Abstract
Background
Discoid meniscus (DM) is a congenital variant that typically affects the lateral meniscus of the knee. Despite the presence of multiple classification systems for DM, no system has demonstrated utility in treatment planning. We performed a retrospective review of discoid menisci that underwent surgical intervention at our institution to assess for discoid meniscus characteristics associated with surgical treatment with or without repair. Next, we propose a new method of classifying lateral discoid meniscus and demonstrate its efficacy in predicting surgical treatment.
Methods
We searched the orthopedic database at our institution for patients who underwent surgical treatment for lateral discoid meniscus between 1991 and 2017. Clinical records were reviewed to determine the type of DM surgery performed (surgery with or without repair) as well as DM morphology, stability, tear presence, tear location, and tear type. Univariate associations between DM characteristics and surgery type were calculated and multivariable logistic regression models of surgery type were created. Based on the univariate analysis and logistic regression models, we propose a new classification system for discoid meniscus.
Results
There were 434 knees with discoid lateral menisci that received surgical treatment at our institution between 1991 and 2017. In univariate analysis, unstable menisci (93%, p<0.001) and menisci with a tear including the periphery (75%, p<0.001) were more likely to receive surgery with repair. By itself, instability demonstrated 89% sensitivity and 94% specificity in predicting surgery with repair. The main effects logistic regression model including stability and tear showed that that the odds of unstable lateral discoid menisci receiving surgery with repair was 133.1 times higher than stable menisci (p<0.001) while lateral discoid menisci that had a tear including the periphery had 6.54 times higher odds of receiving repair than those that had a tear in the central portion only (p<0.001).
Conclusions/Significance
Lateral discoid menisci stability and tear location were associated with surgical treatment type in both univariate analysis and logistic regression models (Table 1). Based on our results and clinical relevance, we propose a classification system for DM with utility in surgical treatment planning. Menisci are classified by stability (Stable (S0), Unstable (S1)), morphology (Incomplete (M0), Complete (M1)), and Tear (No tear (T0), Central tear (T1), or Peripheral tear (T2)) (Figure 1).
Tables and Figures
Logistic Regression Models showing the incremental predictive value of stability, tear location, and their interaction over the Watanabe classification (n=378).
| Model | Variable | P-Value | Log Likelihood (-2) | P-Value Comparing Models |
|---|---|---|---|---|
| (1) Watanabe Only | Watanabe | <0.001 | 488.3 | |
| (2) Watanabe + Stability | Watanabe | 0.290 | 204.8 | <0.001 (1 vs 2) |
| Stability | <0.001 | |||
| (3) Stability | Stability | <0.001 | 208.8 | 0.261 (2 vs 3) |
| (4) Stability + Tear Location | Stability | <0.001 | 192.6 | 0.002 (3 vs 4) |
| Tear Location | <0.001 | |||
| (5) Interaction Model | Interaction | 0.024 | 159.6 | <0.001 (4 vs 5) |
Discoid meniscus classification system based off of stability, morphology, and tear. Diagram shows the total number of patients studied per classification and the percentage of patients per classification receiving surgery with repair (bolded). (*) Indicates the predicted probability estimates (with 95% Confidence Interval) of surgery with repair based on our logistic model using stability, morphology, and tear. Instability was the main driving force behind receiving surgery with repair. Of note, 100 (23%) of the 434 patients had incomplete data.