Tendon Graft: Sources, Surgery, and the Healing Process

A tendon graft replaces a damaged or torn structure with a strip of tendon harvested from elsewhere in your own body or obtained from a tissue donor. The procedure is most commonly associated with anterior cruciate ligament (ACL) reconstruction in the knee, though tendon grafts also appear in shoulder, hand, ankle, and elbow surgeries. Which tendon gets used, how surgeons anchor it, and what your body does with the transplanted tissue over months and years all affect the outcome. The biology behind graft healing turns out to be surprisingly dynamic, and the choices made before surgery can matter as much as what happens during rehab.

Autograft Versus Allograft

The first major decision is whether the graft comes from you or from a deceased donor. An autograft is tissue taken from your own body during the same operation. An allograft is donor tissue processed and stored in a tissue bank. Each has trade-offs, and the debate has shifted over the past two decades as better data has accumulated.

The case for autografts rests largely on lower failure rates. A large cohort study found that the odds of tearing an allograft ACL reconstruction were roughly four times higher than the odds of tearing an autograft, after adjusting for age.¹ A more recent study with nearly six years of follow-up reported graft failure in about 9% of autograft patients compared with roughly 21% of allograft patients, with five-year graft survival of about 91% for autografts and 77% for allografts.² That same study found that about 82% of autograft patients returned to their preinjury sport, compared with roughly 61% of allograft patients.

The picture is slightly more nuanced once you look at how allografts are processed. A systematic review focusing specifically on nonirradiated allografts found that while failure rates still ranged higher (up to about 27% in some studies, versus under 10% for autografts), functional scores between the two groups were not significantly different.³ Two of the included studies pointed to younger, more active patients as the ones most at risk for allograft failure. This is why many surgeons reserve allografts for older, less active patients or for revision surgeries where donor-site tissue has already been used.

When it comes to revision ACL surgery, autograft again outperforms allograft. A systematic review of revision cases found that about two-thirds of autograft patients returned to sport, compared with under half of allograft patients.⁴

Choosing an Autograft Tendon

If you and your surgeon go with an autograft, three main tendon options exist: the patellar tendon (a strip from below the kneecap, often taken with small bone plugs at each end), the hamstring tendons (typically the semitendinosus and sometimes the gracilis, from the back of the thigh), and the quadriceps tendon (from the front of the thigh above the kneecap). Each has a different risk profile at the donor site, meaning the spot where the tendon was removed.

A network meta-analysis of randomized controlled trials found that hamstring grafts had about 69% lower odds of donor-site problems compared with patellar tendon grafts, and quadriceps tendon grafts had about 88% lower odds.⁵ A separate meta-analysis confirmed that the quadriceps tendon produced significantly less donor-site morbidity than both the patellar tendon and the hamstring, without any meaningful differences in graft failure rates or knee stability between the three options.⁶

On a practical level, patellar tendon grafts tend to cause more kneeling pain, which can matter if your job or sport involves kneeling. Hamstring grafts are associated with numbness on the inside of the shin and some degree of muscle wasting, both confirmed in a recent study that compared hamstring and quadriceps donor sites.⁷ The quadriceps tendon is increasingly popular because it offers a thick graft with less residual pain and numbness, though it is still the newest of the three and its long-term track record is shorter.

What Happens to the Donor Site After Harvest

One of the most common patient questions is whether the harvested tendon grows back. For hamstring tendons, the answer is often yes, at least partially. A systematic review found that the semitendinosus tendon regenerated in about 79% of cases, and the gracilis in about 72%, when assessed more than a year after harvest.⁸ A separate study confirmed a 72% regeneration rate for the semitendinosus and found that the regrown tendons closely resembled normal anatomy under the microscope, though they often reattached lower on the leg than the original insertion point.⁹ The new tendon is not identical to the original, and functional deficits in deep knee flexion strength can persist, but the regeneration is a meaningful consolation for patients worried about permanently losing a tendon.

How the Graft Is Fixed Inside the Bone

Once the surgeon has prepared the graft, it must be threaded through tunnels drilled in the shinbone and thighbone and then locked in place. Two broad categories of fixation exist. Interference screws are placed directly alongside the graft inside the bone tunnel, pressing the tissue against the tunnel wall. Suspensory devices sit outside the tunnel, looping over a button or post on the bone surface and holding the graft in tension like a suspension bridge.

Biomechanically, the two approaches are surprisingly similar. A large-animal study comparing suspensory and interference screw fixation found no significant difference in mechanical properties between the two.¹⁰ Lab testing on biceps tendon fixation similarly found comparable failure loads and stiffness between suspensory buttons and screws, though the suspensory device allowed more cyclic displacement before failure.¹¹ In practice, many surgeons use a combination: a suspensory device on the femoral side and an interference screw on the tibial side.

Tunnel placement matters at least as much as the fixation method. When the bone tunnels are positioned within the original ligament’s footprint (an “anatomical” reconstruction), the risk of the graft rubbing against the edge of the bony notch is actually manageable, provided tunnel placement is precise.¹² Cadaver studies have also shown that removing too much bone from the back of the notch can increase forces on the graft during bending, suggesting that surgeons should be conservative with how much bone they reshape.¹³

The Graft’s Biomechanical Behavior in Surgery

Between harvest and fixation, the graft undergoes a process called preconditioning: it is cycled under tension to remove slack and settle the tissue. This step matters more than patients realize, because tendons behave differently depending on how they are loaded in the first few minutes. One study found that higher preconditioning tension produced grafts that maintained more residual tension and stiffness hours later.¹⁴

Patellar tendon and hamstring tendon grafts also behave differently on the bench. While stiffness ends up similar between the two, patellar tendon grafts retain considerably more tension over time. In one comparison, patellar tendon grafts held about 59 newtons at the 15-minute mark while hamstring grafts held only about 48, and by four hours the patellar grafts still held roughly 50 newtons versus about 24 for hamstrings.¹⁵ Graft cross-sectional area also plays a role: modeling suggests that initial stress after fixation ranges from about 4 to 12 megapascals depending on the graft size, and all grafts lose roughly 13 to 17% of that initial stress through relaxation.¹⁶ These numbers help explain why surgeons pay such close attention to graft diameter and initial tensioning: the graft will lose some of whatever tension it is given.

How the Graft Heals and Remodels

A transplanted tendon does not simply sit in place as an inert cable. Your body tears it down and rebuilds it over a timeline that stretches well beyond a year. This process, often called ligamentization, unfolds in overlapping phases.

In the first few weeks, the graft’s core is largely avascular and begins to die off. Meanwhile, cells at the surface start proliferating. A study of human patellar tendon grafts found proliferating cells at the graft surface within two weeks, but no new blood vessels yet at that point. A signaling protein called vascular endothelial growth factor (VEGF) peaked around two to three weeks, and new blood vessels then grew progressively from weeks three through eight.¹⁷ Animal studies confirm that the revascularization process invades from the periphery inward, gradually replacing the dead central tissue. By about 12 weeks, the entire graft diameter showed increased blood vessel density, and by a year, VEGF signaling had largely quieted down.¹⁸

Biochemically, the collagen crosslink profile of the graft gradually shifts to resemble a native ligament rather than the tendon it started as. A study of human biopsies found that hamstring tendon grafts changed their crosslink chemistry substantially by four to six months and were biochemically close to a native ACL by about a year after surgery.¹⁹ MRI mapping of graft composition suggests that substantial internal changes occur in the first six months, followed by a relatively stable period, and then continued maturation from one to two years.²⁰ This is the biological reason behind the long rehab timelines: even when a patient feels strong at six months, the graft’s internal architecture is still maturing.

Where the graft meets bone, a different challenge arises. The native ACL attaches through a specialized transitional zone of fibrocartilage, and recreating that interface is one of the hardest parts of healing. Animal research has shown that combining certain growth factors with platelet-rich plasma can promote the formation of a cartilage-like transition zone at the tendon-bone junction, with visible fibrocartilage maturing progressively over 12 weeks.²¹ This remains an active area of research rather than standard clinical practice.

Rehabilitation and Early Motion

Controlled movement after surgery helps the healing graft become stronger. This principle has been demonstrated repeatedly in animal models. One study on Achilles tendon repair found that early mobilization restored the tendon’s functional properties more rapidly than keeping it completely immobilized.²² Another showed that controlled passive motion produced tendons with significantly greater rupture load, stress capacity, and energy absorption compared with immobilized tendons, without slowing the rate of healing.²³

In modern ACL rehab, this translates to a careful escalation: gentle range-of-motion exercises start within days, weight-bearing progresses over the first few weeks, and strengthening builds gradually. The exact timeline varies by surgeon and graft type, but the underlying principle is the same. Mechanical loading stimulates the cells repopulating the graft to lay down collagen in a more organized pattern. Too little load, and the graft becomes disorganized and weak. Too much too early, and you risk stretching or tearing the still-immature tissue.

Tunnel Widening and Other Complications

One complication that can develop silently after surgery is bone tunnel widening. The drilled tunnels that house the graft can enlarge over time, driven by a combination of biological and mechanical factors. Biological causes include inflammatory molecules in joint fluid seeping into the gap between graft and bone. Mechanical causes include repetitive motion of the graft within the tunnel, sometimes described as a “windshield-wiper effect.”²⁴

A CT-based study found that the tibial tunnel increased in diameter by about 11% over 10 months (from roughly 9 mm to 10 mm), while the femoral tunnel expanded about 3%.²⁵ Reassuringly, that same study found no significant connection between the amount of tunnel widening and clinical outcomes or knee stability. Still, significant widening can become a problem if a revision surgery is ever needed, because the surgeon may have to fill or bypass the enlarged tunnels before placing a new graft.

How Allografts Are Sterilized and Processed

If a donor tendon is used, it must be sterilized without destroying its mechanical properties. A systematic review evaluated the available sterilization methods and recommended freezing combined with gamma irradiation or electron-beam treatment in the range of about 15 to 29 kilograys as the best approach, balancing full sterilization with preserved or even improved mechanical properties.²⁶ Other methods, including ethylene oxide and supercritical carbon dioxide, were found to weaken the tissue and are not recommended.

A direct comparison of gamma irradiation, electron-beam sterilization, and chemical sterilization found that all three left the tendons biomechanically similar to untreated controls, though high-dose irradiation caused more collagen damage at the molecular level.²⁷ A separate lab study confirmed that electron-beam sterilized tendons were mechanically equivalent to unsterilized controls across multiple measures including maximum load, stiffness, and elongation.²⁸ The clinical implication of all this: how an allograft is sterilized can matter as much as which tendon it came from.

What If the Graft Gets Dropped on the Floor

It sounds like a nightmare scenario, but a graft accidentally touching a nonsterile surface during surgery does happen. The question is whether a contaminated graft can be salvaged. A study that deliberately dropped autograft tendons on operating room floors found contamination in 40% of the untreated samples. Soaking in chlorhexidine or in a bacitracin solution brought contamination rates down to 8% and 4%, respectively.²⁹ A broader review of contamination protocols found that chlorhexidine and a polymyxin B-bacitracin solution with mechanical agitation achieved 100% sterility in experimental conditions, while povidone-iodine (the classic brown antiseptic) was far less effective at only about 48%.³⁰ Knowing these protocols exist is reassuring, but the practical takeaway is that not all decontamination agents are equal. Chlorhexidine-based soaks clearly outperform iodine-based ones.

Tendon Grafts in Children

Reconstructing a torn ACL in a child whose bones are still growing introduces a unique concern: the drill tunnels could damage the growth plates and cause a limb-length discrepancy or angular deformity. Physeal-sparing techniques address this by routing the tunnels entirely through the epiphysis (the end portion of the bone beyond the growth plate) rather than crossing it. A long-term study of 57 children who had this technique, with an average age of about 12, found no clinical or radiological evidence of growth disturbance after an average of about 20 centimeters of subsequent growth. Functional outcomes were excellent, with 95% of patients rated good or very good.³¹ Quadriceps tendon autograft is also emerging as an option for pediatric cases using all-epiphyseal tunnel placement.³²

Biologic Augmentation and Tissue Engineering

Researchers are actively looking for ways to speed up graft healing or improve the tendon-bone junction. One approach combines stem cells with platelet-rich plasma applied directly to the graft. An in vivo study found that grafts treated with adipose-derived stem cells on a platelet-rich plasma scaffold showed significantly greater mechanical strength at two weeks compared with untreated grafts, and that only grafts additionally coated with an extracellular matrix hydrogel maintained that strength advantage through eight weeks.³³ A separate animal study found that combining stem cells with platelet-rich plasma significantly increased tendon breaking strength by two months post-surgery.³⁴

Looking further ahead, fully synthetic or animal-derived scaffold grafts could one day eliminate the need for donor tissue altogether. Lab-engineered scaffolds that mimic tendon’s hierarchical fiber structure have shown promising tissue ingrowth in small animal models.³⁵ Decellularized animal tendons are another avenue. Researchers have processed pig tendons to remove the sugar molecules that trigger human immune rejection and partially cross-linked the tissue to control the pace at which the body breaks it down and replaces it with native cells.³⁶ When tested in sheep, a decellularized pig tendon graft integrated into the surrounding bone with formation of specialized anchor fibers, and by 26 weeks its biomechanical properties were comparable to a same-species graft.³⁷ None of these approaches are ready for routine clinical use, but they represent the direction the field is heading: off-the-shelf graft materials that avoid both donor-site pain and the limitations of cadaver tissue.