Rehabilitation After Anterior Cruciate Ligament Reconstruction

After primary anterior cruciate ligament (ACL) reconstruction, rehabilitation advances by function and symptoms; when returning to high-load or contact sport, a minimum time threshold and functional criteria must both be met.

1. Approximate Timeline of Recovery

From surgery to return to contact sport, most guidelines give a window of 9 to 12 months. Within that period one must successively address swelling, extension, quadriceps activation, normal gait, strength recovery, running, jumping and cutting, and finally sport-specific training and competition. Guidelines agree that progression to the next stage should be criterion-based; how quickly patients who advance by criteria actually enter the next stage is rarely reported in public data.

Brinlee et al. (2022), from the University of Delaware physical therapy clinic, compiled time-to-milestone data for 75 patients treated there between 2013 and 2017. Of these, 57 had no meniscal suture (a footnote to original Table 1 notes that this group may have included partial meniscectomy); 18 had concomitant meniscal repair. Original Table 1 lists mean weeks to each milestone by meniscal management, with quadriceps strength further subdivided by graft:

Milestone	No meniscal suture (n=57; may include partial meniscectomy)	Concomitant meniscal repair (n=18)
Effusion reduced to trace	Week 13	Week 13
Active full extension	Week 2	Week 1
Passive full extension	Week 1	Week 2
KOS-ADLS >80	Week 13	Week 11
Quadriceps strength index >80%	BTB graft weeks 12–14; hamstring or allograft week 9	BTB weeks 16–20; hamstring or allograft week 12
Quadriceps strength index >90%	BTB week 30; hamstring or allograft week 15	BTB weeks 26–35; hamstring or allograft week 13

The weeks in the table are means; the original report gave neither standard deviations nor ranges, and did not report how many patients never reached each milestone. Patellar-tendon grafts numbered only 19 in the non-suture group and 5 in the repair group. The table therefore reflects mean time-to-milestone in one clinic that followed a fixed protocol; it cannot be used to judge whether any individual patient is recovering fast or slowly. The most useful figures are those for strength and effusion: in the non-suture group, quadriceps strength index reached 80% at a mean of 9 to 14 weeks, and with a patellar-tendon graft reached 90% at a mean of week 30; both groups needed a mean of 13 weeks for effusion to fall to trace.

Under that clinic’s protocol, running begins at 12–16 weeks after surgery; return to sport is no earlier than 9 months (12 months for allograft), and requires quadriceps strength index, hop testing, and KOS-ADLS all above 90%, with ACL-RSI psychological readiness above 80%.

2. Early Postoperative Period: Extension, Swelling Control, and Quadriceps Activation

Extension

The first priority after surgery is to restore and maintain full extension, including any hyperextension present on the uninjured side. Noll et al. (2015) listed in their introduction the consequences of extension deficit: patients walk with a flexed-knee gait, joint loading and patellofemoral pain increase, and the quadriceps are weaker. Guidelines set the early extension goal as “symmetry with the uninjured side.” The International Knee Documentation Committee (IKDC) physical-examination form grades extension deficit by the difference from the uninjured side: grade A <3°, grade B 3–5°, grade C 6–10°, grade D >10° (Wackerle et al., 2026 Panther expert opinion, Part 1, p. 523).

The relationship between early and later extension deficit was observed prospectively by Noll et al. in 74 patients. Mean age was 17.7 years; 89.3% received a patellar-tendon graft; meniscal repair was excluded. The side-to-side extension deficit narrowed over time: 9.6°±5.7° at the initial visit, 4.7°±3.2° at 4 weeks, 2.5°±2.5° at 8 weeks, and 2.0°±2.6° at 12 weeks (original Table 2). The correlation between 4-week and 12-week extension was r=0.639. In regression analysis, the 4- and 8-week values together explained 45.7% of the variance at 12 weeks, of which the 4-week value accounted for 25.9%; adding the initial-visit value raised the proportion explained to 61.1%.

The authors therefore set reaching “nearly normal” by 4 weeks as a target—that is, within 3–5° of the uninjured side, corresponding to IKDC grade B above; reaching grade A (normal) likewise meets the target. This is a correlational finding: it indicates that deficit at 4 weeks warrants careful attention, but it does not prove that failure to meet the target at 4 weeks leaves a permanent deficit. The cohort was young and almost exclusively received patellar-tendon grafts; recovery curves in adults and with hamstring grafts may differ. In original Table 3, which tabulated cases by IKDC grade, the grades at 4 weeks summed to only 35 and those at 12 weeks to 61, both short of 74, without explanation in the paper; this chapter therefore does not use that table to estimate the proportion in each grade and relies only on the correlation and regression results already reported above.

Consequences of persistent extension deficit at 4 weeks are quantified in the retrospective series of Mauro et al. (2008; based on the abstract). Among 229 reconstructions between 1995 and 2000, 58 patients (25.3%) had extension ≥5° short of the uninjured side at 4 weeks, or required repeat arthroscopy or a drop-out cast for correction. This extension-limited outcome was statistically associated with preoperative extension, time from injury to surgery, and use of an autograft; the abstract did not give the direction or magnitude of the associations. Twenty-eight patients (12.2%) underwent arthroscopic lysis of adhesions; the authors wrote that “48% of patients who lacked extension at 4 weeks eventually underwent arthroscopic debridement.” At final follow-up, extension ≥5° short of the uninjured side was present in 1.7%. These figures come from one hospital’s own management pathway. They show that a substantial proportion of knees still ≥5° short at 4 weeks eventually undergo further surgery; they do not show that recovery is impossible without debridement.

Practical methods are similar across protocols: supine heel props with the knee hanging free, prone hanging with a light weight on the leg, therapist-applied extension force, combined with icing. The Delaware home program starts in the first postoperative week with heel props plus ice, five times daily for 10 minutes each.

If extension stalls or regresses after 4 weeks, or if a hard block appears at terminal extension, a mechanical cause should be considered first—for example a scar nodule in the anterior notch (cyclops lesion)—and the operating surgeon should reassess. Pain, effusion, and quadriceps inhibition likewise limit extension; their management differs (see Section 13).

Swelling

Joint effusion inhibits quadriceps contraction; reducing swelling is itself part of restoring quadriceps function. The Delaware protocol grades effusion with the sweep test: fluid is first milked upward from the medial gutter, then swept downward from the lateral side while observing whether fluid returns into the medial dimple. Grade 0: no return; trace: a small amount of return; grade 1: no spontaneous return after milking, return on the lateral sweep; grade 2: immediate refill after milking; grade 3: unable to milk away. The protocol records grades 1–3 as 1+, 2+, and 3+; “grade 2” and “2+” below refer to the same grade.

Three accompanying rules apply: do not progress training when effusion exceeds grade 1 (i.e., grade 2 or 3); contact the physician if effusion remains above grade 2 for a prolonged period; if effusion rises suddenly by two grades, or reappears after having been absent, first reduce activity volume and then resume gradually (Brinlee et al., 2022, Table 2; protocol Appendix 2). This grading system comes from the Adams et al. (2012) protocol; interrater reliability has been reported (Sturgill et al., 2009, cited via Fältström et al., 2026); there are no controlled studies of its effect when used as a progression criterion.

Cryocompression devices have randomized-trial support, but the evidence is not strong. The OPTIKNEE overview (Culvenor et al., 2022) concluded that low-certainty evidence from two reviews totaling 11 randomized trials shows that cryocompression in the first 1–2 postoperative days reduces pain and reduces analgesic use for about one week. The only review that performed a meta-analysis estimated an approximately 14% reduction in pain at 48 hours versus no cold therapy (95% CI 12%–17%). Effects of cold therapy on range of motion and quality of life were not shown, and there is no evidence beyond one week. The German Knee Society consensus (Häner et al., 2026) lists cold therapy within 3 days after surgery as an A2 recommendation of moderate certainty—one grade higher than OPTIKNEE. Cold therapy is for early postoperative analgesia; it should not be expected to alter the recovery trajectory.

Continuous passive motion (CPM) is likewise A2 with moderate certainty in the German consensus, but the conclusion is the opposite of that for cold therapy: “likely no meaningful improvement”; OPTIKNEE likewise does not encourage routine use.

For managing post-training soreness, the Delaware protocol provides a “soreness rules” table (Brinlee et al., 2022, Table 3):

Soreness pattern	Action
Soreness during warm-up that persists	Rest 2 days; drop one level
Soreness during warm-up that resolves during warm-up	Stay at current level
Resolves during warm-up, recurs during training	Rest 2 days; drop one level
Soreness the day after training (not muscle soreness)	Rest 1 day; do not progress
No soreness	Advance one level per week

These are empirical rules; no trial has compared their outcomes with other progression schemes.

Quadriceps

The goal in the first postoperative week is active quadriceps contraction with superior patellar glide, and no lag on straight-leg raise. Training includes isometric quadriceps contractions, straight-leg raises, and seated knee extension (90° to 0°), performed multiple times daily. The Delaware home table lists a large isometric volume: 5–10 sets of 100 repetitions daily, each held 5–10 seconds. That volume has not been compared with smaller doses and need not be treated as a mandatory quota; the point is multiple daily sessions with each contraction producing clear superior patellar glide. Most protocols begin neuromuscular electrical stimulation at this stage (see Section 6).

Flexion

Flexion is generally not difficult to restore, and targets are similar across protocols. Panther expert opinion aims for 90° by 2–3 weeks and 120° by 4–6 weeks; the Delaware protocol aims for 0–90° in week 1, 0–115° at 2–3 weeks, and within 10° of the uninjured side by 4–6 weeks. Concomitant meniscal repair imposes additional flexion restrictions (see Section 11).

3. Weight-Bearing, Crutches, and Bracing

Timing of Weight-Bearing

Authoritative documents disagree on when weight-bearing may begin. The Aspetar clinical practice guideline (Kotsifaki et al., 2023) recommends weight-bearing as tolerated from the first postoperative week. That guideline included only randomized trials—140 trials, 5231 patients (70% male; mean age 27.9 years; 94% with patellar-tendon or hamstring autograft). The authors declared no conflicts of interest and no dedicated funding; all recommendations were voted by experts from a single institution, a limitation the authors themselves listed (p. 510).

Item 1 of the 2026 German Knee Society consensus states “partial weight-bearing for 2 weeks after isolated ACL reconstruction,” evidence grade B2, moderate certainty, 90% expert agreement; item 2, “prolonged partial weight-bearing with concomitant injury,” grade C, low certainty, 100% agreement.

Panther expert opinion takes a middle position: early weight-bearing after isolated reconstruction; with concomitant meniscal repair or cartilage surgery, “many surgeons recommend partial weight-bearing for 6 weeks with flexion limited to 90°,” based on surgical experience. The body of that document states “no conflicts of interest,” yet the disclosures list multiple authors’ consulting, research funding, or royalty relationships with Arthrex, Smith & Nephew, AlterG, Bauerfeind, and others—an inconsistency between the two statements.

The German consensus’s rationale for 2 weeks of partial weight-bearing is that several studies observed a tendency toward tunnel enlargement with immediate full weight-bearing. The consensus did not report effect sizes from those studies, nor did it relate tunnel enlargement to laxity or revision rates. The wording of the conclusion for this item in its Table 2 is “may allow safe early weight-bearing without increasing revision”; that wording does not entail a requirement to restrict weight-bearing for two weeks.

Among randomized trials that directly compared immediate with delayed weight-bearing, the most often cited is Tyler et al. (1998). Forty-nine patients with patellar-tendon autografts, operated on by a single surgeon, were randomized to weight-bearing as tolerated immediately or to non-weight-bearing for 2 weeks; the latter group was forbidden to wear a shoe, to enforce non-weight-bearing. Concomitant meniscal repair did not alter allocation. Measurements were obtained preoperatively, at 2 weeks, and at 6–14 months (mean 7.3 months). At follow-up there were 25 in the immediate and 20 in the non-weight-bearing group, with 2 lost to follow-up in each; one woman in the immediate group ruptured the contralateral ACL at 9 months, and her approximately one-year side-to-side data were excluded.

At 2 weeks, 17 of 22 in the non-weight-bearing group and 10 of 27 in the immediate group had extension ≥5° short of the uninjured side (p<0.01); at follow-up the figures were 3/20 and 5/25, with no detected difference. Vastus medialis EMG at 2 weeks was higher in the immediate group (p=0.002); at follow-up both groups approached normal. KT-1000 maximum manual side-to-side difference was 2.4±0.70 mm in the non-weight-bearing group and 2.8±0.5 mm in the immediate group (between-group p=0.70); page 143 of the original states that means are reported with standard error of the mean. Anterior knee pain at follow-up was reported by 7/20 (35%) in the non-weight-bearing group and 2/25 (8%) in the immediate group (p=0.03); Lysholm improvement was also greater in the immediate group (p=0.03).

Randomization sequence and allocation concealment were not described; the sample was small, and follow-up ranged from 6 to 14 months. Group numbers contradict one another across the paper. Methods state only that 49 patients were randomized into two groups, without listing group sizes. On page 144, cases classified by time from injury sum to 22 immediate and 27 non-weight-bearing; the denominators used for 2-week extension on pages 144–145 are instead 27 immediate and 22 non-weight-bearing, and the follow-up counts on page 143 (25 and 20) plus 2 lost per group also yield 27 and 22. The two group sizes are simply swapped in different places; without a flowchart one cannot tell which is wrong. Both groups included patients who had meniscal procedures and preoperative collateral laxity, so this was not strictly isolated ACL reconstruction. With about twenty patients per group, failure to detect a laxity difference does not exclude a laxity risk.

The meta-analysis by Fan et al. (2022) defined “accelerated weight-bearing” as full weight-bearing within 2 weeks after surgery and included 9 studies. Original Figure 7 pooled KT-1000 side-to-side difference from 6 studies—5 randomized trials and the Tajima 2019 prospective cohort (accelerated n=116, delayed n=106): the accelerated group was 0.29 mm higher than the delayed group (95% CI 0.02–0.56 mm), I²=0%, with 55.9% of the weight from Tyler 1998. On that basis the authors wrote in the abstract and conclusion that accelerated weight-bearing “causes greater knee laxity and tunnel enlargement” (tunnel data from 2 studies); they also reported higher IKDC scores in the accelerated group.

That laxity conclusion rests on a verifiable data-entry error and therefore cannot stand. Figure 7 entered Tyler’s 0.70 and 0.5 as standard deviations, whereas page 143 of Tyler states that means are reported with standard error of the mean, and the 2.4±0.70 and 2.8±0.5 on page 145 are standard errors. Once entered as standard deviations, meta-analysis software divides each by the square root of sample size to obtain the standard error of the group mean; because 0.70 and 0.5 were already standard errors, this amounts to dividing a second time and overstates the precision of the Tyler row. Figure 7 computed the difference as delayed minus accelerated, so a negative value means a larger side-to-side difference in the accelerated group; Tyler’s confidence interval was compressed to −0.76 to −0.04 and its weight inflated to 55.9%, whereas Tyler’s own between-group p value was 0.70.

Using the six rows of Figure 7, fixed-effect inverse-variance pooling reproduces the original summary (−0.29, −0.56 to −0.02). Converting only the Tyler row’s standard errors back to standard deviations (0.70×√20≈3.1; 0.5×√25=2.5) and leaving the other five rows unchanged drops Tyler’s weight to 5.6% and changes the pooled difference to 0.16 mm higher in the accelerated group (−0.24 to 0.56), p≈0.43. This is a sensitivity calculation that corrects only one row; the entry of the other five studies was not checked item by item in this chapter. It is not a new meta-analysis, and it cannot be inverted to claim that early weight-bearing is “harmless” for laxity. What it does show is that the statistical significance of the original pooled estimate came from this entry error, and that Figure 7’s pooled value and the judgment that “accelerated weight-bearing causes more laxity” cannot be used.

The interventions pooled in that meta-analysis were also heterogeneous: Luo 2016 compared a body-weight-supported treadmill, Henriksson 2002 compared a brace with a cast, Beynnon 2005 compared 19- versus 32-week rehabilitation programs, and Christensen 2013 compared “early aggressive rehabilitation”—all treated as “accelerated versus delayed.” The text states that laxity included 7 studies; the figure shows only 6. Other outcomes such as IKDC were not re-pooled in this chapter, so one cannot assert that they too contain errors; the review’s study list remains usable. The work was funded by the National Natural Science Foundation of China.

The systematic review by Bafrouei et al. (2026) defined “accelerated rehabilitation” more broadly—early weight-bearing, early motion, brace-free care, and shortened programs all counted—and included 10 studies. That review contains serious errors on several key outcomes; its pooled results and judgments of “clinical meaningfulness” cannot be used.

Basic numbers are inconsistent: samples in original Table 1 sum to 2178, while the text states 2180; the text gives the median publication year of included studies as 2007, whereas Table 1 yields 2016.5.

Laxity pooling. The summary in Figure 3 (mean difference −0.508, 95% CI −0.656 to −0.360) is almost entirely driven by De Carlo 1992: that study was entered with a mean difference of −0.530, standard error 0.007, and Z of −76.8. Against De Carlo’s original Table 7, one-year KT-1000 values were 2.08±0.11 mm versus 2.59±0.17 mm (labeled mean±standard error), a difference of 0.51 mm; approximating the standard error of the difference from the two independent standard errors yields about 0.20 mm—nearly 30 times 0.007. How the calculation went wrong cannot be confirmed from the review; what can be confirmed is that this row does not match the source data. The review authors’ own sensitivity analysis (p. 5) states that after removing De Carlo the laxity difference had p=1.00.

Proprioception pooling. The summary in Figure 9 (0.202, 0.104 to 0.300) likewise gives De Carlo 1992 the greatest weight, yet De Carlo measured only range of motion, isokinetic strength, and KT-1000 and reported no proprioception outcome; that datum does not exist in the source literature.

Source of the minimal clinically important difference. When the review cites a “KT-1000 minimal clinically important difference of 1.5 mm,” the source given is Robert et al. (2009) on the GNRB arthrometer. That paper’s 1.5 mm is a diagnostic threshold for distinguishing partial tears at 134 N (sensitivity 80%, specificity 87%)—a different instrument and a different meaning.

Misplaced citations. On page 13 three citations do not match: [40] is a study of a Persian low-back-pain questionnaire, cited as the rule “stop if pain exceeds 2/10”; [41] is a study of a Persian neck-pain questionnaire, cited as a jump-training review; [36] is a randomized trial after total knee arthroplasty, cited as evidence on sensorimotor pain. In addition, methodological quality was rated by a single author, and the article claims Level I evidence.

Other outcomes were not re-pooled item by item in this chapter, so one cannot assert that they too contain errors; but a review that errs in this way on laxity and proprioception, and that misplaces citations in multiple places, is not suitable support for any specific prescription. Its list of 10 studies remains usable as an index to the primary literature.

Early material on “accelerated rehabilitation” comes from De Carlo et al. (1992) and Shelbourne et al. in the same year. De Carlo et al. compared patients of the same surgical group across two eras: those treated in 1982–1986 were immobilized in a cast at 30° for 2 weeks, non-weight-bearing for 8–10 weeks, and allowed open-chain training; those treated in 1987–1989 were weight-bearing as tolerated, used CPM, emphasized full extension, and performed only closed-chain training. The paper states group sizes of 600 and 1052 in one place and “1508 patients” in another; the data-analysis section notes that 1652 score sheets were obtained from multiple evaluations within one year, counting repeat evaluations of the same patient. 600 plus 1052 equals 1652—matching the number of score sheets but not “1508 patients.” How many patients contributed to each outcome is not itemized in the tables.

At one year, KT-1000 side-to-side difference was 2.08 mm in the accelerated group and 2.59 mm in the traditional group (standard errors 0.11 mm and 0.17 mm; p<0.05); the proportions exceeding 3 mm were 16% and 22%, and range-of-motion recovery was also faster in the accelerated group. This is a retrospective era comparison: fixation, weight-bearing, CPM, and training content all changed between eras, as did surgical technique, so one cannot determine which element was responsible; all grafts were patellar tendon, performed by three surgeons. Shelbourne et al.’s update that year (more than 880 cases) gave 2.1 mm for accelerated rehabilitation and 2.8 mm for those treated before the accelerated protocol. That protocol limited everyday walking and elevated the limb for the first two weeks, allowed crutch-protected weight-bearing as tolerated, and began extension, flexion, and quadriceps training immediately after surgery; return to light competition as early as 2 months—a timeline no current guideline adopts.

These two papers show that early weight-bearing and early motion protocols were not accompanied by looser knees in the cases of that era; they do not prove that early weight-bearing carries no laxity risk, still less that it is better than any specific “two weeks of partial weight-bearing” protocol.

Synthesizing the above, after isolated ACL reconstruction weight-bearing as tolerated under crutch protection may begin in the first week. The directly comparative randomized trial (Tyler) detected no cost in laxity or extension; the immediate group had better quadriceps activation at 2 weeks and less anterior knee pain at follow-up; but the trial was small and underpowered for small differences. Pooled summaries that combine heterogeneous accelerated protocols show no laxity difference after correction of one entry error; that calculation is also not a complete re-analysis.

The German consensus’s two weeks of partial weight-bearing rests on observations of tunnel enlargement, without relating those observations to laxity or revision; Aspetar’s early weight-bearing recommendation rests on laxity and functional results from randomized trials and is graded D. Grades A through D in that guideline denote certainty of evidence from high to very low; grade D means that confidence in the effect estimate for this comparison is very low, which is not contradicted by the evidence coming from randomized trials. Both approaches lie within what the evidence permits; in practice they differ by only two weeks. With concomitant meniscal repair, cartilage surgery, or collateral ligament injury, weight-bearing limits are set by the surgery (see Section 11).

Crutches

The Delaware protocol’s criteria for discontinuing crutches are normal gait, no lag on straight-leg raise, and effusion ≤2+. On the same page, stage milestones write effusion <2+, and the training-monitoring section writes that training is not progressed with 2+ effusion—three statements that are not fully consistent. The safer reading is: when effusion reaches grade 2, crutches should not be discontinued and training volume should not be increased. Panther expert opinion emphasizes that “premature self-discontinuation of crutches often leaves an abnormal gait that persists later”—clinical experience without controlled data, but consistent with most therapists’ observations. Timing of crutch discontinuation should be set by whether walking is limp-free, not by a fixed number of days.

Bracing

Whether a rigid brace is needed after isolated reconstruction has been addressed by a dedicated randomized trial. Schoepp et al. (2023), in a German noninferiority trial, enrolled patients with isolated ACL reconstruction using hamstring autograft, excluding those needing meniscal suture, concomitant other ligament injury, or cartilage repair. One hundred thirty-nine patients were randomized to brace-free (n=71) or braced (n=68) care. Both groups followed the same protocol: a nighttime extension splint for the first postoperative week to prevent extension loss, partial weight-bearing of 20 kg for about 14 days, and full weight-bearing by 4 weeks; the braced group additionally wore a 0-0-90 range-limiting brace for 6 weeks. The primary outcome was IKDC subjective score at one year, with a noninferiority margin of 9 points.

At one year, 114 patients entered the analysis (56 brace-free, 58 braced), yet the abstract writes “114 randomized.” The abstract gives a between-group difference of 3.29 points (95% CI −1.39 to 7.97); the results section gives 1.57 (−2.56 to 5.70); why the two differ is unexplained. Methods state that “the confidence interval must contain zero” as the condition for proving noninferiority—an incorrect formulation of the noninferiority test; the correct comparison is against the prespecified 9-point margin. Both intervals lie within ±9 points, so whichever reported value is used the conclusion is noninferiority; the inconsistency between analysis numbers and the abstract remains. Sample size was calculated for 68 complete cases per group; the authors wrote that they planned to randomize 106 and actually randomized 139.

On secondary outcomes, one-year Lysholm was 88±4.0 in the braced group and 85±3.9 in the brace-free group; there were 4 cyclops lesions in the braced group and 0 in the brace-free group; each group had 2 retears. The trial was funded by the German statutory accident insurance institution.

What this trial answers is that after isolated reconstruction with a hamstring graft, a daily rehabilitation brace for 6 weeks did not improve one-year subjective function, the difference lying within the 9-point margin. It did not study concomitant repair procedures, and it does not show that the nighttime extension splint in the first postoperative week can be omitted.

The meta-analysis by Lin et al. (2025) included 11 studies and likewise concluded that bracing does not improve outcomes, but internal numbers in that review contradict one another and specific pooled values should not be cited:

• Sample size is given three ways: 785; 387+388=775; Table 2 summing to 399+409=808.
• In the abstract, the Lysholm effect size is identical to that for KT-1000 (−0.30, −0.72 to 0.11); the body gives two Lysholm figures (random-effects −0.31, −2.31 to 1.69; fixed-effects after removing one study −0.93, −2.53 to 0.66); Figure 4 gives a third (−0.74, −1.86 to 0.38).
• Figure 4 labels Schoepp “not estimable,” yet Figure 9’s subgroup by follow-up duration includes Schoepp, using precisely its one-year Lysholm of 88±4.0 versus 85±3.9.
• Figure 9 places Yapıcı 2022 in “follow-up <2 years,” yet Table 2 lists that study’s follow-up as 40.2 and 36.9 months; therefore the subgroup conclusion that “beyond 2 years of follow-up the braced group’s Lysholm is 2.69 points lower (−5.01 to −0.37)” rests on a flawed grouping.
• The effect size for IKDC objective grade is called an odds ratio (OR) in the abstract and body and a risk ratio (RR) in Figure 3.
• Group sizes for the Gunadham study are entered in reverse.

The direction of this review is consistent with the Schoepp trial and with the meta-analysis by Yang et al. (2019; 7 randomized trials, 440 patients). OPTIKNEE’s appraisal of the latter is that bracing showed no benefit for laxity or physical function (moderate certainty) and no detected effect on self-reported function or pain (low certainty). German consensus item 8 states that a rigid brace after isolated reconstruction “may not be needed” (B1, moderate certainty); item 9 states that with concomitant meniscal injury a brace “may improve outcomes” (C, low certainty).

Current uses of bracing are two: the surgeon requires protection of a meniscal or collateral ligament repair; or early quadriceps control is poor and the patient gives way while walking, needing protection. The former is decided by the surgeon; the latter can use the same criteria as crutch discontinuation to decide when to remove the brace.

Driving and Stair Climbing

Aspetar’s advice on driving is not to drive until the patient can safely depress the brake in a simulated emergency—typically about 4–6 weeks after right-knee surgery and 2–3 weeks after left-knee surgery—based on the systematic review by Salem et al. (2021) on brake-response time. The Delaware protocol lists reciprocal stair climbing as a milestone at 4–6 weeks. These are empirical arrangements without comparative trials.

4. Open- and Closed-Chain Training

Whether seated knee extension (open-chain extension) loosens the graft has been debated in postoperative rehabilitation for thirty years. The concern comes from anterior tibial translation as seated extension approaches terminal extension. Mechanical data show that the ACL strain produced by open-chain extension is not unusually high compared with everyday activities. Beynnon et al., measuring strain in the intact ACL arthroscopically with an implantable sensor (cited via Brinlee et al., 2022, p. 772), reported 4.4% for isometric extension against 30 N·m at 15° of flexion, 3.8% for open-chain extension with a 45 N load, 3.7% for the Lachman test itself, and similar values for squatting and open-chain extension (4%–5%).

Brinlee et al. also wrote that strain during walking can reach 13%, citing Roldán et al. (2017). According to the methods, discussion, and Table 1 of that study’s author-accepted manuscript, 12 young adults without ACL injury history (7 men, 5 women) underwent motion capture; investigators estimated ACL length and strain with a three-degree-of-freedom knee model and then converted force from the force–strain relationship. In Table 1 the walking row lists peak force 0.774±0.064 N/body weight and peak strain 0.132±0.00248 (dimensionless); the discussion states peak walking strain of 13%. Brinlee et al.’s paraphrase “0.132 ± 0.00248 N/BW” is exactly this strain value, but paired with units of force.

More importantly, these numbers come from different studies, methods, and populations (intact ligaments, healthy-subject models); none was measured on a healing graft after reconstruction, and none compared angle-restricted extension within the same patients. They show that the strain seated extension places on the intact ligament is of the same order as everyday activity, and therefore open-chain training cannot be forbidden on mechanical grounds alone; whether it is safe for a healing graft depends on laxity results from randomized trials.

Several guidelines reach the same conclusion with slightly different wording. Aspetar’s evidence summary states that earlier versus later start of open-chain training showed no detected differences in laxity, strength, pain, range of motion, function, or balance; hamstring grafts may be more susceptible to early open-chain training than patellar-tendon grafts; there is no evidence for starting open-chain before week 4 (grade D). Its recommendation is “open-chain training may begin from postoperative week 4 within a restricted 90°–45° range without affecting stability; watch for anterior knee pain and adjust load accordingly,” with 88.8% agreement.

The Dutch guideline by van Melick et al. (2016) is the most specific: from week 4, train within 90°–45°; patellar-tendon grafts may add resistance (e.g., a knee-extension machine); hamstring grafts receive no external load for the first 12 weeks, to prevent graft elongation. For both grafts the range is opened week by week: week 5, 90°–30°; week 6, 90°–20°; week 7, 90°–10°; week 8, full range. German consensus item 20, “early open-chain from week 4,” is A2 with moderate certainty. The OPTIKNEE overview cites the meta-analysis by Perriman et al. (2018; 10 studies): open- versus closed-chain showed no detected differences in laxity, strength, or patient-reported function, with evidence quality low to moderate.

The hamstring-graft restriction rests on the randomized trial by Heijne and Werner (2007). Sixty-eight patients were first stratified by graft into patellar-tendon and hamstring groups (34 each), then each was randomized to start open-chain extension at week 4 or week 12, yielding four groups—P4, P12, H4, H12—of 15–19 patients each. All four groups were weight-bearing as tolerated immediately and wore no brace.

Progression in the early-start groups was: week 4, seated extension 90°–40° without resistance; week 5 range is written differently in two places—methods give 90°–0°, Table 2 gives 90°–20°, and which was actually used cannot be determined; week 6, 90°–0°; from week 7, “external resistance added according to symptoms and tolerance, with no upper limit.” Late-start groups began full-range open-chain at week 12, without resistance for the first week.

Laxity was measured with KT-1000 maximum manual force; assessors were blinded to rehabilitation group; contralateral test–retest ICC was 0.77. Side-to-side differences by group were as follows (original Table 5, mm):

Group	3 months	5 months	7 months
H4 (hamstring, open-chain from week 4)	2.3	2.5	2.3
H12 (hamstring, open-chain from week 12)	0.9	1.3	1.2
P4 (patellar tendon, open-chain from week 4)	1.1	1.6	1.3
P12 (patellar tendon, open-chain from week 12)	1.2	1.3	1.3

Mixed-model estimates showed that H4’s mean side-to-side difference across the three time points was 1.2 mm higher than H12 (95% CI 0.37–2.04, p=0.02) and 1.0 mm higher than P4 (0.22–1.84, p=0.04); abstract figures differ slightly from the body: H4 versus H12 is written 0.37–2.1, p=0.01, and H4 versus P4 as 0.18–1.86. Pivot-shift positivity was also greater in H4; comparisons with P4 at 3 and 7 months both had p=0.04. Range of motion, anterior knee pain, and balance showed no detected differences among the four groups; timing of open-chain did not change quadriceps torque—graft type did.

The authors themselves acknowledged two limitations: laxity was not measured immediately after surgery, so it is unknown whether the 1 mm difference was present from the outset; twenty surgeons participated, two of whom performed half of the hamstring cases, with their patients evenly distributed between H4 and H12. During follow-up, 3 patients missed two time points and 16 missed one. An erratum (same volume, 2007, pp. 472–473) adds only p-value notes for the strength figures and does not affect the laxity results.

Fukuda et al. (2013) used a different early-start approach in a randomized trial. Forty-nine patients with hamstring grafts: the early group began open-chain on a knee-extension machine from week 4, with the machine locked to restrict motion to 90°–45°; the late group trained through 0°–90° from week 12. Both groups began partial weight-bearing at postoperative week 2, wore no brace, and received about 70 supervised sessions at three per week. At 17 months, 18 early and 17 late patients completed follow-up; each group had 2 graft failures that underwent revision reconstruction.

Rolimeter laxity was 2.7 mm both before and after in the early group and rose from 2.6 mm to 3.5 mm in the late group, with no detected between-group difference at any time point. At 17 months, side-to-side difference of 3–5 mm was present in 7/18 (39%) early and 10/17 (59%) late; more than 5 mm occurred in 2 patients in each group. Quadriceps strength (injured as a percentage of uninjured) at 12, 19, and 25 weeks and at 17 months was 81.2, 91.8, 94.1, and 99.7 in the early group and 81.6, 87.0, 89.5, and 95.1 in the late group; within-group improvement appeared earlier in the early group, without statistically significant between-group differences. At 12 weeks the numeric pain rating scale (NPRS, 0–10, worst pain in the past week) was 1.1 early and 2.3 late. The abstract’s “No difference (P<.05)” is a typographical error and should read P>.05. Sample size was calculated to detect a 2 mm laxity difference with 15 per group and was underpowered for smaller differences.

Neither trial compared the two angle-and-load progressions within the same trial; the following interpretation is a reasonable inference whose cause has not been confirmed. When hamstring grafts begin open-chain at week 4, reach full range within two weeks, and soon add external resistance (Heijne’s protocol), mean laxity is about 1 mm greater and pivot-shift positivity is higher; when range is locked at 90°–45° on a machine (Fukuda’s protocol), no laxity difference is detected over 17 months. The differing results may lie in angle and load rather than in the “week 4” time point itself. Heijne compared an entire early-start package with a week-12 start and did not isolate “no external load for the first 12 weeks.” van Melick et al.’s “no load for the first 12 weeks” and Aspetar’s “90°–45°” are consistent in direction with these two trials.

The systematic review by Fontanier et al. (2025) included 14 studies and 729 patients, only 3 at low risk of bias. It classified the two trials above as “early versus late comparisons,” summarized early start as advantageous only for pain, with no unfavorable laxity findings among low-bias studies, and noted that one of the two trials was low-bias and one high. On the original numbers, Heijne’s hamstring group did show a laxity difference; that review’s summary cannot be used to negate this result. Another source of unfavorable laxity findings in the review is Bynum 1995: open-chain alone, started on postoperative day 1, high risk of bias; according to its abstract, the closed-chain group had smaller KT-1000 side-to-side differences and less patellofemoral pain. The authors’ affiliation, Medinetic Learning, is a training company; the article declares no funding—a fact that does not cast doubt on the results.

Among the three low-bias trials listed in the review, graft ruptures were 0 and 0, 0 and 0, and 1 and 1 in the two arms; the last is Fukuda 2013. Fukuda’s original report, however, describes 2 graft failures with revision reconstruction in each group; both the event counts and the outcome name in the review disagree with the original, for reasons that cannot be judged from the two papers. So few events cannot establish whether early open-chain training affects graft safety.

The Delaware protocol is the most aggressive: full-range seated extension 90°–0° from the first postoperative week, with light or no resistance, switching to heavy sandbags or a knee-extension machine at 2–3 weeks. Its basis is the strain data above and the clinic’s own cohort; no randomized trial supports “full-range loading from week 1” as superior to “restricted range from week 4.” A compromise is: full-range extension without resistance in week 1, to maintain motion and muscle activation; resisted work from week 4 within 90°–45°, opening the range on the guideline schedule; and no increase in load for hamstring grafts before week 12. The last point is consistent in direction with Heijne’s results, but the trial itself did not isolate it.

Other graft-specific adjustments were surveyed in a scoping review by Briem et al. (2026): 17 papers from 2014–2024, mostly Level V (expert opinion), with only Brinlee et al. covering all three grafts. Suggestions summarized in its Table 3 include: no dynamic resisted knee flexion for hamstring grafts before week 8, jumping deferred to week 16, and no open-chain load from weeks 4–12; for patellar-tendon grafts, attention to patellofemoral load at 45°–60°; for quadriceps-tendon grafts, training of rectus femoris with the hip extended. These recommendations rest mainly on biomechanical inference and clinical experience, as the review authors themselves characterize them. Panther expert opinion Table 4 is similar: limit resisted knee flexion for hamstring grafts before week 12; patellar-tendon strain is greatest at 90° of flexion, so early resisted work after patellar-tendon harvest should avoid heavy load in deep flexion.

5. Load and Measurement in Strength Training

Slow recovery of quadriceps strength is the most common residual problem after surgery; per the table in Section 1, patellar-tendon grafts take a mean of 30 weeks to reach 90%. Guidelines agree on load magnitude: it must be heavy enough. Aspetar states that both hamstring and patellar-tendon grafts may begin leg press at 3 weeks (grade D, 88.3% agreement), with eccentric training also from week 3; German consensus item 23, “high-intensity training,” is B2 with moderate certainty, on the rationale “improves strength without affecting stability.” The principal randomized trial that directly compared load intensities is that of Bieler et al.

Bieler et al. (2014) randomized 50 patients to high-intensity (n=24) or low-intensity (n=26) training from postoperative weeks 8–20; 38 completed. Load progressed per original Table 1: high-intensity weeks 8–9 at 20RM (familiarization), weeks 10–11 at 15RM, weeks 12–13 at 12RM, weeks 14–20 at 8RM (3 sets of 8); low-intensity weeks 8–9 at 30RM, thereafter always 20RM (2 sets of 20). Load already differed from week 8, so the power difference at week 14 was measured after 6 weeks of different loading. The two groups also performed different exercises—not the same movements at two weights.

Results are expressed as injured-side percentage of the uninjured side. Leg-extension power was 84% versus 73% at week 14 (p=0.027) and 97.5% versus 83.5% at week 20 (p=0.006); no laxity difference was detected. At week 20, single-leg hop was 69.1% versus 65.1% and triple hop 75.3% versus 68.1%, neither statistically significant. Baseline imbalance: time from injury to surgery was 40.3 months in the high-intensity group and 16.8 months in the low-intensity group (p=0.043); the direction of this difference’s effect on the results was not verified. The study was funded by the Danish Ministry of Culture.

Aspetar’s appraisal of this trial is “evidence insufficient to support which intensity is better,” grade D; the German consensus assigned moderate certainty to similar evidence. What the trial supports is that progressive loading from two months after surgery, reaching 8RM by three and a half months, produced no detected laxity change in these 38 people and faster recovery of leg-extension power. It does not support “high intensity is necessarily better,” because the exercises differed, the sample was small, and hop-test differences did not reach statistical significance.

Load intensity may follow Delaware parameters: weeks 7–9, concentric and eccentric load at 60%–75% of 1RM, 5–8 repetitions, 3–5 sets; weeks 10–12, 60%–85%; from weeks 13–16 onward, 75%–90%, 5–8 repetitions, 4–5 sets. Alternatively, use repetitions in reserve: choose a weight that allows 5–8 repetitions with 2–3 left in reserve; if 3–4 remain after the set, increase the weight. These are general training parameters, not compared with other parameters in ACL populations; their value is converting “heavy enough” into concrete numbers.

When assessing strength recovery, measurement method is more error-prone than training content. Isokinetic or isometric dynamometry is most reliable; handheld dynamometers must be fixed; a 1RM test on a knee-extension machine through 90°–45° is acceptable; a 1RM on the leg press is not. Sinacore et al. (2017; based on the abstract) compared several alternative methods in 56 people with impaired knee function: handheld dynamometry and 90°–45° extension 1RM agreed reasonably with isometric dynamometry but both tended to overestimate symmetry; leg press was worse. Brinlee et al. paraphrased Sinacore’s finding: estimating symmetry from leg-press 1RM misclassifies patients who have not actually met criteria as having met them—23.6% at the running stage and 27.3% at return-to-sport. The reason is that leg press is shared by hip, knee, and ankle, so quadriceps weakness on the injured side can be masked by other muscle groups.

Another common error is in the denominator. The uninjured side also weakens after injury, and preoperative uninjured values are often higher than postoperative ones. The Delaware protocol requires using preinjury data or the highest uninjured value during rehabilitation as the denominator; otherwise the symmetry index is inflated. Related data from Wellsandt et al. appear in Section 10.

Absolute strength and the symmetry index answer different questions. Panther expert opinion notes that the symmetry index does not reflect absolute level, recommends also examining body-weight–normalized peak torque, and gives 3.0 Nm/kg as a reference threshold; the document does not attach validation of this figure against function or reinjury.

6. Neuromuscular Electrical Stimulation

The purpose of electrical stimulation is to make the quadriceps work when voluntary contraction is inhibited early after surgery. Among adjuncts, its evidence is relatively strong. The OPTIKNEE overview cites the systematic review by Hauger et al. (2018; 11 studies, 6 pooled): within 4–12 weeks after surgery, stimulation plus training improved quadriceps strength more than training alone, standardized mean difference 0.73 (95% CI 0.29–1.16); included trials had PEDro scores of 3–7, and the overview rated certainty moderate. Aspetar rated “moderate improvement,” grade C; German consensus item 15 is A1 with moderate certainty.

The meta-analysis by Li et al. (2025) pooled trials totaling 202 in the stimulation arm and 200 in the control arm; the abstract’s “N=202” treats one arm’s count as the total. The pooled standardized mean difference for quadriceps strength was 0.55 (0.33 to 0.76), fixed-effect model, I²=22%. “Long-term” follow-up in the paper means more than 6 weeks, not years.

In the subgroup by start time, the effect size for starting within one week after surgery was 1.48 (0.80 to 2.17)—apparently large, but it comes from three very small trials: Delitto 1988 (10 vs 10), Snyder-Mackler 1991 (5 vs 5), and Toth 2020 (9 vs 7). This is a between-trial comparison; no trial randomized early versus late start, so this subgroup result cannot serve as evidence that “earlier is better.” Patient-reported Lysholm showed no pooled difference.

Toth et al. (2020) began intervention early after injury and used a sham-stimulation control. Twenty-five patients with acute ACL rupture were randomized to stimulation (n=14) or sham (n=11). Intervention began within 3 weeks of injury and continued to 3 weeks after surgery (stopped preoperatively, resumed within 72 hours postoperatively); mean intervention duration was 58 days. Patients used a portable unit at home 5 days a week for 60 minutes daily; pulse width 400 μs, frequency 50 Hz, 10 seconds on and 30 seconds off, isometric contraction at about 40° of flexion, intensity aimed at tetanic contraction within pain tolerance. The sham group used an identical-appearing unit that delivered no current and was told it was microcurrent for pain. Adherence was 85% in the stimulation group and 63% in the sham group.

The primary outcome was vastus lateralis biopsy muscle fibers at 3 weeks after surgery: stimulation attenuated type II fiber atrophy and preserved type I contractile function. Secondary outcomes at 6 months (16 completers): CT quadriceps cross-sectional area as an injured-to-uninjured ratio was actually lower in the stimulation group (p<0.05); isometric and isokinetic strength corrected for cross-sectional area, patient-reported outcomes, and single-leg hop showed no between-group differences. Li et al.’s Table 4 lists Toth’s cross-sectional-area data as 72.2 versus 82.8 at 24 weeks, difference −10.6 (p=0.023), and labels the dispersion as standard deviation, whereas Toth reports mean±standard error—an inconsistency of labeling.

Thus Toth shows that high-intensity stimulation early after injury can attenuate atrophy at the cellular level; by 6 months, neither the CT injured-to-uninjured quadriceps cross-sectional-area ratio nor strength showed a benefit.

Stimulation parameters vary widely across trials and are a source of meta-analytic heterogeneity. Delaware parameters may serve as a reference. Electrodes are placed proximally lateral and distally medial on the quadriceps; pulse width 400 μs, frequency 75 Hz, 2-second ramp, 12 seconds on and 50 seconds off. Each session delivers 15 contractions, three times weekly. Intensity should reach at least 50% of the patient’s maximal voluntary isometric contraction and be increased as far as tolerated. Position is 60° of flexion for isometric contraction; with patellar-tendon or quadriceps-tendon grafts, begin at 45° and move to 60° once pain decreases. Discontinue when the quadriceps strength index reaches 80%.

The most critical parameter is intensity. Toth’s trial achieved tetanic contraction at home, so the issue is not whether the unit is at home or in clinic, but whether it can deliver enough current and whether someone ensures intensity is pushed to the tolerance limit.

7. Blood-Flow Restriction Training

Blood-flow restriction (BFR) training uses an inflatable cuff at the proximal thigh to restrict venous return; the training premise is that light loads of 20%–30% 1RM produce muscle stimuli approaching those of heavy loads. Whether that premise holds in postoperative patients must be tested in trials. When early postoperative patients cannot tolerate heavy loads, the method is attractive, and recent recommendations have outrun the evidence. German consensus item 17 rates “low-load resistance plus blood-flow restriction” A1 with moderate certainty, concluding that it “may improve strength and muscle volume”; Panther expert opinion suggests training at 20%–30% 1RM from 2–3 weeks after surgery; the Delaware protocol lists it as an option for patients with pain or poor load tolerance. On the other hand, Aspetar rates it grade D, and the OPTIKNEE overview, citing Charles et al. (2020; 4 randomized trials, 98 patients), rates certainty very low.

The highest-quality trial detected no benefit of BFR. Erickson et al. (2025) conducted a double-blind, sham-controlled randomized trial. The abstract reports 48 patients randomized to BFR or sham; Table 1 lists 25 sham and 23 BFR, with still fewer in outcome analyses (e.g., MRI physiologic cross-sectional area: 19 sham, 18 BFR). Intervention ran from 1 month preoperatively to 4–5 months postoperatively, three times weekly. The BFR group used 20%–30% 1RM at 60% limb occlusion pressure (Delfi system); the sham group used an identical cuff inflated to 20 mmHg, with load 60%–70% 1RM, 3 sets of 10. Both groups also received electrical stimulation and open-chain training; adherence was 93.5% versus 93.7%.

No between-group differences were detected for: isometric and isokinetic peak torque, rate of force development, gait, MRI quadriceps physiologic cross-sectional area, vastus lateralis T1rho (to estimate muscle fibrosis), or biopsy fiber type and cross-sectional area, satellite cells, collagen, and fibrogenic progenitors. Grafts were mostly patellar tendon (23 sham, 22 BFR), with 3 hamstring; almost all had concomitant meniscal injury. The trial was funded by the U.S. National Institutes of Health; acknowledgments disclose that author Darren L. Johnson receives royalties and consulting fees from Smith & Nephew.

This trial was not an equivalence design; “no detected difference” does not mean the two training modes are equally effective. What it shows is that under supervision, low-load BFR did not outperform direct use of moderate-to-high loads. What it does not answer is whether, for patients who cannot tolerate moderate-to-high loads at all, BFR is better than the same low load without a cuff.

Zheng et al. (2026) compared cuff versus no cuff at the same low load—closer to that unanswered question—but in general reconstruction patients at postoperative weeks 13–20, not in a population unable to tolerate moderate-to-high loads, and therefore did not directly answer it. The trial was positive.

Forty-eight patients were randomized; 43 were analyzed per protocol. From weeks 13–20 both groups performed the same 30% 1RM leg press, 4 sets, twice weekly; the BFR group pressurized at 40% arterial occlusion pressure, the sham group below 40 mmHg. How pressure was determined is stated inconsistently: methods say 40% of resting arterial occlusion pressure measured before training, while Table 2 gives fixed pressures of 80–140 mmHg by thigh circumference; which was actually used is not stated. Baseline outcome values before intervention were also not reported.

At week 24, IKDC was 73.9±2.7 versus 69.4±3.6, Tegner 4 versus 3, knee-extension strength 2.23±0.11 versus 1.92±0.23 N/kg (handheld dynamometer), and Y-balance 96.6% versus 94.6%; muscle thickness showed no detected difference. The control “standard protocol” (Table 1) was quite conservative: toe-touch weight-bearing for postoperative weeks 0–2, brace limited to 0–90° through week 5, “avoid hyperextension for 12 weeks.” Sample size was based on an effect size of 1.1 from a small pilot. The study was funded by a Fujian provincial grant; the cuff was a TheraTools brand, and the paper does not state whether the manufacturer supplied or funded it.

This trial shows that adding BFR to the same low load in mid-stage rehabilitation yielded higher function scores and knee-extension strength at 24 weeks; it compared cuff versus no cuff, not BFR versus adequate heavy-load training.

The meta-analysis by Hu et al. (2026) pooled 9 randomized trials from 2019–2024 (372 patients), was unregistered, declared no conflicts of interest, and listed the authors’ affiliation as Bayi Orthopedic Hospital (under China Rongtong Medical Health Group Co., Ltd.). Several problems in this review affect different results.

On citation: pages 2 and 12 cite Erickson et al. (2025) as “confirming that BFR increases muscle-fiber cross-sectional area and mitochondrial markers,” whereas that trial detected no difference; at the same time, this eligible trial was excluded without stated reason.

On pooled values: the peak-torque summary of 8.34 (7.47 to 9.22) draws 84.5% of its weight from one 2022 study. That study’s patients had a mean age of 44.7 years; the review lists 39.01±1.99 versus 30.43±1.76 with units inconsistent with other studies—this pooled value cannot be used. The KOOS Symptoms summary of 8.08 (7.59 to 8.58) draws 99.6% of its weight from one study, listed as 100±0.01 versus 91.92±0.88—almost no dispersion on a 0–100 scale; original scores were not checked in this chapter, so data error cannot be asserted, but until clarified this estimate cannot be relied on. IKDC is 5 studies, 5.25 (−3.01 to 13.51) in Figure 4 and 4 studies, 5.75 (−3.24 to 14.74) in Table 3, without stating which is the authors’ final analysis. The Lysholm summary of 5.75 (0.60 to 10.90) has I²=94%, with one 2024 study at −7.20 favoring control; that pooled value is not thereby invalid, but trial results are inconsistent. In addition, reference numbering is shifted by one from some point onward.

Other outcomes were not checked item by item; the review’s list of 9 trials and their intervention parameters remain usable. On safety, the review paraphrases a 2023 Chinese trial reporting postoperative complications including lower-limb venous thrombosis, chronic pain, and lower-limb swelling, with a lower overall complication rate in the BFR group; the review does not give counts or group assignment, and the primary trial was not checked here.

Taken together: when early postoperative patients cannot add weight, BFR is a reasonable option, and Zheng’s trial supports benefit from adding a cuff at the same low load; the most credible trial did not detect that BFR outperforms adding weight directly. The German consensus’s A1 rating diverges too far from Aspetar and OPTIKNEE; the latter two align better with existing trial results and are adopted here. In use, set pressure as a percentage of limb occlusion pressure with a cuff that can measure pressure; patients with a history of venous thrombosis should consult a physician first. These are operational common sense, not trial conclusions.

8. Neuromuscular Training, Proprioception, and Jump Progression

Balance, landing control, and cutting drills appear in every protocol. What they add beyond strength training alone is addressed in the materials for this chapter by only one directly comparative randomized trial—Risberg et al. (2007). Seventy-four patients with patellar-tendon grafts were randomized to neuromuscular training (n=39) or strength training (n=35), each for 6 months. At 6 months the Cincinnati knee score was 80.5 (SD 12.3) in the neuromuscular group and 73.4 (9.6) in the strength group, p=0.01 (a table footnote writes P=0.05, inconsistent with the body); overall function on a visual analog scale was 72.4 versus 59.3, p=0.02. Strength, hops, balance, proprioception, and KT-1000 (3.0 mm versus 3.4 mm) showed no detected differences.

The strength group actually received more treatment sessions and hours (57.6 sessions, 62.9 hours versus 42.2 sessions, 43.8 hours); adherence was 91% versus 71%. Two clinics participated; balance and proprioception testing equipment failed after patients 47 and 51, so later patients lacked those data.

How clinically meaningful a 7.1-point between-group mean difference is remains hard to judge. Harris et al. (2017) Table 3 give a 6-month minimal clinically important difference of 14 points for the Cincinnati score—a threshold for individual change that cannot be applied directly to a between-group mean difference. What can be said is that self-reported function was somewhat higher after neuromuscular training, with a limited magnitude; the strength group trained more often and longer yet was not stronger. Mixing the two forms of training is reasonable; one need not choose only one.

German consensus items 24 (proprioceptive training) and 25 (jump training) are both B2 with moderate certainty; Aspetar’s appraisal is that jump and agility training may further improve subjective knee function without additional effect on strength.

When jump training begins is set by entry criteria in each protocol, not by a fixed date. The Delaware protocol starts light hops on a leg-press machine or slide board after week 12, and landing-control drills from a 2-inch (about 5 cm) box at weeks 13–16. High-intensity jumps such as box jumps, drop jumps, and standing long jumps require quadriceps strength index and hop tests both above 85%, effusion no more than trace, and completion of running progression and stage-1 sprint progression; progression order is from two-foot to single-leg landing and from single-plane to multiplane; hamstring grafts defer jumping to week 16. The adolescent consensus (van Melick et al., 2023) requires for starting jumps and running in stage 2: symmetric range of motion, no effusion, and symmetry index above 70% (85% agreement), with movement quality placed first. These numbers come from consensus and protocols; no trial has compared 70% with 85%.

9. Return to Running

Starting to run at 12 weeks after surgery is the most common arrangement in the literature; that timing comes from custom. Rambaud et al. (2018) retrieved 212 rehabilitation timelines from 201 studies: the median time to start running was 12 weeks after surgery (interquartile range 3.3 weeks, range 5–39 weeks), with 8 weeks in 11%, 12 weeks in 42%, and 16 weeks in 12%; only 36 studies (18%) used criteria beyond time. The authors explicitly wrote that no study has tested whether starting to run at 12 weeks is safe.

Aspetar places return-to-running criteria in Box 5, with 87.8% agreement; the box notes that these are suggestions “in the absence of research findings,” and the guideline also states that regarding safety of starting at 12 weeks or earlier there are “no definitive results.” Criteria across documents are as follows:

Source	Strength and hops	Other criteria
Non-time criteria summarized by Rambaud et al.	Isometric quadriceps and hamstring strength symmetry index >70%; hop-test symmetry index >70%	Pain <2/10, flexion ≥95% of uninjured, full extension, no or trace effusion
Aspetar guideline Box 5	Quadriceps strength symmetry index >80%; countermovement-jump eccentric impulse symmetry index >80% (requires a force plate)	Flexion ≥95% of uninjured, full extension, no or trace effusion; pain-free aquatic or body-weight-supported treadmill running; pain-free continuous single-leg pogo hops
Delaware protocol	Quadriceps strength index >80%	Effusion no more than trace; patient understands soreness rules
Panther expert opinion Part 2 (adopting one set of Rambaud criteria)	Isometric strength symmetry index >80%, isokinetic >70%	Full extension, mild effusion, pain score <2

Delaware timing is postoperative weeks 13–16, preceded by two to three weeks of body-weight-supported treadmill or pool running.

How to progress volume after starting to run is specified in the Delaware running progression—one of the few protocols that can be followed directly. Level 1 is walk 0.1 mile, jog 0.1 mile (1 mile ≈1.6 km), repeated 10 times; thereafter the jog segment is lengthened stepwise to 0.4 mile; level 5 is continuous jogging for 2 miles; then up to 3 miles, then increased speed. For the first two weeks, 2 days of rest are required between running sessions, with no more than two level advances per week. Soreness rules: soreness during warm-up → rest 2 days and drop one level; soreness during the run → rest 1 day and drop one level; soreness after the run → stay at the current level. These are empirical rules intended to give progression a stated rhythm.

Short-term response after returning to running was examined in a small cohort by Pairot de Fontenay et al. (2022). Thirty-five patients with hamstring grafts were recruited through a university email list, all clinically cleared, and at a mean of 3.8±0.7 months after surgery (range 2.8–5.4) began a 14-day, 10-session walk–jog program completed at home without pace monitoring. Thirty-four entered analysis; 33 completed training.

Eighteen who did not need to adjust progression because of symptom flare during training were defined as responders; the remaining 16 were non-responders. Pre-training IKDC was 67.1 in responders and 59.3 in non-responders. With a cut-point of 63.7, the area under the receiver-operating-characteristic (ROC) curve was 0.804 (95% CI 0.625–0.955); those above the cut-point were 3.11 times as likely to complete training uneventfully as those below (1.29 to 7.53). An area of 0.5 equals chance; 1.0 indicates perfect discrimination.

This cut-point was derived and tested in the same sample, without independent validation; the confidence interval around 34 people is also wide. No one developed joint swelling. The authors wrote in the discussion that other variables were far from significance and that a larger sample would not change the result—an inference without basis; nonsignificance does not mean that strength and related factors are unimportant. What the study shows is that when running begins 3–4 months after surgery, about half have short-term symptom fluctuation, more so among those with lower subjective function scores. It cannot serve as a standard that “IKDC below 63.7 means one should not start running.”

10. Uses and Limits of the Limb Symmetry Index

Among clearance criteria for running, jumping, and return to sport, the most used metric is the injured side as a percentage of the uninjured side—the limb symmetry index (LSI)—most often with a 90% cut-off. That cut-off is empirical; its limits already have data.

Wellsandt et al. (2017) compared two algorithms in 70 athletes undergoing return-to-sport testing (28 hamstring autografts, 42 allografts). One was the conventional symmetry index, with the denominator the uninjured side on the test day; the other, termed “estimated preinjury capacity” (EPIC), used as denominator the uninjured-side value measured 1.5–1.9 months after injury and before surgery. The proportion meeting 90% on quadriceps strength and all four hop tests was 57.1% (40) by the former and 28.6% (20) by the latter; 24 met the symmetry index but not preinjury capacity.

Over at least two years of follow-up, 11 sustained a second ACL injury (median time from surgery to reinjury 78 weeks, range 27–276). Of these, 8 had met the symmetry index at the time, and 6 of those 8 had not met preinjury capacity. Sensitivity for reinjury was 0.273 (95% CI 0.010–0.566) for the symmetry index and 0.818 (0.523–0.949) for preinjury capacity; but specificity of preinjury capacity was 0.305 and the positive likelihood ratio 1.18—that is, it also classified most who did not later reinjure as not meeting criteria. Of the original cohort of 182, 70 entered analysis; among those excluded, 32 were excluded because of a change in dynamometry equipment. The study was funded by the U.S. National Institutes of Health.

This study shows that the uninjured side also weakens after injury, and using it as the denominator overestimates recovery. The preinjury-capacity alternative identifies some additional later reinjuries at the cost of holding more people out; it has itself not been validated as a clearance standard.

The Delaware–Oslo cohort of Grindem et al. (2016) provides the most often cited numbers on the relationship between the symmetry index and reinjury. Here “reinjury” means any acute knee injury, not limited to the ACL. Of 106 reconstruction patients, 100 had follow-up data; 74 returned to level-I sport (handball, soccer, basketball, floorball) within two years.

Among 45 whose quadriceps symmetry index was below 90% before return to sport, 15 (33.3%) reinjured; among 24 above 90%, 3 (12.5%) reinjured; each 1-percentage-point higher symmetry index was associated with a 3% relative reduction in reinjury risk (hazard ratio 0.97, 0.94 to 0.99). Of 18 who passed all seven criteria (four hops, quadriceps strength, and two questionnaires all above 90), 1 (5.6%) reinjured; of 55 who did not pass, 21 (38.2%) reinjured—hazard ratio 0.16 (0.02 to 1.20), p=0.075, not statistically significant; the “84% reduction” in the article title refers to this estimate. Associations of the four hop tests alone with reinjury did not reach statistical significance (hazard ratios 1.01 to 1.04). Testing occurred a mean of 2.3±2.4 months before return to sport; the authors also noted that if testing were repeated closer to return, more would pass.

The Swedish registry cohort of Beischer et al. (2020) examined second ACL injury (graft rupture or contralateral rupture). Among 159 athletes aged 15–30 years, those who met 90% on all five muscle-function tests versus those who did not showed no statistically significant difference in second ACL injury rate (hazard ratio 1.31, 0.47 to 3.67); each 10-percentage-point higher quadriceps symmetry index had a hazard ratio of 0.96 (0.92 to 1.01).

This study had only 18 events. The 159 included were only 32% of the 494 eligible; most of the rest were excluded for lack of testing near return to sport. Testing occurred a mean of 65±47 days before return; second injuries relied on clinical diagnosis without requiring MRI confirmation. A July 2020 corrigendum notes that 22 people were misrecorded as female, changing the female proportion from 64% to 50%, with Tables 4 and 6 and the appendix corrected accordingly. This study cannot negate Grindem’s results—the outcomes, populations, event counts, and testing times differ; it shows that the relationship between meeting the symmetry index and reinjury is unstable in small samples.

Synthesizing the two cohorts, 90% is a minimum requirement; meeting it does not guarantee freedom from reinjury. The cohorts differ in population and outcome and cannot be combined into a single prognostic figure; what can be said is that reinjury rates among those who met criteria were nonzero in both cohorts, and those who did not meet criteria had clearly higher rates in the Grindem cohort. The symmetry index also has known gaps: when the uninjured side weakens, the denominator shrinks; leg-press 1RM overestimates; and the relationship between hop-test symmetry and reinjury did not appear in either cohort. van Melick et al. (2016) recommend a symmetry index of 100% for cutting and contact sports—expert opinion. No figure more thoroughly validated than 90% appears in the materials for this chapter, so it remains the general standard; absolute strength, testing method, and testing time should be considered alongside it.

11. Differences Related to Graft, Meniscus, and Age

Graft

Earlier sections already noted graft effects on open-chain load and jump timing. On overall pace, Brinlee et al.’s table in Section 1 shows clearly slower quadriceps strength recovery with patellar-tendon grafts: in the non-suture meniscus group, strength index reached 80% at a mean of 12–14 weeks and 90% at a mean of 30 weeks; hamstring or allograft reached 80% at a mean of 9 weeks and 90% at a mean of 15 weeks. Harvest-site pain may contribute; Brinlee et al. discussed this without causal analysis. Delaware management of patellar-tendon grafts places isometric testing and electrical stimulation at the most comfortable angle between 45° and 60°, and manages anterior knee pain with tendinopathy-style pain monitoring to adjust load; quadriceps-tendon grafts are handled on the same logic and also recover strength slowly.

Allografts have no harvest-site pain, yet protocols generally defer running, jumping, cutting, and return to sport, placing return after 12 months on the rationale that graft healing is slower—a consensus-level judgment. What is distinctive about hamstring grafts is the harvest site and laxity: the Delaware protocol requires no isometric resisted knee flexion before week 6, no dynamic resisted knee flexion before week 8, and no further restriction after week 12; load limits for open-chain extension appear in Section 4.

Meniscus

With concomitant meniscal surgery, weight-bearing and flexion limits depend on how the meniscus was treated; this part rests almost entirely on expert opinion. The EU–US meniscus rehabilitation consensus of Pujol et al. (2025) comprises 29 statements: evidence grade A in 1, B in 2, C in 9, and D in 17. After partial meniscectomy, immediate full weight-bearing is allowed (grade A); after repair of a longitudinal tear, weight-bearing is allowed but flexion is limited for about 6 weeks (grade C); after repair of complex, horizontal, radial, or root tears, both weight-bearing and flexion are limited for 4–6 weeks (grade C), with squatting, jumping, and rotation avoided for at least 4 months.

Question 7 of the consensus specifically addresses concomitant ACL reconstruction: the rehabilitation plan is similar to that for isolated meniscal repair; return to sport may be deferred to 9 months because of the ACL reconstruction; a stable longitudinal repair does not require changing the ACL rehabilitation plan (mean agreement 8.4, median 9, strong). For ramp lesions, the systematic review cited by the consensus gives a range of at least 2 weeks of non-weight-bearing and full weight-bearing by 4–12 weeks. Time to return to sport after meniscal repair is at least 4 months for longitudinal repairs and 6–9 months for complex and root repairs. Authors of this consensus extensively disclosed consulting and patent relationships with device companies; the statements themselves do not name specific products.

Rehabilitation after ACL reconstruction with concomitant lateral meniscus posterior root repair was summarized by Benelli et al. (2025). All 12 studies were Level IV (case series), totaling 277 patients, with mean MINORS methodological quality of 10. On weight-bearing, 7 used non-weight-bearing for 4–6 weeks, 4 used partial weight-bearing for 2–12 weeks, and 1 used immediate full weight-bearing (Thomas et al., 2023, n=50). On flexion, 8 limited to 0–90° for 4–6 weeks, 1 continued for 16 weeks, and 2 had no limit. On bracing, 5 used a brace, 1 did not, and 6 did not report. There are no comparable outcome data across these studies with which to judge which restriction is better. Differences among protocols may reflect surgeons’ habits or differences in tear type and fixation; the review itself cannot distinguish. Patients should ask their surgeon for the rationale behind the restrictions given.

Delaware adjustments for concomitant procedures can serve as one concrete reference. Isolated repair: immediate weight-bearing as tolerated in a brace or immobilizer locked in extension for 4 weeks; range of motion to 0–90° by week 2, then progressed as tolerated. Complex or root repair: non-weight-bearing for 6 weeks, full weight-bearing by week 9; weight-bearing flexion limited to 45° at weeks 8–12 and to 70° before week 16; no isolated hamstring resistance within 16 weeks. Partial meniscectomy: no adjustment. Microfracture: non-weight-bearing for 2–4 weeks, with training adjusted by lesion site and size. Medial collateral ligament injury: sagittal-plane motion only for 4–6 weeks. This is one clinic’s pathway and has not been validated. German consensus item 2, “prolonged partial weight-bearing with concomitant injury,” has 100% agreement and evidence grade C.

Adolescents

Rehabilitation for patients aged 10–18 years was the subject of a Delphi consensus organized by van Melick et al. (2023), with 20 experts and consensus defined as ≥70% agreement. The consensus treats ages 10–16 and 17–18 separately and uses Tanner staging for developmental stage; for adolescents it places more emphasis on neuromuscular training and movement quality and less on resistance (75% agreement), and requires parental involvement (80%).

Criteria for progressing from stage 1 to stage 2 are no or trace effusion, control of terminal extension (no lag on straight-leg raise), and emotional readiness to begin resisted training. Starting jumps and running in stage 2 requires symmetric range of motion, no effusion, and symmetry index above 70%. Progressing from stage 2 to stage 3 uses a single-leg sit-to-stand test with symmetry index above 80% in prepubertal children and isokinetic strength in mid-to-late puberty. Return to sport requires no effusion (100% agreement), full range of motion (100%), single-leg sit-to-stand or isokinetic strength and hop-test symmetry index above 95% (70%–85% agreement by item), and assessment of psychological readiness.

Return-to-sport timing for ages 10–16 is set at a minimum of 12 months (75% agreement); page 8 of the consensus explicitly states that this is expert opinion, with no evidence that 12 months is safer than 9. The background for extending time is high reinjury rates in the young: Beischer et al. cite an estimate that about one quarter of those under 25 who return to high-risk sport sustain a second ACL injury; in that study’s own cohort, athletes with a preinjury Tegner score of 10 had a second-injury rate of about 25%.

12. Supervision, Home Programs, Apps, and Psychological Support

How many clinic visits per week and how long each lasts involve a trade-off between cost and effect. The OPTIKNEE overview cites the meta-analysis by Gamble et al. (2021): home-based programs with 3–17 supervised sessions, compared with 14–36 supervised sessions, showed no detected outcome differences (moderate certainty); the overview also states that whether completely unsupervised, unstructured home programs are equivalent has no evidence. Aspetar’s appraisal is consistent: unsupervised versus supervised programs showed no detected differences in laxity (grade C) or subjective function, and for patients who cannot afford care or travel, unsupervised training is a feasible option. German consensus item 3, “unsupervised rehabilitation,” is evidence grade B2 with moderate certainty but only 36% expert agreement—no consensus; item 4, “combined supervised and independent training,” is A2 with moderate certainty. In other words, experts accept fewer supervised sessions but not complete absence of supervision.

The systematic review by Khubzan et al. (2025) included 12 studies and 711 patients, 7 of which entered meta-analysis. All pooled directions in original Table 3 favor the supervised group. The 6-month Lysholm standardized mean difference was −0.06 (−0.81 to 0.70, high heterogeneity, low certainty) and the 12-month −0.33 (−0.68 to 0.02, moderate certainty), neither statistically significant. Isometric hamstring strength at 8–9 months (−0.48, −0.88 to −0.08) and isokinetic hamstring strength at 12 months (−1.34, −1.89 to −0.79) reached statistical significance but each came from 2 studies with fewer than 50 per arm; the authors rated certainty moderate by GRADE. Pooled isometric and isokinetic quadriceps strength did not reach statistical significance.

The abstract and body write the hamstring strength effect sizes as unsigned 0.48; direction should follow Table 3 and the forest plots. Supervised-session counts in included studies ranged from 6 to 64, and the “home” versus “supervised” distinction itself is coarse. The “19%” cited in the body is that review’s paraphrase of a single study by Revenäs—home-group Lysholm improved 19% and supervised 1% at 12 months—not a pooled result; Syed et al. (2024) reported return-to-sport rates of 76.6% versus 53.3% in a pilot. Several figure legends cite wrong reference numbers. The review declares no funding. Its direction is consistent with OPTIKNEE: function scores are similar, strength may favor more supervision—especially the hamstrings; evidence comes from small samples with low-to-moderate certainty.

Evidence for smartphone apps and remote rehabilitation needs separate appraisal. German consensus item 14, “digital applications,” is B1 with moderate certainty, concluding possible improvement in KOOS and pain; Panther expert opinion Part 2 judges the opposite—that digital tools and AI-assisted rehabilitation currently lack evidence. The German consensus lists the studies and digital intervention types it relied on but does not name specific products; the two documents may not be appraising the same class of product.

Evidence for psychological support has two layers. An OPTIKNEE-cited review (4 randomized trials) showed that adding guided imagery, relaxation, coping modeling, and similar psychological interventions at 6–12 weeks after surgery improved psychological indices (low certainty) without consistent effects on self-reported function, quality of life, pain, or quadriceps strength. The same overview paragraph also lists knee laxity among improved outcomes, likewise low-certainty evidence; that item is also paraphrased from that review and was not checked against the review or its primary trials in this chapter.

One randomized trial used return to sport as an outcome. The BANG trial by Fältström et al. (2026) enrolled 161 patients aged 15–30 who had participated in cutting sport at least twice weekly before injury, randomized to usual rehabilitation plus 24 weeks of app-based self-directed psychological support grounded in cognitive-behavioral principles (n=80) or usual rehabilitation (n=81). The primary outcome was return at 12 months to the preinjury sport at the preinjury level with full participation in training and matches: 36% in both groups, risk ratio 1.01 (0.59 to 1.73). Psychological readiness, self-reported function, and quality of life showed no detected differences at any time point; each group had 6 new ACL injuries.

The trial planned 222 participants and stopped early because of the pandemic and closure of the app platform in June 2022; only 115 had return-to-sport data at 12 months, and the authors explicitly wrote that the trial was underpowered. In 12-month clinical testing (n=80), the app group’s 180°/s quadriceps and hamstring symmetry indices were lower than control (87% versus 93%, p=0.018; 89% versus 95%, p=0.046)—secondary outcomes the authors did not explain; the sample is small and overinterpretation is unwarranted. The study was funded by Swedish and Australian research grants; one author is editor-in-chief of the publishing journal and recused from the editorial process. German consensus item 7, “postoperative psychological and social interventions,” is evidence grade E with very low certainty and 73% agreement—likewise no consensus.

Taken together: psychological readiness is associated with return to sport—an observation Brinlee et al. paraphrase from ACL-RSI research—but association does not show that psychological intervention raises return-to-sport rates; early psychological intervention may improve fear and anxiety, with weak evidence; one self-directed app in an underpowered trial showed no detected effect on return-to-sport rate; the materials for this chapter contain no trial of face-to-face psychological support on return-to-sport outcomes.

13. Work-Up When Recovery Stalls

When rehabilitation does not go smoothly, common presentations are regression of extension, persistent swelling, strength stuck at a plateau, a change in pain location, or giving way while walking. Each presentation has more than one explanation; the order of work-up depends on recent sprain history, how abruptly effusion appeared, whether passive extension is equally limited, a hard block at terminal extension, recurrent giving-way sensations, and stability-examination findings. Interpretations and differentiating clues in this section are clinical inferences and have not been validated.

Extension regresses, or remains ≥5° short after 4 weeks. Extension limited by pain and effusion usually improves as swelling settles. Active extension short of a passively achievable range suggests quadriceps inhibition. When passive extension is also short, with a hard terminal block and infrapatellar pain on extension, mechanical causes (cyclops lesion, arthrofibrosis) should be considered first; further intensifying extension training then yields limited returns, and the patient should return to the operating surgeon for assessment. Mauro’s data show that among knees still ≥5° short at 4 weeks, about half eventually underwent arthroscopic debridement under that hospital’s pathway—not every case does.

Effusion remains above grade 2 for a prolonged period, or reappears after resolving. Effusion in the first few postoperative weeks is usually related to the surgery itself. Effusion that appears later or recurs after resolving must be judged in light of the course. Recurrent effusion beyond three months should first be related to load progression: swelling the day after training that settles with reduced volume supports the explanation that load exceeds current capacity, but that response alone cannot determine the cause; effusion unrelated to training volume, with locking or a new pain site, raises concern for meniscal or cartilage problems; abrupt large effusion with an injury history raises concern for new injury. The Delaware rule is to contact the physician when effusion remains above grade 2 for a prolonged period.

Unable to activate the quadriceps; lag on straight-leg raise. First check effusion and pain, then whether stimulation intensity is adequate (parameters in Section 6), then training volume. The goal at this stage is restoring contraction; weight-bearing is not yet required.

Pain location changes. For anterior knee pain after patellar-tendon harvest and harvest-site pain after quadriceps-tendon harvest, Delaware uses tendinopathy-style load management: adjust angle and load, observe the response to unloading, and generally do not stop training entirely; if symptoms still worsen after unloading, or if effusion accompanies them, reassess. Posterior thigh stretch pain in the first two to three months after hamstring harvest should first be checked against protocol limits at that stage for what was recently added: isometric resisted knee flexion should not have been present before week 6, dynamic resisted knee flexion before week 8, and no sprinting. If load exceeded limits, drop back to an allowed level and observe; if not, look elsewhere. This item is a clinical inference without trial data.

Strength no longer progresses. First check the testing method: leg-press 1RM overestimates, and a weakened uninjured side shrinks the denominator (see Sections 5 and 10). Then check whether load is heavy enough; Delaware’s 60%–85% 1RM, 5–8 repetitions per set, may serve as a reference. Strength plateaus at 3–6 months after surgery are common; patellar-tendon grafts take a mean of 30 weeks to reach 90%, and slow recovery is not the same as a plateau. Only after confirming a true plateau rather than merely slow progress should one revisit effusion and pain.

Giving way, sudden instability, or new injury. Early giving way is usually related to quadriceps weakness and effusion and commonly lessens as strength recovers, though this cannot be guaranteed for everyone. Recurrent giving way despite recovered strength, or a new injury history, warrants Lachman and pivot-shift testing and, if needed, MRI to assess graft integrity. This lies beyond rehabilitation adjustment.

Fear, reluctance to push. Scales such as ACL-RSI can quantify such concerns. Section 12 already noted that a self-directed app in an underpowered trial showed no detected effect on return-to-sport rate, and the materials for this chapter contain no trial of face-to-face psychological support. On present evidence, a feasible approach is to let the patient voice concerns and have the therapist break training down to a difficulty the patient can complete.

14. Duration of Rehabilitation and Return to Sport

Guidelines agree on rehabilitation duration: van Melick et al. give 9–12 months, the German consensus at least 9 months, Delaware at least 9 months (12 months for allograft), and the adolescent consensus 12 months. The “9 months” figure comes mainly from two observational cohort studies.

In the Delaware–Oslo cohort of Grindem et al. (2016; see Section 10), 22 of 74 (29.7%) who returned to level-I sport sustained a knee reinjury; 2 of 26 (7.7%) who did not return to level-I sport reinjured. After age adjustment, those who returned to level-I sport had a reinjury rate 4.32 times that of those who did not (95% CI 1.01 to 18.40, p=0.048).

Among those who returned to sport before 9 months after surgery, each month of delay reduced the reinjury rate by 51% (hazard ratio 0.49, 0.34 to 0.70); further delay after 9 months did not reach statistical significance for additional reduction (0.75, 0.49 to 1.14). Of 38 who returned before 9 months, 15 (39.5%) reinjured; of 36 who returned after 9 months, 7 (19.4%) reinjured; all 4 who returned within 5 months reinjured within two months.

“Reinjury” in the Grindem cohort was any acute knee injury—28 events in 24 people, including 8 graft ruptures, 8 medial meniscal injuries, and 2 contralateral ACL ruptures. The German consensus cites this study as “at least 9 months reduces retear,” narrowing the outcome. The risk model adjusted only for age; return to sport was defined as first participation, without distinguishing training from competition—a principal limitation the authors themselves listed.

The Swedish cohort of Beischer et al. (2020) enrolled 159 athletes aged 15–30 years. Those who returned before 9 months to knee-demanding sport (Tegner ≥6) had a hazard ratio for second ACL injury of 6.7 (2.6 to 16.7); 10 of 33 early returners were injured. The authors performed two sensitivity analyses. After removing the most influential 10% of events among these 159, the hazard ratio fell from 6.7 to 5.6 (2.1 to 16.7) and the association remained. Including eligible people who had not undergone strength testing (264 people, 20 events) gave a hazard ratio of 2.7 (1.1 to 6.7); after again removing 10% of events in this expanded sample, the association was no longer significant. Of 18 events, 12 occurred in people who returned to sport at postoperative months 8–9.

Time of return to sport relied on retrospective questionnaire recall; median follow-up was 15.5 months; only one third of those eligible entered the primary analysis; the sex-recording error noted earlier was corrected in July 2020. These limitations affect the precision of the multiples, not the direction: both cohorts show more reinjury among those who returned to high-load sport before 9 months.

By general GRADE rules, observational studies start at low certainty. German consensus item 5 is labeled A1 with high certainty in Table 2, yet the body on page 7 writes moderate certainty—an inconsistency without stated reasons for upgrading.

Nine months is a justified minimum time threshold; passing 9 months does not equal safety: 19.4% of those who returned after 9 months in the Grindem cohort still reinjured, and the event rate among post-9-month returners in the Beischer cohort was 3.9 per 100 person-years. Time must be used together with functional criteria: in the Grindem cohort, those who passed all criteria had a 5.6% reinjury rate versus 38.2% among those who did not—a difference that did not reach statistical significance; Kyritsis et al. (2016; based on the abstract) observed 26 graft ruptures (16.5%) among 158 male professional athletes who returned to their prior level, with rupture risk 4.1 times higher (95% CI 1.9–9.2) among those who had not met six discharge criteria than among those who had.

Continuing prevention training after return to sport is a shared recommendation across protocols; Delaware requires twice weekly and a knee-extension 1RM before and after each season. Brinlee et al. cite reinjury rates from this clinic’s ACL-SPORTS trial: 2.5% in men, 22.8% in women.

The women’s report (Johnson et al., 2020) enrolled 40 female athletes; 39 entered analysis; within two years after surgery, 9 sustained a second ACL injury (4 graft ruptures, 5 contralateral)—that is, 9/39=23%. All 9 occurred in patients who were under 20 at surgery and received a hamstring autograft.

That report gives different reinjury rates for the same subgroup. In age-stratified Table 3, 9 of 35 (25.7%) under 25, 9 of 32 (28.1%) under 20, and 8 of 26 (30.8%) under 18 reinjured. Table 4’s comparison with prior literature, however, writes the overall second-injury rate in the under-25 subgroup as 22.8%. Table 4’s contralateral-rupture column writes 14.2%, with a note “11.4% excluding contact injuries”; graft rupture of 11.4% plus that 11.4% equals 22.8%, so Table 4’s overall figure may come from that, but the table does not state that the overall column excludes contact injuries. Counting all second ACL injuries, the under-25 subgroup should be Table 3’s 9/35 (25.7%); Brinlee’s cited female reinjury rate of 22.8% is exactly Table 4’s figure. The discussion also writes an “overall second-injury rate of 26%” without a denominator, close to Table 3’s 25.7%, without stating which group is meant.

That trial compared two programs that both included prevention training (strength, agility, and jump training plus prevention, with or without perturbation training) and had no control arm without prevention training, so it cannot prove the preventive efficacy of this program.

The men’s report (Arundale et al., 2018) enrolled 40 male athletes with entry criteria of 3–9 months after surgery, quadriceps strength index above 80%, minimal effusion, no pain, normal range of motion, and completed running progression; the two training arms were combined for descriptive analysis. Patients passed return-to-sport criteria at a mean of 232±99 days after surgery; all had returned to sport by two years; within two years there was 1 second ACL injury (2.5%): a 32-year-old patient’s allograft ruptured late in the second postoperative year. Both reinjury and return to sport relied on self-report, without exposure data and without an untrained control.

Entry timing, grafts, and age composition differed between the men’s and women’s groups; the difference between 1/40 and 9/39 cannot be read as training being more effective for one sex. The most certain information from these two reports is that reinjury rates remain high when young women return to cutting sport.

For most patients the approximate course is: the first three months focus on restoring full extension, controlling swelling, and reactivating the quadriceps; months three to six progress strength training and begin running; months six to nine add jumping, cutting, and sport-specific work; after nine months, functional test results decide whether return to competition is allowed—if time has elapsed but tests are not met, clearance is still withheld. In practice the most common problems are a compressed strength-training phase and unreliable strength-testing methods.

Data Scope

Materials come from a targeted search in September 2026, centered on guidelines, consensus statements, and systematic reviews from 2023–2026, with selected primary studies and earlier literature. Primary references were based on the full text. Among supplementary and secondary citations, Heijne 2007, Toth 2020, Johnson 2020, Arundale 2018, and Tyler 1998 were based on the full text; Roldán 2017 on the methods, discussion, and Table 1 of the author-accepted manuscript; Mauro 2008, Sinacore 2017, Kyritsis 2016, Hauger 2018, Perriman 2018, Yang 2019, and Bynum 1995 on the abstract only; remaining items rest only on secondary citation and were not checked against the original, with sources noted at each entry. This is not a systematic review; studies not retrieved are outside the discussion; statements such as “only one trial” in the text are limited to the materials above.

References

Primary References

1. [1]

   Bafrouei MJ, et al. Effects of accelerated rehabilitation after ACL reconstruction on clinical and functional outcomes: A systematic review and meta-analysis evaluating clinical meaningfulness. J Exp Orthop. 2026. PMID 42565031; DOI 10.1002/jeo2.70878.

2. [2]

   Beischer S, Gustavsson L, Hamrin Senorski E, Karlsson J, Thomeé C, Samuelsson K, Thomeé R. Young Athletes Who Return to Sport Before 9 Months After Anterior Cruciate Ligament Reconstruction Have a Rate of New Injury 7 Times That of Those Who Delay Return. J Orthop Sports Phys Ther. 2020;50(2):83-90. PMID 32005095; DOI 10.2519/jospt.2020.9071.

3. [3]

   July 2020 Corrigendum (correction to the preceding entry). J Orthop Sports Phys Ther. 2020;50(7):411. PMID 32605465; DOI 10.2519/jospt.2020.50.7.411.

4. [4]

   Benelli C, et al. Restricted postoperative range-of-motion and weight-bearing protocols prevail after lateral meniscus posterior root tear repair with anterior cruciate ligament reconstruction: A systematic review. J Exp Orthop. 2025. PMID 41245733; DOI 10.1002/jeo2.70546.

5. [5]

   Bieler T, et al. The effects of high-intensity versus low-intensity resistance training on leg extensor power and recovery of knee function after ACL-reconstruction. Biomed Res Int. 2014;2014:278512. PMID 24877078; DOI 10.1155/2014/278512.

6. [6]

   Briem K, et al. Rehabilitation guidelines after autograft anterior cruciate ligament reconstruction need more graft-specific exercise recommendations-A scoping review. Knee Surg Sports Traumatol Arthrosc. 2026. PMID 40178127; DOI 10.1002/ksa.12666.

7. [7]

   Brinlee AW, Dickenson SB, Hunter-Giordano A, Snyder-Mackler L. ACL Reconstruction Rehabilitation: Clinical Data, Biologic Healing, and Criterion-Based Milestones to Inform a Return-to-Sport Guideline. Sports Health. 2022;14(5):770-779. PMID 34903114; DOI 10.1177/19417381211056873. Its supplementary material is the University of Delaware physical therapy clinic document Rehabilitation after ACL Reconstruction: Practice Guidelines (updated April 2021); “Delaware protocol” in this chapter refers to that document throughout.

8. [8]

   Culvenor AG, et al. Rehabilitation after anterior cruciate ligament and meniscal injuries: a best-evidence synthesis of systematic reviews for the OPTIKNEE consensus. Br J Sports Med. 2022;56(24):1445-1453. PMID 35768181; DOI 10.1136/bjsports-2022-105495.

9. [9]

   De Carlo MS, Shelbourne KD, McCarroll JR, Rettig AC. Traditional versus Accelerated Rehabilitation following ACL Reconstruction: A One-Year Follow-Up. J Orthop Sports Phys Ther. 1992;15(6):309-316. PMID 18781004; DOI 10.2519/jospt.1992.15.6.309.

10. [10]

    Erickson LN, et al. The Efficacy of Blood Flow Restriction Training to Improve Quadriceps Muscle Function after Anterior Cruciate Ligament Reconstruction. Med Sci Sports Exerc. 2025. PMID 39350350; DOI 10.1249/MSS.0000000000003573.

11. [11]

    Fältström A, Hedevik H, Ardern CL, Kvist J. Adding Self-directed Psychological Support to Usual Rehabilitation Following Anterior Cruciate Ligament Reconstruction Did Not Increase Return-to-Sport Rates at 12 Months. Results From the BAck iN the Game (BANG) Randomized Controlled Trial. J Orthop Sports Phys Ther. 2026;56(5):310-320. PMID 42063300; DOI 10.2519/jospt.2026.13749.

12. [12]

    Fan Z, et al. Delayed versus Accelerated Weight-bearing Rehabilitation Protocol Following Anterior Cruciate Ligament Reconstruction: A Systematic Review and Meta-analysis. J Rehabil Med. 2022. PMID 35037693; DOI 10.2340/jrm.v53.1438.

13. [13]

    Fontanier V, et al. Effect of open kinetic chain exercises during the first weeks of anterior cruciate ligament reconstruction rehabilitation: A systematic review and meta-analysis. Phys Ther Sport. 2025. PMID 39985872; DOI 10.1016/j.ptsp.2025.02.005.

14. [14]

    Fukuda TY, Fingerhut D, Moreira VC, et al. Open kinetic chain exercises in a restricted range of motion after anterior cruciate ligament reconstruction: a randomized controlled clinical trial. Am J Sports Med. 2013;41(4):788-794. PMID 23423316; DOI 10.1177/0363546513476482.

15. [15]

    Grindem H, Snyder-Mackler L, Moksnes H, Engebretsen L, Risberg MA. Simple decision rules can reduce reinjury risk by 84% after ACL reconstruction: the Delaware-Oslo ACL cohort study. Br J Sports Med. 2016;50(13):804-808. PMID 27162233; DOI 10.1136/bjsports-2016-096031.

16. [16]

    Häner M, et al. ACL rehabilitation after reconstruction: Evidence-based consensus recommendations from the German Knee Society-Part IV. J Exp Orthop. 2026. PMID 42549304; DOI 10.1002/jeo2.70875.

17. [17]

    Harris JD, Brand JC, Cote MP, Faucett SC, Dhawan A. Research Pearls: The Significance of Statistics and Perils of Pooling. Part 1: Clinical Versus Statistical Significance. Arthroscopy. 2017;33(6):1102-1112. PMID 28454999; DOI 10.1016/j.arthro.2017.01.053.

18. [18]

    Hu Y, et al. Blood Flow Restriction Training After Anterior Cruciate Ligament Reconstruction: A Systematic Review and Meta-analysis of Randomized Controlled Trials. Orthop J Sports Med. 2026. PMID 42428806; DOI 10.1177/23259671261440196.

19. [19]

    Khubzan WD, et al. Comparison between home-based and supervised rehabilitation protocols after anterior cruciate ligament reconstruction: a systematic review and meta-analysis. EFORT Open Rev. 2025. PMID 40905919; DOI 10.1530/EOR-2024-0216.

20. [20]

    Kotsifaki R, Korakakis V, King E, et al. Aspetar clinical practice guideline on rehabilitation after anterior cruciate ligament reconstruction. Br J Sports Med. 2023;57(9):500-514. PMID 36731908; DOI 10.1136/bjsports-2022-106158.

21. [21]

    Li Z, et al. Effects of Neuromuscular Electrical Stimulation on Quadriceps Femoris Muscle Strength and Knee Joint Function in Patients After ACL Surgery: A Systematic Review and Meta-analysis of Randomized Controlled Trials. Orthop J Sports Med. 2025;13(1). PMID 39811154; DOI 10.1177/23259671241275071.

22. [22]

    Lin Q, et al. Effects of postoperative knee bracing on knee function and stability after anterior cruciate ligament reconstruction: A systematic review and meta-analysis. Knee Surg Sports Traumatol Arthrosc. 2025. PMID 41144771; DOI 10.1002/ksa.70098.

23. [23]

    Noll S, Garrison JC, Bothwell J, Conway JE. Knee Extension Range of Motion at 4 Weeks Is Related to Knee Extension Loss at 12 Weeks After Anterior Cruciate Ligament Reconstruction. Orthop J Sports Med. 2015;3(5). PMID 26675061; DOI 10.1177/2325967115583632.

24. [24]

    Pairot de Fontenay B, et al. Reintroduction of Running After Anterior Cruciate Ligament Reconstruction With a Hamstrings Graft: Can We Predict Short-Term Success? J Athl Train. 2022;57(6). PMID 34623431; DOI 10.4085/1062-6050-0407.21.

25. [25]

    Pujol N, Giordano AO, et al. The formal EU-US Meniscus Rehabilitation 2024 Consensus: An ESSKA-AOSSM-AASPT initiative. Part I-Rehabilitation management after meniscus surgery (meniscectomy, repair and reconstruction). Knee Surg Sports Traumatol Arthrosc. 2025. PMID 40353298; DOI 10.1002/ksa.12674. (The same consensus also appears in Orthop J Sports Med, PMID 40416991.)

26. [26]

    Rambaud AJM, Ardern CL, Thoreux P, Regnaux JP, Edouard P. Criteria for return to running after anterior cruciate ligament reconstruction: a scoping review. Br J Sports Med. 2018;52(22):1437-1444. PMID 29720478; DOI 10.1136/bjsports-2017-098602.

27. [27]

    Risberg MA, Holm I, Myklebust G, Engebretsen L. Neuromuscular training versus strength training during first 6 months after anterior cruciate ligament reconstruction: a randomized clinical trial. Phys Ther. 2007;87(6):737-750. PMID 17442840; DOI 10.2522/ptj.20060041.

28. [28]

    Robert H, Nouveau S, Gageot S, Gagnière B. A new knee arthrometer, the GNRB: experience in ACL complete and partial tears. Orthop Traumatol Surg Res. 2009;95(3):171-176. PMID 19423416; DOI 10.1016/j.otsr.2009.03.009.

29. [29]

    Schoepp C, et al. Brace-Free Rehabilitation after Isolated Anterior Cruciate Ligament Reconstruction with Hamstring Tendon Autograft Is Not Inferior to Brace-Based Rehabilitation-A Randomised Controlled Trial. J Clin Med. 2023;12(5):2074. PMID 36902868; DOI 10.3390/jcm12052074.

30. [30]

    Shelbourne KD, Klootwyk TE, DeCarlo MS. Update on accelerated rehabilitation after anterior cruciate ligament reconstruction. J Orthop Sports Phys Ther. 1992;15(6):303-308. PMID 18781000; DOI 10.2519/jospt.1992.15.6.303.

31. [31]

    van Melick N, van Cingel REH, Brooijmans F, et al. Evidence-based clinical practice update: practice guidelines for anterior cruciate ligament rehabilitation based on a systematic review and multidisciplinary consensus. Br J Sports Med. 2016;50(24):1506-1515. PMID 27539507; DOI 10.1136/bjsports-2015-095898.

32. [32]

    van Melick N, et al. Anterior Cruciate Ligament Rehabilitation for the 10- to 18-Year-Old Adolescent Athlete: Practice Guidelines Based on International Delphi Consensus. Orthop J Sports Med. 2023;11(7). PMID 37492781; DOI 10.1177/23259671231172454.

33. [33]

    Wackerle AM, et al. Freddie Fu Panther Symposium Expert Group 2024: Rehabilitation and return to sport after anterior cruciate ligament reconstruction Part 1: Early and intermediate phases of rehabilitation. Knee Surg Sports Traumatol Arthrosc. 2026. PMID 41186026; DOI 10.1002/ksa.70115.

34. [34]

    Wackerle AM, et al. Freddie Fu Panther Symposium Expert Group 2024: Rehabilitation and return to sport after anterior cruciate ligament reconstruction Part 2: Concepts and emerging technology in return to sport. Knee Surg Sports Traumatol Arthrosc. 2026. PMID 41169247; DOI 10.1002/ksa.70110.

35. [35]

    Wellsandt E, Failla MJ, Snyder-Mackler L. Limb Symmetry Indexes Can Overestimate Knee Function After Anterior Cruciate Ligament Injury. J Orthop Sports Phys Ther. 2017;47(5):334-338. PMID 28355978; DOI 10.2519/jospt.2017.7285.

36. [36]

    Zheng QY, et al. Blood flow restriction as an adjunct during mid-stage rehabilitation after ACL reconstruction: a randomized sham-controlled study. BMC Musculoskelet Disord. 2026. PMID 42092903; DOI 10.1186/s12891-026-09908-5.

Supplementary and Secondary Citations

• •

  Adams D, Logerstedt D, Hunter-Giordano A, Axe MJ, Snyder-Mackler L. Current concepts for anterior cruciate ligament reconstruction: a criterion-based rehabilitation progression. J Orthop Sports Phys Ther. 2012;42(7):601-614. PMID 22402434; DOI 10.2519/jospt.2012.3871. (cited via Brinlee et al., 2022, and its supplementary material)

• •

  Arundale AJH, Capin JJ, Zarzycki R, Smith AH, Snyder-Mackler L. Two year ACL reinjury rate of 2.5%: outcomes report of the men in a secondary ACL injury prevention program (ACL-SPORTS). Int J Sports Phys Ther. 2018;13(3):422-431. PMID 30038828; DOI 10.26603/ijspt20180422 (as listed in the journal running head; not listed in the PubMed citation). (full text)

• •

  Beynnon BD, Fleming BC, Johnson RJ, Nichols CE, Renström PA, Pope MH. Anterior cruciate ligament strain behavior during rehabilitation exercises in vivo. Am J Sports Med. 1995;23(1):24-34. PMID 7726347; DOI 10.1177/036354659502300105. (reference 12 in Brinlee et al., 2022; original unread)

• •

  Beynnon BD, et al. Rehabilitation after anterior cruciate ligament reconstruction: a prospective, randomized, double-blind comparison of programs administered over 2 different time intervals. Am J Sports Med. 2005;33(3):347-359. PMID 15716250; DOI 10.1177/0363546504268406. (cited via Fan et al., 2022)

• •

  Bynum EB, Barrack RL, Alexander AH. Open versus closed chain kinetic exercises after anterior cruciate ligament reconstruction. A prospective randomized study. Am J Sports Med. 1995;23(4):401-406. PMID 7573647; DOI 10.1177/036354659502300405. (abstract)

• •

  Charles D, et al. A systematic review of the effects of blood flow restriction training on quadriceps muscle atrophy and circumference post ACL reconstruction. Int J Sports Phys Ther. 2020;15(6):882-891. PMID 33344004; DOI 10.26603/ijspt20200882. (cited via Culvenor et al., 2022)

• •

  Christensen JC, Goldfine LR, West HS. The effects of early aggressive rehabilitation on outcomes after anterior cruciate ligament reconstruction using autologous hamstring tendon: a randomized clinical trial. J Sport Rehabil. 2013;22(3):191-201. PMID 23579334; DOI 10.1123/jsr.22.3.191. (cited via Fan et al., 2022)

• •

  Delitto A, Rose SJ, McKowen JM, et al. Electrical stimulation versus voluntary exercise in strengthening thigh musculature after anterior cruciate ligament surgery. Phys Ther. 1988;68(5):660-663. PMID 3258994; DOI 10.1093/ptj/68.5.660. (cited via Li et al., 2025)

• •

  Gamble AR, et al. Intensive supervised rehabilitation versus less supervised rehabilitation following anterior cruciate ligament reconstruction? A systematic review and meta-analysis. J Sci Med Sport. 2021;24(9):862-870. PMID 33736965; DOI 10.1016/j.jsams.2021.03.003. (cited via Culvenor et al., 2022)

• •

  Hauger AV, Reiman MP, Bjordal JM, Sheets C, Ledbetter L, Goode AP. Neuromuscular electrical stimulation is effective in strengthening the quadriceps muscle after anterior cruciate ligament surgery. Knee Surg Sports Traumatol Arthrosc. 2018;26(2):399-410. PMID 28819679; DOI 10.1007/s00167-017-4669-5. (abstract)

• •

  Heijne A, Werner S. Early versus late start of open kinetic chain quadriceps exercises after ACL reconstruction with patellar tendon or hamstring grafts: a prospective randomized outcome study. Knee Surg Sports Traumatol Arthrosc. 2007;15(4):402-414. PMID 17219226; DOI 10.1007/s00167-006-0246-z. (full text) Erratum: Knee Surg Sports Traumatol Arthrosc. 2007;15(4):472-473. PMID 27520629; DOI 10.1007/s00167-007-0313-0.

• •

  Henriksson M, Rockborn P, Good L. Range of motion training in brace vs. plaster immobilization after anterior cruciate ligament reconstruction: a prospective randomized comparison with a 2-year follow-up. Scand J Med Sci Sports. 2002;12(2):73-80. PMID 12121424; DOI 10.1034/j.1600-0838.2002.120203.x. (cited via Fan et al., 2022)

• •

  Johnson JL, Capin JJ, Arundale AJH, Zarzycki R, Smith AH, Snyder-Mackler L. A Secondary Injury Prevention Program May Decrease Contralateral Anterior Cruciate Ligament Injuries in Female Athletes: 2-Year Injury Rates in the ACL-SPORTS Randomized Controlled Trial. J Orthop Sports Phys Ther. 2020;50(9):523-530. PMID 32741328; DOI 10.2519/jospt.2020.9407. (full text)

• •

  Kyritsis P, Bahr R, Landreau P, Miladi R, Witvrouw E. Likelihood of ACL graft rupture: not meeting six clinical discharge criteria before return to sport is associated with a four times greater risk of rupture. Br J Sports Med. 2016;50(15):946-951. PMID 27215935; DOI 10.1136/bjsports-2015-095908. (abstract)

• •

  Luo Y, et al. Treadmill training with partial body-weight support after anterior cruciate ligament reconstruction: a randomized controlled trial. J Phys Ther Sci. 2016;28(12). PMID 28174445; DOI 10.1589/jpts.28.3325. (cited via Fan et al., 2022)

• •

  Mauro CS, Irrgang JJ, Williams BA, Harner CD. Loss of extension following anterior cruciate ligament reconstruction: analysis of incidence and etiology using IKDC criteria. Arthroscopy. 2008;24(2):146-153. PMID 18237697; DOI 10.1016/j.arthro.2007.08.026. (abstract)

• •

  Paternostro-Sluga T, et al. Neuromuscular electrical stimulation after anterior cruciate ligament surgery. Clin Orthop Relat Res. 1999. PMID 10613165. (cited via Li et al., 2025)

• •

  Perriman A, Leahy E, Semciw AI. The Effect of Open- Versus Closed-Kinetic-Chain Exercises on Anterior Tibial Laxity, Strength, and Function Following Anterior Cruciate Ligament Reconstruction: A Systematic Review and Meta-analysis. J Orthop Sports Phys Ther. 2018;48(7):552-566. PMID 29685058; DOI 10.2519/jospt.2018.7656. (abstract)

• •

  Revenäs Å, Johansson A, Leppert J. A randomized study of two physiotherapeutic approaches after knee ligament reconstruction. Adv Physiother. 2009;11(1):30-41. DOI 10.1080/14038190801999497 (Crossref registration). (cited via Khubzan et al., 2025)

• •

  Roldán E, Reeves ND, Cooper G, Andrews K. In vivo mechanical behaviour of the anterior cruciate ligament: A study of six daily and high impact activities. Gait Posture. 2017;58:201-207. PMID 28802221; DOI 10.1016/j.gaitpost.2017.07.123. (author-accepted manuscript)

• •

  Salem HS, Park DH, Friedman JL, et al. Return to Driving After Anterior Cruciate Ligament Reconstruction: A Systematic Review. Orthop J Sports Med. 2021;9(1). PMID 33553439; DOI 10.1177/2325967120968556. (cited via Kotsifaki et al., 2023)

• •

  Sinacore JA, Evans AM, Lynch BN, Joreitz RE, Irrgang JJ, Lynch AD. Diagnostic Accuracy of Handheld Dynamometry and 1-Repetition-Maximum Tests for Identifying Meaningful Quadriceps Strength Asymmetries. J Orthop Sports Phys Ther. 2017;47(2):97-107. PMID 28142362; DOI 10.2519/jospt.2017.6651. (abstract)

• •

  Snyder-Mackler L, et al. Electrical stimulation of the thigh muscles after reconstruction of the anterior cruciate ligament. Effects of electrically elicited contraction of the quadriceps femoris and hamstring muscles on gait and on strength of the thigh muscles. J Bone Joint Surg Am. 1991;73(7):1025-1036. PMID 1874764. (cited via Li et al., 2025)

• •

  Sturgill LP, Snyder-Mackler L, Manal TJ, Axe MJ. Interrater reliability of a clinical scale to assess knee joint effusion. J Orthop Sports Phys Ther. 2009;39(12):845-849. PMID 20032559; DOI 10.2519/jospt.2009.3143. (cited via Fältström et al., 2026)

• •

  Syed RIB, et al. Comparative Effectiveness of Supervised and Home-Based Rehabilitation after Anterior Cruciate Ligament Reconstruction in Competitive Athletes. J Clin Med. 2024;13(8):2245. PMID 38673520; DOI 10.3390/jcm13082245. (cited via Khubzan et al., 2025)

• •

  Tajima T, Yamaguchi N, Nagasawa M, Morita Y, Nakamura Y, Chosa E. Early weight-bearing after anterior cruciate ligament reconstruction with hamstring grafts induce femoral bone tunnel enlargement: a prospective clinical and radiographic study. BMC Musculoskelet Disord. 2019;20:274. PMID 31159789; DOI 10.1186/s12891-019-2653-6. (cited via Fan et al., 2022)

• •

  Thomas B, et al. Patients with meniscus posterolateral root tears repair during ACL reconstruction achieve comparable post-operative outcome than patients with isolated ACL reconstruction. Knee Surg Sports Traumatol Arthrosc. 2023;31(8):3405-3411. PMID 37079048; DOI 10.1007/s00167-023-07415-x. (cited via Benelli et al., 2025)

• •

  Toth MJ, Tourville TW, Voigt TB, et al. Utility of Neuromuscular Electrical Stimulation to Preserve Quadriceps Muscle Fiber Size and Contractility After Anterior Cruciate Ligament Injuries and Reconstruction: A Randomized, Sham-Controlled, Blinded Trial. Am J Sports Med. 2020;48(10):2429-2437. PMID 32631074; DOI 10.1177/0363546520933622. (full text)

• •

  Tyler TF, McHugh MP, Gleim GW, Nicholas SJ. The effect of immediate weightbearing after anterior cruciate ligament reconstruction. Clin Orthop Relat Res. 1998;(357):141-148. PMID 9917711; DOI 10.1097/00003086-199812000-00019. (full text; original journal scan released by the authors’ institution, NISMAT)

• •

  Wiggins AJ, Grandhi RK, Schneider DK, Stanfield D, Webster KE, Myer GD. Risk of Secondary Injury in Younger Athletes After Anterior Cruciate Ligament Reconstruction: A Systematic Review and Meta-analysis. Am J Sports Med. 2016;44(7):1861-1876. PMID 26772611; DOI 10.1177/0363546515621554. (cited via Beischer et al., 2020)

• •

  Yang XG, Feng JT, He X, Wang F, Hu YC. The effect of knee bracing on the knee function and stability following anterior cruciate ligament reconstruction: A systematic review and meta-analysis of randomized controlled trials. Orthop Traumatol Surg Res. 2019;105(6):1107-1114. PMID 31279767; DOI 10.1016/j.otsr.2019.04.015. (abstract)