We aimed to determine differences in functional leg performance in hops, knee proprioception, and dynamic postural control within and between patients 2 years after InternalBrace™-augmented ACL repair, patients 2 years after ACL-R, and controls using statistical, clinical, and methodological thresholds, as well as to compare the number of participants achieving normal leg symmetry and normal leg performance between groups. According to the number of exceeded predefined thresholds (see method section), we observed no-to-moderate leg differences within ACL-IB and within ACL-R. We noted no-to-small differences within the legs of controls, no leg differences between ACL-IB and ACL-R (involved legs), no-to-small leg differences between ACL-IB and controls, and no leg differences between ACL-R and controls in functional performance including hops, knee proprioception, and dynamic postural control 2 years after surgery. Moreover, in each patient group, two patients were able to hop 40 cm only with their uninvolved leg but not with their involved leg, and in SLH, fewer patients (ACL-IB and ACL-R) than controls reached leg symmetry and leg performance of the non-dominant leg of healthy controls.

Differences within and between ACL-IB, ACL-R, and healthy controlsHop performance

In agreement with our study, Leister et al. [25] reported no clinically relevant difference in the LSI of ACL-IB in SLH, but in contrast to our study, a clinically relevant difference in LSI of matched ACL-R (hamstring autografts) in SLH 13 months postoperatively. Also similarly to our results, these authors observed a clinically relevant difference in LSI of less than 90% in ACL-IB and ACL-R in SH [25]. Values in this range have been associated with lower patient-reported knee function and degenerative changes [33] or were used as no return to sports criteria [12]. Nevertheless, no significant or relevant differences were found between the matched patient groups in either this study or ours. Moreover, a comparable amount of ACL-IB and ACL-R (6.9% versus 7.4%) passed the hop pretests (jumping laterally over 40 cm with one leg) with their uninvolved leg only, but not with their involved leg. This corresponds to an actual deficit of 100% in the involved leg compared with the uninvolved leg and controls (all controls completed all pretests). While Leister et al. [25] did not report such high deficits in the involved leg for patients after ACL-IB or ACL-R, other authors reported such high deficits for 4 of 81 (4.9%) patients 1 year after ACL-R (hamstring autografts) in SH [33]. Unfortunately, the reasons for not jumping were not described [33]. Our finding that some patients were unable to jump with at least one leg could be explained by the lower activity level (median, TAS 4) compared with the other study investigating hop performance in ACL-IB and matched ACL-R (median, TAS 6.0 [25]).

Compared with controls, only ACL-IB achieved a clinically relevant but statistically nonsignificant threshold in SH, whereas patients after ACL-R just missed these thresholds, with only one more hop. These differences could be a consequence of the possibly underpowered ACL-R study group compared with ACL-IB or control group. Furthermore, if our findings on normal leg symmetry and performance are taken into account, it appears that both groups of patients (also ACL-R) have deficits not only in the involved leg compared with the uninvolved leg, but also with the matched control legs. Consistent with this assumption, significantly lower leg symmetry in LSI in SLH has also been reported for both ACL groups compared with LSI of controls in the literature [25].

In agreement with the literature, we consider the observed failures in hop pretests, small-to-moderate leg differences within ACL-IB, small differences within ACL-R, and small differences compared with controls as relevant deficits with functionally compromised hop performance 2 years after ACL surgery.

Proprioception

We observed small differences within ACL-IB, small-to-moderate differences within ACL-R, and no-to-small differences in controls in the JPS test. Higher reproduction errors have been reported in the literature for active and passive JPS after ACL-R (with different types of autografts) compared with controls, indicating impaired proprioceptive performance up to 2 years postoperatively [8]. However, the results in active–active JPS 9–24 months after ACL-R with only hamstring tendon autografts are controversial [9, 42, 47]. The differences between the studies that found a significant difference between ACL-R and controls [42, 47] and our study results without significance after ACL-R compared with controls may be explained by methodological differences or different activity levels of patients [39, 46]. Since JPS reproduction error describes very small values with a very high relative variability, and only 38–52% of controls achieved an LSI ≥ 90% and a leg performance ≥ 90% of their average performance in the JPS, we consider the use of the LSI, including a 90% cut-off, as clinically not useful. Clinically relevant thresholds for the absolute achieved reproduction error of a leg (< 5°) have been described in the literature [38]. The average of all legs in our study was well below this value, including both legs of our patients after ACL-R, despite significant differences in JPS error at 30° knee flexion. Moreover, the SSDs were comparable to those reported in a review paper [10]. These authors and others [38] questioned whether the reproduction error or its SSDs in the JPS are relevant to knee function and performance. Therefore, we do not consider the observed leg differences in patients and controls to be relevant. These results suggest that there are no relevant differences in proprioceptive function after ACL-IB and ACL-R, and they neither support nor refute the suggestion that proprioceptive function is preserved after ACL-IB [50], nor that proprioception is impaired after ACL-R using hamstring tendon autografts. To the best of our knowledge, the (negative or positive) influence of the synthetic InternalBrace™ on the mechanosensory receptors in the ACL or on the transmission of stimuli has not been investigated and may be an area for further research.

Nevertheless, despite the lack of differences in assessment of methodological thresholds across all comparisons and without relevant differences in JPS, we must consider that the JPS test alone may not be sensitive enough to detect proprioceptive deficits. Proprioception is presumably also facilitated by proprioceptors in other joint tissues or surrounding muscles, and the JPS reflects only one aspect of proprioception [40]. Moreover, weight-bearing assessment of proprioception may be more functional [46] and could explain the possible decreased performance in mediolateral hops despite good proprioceptive results. Hence, further investigation of proprioception in ACL-IB compared with ACL-R and controls is warranted.

Dynamic postural control

The similar YB performance in both legs after ACL-R is consistent with recent literature 2 to 3 years after ACL-R surgery [31]. Although statistically significant, SSD in YB posteromedial and composite scores after ACL-IB were well above clinical and below methodological thresholds. In addition, conflicting results regarding an association between asymmetry in the YB posteromedial score of ≥ 4 cm and asymmetry in the composite score of ≥ 12 cm with future injuries were reported in one review [36]. The absolute asymmetry in our ACL-IB group was lower than these values (for comparison, absolute SSD: posteromedial: 2.1 cm; composite: 3.8 cm). Therefore, we consider the observed small leg differences within ACL-IB to be irrelevant. In contrast to our results, other authors reported lower anterior YB score in the involved leg compared with controls after ACL-R (patellar or hamstring tendons) [31]. This was explained by the inability to resist anterior tibial translation, possibly due to higher quadriceps activity in this direction [21] or the inability to compensate for this translation by presumably weakened hamstring muscles after ACL-R with hamstring tendon autografts [35]. Although we could not confirm these results, the largest discrepancy between both ACL patient groups and control subjects in this YB direction in comparison of number of participants achieving normal leg symmetry and performance may indicate difficulties for patients. To achieve comparable anterior distance as the control subjects, our patients may have used different movement strategies, as reported in patients 2.6 years after ACL-R when performing the anterior reach distance test of the modified Star Excursion Balance test [5]. Hence, examining not only performance, but also kinematics during such a complex task may provide better insight into potentially different mechanisms or strategies between ACL-IB and ACL-R.

Normal leg symmetry and performance

Across all tasks, the lowest percentage of patients with leg symmetry and normal leg performance were observed in hops, with lower percentage in SH than SLH in ACL-IB and controls. This has also been reported in the literature when only leg symmetry was studied in patients 1 year after ACL-R [33]. In that study, only 36% of patients achieved 90% of the performance of the contralateral leg in SH, whereas 62% achieved this level in SLH [33]. These findings reflect the higher complexity of lateral jumps, which may be supportive or even more informative for identifying deficits after ACL surgery. The finding in our study that in SLH fewer patients than controls reached leg symmetry and fewer patients achieved leg performance of the non-dominant leg in controls suggests incomplete rehabilitation in hops 2 years after ACL surgery, regardless of ACL-IB or ACL-R.

Overall, none of the differences examined were below the respective SDC of any test, which calls into question the differences we considered relevant based on reliability and detectability. Further studies defining clear criteria for relevance and methodological verifiability in the specific ACL-R and ACL-IB populations are needed. In addition, complementary analyses of movement execution (representing movement quality) would be useful to provide further information not only on the extent of performance (quantity), but on how performance was achieved.

Strengths and limitations

The main strengths of our study were that all groups were matched for age and sex and did not differ in their knee-specific activity level (TAS). Furthermore, the number and type of concomitant injuries and surgeries in ACL-IB and ACL-R were comparable between groups, as reported in another manuscript related to the umbrella project [30]. Consequently, differences between patients can mainly be attributed to the surgery and not to concomitant injuries or surgeries. For the first time, we provided data on functional leg performance (in terms of proprioception and dynamic postural control) after ACL-IB, and assessed the performance outcomes using statistical, clinical, and methodological thresholds. A limitation of our study is that we excluded patients from the hop test analysis if they failed the pretest to obtain consistent data for our analysis within and between groups. This may have resulted in an overestimation of patients’ hop performances and masked the presence of positive results on the other defined hop parameters thresholds. Moreover, the SDCs were derived from the literature (e.g., healthy population) and hence may not be representative for the patient populations studied. The exclusive inclusion of patients with proximal ACL tears in ACL-IB and all types of ruptures in ACL-R may have biased our results. Because of the moderate knee-related activity level of our participants and the inclusion criteria for patients, the results may not be generalizable to highly active or professional athletes or patients after ACL reconstruction with other tendon grafts. Finally, although patients completed standard physical therapy, the duration and adherence to therapy were not recorded, which may have influenced our results and the activity level of participants.

Clinical implications

Our results show no differences (assumed to be relevant) in functional leg performance in hops, proprioception, and dynamic postural stability between ACL-IB for proximal ruptures and ACL-R with hamstring tendon autografts. The comparable functional performance outcome, but less invasive ACL repair surgery, highlights the potential of augmented ACL repair next to ACL-R for a specific subgroup of patients represented in our study (e.g., proximal ruptures, adults with moderate knee-specific activity level, for example, TAS of 4). The presumed advantage of augmented ACL repair providing comparable functional performance with preserved knee structures (i.e., native ligament and muscle–tendon complex) must be weighed against the reported higher rerupture rates compared with the gold standard ACL-R [7]. However, with careful patient selection (i.e., patients aged ≥ 21 years [7], TAS score ≥ 7 [15], corresponding to recreational activities such as soccer, rugby, ice hockey, and basketball, or competitive activities such as skiing and gymnastics) it might be possible to reduce the rerupture rates. Apart from this, in our study, neither ACL-IB nor ACL-R achieved the full functional performance level of healthy controls 2 years postoperatively, questioning whether ACL-IB or ACL-R surgery can restore full lower leg function after ACL tear.