ACL Graft Choice: BTB vs Hamstring vs Quadriceps Tendon — Evidence and the Exam Answer

Every ACL reconstruction begins with a decision the patient lives with for decades: which tendon to take. ACL graft choice between bone–patellar tendon–bone (BTB), four-strand hamstring and quadriceps tendon autograft is a standard viva question, and the evidence answers it more clearly than the argument around it suggests. All three restore stability and function to a similar degree. BTB has a small, consistent revision advantage and the most kneeling pain. Hamstring has little anterior morbidity and a higher failure risk than BTB in young pivoting athletes. Quadriceps tendon looks comparable on shorter follow-up and has a learning curve. The background is on the ACL injuries topic page.

What does each autograft give, and what does it cost?

Each graft trades fixation and healing against donor-site morbidity. The table sets out the trade; the sections after it give the evidence.

Allograft sits outside the table because, in a young patient, it is not an equal choice. That is covered below.

Does BTB fail less often than hamstring?

Slightly, and the answer depends on which kind of study you read.

Randomised trials see no difference in re-rupture. The Cochrane review by Mohtadi pooled 19 trials and 1,597 young to middle-aged adults. There was no significant difference in hop test, return to activity, Tegner, Lysholm, IKDC or re-rupture. Every static stability test (instrumented, Lachman and pivot shift) favoured patellar tendon. Patellar tendon patients had more anterior knee problems, especially with kneeling, and a significant loss of extension; hamstring patients had a significant loss of flexion strength. Many of the trials were at high risk of bias, and long-term data were too thin to judge osteoarthritis.

Registries see a small, consistent difference in revision. Gifstad combined the Scandinavian registries: 45,998 primary reconstructions, of which 84.1% used hamstring and 14.6% patellar tendon. The revision hazard rate ratio for patellar tendon against hamstring was 0.63 (95% CI 0.53–0.74), consistent across sex, age and cartilage-injury subgroups. The effect looked slightly stronger after soccer, handball and alpine injuries (0.57 against 0.81), but the test for interaction did not reach significance (P = .058). Persson, from the Norwegian registry (12,643 patients), found an adjusted hazard ratio for revision of 2.3 (95% CI 1.8–3.0) for hamstring against patellar tendon. In the same model, patients aged 15–19 had a hazard ratio of 4.0 against those aged 30 or over, and sex had no effect.

Pooled, the difference is real and small. Samuelsen's meta-analysis of 47,613 reconstructions (14 RCTs, 10 prospective cohorts and one registry) found ruptures in 2.80% of BTB and 2.84% of hamstring grafts, with an odds ratio of 0.83 (95% CI 0.72–0.96). The number needed to treat was 235: that many patients would need BTB rather than hamstring to prevent one rupture. Laxity, pivot shift and Lachman did not differ significantly.

Why trials and registries disagree is partly arithmetic. Revision is a rare outcome, and 1,597 patients cannot detect a hazard ratio near 0.6 that tens of thousands can. Registries pay for that size with confounding: they record revision rather than rupture, and they cannot separate the graft from the surgeon, the fixation and the patient who chose it. The quadriceps tendon operative guide puts it plainly: registry revision rates are useful signals, but they cannot establish that graft choice alone caused failure. Hold both findings together. BTB revises a little less, and age and activity matter more than the graft.

The guideline wording follows the evidence. The AAOS guideline (2022 update, summarised by Brophy and Lowry in 2023) says that for skeletally mature patients surgeons may favour BTB to reduce the risk of graft failure or infection, or hamstring to reduce the risk of anterior or kneeling pain. That is a moderate-strength recommendation, a clarification of the 2013 guideline, which had said on strong evidence that outcomes were similar.

Why does hamstring graft diameter matter?

Because small hamstring grafts fail more, and the effect concentrates in the patients most likely to fail anyway.

Magnussen followed 256 hamstring reconstructions. Revision was needed in 1 of 58 (1.7%) grafts over 8 mm, 9 of 139 (6.5%) at 7.5–8 mm and 8 of 59 (13.6%) at 7 mm or less (PubMed (opens in a new tab)). Sixteen of the 18 revisions were in patients under 20 with grafts of 8 mm or less. In the MOON cohort reported by Mariscalco, 14 of 199 (7.0%) grafts of 8 mm or less were revised within 2 years, against 0 of 64 larger grafts. In patients aged 18 or younger the figures were 13 of 71 (18.3%) against 0 of 14. Both are Level III cohorts with short follow-up, and Mariscalco's authors say a larger sample is needed to confirm the revision relationship. The direction is consistent enough to act on.

The practical rule is to size the graft before drilling. If a quadrupled hamstring falls short, the hamstring operative guide lists the options: add strands, or switch to quadriceps tendon or BTB. The same guide notes that small grafts are commonest in smaller patients, so the problem is better anticipated at consent than discovered on the back table.

Where does the quadriceps tendon fit?

Quadriceps tendon is used more and more because it offers a large soft-tissue graft without the kneeling cost of BTB. The comparative evidence supports it as an equal option. It does not show it to be better.

Mouarbes pooled 27 studies (2,856 patients). Against BTB (581 against 514 grafts) there was no significant difference in instrumented laxity, Lachman, pivot shift, Lysholm, subjective IKDC or graft failure, and donor-site pain was a quarter as common with quadriceps tendon (risk ratio 0.25, 95% CI 0.18–0.36). Against hamstring (181 against 176) the only difference was a slightly higher mean Lysholm score with quadriceps tendon (mean difference 3.81 points).

Ebert's randomised trial allocated 112 patients to quadriceps or hamstring autograft. At 24 months laxity and patient-reported outcomes did not differ, except that the ACL Return to Sport after Injury score was better in the hamstring group at 3, 6 and 12 months. Each graft left a weakness at its own donor site. The hamstring group had better quadriceps strength symmetry at 6 and 12 months and better hop symmetry, mainly at 6 months; the quadriceps group had better hamstring strength symmetry at 6, 12 and 24 months. One graft re-tore within 2 years, in the quadriceps group. Rehabilitation and return-to-sport testing should expect an extensor deficit after quadriceps harvest; the quadriceps tendon operative guide recommends serial strength testing rather than reading early weakness as failure.

The registry caution comes from Denmark. Lind compared 531 quadriceps, 14,213 hamstring and 1,835 patellar tendon reconstructions from 2005–2017. Two-year revision was 4.7% for quadriceps tendon against 2.3% for hamstring and 1.5% for patellar tendon, with slightly more laxity and positive pivot shift. The same group then split quadriceps cases by clinic routine. Clinics with 100 or fewer procedures over 2012–2019 had a revision rate of 6.4%, against 2.9% in higher-volume clinics (adjusted hazard ratio 2.3). Their conclusion was that quadriceps tendon is not associated with a high revision rate when performed routinely, and that the learning curve is an important factor. That is a registry inference, not a trial result. The useful message for a surgeon adopting the graft is to expect a learning curve and audit the early cases.

What about allograft?

In a young patient, allograft is the wrong answer in the exam and, on current evidence, usually in practice.

In a MOON prospective cohort, Kaeding found that the odds of graft rupture after allograft were about four times those after autograft, and that the odds rose 2.3-fold for each 10-year decrease in age (PubMed (opens in a new tab)). Patients aged 10–19 had the highest failure rate. The AAOS guideline recommends that surgeons consider autograft over allograft, particularly in young and/or active patients, and grades it strong; the 2013 version had treated the two as similar. At revision, the MARS cohort of 1,205 cases found autograft 2.78 times less likely to re-rupture than allograft at 2 years.

Allograft still has a place: no donor site and a shorter operation for the selected older, lower-demand patient or the multiligament knee that needs several grafts, after an informed discussion.

Does a lateral extra-articular tenodesis change the choice?

It changes what a hamstring graft can be asked to do. STABILITY randomised 618 patients aged 25 or under (mean 18.9 years) to hamstring reconstruction with or without a lateral extra-articular tenodesis (LET). All had at least two of three risk factors: a grade 2 or greater pivot shift, a plan to return to pivoting sport, or generalised ligamentous laxity. At 2 years graft rupture fell from 11% to 4% and the composite of clinical failure from 40% to 25%. The number needed to treat to prevent one rupture was 14.3. The price was early pain: patients without LET had less pain at 3 months, and KOOS pain still favoured them at 6 months; there were no clinically important differences in patient-reported outcomes at other time points. The detail of the tenodesis and its indications is on the anterolateral ligament and LET topic page.

Marmura used the STABILITY data to validate the MOON graft-rupture risk calculator. Age, high-grade preoperative laxity and graft type predicted rupture, and both BTB and hamstring plus LET were protective against hamstring alone. The authors concluded that isolated hamstring grafts should be avoided in young active patients. The AAOS guideline treats ALL reconstruction or LET with a hamstring graft as something that could be considered in selected patients, on moderate evidence, with long-term outcomes not yet known.

What STABILITY did not do is compare hamstring plus LET with BTB. Neither is proven better than the other for the high-risk young athlete; both are defensible, and both beat isolated hamstring in that group.

Which graft for which patient?

Put the evidence together and the decision follows the patient, not a house preference.

Children are a separate decision. Over open physes the graft is usually soft tissue, and BTB is generally avoided while substantial growth remains because bone blocks can injure the physis. In a paediatric meta-analysis of 12 non-randomised studies (205 quadriceps, 454 hamstring), graft rupture was 3.5% with quadriceps tendon against 12.4% with hamstring (PubMed (opens in a new tab)). The confidence intervals were wide and the groups unmatched, so the paediatric ACL injury topic page treats it as promising rather than settled. Our post on physeal-safe paediatric ACL reconstruction covers the techniques.

The graft does not shorten rehabilitation. Whatever the tendon, return to pivoting sport should wait for objective criteria and, in most protocols, at least 9 months; the return to sport criteria page and our post on return-to-sport testing after ACL reconstruction give the battery.

In the exam

Examiners want a reasoned choice, not a favourite. Name the options, quote one trial-level and one registry-level finding, and match the graft to the patient in front of you. If the question opens with the examination, our post on the Lachman and pivot shift covers the grading that feeds the LET decision.