2

C OPYRIGHT 2014
BY

T HE J OURNAL

OF

B ONE

AND J OINT

S URGERY, I NCORPORATED

The Risk of Knee Arthroplasty Following Cruciate Ligament Reconstruction

A Population-Based Matched Cohort Study
Timothy Leroux, MD, MEd, Darrell Ogilvie-Harris, MBBS, FRCSC, Tim Dwyer, MD, FRCSC, FRACS, Jaskarndip Chahal, MD, MSc, FRCSC, Rajiv Gandhi, MD, MSc, FRCSC, Nizar Mahomed, MD, ScD, FRCSC, and David Wasserstein, MD, MSc, FRCSC
Investigation performed at the Institute for Clinical Evaluative Sciences, Toronto, Ontario, Canada, and Toronto Western Hospital (University Health Network), Toronto, Ontario, Canada

Background: Evidence regarding the risk of end-stage osteoarthritis following cruciate ligament reconstruction is based upon small sample sizes and radiographic, rather than clinical, criteria. The goals of this study were to determine the risk of knee arthroplasty, a surrogate for end-stage osteoarthritis, following cruciate ligament reconstruction, and to identify patient, provider, and surgical factors that inuence knee arthroplasty risk. Methods: Using administrative databases, we identied all patients who were sixteen to sixty years of age and had undergone cruciate ligament reconstruction in Ontario from July 1993 to March 2008. Case patients were matched by demographic variables to ve individuals without knee injury from the general population of Ontario, Canada, who had not undergone previous knee surgery, including cruciate ligament reconstruction. The main outcome was knee arthroplasty. Kaplan-Meier survival curves were generated for both cohorts. A Cox proportional hazards model determined those factors that inuenced knee arthroplasty risk. Results: We identied 30,301 eligible patients who had undergone cruciate ligament reconstruction; of these patients, 30,277 were matched to 151,362 individuals from the general population; the median patient age was twenty-eight years and 65% of the patients were male. Primary anterior cruciate ligament reconstruction accounted for >98% of index cases. During the follow-up period, there was a signicant difference (p < 0.001) between matched case and control cohorts with respect to the number of patients who underwent knee arthroplasty during the study period; in the matched case cohort, 209 patients underwent knee arthroplasty (event rate, 0.68 of 1000 person-years), and in the control cohort, 125 patients underwent knee arthroplasty (event rate, 0.10 of 1000 person-years). Moreover, fteen years after cruciate ligament reconstruction (case cohort) or study enrollment (control cohort), there was a signicant difference (p < 0.001) in the cumulative incidence of knee arthroplasty between the case cohort (1.4%) and the control cohort (0.2%). Age of fty years or more (hazard ratio, 37.28; p < 0.001), female sex (hazard ratio, 1.58; p = 0.001), comorbidity score of 5 points (hazard ratio, 5.91; p = 0.002), surgeon continued
Disclosure: None of the authors received payments or services, either directly or indirectly (i.e., via his or her institution), from a third party in support of any aspect of this work. One or more of the authors, or his or her institution, has had a nancial relationship, in the thirty-six months prior to submission of this work, with an entity in the biomedical arena that could be perceived to inuence or have the potential to inuence what is written in this work. No author has had any other relationships, or has engaged in any other activities, that could be perceived to inuence or have the potential to inuence what is written in this work. The complete Disclosures of Potential Conicts of Interest submitted by authors are always provided with the online version of the article. Disclaimer: The Institute for Clinical Evaluative Sciences (ICES) supported the following study. The ICES is funded by an annual grant from the Ontario Ministry of Health and Long-Term Care (MOHLTC). The opinions, results and conclusions reported in this paper are those of the authors and are independent from the funding sources. No endorsement by ICES or the Ontario MOHLTC is intended or should be inferred. A commentary by Robert A. Magnussen, MD, is linked to the online version of this article at jbjs.org.

J Bone Joint Surg Am. 2014;96:2-10

http://dx.doi.org/10.2106/JBJS.M.00393

Downloaded From: http://jbjs.org/ by a CURTIN UNIV OF TECHNOLOGY User on 01/27/2014

3
TH E JO U R NA L O F B O N E & JO I N T SU RG E RY J B J S . O RG V O L U M E 96-A N U M B E R 1 J A N UA RY 1, 2 014
d d d

T H E R I S K O F K N E E A R T H R O P L A S T Y F O L L O W I N G C R U C I AT E LIG AMENT RECONSTRUCTION

annual volume of cruciate ligament reconstruction of twelve or fewer cases per year (hazard ratio, 2.53; p < 0.001), and cruciate ligament reconstruction undertaken in university-afliated hospitals (hazard ratio, 1.51, p = 0.008) increased the odds of knee arthroplasty; however, male sex (hazard ratio, 0.63; p = 0.001) and patient age of less than twenty years (hazard ratio, 0.07; p = 0.009) were protective. Concurrent meniscal repair or debridement did not increase arthroscopy risk.

Conclusions: After fteen years, the cumulative incidence of knee arthroplasty following cruciate ligament reconstruction was low (1.4%); however, it was seven times greater than the cumulative incidence of knee arthroplasty among matched control patients from the general population (0.2%). Older age, female sex, higher comorbidity, low surgeon annual volume of cruciate ligament reconstruction, and cruciate ligament reconstruction performed in a universityafliated hospital were factors that increased knee arthroplasty risk. Level of Evidence: Prognostic Level III. See Instructions for Authors for a complete description of levels of evidence.

Peer Review: This article was reviewed by the Editor-in-Chief and one Deputy Editor, and it underwent blinded review by two or more outside experts. It was also reviewed by an expert in methodology and statistics. The Deputy Editor reviewed each revision of the article, and it underwent a nal review by the Editor-in-Chief prior to publication. Final corrections and clarications occurred during one or more exchanges between the author(s) and copyeditors.

vidence pertaining to the risk of osteoarthritis following cruciate ligament reconstruction is conicted. A recent systematic review showed the prevalence of osteoarthritis following anterior cruciate ligament reconstruction to be 29% to 51% at a minimum ten-year follow-up1. However, the authors of this systematic review noted that, among the highestquality studies (as determined by a modied version of the Coleman methodology score2), the prevalence of osteoarthritis at a minimum ten-year follow-up was only 0% to 13%1. The authors also criticized the existing literature for its heterogeneity, small sample sizes, and the use of radiographic, rather than clinical, criteria to diagnose osteoarthritis1. There remains a need to determine the risk of end-stage osteoarthritis following cruciate ligament reconstruction and to evaluate long-term patient symptoms and function using meaningful osteoarthritis outcome measures. A commonly performed and effective surgical intervention for end-stage osteoarthritis is prosthetic knee arthroplasty, including total knee arthroplasty and unicompartmental knee arthroplasty3. The indications for arthroplasty are largely clinical and include patient symptoms, patient function, and failure of alternative non-surgical and surgical interventions. For these reasons, knee arthroplasty is a reasonable surrogate for clinical end-stage osteoarthritis. In the present study, two matched population cohorts were developed to compare the rate of knee arthroplasty following cruciate ligament reconstruction with the rate of knee arthroplasty in the general population. Moreover, we identied and explored the patient, provider, and surgical factors that increased knee arthroplasty risk following cruciate ligament reconstruction. Materials and Methods Case Cohort Development

vices to all Ontario residents, and OHIP fee codes are highly accurate (>96%) . For each patient, a unique and anonymous identication number was used to conrm a relevant index hospital admission in the Canadian Institute for Health Information (CIHI) databases (Same Day Surgery Database [SDS] or Discharge Abstract Database [DAD]). Exclusion criteria (see Appendix) included age less than sixteen years and age greater than sixty years, non-Ontario residents, nonelective cruciate ligament reconstruction, billing anomalies, prior knee arthroplasty or cruciate ligament reconstruction, and potential confounders of heightened osteoarthritis risk (chondral surgery, lower-extremity osteotomy, or fracture xation about the knee). Revision cruciate ligament reconstruction cases were also excluded following the introduction of a relevant OHIP billing code on July 1, 2003. Of note, database limitations precluded the exclusion of patients who had undergone posterior cruciate ligament reconstruction prior to 2002. As such, all patients who had undergone cruciate ligament reconstruction (anterior cruciate ligament reconstruction or posterior cruciate ligament reconstruction) were included to lengthen the duration of followup and to capture the infrequent event of knee arthroplasty.

Control Cohort Development

Case patients were randomly matched to ve control patients from Ontario, Canada. Control subjects met the aforementioned exclusion criteria, had valid OHIP coverage, and were matched to the corresponding case patient on the basis of patient demographic variables (age [within one year], sex, income quintile, urban or rural address, and comorbidity score). A case patient was unmatched if ve control patients sharing similar demographics could not be identied.

Main Outcome
The main outcome was knee arthroplasty (total knee arthroplasty or unicompartmental knee arthroplasty) and was identied using OHIP fee codes (see Appendix). Laterality was not available for the index event or the main outcome.

Covariates
Available patient, surgical, and provider factors were considered in the analysis. Patient factors included age, sex, neighborhood income quintile, comorbidity score, and residence. Neighborhood income quintile was calculated 5 through an established technique with use of Statistics Canada census data . Patient comorbidity was estimated with use of the Collapsed Aggregated Di6 agnosis Groups (CADG) method . Specically, CIHI databases and International Classication of Diseases (ICD) codes were used to identify comorbid conditions associated with admissions to Canadian hospitals within two years of the index event. On the basis of the number of comorbid conditions, patients were then assigned an overall CADG score and were assigned to one of two 7 categories: 0 to 4 points and 5 points . The postal code for each patient was used

n this matched cohort study, all patients who underwent cruciate ligament reconstruction (anterior cruciate ligament reconstruction and posterior cruciate ligament reconstruction) from July 1993 to March 2008 were identied from Ontario Health Insurance Plan (OHIP) physician billings and were accessed through the Institute for Clinical Evaluative Sciences (ICES; www.ices. on.ca). In Ontario, OHIP provides universal health coverage for medical ser-

Downloaded From: http://jbjs.org/ by a CURTIN UNIV OF TECHNOLOGY User on 01/27/2014

4
TH E JO U R NA L O F B O N E & JO I N T SU RG E RY J B J S . O RG V O L U M E 96-A N U M B E R 1 J A N UA RY 1, 2 014
d d d

T H E R I S K O F K N E E A R T H R O P L A S T Y F O L L O W I N G C R U C I AT E LIG AMENT RECONSTRUCTION

TABLE I Cumulative Incidence of Knee Arthroplasty: Case Versus Control Cohorts Time Since Cruciate Ligament Reconstruction (yr) 2 5 10 15 Cumulative Incidence of Knee Arthroplasty* Case Cohort 0.06 (0.04 to 0.09) 0.15 (0.11 to 0.20) 0.40 (0.32 to 0.49) 1.44 (1.22 to 1.69) Control Cohort 0.002 (0.0004 to 0.006) 0.02 (0.01 to 0.02) 0.06 (0.05 to 0.09) 0.21 (0.17 to 0.26) Cumulative Incidence Comparison (Case:Control) 30:1 8:1 7:1 7:1

*The values are given as the percentage, with the 95% CI in parentheses.

TABLE II Within-Group Comparison: Case Cohort (N = 30,301) Variable Group size* Age (yr) Age distribution Sixteen to nineteen years of age Twenty to twenty-four years of age Twenty-ve to twenty-nine years of age Thirty to thirty-four years of age Thirty-ve to thirty-nine years of age Forty to forty-nine years of age Fifty to sixty years of age Male sex CADG score of 5 points Surgeon annual volume in the year preceding cruciate ligament reconstruction None One to twelve Thirteen to fty Fifty-one to 100 More than 100 Year of index date distribution (non-fragmented) 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 With Knee Arthroplasty 212 42 (35 to 47) 0% 2% 6% 14% 17% 42% 17% 54% 1% 4% 28% 36% 18% 13% 14% 11% 12% 10% 11% 3% 5% 5% 7% 6% 3% 1% 3% 1% Without Knee Arthroplasty 30,089 28 (20 to 35) 22% 18% 17% 16% 13% 13% 2% 65% 0% 3% 19% 33% 24% 21% 5% 5% 6% 6% 6% 6% 6% 6% 7% 7% 8% 9% 9% 9%

*The values are given as the number of patients. The values are given as the median, with the interquartile range in parentheses. The differences were signicant at p < 0.05. The values are given as the percentage of patients.

Downloaded From: http://jbjs.org/ by a CURTIN UNIV OF TECHNOLOGY User on 01/27/2014

5
TH E JO U R NA L O F B O N E & JO I N T SU RG E RY J B J S . O RG V O L U M E 96-A N U M B E R 1 J A N UA RY 1, 2 014
d d d

T H E R I S K O F K N E E A R T H R O P L A S T Y F O L L O W I N G C R U C I AT E LIG AMENT RECONSTRUCTION

TABLE III Cox Proportional Hazard Model: Knee Arthroplasty Risk Factors in Case Cohort Variable Age Sixteen to nineteen years versus twenty-ve to twenty-nine years Twenty to twenty-four years versus twenty-ve to twenty-nine years Thirty to thirty-four years versus twenty-ve to twenty-nine years Thirty-ve to thirty-nine years versus twenty-ve to twenty-nine years Forty to forty-nine years versus twenty-ve to twenty-nine years Fifty to sixty years versus twenty-ve to twenty-nine years Sex Female versus male Male versus female Income quintile 1 versus 3 2 versus 3 4 versus 3 5 versus 3 CADG 5 points versus 0 to 4 points Surgeon annual volume in the year preceding cruciate ligament reconstruction None versus more than 100 One to twelve versus more than 100 Thirteen to fty versus more than 100 Fifty-one to 100 versus more than 100 Meniscal debridement Yes versus no Meniscal repair Yes versus no Academic health-care center (site of index cruciate ligament reconstruction) Yes versus no Hazard Ratio* P Value

0.07 (0.01 to 0.50) 0.40 (0.14 to 1.13) 2.43 (1.26 to 4.68) 4.35 (2.31 to 8.19) 13.77 (7.64 to 24.79) 37.28 (19.59 to 70.93) 1.58 (1.20 to 2.07) 0.63 (0.48 to 0.83) 0.83 (0.50 to 1.38) 1.11 (0.72 to 1.72) 0.87 (0.58 to 1.32) 0.77 (0.52 to 1.14) 5.91 (1.87 to 18.64) 1.88 (0.88 to 4.03) 2.53 (1.56 to 4.11) 2.15 (1.37 to 3.37) 1.17 (0.72 to 1.91) 0.72 (0.52 to 0.99) 0.89 (0.48 to 1.65) 1.51 (1.12 to 2.04)

0.009 0.083 0.008 <0.001 <0.001 <0.001 0.001 0.001 0.467 0.623 0.522 0.192 0.002 0.103 <0.001 <0.001 0.534 0.046 0.713 0.008

*The values are given as the hazard ratio, with the 95% CI in parentheses. Signicance was set at p = 0.05.

to determine urban or rural residence. Of note, age categories were determined a priori and were used for ease of interpretation (sixteen to nineteen years, twenty to twenty-four years, twenty-ve to twenty-nine years, thirty to thirty-four years, thirty-ve to thirty-nine years, forty to forty-nine years, and fty to sixty years). Provider factors included surgeon annual volume of cruciate ligament reconstruction and hospital academic status. The surgeon for each index event was identied anonymously, and the number of cruciate ligament reconstructions performed in the year preceding the index event (July 1 to June 30) was determined. Volume categories were determined a priori on the basis of clinical relevance (none, one to twelve, thirteen to fty, fty-one to 100, and more than 100 8 cruciate ligament reconstructions per year) . The index hospital was categorized as either university-afliated (active in research and/or health-care professional training) or non-university-afliated on the basis of membership in the Council of Academic Hospitals of Ontario (www.caho-hospitals.com). Surgical factors included concomitant meniscal surgery, specically meniscectomy or meniscal repair (see Appendix). Database limitations precluded the identication of meniscal tears treated nonoperatively.

Statistical Analysis
An overall knee arthroplasty event rate per 1000 person-years was calculated for each cohort and was compared via t test. Within the case cohort only, a t test

(continuous variables) or chi-square test (categorical variables) was used to compare demographics between the group of patients who underwent knee arthroplasty and the group of patients who did not. A Kaplan-Meier survival analysis (index event to censorship) was generated for each cohort and was compared with use of a log-rank test. Both the survival rate and the cumulative incidence (the inverse of the survival rate) at two, ve, ten, and fteen years from the index event (case cohort) or study inclusion (control cohort) were calculated. Censorship included knee arthroplasty, death, emigration, loss of OHIP coverage, and study end (March 31, 2012). Two Cox proportional hazards models were calculated. The rst Cox model (only matched case and control patients) quantied the magnitude of effect for cruciate ligament reconstruction as a risk factor for knee arthroplasty. The second Cox model (all case patients) identied factors that inuenced knee arthroplasty risk following cruciate ligament reconstruction. Unmatched case patients were included in the second Cox model to minimize bias from their exclusion. Reference values for either Cox model were the median (reference age, twenty-ve to twenty-nine years; reference quintile, third) or most common categories (reference sex, male; reference comorbidity score, CADG 0 to 4; reference hospital status, non-university afliated). The reference surgeon volume category (more than 100 cruciate ligament reconstructions per year) was selected on the basis of evidence in the orthopaedic literature that

Downloaded From: http://jbjs.org/ by a CURTIN UNIV OF TECHNOLOGY User on 01/27/2014

6
TH E JO U R NA L O F B O N E & JO I N T SU RG E RY J B J S . O RG V O L U M E 96-A N U M B E R 1 J A N UA RY 1, 2 014
d d d

T H E R I S K O F K N E E A R T H R O P L A S T Y F O L L O W I N G C R U C I AT E LIG AMENT RECONSTRUCTION

correlates superior outcomes to higher surgeon volume . The Cox models generated hazard ratios (HRs) with 95% condence intervals (95% CIs). All statistical analyses were performed with use of SAS version 9.1 for UNIX (SAS Institute, Cary, North Carolina) and signicance was set at p = 0.05.

9-11

Source of Funding
The entire study was funded by an internal source through the Arthritis Research Unit at the Toronto Western Hospital (University Health Network) in Toronto, Ontario, Canada. Four authors of this study (D.O.-H., J.C., R.G., and N.M.) are employees at Toronto Western Hospital and afliates of the Arthritis Research Unit. Funds were used to pay for administration fees, analyst salaries, and data storage fees at the Institute for Clinical Evaluative Sciences.

Results e identied 34,786 patients who underwent a cruciate ligament reconstruction from July 1993 to March 2008. Following application of the exclusion criteria, 30,301 patients were eligible for study enrollment (see Appendix). The median patient age was twenty-eight years (interquartile range [IQR], twenty to thirty-six years), and the distribution of patient age was signicantly skewed to a younger age (p < 0.001) (see Appendix). Moreover, 65% of case patients were male (p < 0.001),

89% lived at an urban address (p < 0.001), and the great majority (99.9%) had few medical comorbidities (CADG, 0 to 4) (p < 0.001). Notably, the number of cruciate ligament reconstructions performed annually increased signicantly from 1626 procedures in 1994 to 2799 procedures in 2007 (p < 0.001) (see Appendix). We matched 151,362 individuals from the general population to 30,277 patients from the case cohort; there were twenty-four unmatched case patients. The twenty-four unmatched case patients were compared with the 30,277 matched case patients. There was a signicant difference in age (p = 0.04) between unmatched case patients (median, thirty-three years [IQR, twenty-seven to forty years]) and matched case patients (median, twenty-eight years [IQR, twenty to thirty-six years]). There was also a signicant difference (p < 0.001) in comorbidity score (CADG, 5 points) between unmatched case patients (95.8%) and matched case patients (0%). In addition, there was a signicant difference (p < 0.001) in rural address between unmatched case patients (33.3%) and matched case patients (10.7%). There were no signicant differences with respect to the distribution of sex or income quintile.

Fig. 1

Kaplan-Meier (KM) survival curve of knee arthroplasty: case versus control cohorts. The y axis depicts survival, specically the fraction of patients who have yet to be censored from the study (knee arthroplasty, death, emigration, or study end), and the x axis depicts the number of years since the index cruciate ligament reconstruction (case cohort) or study enrollment (control cohort). The difference in the event-free survival rate between case and control patients was signicant (p < 0.001).

Downloaded From: http://jbjs.org/ by a CURTIN UNIV OF TECHNOLOGY User on 01/27/2014

7
TH E JO U R NA L O F B O N E & JO I N T SU RG E RY J B J S . O RG V O L U M E 96-A N U M B E R 1 J A N UA RY 1, 2 014
d d d

T H E R I S K O F K N E E A R T H R O P L A S T Y F O L L O W I N G C R U C I AT E LIG AMENT RECONSTRUCTION

During the follow-up period, the matched case cohort had 209 knee arthroplasty events, 181 (86.6%) of which were total knee arthroplasties, and the control cohort had 125 knee arthroplasty events, 111 (88.8%) of which were total knee arthroplasties. The overall knee arthroplasty event rate per 1000 person-years was 0.68 (95% CI, 0.59 to 0.78) for the matched case cohort and 0.10 (95% CI, 0.08 to 0.12) for the control cohort (p < 0.001). According to the rst Cox model, patients who underwent cruciate ligament reconstruction had signicantly increased odds (p < 0.001) of undergoing knee arthroplasty at a later date (HR, 7.26 [95% CI, 5.79 to 9.11]). Moreover, the cumulative incidence of knee arthroplasty at two, ve, ten, and fteen years was greater among case patients as compared with control patients (Table I), and the log-rank test revealed a signicant difference in the case and control Kaplan-Meier survival curves (p < 0.001) (Fig. 1). Among the case cohort, the median time to knee arthroplasty following cruciate ligament reconstruction was 11.0 years (IQR, six to fourteen years). Among all 30,301 patients who had undergone cruciate ligament reconstruction, there were 212 knee arthroplasty events. Among those case patients who underwent knee arthroplasty, the median age at the time of cruciate ligament reconstruction was forty-two years (IQR, thirty-ve to forty-seven years), and the median age of case patients who did not undergo knee arthroplasty was twenty-eight years (IQR, twenty to thirty-ve years) (p < 0.001). There were also signicant differences in sex, comorbidity score, surgeon volume, and year of the index event (Table II). The second Cox model (all case patients) revealed that older patient age, female sex, higher comorbidity score, low annual surgeon volume, and cruciate ligament reconstruction performed at a university-afliated hospital increased knee arthroplasty risk (Table III). Neither meniscal repair or meniscal debridement performed concurrently with the index event increased the risk of knee arthroplasty following cruciate ligament reconstruction in this cohort (Table III).

Discussion he cumulative incidence of knee arthroplasty at fteen years following cruciate ligament reconstruction was low (1.4%); however, it was sevenfold greater than the cumulative incidence of knee arthroplasty among a matched control cohort from the general population. This is an important nding that is not only based upon the largest data set of its kind to date, but is also reective of general orthopaedic practice. A notable limitation of past studies has been the use of radiographic classication systems to diagnose osteoarthritis. In the available literature, seven different classication methods have been used in either isolation or combination1,12-16. A major downfall of this approach is that radiographic evidence of osteoarthritis may not reect patient symptomatology, nor does it dictate surgical treatment. The main outcome in this study was knee arthroplasty, a reasonable surrogate for both symptomatic and radiographic end-stage osteoarthritis. Certainly, our nding that patients who have undergone cruciate ligament reconstruction are at

an increased risk of arthroplasty as compared with the general population is interesting; however, we found that the cumulative incidence of knee arthroplasty among patients who had undergone cruciate ligament reconstruction was in keeping with a recent systematic review suggesting that osteoarthritis prevalence among patients who had undergone cruciate ligament reconstruction was low1. Unfortunately, heterogeneity in past studies, particularly osteoarthritis diagnosis, precluded our attempts to draw any further reliable comparisons to the literature. In this study, older age, female sex, higher comorbidity score, low surgeon volume (cruciate ligament reconstruction), cruciate ligament reconstruction undertaken in universityafliated hospitals, and an earlier cruciate ligament reconstruction year increased the likelihood of knee arthroplasty following cruciate ligament reconstruction. However, young age and male sex were associated with a decreased likelihood of arthroplasty. Several of these ndings are consistent with previously identied risk factors, including age17-22 and sex17,23. Not surprisingly, patients who underwent cruciate ligament reconstruction earlier in the study were at increased risk of undergoing arthroplasty. Unfortunately, cruciate ligament reconstruction technique could not be evaluated in this study. The nding that provider factors such as surgeon volume and hospital status inuence osteoarthritis risk following cruciate ligament reconstruction is novel. The relationship between low surgeon volume and poor outcomes has been well documented in elective orthopaedic surgery practice9-11. With respect to cruciate ligament reconstruction, a 2009 population study failed to correlate reoperation following anterior cruciate ligament reconstruction with low surgeon volume, and the authors postulated that this may reect the propensity for higher-volume surgeons to take on more complex cases24. In contrast, we demonstrated that cruciate ligament reconstruction performed by high-volume surgeons translates into a lower risk of eventual knee arthroplasty, and we believe that this nding reects an underlying relationship between experience and technical prociency. The observation that osteoarthritis risk is elevated following cruciate ligament reconstruction performed in a university-afliated hospital is more difcult to interpret. It may be that this relationship reects a difference in patient practice and a tendency for university-afliated orthopaedic surgeons to take on more complex cases, but the potential negative inuence of the surgical trainee on cruciate ligament reconstruction outcomes, specically osteoarthritis, must also be considered. In fact, current evidence suggests that the learning curve for arthroscopy is considerable25, and most trainees are more comfortable with open rather than arthroscopic procedures at graduation26. The relationship between meniscal pathology and osteoarthritis development has been extensively studied27. In the present study, meniscal repair did not inuence knee arthroplasty risk, a nding consistent with the current literature18,28,29. However, many studies have demonstrated that meniscectomy performed with cruciate ligament reconstruction is a risk factor for the development of osteoarthritis18-20,22,23,28-40, including a

Downloaded From: http://jbjs.org/ by a CURTIN UNIV OF TECHNOLOGY User on 01/27/2014

8
TH E JO U R NA L O F B O N E & JO I N T SU RG E RY J B J S . O RG V O L U M E 96-A N U M B E R 1 J A N UA RY 1, 2 014
d d d

T H E R I S K O F K N E E A R T H R O P L A S T Y F O L L O W I N G C R U C I AT E LIG AMENT RECONSTRUCTION

systematic review that found the prevalence of osteoarthritis following meniscectomy and cruciate ligament reconstruction to be 21% to 48%1. Our data do not support a relationship between cruciate ligament reconstruction with meniscectomy and an increased long-term arthroplasty risk. The discrepancy between past and present data could be explained by a difference in osteoarthritis diagnosis, small sample sizes in previous studies, or a variation in the amount of meniscus resected. In fact, osteoarthritis risk following total meniscectomy is increased as compared with partial meniscectomy41, and a notable limitation of past studies (including our own) is the failure or inability to report meniscectomy details. There were several limitations to this study. First, cruciate ligament reconstruction as we have dened it includes primary anterior cruciate ligament reconstruction, primary posterior cruciate ligament reconstruction, and some revision procedures. Prior to 2002, hospital admission databases did not allow us to distinguish between anterior cruciate ligament reconstruction and posterior cruciate ligament reconstruction. However, we do know that posterior cruciate ligament reconstruction represented only ;0.7% of all cruciate ligament reconstructions in Ontario since 200242, a gure consistent with other population studies24. Similarly, a revision cruciate ligament reconstruction OHIP fee code was not available prior to July 1, 2003. In our cohort of patients, we identied 312 revision cruciate ligament reconstruction cases (2.8% of all cases since 2003) (see Appendix). This nding is similar to another large population study with similar inclusion parameters43. Despite the fact that primary posterior cruciate ligament reconstruction and revision cruciate ligament reconstruction represent <2% of all cruciate ligament reconstruction cases included in this study, their potential inuence on the study ndings is unknown. A potential resolution to this limitation would be the inclusion of only anterior cruciate ligament reconstruction cases after 2003; however, this would have lessened the follow-up period, the sample size, and our ability to detect what we have shown to be a signicant, yet uncommon, event. Another limitation of this study was the unknown inuence of the cruciate ligament reconstruction technique. We demonstrated that patients managed early in the study had a higher incidence of subsequent arthroplasty, and the potential inuence of previously used techniques cannot be ignored. The cruciate ligament reconstruction technique has evolved since 1993, and it is possible that modern techniques could diminish osteoarthritis risk. Unfortunately, a comparison between outdated and modern cruciate ligament reconstruction techniques was beyond the scope of this study. The inuence of chondral injury on the development of osteoarthritis following cruciate ligament reconstruction is an important consideration. Although we could not directly study this relationship, given the limitations of our databases, we did attempt to exclude patients with a history of surgery that would potentially increase osteoarthritis risk, such as chondral or meniscal surgery. Despite this, it remains possible that both case and control patients who underwent knee arthroplasty had increased chondral pathology and/or preexisting osteoarthritis

at the time of study inclusion, and that the difference in native knee survivorship between the case and control cohorts can be accounted for by a difference in the rate of chondral pathology and/or preexisting osteoarthritis at the time of study inclusion. The utilization of knee arthroplasty in the treatment of end-stage osteoarthritis can be inuenced by social, demographic, medical, and geographic factors. Despite similar disease severity, studies have shown that older44, male45 patients from a higher socioeconomic status46,47 are more likely to undergo total knee arthroplasty. In Canada, health-care resources can also inuence total knee arthroplasty utilization rates, as exemplied by the dramatic increase in total knee arthroplasty volumes that followed the Ontario Wait Times Strategy initiative in 200448. In this study, we controlled for utilization factors by developing a comparison matched cohort; however, patient age remains an important inuence on our study ndings. Specically, surgeons are less inclined to perform knee arthroplasty in younger patients, and given that patients in our study were younger than the typical patient undergoing arthroplasty, we may have underestimated the burden of disease. Furthermore, our follow-up period was limited to fteen years, and although we found older patients to be at increased knee arthroplasty risk within this time period, a considerably longer follow-up is needed to determine the overall knee arthroplasty risk among younger patients undergoing cruciate ligament reconstruction. Nevertheless, the use of a matched cohort diminishes intra-cohort utilization factor inuence on eventual knee arthroplasty, and the best interpretation of our ndings is in the comparison between the case and control cohorts. The development of osteoarthritis following cruciate ligament injury is asymmetric49. In this study, unicompartmental knee arthroplasty was included in the main outcome to account for this possibility, although we have demonstrated that it is a relatively uncommon procedure as compared with total knee arthroplasty. In fact, >85% of all knee arthroplasty procedures identied in this study were total knee arthroplasty, and there was no difference in the distribution of total knee arthroplasty and unicompartmental knee arthroplasty events in both the case and control cohorts. Two additional limitations were inherent to this type of study. First, it was possible that some surgical cohort or control cohort patients elected to undergo arthroplasty in a jurisdiction outside of Ontario. However, patients would have paid for this surgery out-of-pocket, and we presume that this is an infrequent occurrence. Second, determination of the involved side was not possible in this study, and the study ndings reect contributions from the ipsilateral involved knee and the contralateral knee. A careful matched cohort study design and the exclusion of patients with prior cruciate ligament reconstruction or knee arthroplasty minimized this inuence; however, the baseline rate of knee arthroplasty in the cruciate ligament reconstruction cohort may be inated secondary to contributions from the contralateral side. The comparative risk between case patients and control patients is the most accurate interpretation of our ndings, in which the patients who have

Downloaded From: http://jbjs.org/ by a CURTIN UNIV OF TECHNOLOGY User on 01/27/2014

9
TH E JO U R NA L O F B O N E & JO I N T SU RG E RY J B J S . O RG V O L U M E 96-A N U M B E R 1 J A N UA RY 1, 2 014
d d d

T H E R I S K O F K N E E A R T H R O P L A S T Y F O L L O W I N G C R U C I AT E LIG AMENT RECONSTRUCTION

undergone cruciate ligament reconstruction are seven times more likely to undergo a knee arthroplasty after fteen years. Lastly, the intent of this study was not to uncover the etiology of osteoarthritis following cruciate ligament reconstruction. Rather, the sole intent was to determine knee arthroplasty risk following cruciate ligament reconstruction. Certainly, the initial cruciate ligament injury cannot be overlooked as a causative agent in the development of osteoarthritis following cruciate ligament reconstruction, including chondral injury, and we recognize the importance of drawing comparisons to patients who are cruciate ligament-decient, but do not undergo reconstruction to make such claims. Unfortunately, this was beyond the scope of our study. In conclusion, after fteen years, the cumulative incidence of knee arthroplasty following cruciate ligament reconstruction was low (1.4%); however, it was seven times greater than the cumulative incidence of knee arthroplasty among a cohort of matched control patients from the general population. Factors that increase risk include older age, female sex, higher comorbidity, low surgeon volumes, and treatment at university-afliated hospitals. Meniscal surgery, including debridement and repair, did not increase arthroplasty risk. Appendix Figures showing a bar graph of the age distribution of case cohort patients and a line graph showing the annual vol-

ume of cruciate ligament reconstruction procedures performed in Ontario, Canada, during each non-fragmented study year and tables showing cohort development and administrative codes are available with the online version of this article as a data supplement at jbjs.org. n

Timothy Leroux, MD, MEd Darrell Ogilvie-Harris, MBBS, FRCSC Tim Dwyer, MD, FRCSC, FRACS Jaskarndip Chahal, MD, MSc, FRCSC University of Toronto Orthopaedic Sports Medicine, 149 College Street, Room 508-A, Toronto, ON M5T 1P5, Canada. E-mail address for T. Leroux: timothy.leroux@mail.utoronto.ca Rajiv Gandhi, MD, MSc, FRCSC Nizar Mahomed, MD, ScD, FRCSC Arthritis Research Unit, Toronto Western Hospital (University Health Network), 399 Bathurst Street, Suite 1E-435, Toronto, ON M5T 2S8, Canada David Wasserstein, MD, MSc, FRCSC Sunnybrook Health Sciences Center, 2075 Bayview Avenue, Room MG 301, Toronto, ON M4Y 1H1, Canada. E-mail address: david.wasserstein@mail.utoronto.ca

References
1. iestad BE, Engebretsen L, Storheim K, Risberg MA. Knee osteoarthritis after anterior cruciate ligament injury: a systematic review. Am J Sports Med. 2009 Jul;37(7):1434-43. 2. Coleman BD, Khan KM, Maffulli N, Cook JL, Wark JD; Victorian Institute of Sport Tendon Study Group. Studies of surgical outcome after patellar tendinopathy: clinical signicance of methodological deciencies and guidelines for future studies. Scand J Med Sci Sports. 2000 Feb;10(1):2-11. 3. Cram P, Lu X, Kates SL, Singh JA, Li Y, Wolf BR. Total knee arthroplasty volume, utilization, and outcomes among Medicare beneciaries, 1991-2010. JAMA. 2012 Sep 26;308(12):1227-36. 4. Williams JI, Young W. Patterns of health care in Ontario: In: Goel V, Williams JI, Anserson GM, Blackstein-Hirsch P, Fooks C, Naylor CD, editors. The ICES Practice Atlas. 2nd ed. The Canadian Medical Association; 1996. p. 339-46. 5. Borugian MJ, Spinelli JJ, Mezei G, Wilkins R, Abanto Z, McBride ML. Childhood leukemia and socioeconomic status in Canada. Epidemiology. 2005 Jul;16(4): 526-31. 6. The John Hopkins University Bloomberg School of Public Health. The Johns Hopkins ACG Case-Mix System Version 6.0 Release Notes. 2003. http://s3. amazonaws.com/zanran_storage/www.acg.jhsph.edu/ContentPages/45990640. pdf. Accessed 2013 Oct 21. 7. Baldwin LM, Klabunde CN, Green P, Barlow W, Wright G. In search of the perfect comorbidity measure for use with administrative claims data: does it exist? Med Care. 2006 Aug;44(8):745-53. 8. Wasserstein D, Dwyer T, Gandhi R, Austin PC, Mahomed N, Ogilvie-Harris D. A matched-cohort population study of reoperation after meniscal repair with and without concomitant anterior cruciate ligament reconstruction. Am J Sports Med. 2013 Feb;41(2):349-55. Epub 2012 Dec 20. 9. Hervey SL, Purves HR, Guller U, Toth AP, Vail TP, Pietrobon R. Provider Volume of Total Knee Arthroplasties and Patient Outcomes in the HCUP-Nationwide Inpatient Sample. J Bone Joint Surg Am. 2003 Sep;85(9):1775-83. 10. Katz JN, Losina E, Barrett J, Phillips CB, Mahomed NN, Lew RA, Guadagnoli E, Harris WH, Poss R, Baron JA. Association between hospital and surgeon procedure volume and outcomes of total hip replacement in the United States Medicare population. J Bone Joint Surg Am. 2001 Nov;83(11):1622-9. 11. Jain N, Pietrobon R, Hocker S, Guller U, Shankar A, Higgins LD. The relationship between surgeon and hospital volume and outcomes for shoulder arthroplasty. J Bone Joint Surg Am. 2004 Mar;86(3):496-505. 12. Fairbank TJ. Knee joint changes after meniscectomy. J Bone Joint Surg Br. 1948 Nov;30B(4):664-70. 13. Ahlb ack S. Osteoarthrosis of the knee. A radiographic investigation. Acta Radiol Diagn (Stockh). 1968;Suppl 277:7-72. 14. Kellgren JH, Lawrence JS. Radiological assessment of osteo-arthrosis. Ann Rheum Dis. 1957 Dec;16(4):494-502. 15. Altman RD, Hochberg M, Murphy WA Jr, Wolfe F, Lequesne M. Atlas of individual radiographic features in osteoarthritis. Osteoarthritis Cartilage. 1995 Sep;3(Suppl A):3-70. 16. Irrgang JJ, Ho H, Harner CD, Fu FH. Use of the International Knee Documentation Committee guidelines to assess outcome following anterior cruciate ligament reconstruction. Knee Surg Sports Traumatol Arthrosc. 1998;6(2):107-14. 17. Asano H, Muneta T, Ikeda H, Yagishita K, Kurihara Y, Sekiya I. Arthroscopic evaluation of the articular cartilage after anterior cruciate ligament reconstruction: a short-term prospective study of 105 patients. Arthroscopy. 2004 May;20(5):474-81. 18. Brophy RH, Wright RW, David TS, McCormack RG, Sekiya JK, Svoboda SJ, Huston LJ, Haas AK, Steger-May K; Multicenter ACL Revision Study (MARS) Group. Association between previous meniscal surgery and the incidence of chondral lesions at revision anterior cruciate ligament reconstruction. Am J Sports Med. 2012 Apr;40(4):808-14. Epub 2012 Feb 28. 19. Keays SL, Newcombe PA, Bullock-Saxton JE, Bullock MI, Keays AC. Factors involved in the development of osteoarthritis after anterior cruciate ligament surgery. Am J Sports Med. 2010 Mar;38(3):455-63. Epub 2010 Jan 5. 20. Lebel B, Hulet C, Galaud B, Burdin G, Locker B, Vielpeau C. Arthroscopic reconstruction of the anterior cruciate ligament using bone-patellar tendon-bone autograft: a minimum 10-year follow-up. Am J Sports Med. 2008 Jul;36(7):1275-82. Epub 2008 Mar 19. 21. iestad BE, Holm I, Engebretsen L, Aune AK, Gunderson R, Risberg MA. The prevalence of patellofemoral osteoarthritis 12 years after anterior cruciate ligament reconstruction. Knee Surg Sports Traumatol Arthrosc. 2013 Apr;21(4):942-9. Epub 2012 Aug 17.

Downloaded From: http://jbjs.org/ by a CURTIN UNIV OF TECHNOLOGY User on 01/27/2014

10
TH E JO U R NA L O F B O N E & JO I N T SU RG E RY J B J S . O RG V O L U M E 96-A N U M B E R 1 J A N UA RY 1, 2 014
d d d

T H E R I S K O F K N E E A R T H R O P L A S T Y F O L L O W I N G C R U C I AT E LIG AMENT RECONSTRUCTION

22. Seon JK, Song EK, Park SJ. Osteoarthritis after anterior cruciate ligament reconstruction using a patellar tendon autograft. Int Orthop. 2006 Apr;30(2):94-8. Epub 2006 Jan 25. 23. Li RT, Lorenz S, Xu Y, Harner CD, Fu FH, Irrgang JJ. Predictors of radiographic knee osteoarthritis after anterior cruciate ligament reconstruction. Am J Sports Med. 2011 Dec;39(12):2595-603. Epub 2011 Oct 21. 24. Lyman S, Koulouvaris P, Sherman S, Do H, Mandl LA, Marx RG. Epidemiology of anterior cruciate ligament reconstruction: trends, readmissions, and subsequent knee surgery. J Bone Joint Surg Am. 2009 Oct;91(10):2321-8. 25. Guttmann D, Graham RD, MacLennan MJ, Lubowitz JH. Arthroscopic rotator cuff repair: the learning curve. Arthroscopy. 2005 Apr;21(4):394-400. 26. Hall MP, Kaplan KM, Gorczynski CT, Zuckerman JD, Rosen JE. Assessment of arthroscopic training in U.S. orthopedic surgery residency programsa resident selfassessment. Bull NYU Hosp Jt Dis. 2010;68(1):5-10. 27. Englund M, Roemer FW, Hayashi D, Crema MD, Guermazi A. Meniscus pathology, osteoarthritis and the treatment controversy. Nat Rev Rheumatol. 2012 Jul;8(7):412-9. Epub 2012 May 22. 28. Magnussen RA, Mansour AA, Carey JL, Spindler KP. Meniscus status at anterior cruciate ligament reconstruction associated with radiographic signs of osteoarthritis at 5- to 10-year follow-up: a systematic review. J Knee Surg. 2009 Oct;22(4):347-57. 29. Wu WH, Hackett T, Richmond JC. Effects of meniscal and articular surface status on knee stability, function, and symptoms after anterior cruciate ligament reconstruction: a long-term prospective study. Am J Sports Med. 2002 Nov-Dec;30(6):845-50. 30. A t Si Selmi T, Fithian D, Neyret P. The evolution of osteoarthritis in 103 patients with ACL reconstruction at 17 years follow-up. Knee. 2006 Oct;13(5):353-8. Epub 2006 Aug 28. 31. Borchers JR, Kaeding CC, Pedroza AD, Huston LJ, Spindler KP, Wright RW; MOON Consortium and the MARS Group. Intra-articular ndings in primary and revision anterior cruciate ligament reconstruction surgery: a comparison of the MOON and MARS study groups. Am J Sports Med. 2011 Sep;39(9):1889-93. Epub 2011 Jun 6. 32. Cohen M, Amaro JT, Ejnisman B, Carvalho RT, Nakano KK, Peccin MS, Teixeira R, Laurino CF, Abdalla RJ. Anterior cruciate ligament reconstruction after 10 to 15 years: association between meniscectomy and osteoarthrosis. Arthroscopy. 2007 Jun;23(6):629-34. 33. Ferretti A, Conteduca F, De Carli A, Fontana M, Mariani PP. Osteoarthritis of the knee after ACL reconstruction. Int Orthop. 1991;15(4):367-71. 34. Friederich NF, OBrien WR. [Gonarthrosis after injury of the anterior cruciate ligament: a multicenter, long-term study]. Z Unfallchir Versicherungsmed. 1993;86(2):81-9. German. 35. Hertel P, Behrend H, Cierpinski T, Musahl V, Widjaja G. ACL reconstruction using bone-patellar tendon-bone press-t xation: 10-year clinical results. Knee Surg Sports Traumatol Arthrosc. 2005 May;13(4):248-55. Epub 2005 Feb 3. 36. Jomha NM, Borton DC, Clingeleffer AJ, Pinczewski LA. Long-term osteoarthritic changes in anterior cruciate ligament reconstructed knees. Clin Orthop Relat Res. 1999 Jan;(358):188-93.

n M, Sernert N, Rostgrd-Christensen L, Kartus C, Ejerhed L. Osteoar37. Lide thritic changes after anterior cruciate ligament reconstruction using bonepatellar tendon-bone or hamstring tendon autografts: a retrospective, 7-year radiographic and clinical follow-up study. Arthroscopy. 2008 Aug;24(8):899908. Epub 2008 May 19. 38. Louboutin H, Debarge R, Richou J, Selmi TA, Donell ST, Neyret P, Dubrana F. Osteoarthritis in patients with anterior cruciate ligament rupture: a review of risk factors. Knee. 2009 Aug;16(4):239-44. Epub 2008 Dec 20. 39. Neuman P, Englund M, Kostogiannis I, Frid en T, Roos H, Dahlberg LE. Prevalence of tibiofemoral osteoarthritis 15 years after nonoperative treatment of anterior cruciate ligament injury: a prospective cohort study. Am J Sports Med. 2008 Sep;36(9):1717-25. Epub 2008 May 15. 40. Salmon LJ, Russell VJ, Refshauge K, Kader D, Connolly C, Linklater J, Pinczewski LA. Long-term outcome of endoscopic anterior cruciate ligament reconstruction with patellar tendon autograft: minimum 13-year review. Am J Sports Med. 2006 May;34(5):721-32. Epub 2006 Jan 6. 41. Papalia R, Del Buono A, Osti L, Denaro V, Maffulli N. Meniscectomy as a risk factor for knee osteoarthritis: a systematic review. Br Med Bull. 2011;99:89-106. Epub 2011 Jan 19. 42. Wasserstein D, Mohamed N, Gandhi R, Ogilvie-Harris D. Predictors of complications and re-operation after knee cruciate ligament reconstruction in Ontario 1992-2008. In: Proceedings of the Annual Meeting of the American Orthopaedic Society for Sports Medicine; 2011 Jul 7-10; San Diego, CA. 43. Lind M, Menhert F, Pedersen AB. Incidence and outcome after revision anterior cruciate ligament reconstruction: results from the Danish registry for knee ligament reconstruction. Am J Sports Med. 2012 Jul;40(7):1551-7. 44. Dunlop DD, Manheim LM, Song J, Sohn MW, Feinglass JM, Chang HJ, Chang RW. Age and racial/ethnic disparities in arthritis-related hip and knee surgeries. Med Care. 2008 Feb;46(2):200-8. 45. Hawker GA, Wright JG, Coyte PC, Williams JI, Harvey B, Glazier R, Badley EM. Differences between men and women in the rate of use of hip and knee arthroplasty. N Engl J Med. 2000 Apr 6;342(14):1016-22. 46. Hawker GA, Guan J, Croxford R, Coyte PC, Glazier RH, Harvey BJ, Wright JG, Williams JI, Badley EM. A prospective population-based study of the predictors of undergoing total joint arthroplasty. Arthritis Rheum. 2006 Oct;54(10): 3212-20. 47. Hawker GA, Wright JG, Glazier RH, Coyte PC, Harvey B, Williams JI, Badley EM. The effect of education and income on need and willingness to undergo total joint arthroplasty. Arthritis Rheum. 2002 Dec;46(12):3331-9. 48. Trypuc J, Hudson A, MacLeod H. The pivotal role of critical care and surgical efciencies in supporting Ontarios Wait Time Strategy: Part 3. Healthc Q. 2006;9(4):37-45: 2. 49. Stein V, Li L, Lo G, Guermazi A, Zhang Y, Kent Kwoh C, Eaton CB, Hunter DJ. Pattern of joint damage in persons with knee osteoarthritis and concomitant ACL tears. Rheumatol Int. 2012 May;32(5):1197-208. Epub 2011 Jan 21.

Downloaded From: http://jbjs.org/ by a CURTIN UNIV OF TECHNOLOGY User on 01/27/2014