-6.0) the HHS were significantly improved over the pre-operative scores at 88.7 (SD 13.6, range 50-100, p<0.001). The SF-12 PCS increased significantly from the pre-operative level to 46.7 (Table 2, p<0.001), while the MCS remained relatively high at 55.4. The average angle of inclination for the cup was 47.3 (SD 5.7, range 46.8-59). There were no clinically significant findings in either stem or cup radiographs.
Table 2.
Comparison of outcome measures (average, standard deviation and range) between this study and similar components described in the literature(1, 2)
| Pre-operative | 5 years | 9 years | |
|---|---|---|---|
| SF-12 Mental Component | 54.4 (SD 10.9, 31.5-66.5) | 55.4 (SD 6.6, 36.1-64.7) | 58.6 (SD 3.1, 47-64) |
| SF-12 Physical Component | 26.6 (SD 4.5, 18.9-37.5) | 46.7 (SD 10.1, 26.1-56.2) | 47.9 (SD 7.3, 30.6-55.5) |
| HHS: Current Study | 43.5 (SD 9.0; 24-61.5) | 88.7 (SD 13.6, 50-100). | 91.9 (SD 5.3; 74.9-100) |
| HHS: Ceramic Comparable I1 | 50.6 (SD 11.4; 26.95-78.03) | not reported | 96.86 (SD 8.0; 48.95-100) |
| HHS: Ceramic Comparable II1 | 47.89 (SD 10.7; 28.85-75.23) | not reported | 96.176 (SD 5.9; 74-100) |
| HHS: Ceramic Non-Sandwiched2 | 56 (SD not avail; 17-89) | not reported | 94 (SD not avail; 57-100) |
| HHS: Metal-on-Poly Comparable1 | 49.83 (SD 12.6; 21.35-87.25) | not reported | 96.49 (SD 5.6; 74-100) |
At the 10 year follow-up (9.15 yrs, SD 1.19, 6.04-10.48) the average HHS, PCS and MCS scores remained high. Harris Hip Scores and SF-12 PCS increased significantly from the pre-operative state to the longest post-operative follow-up (p<0.001), but the scores did not significantly differ from those at the five year time point. The largest increase in the Harris Hip Score was 69.0 while lowest increase was 25.3. The MCS and PCS scores for the study group were significantly higher than those of the national population16 (p<0.001 and p=0.001, respectively). The greatest differences were exhibited in the mental health component with an average difference of 8.5 points (2.25 points greater than the 95th percentile) and 88% of patients showing improved mental health. No cup migration or stem subsidence was observed at any time. There were no signs of osteolysis or cup loosening at last follow-up in February of 2013.
Discussion
The mid-term follow-up data, with the exception of the failed ceramic liner, was excellent for patients with the Hedrocel ceramic-on-ceramic bearing THA. These results are similar to other studies of ceramic bearing THA systems8,9,10,11,12. The functionality and pain relief as measured by the HHS were similar to that of other traditional and ceramic bearing surfaces and the SF-12 scores were above average for age-matched controls. The acetabular component showed no signs of migration in the time frame studied. In addition, no evidence of osteolysis was observed, similar to findings in other ceramic on ceramic designs11,9,12. Perhaps the most significant finding, however, was that no new failure mode was observed and the rate of device failure did not appear to be increasing.
Delamination of the ceramic liner from the polyethylene was the mode of failure identified in this study. This mode has been described in earlier reports for this design8 and in other similar designs3,13. While no patient experienced ceramic liner fracture during the data collection period of the current study, following the conclusion of the study period, an atraumatic ceramic fracture was identified in one patient at 10.8 yrs. Again the failure mode appeared to be partial delamination and fracture of the ceramic liner. Revision surgery revealed well-fixed acetabular and femoral components. The implant was revised to a ceramic head on HXPLE implant to decrease the effect of third-body wear.
Two failure mechanisms for ceramic bearing surfaces have been proposed in the literature: impingement and physical property differences between the ceramic and polyethylene. The Hedrocel device design sought to decrease the incidence of device failures by eliminating the direct contact of the femoral stem with the ceramic edge by recessing the ceramic inlay and creating a “bumper” of polyethylene around the rim (Image 1). Physical property differences between the hydrophilic and absorbent ceramic on the hydrophobic polyethylene in a moist environment has been raised as a contributing mechanism of failure2,20.
The revision rate for this study is similar to other studies of sandwich design acetabular components (Table 3)3,4,7,8,10,20. Though the revision rate is higher than some comparable studies, the study period is longer allowing time for more device failures. These findings may discourage further development of polyethylene sandwich ceramic-on-ceramic designs. Indeed, the 8% revision rate for the study group exceeded the <3% norm for traditional metal-on-polyethylene systems based on data from the National Joint Registry of England and Wales21. However, while the “sandwich” design appears sensitive to non-traumatic failures, those patients who did not experience the failure demonstrate the potential for long-term success. A recent report finds that a nonsandwich design ceramic cup provides long-term function without osteolysis or ceramic fracture 11. Sugano et al.11 conclude “cementless THA with the third-generation alumina COC hip bearings provided an excellent survivorship of 95.7% at 14 years and eliminated periprosthetic osteolysis for 11 to 14 years”. These data suggest that this bearing surface merits further consideration.
Table 3.
Ceramic liner fracture rates compared to published data of other ceramic-on-ceramic sandwich design devices
| Study | Overall Revision Rate | Fracture Rate | Avg. Time to Fx | Avg. Follow-up time |
|---|---|---|---|---|
| This study | 7% (2/28) | 0% (0/28) | 5.25 yrs (3.5-7) | 9.2 yrs (6.1-10.5) |
| Park et al. (7) | 1.7% (6/357) | 1.7% (6/357) | 3.1 yrs (1.3-4.8) | 3.9 yrs (3-6) |
| Poggie et al. (8) | 4.4% (14/315) | 3.8% (12/315) | 2.1 yrs (.67-3.5) | 2-5.5 yrs |
| Hasegawa et al. (4) | 9% (3/33) | 6% (2/33) | 1.7 yrs (0.2-3.4) | 5.8 yrs (5-6.5) |
| Lombardi et al. (10) | 1.5% (1/65) | 1.5% (1/65) | 6yrs | 6.1yrs (2.2-9) |
| Lopes et al. (19) | not reported | 2% (7/353) | 4.3yrs (1.3-7.6) | 3.4yrs (0.5-8.8) |
| Viste et al. (3) | 8.6% (13/151) | 3.3% (5/151) | 7 yrs (4.5-8.5) | 9.9 yrs (8.5-11.5) |
In the SF-12 data collected in the course of this study our patient group reported a more positive mental state as compared with age-matched US norms. This difference may be due to benefits realized by those who undergo hip replacement versus the general population, i. e., reduced pain and increased mobility compared to national norms. While the SF-12 has been utilized as a broad clinical measure of a patient's mental and physical state, it has not been studied specifically in a large group of persons who receive a hip replacement. Recent work using the SF-3622 found significant emotional score improvement in a hip and knee replacement group over a seven year post-operative period. However, as the data was not compared to a control, the finding may not be directly attributable to the surgical intervention. Future efforts would be required to investigate whether the positive mental state observed in this study persists in a larger study group and whether it might be attributed to factors related to the hip replacement.
This study was limited for several reasons. First, the small sample size limits the power of this study. The initial investigational study was performed at 22 different sites with an average of 26.8 patients enrolled per site. Logistical limitations precluded inclusion of patients from other sites and further patient enrollment at this site was limited by cancelation of the device. Second, all the procedures were performed by two surgeons. While only two surgeons participated in the study at this site this also decreases the confounders. All the patients had surgery performed at one institution with the same operative approach utilized over a short period of time thus the operating room ancillary staff and postoperative care was similar. Third, the results were not compared with age-matched controls. The initial study utilized controls at a ratio of 2:1, however the controls were not followed beyond the conclusion of the trial. Comparison to metal-on-polyethylene devices of similar design were obtained from the literature in lieu of a control population. These provided larger sample sizes and comparable time frames.
In conclusion, the mechanical properties of ceramic make it a bearing surface with low wear and consequently a low incidence of osteolysis, with the primary drawback being the propensity to fracture. The Trabecular Metal acetabular cup provided a well-fixed bone-implant interface and the ceramic-on-ceramic bearing surface functions well, however the challenge of fracture or dislocation of the ceramic inlay continues to pose a concern. Further research into ceramic and ceramic like components that would better resist fracture as a mode of failure should be pursued as the number of young patients receiving THA continues to grow.
Conflict of Interest Statement
The current study was performed independent of financial support. Authors Norman Walter and Stephen Burton received financial support from Zimmer for their role in the initial device trial. No study member has a significant financial interest in the trialed device or Zimmer.
References
- 1.Bozic KJ, Kurtz S, Lau E, Ong K. The Epidemiology of bearing surface usage in total hip arthroplasty in the United States. J Bone Joint Surg Am. 2009;91:1614–20. doi: 10.2106/JBJS.H.01220. [DOI] [PubMed] [Google Scholar]
- 2.Petsatodis G, Papadopoulos P, Papavasiliou K, Hatzokos I, Agathangelidis F, Christodoulou A. Primary cementless total hip arthroplasty with an alumina ceramic-on-ceramic bearing. J Bone Joint Surg Am. 2010;92:639–44. doi: 10.2106/JBJS.H.01829. [DOI] [PubMed] [Google Scholar]
- 3.Viste A, Chouteau J, Desmarchelier R, Fessy M. Fractures of a sandwich ceramic liner at ten year follow up. Int Orthop. 2012;36(5):955–60. doi: 10.1007/s00264-011-1375-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Hasegawa M, Sudo A, Uchida A. Alumina ceramic-on-ceramic total hip replacement with a layered acetabular component. J Bone Joint Surg Br. 2006;88(7):877–82. doi: 10.1302/0301-620X.88B7.17675. [DOI] [PubMed] [Google Scholar]
- 5.Streit M, Schroder K, Korber M, Merle C, Gotterbarm T, Ewerbeck V, Aldinger P. High survival in young patients using a second generation uncemented total hip replacement. Int Orthop. 2012;36(6):1129–36. doi: 10.1007/s00264-011-1399-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Bascarevic Z, Vukasinovic Z, Slavkovic N, Dulic B, Trajkovic G, Bascarevic V, Timotijevic S. Alumina-on-alumina ceramic versus metal-on-highly cross-linked polyethylene bearing in total hip arthroplasty: a comparative study. Int Orthop. 2010;34:1129–35. doi: 10.1007/s00264-009-0899-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Park Y, Hwang S, Choy W, Kim Y, Moon Y, Lim S. Ceramic failure after total hip arthroplasty with an alumina-on-alumina bearing. J Bone Joint Surg Am. 2006;88(4):780–7. doi: 10.2106/JBJS.E.00618. [DOI] [PubMed] [Google Scholar]
- 8.Poggie R, Turgeon T, Coutts R. Failure analysis of a ceramic bearing acetabular component. J Bone Joint Surg Am. 2007;89(2):367–75. doi: 10.2106/JBJS.F.00148. [DOI] [PubMed] [Google Scholar]
- 9.D'Antonio J, Capello W, Naughton M. Ceramic bearing for total hip arthroplasty have high survivorship at 10 Years. Clin Orthop Relat Res. 2012;470(2):373–81. doi: 10.1007/s11999-011-2076-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Lombardi A, Berend K, Seng B, Clarke I, Adams J. Delta ceramic-on-alumina ceramic articulation in primary THA. Clin Orthop Relate Res. 2010;468:367–374. doi: 10.1007/s11999-009-1143-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Sugano N, Takao M, Sakai T, Nishii T, Miki H, Ohzono K. Eleven-to 14-year follow-up results of cementless total hip arthroplasty using a third-generation alumina ceramic-on-ceramic bearing. J Arthroplasty. 2012;27(5):736–41. doi: 10.1016/j.arth.2011.08.017. [DOI] [PubMed] [Google Scholar]
- 12.Yeung E, Bott P, Chana R, Jackson M, Holloway I, Walter WL, Zicat B, Walter WK. Mid-term results of third-generation alumina-on-alumina ceramic bearing in cementless total hip arthroplasty. J Bone Joint Surg Am. 2012;18(2):138–44. doi: 10.2106/JBJS.J.00331. [DOI] [PubMed] [Google Scholar]
- 13.Iwakiri K, Iwaki H, Minoda Y, Ohashi H, Takaoka K. Alumina inlay failure in cemented polyethylene-backed total hip arthroplasty. Clin Orthop Relat Res. 2008;466:1186–92. doi: 10.1007/s11999-008-0168-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Harris William. Traumatic arthritis of the hip after dislocation and acetabular fractures: treatment by mold arthroplasty. J Bone Joint Surg Am. 1969;51(4):737–55. [PubMed] [Google Scholar]
- 15.Ware J, Kosinski M, Keller S. A 12-Item short-form health survery: construction of scales and preliminary tests of reliability and validity. Med Care. 1996;34:220–3. doi: 10.1097/00005650-199603000-00003. [DOI] [PubMed] [Google Scholar]
- 16.Office of Public Health Assessment, Utah. Office of Public Health Assessment. Salt Lake City, Utah: Center for Health Data, Utah Department of Health; 2001. Utah Health Status Survey Report Health Status in Utah: The Medical Outcomes Study SF-12, 2001. [Google Scholar]
- 17.Gustke K. O'Connor M, editor. Limited incision muscle-sparing anterolateral Watson-Jones approach for total hip arthroplasty. Minimally Invasive Hip Surgery. AAOS. 2007:27–36. [Google Scholar]
- 18.Gruen T, McNeice G, Amstutz H. “Modes of failure” of cemented stem-type femoral components: a radiographic analysis of loosening. Clin Orthop Relate Res. 1979;141:17–27. [PubMed] [Google Scholar]
- 19.DeLee J, Charnley J. Radiological demarcation of cemented sockets in total hip replacement. Clin Orthop Relat Res. 1976;121:20–32. [PubMed] [Google Scholar]
- 20.Lopes R, Philippeau J, Passuti N, Gouin F. High rate of ceramic sandwich liner fracture. Clin Orthop Relat Res. 2012;470(6):1705–10. doi: 10.1007/s11999-012-2279-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Smith A, Dieppe P, Vernon K, Porter M, Blom A. Failure rates of stemmed metal-on-metal hip replacements: analysis of data from National Joint Registry of England and Wales. Lancet. 2012;379(9822):1199–204. doi: 10.1016/S0140-6736(12)60353-5. [DOI] [PubMed] [Google Scholar]
- 22.Bruyère O, Ethgen O, Neuprez A, Zégels B, Gillet P, Huskin JP, Reginster JY. Health-related quality of life after total knee or hip replacement for osteoarthritis: a 7-year prospective study. Arch Orthop Trauma Surg. 2012 doi: 10.1007/s00402-012-1583-7. Jul 28. [DOI] [PubMed] [Google Scholar]