Assessment Of Early Re-Interventions For Persistent Pain Following Phase 3 Oxford Unicondylar Knee Replacement

David M, Datta A, Baloch K
Royal Orthopaedic Hospital Birmingham

Address for Correspondence:
Michael David
ST3 West Midlands Deanery
‘Ashiana’, 35 Wrekin Road
Sutton Coldfield
West Midlands, B73 5SU, UK

Phone : +44 (0)7966 765712
E-mail : michaeldavid@nhs.net

Abstract:

Introduction: unicompartmental knee replacement (UKR) is a popular alternative to total knee replacement (TKR) in medial compartment disease.  Revision to TKR is well described as a late endpoint.  Our objective is to investigate early surgical management of persistent pain following UKR, identify common themes and effectiveness of re-interventions.
Methods: 381 UKRs implanted over 5 years included, and patients requiring re-operation reviewed retrospectively.
Findings: 27 re-interventions performed on 17 patients at a mean 16.8 months (95% CI 9.5 to 24.1), with symptom onset post-operatively at 9 months (95% CI 4-14).  There were 10 arthroscopies, 10 revisions to TKR, 4 manipulations under anaesthesia (MUA), 2 bearing exchanges, and 1 tibial-plateau fracture fixation.  MUA improved stiffness 2 of 3 patients.  Arthroscopy was successful in 2 patients with loose cement-bodies, not providing a diagnosis in 8, of whom 7 required subsequent revision.  Overall there were 10 revisions: 9 were performed for persistent pain and 9 reported symptom improvement.  Intra-operative findings included aseptic loosening (n=4), synovitis (n=2), increased posterior slope of the tibial cut (n=1), dislocated bearing (n=1), and no cause of failure (n=2). There were no deep infections.
Conclusions:   Our unit’s early re-intervention rate is 4.5% (95% CI 2.4 to 6.5), with a revision rate of 2.6% (95% CI 1.0 to 4.2) after a mean (±SD) follow-up of 40 (±16) months.  Arthroscopy is a poor diagnostic and therapeutic option against persistent pain following UKR.  In contrast, the decision to revise, although initially disappointing for both patient and surgeon, gave symptom improvement in 90%.

J.Orthopaedics 2010;7(3)e5

Keywords:

Introduction:

The mobile bearing unicondylar knee replacement (UKR) was developed in 1978, with the first Oxford partial knee arthroplasty performed in 1982.¹ Survival at ten years is quoted at 98% (95% CI 93% to 100%) by the design group² and 95% (95% CI 90.8% to 99.3%) quoted by an independent study.³ These excellent long term results have encouraged an increased use of this implant for unicompartmental arthritis of the knee in recent years.⁴

The Oxford group suggest one in three of all knees requiring arthroplasty are suitable for a UKR,⁵ despite this only approximately 8% of all knee arthroplasties are unicompartmental replacements.⁴ This may be because there is concern following UKR of persistent or new onset post-operative pain. The pain could indicate an incorrect initial choice of implant, disease progression to unresurfaced compartments, aseptic loosening or infection.⁴ Further management of this group of patients presents a clinical dilemma, especially if laboratory and radiological investigations do not highlight a potential cause for their symptoms. Conservative measures are usually recommended to patients, with advice that the symptoms are likely to settle with time or that the symptoms are not amenable to surgical intervention.  This reflects the established management of similar pain following total knee replacement (TKR).⁶  Despite these reassurances there are a small number of patients who consider their symptoms debilitating enough to warrant further surgery.

The Royal Orthopaedic Hospital (Birmingham) is a specialised Orthopaedic tertiary referral centre in the United Kingdom dealing with a large throughput of arthroplasty patients, and having on-site dedicated knee arthroplasty surgeons as well as orthopaedic nurses and physiotherapists.  The objectives of this study are to establish early re-operation rates following UKR and assess the results of subsequent operative intervention in patients with debilitation pain following primary UKR. 

Materials and Methods:

All Phase 3 Oxford UKRs (Biomet, Bridgend, UK) implanted at our institution between November 2002 and December 2007 were included in the study. 40 patients underwent bilateral surgery. Surgery was performed under the supervision of seven consultant arthroplasty surgeons as described in the manual of surgical technique.⁷  Patients who required any subsequent surgery on the ipsilateral knee were highlighted retrospectively using our proprietary database, and their medical notes individually reviewed.

Results :

383 UKRs were performed on 343 patients over this five year period, with a mean (±SD) follow-up of 40.1 months (±16 months).  Two patients were excluded as they underwent simultaneous ACL reconstruction leaving a cohort of 381 UKRs.

There were 27 re-operations performed in 17 patients [see table 1], equivalent to a re-intervention rate of 4.5% (95% CI 2.4 to 6.5).  The mean time to the first re-operation was 16.8 months (95% CI 9.5 to 24.1) from the primary procedure.  Procedures undertaken were: 10 revisions to total knee replacements (1 performed in two stages), 10 arthroscopies, 4 manipulations under anaesthetic (MUA), 2 exchange of mobile bearings and 1 open reduction internal fixation of a tibial plateau fracture.

Manipulation under anaesthetic

Three patients with ongoing limited range of movement were manipulated.  Satisfactory range of movement was achieved for two patients.  Despite two manipulations the final patient did not improve until the mobile bearing was exchanged for a smaller component.  This resulted in a satisfactory outcome.

Table 1: Summary of patients requiring re-intervention

Arthroscopies

Ten patients were managed with arthroscopic procedures, but only two (20%) had significant resolution of their symptoms.  In both these cases loose cement fragments were seen in post-operative radiographs and arthroscopic removal settled their symptoms.  Arthroscopic assessment of the remaining eight patients revealed no gross loosening of the implants when probed and no gross disease progression.  One patient was noted to have lateral partial thickness fissuring in the lateral compartment, while four patients had evidence of early patello-femoral wear.  All eight were subsequently revised to total knee replacements, with symptom improvement in seven.  Mean interval between arthroscopy and revision was 17.1 months (95% CI 10.1 to 24.1).

 Revision to total knee replacement

In total ten UKRs were revised to total knee replacements, eight as previously discussed were following initial arthroscopic surgery, and two following radiographic suspicion of aseptic loosening.  Mean time from primary UKR to revision was 38.2 months (95% CI 28.1 to 48.3). Aseptic loosening of the components was demonstrated intraoperatively in four patients (3 in the femoral component and 1 in the tibial component).  In those UKRs with stable implants, intraoperative findings included: synovitis with histology consistent with a prominent foreign body reaction (n=2), and increased posterior slope of the tibial cut (n=1).  No clear cause was identified in two patients.

Nine of the ten revisions were single-stage procedures.  The only two-stage revision was initially planned as a one stage revision but converted after intra-operative findings of widespread haemosiderin deposits.  Subsequent microbiological and histological investigations excluded infection.  A primary prosthesis was implanted in eight cases (7 AGCs Biomet, Bridgend UK & 1 Kinemax Stryker Howmedica ), and a revision implant with 10mm medial augments was required in two patients for associated bone loss (Maxim Biomet, Warsaw, IN).  There were no revisions for infection.

Amongst the ten patients revised for ongoing pain nine had a significant improved in symptoms, with complete resolution of symptoms in seven.  One patient had non-specific ongoing symptoms whilst another complained non-dermatomal pain.  Both considered these symptoms acceptable.  The final patient with a poor outcome had an encouraging improvement in symptoms initially, but after a fall within eight weeks of his revision developed symptoms consistent with complex regional pain syndrome.

The early re-operation rate following Oxford unicompartmental knee replacement in our unit is 4.5% (95% CI 2.4 to 6.5), with a revision rate of 2.6% (95% CI 1.0 to 4.2) after a mean (±SD) follow-up of 40.1 (±16) months.

Discussion :

The ideal indication for unicompartmental knee replacement is “anteromedial arthritis”.⁸  Clinically this equates to an intact anterior cruciate ligament, varus deformity that is correctable at 20° flexion, functionally normal medial collateral ligament and full thickness cartilage in the lateral compartment.  Chondral ulcers on the medial aspect of the lateral femoral condyle and patellofemoral arthritis are not considered contraindications by the Oxford group.⁹  The Oxford group state that using the above criteria, 1 in 3 knees suitable for arthroplasty are appropriate for their implant.⁵  The current phase III Oxford implant was introduced in 1998 and is implanted using minimally invasive techniques.  In the Oxford knee this is associated with half the recovery time of the full arthrotomy, and one-third the recovery time of a total knee arthroplasty.¹⁰

Despite the encouraging results with the Oxford unicompartmental knee replacement (UKR) only 8% of all primary knee arthroplasty are unicompartmental, of which 72% are the Oxford phase III implant.⁴  There are a combination of design, philosophical and clinically-related reasons that some surgeons limit their own indications for the implantation of a unicompartmental knee replacement. Technically, unicompartmental arthroplasty is more challenging than a total arthroplasty as exact soft tissue balancing is required.¹¹  The addition of a mobile bearing to the Oxford design, with the potential for dislocation, reduces the margin for error further.  Studies have shown that technical errors are the main cause of early failure, long term studies have shown that if survival is good at seven years then long term results are likely to be good.¹⁰  This pattern of failure is unique to congruent mobile bearing designs as wear rates are low, which appears to have resulted in improved long term results in comparison to other designs.²

The potential problem of persistent postoperative pain, with its subsequent management challenges, has also raised concerns.  Failure to fully resolve all symptoms following a TKR is also well recognised and once infection, aseptic loosening and gross malalignment are excluded it is usually managed conservatively.  This course of action is realistically harder to follow in any form of partial replacement, as the potential always exists for further surgery to the unresurfaced components.  However, conservative measures are still usually recommended to recipients of a UKR as they are to their TKR counterparts, with advice that the symptoms are likely to settle with time or that the symptoms are not amenable to surgical intervention.   Despite these reassurances there are a small number of patients who consider their symptoms debilitating enough to warrant further surgery.  Clinically this group is often difficult to manage appropriately.

The source of the persistent debilitating symptoms is not always clear despite clinical examination, serial laboratory and radiological investigations.  In our specialist unit the preferred primary re-operation chosen for persistent pain was knee arthroscopy.  Symptom improvement was only achieved in 2 out of 10 patients, both of whom had a clear pre-arthroscopy diagnosis of loose cement fragments. In the remaining patients symptoms did not improve significantly enough to prevent revision to a TKR.  Arthroscopic assessment further failed to identify loose prostheses that were apparent at the time of revision.  Our findings would suggest that the role of arthroscopy is limited in the management of a painful UKR as it is a poor diagnostic and therapeutic modality.

The decision to revise to a total knee replacement is never straightforward, for the patient it involves further major surgery with no guarantee of symptom resolution and for the surgeon it confirms failure of the original UKR.  This difficult decision is demonstrated in the mean 17-month interval between the initial arthroscopy and subsequent revision.  Despite these misgivings our results are encouraging, with nine (90%) of patients revised to TKR achieving clinically significant symptom improvement, of which seven (70%) are pain free.  Aseptic loosening was the most consistent intra-operative finding identified in four patients.  The bone loss associated with loosening was minimal, and only two cases required a revision implant.  Our findings would suggest that although the decision to revise is initially disappointing there is a marked improvement in symptoms.

Pandit⁵ reviewed 688 consecutive Oxford knee replacements from the Oxford group: there were eight (1.2%) revisions within 2.5 years of the index procedure and seven other re-operations identified over an unspecified time period.  Indications for revision were infection in four cases, two dislocated bearings and two for unexplained pain.  Despite revision surgery, pain persisted in one of these patients.  Re-operation was required for stiffness in four cases, pain resulted in arthroscopy in two patients and one patient underwent debridement for a superficial infection.  It is not clear from their data if the patients that were formally revised to a total knee replacement underwent other procedures prior to revision.  Extrapolating their data, the presumed “best-case” scenario for early re-operation is fifteen cases (2.2%) with a revision rate of 1.2%.

Svärd’s³ series of 103 Oxford replacements identified three revisions (2.9%) and one re-operation (0.9%) within two years of the index procedure. The indications for revision were pain and clinical suspicion of infection respectively. The re-operation was for a suspected loose body, this was not located but symptoms improved.  No mention was made in the paper with respect to complications of stiffness and the requirement for manipulation under anaesthetic.  Again it is not clear from their data if the patients that were formally revised to a total knee replacement underwent other procedures prior to revision.  Extrapolating their data the presumed “best-case” scenario for early re-operation is four cases (3.8%) with a revision rate of 2.9%.

Reviewing the outcome of 381 Oxford knee replacements in our institution, seventeen (4.5%) patients required re-operations for persistent symptoms. In ten (2.6%) patients a decision was been made to revise the primary implant to a total knee replacement.

Our figures are higher than the extrapolated “best-case” scenario data published from the design centre in Oxford⁵ (2.2% and 1.2%) and broadly comparable with the independent series published by Svärd³ (3.8% and 2.9%).

Conclusions:

4.5% (95% CI 2.4% to 6.5%) of patients required a re-operation for ongoing symptoms following an Oxford UKR.  Our revision rate is 2.6% (95% CI 1.0% to 4.2%), which is comparable to current published data.  Management of debilitating pain following unicondylar knee surgery can be difficult.  Arthroscopy is a poor diagnostic and therapeutic modality for this clinical problem, and only recommended if a correctible problem is identified clearly prior (e.g. a loose body).  Our results suggest that if appropriate conservative measures have failed, persistent pain is managed most reliably with revision to a total knee replacement.

Reference:

1. Goodfellow J, O’Connor J. The mechanics of the knee and prosthesis design. J Bone Joint Surg (Br), 1978; 60-B:358-369
2. Murray D, Goodfellow J, O’Connor J. The Oxford medial unicompartmental arthroplasty – a ten year survival study. J Bone Joint Surg (Br), 1998; 80-B (6):983-989
3. Svärd U, Price A. Oxford medial unicompartmental knee arthroplasty. A survival analysis of an independent series. J Bone Joint Surg (Br), 2001; 83:191-194
4. National Joint Registry for England and Wales 5th annual Report 2008
5. Pandit H, Jenkins C, Barker K, Dodd C, Murray D. The Oxford medial unicompartmental knee replacement using a minimally-invasive approach. J Bone Joint Surg (Br), 2006; 88-B:54-60
6. Elsom DW, Brenkel IJ. A Conservative approach is feasible in unexplained pain after knee replacement. J Bone Joint Surg (Br), 2007; 89-B:1042-1045
7. Oxford™ Unicompartmental Knee: Manual of the Surgical Technique, Biomet UK Ltd (UK)
8. White S, Ludkowski P, Goodfellow J. Anteromedial osteoarthritis of the knee. J Bone Joint Surg (Br), 1991; 73-B:582-586
9. D. J. Beard, H. Pandit, H. S. Gill, D. Hollinghurst, C. A. F. Dodd, D. W. Murray. The influence of the presence and severity of pre-existing patellofemoral degenerative changes on the outcome of the Oxford medial unicompartmental knee replacement. J Bone Joint Surg Br, Dec 2007; 89-B: 1597 - 1601.
10. Price A, Webb J, Topf H, Dodd CA, Goodfellow JW, Murray DW. Rapid Recovery after Oxford unicompartmental arthroplasty through a short incision. J Arthroplasty, 2001; 16:970-976
11. Robertsson O, Knutson K, Lewold S, Lidgren L. The routine of surgical management reduces failure after unicompartmental knee arthroplasty. J Bone Joint Surg Br, 2001, 83(1):45-4

This is a peer reviewed paper 

Please cite as: Michael David: Assessment Of Early Re-Interventions For Persistent Pain Following Phase 3 Oxford Unicondylar Knee Replacement

J.Orthopaedics 2010;7(3)e5

URL: http://www.jortho.org/2010/7/3/e5

ANNOUNCEMENTS

Arthrocon 2011


Refresher Course in Hip Arthroplasty

13th March,  2011

At Malabar Palace,
Calicut, Kerala, India

Download Registration Form

For Details
Dr Anwar Marthya,
Ph:+91 9961303044

E-Mail:
anwarmh@gmail.com

Powered by
VirtualMedOnline