Geometrical characteristics of the knee are associated with the injury of the meniscus

To assess the geometrical risk factors for meniscal injuries. Our hypothesis was that the narrowness of the intercondylar notch and the smaller tibial spine could increase the risk of meniscal injuries. We retrospectively studied two hundred and seven patients examined for knee magnetic resonance images. The severity of meniscal injuries was evaluated by two experienced orthopedists. The notch width, bicondylar notch width, notch width index, condyle width of the femur, tibial spine height, and intercondylar angle were measured in magnetic resonance image slides by two blinded orthopedists. difference was found between the grade 1 and the control group. Subscript b indicates the signicant difference was found between the grade 2 and the control group. Subscript c indicates the signicant difference was found between the grade 3 and the control group. Subscript d indicates a signicant difference was found between grade 1 and grade 2. Subscript e indicates the signicant difference was found between grade 1 and grade 3. Subscript f indicates a signicant difference was found between grade 2 and grade 3.


Abstract Background
To assess the geometrical risk factors for meniscal injuries. Our hypothesis was that the narrowness of the intercondylar notch and the smaller tibial spine could increase the risk of meniscal injuries.

Methods
We retrospectively studied two hundred and seven patients examined for knee magnetic resonance images. The severity of meniscal injuries was evaluated by two experienced orthopedists. The notch width, bicondylar notch width, notch width index, condyle width of the femur, tibial spine height, and intercondylar angle were measured in magnetic resonance image slides by two blinded orthopedists.

Results
In all two hundred and seven patients, 112 patients with a meniscus injury and 95 patients were as healthy control. The NWI (P = 0.027) in patients with meniscus injuries was signi cantly different from the control group. A 1 SD (0.04 mm) increase in NWI was associated with a 0.4-fold increase in the risk of meniscal injury. A 1 SD (0.04 mm) increase in NWI was associated with a 0.64-fold increase in the risk of grade 3 meniscal injury. Furthermore, NWI and medial spine height are decreased signi cantly in grade 2 (P < 0.05) meniscal injury than in other grades. The medial spine height was signi cantly decreased in the meniscal injury group (P = 0.025), and the decrease of medial spine height would increase the risk of meniscal injury (OR = 0.77) and grade 3 meniscal injury (OR = 0.8).

Conclusions
The stenosis of the femoral intercondylar notch and small medial tibial spine are risk factors of meniscal injury. The decreased NWI and the decreased medial tibial spine height were also associated with the severity of the meniscal injury.

Background
The injury of the meniscus is one of the most common orthopedic issues worldwide. It is commonly seen in athletes with pivoting maneuvers. Sometimes patients performing low-impact sports can also be injured (1).
This study set out to investigate the relationship between the anatomical characteristics of the knee and the meniscal injury using magnetic resonance imaging (MRI) data. We hypothesized that the geometrical features of the knee such as NW, BCW, NWI, condyle width, spine height, and intercondylar angle are associated with the injury of the meniscus.

Study participants
This study was a retrospective case-control study including 207 patients with or without the injury of meniscal from January 2015 to January 2020. All patients were diagnosed by MR images and con rmed by two experienced orthopaedists. Patients without any sign of meniscal injury were included as the control group. The excluding criteria were as follows: patients with femoral or tibial fractures, previous surgery of knee or ligaments, deformity of the knee. Age, sex, and with or without ACL injury or rupture were recorded for all participants. The ethics committee approved this study of China-Japan Friendship Hospital.

Measurement of magnetic resonance image
For each participation included, the 1.5 Tesla Knee MRI were collected. The thickness of the slides was 4mm. DICOM MRI images were viewed and measured using OsiriX Software (Pixmeo, Geneva, Switzerland, version 12.0.2).
Two experienced orthopaedists examined meniscal injuries, and the high signal intensities of meniscus were classi ed using the three-grade staging system introduced by Fishcher et al. (25). Grade 1: the highintensity signal on T2-weighted images was limited in a small focal area. Grade 2: the high-intensity signal presented as a linear area but did not extend to the articular surface. Grade 3: the abnormal signal extended to at least one articular surface, which indicating meniscal injury.
The NW and BCW were measured using the method introduced by Domzalski et al. (26). The BCW was measured at the level of the popliteal grove in a T2-weighted coronal magnetic resonance image. The line was drawled from the lateral condyle of the femur, parallel to the joint line, and connected to the distal femoral condyles. The NW was measured from the most inferior margins of the borders of the intercondylar notch. Both the width of the medial and lateral femoral condyle was measured at the same level (Fig. 1A). The NWI represented the ratio of NW to the BCW. The height of tibial spines was measured on the T2-weighted mid-coronal magnetic resonance image, de ned as the single slice that presented the largest area of the tibial spine. The greatest width of the tibial was chosen if the size of the spine in the two slices were similar (Fig. 1B). The intercondylar angle was shaped by two lines from the top of the intercondylar notch to the most inferior border of the intercondylar notch at both medial and lateral condyles in proton density-weighted axial magnetic resonance images (Fig. 1C).
All the measurements were made twice by two experienced orthopedists. The average of each measurement was calculated.

Statistical Analysis
For continuous variables, the normal distribution test was performed using the Kolmogorov-Smirnov test.
The unpaired t-test was used if the variable were for to compare the difference between the group of meniscal tear and the normal group, and the one-way analysis of variance (ANOVA) with post-hoc test (LSD) was used to explore the underlying relationship between the grade of the meniscal tear and the age, NW, BCW, NWI, the height of medial spine peak, condyle width and intercondylar angle which was presented as mean ± standard deviation. For categorical data, chi-square tests were used the analyze the relationship of meniscal injuries and the sex, the number of patients that with injury or rupture of ACL, which presented as amount and proportions. To analyze the most highly associated factors associated with the meniscal injury and the grade of meniscal injury, binary and ordinal logistic regression with the Enter method were used. Correlated independents were excluded to avoid the in uence of these independents to the risk of meniscal injury or the grade of meniscal injury. The variable included in the logistic regression were either associated with the risk of meniscal injury or the grade of meniscal injury independently or had clinical associations with the dependent. All data were analyzed using SPSS software, version 27.0 (SPSS Inc., Chicago, IL). The signi cance level for all analyses was set at P < 0.05.
The interclass correlation coe cient (ICC) was calculated to evaluate the intra-and interobserver reliability. The ICC value greater than 0.9 was considered excellent, and a good value was de ned between the 0.8-0.9 (27).

Patients characteristics
In total, 207 patients were included in this study, 112 patients with the injured meniscus and 95 patients were included as the control group. Among them, 119 were female, and 88 were male, with a mean age of 48.63 ± 15.53. Among the patients being studied, 59 (32.37%) with an ACL injury,45 (21.74%) with ACL rupture were observed in 74 (35.75%) patients ( Table 1). The reproducibility of all the measurements in magnetic resonance images was excellent, with an average ICC of 0.96 (from 0.91 to 0.99). Details were shown in Supplemental Table 1. Correlation between the meniscus tear and knee morphometrics Compared with the control group), the NWI (P = 0.027) in patients with meniscus injuries was signi cantly different than that in patients without it. In contrast, there was no signi cant difference between the two groups regarding NW (P = 0.081), BCW (P = 0.574), both medial and lateral condyle width (P = 0.145, P = 0.199), and lateral spine height (P = 0.925). Furthermore, the medial spine height (P = 0.025) in meniscal injury was signi cantly different from the control group (Table 1). In the binary logistic regression modeling adjust for gender, age, NWI, medial condyle width, lateral condyle width, medial spine height, lateral spine height, and intercondylar angle (Table 2), ve variables were found to be associated with increased odds of meniscal injury: gender (OR 3.43, 95% CI 1.23-9.55), age (OR 1.06, 95% CI 1.04-1.09), NWI (OR 0.4, 95% CI 0.2-0.78) and medial spine height (OR 0.77, 95% CI 0.61-0.98). A 1 SD (0.04 mm) increase in NWI was associated with a 0.4-fold increase in the risk of meniscal injury. Correlation between the grade of meniscus injury and knee morphometrics The pairwise comparison was performed to identify the difference between the grades of meniscal injury and the control group. Signi cant differences were observed in age, ACL rupture, NWI, medial condyle width, and medial spine height (Table 3). All three grades showed signi cant differences compared with the control group regarding the age of patients, respectively. The number of patients with grade 3 meniscal injury and ACL rupture was signi cantly different from grade 1. The NWI in patients with grade 2 meniscal injury was signi cantly lower than that of the control group. No signi cance was found among three degrees. Furthermore, we found signi cant differences in the medial condyle width between grade 3 and the control group and grade 3 and grade 2. In addition, medial spin height also showed a signi cant difference between grade 2 and the control group, grade 2 and grade 1, grade 2 and grade 3, respectively. In the ordinal logistic regression modeling adjust for gender, age, NWI, medial condyle width, lateral condyle width, medial spine height, lateral spine height, and intercondylar angle (  *Statistical analysis was performed using the independent t test for continuous variables and using the chi-square test for categorical variables. Subscript a indicates the signi cant difference was found between the grade 1 and the control group. Subscript b indicates the signi cant difference was found between the grade 2 and the control group. Subscript c indicates the signi cant difference was found between the grade 3 and the control group. Subscript d indicates a signi cant difference was found between grade 1 and grade 2. Subscript e indicates the signi cant difference was found between grade 1 and grade 3. Subscript f indicates a signi cant difference was found between grade 2 and grade 3.
Abbreviations: ACL anterior cruciate ligament, NW the femoral notch width, BCW the bicondylar width, NWI the notch width index Abbreviations: NWI the notch width index, OR odds ratio, CI con dence interval

Discussion
The major discovery of our study was the correlation of the geometrical parameters of the knee and the meniscal injuries. To be speci c, the decrease of NWI and medial spine height was signi cantly correlated with the increased risk of meniscal injury compared with the control group. Moreover, NWI and medial spine height are also signi cantly related to the severity of the meniscal injury. Although statistical signi cance was not found in grade 1 and grade 3 comparing with the control group, the result showed lower NWI in these two groups. Furthermore, we also con rmed the meniscal injury was correlated with sex and age. Ordinal regression also veri ed the relationship between genders, and age, which is related to the severity of the meniscal injury.
The ndings of our study reveal that meniscal injury and its grade were associated with NW and NWI, which are similar to the ndings in previous results on ACL injuries (12,14). In 1938, Palmer rst recognized the narrowness of intercondylar was associated with the ACL injury (28). Souryal et al. portrayed a method of measuring the intercondylar width, the NWI on plain radiograph (13), and found the correlation between the stenosis of the intercondylar notch and the injury of ACL. Previously studies have concluded that the femoral NW and NWI, as two-dimensional parameters, can evaluate the size of the femoral notch effectively (29). Previous studies demonstrated that the stenosis of intercondylar notch increasing the risk of ACL injury (12,14). A narrow space of intercondylar notch tends to house a relatively small volume ACL, which the strength of the ligament was decreased and consequently led to the predisposition of ACL injury (19,20). As the knee over-bent or rotated, a narrow intercondylar notch tends to lead to impact between the lateral wall of the femoral intercondylar and ACL. This phenomenon was more evident when the knee was externally rotated or at the position of exion-valgus. As the impact of ACL harms the ber bundles, the intensity of ACL decreased and prone to be injured (19,20). On the other hand, the correlation between the NW or NWI and the risk of ACL injury was controversial (12,30 (13,14,26).
The tibial spine is located at the center of the tibial femoral articular surface through weight-bearing activity. The medial tibial spun has the highest contact pressure in a load-bearing knee. Previous studies described the anatomical relationship between the ACL and the tibial spine (33,34). Oka  studies in ACL injuries, we found the decrease of medial tibial spine height can increase the risk of meniscal injuries. And in patients with grade 2 meniscal injury, the height of the medial tibial spine is signi cantly decreased compared to another degree of injury and the control group.
The intercondylar angle was another parameter that described the intercondylar notch. However, the investigation of intercondylar angle was poor in ACL-related papers. Alentorn et al. (17) found the decrease of intercondylar angle would increase the risk of ACL injuries, and they suggested a 50° of cutoff value. The same cut-off value was accepted by Stein et al. (18) but they found no association between the angle and the risk of ACL injuries. Alentorn et al. found the intercondylar notch angle was signi cantly narrower in ACL injury patients. Therefore, they considered the intercondylar angle as a more useful parameter to describe the narrow intercondylar notch (39). A similar result was concluded by Raja et al. (40). In this study, we found intercondylar angle did not associate with meniscal injury.
Females are apt to have ACL injuries than males, and the anatomical structures were different in females compared to males (41). The results of Wolters et al. (42) showed a narrower intercondylar notch in women, whereas results from Eck et al. (43) concluded that there were no differences in NWI between genders. In addition, the risk of ACL injury increased with age, as the results from Snoeker et al. (44). The difference in age and gender can also be found in the meniscus injury. The prevalence of meniscal injury increased with the age of patients (45). On the other hand, the incidence of acute meniscal injury decreased with age (46). The difference between gender in patients with meniscal injury remains controversial. In athletes, the meniscal injury was more easily found in males than females (47). In contrast, female athletes have a higher risk for medial meniscus posterior root tears (48). In our study, we found the number of female patients is larger than male, although no signi cant difference was found between the groups and the severity of the injury. However, the opposite result was found after analyzed by different severity of the injury. It may occur due to the relatively small amount of patients in each grade of injury and led to potential bias.
The ACL and meniscus have inseparable correlations when it comes to the injury of the knee (2-6, 12). Meniscus tears have been reported in 40%-82% of ACL tears, and the medial meniscus was more likely to be injured compared with the lateral meniscus (45). Previous studies considered the medial meniscus also has a restraining effect on the anterior tibial translation ACL. Shybut et al. (4) found the tibial translation changed increased signi cantly with ACL-de ciency knee, which can lead to the injury of the meniscus. With the de cient ACL, patients were prone to have larger internal rotation, and the meniscal translation increased compared to in the intact state (5). Levy et al. showed that compared to the lateral meniscus, the medial meniscus has a signi cant posterior wedge effect and is rmly connected to the tibial plateau with capsular attachments (2). After performing lateral meniscectomy in an ACL-de cient knee, the anterior tibial translation insigni cantly increased (3). Similar to the results of Levy et al., multiple prior studies have concluded the medial meniscus was a secondary stabilizer to ACL at the process of anterior translation (8, 9,12). After studying cadavers with both lateral and medial meniscectomies, Musahl et al. (49) found medial meniscus played a more important role in restraining the anterior tibial translation but have no effect on pivot shift. In comparison, lateral meniscus exerted its effect on preventing rotational disability and cannot inhibit the anterior tibial translation. Arner et al. (5) found that with the decreased strength of ACL, the lateral meniscus had more mobility and are more likely to injure. The lateral meniscus tears often presented at the acute stage of ACL injury, whereas the medial meniscus tears were more likely to develop at the chronic stage. This can be explained by the greater translation in the lateral meniscus and the greater stress conducted to the medial meniscus (6). On the other hand, the injury of the meniscus can also harm the stability of the ACL-de ciency knee. Shybut et al. (4) found the meniscal posterior root tears can further decrease the stability of the knee with ACL de ciency. The underlying mechanism of that was increased pivot-shift instability in those with injury of the lateral meniscus. The position of the meniscus could be altered after ACL reconstruction, which was reported in multiple studies (50,51) and the ACL reconstruction can also restore the abnormal biomechanics such as meniscal shift (9). These ndings indicate the injury of the meniscus and the pathological extrusion were closely associated with ACL and can be in uenced by the abnormality of the ACL.
Our studies have several limitations. First, our study measured the notch parameters and the tibial spine on segments of MRI, which only represent the intercondylar notch dimension at one slice. It cannot fully embody the overall volume of the intercondylar notch. Although a previous study found the two-dimensional measurement can effectively evaluate the volume of intercondylar notch (29), this measurement can lead to potential bias. However, due to the limitation of technology, the application of a three-dimensional measurement is restricted. Besides, Our study was a retrospective case-control study, and the imaging data were acquired after the injury. The cause and consequences cannot be elucidated.
For example, in those ACL-injured patients accompanied with meniscal injury, whether the narrowed intercondylar notch led to ACL injury rst, therefore, led to meniscal injury or the narrowed intercondylar notch in uenced the ACL and meniscus separately and directly remained unknown. Moreover, Due to the characteristic of our study, the number of female and male participants was unequal. The difference in gender can lead to different results in the prevalence of ACL injuries and meniscal injuries as previously reported (41,44,45). Restricted by the limited time and resources, our study did not include the height and weight of subjects. Consequently, the BMI cannot be calculated. As previously reported, BMI is a risk factor of meniscal injury (44) and it may have a potential in uence on our results. Future studies should take it into consideration. Only Chinese subjects were included in this study. The potential in uence of ethnicity cannot be analyzed and should be considered in future studies.

Conclusion
Our study shows that the stenosis of the femoral intercondylar notch and small medial tibial spine have an association with the increased risk of meniscal injury. The decreased NWI and the decreased medial tibial spine height were associated with the severity of the meniscal injury. The ethics committee of China-Japan Friendship Hospital approved this study on January 22, 2020, under the identi cation number 2020-QGW-078. No consent was obtained from participants for this is a retrospective study, and participants were anonymized.

Consent for publication
Not applicable Availability of data and materials All data generated or analyzed during this study are included in this published article and its supplementary information les.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download.