The Evaluation of Refrigerated and Frozen Osteochondral Allografts in the Knee

Between 1998 and 2002, 25 patients who were treated with a refrigerated or frozen allograft were evaluated. The mean patient age was 48 years. The mean lesion size was 4.5 cm. Validated outcome instruments [Knee Society Score, Western Ontario and McMaster University Score] were used. Clinical and radiographic evaluations were performed pre-operatively and at the most recent follow-up. Histological and electron microscopic analysis was performed on grafts prior to implantation. Clinical follow-up averaged 46 months (range 24 60 months). The Western Ontario and McMaster University Score improved from 46 + 24 to 66 + 22 (p = 0.003). The Knee Society Score improved from 104 + 43 to 132 + 42 (p = 0.01). No correlation was noted between graft type and histological or electron microscopy scoring. Post-operative mechanical alignment was not correlated with an improvement in Western Ontario and McMaster University Score (p = 0.19) or Knee Society Score (0.27). Six patients (24%), all refrigerated allografts, were failures and underwent knee arthroplasty. Seventy-six percent of implanted frozen and refrigerated osteochondral allografts are in place 4 years after surgery. Frozen allografts appear to be surviving as well as refrigerated grafts. The use of magnetic resonance imaging may enable the evaluation of graft incorporation and articular cartilage integrity.

Previously, authors have published the viability results of stored refrigerated allografts [3,29,30].However, the authors are unaware of data correlating patients' functional outcome with radiographic evaluation and histological/ electron microscopy grading of refrigerated and frozen allografts at the time of implantation.The purpose of the current study was to clinically and radiographically evaluate patients who underwent refrigerated or frozen allograft transplantation.In addition, histological/electron microscopy grading of the allograft and radiographic evaluation of the affected knee was analyzed in relationship to functional outcome at an average of 46 months follow-up.

Patient Data
The current study was approved by the Institutional Review Board at our institution prior to implementation.All study patients gave informed consent prior to being enrolled in the study.Between 1998 and2002, 26 patients underwent osteoarticular transplantation of the femur and/or patella with a refrigerated or frozen allograft from which histologic and electron microscopic data was available at the time of implantation.Inclusion criteria were as follows: 1) Tegner 3 or greater activity level; and 2) a contained articular cartilage defect amenable to a non-structural osteoarticular graft; 3) articular cartilage damage limited to 1 or 2 compartments; 4) biomechanical knee alignment that was less than 5° of varus or valgus or correctable with a distal femoral or proximal tibial osteotomy; and 5) failure of conservative measures including non-steroidal anti-inflammatory medications and physical therapy for a minimum of 3 months.Body mass index greater than 30 and chronological age were not used as exclusionary criteria from surgery.Patients whose lifestyle included less demanding activities were encouraged to undergo a unicondylar or total knee arthroplasty.
Prior to surgery, all patients underwent clinical evaluation, standing radiographic imaging of both knees, and magnetic resonance imaging of the affected knee.Specific magnetic resonance imaging sequences (fast spin echo) were performed to assess the articular cartilage of the patella, femur and tibia.All images were reviewed by the senior author and a board certified, fellowshiptrained radiologist.

Operative Technique
All allograft transplants were performed through a mini-arthrotomy unless a tibial tubercle osteotomy was performed for a patellar and/or trochlear defect.After the recipient defect was measured, the bone was reamed to a depth of 8 and 12 mm.The donor allograft was cored with a coring device (Arthrex, Inc., Naples, Florida) with a minimum diameter of 18 mm.The graft was press-fit into the recipient site with the articular surfaces congruent.All grafts were placed with less than 1 mm of donor-recipient articular surface incongruity.Early in the study, 1 graft (patella) was secured with an intra-articular screw.No subsequent grafts underwent fixation.
All tibial tubercle osteotomies were performed according to the technique described by Fulkerson and secured with two 3.5 mm or 4.5 mm screws [24] .All high tibial osteotomies were performed with a lateral closing or medial opening wedge.All distal femoral osteotomies were performed with a lateral opening wedge.For all opening wedge osteotomies, frozen allograft corticocancellous bone graft was used to fill the osteotomy defect.

Postoperative Treatment
All patients undergoing an isolated mini-arthrotomy were discharged the following day.Patients undergoing a tibial tubercle osteotomy, high tibial osteotomy or distal femoral osteotomy remained in the hospital 48 to 72 hours.For all inpatients, continuous passive motion was instituted on the first post-operative day.Continuous passive motion was continued on an outpatient basis 3 to 6 hours per day for 3 weeks post-operatively.Patients undergoing isolated allograft transplantation were kept toe-touch weight bearing for 4 weeks and then progressed to full weight bearing over 2 weeks.All patients undergoing a tibial tubercle osteotomy, high tibial osteotomy or distal femoral osteotomy were kept non-weight bearing for 6 weeks and progressed to full weight-bearing over the next 2 weeks.Heavy labor and athletic activity were delayed until 6 months after surgery.

Clinical and Radiographic Assessment
All patients underwent radiographic evaluation preoperatively and at yearly intervals.Two standardized radiographic views (standing antero-posterior and skyline) were obtained by a trained radiological technician.In addition, standing long-leg antero-posterior radiographs were obtained pre-operatively and at the most recent follow-up.A trained research assistant blinded to the patient's clinical data measured each film.Each film was assessed for the area of greatest joint space narrowing (Grade 3 = > 3mm joint space, Grade 2 = < 3mm of joint space but not bone on bone, Grade 1= bone on bone) in specific areas (Figure 1).Based upon these measurements, the patient was given a composite score ranging from 3 to 6 which was used for later analysis.Long leg weight-bearing radiographs were measured by one of the authors.The biomechanical axis was determined by the intersection of a line drawn from the center of the femoral head to the center of the tibial plateau and a line drawn from the center of the talus to the center of the tibial plateau.
All patients underwent clinical evaluation including knee range of motion.Patients completed the Knee Society Score, Western Ontario and McMaster University (WOMAC) Score preoperatively and at yearly intervals after surgery.Failure was defined as conversion of a transplanted graft to a unicondylar or total knee arthroplasty.All failures were included in the overall analysis and also analyzed separately.

Refrigerated Allografts
All grafts were aseptically processed within 72 hours of Copyright © 2011 SciRes.SS Harvested osteochondral grafts were aseptically cleansed with saline pulse lavage in order to remove blood and fat from the cancellous bone.All grafts were pre--soaked in antibiotic storage medium.Grafts were transferred to double sterile pouches containing storage media to maximize chondrocyte viability during storage (IN-CELL Corp., San Antonio, Texas).Only grafts demonstrating no growth after 7 day microbial culturing were released for distribution.Grafts were refrigerated and stored at 2° -8° Celsius with a shelf life of 6 weeks from the date of processing (Regeneration Technologies Inc., Alachua, Florida).

Frozen Allografts
All grafts were aseptically processed within 72 hours of the donor's death and procedures were performed in ac-cordance with the guidelines established by the American Association of Tissue Banks.Donors' blood was screened for malignancies, autoimmune and certain neurological disorders, and any high-risk behavior.Donor blood and tissue was screened for human immunodeficiency virus-1 and human immunodeficiency virus-2 antibodies, Hepatitis B and C antibodies, human Tlymphotropic Virus I and II, syphilis, and polymerase chain reaction testing for human immunodeficiency virus-1.All grafts were stored at -70° Fahrenheit until the day of use.

Cartilage Preparation and Histologic Scoring
Immediately upon opening the allograft in the operating room, three 5 mm plugs were sterilely harvested from a peripheral area of articular cartilage with an Osteochondral Autograft Transfer System (OATS) harvester (Arthrex, Naples, Florida).The sample site was free from any visual cartilage defects.All plugs were placed in a sterile container with a small amount of normal saline and sealed for transport to the pathology and flow cytometry laboratories for evaluation.The time from plug harvest to processing was approximately 30 minutes for all specimens.
For histology a portion of fresh cartilage was cut from the bone surface, fixed in 10% buffered formalin, processed to paraffin, sectioned at 5 um, and stained with hematoxylin and eosin.Three slides were prepared from the plugs and sent for evaluation.The slides were scored blindly and the scores averaged.The average score was used for data analysis.The following scoring system was utilized for each slide: 0: all cells appeared lethally injured; 1+: majority of cells exhibited marked to lethal injury; 2+: minority of cells exhibited marked to lethal injury; 3+: all cells appear viable.Severe pyknosis and cell lysis were judged to represent marked to lethal injury.
For electron microscopy, fresh tissue was placed in 3% glutaraldehyde (0.1 N sodium cacodylate buffer, pH 7.2), post-fixed in 1% osmium tetroxide, and embedded in epoxy resin.Sections (1 um) were stained with toluidine blue.Thin sections were stained with uranyl acetate and lead citrate and examined in a Phillips CM100 electron microscope (FEI Company, Hillsboro, Oregon).The scoring scheme above was also applied to the ultrastructural findings.Features such as severe pyknosis, obscured detail of the organelles, and rupture of the plasma membrane were considered indicators of lethal injury, and lesser degrees of cytoplasmic contraction, cytoplasmic blebs, and accumulation of myelin figures were considered indications of marked injury (Figures 2(a)-2(c)).

Statistical Analysis
Statistical analyses were performed using appropriate procedures of the JMP software system (SAS Inc., Carey, NC).Means for demographics, graft size, follow-up, shelf time (time from harvest to implantation), harvest time (time of death to time of harvest), electron microscopy score and histology score were calculated.Correlation analysis were performed to determine if a relationship existed between electron microscopy score, histology score, outcome score and any other variable.Significance was determined at the 0.05 level.

Results
Twenty-five patients (96%) with complete data were available for follow-up.The average age for the overall group was 48 years.The average follow-up for the group was 46 months.The average body mass index of the group was 32.The average graft size for the group was 4.5 cm 2 .Pre-operative knee range of motion was less in the frozen group when compared to the refrigerated group (p = 0.01) (Table 1).
Both groups improved after surgery.Post-operative WOMAC Score, Knee Society Score and knee range of motion values were 66, 132, and 116 respectively (Table 2

) (Figures 2(a)-2(c)).
Post-operative improvement in WOMAC Scores and Knee Society Scores, and knee range of motion was analyzed in relation to graft type.Greater improvement in WOMAC and Knee Society Scores was noted in the frozen group when compared to the refrigerated group (p = 0.07) (Figures 3(a)-3(b)).An improvement in knee range of motion was noted in frozen patients compared refrigerated allograft patients (p = 0.02) (Table 2).Radiographic measurements were made on standing antero-posterior (AP) and skyline views.Medial and lateral tibio-femoral joint spaces were given a score from 1 to 3, while medial and lateral patello-femoral joint spaces were given a score from 1 to 3. The baseline group AP radiographic score was 5.1 which increased at follow-up to 5.3 (p = 0.06).The baseline patella score for the group was 5.6 which increased at follow-up to 5.7 (Table 3).
Mechanical axis was measured in all patients without a tibial tubercle osteotomy.The average pre-operative and post-operative mechanical axes were 174° and 177° respectively (p = 0.08).When post-operative alignment was analyzed in relation to outcome improvement, no correlation was noted (Table 4).
Due to the small numbers, the 4 histological and electron microscopy scores were combined (scores zero and one = Grade 1/scores two and three = Grade 2).When pre-operative WOMAC and Knee Society Scores were compared, there was no difference between groups.No correlation was noted between graft type and histological or electron microscopy scoring.Therefore, no difference in chondrocyte viability was noted on histological or electron microscopy between refrigerated and frozen allografts at the time of implantation.No correlation was noted between the histological or electron microscopy grading systems and any outcome tool (Table 5).
Improvement in outcome was analyzed in relation to the number of sites undergoing grafting.No difference in outcome improvement was noted between single or multiple site grafts (Knee Society Score p = 0.7; WOMAC Score p = 0.5; range of motion p = 0.9).
No difference in post-operative scores was noted between the 2 groups (p = 0.8, p = 0.7).However, poorer   pre-operative knee range of motion, Knee Society and WOMAC scores in patients receiving a frozen allograft (p = 0.01, p = 0.1, p = 0.1) resulted in a greater net change in outcome in the frozen allograft group.Six patients (24%) who underwent a knee arthroplasty were considered failures and analyzed separately.All failures were refrigerated allografts.We evaluated histological and electron microscopy findings at the time of implantation with success or failure of the graft.No failures were noted if the histology score was Grade 2 (p = 0.02) or if the electron microscopy score was Grade 2 (p = 0.4).

Discussion
The use of fresh allografts for the treatment of full thickness articular defects is well documented [2,15,[31][32][33][34].Fresh osteochondral shell allografts (<1 cm of subchondral bone) provide the greatest likelihood of chondrocyte survivability, while reducing immunogenicity by decreasing the exposure of white cells found in cancellous bone [1].Various authors have indicated that the long-term survival of an allograft is dependent upon viable chondrocytes with a high level of donor deoxyribonucleic acid (DNA) to replenish the transplanted matrix [35].Currently, fresh allografts provide the largest population of viable donor chondrocytes for transplantation.
Although fresh osteochondral allografts provide an excellent source of live chondrocytes, their use is not without drawbacks.Foremost is the need for the graft to be as fresh as possible to maintain chondrocyte viability, yet not expose the recipient to infection at the time of implantation.Currently, all published reports of osteochondral grafts using the term "fresh" have cited a time from graft harvest to implantation of 7 days or less [4,[36][37][38][39][40][41][42][43][44].In the current study, the time from graft harvest to implantation was greater than 7 days for all non-frozen grafts.Noting this, we chose to label these grafts as "refrigerated", rather than "fresh".
Recent data have indicated good results with the use of refrigerated osteochondral allografts.Emmerson et al. reported 72% good/excellent results at 7.7 years with refrigerated allografts averaging 7.5 cm in size implanted for osteochondritis dissecans of the femoral condyle.[45].McCulloch et al. reviewed 25 consecutive patients who underwent refrigerated osteochondral allograft transplantation for defects in the femoral condyle.Overall, patients reported 84% satisfaction with their results [46].Williams et al. reported on 19 patients who were treated with "fresh stored" allografts maintained for an average of 30 days prior to implantation.Mean lesion size was 602 mm 2 .At the most recent follow-up, the mean Short Form-36 score improved to 66 (+/-24) and the mean score on the Activities of Daily Living Scale increased to 70 (+/-22) [47].
The outcome tools used in the current study were not identical to those used by other authors [46,47].However, both the WOMAC and Knee Society Scores are validated outcome instruments reported in the literature for the evaluation of patients with arthritic conditions [48][49][50].Indeed, previous studies of function in osteoarthritis have shown that WOMAC Score is more sensitive to change and has greater efficiency than other instruments used to assess osteoarthritis, including the SF-36 Health Survey [51,52].
In the current study we noted clinical improvement in patients implanted with a refrigerated or frozen allograft.Many of these patients had malalignment or patellofemoral disease requiring a concomitant osteotomy.Whe-ther the osteotomies themselves played a larger role then the transplantation in improved patient outcome is unclear.
Numerous animal studies have indicated that harvested osteochondral allografts can be safely preserved up to 28 days with overall maintenance of biomechanical properties, cell matrix collagen content, and permeability [53][54][55][56].Wayne et al. demonstrated that osteochondral shell grafts could be stored in culture medium at 4° Celsius for up to 60 days without deterioration of collagen content, proteoglycan amount, or histologic appearance [56].Pearsall et al. reported on 16 refrigerated osteochondral allografts that underwent histologic and ultrastructural examination prior to implantation.The authors found an inverse correlation between matrix staining and time to implantation (p < 0.05) and reported an average time from harvest to implantation of 30 days (range: 17 to 44 days) [57].Despite the aforementioned animal and histologic/electron microscopy data, the authors are unaware of published data correlating pre-implantation histologic and electron microscopy findings with clinical outcome.
The use of deep frozen allografts for the treatment of osteoarticular defects has also been reported, with cited failure rates as high as 25% [2,15,58].Reasons cited for failure include slow incorporation, diminished chondrocyte survival, and subsequent matrix degeneration [1,58].Pritzker and others noted that freezing cartilage kills chondrocytes [4,23,[59][60][61].Despite reports of poor chondrocyte viability at the time of frozen allograft implantation, we detected no significant difference in outcome between our refrigerated and frozen allograft patients.Moreover, when pre-operative scores were examined, the frozen allograft group had significantly worse scores at the time of implantation, indicating that these patients were actually doing better at the latest follow-up.
There are several weaknesses to the current study.The analysis included a limited number of patients.Therefore, statistical significance could not be determined for several variables.There was no follow-up magnetic resonance imaging of the study patients to assess incorporation of the implanted graft.Previous authors have hypothesized that early magnetic resonance imaging abnormalities represent non-specific post-operative change, whereas persistent abnormalities represent immune-related injury [62].Further follow-up of the current patient cohort will include magnetic resonance imaging to evaluate graft incorporation and analyze the findings in relation to clinical outcome.The histological and electron microscopy evaluation was performed by a single pathologist.However, the evaluator has over 20 years of expertise in this area and interpreted all specimens blinded.Although several of the electron microscopy and histologic scores were low, these findings may be attributable to the need to refine the technique of electron microscopy and histologic assessment in articular cartilage.Finally, although the current study assessed chondrocyte viability, it did not evaluate chondrocyte function.Various authors have reported on the mechanical properties of articular cartilage, including an assessment of cartilage stiffness [63][64][65].
In conclusion, 76% of implanted frozen and refrigerated osteochondral allografts in the current study are still in place at 4 years after surgery.At the latest follow-up, frozen allografts are surviving as well as refrigerated grafts.The reason for this finding is unclear, as the majority of these grafts were used in salvage cases in older patients.Long term follow-up is needed to assess clinical outcome, while the use of magnetic resonance imaging may be beneficial to evaluate graft incorporation and articular cartilage integrity.

Figure 1 .
Figure 1.Knee radiographic grading form.A grade from one to three is given for the medial and lateral joint spaces within the knee and patellofemoral joint.the donor's death and procedures were performed in accordance with the guidelines established by the American Association of Tissue Banks.Donors' blood was screened for malignancies, autoimmune and certain neurological disorders, and any high-risk behavior.Donor blood and tissue was screened for human immunodeficiency virus-1 and human immunodeficiency virus-2 antibodies, Hepatitis B and C antibodies, human T-lymphotropic Virus I and II, syphilis, and polymerase chain reaction testing for human immunodeficiency virus-1.Harvested osteochondral grafts were aseptically cleansed with saline pulse lavage in order to remove blood and fat from the cancellous bone.All grafts were pre--soaked in antibiotic storage medium.Grafts were transferred to double sterile pouches containing storage media to maximize chondrocyte viability during storage (IN-CELL Corp., San Antonio, Texas).Only grafts demonstrating no growth after 7 day microbial culturing were released for distribution.Grafts were refrigerated and stored at 2° -8° Celsius with a shelf life of 6 weeks from the date of processing (Regeneration Technologies Inc., Alachua, Florida).

Figure 2 .
Figure 2. (a)-(c) Pre-operative and post-operative radiographs and intraoperative radiographs and intra-operative images of an active male farmer who underwent a distal femoral opening wedge osteotomy and lateral femoral condyle refrigerated osteoarticular allograft.(a) Pre-operative standing antero-posterior radiograph.The patient's preoperative KSS score was 175.(b) Intra-operative image from the previous patient showing two refrigerated osteoarticular plugs in place in the lateral femoral condyle.(c) Post-operative standing antero-posterior radiograph of the same patient.His post-operative KSS score at the latest follow-up was 190.

Figure 3 .
Figure 3. (a), (b) Pre-operative and post-operative radiographs of a homemaker with 22 degrees of combined femoral and tibial varus.She underwent combined distal femoral closing wedge and medial tibial opening osteotomies.A frozen osteoarticular allograft was used to reconstruct the medial femoral condyle.(a).Pre-operative stan-ding antero-posterior radiograph.The patient's pre-ope-rative KSS score was 59. (b).Post-operative standing antero-posterior radiograph of the same patient.Her post-ope-rative KSS score at the latest follow-up was 185.

Table 2 . Overall pre and post-operative outcome scores and outcome improvement in refrigerated and frozen allograft patients.
*Statistically significant.