To identify risk factors for knee osteoarthritis (OA) 10–15 years after anterior cruciate ligament (ACL) reconstruction. We hypothesized that quadriceps muscle weakness after ACL reconstruction would be a risk factor for radiographic and symptomatic radiographic knee OA 10–15 years later.
Subjects with ACL reconstruction (n = 258) were followed for 10–15 years. Subjects with unilateral injury at the 10–15-year followup were included in the present study. Outcomes included the Cincinnati knee score, knee joint laxity, hop performance, and isokinetic muscle strength tests at 6 months, 1 year, and 2 years postoperatively. At the 10–15-year followup, radiographs were taken and graded according to the Kellgren/Lawrence classification (range 0–4).
Of the 212 subjects (82%) assessed at the 10–15-year followup, 164 subjects had unilateral injury. The mean ± SD age at ACL reconstruction was 27.4 ± 8.5 years. Increased age (odds ratio [OR] 1.06, 95% confidence interval [95% CI] 1.01–1.11) and meniscal injury and/or chondral lesion (OR 2.05, 95% CI 1.00–4.20) showed significantly higher odds for radiographic knee OA. Low self-reported knee function 2 years postoperatively (OR 0.95, 95% CI 0.92–0.98) and loss of quadriceps strength between the 2-year and the 10–15-year followup (OR 1.00, 95% CI 1.00–1.01) showed significantly higher odds for symptomatic radiographic knee OA. Quadriceps muscle weakness after ACL reconstruction was not significantly associated with knee OA.
This study detected no association between quadriceps weakness after ACL reconstruction and knee OA as measured 10–15 years later.
Knee injuries, including anterior cruciate ligament (ACL) injuries and meniscal injuries, have been shown to be some of the most important risk factors for the development of knee osteoarthritis (OA) (1, 2). However, the causation from the healthy cartilage and bone structures in the knee joint before the injury to the development of knee OA after the injury is still not fully understood. Studies have shown that damage to the cartilage at the time of the injury may initiate disruption of the cartilage matrix, changes in cell metabolism, and also death of chondrocytes (3). The development of knee OA following a knee injury may be influenced by mechanical components such as altered joint loading due to reduced mechanical stability, malalignment, or reduced shock absorption (3, 4). Several risk factors for the development of tibiofemoral knee OA have been identified in subjects with ACL injuries, but few studies have examined the association between early impaired knee function and knee OA in long-term followup studies (5). Such factors may be important to identify in the early phase after ACL reconstruction to further be able to prevent the onset of knee OA. Meniscal injury and subsequently partial meniscal resection have been shown to be important risk factors for knee OA (6). Quadriceps muscle weakness, which is often seen after ACL injuries (7), has been shown to increase the knee joint loading patterns with a reduced ability of shock absorption, and thereby has been suggested as a significant risk factor for the development of knee OA (8–11). To our knowledge, no prospective studies with more than 10 years of followup after ACL reconstruction have investigated quadriceps muscle weakness as a potential risk factor for tibiofemoral OA. Furthermore, prospective long-term studies aiming at detecting risk factors for symptomatic radiographic knee OA in subjects with ACL injuries are lacking. The aim of the present study was therefore to identify risk factors associated with radiographic and symptomatic radiographic OA in the tibiofemoral joint 10–15 years after ACL reconstruction. We hypothesized that quadriceps muscle weakness after ACL reconstruction was a significant risk factor for radiographic and symptomatic radiographic tibiofemoral OA 10–15 years later.
SUBJECTS AND METHODS
The present cohort study involved 258 subjects with an ACL rupture. The subjects were consecutively included in the time period between 1990 and 1997 in 4 prospective cohort studies with the same inclusion and exclusion criteria (12–14). The inclusion criteria were: between ages 14 and 50 years, isolated ACL injury or combined with meniscal and/or medial collateral ligament (MCL) injury and/or chondral lesions, and candidates for ACL reconstruction with bone-patellar tendon-bone (BPTB) autograft or hamstrings tendon autograft. Subjects were excluded if they had experienced other major injuries to both the lower extremities less than 1 year before the surgery or if they had experienced ligament injuries in the contralateral knee. All of the included subjects went through a supervised rehabilitation program for 6 months postoperatively. The surgical procedures and the rehabilitation programs are described in previous studies (12–14). The included subjects have been prospectively followed and have been through clinical and functional examinations at 6 months, 1 year, 2 years, and 10–15 years after the ACL reconstruction. The prospective data on clinical and functional outcomes were collected by the same research team and in the same way for all of the study participants. On the basis of the similar inclusion and exclusion criteria in the 4 original studies, the materials have been considered to constitute one prospective cohort of subjects.
In the present study, subjects with known ACL or meniscal injuries in the contralateral knee experienced during the followup period were excluded to use the contralateral knee as a control knee for muscle strength tests, hop tests, and knee joint laxity tests. The additional injuries in the involved knee included MCL injuries and chondral lesions identified at the time of the reconstruction, and meniscal injuries experienced at the time of the ACL injury or during the followup period. Data of the meniscal injuries, the MCL injuries, and the chondral lesions were collected at the 10–15-year followup by reading surgical files of all of the included subjects from the index operation and for re-injuries experienced during the followup period.
The subjects were informed by written consent before participation at the 10–15-year followup evaluation. The study has been evaluated by the Regional Committees for Medical and Health Research Ethics in Eastern Norway.
Assessment of knee function.
Knee joint laxity was measured with the KT-1000 knee arthrometer (MEDmetric) at manual maximum force (15). The difference in displacement between the two knees was calculated and expressed in millimeters. The Cincinnati knee score was included to examine self-reported knee function (16). A score of 100 indicated normal knee function. This self-reported questionnaire has been validated for measuring knee function in ACL-injured subjects (17–19).
Evaluation of muscle strength, including knee extension (quadriceps strength) and knee flexion (hamstrings strength), was performed with the Cybex 6000 (Cybex). The isokinetic test protocol consisted of 5 concentric repetitions at 60°/second. Muscle strength performance was recorded as total work for all repetitions. The muscle strength values were presented in joules and joules normalized to body weight (%BW), calculated with the formula: [(joules/BW) × 100]. Isokinetic muscle strength measurement has been shown to be reliable (20), and has been widely used to measure muscle performance in subjects with ACL injury (21).
The triple jump test (recorded in meters) and the stair hop test (recorded in seconds) were included at the 6-month, 1-year, and 2-year followup periods (22, 23). Body mass index (BMI) was measured and calculated with the formula kg/m2.
Radiologic examination was performed only at the 10–15-year followup. The SynaFlexer frame (Synarc) for standardized fixed flexion positioning (20° knee flexion and 5° external foot rotation) was used for the radiograph procedure. This frame was validated for measurement of joint space width (24). The pictures were taken bilaterally from a posteroanterior view.
The Kellgren/Lawrence (K/L) classification system (25, 26) was used for assessing radiographic changes in the tibiofemoral joint. A K/L score of ≥2 was used to define radiographic knee OA according to previous literature (27). Radiographs were read by a single radiologist (RG). The clinical and functional results and the type of graft were unknown to the radiologist. An intrarater reliability test was performed by the radiologist on 35 radiographs (70 knees).
Symptomatic radiographic knee OA was defined on the basis of if the subjects answered positively to a question about whether or not they had experienced knee pain during the last 4 weeks, and in addition had a K/L score of ≥2. The question of knee pain was derived from a 2-step telephone interview to screen for symptomatic OA developed by Roux et al (28, 29). Knee pain has been shown to be the single symptom that associates most strongly with radiographic OA (28), and has been included in several studies to define symptomatic radiographic knee OA (30–32).
The Statistical Package for Social Sciences, version 16.0 (SPSS), was used for analyzing the data. Means and SDs were presented for descriptive statistics. The chi-square test was used for comparison of categorical variables. The Mann Whitney U test was performed for group comparisons of data that were not normally distributed. Binary logistic regression models with measurement of odds ratios (ORs) and 95% confidence intervals (95% CIs) were used to evaluate potential risk factors for radiographic and symptomatic radiographic knee OA. First, univariate analyses were performed for radiographic and symptomatic radiographic knee OA and potential risk factors that included age, sex, additional injury, graft type, time from injury to surgery, BMI, KT-1000 manual maximum tests (difference), and knee function variables at 6 months, 1 year, and 2 years postoperatively (the Cincinnati knee score, the triple jump test, the stair hop test, and the muscle strength tests). The variable “additional injury” constituted meniscal injuries and/or MCL injury and/or chondral lesion shown at the 10–15-year followup, and was dichotomized into “additional injury” or “isolated ACL injury.” Those variables that showed a P value of less than 0.20 in the univariate analyses were included in a second analysis with adjustment for age, sex, additional injury, and graft type. The final regression models included variables that were significantly associated with radiographic or symptomatic radiographic knee OA in the second step. A P value less than 0.05 was considered to be statistically significant.
Two hundred fifty-eight subjects in the 4 cohorts were included at the time of ACL reconstruction (12–14). The sex distributions in the 4 cohorts showed slightly more than 50% men than women (ranges 53–61% and 39–47%, respectively). The mean ± SD time between injury and surgery for the entire cohort was 25.4 ± 50 months. The original cohorts showed a mean ± SD time between injury and surgery of 27 ± 46 (the 2 first studies), 43 ± 64, and 9 ± 8 months, respectively. Approximately 50% had additional injuries in the 4 cohorts at the time of ACL reconstruction (range 45–60%).
Two hundred twelve subjects (82%) participated at the 10–15-year followup. Of these, 164 had an uninjured contralateral knee and were therefore included in the analyses in the present study (Figure 1). The mean ± SD age at ACL reconstruction was 27.4 ± 8.5 years (n = 164), the mean ± SD time between injury and surgery was 27.2 ± 53.0 months, and the mean ± SD time between the ACL reconstruction and the 10–15-year followup was 12.1 ± 1.4 years. Subject characteristics are described in Table 1.
Activities performed at the time of the ACL injury
Additional injury at the 10–15-year followup in the involved knee
Meniscus and MCL
Meniscus and chondral lesion
Eighty-two subjects (50%) had no additional injuries at the time of the ACL reconstruction. The additional injuries revealed at the time of ACL reconstruction included 36 medial meniscus injuries (22%), 24 lateral meniscus injuries (15%), 7 MCL and meniscus injuries (4%), 13 menisci injuries (8%), and 2 MCL injuries (1%). Twenty-seven subjects (16%) had chondral lesions at the time of the ACL reconstruction. At the 10–15-year followup, isolated ACL injury was shown in 70 subjects (43%) and 94 subjects (57%) had additional meniscal and/or MCL injury and/or chondral lesion. A total of 94 partial meniscal resections were performed in the 88 subjects with additional meniscal injury, either before the ACL reconstruction (18%), during the ACL reconstruction (55%), or during the followup period (27%). Furthermore, 9 meniscal tears were sutured before the ACL reconstruction (n = 3), at the ACL reconstruction (n = 4), or during followup (n = 2).
Radiographic knee OA (K/L grade ≥2) was detected in 113 subjects (69%) (Table 2). Of the 77 subjects that reported knee pain at the 10–15-year followup, 58 subjects (75%) had a K/L grade ≥2 and 19 subjects (25%) had no radiographic changes (Table 2).
Table 2. Frequencies (%) of the Kellgren/Lawrence grades and knee pain
Quadriceps weakness measured after the ACL reconstruction both in absolute values (joules) or absolute values normalized to BW (%BW) was not significantly associated with radiographic or symptomatic radiographic knee OA identified 10–15 years later (Tables 3 and 4). Furthermore, no other functional test results were significantly associated with radiographic knee OA (Table 3).
Table 3. Binary logistic regression analyses of the association between radiographic knee OA and potential risk factors*
Dependent variable: Kellgren/Lawrence grade of ≥2 (OA: n = 113, no OA: n = 51). The strength tests and the hop tests are given for the injured leg. OA = osteoarthritis; OR = odds ratio; 95% CI = 95% confidence interval; BW = body weight.
Adjusted for age, sex, additional injury, and graft type.
The reference categories are female, isolated anterior cruciate ligament injury, hamstrings tendon graft, and quadriceps strength ≤80%.
Dependent variable: symptomatic radiographic OA (OA: n = 58, no OA: n = 106). The strength results and the hop tests are given for the injured leg. OA = osteoarthritis; OR = odds ratio; 95% CI = 95% confidence interval; BW = body weight.
Adjusted for sex, age, additional injury, and graft type.
The reference categories are female, isolated anterior cruciate ligament injury, hamstrings tendon graft, and quadriceps strength ≤80%.
Low Cincinnati knee score at the 2-year followup and loss of quadriceps strength between 2 and 10–15 years were significantly associated with symptomatic radiographic knee OA, adjusted for age, sex, additional injury, and graft type (Table 4).
The final regression models for the risk factor analyses included variables that were significantly associated with radiographic or symptomatic radiographic knee OA, adjusted for age, sex, additional injury, and graft type (Table 5). Subjects with increased age at surgery (OR 1.06, 95% CI 1.01–1.11) and additional injury (OR 2.05, 95% CI 1.00–4.20) had significantly higher odds for radiographic knee OA. Furthermore, subjects with impaired self-reported knee function 2 years postoperatively had significantly higher odds for symptomatic radiographic knee OA (OR 0.95, 95% CI 0.92–0.98). The odds for symptomatic radiographic knee OA increased with 5% for each unit decrease in the Cincinnati knee score at 2 years after the ACL reconstruction. Also, loss of quadriceps strength between 2 and 10–15 years showed significantly higher odds for symptomatic radiographic knee OA (OR 1.00, 95% CI 1.00–1.01). Men tended to have higher odds for symptomatic radiographic knee OA compared with women (OR 2.19, 95% CI 1.00–4.80).
Table 5. The final logistic regression models, including potential risk factors for radiographic OA (n = 164) and symptomatic radiographic OA (n = 141), in subjects with anterior cruciate ligament reconstruction*
OR (95% CI)
OA = osteoarthritis; OR = odds ratio; 95% CI = 95% confidence interval.
Radiographic OA (n = 113)
Age at surgery, years
Symptomatic radiographic OA (n = 50)
Cincinnati knee score at 2 years
Loss of quadriceps strength, 2 to 10–15 years, joules
Previous studies have highlighted that quadriceps weakness may be a risk factor for development of knee OA (8, 9, 34). Our hypothesis that quadriceps weakness after ACL reconstruction was a risk factor for knee OA 10–15 years later was not confirmed. Risk factors associated with radiographic knee OA were increased age at the time of surgery and meniscal injury and/or chondral lesion. Furthermore, factors that were associated with symptomatic radiographic knee OA included self-reported knee function 2 years postoperatively and loss of quadriceps strength between the 2-year and the 10–15-year followup.
To our knowledge, this is the first prospective long-term followup study evaluating quadriceps weakness as a risk factor for knee OA in subjects who have undergone ACL reconstruction. However, a few population-based studies have evaluated the association between quadriceps weakness and knee OA: Slemenda et al (9) suggested that quadriceps weakness was a risk factor for radiographic knee OA in women, but not in men, in a study of elderly subjects with no known knee injuries. They found that subjects who developed knee OA 30 months later were 18% weaker at baseline than those who did not develop knee OA (P = 0.053). Our results revealed no significant differences in quadriceps strength values between those with radiographic knee OA compared with those without radiographic knee OA either 6 months (0%), 1 year (3%), or 2 years (5%) postoperatively. However, our cohort consisted of younger individuals with previous knee injuries, and the subjects in our study had gone through a rehabilitation program aiming at retaining muscle strength after the ACL reconstruction. Therefore, a comparison between the two studies cannot be performed. The study by Slemenda et al (9) is widely cited for supporting the fact that quadriceps weakness relative to body weight is a risk factor for the development of radiographic knee OA, but few studies have reproduced similar results (33, 34). Nevertheless, their cohort of subjects with knee OA after 31 months of followup consisted of only 13 subjects, in whom 7 had unilateral knee OA at baseline. In addition, the analyses did not include adjustment for potential confounding factors such as age or knee injuries experienced during the followup period.
Segal et al (35) studied the effect of thigh muscle strength on knee OA in subjects between ages 50 and 79 years (mean ± SD age 62 ± 8 years). They could not document that quadriceps weakness was a risk factor for radiographic knee OA 30 months later. However, they concluded that quadriceps weakness seemed to predict symptomatic radiographic knee OA. Their cohort included subjects with known risk factors for knee OA such as obesity or prior knee injuries. Currently, there is no evidence showing that quadriceps weakness as a single factor is a risk factor for the development of knee OA in subjects with ACL injury.
Based on recent studies (34, 35) and our study, quadriceps muscle weakness did not seem to be a risk factor for radiographic knee OA in different populations. Nevertheless, our results showed that subjects who lost quadriceps strength between 2 years and 10–15 years after the ACL reconstruction had higher odds for symptomatic radiographic knee OA. Because we had no radiographic data before the 10–15-year followup, we do not know what occurred first, the quadriceps weakness or the symptomatic radiographic knee OA. However, quadriceps weakness has been shown to correlate with knee pain (36); therefore, the association between the loss of quadriceps strength between 2 years and 10–15 years after the ACL reconstruction and symptomatic radiographic OA may be an association with the knee pain only. The fact that quadriceps weakness at 6 months, 1 year, and 2 years after ACL reconstruction was not associated with symptomatic radiographic OA may indicate that the loss of quadriceps strength during the long-term followup has been a consequence of the knee pain. Pain inhibition and thereby activation failure has been shown to reduce muscle function in subjects with knee OA (4). Therefore, the loss of quadriceps strength for subjects with symptomatic radiographic knee OA seen in our cohort may be due to inhibition triggered by pain. Abnormal muscle function influences the magnitude of the knee joint loading, and abnormal loading patterns during walking have been associated with the onset of knee OA (4). Normal muscle function, including muscle strength, activation patterns, and proprioceptive acuity, is a key factor to sustain the activity level and to reduce pain in all age groups (37), and to restore normal muscle function after ACL injuries should be one of the main long-term aims after ACL reconstruction.
Several risk factors have been associated with knee OA in subjects with previous ACL injury, such as meniscal injury, BPTB graft, chondral lesions, loss of knee extension, increased knee joint laxity, increased time between the injury and the surgery, and increased age at injury (6). Other factors that have been associated with knee OA include obesity, <90% performance on the single leg hop test compared with the uninjured side 1 year after surgery, high level of sports activity, OA of the contralateral knee, and time duration of followup (5, 11, 38, 39). Our results supported that meniscal injury alone or combined with chondral lesion and increased age was associated with radiographic knee OA, which is also in line with a newly published study by Keays et al (11). In the present study, most of the subjects with chondral lesions also had meniscal injuries; therefore, chondral lesions could not be studied as a separate risk factor. However, strong evidence that meniscal injuries and subsequently partial meniscal resections are risk factors for the development of knee OA following an ACL injury exists (6). Also, subjects with isolated meniscal injury have shown a high prevalence of knee OA (2). The menisci function as shock absorbers and transmit load in the knee joint during movement and static loading (40). Removal of parts of one or both of the menisci leads to altered loading on the cartilage, and consequently may initiate the onset of OA. More effort should therefore be put on prevention of meniscal injuries, but also treatment strategies, including less resection procedures, in order to sustain the role of the menisci after the injury. In our study, more than 95% of those with meniscal injuries were partially meniscectomized. Consequently, we were not able to evaluate the association between the type of meniscal treatment and knee OA.
The results in the present study revealed differences in the risk factors reported for subjects with previous knee injuries compared with older subjects with knee OA. For instance, obesity has been reported to be an important risk factor for knee OA (41). However, our cohort showed low mean BMI and may therefore not be comparable with population-based studies. Furthermore, women have shown to have a higher prevalence of knee OA than men, but in the present study, the men tended to show higher odds for symptomatic radiographic knee OA compared with the women. We have no good explanation for this difference, but the higher prevalence of knee OA seen for men may be due to unknown confounders such as malalignment, knee demanding occupations, or higher activity level. However, no differences were detected between the women and the men on the Tegner Activity Scale at the 10–15-year followup. BPTB graft has been significantly associated with radiographic knee OA in subjects with ACL reconstruction (5, 11). Our analysis detected a trend toward higher odds for radiographic knee OA for subjects with BPTB graft compared with those with hamstrings tendon graft (P = 0.07). Nevertheless, few subjects with hamstrings tendon graft were included in the study, which may have influenced the results.
The subjects with impaired self-reported knee function 2 years after ACL reconstruction had higher odds for symptomatic radiographic knee OA. Therefore, those with impaired knee function 2 years postoperatively seemed to be at risk for symptomatic radiographic knee OA. No radiologic assessment was included at the 2-year followup; therefore, it is difficult to state the onset of the radiographic changes in the tibiofemoral joint. Although 16% had chondral lesions at the time of injury, it is reasonable to believe that few subjects had radiographic knee OA 2 years postoperatively in subjects with a mean ± SD age of 27.4 ± 8.5 years (42).
This prospective study is the first to provide important knowledge on the association between quadriceps muscle weakness and knee OA in subjects with ACL reconstruction. However, the study did not include radiographic evaluation at all of the followup periods, resulting in a lack of information on the onset of radiographic knee OA. Therefore, no conclusion on causality between quadriceps muscle weakness and knee OA can be drawn. Furthermore, our results on the association between quadriceps weakness and knee OA cannot be generalized to subjects in the same age group without knee injuries. The study cohort revealed a high prevalence of mild radiographic knee OA and few subjects had severe radiographic knee OA. This may have influenced the results. Our study included several potential risk factors, but our study did not include data on activity level, malalignment, bone mineral density, biochemical markers, nutritional factors, or socioeconomic factors, which have previously been associated with knee OA (43, 44). Finally, a definition of symptomatic radiographic knee OA according to one question on knee pain during the last 4 weeks in addition to radiographic signs may have overestimated the amount of subjects with symptomatic radiographic knee OA because the knee pain could be caused by other factors unrelated to OA. Contrarily, subjects with radiographic knee OA may have been misclassified as not having symptomatic radiographic OA if the subjects had no pain during the last 4 weeks.
Increased age at the time of ACL reconstruction and meniscal injury and/or chondral lesion was a significant risk factor for radiographic knee OA. Subjects with impaired knee function at 2 years after ACL reconstruction had significantly higher odds for symptomatic radiographic knee OA 10 years later. Quadriceps muscle weakness after ACL reconstruction was not a risk factor for radiographic or symptomatic radiographic knee OA 10–15 years after ACL reconstruction.
All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Ms Øiestad had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study conception and design. Øiestad, Holm, Risberg.
Acquisition of data. Øiestad, Holm, Gunderson, Myklebust, Risberg.
Analysis and interpretation of data. Øiestad, Holm, Gunderson, Myklebust, Risberg.
We thank Leiv Sandvig, MD, PhD, for statistical advice and Lars Engebretsen, MD, PhD, Arne Kristian Aune, MD, PhD, and Merethe Aarsland Fosdahl, PT, MSc, for excellent help with the data collection.