Table 1.

Studies that have adopted CoD tests to assess knee function following ACLr

Study	Subjects/period of assessment	Aim	Testing protocol/measurement method	Main results
Clark et al. [19]	Ten male athletes from Division I University with ACLr; 18–30 years; assessed after full clearance to sports participation (> 9-month post-surgery)	Comparison between involved and uninvolved limb	90° cutting; eight-camera 3-D motion analysis	A 5° measurement of error off of neutral was utilized as reference to indicate meaningful kinematics changes. 80% of the athletes demonstrated significant knee valgus (> 5°) of the involved knee during the cutting task. However, significant knee valgus (> 5°) was also observed in the uninvolved knee for 60% of the athletes. This results question whether returning an involved limb to the standard of the uninvolved limb should be considered as gold standard
King et al. [41]	156 male athletes with ACLr from multidirectional sports and 62 healthy control; 18–35 years; assessed 9-month after surgery	Examine differences in asymmetry of biomechanical and performance variables during PLA and UNP CoD testing between ACLr athletes and matched healthy cohort	90° cutting; eight-camera motion analysis system, two force platforms and light timing system	There was a significant difference in asymmetry of CoD times between groups for both PLA (p = 0.004) and UNP (p = 0.008) conditions, with greater asymmetry in the ACLr group compared with healthy matched control; however, the magnitude of the difference had a small effect size (0.4). During PLA CoD, ACLr group exhibited greater asymmetry in vertical, medial and posterior GRF as well as greater asymmetry for hip abduction moment after initial contact compared to healthy control group. During UNP CoD, ACLr group exhibited greater asymmetry in vertical and medial GRF as well as knee flexion angle compared to healthy control group. However, there was greater asymmetry in the healthy control group for trunk-on-pelvis flexion angle. The ACLr group was more asymmetrical compared to healthy counterparts, suggesting incomplete restoration of normal movement 9 months after ACLr
King et al. [42]	156 male athletes (Gaelic Football, Soccer, Hurling, Rugby) with ACLr; 18–35 years; assessed 9-month after surgery	Comparison between involved and uninvolved limb and between PLA and UNP conditions	90° cutting; eight-camera motion analysis system, two force platforms and light timing system	Involved limb exhibited different biomechanical responses (e.g., less knee flexion/extension moment, knee internal/external rotation moments) compared to uninvolved limb during both PLA and UNP condition. Unplanned CoD elicited less contralateral pelvis rotation, distance from center of mass to the ankle in frontal plane, posterior GRF and greater hip abduction compared to PLA. No changes in time to complete the CoD testing were observed for both involved and uninvolved limb and between PLA and UNP conditions, suggesting that performance-based criteria may not be the most sensitive criterion to discharge ACLr patients back to sports participation
Kyritsis et al. [44]	158 male professional athletes (Football, Handball, others) with ACLr; 22.0 ± 5.0 years; assessed at the end of rehabilitation prior to discharge	Evaluate discharge criteria and its association with ACL re-injury after RTS	Standardized t test amongst others (e.g., strength, hop); Stopwatch	A cut-off of < 11 s during t test was established as CoD criterion to discharge patients back to sports participation. There was no association with future injury risk as an independent test but formed part of a testing battery where completion resulted in lower re-injury risk
Jang et al. [34]	67 male athletes (Football, Basketball, Volleyball, others) with ACLr. RTS (n = 51, 21.9 ± 4.0 years) and non-RTS (n = 16, 21.8 ± 3.5 years); assessed at 6, 12, 24, and 36-month after surgery	Comparison between RTS and non-RTS groups	Co-contraction, Carioca, and Shuttle run tests; Stopwatch	The RTS group exhibited higher performance in co-contraction (RTS = 14.2 ± 1.4 vs. non-RTS = 15.8 ± 2.2, p = 0.010) and carioca (RTS = 8.5 ± 1.5 vs. non-RTS = 9.9 ± 2.6, p = 0.045) tests compared to non-RTS counterparts, suggesting that these tests may assess rotational stability on the ACLr knee. No differences in performance (RTS = 7.5 ± 0.5 vs. non-RTS = 7.8 ± 1.2, p = 0.607) was found between groups for the shuttle run test
Pollard et al. [67]	20 female soccer players with ACLr (n = 10, 23.2 ± 3.4 years) and Healthy control (n = 10, 21.0 ± 1.2 years); assessed 12-month after surgery	Comparison between ACLr and control groups	45° sidestep cutting; Reflective markers (14-mm spheres) with 3-D motion analysis system and floor-embedded force platform	The ACLr players exhibited increased lower extremity variability during the cutting task as compared with the healthy counterparts in the following couplings: hip rotation/knee abduction-adduction (ACLr = 27.2° ± 11.5° vs. control = 19.7° ± 6.8°, p = 0.04); hip flexion-extension/knee abduction-adduction (ACLr = 26.0° ± 13.3° vs. control = 18.6° ± 5.3°, p = 0.05); knee abduction-adduction/knee flexion-extension (ACLr = 13.5° ± 5.7° vs. control = 7.3° ± 2.7°, p < 0.01); and knee abduction-adduction/knee rotation (ACLr = 26.4° ± 10.8° vs. control = 19.3° ± 4.5°, p = 0.03). Increased movement variability during CoD task is likely reflective of altered neuromuscular control as a result of ACLr
Stearns et al. [78]	24 female soccer players with ACLr (n = 12, 23.7 ± 1.9 years) and Healthy control (n=12, 21.3 ± 1.2 years); assessed 12-month after surgery	Comparison between ACLr and control groups	Sidestep cutting; 8-camera 3D motion analysis system and floor-embedded force platform	The ACLr group exhibited increased mean knee abduction angles (ACLr = 3.8° vs. control = 1.8°, p = 0.03), and peak knee adductor moments (ACLr = 1.33 N.m/kg vs. control = 0.80 N.m/kg, p = 0.004) compared with healthy matched control, suggesting higher risk of re-injury upon RTS participation
Kong et al. [43]	60 male patients with ACLr (n = 30, 23.4 ± 3.1 years) and Healthy control (n=30, 24.7 ± 2.1 years); assessed 6-month after surgery	Reliability in the heathy group and correlation between tests in ACLR group	Co-contraction, Carioca, and Shuttle run tests; Stopwatch	High test–retest correlation values were found for co-contraction (r = 0.511, p = 0.025), shuttle run (r = 0.746, p = 0.000), and carioca (r = 0.742, p = 0.000) tests. High correlation was also found between the three tests and strength tests (Isokinetic, hop). ACLr group exhibited lower performance compared to healthy match controls in all the three tests (co-contraction: ACLr = 1.89 ± 2.91 vs. control = 13.3 ± 1.04; shuttle run: ACLr = 7.67 ± 0.97 vs. control = 6.49 ± 0.39; and carioca: ACLr = 9.31 ± 2.43 vs. control = 6.96 ± 0.81)
Myer et al. [56]	36 athletes (Football, Soccer, Basketball, Volleyball) with ACLr (n=18, 16.9 ± 2.1 years) and Heathy control (n=20, 16.9 ± 1.1 years); assessed 12-month after surgery	Test side-to-side symmetry	Modified t-test; Stopwatch	No asymmetries were identified in modified t-test for ACLr group post-surgery. Also LSI was not different between ACLr group and healthy matched control (p > 0.05). This finding opposed the hop tests, which had heightened sensitivity and displayed different LSI scores between groups (ACLr = 92% vs. control = 100%, p < 0.001). The results suggest that the modified version of t-test was not sensitive to identify deficiencies between involved and uninvolved limb
Keays et al. [39]	31 patients with ACLr Male (n = 22) Female (n = 9) 19–38 years (mean = 27 years); assessed 6-month after surgery	Correlation between isokinetic strength test and shuttle run, carioca and side step tests	Shuttle run, Carioca, and Side step tests; Cybex Isokinetic Dynamometer and Stopwatch	Correlation values were found between shuttle run (r = 0.498, p = 0.004), carioca (r = 0.474, p = 0.007), side step (r = 0.528, p = 0.002) tests and isokinetic quadriceps strength test 6 months after ACLr. These post-surgery correlation values were higher compared to pre-surgery. No correlations were found between shuttle run, side step and carioca tests and isokinetic hamstring strength test
Keays et al. [38]	31 patients with ACLr Male (n = 22) Female (n = 9) 19 – 38 years (mean = 27 years); assessed 6-month after surgery	Comparison between involved and uninvolved limb	Shuttle run, Carioca, and Side step tests; Stopwatch	A 10, 17, and 23% improvement in performance were found for Shuttle run (pre-surgery = 9.83 ± 1.46 s vs. post-surgery = 8.86 ± 1.04 s, p < 0.01), Side step (pre-surgery = 13.03 ± 3.48 s vs. post-surgery = 10.86 ± 1.28 s, p < 0.01), and Carioca (pre-surgery = 17.42 ± 5.62 s vs. post-surgery = 13.38 ± 2.23 s, p < 0.001) tests, respectively, 6-month following ACLr
Lephart et al. [48]	41 patients with ACLr Male (n = 32) Female (n = 9) 19–38 years (mean = 22.7 years); assessed within 10–36 months’ post-surgery	Performance comparison between RTS (n = 29) and Non-RTS (n = 12) groups	Co-contraction, Carioca, and Shuttle run tests; Stopwatch	The RTS group exhibited better performance during the shuttle run (7.45 ± 8.2 s vs. non-RTS group = 9.67 ± 3.18 s, p < 0.01), carioca (8.54 ± 1.93 s vs. non-RTS group = 17.31 ± 14.33 s, p < 0.01), and co-contraction (14.96 ± 4.48 s vs. non-RTS group = 20.70 ± 12.42 s, p < 0.05) tests compared to non-RTS counterparts
Tibone et al. [80]	11 patients with ACLr Male (n = 10) Female (n = 1) 18–45 years (mean = 25.5 years); assessed 2-years after surgery	Comparison between involved and uninvolved limb	Straight cut and Cross cut maneuvers; Force plate platform	No differences in cutting index were found between involved and uninvolved limb for both CoD tasks (straight cut: involved = 1348.7 ± 1573.69 vs. uninvolved = 1898.8 ± 2633.65, p = NS; cross cut: involved = 440.5 ± 476.04 vs. uninvolved = 421.7 ± 279.01, p = NS), suggesting mechanical responses were normalized 2-years after surgery

ACL anterior cruciate ligament, ACLr anterior cruciate ligament reconstruction, n number, CoD change of direction, PLA planned, UNP unplanned, RTS return to sport, LSI limb symmetry index, NS non-significant