Meniscal degeneration is prognostic of destabilzing meniscal tear and accelerated knee osteoarthritis: Data from the Osteoarthritis Initiative

The objective of this study was to assess the prognostic potential of magnetic resonance (MR)‐detected meniscal degeneration in relation to incident destabilizing meniscal tears (radial, complex, root, or macerated) or accelerated knee osteoarthritis (AKOA). We used existing MR data from a case‐control study of three groups from the Osteoarthritis Initiative without radiographic KOA at baseline: AKOA, typical KOA, and no KOA. From these groups, we included people without medial and lateral meniscal tear at baseline (n = 226) and 48‐month meniscal data (n = 221). Intermediate‐weighted fat‐suppressed MR images annually from baseline to the 48‐month visit were graded using a semiquantitative meniscal tear classification criterion. Incident destabilizing meniscal tear was defined as progressing from an intact meniscus to a destabilizing tear by the 48‐month visit. We used two logistic regression models to assess whether: (1) presence of medial meniscal degeneration was associated with an incident medial destabilizing meniscal tear, and (2) presence of meniscal degeneration in either meniscus was associated with incident AKOA over the next 4 years. People with the presence of a medial meniscal degeneration had three times the odds of developing an incident destabilizing medial meniscal tear within 4 years compared with a person without medial meniscus degeneration (odds ratio [OR]: 3.03; 95% confidence interval [CI]: 1.40–6.59). People with meniscal degeneration had five times the odds of developing incident AKOA within 4 years compared with a person without meniscal degeneration in either meniscus (OR: 5.04; 95% CI: 2.57–9.89). Meniscal degeneration on MR is clinically meaningful as it relates to future poor outcomes.

incident AKOA over the next 4 years.People with the presence of a medial meniscal degeneration had three times the odds of developing an incident destabilizing medial meniscal tear within 4 years compared with a person without medial meniscus degeneration (odds ratio [OR]: 3.03; 95% confidence interval [CI]: 1.40-6.59).
People with meniscal degeneration had five times the odds of developing incident AKOA within 4 years compared with a person without meniscal degeneration in either meniscus (OR: 5.04; 95% CI: 2.57-9.89).Meniscal degeneration on MR is clinically meaningful as it relates to future poor outcomes.

| BACKGROUND
The menisci play a crucial role in the health of the knee joint by distributing load and absorbing shock. 1 Meniscal pathology is a known risk factor for knee osteoarthritis (KOA) incidence and progression. 1Meniscal degeneration is commonly evaluated with semiquantitative KOA magnetic resonance (MR) assessments and is defined as increased intrameniscal signal alterations that does not extend to an articular surface. 2,3Meniscal degeneration is considered to indicate an early sign that may lead to future meniscal tear.
5][6] Therefore, understanding the relationship between meniscal degeneration and the specific outcome of destabilizing meniscal tears is crucial.Similarly, AKOA is a clinically important outcome because it represents a more painful and physically debilitating disease when compared with the typical, progressive KOA onset. 7The rapid decline in joint health in AKOA is commonly preceded by a destabilizing meniscal tear. 4ile meniscal degeneration likely antedates a meniscal tear and meniscal pathology is a risk for KOA.It remains unclear whether meniscal degeneration, specifically, is prognostic of future incident destabilizing meniscal tears or accelerated knee osteoarthritis (AKOA). 2,3,8This manuscript seeks to address this gap in the literature by assessing the prognostic potential of MR-assessed meniscal degeneration in relation to incident destabilizing meniscal tears or AKOA.
To address this research question, we will use data from the Osteoarthritis Initiative (OAI), which is uniquely suited for studying early-stage KOA as it is the only large cohort to collect annual knee MR images for up to 4 years before disease onset.We hypothesize that people with meniscal degeneration will be at higher risk of incident destabilizing meniscal tear or AKOA over a 4-year period when compared with people with no meniscal degeneration.By elucidating the prognostic potential of MR-assessed meniscal degeneration, this manuscript will make an important contribution to our understanding of the early pathogenesis of KOA and inform strategies to identify and intervene in high-risk individuals.

| Study design
This is a secondary data analysis of a nested case-control study (level of evidence III) that described people who developed AKOA within the initial 4 years of the OAI. 9 The OAI enrolled 4796 people with or at risk for symptomatic KOA across four clinical sites from February 2004 to May 2006: Ohio State University, University of Maryland and Johns Hopkins, Memorial Hospital of Rhode Island, and University of Pittsburgh. 10 All OAI enrolling sites and the OAI coordinating center at the University of California, San Francisco received institutional review boards approval before conducting the study.All participants provided informed consent before participating in the OAI.For our first aim, we considered the presence of medial meniscal degeneration at the OAI baseline visit (yes/no) to be the exposure and incident destabilizing medial meniscal tear during the first 4 years of the OAI to be the outcome.For our second aim, we considered the presence of meniscal degeneration in the medial or lateral meniscus at the OAI baseline visit (yes/no) to be the exposure and incident radiographic onset of AKOA during the first 4 years of the OAI to be the outcome.

| Participant selection
For the original nested case-control study, we identified groups based on the amount of radiographic disease progression during the first 4 years of the OAI. 9 Central OAI readers graded annually collected weight-bearing, fixed-flexion posteroanterior knee radiographs to monitor radiographic disease progression (files: kXR_SQ_BU##_SAS [versions 0.6, 1.6, 3.5, 5.5, and 6.3]; intrarater agreement: weighted κ = 0.70-0.80). 10,11All participants had no radiographic KOA (i.e., Kellgren-Lawrence [KL] grade 0/1) at the  11 We matched people in the typical and no KOA groups by sex to people in the AKOA group (i.e., 125 people per group; 375 total).For this secondary data analysis, we limited our study sample to people with (1) intact menisci (defined as having grade = "normal or meniscal degeneration" on MR images; see Section 2.3) at the OAI baseline visit (n = 226) and (2) 4-year meniscal status data (n = 221).
For the AKOA statistical analysis, we created a single "No AKOA" group by combining the people from the typical and no KOA groups to compare people who did or did not develop AKOA.

| MR acquisition
All OAI sites used identical Siemens Trio 3-Tesla MR systems and used the standardized OAI Imaging Protocol to acquire MR images. 12e following sequences were provided to the two radiologists (R. J. W.: 255 cases, J. W. M.: 120 cases) to perform the semiquantitative meniscal scoring: (1) sagittal intermediate-weighted, turbo spin echo, fat-suppressed MR sequence; (2) coronal intermediate-weighted, turbo spine echo, sequence without fat suppression, threedimensional dual-echo steady-state sequence.These sequences have been described in detail elsewhere. 12

| Semiquantitative meniscus status
The two radiologists used a modified International Society of Arthroscopy, Knee Surgery, and Orthopedic Sports Medicine meniscal tear classification to assess the meniscus status at the OAI baseline and first four annual follow-up visits. 13The readers assessed the status of the body and posterior/anterior horn of the medial and lateral meniscus: (1) normal, (2) meniscal degeneration, (3) radial, (4) maceration, (5) complex tear, (6) horizontal, (7) flap horizontal, (8) vertical longitudinal, (9) morphologic deformity, or (10) vertical flap tear.The readers also indicated if they observed a root tear.We categorized each meniscus into one of three categories: (1) intact meniscus: normal or meniscal degeneration in all three regions; (2) destabilizing meniscal tear: radial, root, maceration, or complex tear in any meniscal region; and (3) other meniscal tear: horizontal, flap horizontal, vertical longitudinal, morphologic deformity, or vertical flap tear in any meniscal region (excluding menisci with a destabilizing meniscal tear).This study only included participants with an intact meniscus at the OAI baseline.
The exposure for this study was the presence of meniscal degeneration at OAI baseline.Meniscal degeneration was defined as an increased signal on a fluid-sensitive sequence that does not extend to an articular surface. 2,3For the destabilizing meniscal tear outcome, incident destabilizing meniscal tear was defined as progressing from an intact meniscus to a destabilizing tear by the 4-year follow-up visit.For the meniscal group statistical analysis, the incident destabilizing meniscal tear group was compared with people that did not present with an incident destabilizing meniscal tear by the 4-year follow-up visit.For aim 1, we looked at the cross-tabs of the presence of meniscal degeneration at baseline and incidence of destabilizing meniscal tear in the medial and lateral meniscus separately.Interrater agreement was substantial (κ = 0.90) for medial and good (κ = 0.63) for lateral meniscal tear. 4

| Clinical data
For baseline demographics, body mass index (BMI), age, Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) pain score, and index knee KL grade were acquired at the OAI baseline visit.The data are publicly available (File: allclinical0#; version 0.2.2). 10

| Statistical analysis
We used a logistic regression model to assess whether people with medial meniscal degeneration (predictor) would be more likely to present with incident medial destabilizing meniscal tear over the next 4 years when compared with people without medial meniscal degeneration (reference group).This analysis was not replicated in the lateral meniscus due to limited sample size of incident lateral destabilizing meniscal tears (n = 7).We used a separate logistic regression model to assess whether people with the presence of meniscal degeneration in either meniscus (predictor) would be more likely to present with incident AKOA over the next 4 years when compared with people without medial or lateral meniscal degeneration (reference group).Each analysis was conducted unadjusted since the main purpose of this manuscript was to identify the prognostic ability of baseline meniscal degeneration.Since the original nested case-control study matched participants in the AKOA group to the other two groups based on sex, we conducted a sensitivity analysis that excludes the entire matched triad if one member was ineligible for the current study due to the restrictions of no meniscal tear at baseline.All analyses were performed with SAS Enterprise 7.15.

| RESULTS
In total, 221 participants were included in this study.Table 1 highlights the demographics for people with and without meniscal degeneration at the OAI baseline visit.People with the presence of a medial meniscal degeneration had three times the odds of developing an incident destabilizing medial meniscal tear within 4 years compared with a person with no medial meniscal degeneration at OAI baseline (odds ratio [OR]: 3.04; 95% confidence interval [CI]: 1.40-6.59;Table 2).The sensitivity of this model was 63%, the specificity was 65%, the positive predictive value was 23%, and the negative predictive value was 91%.
Since only seven people experienced an incident destabilizing lateral meniscal tear within 4 years, we did not perform a statistical analysis.However, six out of the seven people who experienced an incident destabilizing lateral meniscal tear had lateral meniscal degeneration at the OAI baseline visit.
Similarly, people with meniscal degeneration in either the medial or lateral meniscus had five times the odds of developing incident AKOA within 4 years compared with a person with no meniscal degeneration in either meniscus at OAI baseline (OR: 5.04; 95% CI: 2.57-9.89).The sensitivity of this model was 77%, the specificity was 60%, the positive predictive value was 42%, and the negative predictive value was 87%.
For the sensitivity analyses, there were 81 participants among the eligible matched triads from the original case-control study (Table 3).The sensitivity analyses did not significantly alter the main results of this study.People with the presence of a medial meniscal degeneration had three times the odds of developing an incident destabilizing medial meniscal tear within 4 years compared with a person with no medial meniscal degeneration at OAI baseline (OR: 3.04; 95% CI: 0.98-9.45).People with meniscal degeneration in either the medial or lateral meniscus had 8 times the odds of developing incident AKOA within 4 years compared with a person with no meniscal degeneration in either meniscus at OAI baseline (OR: 8.11; 95% CI: 2.64-24.85).

| DISCUSSION
Our results highlight that people with meniscal degeneration are three and five times more likely to develop an incident destabilizing meniscal tear or AKOA, respectively, within the following 4 years.
Therefore, the presence of MR-detected meniscal degeneration should be considered clinically meaningful as it relates to future incidence of clinical outcomes in individuals that have no meniscal pathology or radiographic evidence of KOA at baseline.These findings are clinically significant as they indicate that the presence of meniscal degeneration on MR may be an early biomarker for progression to AKOA or destabilizing meniscal tear.
Identifying prognostic biomarkers that predict the future incidence of AKOA or destabilizing meniscal tears is critical to reducing the high burden attributed to these outcomes. 14,15The results of this study corroborate previous findings that highlight the importance of the menisci in the development of AKOA. 7,9,16,17ecifically, the presence of any medial or lateral meniscal pathology T A B L E 1 Group baseline characteristics.| 2421 at OAI baseline was associated with 2.1 and 2.4 times the odds of developing AKOA, 16 respectively, compared with those that did not develop AKOA.Additionally, at 2 years before the onset of AKOA, the odds of having a destabilizing meniscal tear were greater than four times higher in adults who developed AKOA compared with adults that did not develop AKOA. 4 The current study results demonstrate that among people with intact menisci, meniscal degeneration may be an early marker that places the joint at high risk for accelerated decline.This adds to prior work that highlight the clinical significance of meniscal degeneration as a risk factor for future OA development. 8,18We hypothesize that meniscal degeneration reflects underlying structural changes of the meniscus and loss of meniscal loading capabilities that lead to a destabilizing meniscal tear and AKOA.
The results of this study also highlight the need for future studies to distinguish menisci graded as "normal" and those with "meniscal degeneration."Previous studies often combine normal and meniscal degeneration into the same meniscal category since both grades do not have a discrete meniscal tear. 19,20However, the current study highlights the potential clinical relevance of meniscal degeneration as a risk factor for future destabilizing meniscal tear or AKOA.
Therefore, studies that combine normal and meniscal degeneration into the same group may be obscuring meaningful associations as the reference group (intact meniscus) includes both healthy and pathologic menisci.
While this study provided important information regarding the prognostic capabilities of meniscal degeneration, some limitations should be addressed.First, this study used a small sample size of an existing case-control study of people that develop AKOA.Future studies within the OAI and other longitudinal cohort studies should explore the full prognostic capabilities of meniscal degeneration.This analysis focused on incident destabilizing meniscal tear and AKOA as outcomes and did not use any patient-reported or objective physical function assessments to determine the effects of meniscal degeneration on longitudinal changes in patient-centered outcomes.However, destabilizing meniscal tears and AKOA are dramatic changes associated with poor clinical outcomes. 7,21Meniscal degeneration and AKOA may be mediated by incident destabilizing meniscal tears; however, our statistical analysis did not utilize a mediation analysis to test this conceptual model.Future studies are needed to confirm the pathway between meniscal degeneration and AKOA.Similarly, we intended to determine if a clinician should be concerned when they observe meniscal degeneration, regardless of whether the alterations are a proxy for a confounding factor (e.g., obesity, greater age) or part of a causal pathway.Hence, for these analyses, we did not adjust our statistical models.Our findings warrant future studies to understand why meniscal degeneration are prognostic of destabilizing tears and AKOA.Finally, we used a binary, subjective measure of meniscal degeneration, as this is commonly included in semiquantitative KOA scoring systems as well as clinical practice. 22However, using more complex quantitative imaging assessments (e.g., texture analysis) 23 or compositional MR sequences (e.g., T2 mapping) 24 may provide a more robust approach to characterize signal alterations within the meniscus.
In conclusion, the results of this study highlight that people with meniscal degeneration are at three and five times greater odds of developing an incident destabilizing meniscal tear and AKOA over the next 4 years when compared with people with no meniscal degeneration.These findings are important as they indicate that meniscal degeneration may be an early warning sign of altered functioning of the menisci that may lead to an accelerated decline in joint health.Additionally, we recommend that future studies do not combine people with meniscal degeneration and no meniscal degeneration into a single control group, as combining these two groups may confound results.
Presence of baseline meniscal degeneration predicts future destabilizing meniscal tears and incident accelerated knee osteoarthritis (AKOA).
T A B L E 2