A retrospective review of all patients who underwent a MRI examination of the knee for trauma at the Yaounde Teaching Hospital between July 2012 and December 2013 was performed. This is a national referral hospital with over 40,000 new patients seen per year. Our sample consisted of all patients who came to do an MRI because of knee injury during the study period. Then we excluded those with tumors, infections, gonarthrosis, and previous knee surgery or isolated posterior cruciate ligament tear (Figure 1). Clinical data collected from the electronic medical records were age, sex and the knee’s side. The radiological files were reviewed collegially.

The MRI examinations were performed using a 0.2 T device (Signa Profile Excite Lightspeed VCT, GE Healthcare). Patients were installed supine in the MRI unit with the knee positioned in neutral rotation and extension into a dedicated cylindrical extremity coil. Feet were introduced into the magnet first.

Fast spin echo and proton density T1-weighed and T2-weighed images obtained in sagittal and coronal views were used in this study. Protocols settings were as follows: field of view 18 cm, slice thickness of 3 mm without interslice gap and a matrix of 512 × 512.

Radiological files were stored in DICOM format, transferred to an off-line workstation, and loaded to a software package for image data analysis and visualization (Clear Canvas Workstation 2.0 SP1). Proofreading was done collegially by the main investigator and two senior radiologists on a 21 inch monitor with a maximal resolution of 1368 × 768.

ACL tears were classified as total or partial ruptures [3] . Meniscal lesions were searched according to the criteria proposed by Crues and Stoller [17] . As recommended by Ravey et al. we only reported stage III lesions as post traumatic meniscal tears [12] . Other findings sought were: collateral ligament injuries (contusions or rupture), posterior cruciate ligament tears, bone lesions, quadricipital and patellar ligament lesions, anterior tibial translation and joint effusion.

We obtained ethics approval from the ethics committee of the Yaounde Teaching Hospital. It was considered that this study did not require informed consent from patients.

Statistical testing included the Kolmogorov-Smirnov test for normality, Fisher test for testing gender-related variability and logistic regression for assessing association between ACL tears and other lesions. All statistical analyses were performed using SPSS software (SPSS, version 20.0; SPSS Inc., Chicago, IL., USA). The level of statistical significance was set at p < 0.05.

The study population consisted of 96 patients aged 16 to 59 years. The sample was normally distributed (p = 0.213) with a mean age of 35.36 ± 11.86 years. There were 68 men (70.8%; sex ratio: 2.43). The left and right knees were represented in equal numbers (n = 48) (Table 1). The overall prevalence of ACL tears was 45.8% (44/96) with no gender difference (p = 0.20). Total tears were strongly predominant (36/44; 81.8%).

Twenty-three cases (23/96; 24%) of meniscal tears were found, including 16 cases (36.4%) significantly associated with ACL tears (p = 0.002). Meniscal tears were more frequent on the medial meniscus (p = 0.008) and particularly its posterior horn (p = 0.023). Simultaneous rupture of collateral ligaments (“bi-collateral” tear) was

MM: Medial Meniscus; LM: Lateral Meniscus.

significantly associated with ACL tears (p = 0.016). Joint effusion (49/96, 51.04%; p = 0.82) and bone contusion (29/96, 30.2; p = 0.3) were other frequent joint lesions encountered, but none of them was significantly associated to ACL tears (Table 3).

On multivariate analysis, meniscal tears and “bi-collateral” ligamentous tear were found to be independently associated with ACL tears (Table 4).

The predominance of ACL tears in traumatic knee disorders is now widely accepted [3] [18] . ACL is the main obstacle to anterior tibial translation; it provides secondary stabilization in response to internal tibial rotation and varus or valgusangulation. The main mechanisms of knee trauma are: external tibial rotation, internal rotation of the femur, extreme valgus unintentionally applied to a flexed knee, hyperextension and valgus or direct side impact [2] .

The association between ACL tears and “bi-collateral” tears had not been previously reported. The tibial collateral ligament is more often injured in case of knee trauma due to forced valgus which causes its stretching and tearing, meanwhile the fibular collateral ligament can be broken due to external rotation or direct lateral knee injury [1] [13] . According to Guenoun et al. forced valgus and external rotation are part of the most common ACL tears mechanism, which may justify their association with “bi-collateral” ligamentous tears [3] .

The association between ACL tears and meniscal tears is well established in the literature as stated by Lecouvet et al. [1] . The medial meniscus is usually more injured because it is thinner, narrower and more “opened” than the lateral meniscus [8] . Lerais et al. had found that the posterior horn is preferably teared because meniscal tears often start there and gradually extend to the front, so that isolated lesions of the anterior horn of the medial meniscus does not exist [19] .

We did not find any gender difference in the prevalence of ACL tears which is not consistent with the data of literature recalled by Cimino et al. who stated that there is 1.4 to 9.5 times increased risk of ACL tears in women [4] . Some theories for this predominance in women have been suggested: poorer access to training facilities, hamstring strength, hormonal fluctuations, narrow intercondylar notch, increased Q angle and quadriceps dominance [4] [20] [21] . The absence of gender difference in our study is probably due to the overrepresentation of men who constituted 70.8% of the sample. Men are probably more often the victims of knee injury than women because they are most likely to practice sport which is the first provider of ACL tears [22] .

We performed our examinations with a low field magnet (0.2 T). Several studies comparing low field (0.2 T) and high field (above 1.5 T) devices in the evaluation of ACL tears have found a poorer image quality and a longer examination time with low fields despite a comparable diagnostic accuracy [23] [24] . Thus, we considered this was not a limitation to carry out the study.

Three real limitations to this study should be mentioned: the limited sample size, overrepresentation of men and the lack of clinical data. MRI is a recent and still expensive technology (at least 225 U.S. dollars) in our country; clinicians do not prescribe enough and there are few patients who can afford it. The retrospective design of this study did not allow us to collect more clinical data in order to integrate them into our analysis and verify if, as previously reported, they could improve the detectability of ACL tears [5] . There is a necessity to continue this study prospectively on a larger and balanced sample in terms of gender, in order to validate its findings.

Our results suggest that in case of knee trauma, the presence of a “bi-collateral” ligamentous tear (recognizable on clinical or ultrasound examination) increases the probability of ACL tear. This could be very useful in our country where MRI is not yet attainable for the majority of the population.