Eur J Orthop Surg Traumatol DOI 10.1007/s00590-014-1521-1
ORIGINAL ARTICLE
Cemented Thompson’s hemiarthroplasty in patients with intracapsular neck of femur fractures: survival analysis of 1,670 procedures S. K. Khan • S. S. Jameson • A. Sims • J. A’Court • M. R. Reed • A. Rangan • S. D. Muller
Received: 18 February 2014 / Accepted: 5 August 2014 Ó Springer-Verlag France 2014
Abstract Cemented Thompson’s prostheses have been used to treat elderly patients with displaced intracapsular hip fractures at our two units for the last 15 years, amid growing support for the use of newer implant designs for hip hemiarthroplasty. This provided us with an opportunity to investigate survival of the Thompson’s stem in our patients. A retrospective cohort study was set up to review previously collected data on patients who underwent Thompson’s hemiarthroplasty over a 7-year period. These were linked to surgical notes, clinical letters and radiographs to record post-operative course and subsequent admissions and procedures. The identifiers were then linked to mortality data from the Office of ‘National Statistics. Kaplan–Meier survival analyses were done for implants and patients. A total of 1,632 patients (mean age 82.7 years) underwent 1,670 procedures. Five-year implant survival was 95.4 %. A total of 36 stems were revised, including 14 revisions to total hip arthroplasty and 22 excision arthroplasties. Reasons for revision included infection (2.1 %), dislocation (1.1 %) and aseptic loosening (0.5 %). Symptomatic aseptic loosening and acetabular erosion occurred late (mean time 3.2 and 5.7 years,
S. K. Khan S. S. Jameson J. A’Court M. R. Reed S. D. Muller Northumbria Healthcare NHS Foundation Trust, Ashington NE63 9JJ, UK S. K. Khan (&) Apartment 509, The Bar, St James Gate, Newcastle upon Tyne NE1 4BA, UK e-mail: [email protected]; [email protected] A. Sims A. Rangan South Tees Hospitals NHS Trust, Marton Road, Middlesbrough TS4 3BW, UK
respectively following surgery). Aseptic loosening and erosion following hemiarthroplasty are relatively late complications. Keywords Hemiarthroplasty Thompson’s prosthesis Implant survival Patient survival Revision Dislocation Acetabular erosion
Introduction The incidence of intracapsular neck of femur (NOF) fracture is increasing in the United Kingdom (UK) [1]. Hip arthroplasty (hemiarthroplasty or total hip arthroplasty) is the standard treatment for the majority of these displaced fractures. The Thompson’s prosthesis is a press fit hemiarthroplasty implant designed in the pre-cement era [2, 3]. It is now most frequently used with cement [4, 5], with improved clinical effectiveness compared to uncemented monoblock (e.g. Austin-Moore) stems [6, 7]. The Orthopaedic Device Evaluation Panel (ODEP) produces ratings for hip arthroplasty prostheses used in the UK, based on length of follow-up and quality of evidence [8]. The Thompson’s stem currently does not have an ODEP rating, presumably because of the relative paucity of implant survival data in this at-risk patient group. The National Institute for Health and Clinical Excellence (NICE) offers evidence-based guidance to clinicians in the UK. The recent NICE guidance on NOF fracture management recommends the use of ‘proven’ cemented stem designs with an ODEP rating of at least 3B (97 % survival at 3 years), instead of the Thompson’s prosthesis [9]. Given that none of the available newer designs currently carry an ODEP rating, recommendations against the use of the Thompson’s prosthesis do not seem clearly justified.
123
Eur J Orthop Surg Traumatol
There are conflicting global trends for the use of Thompson’s prosthesis in favour of newer stems. A recent review of the Swedish Hip Arthroplasty Register has reported a decline in the proportion of Thompson’s stems from 9 to 1 % from 2005 to 2009 with an increasing number of unipolar modular prostheses being implanted [10]. The 2011 annual report from Australian National Joint Registry found that the use of the Thompson’s prosthesis had increased in 2010 (30.6 % of all monoblocks, up from 20.9 % in 2003); however, the number of monoblock stems was declining compared to modular and bipolar implants [11]. Cemented Thompson’s prostheses have been used to treat all elderly patients with displaced intracapsular NOF fracture at our two units over the last decade. This provided us with a unique opportunity to investigate outcomes in these patients. Three aims were defined as follows: (1) to examine the survival of the Thompson’s implant, (2) to investigate the indications for revision and (3) to compare patient survival to implant survival.
Methods A two-centre retrospective cohort study was designed to analyse implant and patient survival in patients who underwent cemented Thompson’s hemiarthroplasty for intracapsular NOF fracture between 1 January 2003 and 31 Fig. 1 Flowchart showing constitution of the final data set
123
December 2010. Unit 1 has one of the largest geographical catchment areas in the UK, and provides trauma services at three acute hospitals, while Unit 2 is a regional major trauma centre. Ethical approval was obtained for this study. Patients were identified from the hospitals’ central registers. Previously collected data on patients’ age, date of surgery and length of hospital stay were requested. Prophylactic antibiotic use, implant type, laterality of procedure and use of cement were confirmed from surgical and clinical notes by linking the patient identifiers to the online databases SIRIS (Surgical Information Recording & Interrogation System; Xentec Ltd, Wallsend, UK), CaMIS (Clinical and Management Information System; Ascribe Ltd, Bolton, UK), TheatreMan (Trisoft Limited, Chesterfield, UK) and the picture archiving and communication system (PACS) at the two Trusts (IMPAX, Agfa HealthCare, Brentford, UK). Exclusion criteria from the final subset included patients with pathological fractures due to neoplastic disease, those having the Thompson’s hemiarthroplasty as a revision procedure and those having uncemented procedures (Fig. 1). These linkages were next used to identify subsequent orthopaedic admissions from the Emergency departments or from clinic appointments requested by the patients’ general practitioners. A minimum data set was defined for each admission, including diagnosis (e.g. infection, dislocation, aseptic loosening, acetabular erosion, periprosthetic fractures, contralateral NOF fractures and persistent pain),
Eur J Orthop Surg Traumatol
type of procedures performed (e.g. washout and debridement, closed or open reduction, revision to total hip arthroplasty (THA), excision arthroplasty and fracture fixation) and any further complications. A revision was defined as any procedure in which the original cemented Thompson’s stem was removed, i.e. either excision arthroplasty or revision to THA. Survival (Kaplan–Meier) analyses were performed for implant and patient survival following Thompson hemiarthroplasty. Life tables were produced to report yearly implant survival rates for all procedures included in the study, with numbers entering each year reported [12]. Survival was not reported once the number entering a year fell below 5 % of the original number. In order to generate patient survival data, patient identifiers were linked to the Public Health and Mortality File generated by the Office for National Statistics (ONS; Cardiff, UK) for dates and places of death. All analyses were performed using SPSS version 19 (SPSS Inc, IBM Corporation, Armonk, NY). Fig. 2 Kaplan–Meier curve for implant survivorship
Results One thousand six hundred and thirty-two patients underwent 1,670 cemented Thompson’s hemiarthroplasties during the study period. These included 1,594 patients with unilateral hemiarthroplasties, 37 patients with bilateral non-contemporary procedures and one patient with bilateral simultaneous procedures. The mean age at presentation was 82.7 years (range 52–102 years, SD 7.3 years). There was a notable increase in the mean age over the 7-year study period, rising from 77.6 years in 2004 to 83.4 years in 2010. Implant complications and survival Seventy-six complications (4.5 % complication rate) were identified (Table 1). Infection was the commonest problem, occurring after 36 procedures (2.1 % infection rate). Table 1 Complications identified in current series of 1,670 prostheses Complication
% of all procedures (number)
% of all revisions
1
Infection
2.1 (n = 36)
44.4
2 3
Dislocation Aseptic loosening
1.1 (n = 18) 0.5 (n = 8)
27.7 19.4
4
Periprosthetic fracture
0.4 (n = 6)
5.6
5
Persistent pain
0.3 (n = 5)
2.8
6
Protrusion
0.1 (n = 2)
0
7
Bleeding
0.1 (n = 1)
0
Twenty of these stems were retained; 19 had a surgical debridement and washout in theatre, while one infection settled on antibiotics alone. Sixteen stems were revised, including 12 excision arthroplasties and four stems undergoing two-stage revisions to THA. The mean time to revision for infection was 254 days. Aseptic dislocation occurred in 18 out of 1,670 implants (1.1 % dislocation rate). Six stems were reduced under general anaesthesia and thus retained. The remaining 12 stems were either irreducible or re-dislocated again, necessitating revision. The mean time to revision for dislocation was 31 days. Except for one outlier, all were performed within 60 days of the index procedure. Aseptic loosening was identified in 8 implants in this series (0.5 % aseptic loosening rate). Seven patients returned to theatre for revision, including 6 THAs and one excision arthroplasty. The mean time to revision for aseptic loosening was 3.2 years. Two patients were symptomatic with acetabular protrusion (confirmed on radiographs) when reviewed in clinic. This review occurred at a mean of 5.6 years after the index procedure. Both the patients were offered surgery (intention to revise), but declined. Of all the identified complications, 40 implants were retained and 36 were revised (2.1 % all-cause revision rate). The latter included 22 excision arthroplasties (61 % of revisions) and 14 revisions to THA (39 %). The mean time to all-cause revision was 1.2 years (range 3 days–5.8 years). Implant survival over the study is shown in Fig. 2. Five-year implant survival was 95 % (Table 2).
123
Eur J Orthop Surg Traumatol Table 2 Implant survival for Thompson’s population, by year since index procedure
Interval start time (start year)
Number at start
Failure
Withdrawn
Number at-risk
0
1,670
25
494
1,422
2
98
98
1
1,150
2
401
950
0
100
98
2
747
3
281
607
0
100
98
3
463
2
210
358
1
99
97
4
251
3
134
184
2
98
95
5
114
1
73
78
1
99
94
6 7
40 21
0 0
19 11
31 16
0 0
100 100
94 94
8
10
0
10
5
0
100
94
Fig. 3 Kaplan–Meier curve for patient survivorship
Patient survival Thirty-day and 1-year survival was 91.6 and 69.6 %, respectively. This declined to 35 % at 5 years and just over 20 % at 8 years. Patient survival is demonstrated in Fig. 3.
Discussion This two-centre retrospective study reports patient and implant survival after cemented Thompson’s hemiarthroplasty in a large consecutive series of fragility NOF fractures. There were four key findings: (1) implant survival was 95 % at 5 years, (2) patient survival was 35 % at 5 years, (3) the mean age at presentation with hip fractures
123
Annual failure rate (%)
Annual success rate (%)
Survival rate (%)
is increasing and (4) acetabular erosion and aseptic loosening are relatively late-occurring complications. Our patient mortality and implant revision rates were in agreement with those reported in literature to date [13–15]. This study thus supports the use of Thompson’s prosthesis in the majority of patients with displaced NOF fractures. The dislocation rate (1.1 %) is surprising given that the design of the Thompson’s would intuitively lend itself to a lower risk of dislocation, i.e. a large head that confers a favourable head/neck ratio and increases the excursion distance. Our outcome is however comparable to contemporarily published series [15, 16] and is substantially less than the overall 7 % dislocation rate reported in a previous review of 338 cemented Thompson’s implants [17]. Scandinavian studies have found the posterior approach to be significantly associated with dislocation [17, 18]. All procedures in our series were performed using the anterolateral approach. This is the approach most commonly used in hemiarthroplasty for NOF fracture in the UK and is currently gaining favour in Sweden [19]. Our data also confirms aseptic loosening to be a late complication of hemiarthroplasty. Given that patient survival is poor, it is not surprising that aseptic loosening was rare in this series. Acetabular erosion [20] was similarly found to be an extremely uncommon (0.2 %) late complication. The Thompson prosthesis is now over 60 years old and was designed before the use of cement. Advocates of newer implant designs, e.g. Exeter Trauma Stem (ETS, StrykerTM UK Ltd., Newbury, UK) draw attention to potential advantages of these designs, including lesser soft tissue exposure, optimal restoration of femoral neck offset, anteversion and leg length, ease of implantation and cement-in-cement revision if needed [21]. The ETS is also a monoblock with a surface finish (roughness average) about ten times more than the Exeter arthroplasty stem [22]. This may result in aseptic loosening, like the mattsurface finish did to the Exeter stem in 1970s [23]; it may also influence loosening rates. The perceived advantages of
Eur J Orthop Surg Traumatol
the ETS over the Thompson’s have however not translated into significant differences in short-and medium-term outcomes in the only published randomised controlled trial comparing the two implants [21]. The Thompson’s is cheaper too, costing £130 (€165) less than the ETS [21]. In a historical cohort comparison between the ETS (n = 388) and Thompsons (n = 766), Bidwai and Willett found no difference in surgical or radiological outcomes [24]. The recent Australian joint registry report also showed that short-term revision rates for the cemented Thompson’s and the ETS were comparable (1.04 vs. 1.10 revisions/100 observed component years), despite potential selection bias with slightly younger and fitter patients receiving the ETS [11]. One of the strengths of this study is the large captive population, which historically has remained static, i.e. has continued to reside within the same catchment area postdischarge. We believe capturing all deaths including out of hospital deaths by way of the ONS data is a major strength of this study, making the patient survival analysis reliable. Moreover, 2.2 % patients in our series returned to have a contralateral Thompson’s hemiarthroplasty at a median of 0.8 years (range 14 days–7.2 years); this contralateral fracture rate is comparable to that reported in nationwide population-based Scandinavian studies [25]. Finally, all recorded readmissions with complications and subsequent revisions at our units have been reported, making this the largest reported series on implant survival for Thompson’s stems. We acknowledge potential limitations of this study. Most Hospitals in the English NHS do not arrange routine follow-up for patients undergoing hemiarthroplasty. Survival analysis has certain shortcomings. Our outcome measure for implant ‘survival’ was a revision procedure. However, patients living with a poorly performing or painful prostheses, but who are unable to have (or do not want) further surgery, cannot be deemed a success. We do not have data on surviving patients with problems who were never referred back. Similarly, we do not have data on any patients who may have had revisions outside the region due to relocation or tertiary referral pathways. We believe this is unlikely, given the characteristics of our combined catchment area. Lastly, survival curves in the presence of competing risks (in this case, the large competing risk of death) may overestimate the risk of revision [26]; therefore, the true revision risk for these patients may actually be smaller than we report here. In conclusion, this study found an increasingly ageing fracture population with 35 % patient survival, but 95 % implant survival, at 5 years. Our data show latency in developing aseptic loosening and acetabular erosion, which are the usually cited reasons for revision. The implant clearly satisfies the minimum ODEP 3B rating now
mandated by NICE [27]. In fact in this study, it displayed a profile consistent with an ODEP 5B rating, however this is a matter for ODEP to decide. Abandoning the Thompson’s hemiarthroplasty in favour of newer designs is not sufficiently substantiated by currently available evidence. The clinical and cost effectiveness of the newer hemiarthroplasty stems needs to be ascertained through well-designed comparative studies. Conflict of interest of interest.
The authors declare that they have no conflict
References 1. White SM, Griffiths R (2011) Projected incidence of proximal femoral fracture in England: a report from the NHS Hip Fracture Anaesthesia Network (HIPFAN). Injury 42:1230–1233 2. Thompson FR (1954) Two and a half years experience with a vitallium intramedullary hip prosthesis. J Bone Joint Surg 36A:489–502 3. Hernigou P, Quiennec S, Guissou I (2014) Hip hemiarthroplasty: from Venable and Bohlman to Moore and Thompson. Int Orthop 38:655–661 4. Crossman PT, Khan RJ, MacDowell A, Gardner AC, Reddy NS, Keane GS (2002) A survey of the treatment of displaced intracapsular femoral neck fractures in the UK. Injury 33:383–386 5. The National Hip Fracture Database (2011) Preliminary National Report 2011. http://www.nhfd.co.uk/20/hipfractureR.nsf/resour ceDisplay?openform. Accessed 25 Nov 2013 6. Parker MJ, Gurusamy KS, Azegami S (2010) Arthroplasties (with and without bone cement) for proximal femoral fractures in adults. Cochrane Database Syst Rev 6:CD001706 7. Khan RJ, MacDowell A, Crossman P, Keene GS (2002) Cemented or uncemented hemiarthroplasty for displaced intracapsular fractures of the hip—a systematic review. Injury 33:13–17 8. NHS Supply Chain (2012) Orthopaedic Data Evaluation Panel. http://www.supplychain.nhs.uk/product-areas/orthopaedics/odep. Accessed 25 Nov 2013 9. National Institute for Health and Clinical Excellence (2011) The management of hip fracture in adults (Clinical guideline CG124). http://www.nice.org.uk/nicemedia/live/13489/54919/54919.pdf. Accessed 25 Nov 2013 10. Leonardsson O, Garellick G, Ka¨rrholm J, Akesson K, Rogmark C (2012) Changes in implant choice and surgical technique for hemiarthroplasty. Acta Orthop 83:7–13 11. Australian Orthopaedic Association National Joint Replacement Registry (2011) Annual report 2011. https://aoanjrr.dmac.ade laide.edu.au/annual-reports-2011. Accessed 25 November 2013 12. Murray DW, Carr AJ, Bulstrode C (1993) Survival analysis of joint replacements. J Bone Joint Surg Br 75:697–704 13. Hudson JI, Kenzora JE, Hebel JR, Gardner JF, Scherlis L, Epstein RS, Magaziner JS (1998) Eight-year outcome associated with clinical options in the management of femoral neck fractures. Clin Orthop Relat Res 348:59–66 14. Parker MI, Pryor G, Gurusamy K (2010) Cemented versus uncemented hemiarthroplasty for intracapsular hip fractures: a randomised controlled trial in 400 patients. J Bone Joint Surg Br 92:116–122 15. Ruiz-Iba´n MA, Crespo-Herna´ndez P, Ferna´ndez-Rolda´n S, DiazHerediaa J, Martinez-Urenaa P, Muriel A (2008) Cemented hemiarthroplasty after a femoral neck fracture. Survival analysis. Rev Esp Cir Ortop Traumatol 52:206–212
123
Eur J Orthop Surg Traumatol 16. Kassam AM, Griffiths S, Higgins G (2014) Historical implant or current best standard? Minimum five year follow-up outcomes of cemented Thompson hemiarthroplasties. J Arthroplast (published online 05 May 2014) 17. Pajarinen J, Savolainen V, Tulikoura I, Lindahl J, Hirvensalo E (2003) Factors predisposing to dislocation of the Thompson’s hemiarthroplasty: 22 dislocations in 338 patients. Acta Orthop Scand 74:45–48 18. Enocson A, Tidermark J, Tornkvist H, Lapidus LJ (2008) Dislocation of hemiar-throplasty after femoral neck fracture: better outcome after the anterolateral approach in a prospective cohort study on 739 consecutive hips. Acta Orthop 79:211–217 19. Garellick G, Ka¨rrholm J, Herberts P (2010) Swedish Hip Arthroplasty Register annual report 2009. http://www.shpr.se/ Libraries/Documents/AnnualReport-2009.sflb.ashx. Accessed 25 Nov 2013 20. Phillips TW (1989) Thompson’s hemiarthroplasty and acetabular erosion. J Bone Joint Surg Am 71-A:913–917 21. Parker MJ (2012) Cemented Thompson’s hemiarthroplasty versus cemented Exeter Trauma Stem (ETS) hemiarthroplasty for intracapsular hip fractures: a randomised trial of 200 patients. Injury 43:807–810
123
22. Petheram TG, Bone M, Joyce TJ, Serrano-Pedraza I, Reed MR, Partington PF (2013) Surface finish of the Exeter Trauma Stem: a cause for concern? Bone Joint J 95-B(2):173–176 23. Howie DW, Middleton RG, Costi K (1998) Loosening of matt and polished cemented femoral stems. J Bone Joint Surg [Br] 80-B:573–576 24. Bidwai AS, Willett KM (2012) Comparison of the Exeter Trauma Stem and the Thompson hemiarthroplasty for intracapsular hip fractures. Hip Int 22:655–660 25. Ryg J, Rejnmark L, Overgaard S, Brixen K, Vestergaard P (2009) Hip fracture patients at risk of second hip fracture: a nationwide population-based cohort study of 169,145 cases during 1977–2001. J Bone Miner Res 24:1299–1307 26. Gillam MH, Salter A, Ryan P, Graves SE (2011) Different competing risks models applied to data from the Australian Orthopaedic Association National Joint Replacement Registry. Acta Orthop 82:513–520 27. National Institute for Health and Clinical Excellence (2000) Guidance on the Selection of Prostheses for Primary Total Hip replacement. http://www.nice.org.uk/nicemedia/live/11386/ 32002/32002.pdf. Accessed 25 Nov 2013
ORIGINAL ARTICLE
Cemented Thompson’s hemiarthroplasty in patients with intracapsular neck of femur fractures: survival analysis of 1,670 procedures S. K. Khan • S. S. Jameson • A. Sims • J. A’Court • M. R. Reed • A. Rangan • S. D. Muller
Received: 18 February 2014 / Accepted: 5 August 2014 Ó Springer-Verlag France 2014
Abstract Cemented Thompson’s prostheses have been used to treat elderly patients with displaced intracapsular hip fractures at our two units for the last 15 years, amid growing support for the use of newer implant designs for hip hemiarthroplasty. This provided us with an opportunity to investigate survival of the Thompson’s stem in our patients. A retrospective cohort study was set up to review previously collected data on patients who underwent Thompson’s hemiarthroplasty over a 7-year period. These were linked to surgical notes, clinical letters and radiographs to record post-operative course and subsequent admissions and procedures. The identifiers were then linked to mortality data from the Office of ‘National Statistics. Kaplan–Meier survival analyses were done for implants and patients. A total of 1,632 patients (mean age 82.7 years) underwent 1,670 procedures. Five-year implant survival was 95.4 %. A total of 36 stems were revised, including 14 revisions to total hip arthroplasty and 22 excision arthroplasties. Reasons for revision included infection (2.1 %), dislocation (1.1 %) and aseptic loosening (0.5 %). Symptomatic aseptic loosening and acetabular erosion occurred late (mean time 3.2 and 5.7 years,
S. K. Khan S. S. Jameson J. A’Court M. R. Reed S. D. Muller Northumbria Healthcare NHS Foundation Trust, Ashington NE63 9JJ, UK S. K. Khan (&) Apartment 509, The Bar, St James Gate, Newcastle upon Tyne NE1 4BA, UK e-mail: [email protected]; [email protected] A. Sims A. Rangan South Tees Hospitals NHS Trust, Marton Road, Middlesbrough TS4 3BW, UK
respectively following surgery). Aseptic loosening and erosion following hemiarthroplasty are relatively late complications. Keywords Hemiarthroplasty Thompson’s prosthesis Implant survival Patient survival Revision Dislocation Acetabular erosion
Introduction The incidence of intracapsular neck of femur (NOF) fracture is increasing in the United Kingdom (UK) [1]. Hip arthroplasty (hemiarthroplasty or total hip arthroplasty) is the standard treatment for the majority of these displaced fractures. The Thompson’s prosthesis is a press fit hemiarthroplasty implant designed in the pre-cement era [2, 3]. It is now most frequently used with cement [4, 5], with improved clinical effectiveness compared to uncemented monoblock (e.g. Austin-Moore) stems [6, 7]. The Orthopaedic Device Evaluation Panel (ODEP) produces ratings for hip arthroplasty prostheses used in the UK, based on length of follow-up and quality of evidence [8]. The Thompson’s stem currently does not have an ODEP rating, presumably because of the relative paucity of implant survival data in this at-risk patient group. The National Institute for Health and Clinical Excellence (NICE) offers evidence-based guidance to clinicians in the UK. The recent NICE guidance on NOF fracture management recommends the use of ‘proven’ cemented stem designs with an ODEP rating of at least 3B (97 % survival at 3 years), instead of the Thompson’s prosthesis [9]. Given that none of the available newer designs currently carry an ODEP rating, recommendations against the use of the Thompson’s prosthesis do not seem clearly justified.
123
Eur J Orthop Surg Traumatol
There are conflicting global trends for the use of Thompson’s prosthesis in favour of newer stems. A recent review of the Swedish Hip Arthroplasty Register has reported a decline in the proportion of Thompson’s stems from 9 to 1 % from 2005 to 2009 with an increasing number of unipolar modular prostheses being implanted [10]. The 2011 annual report from Australian National Joint Registry found that the use of the Thompson’s prosthesis had increased in 2010 (30.6 % of all monoblocks, up from 20.9 % in 2003); however, the number of monoblock stems was declining compared to modular and bipolar implants [11]. Cemented Thompson’s prostheses have been used to treat all elderly patients with displaced intracapsular NOF fracture at our two units over the last decade. This provided us with a unique opportunity to investigate outcomes in these patients. Three aims were defined as follows: (1) to examine the survival of the Thompson’s implant, (2) to investigate the indications for revision and (3) to compare patient survival to implant survival.
Methods A two-centre retrospective cohort study was designed to analyse implant and patient survival in patients who underwent cemented Thompson’s hemiarthroplasty for intracapsular NOF fracture between 1 January 2003 and 31 Fig. 1 Flowchart showing constitution of the final data set
123
December 2010. Unit 1 has one of the largest geographical catchment areas in the UK, and provides trauma services at three acute hospitals, while Unit 2 is a regional major trauma centre. Ethical approval was obtained for this study. Patients were identified from the hospitals’ central registers. Previously collected data on patients’ age, date of surgery and length of hospital stay were requested. Prophylactic antibiotic use, implant type, laterality of procedure and use of cement were confirmed from surgical and clinical notes by linking the patient identifiers to the online databases SIRIS (Surgical Information Recording & Interrogation System; Xentec Ltd, Wallsend, UK), CaMIS (Clinical and Management Information System; Ascribe Ltd, Bolton, UK), TheatreMan (Trisoft Limited, Chesterfield, UK) and the picture archiving and communication system (PACS) at the two Trusts (IMPAX, Agfa HealthCare, Brentford, UK). Exclusion criteria from the final subset included patients with pathological fractures due to neoplastic disease, those having the Thompson’s hemiarthroplasty as a revision procedure and those having uncemented procedures (Fig. 1). These linkages were next used to identify subsequent orthopaedic admissions from the Emergency departments or from clinic appointments requested by the patients’ general practitioners. A minimum data set was defined for each admission, including diagnosis (e.g. infection, dislocation, aseptic loosening, acetabular erosion, periprosthetic fractures, contralateral NOF fractures and persistent pain),
Eur J Orthop Surg Traumatol
type of procedures performed (e.g. washout and debridement, closed or open reduction, revision to total hip arthroplasty (THA), excision arthroplasty and fracture fixation) and any further complications. A revision was defined as any procedure in which the original cemented Thompson’s stem was removed, i.e. either excision arthroplasty or revision to THA. Survival (Kaplan–Meier) analyses were performed for implant and patient survival following Thompson hemiarthroplasty. Life tables were produced to report yearly implant survival rates for all procedures included in the study, with numbers entering each year reported [12]. Survival was not reported once the number entering a year fell below 5 % of the original number. In order to generate patient survival data, patient identifiers were linked to the Public Health and Mortality File generated by the Office for National Statistics (ONS; Cardiff, UK) for dates and places of death. All analyses were performed using SPSS version 19 (SPSS Inc, IBM Corporation, Armonk, NY). Fig. 2 Kaplan–Meier curve for implant survivorship
Results One thousand six hundred and thirty-two patients underwent 1,670 cemented Thompson’s hemiarthroplasties during the study period. These included 1,594 patients with unilateral hemiarthroplasties, 37 patients with bilateral non-contemporary procedures and one patient with bilateral simultaneous procedures. The mean age at presentation was 82.7 years (range 52–102 years, SD 7.3 years). There was a notable increase in the mean age over the 7-year study period, rising from 77.6 years in 2004 to 83.4 years in 2010. Implant complications and survival Seventy-six complications (4.5 % complication rate) were identified (Table 1). Infection was the commonest problem, occurring after 36 procedures (2.1 % infection rate). Table 1 Complications identified in current series of 1,670 prostheses Complication
% of all procedures (number)
% of all revisions
1
Infection
2.1 (n = 36)
44.4
2 3
Dislocation Aseptic loosening
1.1 (n = 18) 0.5 (n = 8)
27.7 19.4
4
Periprosthetic fracture
0.4 (n = 6)
5.6
5
Persistent pain
0.3 (n = 5)
2.8
6
Protrusion
0.1 (n = 2)
0
7
Bleeding
0.1 (n = 1)
0
Twenty of these stems were retained; 19 had a surgical debridement and washout in theatre, while one infection settled on antibiotics alone. Sixteen stems were revised, including 12 excision arthroplasties and four stems undergoing two-stage revisions to THA. The mean time to revision for infection was 254 days. Aseptic dislocation occurred in 18 out of 1,670 implants (1.1 % dislocation rate). Six stems were reduced under general anaesthesia and thus retained. The remaining 12 stems were either irreducible or re-dislocated again, necessitating revision. The mean time to revision for dislocation was 31 days. Except for one outlier, all were performed within 60 days of the index procedure. Aseptic loosening was identified in 8 implants in this series (0.5 % aseptic loosening rate). Seven patients returned to theatre for revision, including 6 THAs and one excision arthroplasty. The mean time to revision for aseptic loosening was 3.2 years. Two patients were symptomatic with acetabular protrusion (confirmed on radiographs) when reviewed in clinic. This review occurred at a mean of 5.6 years after the index procedure. Both the patients were offered surgery (intention to revise), but declined. Of all the identified complications, 40 implants were retained and 36 were revised (2.1 % all-cause revision rate). The latter included 22 excision arthroplasties (61 % of revisions) and 14 revisions to THA (39 %). The mean time to all-cause revision was 1.2 years (range 3 days–5.8 years). Implant survival over the study is shown in Fig. 2. Five-year implant survival was 95 % (Table 2).
123
Eur J Orthop Surg Traumatol Table 2 Implant survival for Thompson’s population, by year since index procedure
Interval start time (start year)
Number at start
Failure
Withdrawn
Number at-risk
0
1,670
25
494
1,422
2
98
98
1
1,150
2
401
950
0
100
98
2
747
3
281
607
0
100
98
3
463
2
210
358
1
99
97
4
251
3
134
184
2
98
95
5
114
1
73
78
1
99
94
6 7
40 21
0 0
19 11
31 16
0 0
100 100
94 94
8
10
0
10
5
0
100
94
Fig. 3 Kaplan–Meier curve for patient survivorship
Patient survival Thirty-day and 1-year survival was 91.6 and 69.6 %, respectively. This declined to 35 % at 5 years and just over 20 % at 8 years. Patient survival is demonstrated in Fig. 3.
Discussion This two-centre retrospective study reports patient and implant survival after cemented Thompson’s hemiarthroplasty in a large consecutive series of fragility NOF fractures. There were four key findings: (1) implant survival was 95 % at 5 years, (2) patient survival was 35 % at 5 years, (3) the mean age at presentation with hip fractures
123
Annual failure rate (%)
Annual success rate (%)
Survival rate (%)
is increasing and (4) acetabular erosion and aseptic loosening are relatively late-occurring complications. Our patient mortality and implant revision rates were in agreement with those reported in literature to date [13–15]. This study thus supports the use of Thompson’s prosthesis in the majority of patients with displaced NOF fractures. The dislocation rate (1.1 %) is surprising given that the design of the Thompson’s would intuitively lend itself to a lower risk of dislocation, i.e. a large head that confers a favourable head/neck ratio and increases the excursion distance. Our outcome is however comparable to contemporarily published series [15, 16] and is substantially less than the overall 7 % dislocation rate reported in a previous review of 338 cemented Thompson’s implants [17]. Scandinavian studies have found the posterior approach to be significantly associated with dislocation [17, 18]. All procedures in our series were performed using the anterolateral approach. This is the approach most commonly used in hemiarthroplasty for NOF fracture in the UK and is currently gaining favour in Sweden [19]. Our data also confirms aseptic loosening to be a late complication of hemiarthroplasty. Given that patient survival is poor, it is not surprising that aseptic loosening was rare in this series. Acetabular erosion [20] was similarly found to be an extremely uncommon (0.2 %) late complication. The Thompson prosthesis is now over 60 years old and was designed before the use of cement. Advocates of newer implant designs, e.g. Exeter Trauma Stem (ETS, StrykerTM UK Ltd., Newbury, UK) draw attention to potential advantages of these designs, including lesser soft tissue exposure, optimal restoration of femoral neck offset, anteversion and leg length, ease of implantation and cement-in-cement revision if needed [21]. The ETS is also a monoblock with a surface finish (roughness average) about ten times more than the Exeter arthroplasty stem [22]. This may result in aseptic loosening, like the mattsurface finish did to the Exeter stem in 1970s [23]; it may also influence loosening rates. The perceived advantages of
Eur J Orthop Surg Traumatol
the ETS over the Thompson’s have however not translated into significant differences in short-and medium-term outcomes in the only published randomised controlled trial comparing the two implants [21]. The Thompson’s is cheaper too, costing £130 (€165) less than the ETS [21]. In a historical cohort comparison between the ETS (n = 388) and Thompsons (n = 766), Bidwai and Willett found no difference in surgical or radiological outcomes [24]. The recent Australian joint registry report also showed that short-term revision rates for the cemented Thompson’s and the ETS were comparable (1.04 vs. 1.10 revisions/100 observed component years), despite potential selection bias with slightly younger and fitter patients receiving the ETS [11]. One of the strengths of this study is the large captive population, which historically has remained static, i.e. has continued to reside within the same catchment area postdischarge. We believe capturing all deaths including out of hospital deaths by way of the ONS data is a major strength of this study, making the patient survival analysis reliable. Moreover, 2.2 % patients in our series returned to have a contralateral Thompson’s hemiarthroplasty at a median of 0.8 years (range 14 days–7.2 years); this contralateral fracture rate is comparable to that reported in nationwide population-based Scandinavian studies [25]. Finally, all recorded readmissions with complications and subsequent revisions at our units have been reported, making this the largest reported series on implant survival for Thompson’s stems. We acknowledge potential limitations of this study. Most Hospitals in the English NHS do not arrange routine follow-up for patients undergoing hemiarthroplasty. Survival analysis has certain shortcomings. Our outcome measure for implant ‘survival’ was a revision procedure. However, patients living with a poorly performing or painful prostheses, but who are unable to have (or do not want) further surgery, cannot be deemed a success. We do not have data on surviving patients with problems who were never referred back. Similarly, we do not have data on any patients who may have had revisions outside the region due to relocation or tertiary referral pathways. We believe this is unlikely, given the characteristics of our combined catchment area. Lastly, survival curves in the presence of competing risks (in this case, the large competing risk of death) may overestimate the risk of revision [26]; therefore, the true revision risk for these patients may actually be smaller than we report here. In conclusion, this study found an increasingly ageing fracture population with 35 % patient survival, but 95 % implant survival, at 5 years. Our data show latency in developing aseptic loosening and acetabular erosion, which are the usually cited reasons for revision. The implant clearly satisfies the minimum ODEP 3B rating now
mandated by NICE [27]. In fact in this study, it displayed a profile consistent with an ODEP 5B rating, however this is a matter for ODEP to decide. Abandoning the Thompson’s hemiarthroplasty in favour of newer designs is not sufficiently substantiated by currently available evidence. The clinical and cost effectiveness of the newer hemiarthroplasty stems needs to be ascertained through well-designed comparative studies. Conflict of interest of interest.
The authors declare that they have no conflict
References 1. White SM, Griffiths R (2011) Projected incidence of proximal femoral fracture in England: a report from the NHS Hip Fracture Anaesthesia Network (HIPFAN). Injury 42:1230–1233 2. Thompson FR (1954) Two and a half years experience with a vitallium intramedullary hip prosthesis. J Bone Joint Surg 36A:489–502 3. Hernigou P, Quiennec S, Guissou I (2014) Hip hemiarthroplasty: from Venable and Bohlman to Moore and Thompson. Int Orthop 38:655–661 4. Crossman PT, Khan RJ, MacDowell A, Gardner AC, Reddy NS, Keane GS (2002) A survey of the treatment of displaced intracapsular femoral neck fractures in the UK. Injury 33:383–386 5. The National Hip Fracture Database (2011) Preliminary National Report 2011. http://www.nhfd.co.uk/20/hipfractureR.nsf/resour ceDisplay?openform. Accessed 25 Nov 2013 6. Parker MJ, Gurusamy KS, Azegami S (2010) Arthroplasties (with and without bone cement) for proximal femoral fractures in adults. Cochrane Database Syst Rev 6:CD001706 7. Khan RJ, MacDowell A, Crossman P, Keene GS (2002) Cemented or uncemented hemiarthroplasty for displaced intracapsular fractures of the hip—a systematic review. Injury 33:13–17 8. NHS Supply Chain (2012) Orthopaedic Data Evaluation Panel. http://www.supplychain.nhs.uk/product-areas/orthopaedics/odep. Accessed 25 Nov 2013 9. National Institute for Health and Clinical Excellence (2011) The management of hip fracture in adults (Clinical guideline CG124). http://www.nice.org.uk/nicemedia/live/13489/54919/54919.pdf. Accessed 25 Nov 2013 10. Leonardsson O, Garellick G, Ka¨rrholm J, Akesson K, Rogmark C (2012) Changes in implant choice and surgical technique for hemiarthroplasty. Acta Orthop 83:7–13 11. Australian Orthopaedic Association National Joint Replacement Registry (2011) Annual report 2011. https://aoanjrr.dmac.ade laide.edu.au/annual-reports-2011. Accessed 25 November 2013 12. Murray DW, Carr AJ, Bulstrode C (1993) Survival analysis of joint replacements. J Bone Joint Surg Br 75:697–704 13. Hudson JI, Kenzora JE, Hebel JR, Gardner JF, Scherlis L, Epstein RS, Magaziner JS (1998) Eight-year outcome associated with clinical options in the management of femoral neck fractures. Clin Orthop Relat Res 348:59–66 14. Parker MI, Pryor G, Gurusamy K (2010) Cemented versus uncemented hemiarthroplasty for intracapsular hip fractures: a randomised controlled trial in 400 patients. J Bone Joint Surg Br 92:116–122 15. Ruiz-Iba´n MA, Crespo-Herna´ndez P, Ferna´ndez-Rolda´n S, DiazHerediaa J, Martinez-Urenaa P, Muriel A (2008) Cemented hemiarthroplasty after a femoral neck fracture. Survival analysis. Rev Esp Cir Ortop Traumatol 52:206–212
123
Eur J Orthop Surg Traumatol 16. Kassam AM, Griffiths S, Higgins G (2014) Historical implant or current best standard? Minimum five year follow-up outcomes of cemented Thompson hemiarthroplasties. J Arthroplast (published online 05 May 2014) 17. Pajarinen J, Savolainen V, Tulikoura I, Lindahl J, Hirvensalo E (2003) Factors predisposing to dislocation of the Thompson’s hemiarthroplasty: 22 dislocations in 338 patients. Acta Orthop Scand 74:45–48 18. Enocson A, Tidermark J, Tornkvist H, Lapidus LJ (2008) Dislocation of hemiar-throplasty after femoral neck fracture: better outcome after the anterolateral approach in a prospective cohort study on 739 consecutive hips. Acta Orthop 79:211–217 19. Garellick G, Ka¨rrholm J, Herberts P (2010) Swedish Hip Arthroplasty Register annual report 2009. http://www.shpr.se/ Libraries/Documents/AnnualReport-2009.sflb.ashx. Accessed 25 Nov 2013 20. Phillips TW (1989) Thompson’s hemiarthroplasty and acetabular erosion. J Bone Joint Surg Am 71-A:913–917 21. Parker MJ (2012) Cemented Thompson’s hemiarthroplasty versus cemented Exeter Trauma Stem (ETS) hemiarthroplasty for intracapsular hip fractures: a randomised trial of 200 patients. Injury 43:807–810
123
22. Petheram TG, Bone M, Joyce TJ, Serrano-Pedraza I, Reed MR, Partington PF (2013) Surface finish of the Exeter Trauma Stem: a cause for concern? Bone Joint J 95-B(2):173–176 23. Howie DW, Middleton RG, Costi K (1998) Loosening of matt and polished cemented femoral stems. J Bone Joint Surg [Br] 80-B:573–576 24. Bidwai AS, Willett KM (2012) Comparison of the Exeter Trauma Stem and the Thompson hemiarthroplasty for intracapsular hip fractures. Hip Int 22:655–660 25. Ryg J, Rejnmark L, Overgaard S, Brixen K, Vestergaard P (2009) Hip fracture patients at risk of second hip fracture: a nationwide population-based cohort study of 169,145 cases during 1977–2001. J Bone Miner Res 24:1299–1307 26. Gillam MH, Salter A, Ryan P, Graves SE (2011) Different competing risks models applied to data from the Australian Orthopaedic Association National Joint Replacement Registry. Acta Orthop 82:513–520 27. National Institute for Health and Clinical Excellence (2000) Guidance on the Selection of Prostheses for Primary Total Hip replacement. http://www.nice.org.uk/nicemedia/live/11386/ 32002/32002.pdf. Accessed 25 Nov 2013
