You are on page 1of 5

The Journal of Arthroplasty 29 (2014) 17741778

Contents lists available at ScienceDirect

The Journal of Arthroplasty

journal homepage:

Why Are Total Knee Arthroplasties Failing TodayHas Anything

Changed After 10 Years?
Peter F. Sharkey, MD, Paul M. Lichstein, MD, MS, Chao Shen, MD, Anthony T. Tokarski, BS,
Javad Parvizi, MD, FRCS
The Rothman Institute at Thomas Jefferson University, Philadelphia, Pennsylvania

a r t i c l e i n f o a b s t r a c t

Article history: The purpose of this study was to determine the frequency and cause of failure after total knee arthroplasty
Received 19 April 2013 and compare the results with those reported by our similar investigation conducted 10 years ago. A total of
Accepted 21 July 2013 781 revision TKAs performed at our institution over the past 10 years were identied. The most common
failure mechanisms were: loosening (39.9%), infection (27.4%), instability (7.5%), periprosthetic fracture
(4.7%), and arthrobrosis (4.5%). Infection was the most common failure mechanism for early revision
knee arthroplasty
(b 2 years from primary) and aseptic loosening was the most common reason for late revision. Polyethylene
revision (PE) wear was no longer the major cause of failure. Compared to our previous report, the percentage of
revisions performed for polyethylene wear, instability, arthrobrosis, malalignment and extensor mechanism
deciency has decreased.
2014 Published by Elsevier Inc.

Since its introduction, total knee arthroplasty (TKA) has evolved into A previous study performed 10 years ago at our institution deter-
one of the most successful and thoroughly investigated surgical in- mined the mechanisms of TKA failure between 1997 and 2000 [9].
terventions, especially over the past two decades [1]. Survival rates greater However, surgical technique, prosthetic design, perioperative care,
than 90% after 10 years follow-up highlight the durability of primary TKA and surgical experience have evolved over the past 10 years. Addi-
[2]. With expectations for successful outcomes, the number of primary tionally, it has been suggested that advancements in prosthetic com-
TKA performed annually is projected to increase exponentially in the next ponent design and surgical instrumentation have generated a
few years [3,4]. In 2003 there were approximately 402,100 primary TKA paradigm shift in the etiology of common failure mechanisms [8].
performed and demand is anticipated to increase by 673% to over three The purpose of the current study was to elucidate if the etiology of
million procedures by the year 2030 [5]. Despite long survivorship and failure of TKA has changed over the past decade at our institution.
improved outcomes, the increasing number of primary TKA has been
associated with increased rates of revision TKA procedures [1]. Methods and Materials
An aging population and the acceptance of TKA in young active
patients have contributed to the increasing number of both primary Following Institutional Review Board approval, we performed a
and revision TKA procedures performed annually. The Australian systematic retrospective review of all revision TKA performed at our
national registry reported revision TKA accounted for 8.3% of all knee institution between July 1, 2003 and July 1, 2012. During this time
replacement surgeries conducted in 2011 and this number trends period, 10,003 total knee arthroplasty surgeries were performed, with
higher yearly [6]. The most frequently reported mechanisms of failure 781 (7.8%) revision surgeries. Important information pertinent to
in primary TKA have varied but consistently include periprosthetic demographics such as age, gender, weight, and race was recorded. The
joint infection (PJI), loosening, and instability [7]. Recent studies have interval from primary TKA to revision procedures was obtained in
further analyzed failure mechanisms of primary TKA highlighting addition to whether the primary TKA had been performed at our
not only the role of PJI and aseptic loosening in failure, but also how institution or was referred from elsewhere. Failure mechanisms were
such failure mechanisms may affect revision surgery outcomes [7,8]. determined by review of our institutional prospective revision data-
Knowledge of the etiology of failure mechanisms is of paramount base and corroborated by review of operative records. At our insti-
importance for delivery of appropriate care. tution, data on revision cases are collected prospectively which
includes collection of intraoperative data by a research fellow who is
present during each revision arthroplasty. With the assistance of the
primary surgeon, the research fellow completes questionnaires
The Conict of Interest statement associated with this article can be found at http://
related to the cause of failure and nature of intraoperative ndings.
Reprint requests: Peter F Sharkey MD, The Rothman Institute Research, The All revision TKA patients were subdivided into early and late
Sheridan Building, 125th South 9th & Samson Street, Ste 1000, Philadelphia PA 19107. failure groups depending upon the time interval between primary
0883-5403/ 2014 Published by Elsevier Inc.
P.F. Sharkey et al. / The Journal of Arthroplasty 29 (2014) 17741778 1775

TKA and revision procedure with two years being considered as the exchanged in 81 (10%) cases. Isolated patella resurfacing was per-
cut-off between early and late failures. The cause of overall failure, formed in 9 (1%), (Table).
as well as a comparison in the cause of failure of TKA in 2002 Aseptic loosening of the prosthesis was the most common etiology
(previously reported) and 2012 was conducted. Statistical analysis of failure overall, with component loosening observed in 39.9% of all
was performed using SPSS Version 15.0 (SPSS Inc., Chicago, Illinois) revision procedures. Loosening of the prosthesis was more common in
software and signicance was determined using Chi Squared analy- the late revision group, accounting for 51.4% of patients undergoing
sis for dichotomous comparisons of early, late, and overall failure revision more than two years following index arthroplasty. By con-
mechanisms. Patients were further categorized by the number of trast, loosening accounted for 22.8% of early revision cases. Instability
components involved in the revision procedure and grouped ac- was observed in 6.1% of early and 10.3% of late revision surgeries.
cordingly as complete (three-component revision or resection arthro- The most common failure mechanism in the early revision group
plasty), two components, single component, PE exchange, or isolated was infection, which was responsible for 37.6% of revisions performed
patellar resurfacing. less than 2 years after the primary procedure. Infection also
accounted for 21.9% of revision TKA in the late failure group and for
Results 27.4% of overall revision TKA procedures at our institution.
Complications related to the extensor mechanism were an
There were 781 total knee arthroplasty revisions included in the important cause of revision TKA. The etiology of failure for exten-
present study. Revision surgeries were performed in 318 (41%) sor mechanism related complications in all revision TKA included
patients referred from outside institutions. 453 patients were female loosening of the patellar component (3%), isolated patellar resurfacing
and 291 were male. The average age of male patients was 65.4 years (4%), and extensor mechanism deciency (0.3%).
(range, 3796 years) and the average age of female patients was Cumulatively, the incidence of PE wear, with or without osteolysis,
65.1 years (range, 3486 years) at time of revision TKA. The average was 3.5% for overall revision TKA procedures and accounted for 4.3%
body mass index was 33.06 (kg/m 2) (range, 17.760.7) in female of the patients in the late failure group and 2% of the patients in the
patients and 31.9 (range, 18.662.8) in male patients. Revision sur- early failure group. The overall incidence of other causes for revision
gery was conducted for 667 posterior stabilized, 62 cruciate retaining, TKA was similar in both the early and late revision groups. The over-
49 unicondylar, and 3 constrained primary TKAs. all incidence of arthrobrosis and peri-prosthetic fracture was 4.5%
The average time before revision TKA in the early failure group and 4.7%, respectively.
was 0.84 years (range, 1 day to 1.97 years). In the late failure group, Compared with our results of 10 years ago, the current data sug-
the average time to revision was 6.9 years (range, 2.01 years to gest polyethylene (PE) wear is no longer the primary cause of TKA
30.36 years). Fig. 1 displays the mechanism of the failure and the failure. Figs. 3 and 4 display a comparison of the percentage of patients
corresponding percentage of patients with each failure mode. Fig. 2 in 2002 and 2012 with TKA failure stratied by mechanism and
displays the percentage of patients with each failure mechanism segregated by early and late presentation. The most common indi-
stratied into early, late, and overall failure subgroups. The results of cations in our present cohort for revision TKA in all the patients were
our prior investigation are provided for comparison. Patients under- loosening (39.9%), infection (27.4%), instability (7.5%) and peripros-
going early revision comprised 37.6% of our cohort and 62.4% of the thetic fractures (4.7%). 62.4% of the revision TKA procedures were
revisions were performed more than 2 years after the initial surgery. performed more than 2 years after primary surgery. Additionally,
Of the revision surgeries, 299 (38%) involved exchange of all TKA as depicted in Fig. 5, the overall incidence of loosening and infec-
components. Two components were exchanged in 273 (35%) cases. A tion was seen to have increased since 2002 and predominate in our
PE exchanged was performed in 119 (15%) cases. One component was recent cohort. However, signicant reductions were seen in the

Fig. 1. The percentage of patients with each failure mechanism stratied into early and late subgroups is shown.
1776 P.F. Sharkey et al. / The Journal of Arthroplasty 29 (2014) 17741778

Fig. 2. Percentage of patients with each failure mechanism stratied into early, late, and overall failure subgroups is shown.

incidence of instability, arthrobrosis, malalignment, and extensor methodology. There is debate concerning optimal xation which
mechanism deciency. include cement technique and use of cementless components.
Although cemented xation has been reported to be less technically
challenging and provides initial secure prosthesis stability, cementless
xation offers potential advantages such as less fat embolization and
reliable long-term xation. Recently, Nakama et al demonstrated that
Although the incidence of revision TKA is relatively low in general,
cementless xation had a higher rate of early loosening but low risk
the percentage continues to steadily increase annually [1]. Therefore,
for late aseptic loosening [12]. In contrast, other studies have demon-
we undertook the present investigation to determine the incidence
strated that there was no signicant difference in the incidence of
of various failure modes in contemporary TKA requiring revision
aseptic loosening in primary TKA between cemented and cementless
surgery. An updated study into the evolution of the mechanisms of
prostheses [12,13]. In a meta-analysis conducted by Gandhi et al [13]
failure over the past decade is intended to assist investigators in
randomized controlled trials of cemented and cementless methods
identifying and avoiding common failure modes. Our results indicate
were reviewed and it was concluded that implant survival and func-
aseptic prosthetic loosening, infection and post-operative instability
tional scores were similar in both groups [11]. Regardless of prosthesis
were the three most common reasons for all contemporary revision
type, another study showed revision TKA for aseptic loosening gene-
TKA procedures. In further subgroup analysis, PJI was the most
rated superior patient satisfaction and functional outcomes when
common cause of revision in patients who underwent revision within
compared to revision TKA for all other indications [14].
two years of their primary TKA and loosening of the prosthesis was
According to our study, infection has now become the most
the most common reason for revision in patients undergoing the
common mode of early failed TKA. Although perioperative prophy-
procedure greater than two years from their primary TKA. Less than
lactic antibiotics and other anti-infection modalities have been widely
50% of the revision TKA procedures were done within 2 years
utilized in primary TKA, infection is still one of the most common
following primary TKA compared with 55.6% in our previous report.
major complications related to joint replacement surgery. Not only
PE wear was no longer the most common cause of revision TKA in this
was infection the most likely cause of early failure, it was the second
contemporary cohort of patients, a signicant difference from our
most common cause in the late revision group as well. These results
conclusions of 10 years ago [7], likely highlighting the successful
have also been reported by others. After reviewing 349 failed TKA
development of more wear resistant biomaterials or better locking
cases, Hossain et al [8] demonstrated that infection accounted for 30%
mechanism for the polyethylene into the tibial tray [10].
of all revision procedures and in fact was the most common cause
In the present study, aseptic prosthetic loosening was the most
of failure in both primary and revision TKA. Bozic et al [7] reviewed
common indication for revision TKA in the past ten years. Similar
and analyzed 60,357 revision TKAs in the United States and found a
results have been noted from national joint registry annual reports.
similar incidence of infection. According to the 2012 annual report of
According to the Australian and Swedish 2011 annual reports of knee
the England and Wales national joint registry, infection accounts for
arthroplasty, the most common reason for revision TKA was aseptic
23% of revision TKA procedures. Other annual reports from national
loosening, accounting for more than 20% of all the revision TKA [6,11].
joint registries found similar trends [11]. Two stage revision pro-
Loosening of the prosthesis is related to TKA component xation
cedures for the treatment of infected TKA are considered the gold
standard. Sadly, the outcomes of revision TKA for infection have been
Table shown to be suboptimal. Vanhegan et al [15] suggested that revision
Number of Components Exchanged. TKA is more likely to be successful when the causative organism was
Components Exchanged Number of Cases known. Therefore, research for early and accurate diagnosis of in-
fecting organisms could increase the rate of successful septic revision
1 81 (10%)
2 273 (35%) surgery in the future.
Total 299 (38%) Instability was a consistent mechanism of failure comparing our
PE exchange 119 (15%) current and prior investigations. Although it was again demonstrated
Resurfacing 9 (1%) to be the third most common failure mode in primary TKA, the over-
PE = polyethylene. all incidence has decreased dramatically since our last report. There
P.F. Sharkey et al. / The Journal of Arthroplasty 29 (2014) 17741778 1777

Fig. 3. Comparison of the percentage of patients in 2002 and 2012 with early TKA failure stratied by mechanism. *P b 0.05, **P b 0.01.

were 62 cruciate retaining primary TKAs that underwent revision, are involved in the development of arthrobrosis following TKA
of those, 11 (17.7%) underwent revision for instability. Of the 667 [17,18]. Preoperative knee range of motion, surgical technique, post-
posterior stabilized primary TKAs that underwent revision, instability operative rehabilitation, and patient specic biologic factors all play
was the failure mode in 56 knees (8.4%). Such improvements in a role in stiffness after primary TKA. Most authors agree that limited
stability following TKA may be attributable to advancements in pre-operative range of motion is the strongest predictor of post-
operative technique, but the development and increased utilization of operative stiffness [1921].
posterior stabilized prostheses may also be responsible. The success of Revision TKA is one method used to treat severe stiffness following
CR TKA is more sensitive than PS TKA to optimal anatomic ligament primary TKA [22]. Kim et al [19] reviewed 54 patients who had stiffness
tensioning. This issue may prove more challenging for surgeons after primary TKA and demonstrated that although revision TKA was
performing low volumes of TKA. Of the 11 CR knees that underwent expected to improve both functional outcome and range of motion, this
revision, all (100%) cases were performed at an outside institution. improvement was modest. Furthermore, in a systematic review of the
Establishing stability with a PS design may be more reliable and less treatment of arthrobrosis after primary TKA, revision resulted in the
challenging. Of the 56 knees undergoing revision for instability, 32 least improvement compared to other methods reviewed (manipula-
(57.1%) were originally performed outside of our institution. tion under anesthesia, arthroscopic release and open release) [23]. After
The incidence of patients developing arthrobrosis following pri- reviewing 25 revision TKA patients who exhibited postoperative
mary TKA has decreased in recent years. Several authors have stiffness, Keeney et al [24] suggested that the results of revision TKA
reported that the incidence of arthrobrosis was less than 5% in all for stiffness were less predictable than revisions for other reasons. In our
primary TKA patients [14,16]. It is well accepted that multiple factors study, we found a relationship between arthrobrosis and chronic

Fig. 4. Comparison of the percentage of patients in 2002 and 2012 with late TKA failure stratied by mechanism. **P b 0.01.
1778 P.F. Sharkey et al. / The Journal of Arthroplasty 29 (2014) 17741778

Fig. 5. Comparison of overall incidence of failure between 2002 and 2012 stratied by mechanism. *P b 0.05, *P b 0.01.

infection. This nding has also been reported by others [19,25]. If 2. Abdel MP, Morrey ME, Jensen MR, et al. Increased long-term survival of posterior
cruciate-retaining versus posterior cruciate-stabilizing total knee replacements.
patients exhibit stiffness after primary TKA, a work-up to rule out deep J Bone Joint Surg Am 2011;93:2072.
infection should be performed. 3. Kim H-A, Kim S, Seo YI, et al. The epidemiology of total knee replacement in South
Our study has limitations. First, we only included patients under- Korea: national registry data. Rheumatology (Oxford) 2008;47:88.
4. Singh JA, Vessely MB, Harmsen WS, et al. A population-based study of trends in
going revision TKA at a single institution where referral for chal- the use of total hip and total knee arthroplasty, 19692008. Mayo Clin Proc 2010;
lenging revision is common. The results from a large population or a 85:898.
national registry database might yield different ndings. Another 5. Kurtz S, Ong K, Lau E, et al. Projections of primary and revision hip and knee
arthroplasty in the United States from 2005 to 2030. J Bone Joint Surg Am 2007;
limitation is the follow-up interval for each patient varied and 89:780.
therefore limits some conclusions. The third limitation is that not all of 6. Anon. The Swedish Knee Arthroplasty Register Annual Report; 2011.
the primary TKA procedures were performed at our institute. Despite 7. Bozic KJ, Kurtz SM, Lau E, et al. The epidemiology of revision total knee arthroplasty
in the United States. Clin Orthop Relat Res 2010;468:45.
these limitations, we were able to illustrate the etiological changes of
8. Hossain F, Patel S, Haddad FS. Midterm assessment of causes and results of revision
revision TKA with a large cohort of patients. The strength of our study total knee arthroplasty. Clin Orthop Relat Res 2010;468:1221.
compared to registry data is that the cause of failure is much more 9. Sharkey PF, Hozack WJ, Rothman RH, et al. Insall Award paper. Why are total knee
likely to be accurate as a dedicated research fellow was present during arthroplasties failing today? Clin Orthop Relat Res 2002:7.
10. Kurtz SM, Gawel HA, Patel JD. History and systematic review of wear and osteolysis
these revision arthroplasties to collect the necessary data related to outcomes for rst-generation highly crosslinked polyethylene. Clin Orthop Relat
revision surgery. Res 2011;469:2262.
Our results showed that aseptic loosening, infection, instability, 11. Anon. National Joint Registry for England and Wales Annual Report; 2012.
12. Nakama GY, Peccin MS, Almeida GJM, et al. Cemented, cementless or hybrid
periprosthetic fracture, and arthrobrosis are the most common xation options in total knee arthroplasty for osteoarthritis and other non-
etiologies of failure in TKA over the last ten years for patients revised traumatic diseases. Cochrane Database Syst Rev 2012;10 CD006193.
at our institution. Infection is the most common mechanism of failure 13. Gandhi R, Tsvetkov D, Davey JR, et al. Survival and clinical function of cemented and
uncemented prostheses in total knee replacement: a meta-analysis. J Bone Joint
in patients revised in less than two years and second only to aseptic Surg Br 2009;91:889.
loosening in late failures. Ruling out PJI should be prioritized in 14. Baker P, Cowling P, Kurtz S, et al. Reason for revision inuences early patient
patients with primary TKA failure. In contrast, aseptic loosening is the outcomes after aseptic knee revision. Clin Orthop Relat Res 2012;470:2244.
15. Vanhegan IS, Morgan-Jones R, Barrett DS, et al. Developing a strategy to treat
most common overall reason for failure in TKA patients. Given the established infection in total knee replacement: a review of the latest evidence and
majority of TKA patients revised for loosening were more than two clinical practice. J Bone Joint Surg Br 2012;94:875.
years from primary surgery; stable long term xation is a critical 16. Zmistowski B, Restrepo C, Kahl LK, et al. Incidence and reasons for nonrevision
reoperation after total knee arthroplasty. Clin Orthop Relat Res 2011;469:138.
concern. Instability remains a common failure etiology; however the
17. Bong MR, Di Cesare PE. Stiffness after total knee arthroplasty. J Am Acad Orthop
incidence has decreased dramatically possibly secondary to advance- Surg 2004;12:164.
ments in surgical technique and more common use of prosthetic 18. Freeman TA, Parvizi J, Della Valle CJ, et al. Reactive oxygen and nitrogen species
designs utilizing a posterior stabilized construct. Arthrobrosis induce protein and DNA modications driving arthrobrosis following total knee
arthroplasty. Fibrogenesis Tissue Repair 2009;2:5.
continues to be encountered frequently and although further 19. Kim J, Nelson CL, Lotke PA. Stiffness after total knee arthroplasty. Prevalence of the
investigation into its relationship with infection is necessary, the complication and outcomes of revision. J Bone Joint Surg Am 2004;86-A:1479.
clinician is well advised to screen for PJI in cases of failure secondary 20. Gandhi R, de Beer J, Leone J, et al. Predictive risk factors for stiff knees in total knee
arthroplasty. J Arthroplasty 2006;21:46.
to arthrobrosis. PE wear, which had been shown to be the most 21. Winemaker M, Rahman WA, Petruccelli D, et al. Preoperative knee stiffness and
common etiology in our past investigation, has decreased remarkably total knee arthroplasty outcomes. J Arthroplasty 2012;27:1437.
with the development of more resilient biomaterials and is no longer 22. Kim GK, Mortazavi SMJ, Parvizi J, et al. Revision for stiffness following TKA: a
predictable procedure? Knee 2012;19:332.
the most common indication for revision TKA at our institution. 23. Ghani H, Maffulli N, Khanduja V. Management of stiffness following total knee
arthroplasty: A systematic review. Knee 2012;19:751.
References 24. Keeney JA, Clohisy JC, Curry M, et al. Revision total knee arthroplasty for restricted
motion. Clin Orthop Relat Res 2005;440:135.
1. Cram P, Lu X, Kates SL, et al. Total knee arthroplasty volume, utilization, and 25. Su EP, Su SL, Della Valle AG. Stiffness after TKR: how to avoid repeat surgery.
outcomes among Medicare beneciaries, 19912010. JAMA 2012;308:1227. Orthopedics 2010;33:658.