Allograft Reconstruction of the ACL-Deficient Knee Treatment & Management

A major advantage of allografts is that there are a greater variety of tissues available for reconstruction. Bone-patellar tendon-bone has been used most commonly, and although some advocate its use in primary cases, most are used in revisions. Its popularity stems from its 2 bony attachment sites, which ease fixation. Achilles tendon is also available, but it is used more commonly in posterior cruciate ligament (PCL) reconstruction due to its size, length, relative ease of insertion, and accommodation to being split into 2 bundles as part of an increasing trend for PCL reconstruction. Hamstring, tensor fascia lata, and other tissues, such as anterior and posterior tibial tendons, have also been used with varying success.^[16] Rene Verdonk of Belgium has reported good success in revisions with these tibial tendons with up to an 8-year follow-up.

Following proper thawing or rehydration and implantation, the incorporation of both autograft and allograft follows a similar sequence. The original structure acts as a scaffold for revascularization, cell repopulation, and remodeling.^[17] However, the timing of events varies, as the remodeling and maturation process is prolonged by as much as 50% for allografts. Grafts are weakest during this vascularization and maturation period. This has implications for the stresses that these tissues can withstand in the postoperative period. One study reports a higher graft failure rate with younger patients and with allograft, with a multiplicative effect when combined.^[18]

Once remodeling is complete, implanted allografts appear histologically similar to native ACL. However, this does not necessarily translate into strength or stability. Shino^[4] showed histologic maturity at 18 months, while Arnozky^[19] showed dog allograft histologically resembling normal ACLs at 1 year. Using a goat model, Drez^[20] and Jackson^{[9, 12]} independently showed similarities with native ACL at 26 weeks. Although it is now understood that the goat model is not applicable to humans regarding time of incorporation, Drez showed the maximum load-to-failure of allografts to be 43% of the native ACL, and Jackson showed this failure to be 27% of native ACL versus 62% for autografts.

Infection following any surgical procedure is certainly one of the accepted, yet feared, complications. Recently, however, significant publicity has surrounded 3 infections and subsequent deaths following orthopedic allograft transplants.^[21]

The strain level that damages grafts and the strain level necessary for graft development are not presently known. Proper graft placement certainly plays a critical role. Specifically for allograft, the hydration status or how well thawed a graft is must be considered. If the graft is not allowed to fully recover from its frozen or freeze-dried state, postoperative tensioning and strain characteristics may drastically change soon after surgery.

Long-term published clinical studies comparing allografts to autografts are few. Indelicato^[22] and Shelton^[23] all showed generally good results in comparison of the tissues. One study shows improved long-term outcomes with autograft over allograft as well as with not smoking and with normal body mass index.^[24] In another study, an overall trend of fewer patellofemoral symptoms and better range of motion with allografts was noted. Shelton described a trend of increased pivot glide with allograft, which was not statistically significant. Although happy with their allograft results, they all remained cautious with their outlook, echoing the sentiments of Beynnon that it may take years to see a pattern for overall failure for any graft type.^[25]

Beynnon theorizes that the initial and 2- to 3-year outcome studies may not accurately assess longer-term results.^[26] He showed that reestablishing AP stability is not a predictor of future graft behavior. Using strain gauges in autograft reconstructions, he showed that strain characteristics established at the time of surgery was a more powerful predictor of long-term results. Grafts that varied most from normal strain patterns in the early postoperative period showed long-term failure. This is disturbing when recent bench studies have shown that tensioning allografts in the human cadaver knee to fully achieve AP joint stability increased forces in the graft at all angles of flexion.

Authors have long proclaimed dangerous strain and shearing in terminal extension. Of particular note, the good results that Indelicato^[22] and Shelton^[23] achieved all predated the era of accelerated rehabilitation protocols popularized by Shelbourne.^[27] In fact, the allograft protocols included limited arcs and crutch weightbearing for up to 12 weeks.

With all of this in mind and knowing that allografts take longer to remodel and mature, the following question remains: Should there be concern with allografts in general and specifically in relation to recent trends in accelerated rehabilitation? Although Shelbourne has not suggested this, should his autograft axiom be applied? It allows activity based on the status of rehabilitation and not on graft biology. Alternatively, should these patients be restricted as is commonly done in grafts without bone plugs due to fixation concern? This question is especially important with the potential earlier aggressive rehabilitation and return to activity that allografts allow due to the decreased morbidity compared with autografts.^{[28, 29, 30]}

The risks of disease transmission would seem to have become infinitely small, but, as evidenced by fatal infections noted already, this risk has not been reduced to zero. It is imperative that the surgeon constantly monitors the source of his or her grafts and has a very specific protocol of response in the face of an adverse surgical outcome when infection is a possible diagnosis.^[31]

With a supply of safe graft materials, other than a national graft shortage or insurers or the hospital denying coverage for the additional costs, strength and long-term results become the main concern.

The information above indicates the need to protect these grafts from aggressive early rehabilitation. Protection may include limited weightbearing and stresses placed across the joint. However, no data are available to support this protocol, and prospective comparative studies are needed. For primary cases, weighing the risk of outright allograft failure due to tissue weakness against the morbidities of autograft harvest still leaves the surgeon with a difficult decision. No clear answer exists.

Far from ideal, allografts offer a material off the shelf with a relatively good record. Although prospective long-term results are unknown, many patients have done well clinically with this procedure as a primary reconstruction. However, with improved soft-tissue fixation, tripled semitendinosus without gracilis and Quad tendon grafts are becoming more appealing, as they offer strong autograft materials without the problems associated with patella tendon. For revisions and situations in which no autograft material is available, it offers hope where none might otherwise exist.