Research · ACL Rehabilitation

Return-to-Play Testing After ACL Reconstruction: What Hop Tests and Limb Symmetry Actually Predict

By Muhammad Dzulhisham · MSc Sports Coaching, NSCA-CSCS

A separate article on this site looked at the broad case for return-to-play testing: why "the pain is gone" was never a sufficient standard, and why objective, criteria-based clearance replaced time-based clearance in modern ACL rehabilitation. This article asks a narrower and more clinical question. Once an athlete is standing in front of a battery of hop tests, a limb symmetry index calculation, and an isokinetic dynamometer, how well does passing that specific battery actually predict that they will not tear the graft again, or injure the other knee, in the following one to two years?

The honest answer, built from the peer-reviewed literature rather than from the marketing language around "cleared to return," is more complicated than most clinics let on. Some of the components of the standard battery carry real predictive signal. Others carry far less than their widespread use would suggest, and at least one of them — the 90% limb symmetry index cutoff — has been directly tested for its ability to discriminate a safe return from an unsafe one, and performed close to chance.

What the standard battery actually measures

Across clinics, the objective return-to-sport battery after ACL reconstruction has converged on a fairly consistent set of components, described in a 2019 scoping review of 209 studies that catalogued the criteria used worldwide.⁹ That review found the field organizes criteria into six domains — time, strength, hop testing, clinical examination, patient-reported outcomes, and performance-based measures — but that time and simple impairment-based measures dominate, despite sport being a far more complex physical and cognitive task than a static test can capture.⁹

Hop tests

The most common protocol is a battery of four single-leg hop tests: the single hop for distance, the triple hop for distance, the crossover hop for distance, and the 6-meter timed hop. Each is performed on the surgical and non-surgical limb, and the result is expressed as a percentage of the surgical limb's performance relative to the non-surgical limb.

Limb symmetry index (LSI)

This is the ratio at the center of nearly every hop and strength test: (surgical limb score ÷ non-surgical limb score) × 100. A result of 90% or higher on both hop testing and quadriceps strength has been the de facto discharge threshold in research and clinical practice for over a decade, popularized in large part by cohort work out of the University of Delaware and Oslo.

Isokinetic strength testing

Isokinetic dynamometry measures peak torque of the quadriceps and hamstrings at fixed angular velocities, generating both an LSI for each muscle group and a hamstring-to-quadriceps (H:Q) strength ratio, typically expected in the 55–65% range.

The evidence that started the trend: testing does discriminate risk

The strongest and most widely cited support for objective testing comes from the Delaware-Oslo ACL cohort study, which followed 106 athletes returning to pivoting sport. Athletes who passed a set of discharge criteria — quadriceps strength LSI, hop test LSI, and self-reported knee function, each at or above 90% — had a substantially lower second-injury rate than those who did not, and each month return to sport was delayed toward nine months post-surgery reduced reinjury risk by roughly half up to that point.¹ A separate cohort of 158 male professional athletes, evaluated with isokinetic testing, three single-leg hop tests, and an agility test, found that those who did not meet all six discharge criteria had roughly four times the risk of graft rupture (hazard ratio 4.1, p<0.001) compared with those who met every criterion, and a lower hamstring-to-quadriceps ratio at 60°/s was independently associated with rupture risk.²

These two studies are the foundation of the modern criteria-based return-to-sport model, and they are genuinely important: they showed, for the first time in prospective data, that objective numbers at the point of clearance are not just a formality but carry measurable prognostic value.

Where the picture gets more complicated

The trouble is that neither of those two cohorts, on their own, proves that the testing battery in general use today is a reliable go/no-go gate. Later systematic reviews and larger meta-analyses have tried to pool the evidence across studies, and the results are considerably less reassuring.

A 2019 systematic review and meta-analysis in the Journal of Orthopaedic & Sports Physical Therapy searched for every available study on whether passing return-to-sport criteria was associated with a lower risk of a second ACL injury. Only four studies met inclusion criteria — a sign of how thin the evidence base actually was at the time — and the pooled result found no statistically significant association between passing criteria and reduced second-injury risk. The authors rated the overall quality of evidence as too low to draw a firm conclusion either way.³

A larger 2024 meta-analysis in PeerJ, pooling nine studies and 1,410 patients, reached a more specific and more sobering conclusion. Passing return-to-sport testing was not significantly associated with a reduced risk of overall knee injury, secondary ACL injury, or contralateral ACL injury. The one outcome for which passing testing showed a statistically significant protective association was graft rupture specifically (odds ratio 0.49, 95% CI 0.33–0.75, p=0.001) — a real and meaningful signal, but a narrower one than "passing the tests means you're safe to return."⁴ The authors were candid about why the picture is muddy: RTS test batteries and thresholds vary substantially between studies, most of the underlying data comes from observational cohorts without adjustment for confounders, and results even diverge by region, with protective associations more consistent in European cohorts than in some United States-based studies.⁴

"The use of LSI from tests of muscle function to determine safe return to sport after ACL reconstruction … cannot differentiate between athletes who had a safe return to sport and those who did not, regardless of whether LSI was used as cut-offs, incremental [thresholds], or as deviation from symmetry."

Why limb symmetry, specifically, is the weakest link

Of the three legs of the standard battery, the limb symmetry index has drawn the sharpest and most recent criticism, for a structural reason: it only ever compares the surgical leg to the non-surgical leg. If both legs are weak, the ratio can still look perfect.

A widely cited 2017 study followed 70 athletes through quadriceps strength and four hop tests before injury and again at six months after reconstruction. Just over half (57.1%) met the standard 90% LSI threshold on strength and all four hop tests. But when the same six-month surgical-limb scores were compared not to the non-surgical limb, but to an estimate of the athlete's own pre-injury capacity, only 28.6% cleared that bar — meaning LSI was systematically flattering recovery relative to what the athlete could actually do before getting hurt.⁵ The consequence for injury prediction was stark: of the 11 athletes in that cohort who went on to suffer a second ACL injury, 8 of them (72.7%) had already passed the standard 90% LSI criterion at six months. The pre-injury-capacity comparison was a far more sensitive predictor of who would go on to a second injury (sensitivity 0.818) than LSI was (sensitivity 0.273).⁵

A separate study of 115 young athletes already cleared to return to sport asked a simpler question: of those cleared clinically, how many actually met the full recommended set of quantitative cutoffs? The proportions meeting individual test cutoffs ranged from 43.5% to 78.3%; only 53.0% met all four hop test cutoffs; only 27.8% met all strength cutoffs; and only 13.9% met every cutoff in combination.⁶ Athletes are routinely being cleared using a mix of judgment and partial criteria, not the full quantitative battery the research literature assumes is being applied.

The most direct test yet of LSI's discriminative ability came from a 2024–2025 analysis of 233 athletes that asked whether any way of using LSI — as a fixed cutoff, as an incremental threshold, or as a continuous deviation from perfect symmetry — could actually separate athletes who went on to a safe return from those who suffered a second ACL injury within two years. The best-performing cutoffs it could find had an area under the curve of only 0.50–0.59 (0.50 is equivalent to a coin flip) and Youden's J index values of just 0.09–0.24, indicating very weak discrimination.⁷ Part of the explanation is that LSI is measuring a moving target: the "healthy" reference limb can itself decline in strength and hop performance over the same rehabilitation period, for reasons ranging from detraining to guarding, which can make a genuinely under-recovered surgical limb look symmetrical by comparison.⁷

Isokinetic strength ratios: a signal that doesn't always replicate

The hamstring-to-quadriceps ratio has an intuitive rationale — the hamstrings act as a dynamic restraint to anterior tibial translation, so a low ratio should theoretically leave the reconstructed knee less protected. The professional athlete cohort described above did find an independent association between a lower H:Q ratio and graft rupture risk.² But that finding has not replicated cleanly in larger, more general samples. A 2024 cohort study following 574 patients back to their pre-injury sport level found that a higher H:Q ratio was associated with lower odds of a second ACL injury in unadjusted analysis, but after accounting for other prognostic factors, the association was no longer statistically significant, and the ratio's ability to discriminate who would go on to a second injury was poor (ROC area under the curve of 0.60, only modestly better than chance).⁸ The authors concluded the ratio should not be used in isolation to make return-to-sport decisions.⁸ The likeliest explanation for the discrepancy is context: elite, high-load professional soccer is a different re-injury environment than the general ACL reconstruction population most cohorts are drawn from, and a ratio that carries signal in one setting may carry much less in another.

What this means for how a battery should actually be used

None of this means objective testing is worthless — the opposite would be a worse mistake. Delayed time to return, a rehabilitation program built around measurable strength and hop symmetry, and the discipline of a structured discharge checklist are all associated with better outcomes in the highest-quality cohort data available.^1,2 What the evidence supports is a more modest and more honest claim than "pass these three tests and you're cleared": a testing battery is a useful, partial, and imperfect risk-reduction tool, not a certificate of safety.

In practice, that argues for several adjustments over relying on a single 90% LSI number in isolation:

• Treat quadriceps and hamstring strength relative to estimated pre-injury capacity, not only relative to the uninjured limb, since the uninjured limb is not a fixed reference point.⁵
• Weight isokinetic strength symmetry and absolute strength more heavily than hop test symmetry alone, since hop performance can mask underlying strength deficits and vice versa.
• Use time since surgery as a floor, not a target — the data support a minimum of roughly nine months for pivoting-sport athletes, independent of whether tests are passed earlier.¹
• Build testing into a broader battery that also captures movement quality, sport-specific and reactive/cognitive-loaded tasks, and psychological readiness, rather than the impairment-based measures that still dominate published clearance criteria.^9,10
• Continue monitoring load and mechanics after clearance rather than treating a passed test as the end of risk management, given that even athletes who pass standard criteria retain a meaningful residual re-injury rate.^3,4,7

The research on functional movement screening on this site made a related point about a different tool: a good test can flag risk without being a reliable predictor of any single injury in any single athlete. The same caution applies here, arguably with higher stakes. A hop test, an LSI calculation, and an isokinetic printout are genuinely useful pieces of information. They are not, on the evidence currently available, a number that tells a clinician or an athlete that the knee is safe.