Objectives: To determine the significant risk factors for medial meniscal injury in naturally occurring cranial cruciate ligament rupture and to quantify the risk using multivariate analysis.
Methods: A retrospective case control study was performed of dogs that had undergone surgery for cranial cruciate ligament rupture. Data recorded included patient signalment (age, breed and sex), the duration of the lameness, the extent of the cranial cruciate ligament rupture (complete or partial) and the condition of the medial meniscus. Logistic regression was used to analyse the relationship between these variables and tears in the medial meniscus.
Results: One hundred and sixty-one of 443 stifles (36·3%) in 366 dogs had a medial meniscal tear. The risk of a medial meniscal tear was increased by 12·9 times in association with complete cranial cruciate ligament rupture (OR 12·9; 95% CI 6·8 to 24·2), by approximately 2·6% for each additional week of lameness (OR 1·026; 95% CI 1·009 to 1·043) and by approximately 1·4% for each additional kilogram of bodyweight (OR 1·014; 95% CI 1·000 to 1·028). Golden retrievers and Rottweilers were at increased risk and West Highland white terriers were at reduced risk of medial meniscal tears compared with Labrador retrievers.
Clinical Significance: To minimise the risk of medial meniscal tears, surgical stabilisation should not be unnecessarily delayed.
Damage to the medial meniscus is a frequent finding in dogs suffering from cranial cruciate ligament rupture (CCLR). The prevalence of meniscal tears in dogs undergoing surgery for CCLR is reported to be 10 to 70% (Flo 1993). Meniscal tears are thought to arise through multiple factors, including the degree of stifle instability, duration of the lameness, bodyweight and activity level of the dog.
The relative importance of these risk factors has not been systematically investigated, although there are several previous reports that together present somewhat confusing evidence for apparent relationships between selected variables and meniscal injury. For instance, Timmerman and others (1998) reported an increased risk of medial meniscal tears in dogs in which the duration of lameness was greater than 6 weeks. In contrast, Ralphs and Whitney (2002) found no increased incidence of medial meniscal tears associated with lameness lasting greater than 14 days, but noted that medial meniscal tears were significantly more likely in dogs with complete rather than partial cruciate ligament rupture. Nečcas and Zatloukal (2002) examined multiple factors that may be involved in medial meniscal tears and concluded that dogs weighing less than 24 kg had a significantly lower risk of medial meniscal tears and dogs with more severe osteoarthritis or joint effusion had a higher risk. Furthermore, this study found that duration and severity of lameness, severity of joint instability and extent of CCLR were not significant risk factors for medial meniscal tears. An analysis of these and other studies on naturally occurring CCLR is complicated by the multitude of variables, low number of cases and inconsistent methods of appraising the intra-articular structures.
In 1993 Flo stated, “If cranial cruciate ligament stabilisation was performed earlier the meniscal injury rate would probably be less.” If this were indeed the case, early surgical intervention would be expected to improve patient outcome because there is compelling evidence that meniscal tears, partial or complete meniscectomy or meniscal release is associated with poorer joint function than uncomplicated CCLR in dogs (Cox and others 1975, Innes and others 2000, Luther and others 2009, Thieman and others 2009). In human beings, it is widely accepted that meniscal pathology at the time of cruciate reconstruction is a major prognostic factor in the progression of osteoarthritis within the knee joint (Magnussen and others 2009).
This study was constructed to determine the veracity of Flo’s hypothesis and to investigate the importance of other complicating factors. This study aimed to determine systematically the relationships among breed, bodyweight, lameness duration and extent of CCLR on the prevalence of medial meniscal tears. Our hypothesis was that increasing duration of lameness, greater bodyweight and complete (versus partial) CCLR would each be associated with a greater prevalence of medial meniscal tears.
Materials and methods
Clinical information was retrieved from the records of all dogs that were presented to the authors’ hospital for the surgical treatment of CCLR between 2002 and 2009. Data for analysis included patient signalment (age, breed and sex), duration of lameness, extent of CCLR and condition of the medial meniscus.
The duration of lameness was that reported by the owners or referring veterinary surgeons. The extent of CCLR was determined from the surgical report and stifle manipulation during preoperative examinations under sedation or anaesthesia (McKee and Cook 2006, Carobbi and Ness 2009). “Complete” ruptures were grossly unstable stifles (cranial drawer and cranial tibial thrust positive) with surgical confirmation that no functional ligament remained intact. “Partial” ruptures were diagnosed in stable stifles (i.e. minimal/absent cranial drawer or cranial tibial thrust in extension) with surgical confirmation of incomplete rupture. The condition of the medial meniscus was determined from the surgical report: either “intact” or “torn.” All cases underwent medial or lateral stifle arthrotomy, with or without meniscal probing, by a board-certified surgeon or senior resident. Dogs were excluded if clinical records were incomplete or if previous surgery had been performed on the affected stifle.
Spreadsheet software (Excel 2003, Microsoft) was used for data management before further analysis using statistical software (GraphPad Prism 5.0, GraphPad Software Inc. (San Diego, CA, USA) and SPSS 16.0, SPSS Inc, Chicago, IL, USA). Categorical data were analysed using the chi-squared test. Ratio measurements were summarised as mean (±sd), tested for deviations from normality using the D’Agostino-Pearson test and then compared using either the t test or the Mann-Whitney test. All variables were examined with the null hypothesis of homogeneity between groups, with statistical significance taken as P≤0·05 (two-tailed). Logistic regression was performed on variables thought to be associated with medial meniscal tears using forced-entry regression (entry criterion P<0·15). Continuous variables were duration of lameness and bodyweight; categorical variables were the degree of CCLR (partial or complete) and breed (baseline indicator breed Labrador retriever). Odds ratios (OR) and confidence intervals (CI) were reported for variables significantly associated with the outcome as defined by Wald’s test.
Four hundred and forty-three stifles from 366 dogs met the inclusion criteria and were included in the study. Medial meniscal tears were found in 161 of the 443 stifles (36·3%). Of these 443 stifles, 169 (38·1%) had partial CCLR and 274 (61·9%) had complete CCLR. Medial meniscal injuries were present in 17 of 169 (10·1%) stifles with partial CCLR and 144 of 274 (52·6%) stifles with complete CCLR (chi-squared test, P<0·0001, Table 1). Forty-five dogs that had CCLR surgery during the study period could not be included in the analysis because of incomplete records.
Table 1. Contingency table of surgical findings in 443 stifles from 366 dogs (P<0·0001, chi-squared test)
CCLR cranial cruciate ligament rupture
Intact medial meniscus
Torn medial meniscus
The prevalence of medial meniscal tears varied between different breeds, with a particularly high prevalence in the Rottweiler and golden retriever and a lower prevalence in smaller breeds such as the West Highland white terrier (Table 2). Partial CCLR was found in most breeds in the study, but was particularly common in the boxer, the German shepherd dog and the West Highland white terrier (Table 2). Contingency tables comparing breeds with more than 10 individuals represented demonstrated evidence of breed differences in the prevalence of medial meniscal tears (chi-squared test, P<0·0001) and the prevalence of partial CCLR (chi-squared test, P<0·01).
Table 2. Prevalence of medial meniscal tears and partial CCLR in breeds with more than 10 individuals represented
Total number with medial meniscal tear (%)
Total number with partial CCLR (%)
CCLR cranial cruciate ligament rupture, GSD German shepherd dog
Staffordshire bull terrier
West Highland white terrier
A summary of the ratio measurements for each category of CCLR along with univariate analysis is presented in Table 3; however these data were gathered primarily to allow a more powerful multivariate analysis to be performed, which more accurately reflects the complex interactions occurring between variables in clinical cases. Multivariate analysis using the data from all 443 cases formulated a regression model able to predict the state of the medial meniscus in 77·9% of cases. Gender (P=0·42) and age (P=0·27) were not found to be statistically significant risk factors and were removed from the final models. Factors found to be statistically significantly associated with the outcome according to Wald’s test included complete CCLR (P<0·001), duration of lameness (P=0·003), breed (P=0·010) and bodyweight (P=0·043). Complete CCLR increased the risk of a medial meniscal tear by 12·9 times (OR 12·9; 95% CI 6·8 to 24·2) compared with partial CCLR. The risk of a medial meniscal tear increased by around 2·6% for each week of lameness [duration of lameness (weeks) OR 1·026; 95% CI 1·009 to 1·043]. Increasing bodyweight increased the risk of a medial meniscal tear by around 1·4% per additional kilogram of bodyweight (OR 1·014; 95% CI 1·000 to 1·028). Although breed was significantly associated with medial meniscal tears (P=0·010), few individual breeds were significantly different from the baseline Labrador retriever. The golden retriever (OR 3·0; 95% CI 1·2 to 7·5) and the Rottweiler (OR 2·3; 95% CI 1·0 to 5·4) were at higher risk and the West Highland white terrier (OR 0·20; 95% CI 0·04 to 1·0) at lower risk of a medial meniscal tear than the Labrador retriever. Logistic regression equations were used to plot the probability of meniscal injury against duration of lameness for the Labrador retriever (Fig 1) and the Rottweiler (Fig 2) based on the type of CCLR.
Table 3. Relationship between condition of medial meniscus and duration of lameness, age and bodyweight
*Mann-Whitney test. Bold type indicates statistical significance (P<0.05)
Duration of lameness (weeks)
Duration of lameness (weeks)
Duration of lameness (weeks)
This study demonstrates an association between duration of lameness and medial meniscal injury, particularly for complete CCLR. In addition, for dogs with complete CCLR, those with medial meniscal tears were significantly larger than those with intact menisci. We also found evidence that there are breed differences in the prevalence of medial meniscal tears, with golden retrievers and Rottweilers being at higher risk and West Highland white terriers being at lower risk than the Labrador retriever. This difference in the prevalence of meniscal tears persists even when the effect of bodyweight and duration of lameness are taken into account through the multivariate analysis, suggesting that other factors including differences in activity levels, joint biomechanics, contralateral stifle disease and genetic variation between breeds may also be important.
An association between medial meniscal tears and complete CCLR has been suspected previously (Ralphs and Whitney 2002) and was confirmed in this study. Consequently, the prevalence of medial meniscal tears was significantly lower in association with partial CCLR. The risk of a medial meniscal tear increased by 12·9 times in association with complete CCLR (“unstable joint”) in comparison with a partial CCLR (“stable joint”). The prevalence of partial CCLR was significantly different between different breeds; the boxer and German shepherd dog are more frequently found to have partial CCLR and consequently have a lower prevalence of medial meniscal tears than similarly sized dogs of other breeds.
Limitations of this study include those inherent in all retrospective studies, particularly the dependence on accurate case notes for the collection of patient data. The duration of lameness was often based on client recollection and in some cases an acute deterioration in the lameness was noted in the history. Consequently, the total duration of lameness for complete CCLR may include a variable proportion of time for which a partial CCLR was present. The prevalence of medial meniscal tears may be underestimated due to variation in the skill and experience of the surgeons performing the arthrotomy and inconsistent use of a meniscal probe. Arthroscopic evaluation and probing has recently been shown to be a superior method for evaluating the menisci in some studies (Pozzi and others 2008), but not in others (Ertelt and Fehr 2009). Another factor that requires clarification is the definition of partial CCLR. In this study the extent of the CCLR was defined to reflect the degree of stability present within the joint, as it is the cranial tibial translocation that is postulated to contribute to meniscal damage (Flo 1993). The prevalence of partial CCLR may be increasing because of greater awareness of cruciate ligament disease and higher client expectations. Clearly a spectrum of partial rupture exists and this should be clarified by further research in order to allow comparison between studies and ultimately allow the development of evidence-based treatment recommendations. A high proportion of dogs operated upon during the study period were necessarily excluded from the analysis because of incomplete case records, particularly insufficiently detailed surgical reports. The potential for these exclusions to bias the results is unknown, but likely to be minimal, because they do not represent a systematically different sample population. A further limitation of this study is the bias for larger breeds of dogs. This bias represents a population of dogs referred for treatment and may not be representative of the spectrum of breeds that were presented with CCLR. It is likely that smaller breeds are under-represented as they are less frequently referred for surgical management. A multi-centre study involving first opinion practices may provide a more representative population and increase the power of the study, allowing accurate OR to be calculated for more breeds.
In conclusion, this study presents evidence to support the assertion that if functional stifle stability can be obtained, early surgery for CCLR may reduce the incidence of medial meniscal injury. Bodyweight and complete CCLR were also shown to be significantly associated with an increased risk of medial meniscal injury. To improve patient outcomes, particularly in breeds that may be at higher risk of meniscal injury, surgery should not be unduly delayed. The optimal time for surgery has not been established and remains a contentious issue in the surgery of human beings. Although there is good evidence that delayed surgery results in a greater prevalence of meniscal tears (Keene and others 1993, Millett and others 2002), articular cartilage damage (Smith and others 2002, Nelson and others 2006) and soft-tissue atrophy, some studies suggest that a delay in the index surgery does not affect the short to medium term functional outcome in most human patients (Tambe and others 2006). However, it is generally accepted that the chondral consequences of delayed surgery can be associated with the progression of osteoarthritis (Nelson and others 2006) and a poorer outcome in the longer term in human patients (Marcacci and others 1995, Magnussen and others 2009,). The relationship between functional outcome and timing of CCLR surgery has yet to be investigated fully in dogs.
The authors would like to thank Dr. T. J. McKinley for statistical advice.
Conflict of Interest
None of the authors of this article has a financial or personal relationship with other people or organisations that could inappropriately influence or bias the content of the paper.