The effect of meniscal repair on strength deficits 6 months after ACL reconstruction

Hence, the goal of this study was to analyze the effect of meniscal repair on the strength outcomes 6 months post-ACLR. Second, we performed a subgroup analysis to differentiate the potential outcomes according to the location of the meniscal repair. We hypothesized that the strength deficits would be more pronounced in patients undergoing ACLR with meniscal repair when compared to isolated ACLR.

We screened the medical records of 221 patients that were scheduled for ACL reconstruction. Inclusion criteria for the meniscal repair group were unilateral ACLR with meniscal repair in the same session. The matched control group had undergone unilateral ACLR without meniscal repair. Patients undergoing partial meniscectomy were also included in the control group. Exclusion criteria for both groups were second-stage revision, additional cartilage procedures (microfracturing, matrix associated chondrogenesis) or osteotomies performed on either leg. Furthermore, we excluded all patients that had suffered relevant injuries to either leg like contralateral ACL ruptures and previous tendon or muscular injuries of the lower limbs. Associated treatments like partial meniscectomy or superficial chondroplasty with no effect on the postoperative proceedings were not specifically recorded. Figure 1 summarizes patient recruitment in a flowchart.

A total of 61 patients met our inclusion criteria for the meniscal repair group (ACLR + MR). For the control group, n = 122 was chosen as a propensity-matched control group (ACLR) which was matched for the choice of graft, age decade, sex, height and revision ACLR. This matching resulted in two cohorts with a balanced distribution of covariates (Table 1).

The meniscal repair group was further divided according to the location of meniscal repair, with n = 30 medial meniscal repair (MM), n = 19 lateral meniscal repair (LM) and n = 12 meniscal repair in both compartments (BM) (see below).

Of the 30 patients undergoing medial meniscal repair, 24 of these lesions were in the posterior horn, two bucket-handle lesions and four lesions were primarily in the pars intermedia. Of the 19 patients that had a repair of the lateral meniscus (LM), 14 were in the posterior part, 3 in the pars intermedia, 1 in the anterior horn and 1 meniscal root repair. Of the 12 patients that were repaired on the lateral and medial meniscus, 7 had both lesions in the posterior horn, 3 were not specified separately and 2 underwent repair of a lateral root tear combined with a medial posterior horn repair.

For ACLR, we used a proximal extra-cortical fixation (Endobutton CL Ultra, Smith&Nephew, London, UK) and a tibial hybrid fixation using a bioresorbable interference screw and additionally extra-cortical fixation with the femoral tunnel drilled via the anteromedial portal.

The meniscal repairs were performed as follows: posterior horn repairs were performed using an all-inside technique (FastFix, Smith&Nephew, London, UK), the anterior lesion was repaired using an outside-in technique and the three root tears were arthroscopically reconstructed via an additional transtibial drilling with extracortical button fixation (EndoButton CL Ultra, Smith&Nephew, London, UK). Repairs of bucket-handle lesions were performed combining all-inside techniques (FastFix, Smith&Nephew, London, UK), and inside-out techniques using non-resorbable sutures (PDS 2-0).

The postoperative rehabilitation scheme was highly standardized for all patients where isolated ACLR was performed; in these patients, immediate full weight-bearing was allowed. The knee flexion angle was limited at 90° for 2 weeks and no knee-orthosis was used.

In those patients undergoing an isolated medial meniscus repair, the passive flexion limit was kept at 90° for 6 weeks; however, immediate full weight-bearing was allowed while keeping the leg in full extension and wearing external bracing during mobilization. Only the patients after a bucket-handle repair had partial weight-bearing (15 kg) for 3 weeks.

In lateral meniscal repair, flexion angle was limited according to the location and severity of the tear; partial weight-bearing was recommended for 3 weeks with a flexion limit of 60° for 3 weeks and 90° for another 3 weeks. In the three cases of meniscal root fixation, no weight-bearing was allowed for 6 weeks with passive flexion angles of 60–90° during that time.

Generally, after the first 6 weeks, the progression within the individual rehabilitation scheme was criterion based [23]. One key factor in allowing a progressive weight-bearing was the focused activation of the quadriceps muscle to allow active anterior–posterior stabilization. Furthermore, in cases of postoperative flexion limit, a gradual increase using continuous passive motion machines was recommended until reaching 90° of knee flexion. Before restarting running activity, an adequate stabilization of a single leg stance was required. Running activity was initially supported using an anti-gravity treadmill; generally, most patients returned to running about 4–5 months postoperatively.

The functional testing was performed preoperatively and on average 26 weeks post-surgery. However, those patients undergoing surgery within the first few days after the accident, suffering from meniscal impingement or other reasons of limited preoperative ROM like bucket-handle lesions, did not perform preoperative isokinetic strength measurements. The modalities of the strength measurements used in this study are as previously described and in accordance with the current recommendations in the literature [24]. For all strength measurements, we used an isokinetic dynamometer (Humac Norm, CSMi, Stoughton, USA).

Concentric peak torque in flexion and extension was measured as the average of five repetitions at a 60°/sec dynamometer speed. Prior to strength assessments, three submaximal trials were applied for familiarization. Isokinetic testing was completed with maximal effort and verbal encouragement in concentric–concentric mode. During strength assessment, patients were sitting upright, upper body fixed, hands at the grips, while the leg was tightly fixed at the thigh with the lever arm positioned at two-thirds of the lower leg.

Missing data were explored according to their pattern and cause [25]. The mechanism behind missing data followed a missing completely at random pattern. In logistic regression analyses, predictors for missingness were determined based on demographic and clinical characteristics and those predictors found significant were used to estimate missing data in multiple imputations. All statistical analyses were run as complete case analyses and then contrasted in a sensitivity analysis with multiple imputations of missing data [25].

Prior to statistical analyses, assumptions for independent and dependent samples, Student’s t tests as well as repeated measures analysis of variance to compare outcomes in operated and non-affected limbs in each group and to compare outcomes over time between groups were tested. The presence of normal distributions and the amount of outliers in outcomes were checked using data exploration techniques. To remedy problems with assumptions, outlying observations were shifted to the respective lower and upper ends of 1.5 times the interquartile range to truncate their influence on the data [26].

Two main analyses were run: one to compare outcomes between the ACLR and the matched control group over time, a repeated measures analysis of variance (rmANOVA) was conducted with a main factor for group (ACLR vs. matched control group) and two-level factor time (pre/post) for each outcome, respectively. Mauchly’s test of sphericity was used to check assumptions with the Greenhouse–Geisser correction employed if the assumption of sphericity was violated. The level of significance was defined at p < 0.05. In addition to statistical significance, effect sizes eta squared (η²) and percentage change (observed difference to the total amount of difference over time) were calculated for the pairwise comparisons of the repeated measure factor time. Effect sizes were interpreted following Cohen [27] as small: 0.01, medium: 0.06, and large: 0.12. A second analysis was run comparing strength outcomes between operated and uninjured limbs in each ACLR repair group (medial, lateral, both medial and lateral and no repair group) using dependent sample Student’s t tests. Due to the high number of statistical tests, statistically significant p values were Bonferroni corrected to a p value of p < 0.002.

Due to the retrospective nature of the study, an a priori sample size calculation was not possible. However, we calculated the maximum effect sizes that could be obtained given the data that were available to estimate type II error. For the first analysis using rm-ANOVA, a power of 0.8 and an alpha-error of 0.05 at a medium correlation of 0.5 between the repeated measurements would yield effect sizes of partial η² = 0.02 (f = 0.13). For the second analysis, we performed a sample size calculation for a matched pair t test assuming an alpha-error of 0.05, a power of 0.8 and an effect size of 0.5 which lead to a minimum group size of n = 34.

Statistical analysis was conducted using the Statistical Package for the Social Sciences (SPSS) Version 24 [28] and “R” [29], sample size and sensitivity analyses were conducted using G*Power v. 3.9.1.4. Graphical display was performed using Veusz (Veusz v. 3.0.1).

Table 2 presents the calculations for the between-group analyses and the effect size of changes from pre- to post-surgery. Over the course of rehabilitation, all absolute strength values of the operated limb improved significantly (p < 0.05). The meniscal repair group had a higher preoperative deficit in knee extension strength when compared to controls (p = 0.07). Thus, the improvement during rehabilitation was greater in this group than in controls (19% vs. 5%). For the control group, also the non-affected limb showed significant improvements over time in extension (p = 0.04) and flexion (p = 0.03) strength, whereas the meniscal repair group’s healthy leg did not change. The limb symmetry for extension strength improved significantly in both groups (CON pre 80% to post 85% and MEN pre 63% to post 82%, see Fig. 2). The absolute strength as well as the leg symmetry for knee flexion strength (see Fig. 3) showed comparable values for meniscal repair and CON. There were no significant group x time interactions as shown in Table 2. Also, H/Q-ratio did not change over time and effect sizes were negligible.

At 6 months post-surgery, the cross-sectional analysis of the subgroups for meniscal repair revealed several differences between the different locations of meniscal repair as shown in Table 3. Generally, all groups still show a relevant side-to-side deficit, where the operated leg achieves lower values for extension and flexion strength.

Overall MM repair showed higher limb symmetry in flexion (89%) and extension (82%) strength when compared to LM (81% and 76% resp.) or BM (86% and 82% resp.), but significance was not reached between the subgroups. Also, the values of MM were comparable to the values observed for CON (88% and 85%) and no significant differences between the groups were found. Lateral meniscal repair showed the overall lowest values for limb symmetry while the absolute strength values for the non-affected limb were the highest. Across all groups, the H/Q ratio was higher for the operated limb when compared to the non-affected limb.

Those patients undergoing meniscal repair showed inferior strength and leg symmetry at the time of surgery when compared to controls. This is in line with the literature where patients undergoing meniscal intervention had shown inferior preoperative function and performance [11, 14, 42, 43]. Possibly, the greater amount of damaged tissue causes a more severe arthrogenic inhibition of the periarticular muscles [44, 45]. This inhibition as well as local factors such as pain, swelling and inflammation may disappear once the meniscal integrity is restored [17]. Consequently, the meniscal repair group did not show a pronounced deficit in extension strength at the postoperative testing. However, this somewhat contradicts earlier findings, where it was shown that preoperative quadriceps strength correlates to postoperative strength recovery [22, 42]. Furthermore, it was suggested that the partial weight-bearing during rehabilitation of meniscal repairs reduces quadriceps strength [11, 17, 46]. However, our results indicate that it is possible for the athlete to regain quadriceps strength after ACLR + MR within the same time period as isolated ACLR. Interestingly, limb symmetry in knee flexion strength was generally higher than in quadriceps strength and effect sizes were stronger. Despite harvesting hamstring tendons in the majority of the patients (72–78%) the recovery of knee flexion strength was more symmetric at 6 months postoperatively compared to knee extension strength. This may be due to the finding that arthrogenic muscle inhibition primarily affects the quadriceps muscle which causes a prolonged strength deficit in knee extension compared to flexor strength [45, 47]. All other established factors affecting quadriceps strength after ACLR like age and gender were equally distributed across the groups.

The importance of achieving an adequate limb symmetry in strength before returning to the field is well accepted [6, 24, 30]. Return-to-play criteria mostly require a recovery of > 85% of the healthy limb’s strength [3, 6, 20]. In our subgroup analysis, we were able to show that only medial meniscal repair (MM 87.2%) fulfilled this criterion, while repairs of the lateral meniscus (LM 75.7%) and repair of both menisci (BM 82.3%) still exhibited a greater deficit. Contrarily, a recent analysis revealed that medial as well as lateral repair is associated with reduced quadriceps strength at 6 months postoperatively, while knee flexion strength was not significantly reduced [46]. Cristiani et al. attributed this to the early restriction in range of motion and partial weight-bearing [46]. However, scientific evidence on the effect of these limitations is scarce.

Since strength asymmetry has been linked to subjective knee function post ACLR [48], the strength deficit observed in our study may explain why lateral meniscal repairs also exhibit worse subjective function at 6 months post-surgery [11]. In previous studies, a comparable subjective function after meniscal repair was achieved as late as 1 and 2 years post-surgery [11, 14]. A limb symmetry below the cut-off value of 85% was found in the LM and BM groups, which must be considered clinically relevant [6]. Thus, lateral meniscal repair may require more time before successfully passing this return-to-play criterion. It may be suggested that this is due to the initial restrictions (range of motion and weight-bearing), which may have a persisting negative effect on strength recovery. However, the results of this subgroup analysis are of explorative character and they do not allow for a causative interpretation.

Furthermore, it needs to be stated, that the interindividual variability was relatively high; hence, none of the differences between the subgroups reached statistical significance when adapted for multiple testing. This underlines the observation that the individual recovery from ACLR varies greatly [30]. In line with current publications, this supports a criterion-based rehabilitation over an isolated time-based approach [30, 49, 50].

Limitations of the study include the relatively high rate of missing data from pre-operative isokinetic analysis, which is owed to the fact that patients with effusion, pain or unstable meniscal lesions were excluded from preoperative functional analysis. Furthermore, sample size calculation suggested that a type two error might be present in the subgroup analysis and requires a careful interpretation especially of the results in LM and BM.