VOL 86-B No 7 SEPTEMBER 2004 991

Upper limb

Osteochondral autologous transplantation for the treatment of full-thickness articular cartilage defects of the shoulder

M Scheibel C Bartl P Magosch S Lichtenberg P Habermeyer

From ATOS-Clinic Heidelberg Heidelberg Germany

M Scheibel MD Resident C Bartl MD Resident P Magosch MD Resident S Lichtenberg MD Chief of the Department P Habermeyer MD Chief of the DepartmentDepartment of Shoulder and Elbow Surgery ATOS-Clinic Heidelberg Bismarckplatz 9-15 69115 Heidelberg Germany

Correspondence should be sent to Dr M Scheibel

copy2004 British EditorialSociety of Bone andJoint Surgerydoi1013020301-620X86B714941 $200

J Bone Joint Surg [Br] 200486-B991-7Received 9 October 2003 Accepted after revision 9 December 2003

We performed eight osteochondral autologous transplantations from the knee joint to the

shoulder All patients (six men two women mean age 431 years) were documented

prospectively In each patient the stage of the osteochondral lesion was Outerbridge grade

IV with a mean size of the affected area of 150 mm2 All patients were assessed by using the

Constant score for the shoulder and the Lysholm score for the knee Standard radiographs

magnetic resonance imaging and second-look arthroscopy were used to assess the

presence of glenohumeral osteoarthritis and the integrity of the grafts After a mean of 326

months (8 to 47) the mean Constant score increased significantly Magnetic resonance

imaging revealed good osseointegration of the osteochondral plugs and congruent

articular cartilage at the transplantation site in all but one patient Second-look arthroscopy

performed in two cases revealed a macroscopically good integration of the autograft with

an intact articular surface

Osteochondral autologous transplantation in the shoulder appears to offer good clinical

results for treating full-thickness osteochondral lesions of the glenohumeral joint However

our study suggests that the development of osteoarthritis and the progression of pre-

existing osteoarthritic changes cannot be altered by this technique

The spontaneous capacity for repair of dam-aged articular cartilage is limited1-4 Arthro-scopic lavage shaving or debridement provideonly temporary relief of symptoms and do notstimulate the regeneration of the chondrallesion5-7 In addition a recently publisheddouble-blinded randomised placebo-control-led trial has shown that arthroscopic lavageand debridement is no better than a placeboprocedure8 Bone marrow stimulation tech-niques such as subchondral drilling microfrac-ture and abrasion arthroplasty create afibrocartilage or hyaline-like cartilage coverthat provides less mechanical stability againsttear and shear forces9-11

Transplantation of osteochondral autograftshas become a popular and widely acceptedtechnique for treating circumscribed osteo-chondral lesions Most clinical studies haveconcentrated upon the treatment of focal carti-lage defects of the knee and ankle joints12-20

Focal osteochondral lesions of the shoulder areless common than those of the lower extremi-ties However they are often symptomatic andif left untreated fail to heal and may progressto glenohumeral osteoarthritis After theencouraging results of osteochondral autolo-gous transplantation in the knee and ankle

joints we have started to use the osteochondralautologous transfer system (OATS) on patientswith full-thickness osteochondral lesions of theshoulder Our aim was to evaluate the clinicaland radiological results of the procedure aswell as its risks

Patients and Methods

In a retrospective study we assessed the clinicaland radiological results of eight osteochondralautologous transplantations from the knee tothe shoulder The diagnosis was established byhistory clinical examination standard radio-graphs magnetic resonance imaging and diag-nostic arthroscopy

Inclusion criteria for this study were patientswith circumscribed Outerbridge grade IV21

osteochondral lesions of the humeral head orthe glenoid an affected area greater than 100mm2 and no or mild osteoarthritic changes(le stage I according to Samilson and Prieto22)Patients with superficial chondral lesions (Out-erbridge grades I to III)21 lesions greater than250 mm2 and advanced radiological signs ofglenohumeral osteoarthritis (ge stage II accord-ing to Samilson and Prieto22) were excludedAll patients were informed about the unknownnatural course of the osteochondral lesion the

992 M SCHEIBEL C BARTL P MAGOSCH S LICHTENBERG P HABERMEYER

THE JOURNAL OF BONE AND JOINT SURGERY

documented clinical and radiological results of the osteo-chondral transplantation procedure in the knee and anklejoints as well as the surgical technique and alternative ther-apeutic options

Pre- and post-operatively the clinical evaluation includedthe Constant score for the shoulder and the Lysholm scorefor the knee joint2324 Patients were reviewed at six weekssix months and one year post-operatively and at the mostrecent follow-up The final assessment was performed by anindependent examiner and not by the surgeon Standardradiographs (true anteroposterior true anteroposterior inexternal and internal rotation and axillary views) were usedto assess the presence of glenohumeral osteoarthritis Signsof glenohumeral osteoarthritis were graded according tothe classification of Samilson and Prieto22 MRI was per-formed post-operatively in every patient in order to assessthe integrity of the osteochondral grafts As osteochondral

autologous transplantation has not been fully investigatedin the shoulder a second-look arthroscopy six months post-operatively was recommended to all patients but per-formed in only two The remaining patients refused furthersurgical intervention

Statistical analysis between variables pre- and post-oper-atively was performed with the Wilcoxon signed-rank testfor non-parametric data The level of significance was set atp lt 005 The statistical software used was StatView (Aba-cus Concepts Inc Berkeley California)Operative technique All procedures were performed in thebeach chair position under interscalene block and generalanaesthesia A diagnostic glenohumeral arthroscopy wasperformed using a standard posterior portal According tothe International Cartilage Repair Society the defect wasclassified on the basis of its location size thickness and thecondition of the opposing articular surface In each patient

Table I Characteristics of the patients defects and procedures

Patient GenderAge (yrs) Aetiology Location

Size of defect (mm2)

Number of cylinders used

Additional procedures performed

1 M 57 Traumatic anterior instability

Central (h) 120 1 Labral augmenta-tion capsular shift

2 M 38 Traumatic anterior instability

Posteromedial (h) 150 3 None

3 M 56 Post-traumatic Central (h) 105 2 None4 M 53 Hyperlaxity Posterocentral (h) 105 2 Labral augmenta-

tion capsular shift5 F 43 Traumatic anterior

instabilityAnterocentral (g) 128 1 Labral augmenta-

tion capsular shift6 M 44 Hyperlaxity Central (h) 144 1 None7 M 23 Traumatic posterior

subluxationAnterocentral (h) 250 3 None

8 F 31 Traumatic anterior instability

Posteromedial (h) 200 2 Labral augmenta-tion capsular shift

h humeral g glenoid

Fig 1a Fig 1b

Intra-operative findings A grade IV osteochondral lesion of the humeral head a) before and b) after autologous transplantation with two osteochondralplugs

OSTEOCHONDRAL AUTOLOGOUS TRANSPLANTATION 993

VOL 86-B No 7 SEPTEMBER 2004

the stage of the osteochondral lesion was Outerbridgegrade IV with a mean size of 150 mm2 (105 to 250 mm2) Inseven patients the defect was located on the humeral sideand in one patient it was on the glenoid side In threepatients the defect was central in position in three it wasposteromedial and in one it was posterocentral on thehumeral head The one glenoid defect was positionedanteromedially (Table I)

The transplantation procedure was performed with anosteochondral autograft transfer system (Arthrex NaplesFlorida) The shoulder joint was exposed through a stand-ard deltopectoral approach and the cephalic vein was pre-served and retracted laterally The subscapularis musclewas detached from the lesser tuberosity approximately 05cm from its insertion and stay sutures were attached forsubsequent refixation and the muscle was then retractedmedially If capsular instability was present the subscapula-ris was stripped from the anterior capsule in order to per-form a capsular shift at the end of the procedure Thehumeral head was next exposed and the defect wasinspected (Fig 1a) The size of the lesion was again meas-ured using a range of appropriate colour-coded sizers Arecipient socket was then created in order to provide anadequate press-fit graft fixation The appropriately sizedtubular harvester was introduced into the joint and placedover the affected area taking great care to ensure that theharvester was perpendicular to the articular surface Thedecision to transplant single or multiple osteochondralgrafts was based upon the size and location of the defectEither one (three patients) two (three patients) or three(two patients) osteochondral plugs were used in order to fillthe defects (Table I) Next the osteochondral graft wastaken The knee joint was exposed through a lateral mini-arthrotomy The chosen donor site was in an area along theouter edge of the lateral femoral condyle immediatelyabove the sulcus terminalis This is a low weight-bearingarea and offers a convex articular surface similar to that ofthe central humeral head25 The tube harvester was insertedinto the joint and placed perpendicular to the selected har-vest site It was then driven into the bone to the same depthas the osteochondral defect and the graft was harvestedThe donor cavities were not filled The harvested grafts

were then transferred into the prepared recipient site usinga press-fit technique (Fig 1b)

In order to address any underlying instability or capsularredundancy an additional labral augmentation (Harrymanet al26) and a lateral capsular shift (Matsen et al27) was per-formed in four patients Finally the subscapularis tendonwas reattached anatomically with the arm in 30˚ of abduc-tion and 20˚ of external rotationPost-operative management Post-operatively the patientsunderwent a standard rehabilitation programme Theaffected arm was placed in an abduction pillow for threeweeks Passive range of movement was initiated on thethird post-operative day and was restricted to 60˚ of flex-ion abduction and internal rotation for the first two weeksand then increased to 90˚ up to the five-week point Exter-nal rotation was prohibited up to the six-week point inorder to protect the reconstructed anterior capsule and sub-scapularis tendon Four to six weeks post-operatively thepatient started active movements and a muscle strengthen-ing programme

Results

The mean age of the patients was 431 years (23 to 57) Thedominant shoulder was affected in six patients Full detailsare summarised in Table IClinical Seven patients were available for medium termfollow-up and one for short-term follow-up (Table II)After a mean follow-up of 326 months (8 to 47) the overallConstant score improved significantly from 739 points (57to 896) to 887 points (824 to 954) (p lt 005) All patientshad less pain when compared with their original pre-oper-ative rating There was a significant (p lt 005) increase inthe overall pain rating scale from the initial mean score of87 (5 to 13) to the latest follow-up score of 145 (13 to 15)Six patients were completely free of pain The mean activi-ties of daily living (ADL) rate increased significantly(p lt 005) from 129 (7 to 18) to 191 (18 to 20) All buttwo patients achieved their full work and sporting activitylevels The results for range of movement and strength alsoincreased although these were not statistically significant(p gt 005 for both) and probably due to pain relief Thefunctional outcome of the knee was graded with the

Table II Clinical results

Constant score (pre-operatively)

Follow-up (mths)

Constant score (post-operatively) Lysholm score

Patient Pain ADL ROMdaggerStrength (kg) Total Pain ADL ROM

Strength (kg) Total Total

1 10 9 38 97 772 36 15 20 36 106 931 1002 10 12 40 99 826 41 13 18 40 111 941 1003 5 10 38 84 705 47 13 19 40 69 863 914 13 16 30 50 694 44 15 18 36 69 834 1005 12 17 40 45 784 30 15 20 40 60 875 1006 5 14 40 33 659 24 15 20 40 61 877 1007 10 18 40 104 896 31 15 20 40 99 954 878 5 7 40 27 576 8 15 18 40 47 824 64

ADL activities of daily livingdagger ROM range of movement

994 M SCHEIBEL C BARTL P MAGOSCH S LICHTENBERG P HABERMEYER

THE JOURNAL OF BONE AND JOINT SURGERY

Lysholm score24 Six patients had an excellent result (91 to100 points) one patient had a satisfactory result (77 to 90points) and one patient had a poor result (0 to 76 points)Full clinical results are shown in Table IIRadiographic Pre-operatively four patients had stage Iglenohumeral osteoarthritic changes22 In all but one theosteoarthritis had deteriorated between the pre- and post-operative assessments The patient who did not progresswas only available for short-term follow-up Four patientsdid not show any signs of glenohumeral osteoarthritis pre-operatively (Fig 2a) However by their latest follow-up allfour had developed new inferior humeral osteophyte for-mation (Fig 2b) although this did not influence the finalclinical result Despite an increase in the size of humeralosteophytes the mean glenohumeral distance did not signif-icantly reduce between the pre-operative (44 mm) andpost-operative (42 mm) examinations (p gt 005)

MRI which was performed in the paracoronal trans-axial and parasagittal planes revealed excellent graft via-bility and congruence of the chondral surfaces (Fig 3) in allbut one patient In one patient signs of transplant insuffi-ciency with evidence of avascular necrosis could be seenHowever the patient was clinically asymptomatic and satis-fied with the post-operative result Full radiographic resultsare shown in Table IIISecond-look arthroscopy In two patients a second-lookarthroscopy (Table III) was performed six months post-operatively in order to assess the chondral surfaces In boththere was good macroscopic integration of the grafts with

the original osteochondral defect being completely coveredby chondral tissue (Fig 4) The surrounding chondral sur-face showed some superficial fissuring representing a gradeI chondral lesion according to Outerbridge21

Complications No complications which could be directlyrelated to the surgical procedure in the shoulder were seen

Fig 2a Fig 2b

Anteroposterior radiograph of a shoulder a) before and b) 24 months after autologous osteochondral transplanta-tion Note the increase in inferior osteophyte formation between the pre- and post-operative examinations

Fig 3

Post-operative paracoronal MRI 24 months after auto-logous osteochondral transplantation showing anintact osteochondral plug and a congruent articularsurface

OSTEOCHONDRAL AUTOLOGOUS TRANSPLANTATION 995

VOL 86-B No 7 SEPTEMBER 2004

However one patient reported persistent pain and recur-rent effusions of the donor knee This patient achieved apoor Lysholm score result (lt 75 points) by their latestfollow-up Because of donor site morbidity the patientunderwent two revision arthroscopic surgical procedureswith debridement of the knee joint Although the shoulderwas completely painless eight months post-operatively thepatient was dissatisfied with the procedure and wasunavailable for medium-term follow-up

Discussion

Osteochondral lesions of the shoulder are less commonthan those of the lower extremity but can cause considera-ble symptoms These might include joint pain effusion andmechanical dysfunction Although the natural history ofisolated osteochondral defects has not been well definedclinical experience has shown that these lesions mayprogress to symptomatic degeneration of the joint Conse-quently the treatment of selected isolated articular carti-lage lesions may delay or prevent the development ofosteoarthritis

The aim of all methods of articular cartilage restorationis to reproduce the mechanical structural and biochemicalproperties of the original hyaline articular surfaceAlthough different studies have reported good and excel-lent clinical results from autologous chondrocyte implanta-tion (ACI) osteochondral autograft transplantation iscurrently the only technique that appears to maintain thecharacteristics of hyaline cartilage151628-30 Bobic13 statedthat the main reason for the long-term survival of trans-planted hyaline cartilage was the preservation of an intacttidemark and cancellous bone barrier Transplantation ofarticular cartilage as a part of an osteochondral graft hasbeen shown to be an effective method of replacing focalareas of damaged articular cartilage and reducing pain inboth the knee and ankle joints12-20

Our study shows that osteochondral transplantationfrom the knee to the shoulder results in a good clinical out-come in terms of pain relief and functional recovery Weobserved a significant increase in the overall Constant scorebetween the pre- and post-operative assessments At theirlatest follow-up all but two of our patients were com-

Table III Staging radiographic MRI and arthroscopic results

Patient

Stage of osteoarthritis18 Glenohumeral distance (mm)Graft integrity on MRI

Graft integrity at second-look arthroscopyPre-operative Post-operative Pre-operative Post-operative

1 I II 5 5 Intact Congruent surface2 I II 4 4 Intact Not performed3 None III 4 4 Intact Not performed4 None II 4 3 Avascular necrosis Not performed5 None I 5 5 Intact Not performed6 I II 3 3 Intact Not performed7 None I 5 5 Intact Congruent surface8 I I 5 5 Intact Not performed

Fig 4a Fig 4b

Arthroscopic view of a grade IV osteochondral lesion a) before and b) six months after autologous transplantation

996 M SCHEIBEL C BARTL P MAGOSCH S LICHTENBERG P HABERMEYER

THE JOURNAL OF BONE AND JOINT SURGERY

pletely pain-free and able to perform all activities of dailyliving Two patients had even returned to overhead sportsactivities It is important to understand that the functionalresults which our patients obtained are not only due to theosteochondral transplantation procedure but also to thecorrection of the underlying and additional glenohumeralpathology In four patients with recurrent instability of theshoulder a capsular shift with labral augmentation was per-formed

It is significant that in all of our patients glenohumeralosteoarthritic changes were present at their latest follow-up Although the glenohumeral distance on the antero-posterior radiograph did not change significantly inferiorosteophyte formation was seen in all patients For thosewho had stage I osteoarthritis pre-operatively all but oneshowed a significant deterioration between the pre- andpost-operative assessments although this was never bymore than one stage The one patient who did not progresswas only available for short-term follow-up Four patientshad no signs of glenohumeral osteoarthritis pre-operativelyHowever by their latest follow-up all had developed newinferior humeral osteophyte formation Despite this theseosteoarthritic changes did not correlate with either post-operative pain or the final post-operative result Thisprogression in glenohumeral osteoarthritis still poses thequestion as to whether osteochondral transplantation mayslow down the development of symptomatic glenohumeralosteoarthritis at all In all but one patient MRI follow-uprevealed excellent graft viability and congruence of thechondral surfaces In one patient there were signs of graftinsufficiency probably due to avascular necrosis Howeverthis patient was pain-free and very satisfied with the result

The potential disadvantages of this technique includedonor site morbidity a limited supply of grafts dead spacebetween circular grafts graft integration and the differentmechanical properties and geometry between donor andrecipient hyaline cartilage2-4 Donor site morbidity is per-haps the most important risk factor when performing anosteochondral transplantation from the knee to the shoul-der Studies have shown that donor sites normally refill withcancellous bone and fibrocartilage and do not cause signif-icant problems131931 However persistent pain and recur-rent effusions of the donor knee were seen in one of ourpatients Although the shoulder was pain-free the patientrequired two revision arthroscopic procedures Donor sitemorbidity can clearly be a serious problem and must betaken into consideration when counselling a patient foran osteochondral transplantation from the knee to theshoulder

The ideal osteochondral defect for an osteochondralautologous transplantation in the shoulder is relativelysmall perhaps 10 to 20 mm in diameter13 In our studygroup the mean size of the affected area was 150 mm2 (105to 250) Large osteochondral defects are not suitable forthis technique as osteochondral grafts are limited and it isalso technically difficult to reconstruct a large subchondral

defect A large number of grafts may also lead to instabilityin the transplantation area Other problems with thistechnique are differences in the thickness biomechanicalcomposition and mechanical properties of the articularcartilage of the knee joint when compared with theshoulder

Although we experienced no difficulties in performingthis procedure the final clinical results may be verytechnique-dependent There are many peri-operative pit-falls that need to be considered In particular an appropri-ate length of graft is essential to a successful outcome If toolong or too short an osteochondral cylinder can lead toincongruity of the articular surface Also if the graft is notinserted in an orthograde fashion an adequate press-fitcannot be achieved which may lead to loosening or failedintegration

Our study has certain limitations It includes only a smallnumber of patients so that statistical analysis is restrictedDespite this to our knowledge it is the first study to docu-ment the medium-term clinical and radiographic results forpatients with full-thickness cartilage lesions of the shoulderwho have been treated by osteochondral autologous trans-plantation It must also be remembered that not all of thepatients had the same aetiology for their osteochondraldefect Finally in three patients the associated underlyingpathology was also corrected which must be taken intoaccount when interpreting the clinical results

In summary osteochondral autologous transplantationin the shoulder appears to offer good clinical results for thetreatment of osteochondral lesions of the glenohumeraljoint However the results of our study suggest that thedevelopment of osteoarthritis and the progression of pre-existing osteoarthritic changes cannot be altered by thetechnique

No benefits in any form have been received or will be received from a commer-cial party related directly to the subject of this article

References1 Buckwalter JA Mankin HJ Articular cartilage Part I tissue design and chondro-

cyte-matrix interactions J Bone Joint Surg [Am] 199779-A600-11

2 Buckwalter JA Mankin HJ Articular cartilage II degeneration and osteoarthro-sis repair regeneration and transplantation J Bone Joint Surg [Am] 199779-A612-32

3 Mandelbaum BR Browne JE Fu F et al Articular cartilage lesions of the kneeAm J Sports Med 199826853-61

4 Newman AP Articular cartilage repair Am J Sports Med 199826309-24

5 Baumgaertner MR Cannon WD Jr Vittori JM Schmidt ES Maurer EC Arthro-scopic debridement of the arthritic knee Clin Orthop 1990253197-202

6 Jackson RW Marans HJ Silver RS Arthroscopic treatment of degenerativearthritis of the knee J Bone Joint Surg [Am] 198870-A332

7 Weinstein DM Bucchieris JS Pollock RG Flatow EL Bigliani LU Arthro-scopic debridement of the shoulder for osteoarthritis Arthroscopy 200016471-6

8 Moseley JB OrsquoMalley K Petersen NJ et al A controlled trial of arthroscopic sur-gery for osteoarthritis of the knee N Engl J Med 200234781-8

9 Childers JC Jr Ellwood SC Partial chondrectomy and subchondral bone drilling forchondramalacia Clin Orthop 1979144114-20

10 Insall J The Pridie debridement operation for osteoarthritis of the knee Clin Orthop197410161-7

11 Johnson LL Arthroscopic abrasion arthroplasty historical and pathological perspec-tive present status Arthroscopy 1986254-69

992 M SCHEIBEL C BARTL P MAGOSCH S LICHTENBERG P HABERMEYER

THE JOURNAL OF BONE AND JOINT SURGERY

documented clinical and radiological results of the osteo-chondral transplantation procedure in the knee and anklejoints as well as the surgical technique and alternative ther-apeutic options

Pre- and post-operatively the clinical evaluation includedthe Constant score for the shoulder and the Lysholm scorefor the knee joint2324 Patients were reviewed at six weekssix months and one year post-operatively and at the mostrecent follow-up The final assessment was performed by anindependent examiner and not by the surgeon Standardradiographs (true anteroposterior true anteroposterior inexternal and internal rotation and axillary views) were usedto assess the presence of glenohumeral osteoarthritis Signsof glenohumeral osteoarthritis were graded according tothe classification of Samilson and Prieto22 MRI was per-formed post-operatively in every patient in order to assessthe integrity of the osteochondral grafts As osteochondral

autologous transplantation has not been fully investigatedin the shoulder a second-look arthroscopy six months post-operatively was recommended to all patients but per-formed in only two The remaining patients refused furthersurgical intervention

Statistical analysis between variables pre- and post-oper-atively was performed with the Wilcoxon signed-rank testfor non-parametric data The level of significance was set atp lt 005 The statistical software used was StatView (Aba-cus Concepts Inc Berkeley California)Operative technique All procedures were performed in thebeach chair position under interscalene block and generalanaesthesia A diagnostic glenohumeral arthroscopy wasperformed using a standard posterior portal According tothe International Cartilage Repair Society the defect wasclassified on the basis of its location size thickness and thecondition of the opposing articular surface In each patient

Table I Characteristics of the patients defects and procedures

Patient GenderAge (yrs) Aetiology Location

Size of defect (mm2)

Number of cylinders used

Additional procedures performed

1 M 57 Traumatic anterior instability

Central (h) 120 1 Labral augmenta-tion capsular shift

2 M 38 Traumatic anterior instability

Posteromedial (h) 150 3 None

3 M 56 Post-traumatic Central (h) 105 2 None4 M 53 Hyperlaxity Posterocentral (h) 105 2 Labral augmenta-

tion capsular shift5 F 43 Traumatic anterior

instabilityAnterocentral (g) 128 1 Labral augmenta-

tion capsular shift6 M 44 Hyperlaxity Central (h) 144 1 None7 M 23 Traumatic posterior

subluxationAnterocentral (h) 250 3 None

8 F 31 Traumatic anterior instability

Posteromedial (h) 200 2 Labral augmenta-tion capsular shift

h humeral g glenoid

Fig 1a Fig 1b

Intra-operative findings A grade IV osteochondral lesion of the humeral head a) before and b) after autologous transplantation with two osteochondralplugs

OSTEOCHONDRAL AUTOLOGOUS TRANSPLANTATION 993

VOL 86-B No 7 SEPTEMBER 2004

the stage of the osteochondral lesion was Outerbridgegrade IV with a mean size of 150 mm2 (105 to 250 mm2) Inseven patients the defect was located on the humeral sideand in one patient it was on the glenoid side In threepatients the defect was central in position in three it wasposteromedial and in one it was posterocentral on thehumeral head The one glenoid defect was positionedanteromedially (Table I)

The transplantation procedure was performed with anosteochondral autograft transfer system (Arthrex NaplesFlorida) The shoulder joint was exposed through a stand-ard deltopectoral approach and the cephalic vein was pre-served and retracted laterally The subscapularis musclewas detached from the lesser tuberosity approximately 05cm from its insertion and stay sutures were attached forsubsequent refixation and the muscle was then retractedmedially If capsular instability was present the subscapula-ris was stripped from the anterior capsule in order to per-form a capsular shift at the end of the procedure Thehumeral head was next exposed and the defect wasinspected (Fig 1a) The size of the lesion was again meas-ured using a range of appropriate colour-coded sizers Arecipient socket was then created in order to provide anadequate press-fit graft fixation The appropriately sizedtubular harvester was introduced into the joint and placedover the affected area taking great care to ensure that theharvester was perpendicular to the articular surface Thedecision to transplant single or multiple osteochondralgrafts was based upon the size and location of the defectEither one (three patients) two (three patients) or three(two patients) osteochondral plugs were used in order to fillthe defects (Table I) Next the osteochondral graft wastaken The knee joint was exposed through a lateral mini-arthrotomy The chosen donor site was in an area along theouter edge of the lateral femoral condyle immediatelyabove the sulcus terminalis This is a low weight-bearingarea and offers a convex articular surface similar to that ofthe central humeral head25 The tube harvester was insertedinto the joint and placed perpendicular to the selected har-vest site It was then driven into the bone to the same depthas the osteochondral defect and the graft was harvestedThe donor cavities were not filled The harvested grafts

were then transferred into the prepared recipient site usinga press-fit technique (Fig 1b)

In order to address any underlying instability or capsularredundancy an additional labral augmentation (Harrymanet al26) and a lateral capsular shift (Matsen et al27) was per-formed in four patients Finally the subscapularis tendonwas reattached anatomically with the arm in 30˚ of abduc-tion and 20˚ of external rotationPost-operative management Post-operatively the patientsunderwent a standard rehabilitation programme Theaffected arm was placed in an abduction pillow for threeweeks Passive range of movement was initiated on thethird post-operative day and was restricted to 60˚ of flex-ion abduction and internal rotation for the first two weeksand then increased to 90˚ up to the five-week point Exter-nal rotation was prohibited up to the six-week point inorder to protect the reconstructed anterior capsule and sub-scapularis tendon Four to six weeks post-operatively thepatient started active movements and a muscle strengthen-ing programme

Results

The mean age of the patients was 431 years (23 to 57) Thedominant shoulder was affected in six patients Full detailsare summarised in Table IClinical Seven patients were available for medium termfollow-up and one for short-term follow-up (Table II)After a mean follow-up of 326 months (8 to 47) the overallConstant score improved significantly from 739 points (57to 896) to 887 points (824 to 954) (p lt 005) All patientshad less pain when compared with their original pre-oper-ative rating There was a significant (p lt 005) increase inthe overall pain rating scale from the initial mean score of87 (5 to 13) to the latest follow-up score of 145 (13 to 15)Six patients were completely free of pain The mean activi-ties of daily living (ADL) rate increased significantly(p lt 005) from 129 (7 to 18) to 191 (18 to 20) All buttwo patients achieved their full work and sporting activitylevels The results for range of movement and strength alsoincreased although these were not statistically significant(p gt 005 for both) and probably due to pain relief Thefunctional outcome of the knee was graded with the

Table II Clinical results

Constant score (pre-operatively)

Follow-up (mths)

Constant score (post-operatively) Lysholm score

Patient Pain ADL ROMdaggerStrength (kg) Total Pain ADL ROM

Strength (kg) Total Total

1 10 9 38 97 772 36 15 20 36 106 931 1002 10 12 40 99 826 41 13 18 40 111 941 1003 5 10 38 84 705 47 13 19 40 69 863 914 13 16 30 50 694 44 15 18 36 69 834 1005 12 17 40 45 784 30 15 20 40 60 875 1006 5 14 40 33 659 24 15 20 40 61 877 1007 10 18 40 104 896 31 15 20 40 99 954 878 5 7 40 27 576 8 15 18 40 47 824 64

ADL activities of daily livingdagger ROM range of movement

994 M SCHEIBEL C BARTL P MAGOSCH S LICHTENBERG P HABERMEYER

THE JOURNAL OF BONE AND JOINT SURGERY

Lysholm score24 Six patients had an excellent result (91 to100 points) one patient had a satisfactory result (77 to 90points) and one patient had a poor result (0 to 76 points)Full clinical results are shown in Table IIRadiographic Pre-operatively four patients had stage Iglenohumeral osteoarthritic changes22 In all but one theosteoarthritis had deteriorated between the pre- and post-operative assessments The patient who did not progresswas only available for short-term follow-up Four patientsdid not show any signs of glenohumeral osteoarthritis pre-operatively (Fig 2a) However by their latest follow-up allfour had developed new inferior humeral osteophyte for-mation (Fig 2b) although this did not influence the finalclinical result Despite an increase in the size of humeralosteophytes the mean glenohumeral distance did not signif-icantly reduce between the pre-operative (44 mm) andpost-operative (42 mm) examinations (p gt 005)

MRI which was performed in the paracoronal trans-axial and parasagittal planes revealed excellent graft via-bility and congruence of the chondral surfaces (Fig 3) in allbut one patient In one patient signs of transplant insuffi-ciency with evidence of avascular necrosis could be seenHowever the patient was clinically asymptomatic and satis-fied with the post-operative result Full radiographic resultsare shown in Table IIISecond-look arthroscopy In two patients a second-lookarthroscopy (Table III) was performed six months post-operatively in order to assess the chondral surfaces In boththere was good macroscopic integration of the grafts with

the original osteochondral defect being completely coveredby chondral tissue (Fig 4) The surrounding chondral sur-face showed some superficial fissuring representing a gradeI chondral lesion according to Outerbridge21

Complications No complications which could be directlyrelated to the surgical procedure in the shoulder were seen

Fig 2a Fig 2b

Anteroposterior radiograph of a shoulder a) before and b) 24 months after autologous osteochondral transplanta-tion Note the increase in inferior osteophyte formation between the pre- and post-operative examinations

Fig 3

Post-operative paracoronal MRI 24 months after auto-logous osteochondral transplantation showing anintact osteochondral plug and a congruent articularsurface

OSTEOCHONDRAL AUTOLOGOUS TRANSPLANTATION 995

VOL 86-B No 7 SEPTEMBER 2004

However one patient reported persistent pain and recur-rent effusions of the donor knee This patient achieved apoor Lysholm score result (lt 75 points) by their latestfollow-up Because of donor site morbidity the patientunderwent two revision arthroscopic surgical procedureswith debridement of the knee joint Although the shoulderwas completely painless eight months post-operatively thepatient was dissatisfied with the procedure and wasunavailable for medium-term follow-up

Discussion

Osteochondral lesions of the shoulder are less commonthan those of the lower extremity but can cause considera-ble symptoms These might include joint pain effusion andmechanical dysfunction Although the natural history ofisolated osteochondral defects has not been well definedclinical experience has shown that these lesions mayprogress to symptomatic degeneration of the joint Conse-quently the treatment of selected isolated articular carti-lage lesions may delay or prevent the development ofosteoarthritis

The aim of all methods of articular cartilage restorationis to reproduce the mechanical structural and biochemicalproperties of the original hyaline articular surfaceAlthough different studies have reported good and excel-lent clinical results from autologous chondrocyte implanta-tion (ACI) osteochondral autograft transplantation iscurrently the only technique that appears to maintain thecharacteristics of hyaline cartilage151628-30 Bobic13 statedthat the main reason for the long-term survival of trans-planted hyaline cartilage was the preservation of an intacttidemark and cancellous bone barrier Transplantation ofarticular cartilage as a part of an osteochondral graft hasbeen shown to be an effective method of replacing focalareas of damaged articular cartilage and reducing pain inboth the knee and ankle joints12-20

Our study shows that osteochondral transplantationfrom the knee to the shoulder results in a good clinical out-come in terms of pain relief and functional recovery Weobserved a significant increase in the overall Constant scorebetween the pre- and post-operative assessments At theirlatest follow-up all but two of our patients were com-

Table III Staging radiographic MRI and arthroscopic results

Patient

Stage of osteoarthritis18 Glenohumeral distance (mm)Graft integrity on MRI

Graft integrity at second-look arthroscopyPre-operative Post-operative Pre-operative Post-operative

1 I II 5 5 Intact Congruent surface2 I II 4 4 Intact Not performed3 None III 4 4 Intact Not performed4 None II 4 3 Avascular necrosis Not performed5 None I 5 5 Intact Not performed6 I II 3 3 Intact Not performed7 None I 5 5 Intact Congruent surface8 I I 5 5 Intact Not performed

Fig 4a Fig 4b

Arthroscopic view of a grade IV osteochondral lesion a) before and b) six months after autologous transplantation

996 M SCHEIBEL C BARTL P MAGOSCH S LICHTENBERG P HABERMEYER

THE JOURNAL OF BONE AND JOINT SURGERY

pletely pain-free and able to perform all activities of dailyliving Two patients had even returned to overhead sportsactivities It is important to understand that the functionalresults which our patients obtained are not only due to theosteochondral transplantation procedure but also to thecorrection of the underlying and additional glenohumeralpathology In four patients with recurrent instability of theshoulder a capsular shift with labral augmentation was per-formed

It is significant that in all of our patients glenohumeralosteoarthritic changes were present at their latest follow-up Although the glenohumeral distance on the antero-posterior radiograph did not change significantly inferiorosteophyte formation was seen in all patients For thosewho had stage I osteoarthritis pre-operatively all but oneshowed a significant deterioration between the pre- andpost-operative assessments although this was never bymore than one stage The one patient who did not progresswas only available for short-term follow-up Four patientshad no signs of glenohumeral osteoarthritis pre-operativelyHowever by their latest follow-up all had developed newinferior humeral osteophyte formation Despite this theseosteoarthritic changes did not correlate with either post-operative pain or the final post-operative result Thisprogression in glenohumeral osteoarthritis still poses thequestion as to whether osteochondral transplantation mayslow down the development of symptomatic glenohumeralosteoarthritis at all In all but one patient MRI follow-uprevealed excellent graft viability and congruence of thechondral surfaces In one patient there were signs of graftinsufficiency probably due to avascular necrosis Howeverthis patient was pain-free and very satisfied with the result

The potential disadvantages of this technique includedonor site morbidity a limited supply of grafts dead spacebetween circular grafts graft integration and the differentmechanical properties and geometry between donor andrecipient hyaline cartilage2-4 Donor site morbidity is per-haps the most important risk factor when performing anosteochondral transplantation from the knee to the shoul-der Studies have shown that donor sites normally refill withcancellous bone and fibrocartilage and do not cause signif-icant problems131931 However persistent pain and recur-rent effusions of the donor knee were seen in one of ourpatients Although the shoulder was pain-free the patientrequired two revision arthroscopic procedures Donor sitemorbidity can clearly be a serious problem and must betaken into consideration when counselling a patient foran osteochondral transplantation from the knee to theshoulder

The ideal osteochondral defect for an osteochondralautologous transplantation in the shoulder is relativelysmall perhaps 10 to 20 mm in diameter13 In our studygroup the mean size of the affected area was 150 mm2 (105to 250) Large osteochondral defects are not suitable forthis technique as osteochondral grafts are limited and it isalso technically difficult to reconstruct a large subchondral

defect A large number of grafts may also lead to instabilityin the transplantation area Other problems with thistechnique are differences in the thickness biomechanicalcomposition and mechanical properties of the articularcartilage of the knee joint when compared with theshoulder

Although we experienced no difficulties in performingthis procedure the final clinical results may be verytechnique-dependent There are many peri-operative pit-falls that need to be considered In particular an appropri-ate length of graft is essential to a successful outcome If toolong or too short an osteochondral cylinder can lead toincongruity of the articular surface Also if the graft is notinserted in an orthograde fashion an adequate press-fitcannot be achieved which may lead to loosening or failedintegration

Our study has certain limitations It includes only a smallnumber of patients so that statistical analysis is restrictedDespite this to our knowledge it is the first study to docu-ment the medium-term clinical and radiographic results forpatients with full-thickness cartilage lesions of the shoulderwho have been treated by osteochondral autologous trans-plantation It must also be remembered that not all of thepatients had the same aetiology for their osteochondraldefect Finally in three patients the associated underlyingpathology was also corrected which must be taken intoaccount when interpreting the clinical results

In summary osteochondral autologous transplantationin the shoulder appears to offer good clinical results for thetreatment of osteochondral lesions of the glenohumeraljoint However the results of our study suggest that thedevelopment of osteoarthritis and the progression of pre-existing osteoarthritic changes cannot be altered by thetechnique

No benefits in any form have been received or will be received from a commer-cial party related directly to the subject of this article

References1 Buckwalter JA Mankin HJ Articular cartilage Part I tissue design and chondro-

cyte-matrix interactions J Bone Joint Surg [Am] 199779-A600-11

2 Buckwalter JA Mankin HJ Articular cartilage II degeneration and osteoarthro-sis repair regeneration and transplantation J Bone Joint Surg [Am] 199779-A612-32

3 Mandelbaum BR Browne JE Fu F et al Articular cartilage lesions of the kneeAm J Sports Med 199826853-61

4 Newman AP Articular cartilage repair Am J Sports Med 199826309-24

5 Baumgaertner MR Cannon WD Jr Vittori JM Schmidt ES Maurer EC Arthro-scopic debridement of the arthritic knee Clin Orthop 1990253197-202

6 Jackson RW Marans HJ Silver RS Arthroscopic treatment of degenerativearthritis of the knee J Bone Joint Surg [Am] 198870-A332

7 Weinstein DM Bucchieris JS Pollock RG Flatow EL Bigliani LU Arthro-scopic debridement of the shoulder for osteoarthritis Arthroscopy 200016471-6

8 Moseley JB OrsquoMalley K Petersen NJ et al A controlled trial of arthroscopic sur-gery for osteoarthritis of the knee N Engl J Med 200234781-8

9 Childers JC Jr Ellwood SC Partial chondrectomy and subchondral bone drilling forchondramalacia Clin Orthop 1979144114-20

10 Insall J The Pridie debridement operation for osteoarthritis of the knee Clin Orthop197410161-7

11 Johnson LL Arthroscopic abrasion arthroplasty historical and pathological perspec-tive present status Arthroscopy 1986254-69

OSTEOCHONDRAL AUTOLOGOUS TRANSPLANTATION 993

VOL 86-B No 7 SEPTEMBER 2004

the stage of the osteochondral lesion was Outerbridgegrade IV with a mean size of 150 mm2 (105 to 250 mm2) Inseven patients the defect was located on the humeral sideand in one patient it was on the glenoid side In threepatients the defect was central in position in three it wasposteromedial and in one it was posterocentral on thehumeral head The one glenoid defect was positionedanteromedially (Table I)

The transplantation procedure was performed with anosteochondral autograft transfer system (Arthrex NaplesFlorida) The shoulder joint was exposed through a stand-ard deltopectoral approach and the cephalic vein was pre-served and retracted laterally The subscapularis musclewas detached from the lesser tuberosity approximately 05cm from its insertion and stay sutures were attached forsubsequent refixation and the muscle was then retractedmedially If capsular instability was present the subscapula-ris was stripped from the anterior capsule in order to per-form a capsular shift at the end of the procedure Thehumeral head was next exposed and the defect wasinspected (Fig 1a) The size of the lesion was again meas-ured using a range of appropriate colour-coded sizers Arecipient socket was then created in order to provide anadequate press-fit graft fixation The appropriately sizedtubular harvester was introduced into the joint and placedover the affected area taking great care to ensure that theharvester was perpendicular to the articular surface Thedecision to transplant single or multiple osteochondralgrafts was based upon the size and location of the defectEither one (three patients) two (three patients) or three(two patients) osteochondral plugs were used in order to fillthe defects (Table I) Next the osteochondral graft wastaken The knee joint was exposed through a lateral mini-arthrotomy The chosen donor site was in an area along theouter edge of the lateral femoral condyle immediatelyabove the sulcus terminalis This is a low weight-bearingarea and offers a convex articular surface similar to that ofthe central humeral head25 The tube harvester was insertedinto the joint and placed perpendicular to the selected har-vest site It was then driven into the bone to the same depthas the osteochondral defect and the graft was harvestedThe donor cavities were not filled The harvested grafts

were then transferred into the prepared recipient site usinga press-fit technique (Fig 1b)

In order to address any underlying instability or capsularredundancy an additional labral augmentation (Harrymanet al26) and a lateral capsular shift (Matsen et al27) was per-formed in four patients Finally the subscapularis tendonwas reattached anatomically with the arm in 30˚ of abduc-tion and 20˚ of external rotationPost-operative management Post-operatively the patientsunderwent a standard rehabilitation programme Theaffected arm was placed in an abduction pillow for threeweeks Passive range of movement was initiated on thethird post-operative day and was restricted to 60˚ of flex-ion abduction and internal rotation for the first two weeksand then increased to 90˚ up to the five-week point Exter-nal rotation was prohibited up to the six-week point inorder to protect the reconstructed anterior capsule and sub-scapularis tendon Four to six weeks post-operatively thepatient started active movements and a muscle strengthen-ing programme

Results

The mean age of the patients was 431 years (23 to 57) Thedominant shoulder was affected in six patients Full detailsare summarised in Table IClinical Seven patients were available for medium termfollow-up and one for short-term follow-up (Table II)After a mean follow-up of 326 months (8 to 47) the overallConstant score improved significantly from 739 points (57to 896) to 887 points (824 to 954) (p lt 005) All patientshad less pain when compared with their original pre-oper-ative rating There was a significant (p lt 005) increase inthe overall pain rating scale from the initial mean score of87 (5 to 13) to the latest follow-up score of 145 (13 to 15)Six patients were completely free of pain The mean activi-ties of daily living (ADL) rate increased significantly(p lt 005) from 129 (7 to 18) to 191 (18 to 20) All buttwo patients achieved their full work and sporting activitylevels The results for range of movement and strength alsoincreased although these were not statistically significant(p gt 005 for both) and probably due to pain relief Thefunctional outcome of the knee was graded with the

Table II Clinical results

Constant score (pre-operatively)

Follow-up (mths)

Constant score (post-operatively) Lysholm score

Patient Pain ADL ROMdaggerStrength (kg) Total Pain ADL ROM

Strength (kg) Total Total

1 10 9 38 97 772 36 15 20 36 106 931 1002 10 12 40 99 826 41 13 18 40 111 941 1003 5 10 38 84 705 47 13 19 40 69 863 914 13 16 30 50 694 44 15 18 36 69 834 1005 12 17 40 45 784 30 15 20 40 60 875 1006 5 14 40 33 659 24 15 20 40 61 877 1007 10 18 40 104 896 31 15 20 40 99 954 878 5 7 40 27 576 8 15 18 40 47 824 64

ADL activities of daily livingdagger ROM range of movement

994 M SCHEIBEL C BARTL P MAGOSCH S LICHTENBERG P HABERMEYER

THE JOURNAL OF BONE AND JOINT SURGERY

Lysholm score24 Six patients had an excellent result (91 to100 points) one patient had a satisfactory result (77 to 90points) and one patient had a poor result (0 to 76 points)Full clinical results are shown in Table IIRadiographic Pre-operatively four patients had stage Iglenohumeral osteoarthritic changes22 In all but one theosteoarthritis had deteriorated between the pre- and post-operative assessments The patient who did not progresswas only available for short-term follow-up Four patientsdid not show any signs of glenohumeral osteoarthritis pre-operatively (Fig 2a) However by their latest follow-up allfour had developed new inferior humeral osteophyte for-mation (Fig 2b) although this did not influence the finalclinical result Despite an increase in the size of humeralosteophytes the mean glenohumeral distance did not signif-icantly reduce between the pre-operative (44 mm) andpost-operative (42 mm) examinations (p gt 005)

MRI which was performed in the paracoronal trans-axial and parasagittal planes revealed excellent graft via-bility and congruence of the chondral surfaces (Fig 3) in allbut one patient In one patient signs of transplant insuffi-ciency with evidence of avascular necrosis could be seenHowever the patient was clinically asymptomatic and satis-fied with the post-operative result Full radiographic resultsare shown in Table IIISecond-look arthroscopy In two patients a second-lookarthroscopy (Table III) was performed six months post-operatively in order to assess the chondral surfaces In boththere was good macroscopic integration of the grafts with

the original osteochondral defect being completely coveredby chondral tissue (Fig 4) The surrounding chondral sur-face showed some superficial fissuring representing a gradeI chondral lesion according to Outerbridge21

Complications No complications which could be directlyrelated to the surgical procedure in the shoulder were seen

Fig 2a Fig 2b

Anteroposterior radiograph of a shoulder a) before and b) 24 months after autologous osteochondral transplanta-tion Note the increase in inferior osteophyte formation between the pre- and post-operative examinations

Fig 3

Post-operative paracoronal MRI 24 months after auto-logous osteochondral transplantation showing anintact osteochondral plug and a congruent articularsurface

OSTEOCHONDRAL AUTOLOGOUS TRANSPLANTATION 995

VOL 86-B No 7 SEPTEMBER 2004

However one patient reported persistent pain and recur-rent effusions of the donor knee This patient achieved apoor Lysholm score result (lt 75 points) by their latestfollow-up Because of donor site morbidity the patientunderwent two revision arthroscopic surgical procedureswith debridement of the knee joint Although the shoulderwas completely painless eight months post-operatively thepatient was dissatisfied with the procedure and wasunavailable for medium-term follow-up

Discussion

Osteochondral lesions of the shoulder are less commonthan those of the lower extremity but can cause considera-ble symptoms These might include joint pain effusion andmechanical dysfunction Although the natural history ofisolated osteochondral defects has not been well definedclinical experience has shown that these lesions mayprogress to symptomatic degeneration of the joint Conse-quently the treatment of selected isolated articular carti-lage lesions may delay or prevent the development ofosteoarthritis

The aim of all methods of articular cartilage restorationis to reproduce the mechanical structural and biochemicalproperties of the original hyaline articular surfaceAlthough different studies have reported good and excel-lent clinical results from autologous chondrocyte implanta-tion (ACI) osteochondral autograft transplantation iscurrently the only technique that appears to maintain thecharacteristics of hyaline cartilage151628-30 Bobic13 statedthat the main reason for the long-term survival of trans-planted hyaline cartilage was the preservation of an intacttidemark and cancellous bone barrier Transplantation ofarticular cartilage as a part of an osteochondral graft hasbeen shown to be an effective method of replacing focalareas of damaged articular cartilage and reducing pain inboth the knee and ankle joints12-20

Our study shows that osteochondral transplantationfrom the knee to the shoulder results in a good clinical out-come in terms of pain relief and functional recovery Weobserved a significant increase in the overall Constant scorebetween the pre- and post-operative assessments At theirlatest follow-up all but two of our patients were com-

Table III Staging radiographic MRI and arthroscopic results

Patient

Stage of osteoarthritis18 Glenohumeral distance (mm)Graft integrity on MRI

Graft integrity at second-look arthroscopyPre-operative Post-operative Pre-operative Post-operative

1 I II 5 5 Intact Congruent surface2 I II 4 4 Intact Not performed3 None III 4 4 Intact Not performed4 None II 4 3 Avascular necrosis Not performed5 None I 5 5 Intact Not performed6 I II 3 3 Intact Not performed7 None I 5 5 Intact Congruent surface8 I I 5 5 Intact Not performed

Fig 4a Fig 4b

Arthroscopic view of a grade IV osteochondral lesion a) before and b) six months after autologous transplantation

996 M SCHEIBEL C BARTL P MAGOSCH S LICHTENBERG P HABERMEYER

THE JOURNAL OF BONE AND JOINT SURGERY

pletely pain-free and able to perform all activities of dailyliving Two patients had even returned to overhead sportsactivities It is important to understand that the functionalresults which our patients obtained are not only due to theosteochondral transplantation procedure but also to thecorrection of the underlying and additional glenohumeralpathology In four patients with recurrent instability of theshoulder a capsular shift with labral augmentation was per-formed

It is significant that in all of our patients glenohumeralosteoarthritic changes were present at their latest follow-up Although the glenohumeral distance on the antero-posterior radiograph did not change significantly inferiorosteophyte formation was seen in all patients For thosewho had stage I osteoarthritis pre-operatively all but oneshowed a significant deterioration between the pre- andpost-operative assessments although this was never bymore than one stage The one patient who did not progresswas only available for short-term follow-up Four patientshad no signs of glenohumeral osteoarthritis pre-operativelyHowever by their latest follow-up all had developed newinferior humeral osteophyte formation Despite this theseosteoarthritic changes did not correlate with either post-operative pain or the final post-operative result Thisprogression in glenohumeral osteoarthritis still poses thequestion as to whether osteochondral transplantation mayslow down the development of symptomatic glenohumeralosteoarthritis at all In all but one patient MRI follow-uprevealed excellent graft viability and congruence of thechondral surfaces In one patient there were signs of graftinsufficiency probably due to avascular necrosis Howeverthis patient was pain-free and very satisfied with the result

The potential disadvantages of this technique includedonor site morbidity a limited supply of grafts dead spacebetween circular grafts graft integration and the differentmechanical properties and geometry between donor andrecipient hyaline cartilage2-4 Donor site morbidity is per-haps the most important risk factor when performing anosteochondral transplantation from the knee to the shoul-der Studies have shown that donor sites normally refill withcancellous bone and fibrocartilage and do not cause signif-icant problems131931 However persistent pain and recur-rent effusions of the donor knee were seen in one of ourpatients Although the shoulder was pain-free the patientrequired two revision arthroscopic procedures Donor sitemorbidity can clearly be a serious problem and must betaken into consideration when counselling a patient foran osteochondral transplantation from the knee to theshoulder

The ideal osteochondral defect for an osteochondralautologous transplantation in the shoulder is relativelysmall perhaps 10 to 20 mm in diameter13 In our studygroup the mean size of the affected area was 150 mm2 (105to 250) Large osteochondral defects are not suitable forthis technique as osteochondral grafts are limited and it isalso technically difficult to reconstruct a large subchondral

defect A large number of grafts may also lead to instabilityin the transplantation area Other problems with thistechnique are differences in the thickness biomechanicalcomposition and mechanical properties of the articularcartilage of the knee joint when compared with theshoulder

Although we experienced no difficulties in performingthis procedure the final clinical results may be verytechnique-dependent There are many peri-operative pit-falls that need to be considered In particular an appropri-ate length of graft is essential to a successful outcome If toolong or too short an osteochondral cylinder can lead toincongruity of the articular surface Also if the graft is notinserted in an orthograde fashion an adequate press-fitcannot be achieved which may lead to loosening or failedintegration

Our study has certain limitations It includes only a smallnumber of patients so that statistical analysis is restrictedDespite this to our knowledge it is the first study to docu-ment the medium-term clinical and radiographic results forpatients with full-thickness cartilage lesions of the shoulderwho have been treated by osteochondral autologous trans-plantation It must also be remembered that not all of thepatients had the same aetiology for their osteochondraldefect Finally in three patients the associated underlyingpathology was also corrected which must be taken intoaccount when interpreting the clinical results

In summary osteochondral autologous transplantationin the shoulder appears to offer good clinical results for thetreatment of osteochondral lesions of the glenohumeraljoint However the results of our study suggest that thedevelopment of osteoarthritis and the progression of pre-existing osteoarthritic changes cannot be altered by thetechnique

No benefits in any form have been received or will be received from a commer-cial party related directly to the subject of this article

References1 Buckwalter JA Mankin HJ Articular cartilage Part I tissue design and chondro-

cyte-matrix interactions J Bone Joint Surg [Am] 199779-A600-11

2 Buckwalter JA Mankin HJ Articular cartilage II degeneration and osteoarthro-sis repair regeneration and transplantation J Bone Joint Surg [Am] 199779-A612-32

3 Mandelbaum BR Browne JE Fu F et al Articular cartilage lesions of the kneeAm J Sports Med 199826853-61

4 Newman AP Articular cartilage repair Am J Sports Med 199826309-24

5 Baumgaertner MR Cannon WD Jr Vittori JM Schmidt ES Maurer EC Arthro-scopic debridement of the arthritic knee Clin Orthop 1990253197-202

6 Jackson RW Marans HJ Silver RS Arthroscopic treatment of degenerativearthritis of the knee J Bone Joint Surg [Am] 198870-A332

7 Weinstein DM Bucchieris JS Pollock RG Flatow EL Bigliani LU Arthro-scopic debridement of the shoulder for osteoarthritis Arthroscopy 200016471-6

8 Moseley JB OrsquoMalley K Petersen NJ et al A controlled trial of arthroscopic sur-gery for osteoarthritis of the knee N Engl J Med 200234781-8

9 Childers JC Jr Ellwood SC Partial chondrectomy and subchondral bone drilling forchondramalacia Clin Orthop 1979144114-20

10 Insall J The Pridie debridement operation for osteoarthritis of the knee Clin Orthop197410161-7

11 Johnson LL Arthroscopic abrasion arthroplasty historical and pathological perspec-tive present status Arthroscopy 1986254-69

994 M SCHEIBEL C BARTL P MAGOSCH S LICHTENBERG P HABERMEYER

THE JOURNAL OF BONE AND JOINT SURGERY

Lysholm score24 Six patients had an excellent result (91 to100 points) one patient had a satisfactory result (77 to 90points) and one patient had a poor result (0 to 76 points)Full clinical results are shown in Table IIRadiographic Pre-operatively four patients had stage Iglenohumeral osteoarthritic changes22 In all but one theosteoarthritis had deteriorated between the pre- and post-operative assessments The patient who did not progresswas only available for short-term follow-up Four patientsdid not show any signs of glenohumeral osteoarthritis pre-operatively (Fig 2a) However by their latest follow-up allfour had developed new inferior humeral osteophyte for-mation (Fig 2b) although this did not influence the finalclinical result Despite an increase in the size of humeralosteophytes the mean glenohumeral distance did not signif-icantly reduce between the pre-operative (44 mm) andpost-operative (42 mm) examinations (p gt 005)

MRI which was performed in the paracoronal trans-axial and parasagittal planes revealed excellent graft via-bility and congruence of the chondral surfaces (Fig 3) in allbut one patient In one patient signs of transplant insuffi-ciency with evidence of avascular necrosis could be seenHowever the patient was clinically asymptomatic and satis-fied with the post-operative result Full radiographic resultsare shown in Table IIISecond-look arthroscopy In two patients a second-lookarthroscopy (Table III) was performed six months post-operatively in order to assess the chondral surfaces In boththere was good macroscopic integration of the grafts with

the original osteochondral defect being completely coveredby chondral tissue (Fig 4) The surrounding chondral sur-face showed some superficial fissuring representing a gradeI chondral lesion according to Outerbridge21

Complications No complications which could be directlyrelated to the surgical procedure in the shoulder were seen

Fig 2a Fig 2b

Anteroposterior radiograph of a shoulder a) before and b) 24 months after autologous osteochondral transplanta-tion Note the increase in inferior osteophyte formation between the pre- and post-operative examinations

Fig 3

Post-operative paracoronal MRI 24 months after auto-logous osteochondral transplantation showing anintact osteochondral plug and a congruent articularsurface

OSTEOCHONDRAL AUTOLOGOUS TRANSPLANTATION 995

VOL 86-B No 7 SEPTEMBER 2004

However one patient reported persistent pain and recur-rent effusions of the donor knee This patient achieved apoor Lysholm score result (lt 75 points) by their latestfollow-up Because of donor site morbidity the patientunderwent two revision arthroscopic surgical procedureswith debridement of the knee joint Although the shoulderwas completely painless eight months post-operatively thepatient was dissatisfied with the procedure and wasunavailable for medium-term follow-up

Discussion

Osteochondral lesions of the shoulder are less commonthan those of the lower extremity but can cause considera-ble symptoms These might include joint pain effusion andmechanical dysfunction Although the natural history ofisolated osteochondral defects has not been well definedclinical experience has shown that these lesions mayprogress to symptomatic degeneration of the joint Conse-quently the treatment of selected isolated articular carti-lage lesions may delay or prevent the development ofosteoarthritis

The aim of all methods of articular cartilage restorationis to reproduce the mechanical structural and biochemicalproperties of the original hyaline articular surfaceAlthough different studies have reported good and excel-lent clinical results from autologous chondrocyte implanta-tion (ACI) osteochondral autograft transplantation iscurrently the only technique that appears to maintain thecharacteristics of hyaline cartilage151628-30 Bobic13 statedthat the main reason for the long-term survival of trans-planted hyaline cartilage was the preservation of an intacttidemark and cancellous bone barrier Transplantation ofarticular cartilage as a part of an osteochondral graft hasbeen shown to be an effective method of replacing focalareas of damaged articular cartilage and reducing pain inboth the knee and ankle joints12-20

Our study shows that osteochondral transplantationfrom the knee to the shoulder results in a good clinical out-come in terms of pain relief and functional recovery Weobserved a significant increase in the overall Constant scorebetween the pre- and post-operative assessments At theirlatest follow-up all but two of our patients were com-

Table III Staging radiographic MRI and arthroscopic results

Patient

Stage of osteoarthritis18 Glenohumeral distance (mm)Graft integrity on MRI

Graft integrity at second-look arthroscopyPre-operative Post-operative Pre-operative Post-operative

1 I II 5 5 Intact Congruent surface2 I II 4 4 Intact Not performed3 None III 4 4 Intact Not performed4 None II 4 3 Avascular necrosis Not performed5 None I 5 5 Intact Not performed6 I II 3 3 Intact Not performed7 None I 5 5 Intact Congruent surface8 I I 5 5 Intact Not performed

Fig 4a Fig 4b

Arthroscopic view of a grade IV osteochondral lesion a) before and b) six months after autologous transplantation

996 M SCHEIBEL C BARTL P MAGOSCH S LICHTENBERG P HABERMEYER

THE JOURNAL OF BONE AND JOINT SURGERY

pletely pain-free and able to perform all activities of dailyliving Two patients had even returned to overhead sportsactivities It is important to understand that the functionalresults which our patients obtained are not only due to theosteochondral transplantation procedure but also to thecorrection of the underlying and additional glenohumeralpathology In four patients with recurrent instability of theshoulder a capsular shift with labral augmentation was per-formed

It is significant that in all of our patients glenohumeralosteoarthritic changes were present at their latest follow-up Although the glenohumeral distance on the antero-posterior radiograph did not change significantly inferiorosteophyte formation was seen in all patients For thosewho had stage I osteoarthritis pre-operatively all but oneshowed a significant deterioration between the pre- andpost-operative assessments although this was never bymore than one stage The one patient who did not progresswas only available for short-term follow-up Four patientshad no signs of glenohumeral osteoarthritis pre-operativelyHowever by their latest follow-up all had developed newinferior humeral osteophyte formation Despite this theseosteoarthritic changes did not correlate with either post-operative pain or the final post-operative result Thisprogression in glenohumeral osteoarthritis still poses thequestion as to whether osteochondral transplantation mayslow down the development of symptomatic glenohumeralosteoarthritis at all In all but one patient MRI follow-uprevealed excellent graft viability and congruence of thechondral surfaces In one patient there were signs of graftinsufficiency probably due to avascular necrosis Howeverthis patient was pain-free and very satisfied with the result

The potential disadvantages of this technique includedonor site morbidity a limited supply of grafts dead spacebetween circular grafts graft integration and the differentmechanical properties and geometry between donor andrecipient hyaline cartilage2-4 Donor site morbidity is per-haps the most important risk factor when performing anosteochondral transplantation from the knee to the shoul-der Studies have shown that donor sites normally refill withcancellous bone and fibrocartilage and do not cause signif-icant problems131931 However persistent pain and recur-rent effusions of the donor knee were seen in one of ourpatients Although the shoulder was pain-free the patientrequired two revision arthroscopic procedures Donor sitemorbidity can clearly be a serious problem and must betaken into consideration when counselling a patient foran osteochondral transplantation from the knee to theshoulder

The ideal osteochondral defect for an osteochondralautologous transplantation in the shoulder is relativelysmall perhaps 10 to 20 mm in diameter13 In our studygroup the mean size of the affected area was 150 mm2 (105to 250) Large osteochondral defects are not suitable forthis technique as osteochondral grafts are limited and it isalso technically difficult to reconstruct a large subchondral

defect A large number of grafts may also lead to instabilityin the transplantation area Other problems with thistechnique are differences in the thickness biomechanicalcomposition and mechanical properties of the articularcartilage of the knee joint when compared with theshoulder

Although we experienced no difficulties in performingthis procedure the final clinical results may be verytechnique-dependent There are many peri-operative pit-falls that need to be considered In particular an appropri-ate length of graft is essential to a successful outcome If toolong or too short an osteochondral cylinder can lead toincongruity of the articular surface Also if the graft is notinserted in an orthograde fashion an adequate press-fitcannot be achieved which may lead to loosening or failedintegration

Our study has certain limitations It includes only a smallnumber of patients so that statistical analysis is restrictedDespite this to our knowledge it is the first study to docu-ment the medium-term clinical and radiographic results forpatients with full-thickness cartilage lesions of the shoulderwho have been treated by osteochondral autologous trans-plantation It must also be remembered that not all of thepatients had the same aetiology for their osteochondraldefect Finally in three patients the associated underlyingpathology was also corrected which must be taken intoaccount when interpreting the clinical results

In summary osteochondral autologous transplantationin the shoulder appears to offer good clinical results for thetreatment of osteochondral lesions of the glenohumeraljoint However the results of our study suggest that thedevelopment of osteoarthritis and the progression of pre-existing osteoarthritic changes cannot be altered by thetechnique

No benefits in any form have been received or will be received from a commer-cial party related directly to the subject of this article

References1 Buckwalter JA Mankin HJ Articular cartilage Part I tissue design and chondro-

cyte-matrix interactions J Bone Joint Surg [Am] 199779-A600-11

2 Buckwalter JA Mankin HJ Articular cartilage II degeneration and osteoarthro-sis repair regeneration and transplantation J Bone Joint Surg [Am] 199779-A612-32

3 Mandelbaum BR Browne JE Fu F et al Articular cartilage lesions of the kneeAm J Sports Med 199826853-61

4 Newman AP Articular cartilage repair Am J Sports Med 199826309-24

5 Baumgaertner MR Cannon WD Jr Vittori JM Schmidt ES Maurer EC Arthro-scopic debridement of the arthritic knee Clin Orthop 1990253197-202

6 Jackson RW Marans HJ Silver RS Arthroscopic treatment of degenerativearthritis of the knee J Bone Joint Surg [Am] 198870-A332

7 Weinstein DM Bucchieris JS Pollock RG Flatow EL Bigliani LU Arthro-scopic debridement of the shoulder for osteoarthritis Arthroscopy 200016471-6

8 Moseley JB OrsquoMalley K Petersen NJ et al A controlled trial of arthroscopic sur-gery for osteoarthritis of the knee N Engl J Med 200234781-8

9 Childers JC Jr Ellwood SC Partial chondrectomy and subchondral bone drilling forchondramalacia Clin Orthop 1979144114-20

10 Insall J The Pridie debridement operation for osteoarthritis of the knee Clin Orthop197410161-7

11 Johnson LL Arthroscopic abrasion arthroplasty historical and pathological perspec-tive present status Arthroscopy 1986254-69

OSTEOCHONDRAL AUTOLOGOUS TRANSPLANTATION 995

VOL 86-B No 7 SEPTEMBER 2004

However one patient reported persistent pain and recur-rent effusions of the donor knee This patient achieved apoor Lysholm score result (lt 75 points) by their latestfollow-up Because of donor site morbidity the patientunderwent two revision arthroscopic surgical procedureswith debridement of the knee joint Although the shoulderwas completely painless eight months post-operatively thepatient was dissatisfied with the procedure and wasunavailable for medium-term follow-up

Discussion

Osteochondral lesions of the shoulder are less commonthan those of the lower extremity but can cause considera-ble symptoms These might include joint pain effusion andmechanical dysfunction Although the natural history ofisolated osteochondral defects has not been well definedclinical experience has shown that these lesions mayprogress to symptomatic degeneration of the joint Conse-quently the treatment of selected isolated articular carti-lage lesions may delay or prevent the development ofosteoarthritis

The aim of all methods of articular cartilage restorationis to reproduce the mechanical structural and biochemicalproperties of the original hyaline articular surfaceAlthough different studies have reported good and excel-lent clinical results from autologous chondrocyte implanta-tion (ACI) osteochondral autograft transplantation iscurrently the only technique that appears to maintain thecharacteristics of hyaline cartilage151628-30 Bobic13 statedthat the main reason for the long-term survival of trans-planted hyaline cartilage was the preservation of an intacttidemark and cancellous bone barrier Transplantation ofarticular cartilage as a part of an osteochondral graft hasbeen shown to be an effective method of replacing focalareas of damaged articular cartilage and reducing pain inboth the knee and ankle joints12-20

Our study shows that osteochondral transplantationfrom the knee to the shoulder results in a good clinical out-come in terms of pain relief and functional recovery Weobserved a significant increase in the overall Constant scorebetween the pre- and post-operative assessments At theirlatest follow-up all but two of our patients were com-

Table III Staging radiographic MRI and arthroscopic results

Patient

Stage of osteoarthritis18 Glenohumeral distance (mm)Graft integrity on MRI

Graft integrity at second-look arthroscopyPre-operative Post-operative Pre-operative Post-operative

1 I II 5 5 Intact Congruent surface2 I II 4 4 Intact Not performed3 None III 4 4 Intact Not performed4 None II 4 3 Avascular necrosis Not performed5 None I 5 5 Intact Not performed6 I II 3 3 Intact Not performed7 None I 5 5 Intact Congruent surface8 I I 5 5 Intact Not performed

Fig 4a Fig 4b

Arthroscopic view of a grade IV osteochondral lesion a) before and b) six months after autologous transplantation

996 M SCHEIBEL C BARTL P MAGOSCH S LICHTENBERG P HABERMEYER

THE JOURNAL OF BONE AND JOINT SURGERY

pletely pain-free and able to perform all activities of dailyliving Two patients had even returned to overhead sportsactivities It is important to understand that the functionalresults which our patients obtained are not only due to theosteochondral transplantation procedure but also to thecorrection of the underlying and additional glenohumeralpathology In four patients with recurrent instability of theshoulder a capsular shift with labral augmentation was per-formed

It is significant that in all of our patients glenohumeralosteoarthritic changes were present at their latest follow-up Although the glenohumeral distance on the antero-posterior radiograph did not change significantly inferiorosteophyte formation was seen in all patients For thosewho had stage I osteoarthritis pre-operatively all but oneshowed a significant deterioration between the pre- andpost-operative assessments although this was never bymore than one stage The one patient who did not progresswas only available for short-term follow-up Four patientshad no signs of glenohumeral osteoarthritis pre-operativelyHowever by their latest follow-up all had developed newinferior humeral osteophyte formation Despite this theseosteoarthritic changes did not correlate with either post-operative pain or the final post-operative result Thisprogression in glenohumeral osteoarthritis still poses thequestion as to whether osteochondral transplantation mayslow down the development of symptomatic glenohumeralosteoarthritis at all In all but one patient MRI follow-uprevealed excellent graft viability and congruence of thechondral surfaces In one patient there were signs of graftinsufficiency probably due to avascular necrosis Howeverthis patient was pain-free and very satisfied with the result

The potential disadvantages of this technique includedonor site morbidity a limited supply of grafts dead spacebetween circular grafts graft integration and the differentmechanical properties and geometry between donor andrecipient hyaline cartilage2-4 Donor site morbidity is per-haps the most important risk factor when performing anosteochondral transplantation from the knee to the shoul-der Studies have shown that donor sites normally refill withcancellous bone and fibrocartilage and do not cause signif-icant problems131931 However persistent pain and recur-rent effusions of the donor knee were seen in one of ourpatients Although the shoulder was pain-free the patientrequired two revision arthroscopic procedures Donor sitemorbidity can clearly be a serious problem and must betaken into consideration when counselling a patient foran osteochondral transplantation from the knee to theshoulder

The ideal osteochondral defect for an osteochondralautologous transplantation in the shoulder is relativelysmall perhaps 10 to 20 mm in diameter13 In our studygroup the mean size of the affected area was 150 mm2 (105to 250) Large osteochondral defects are not suitable forthis technique as osteochondral grafts are limited and it isalso technically difficult to reconstruct a large subchondral

defect A large number of grafts may also lead to instabilityin the transplantation area Other problems with thistechnique are differences in the thickness biomechanicalcomposition and mechanical properties of the articularcartilage of the knee joint when compared with theshoulder

Although we experienced no difficulties in performingthis procedure the final clinical results may be verytechnique-dependent There are many peri-operative pit-falls that need to be considered In particular an appropri-ate length of graft is essential to a successful outcome If toolong or too short an osteochondral cylinder can lead toincongruity of the articular surface Also if the graft is notinserted in an orthograde fashion an adequate press-fitcannot be achieved which may lead to loosening or failedintegration

Our study has certain limitations It includes only a smallnumber of patients so that statistical analysis is restrictedDespite this to our knowledge it is the first study to docu-ment the medium-term clinical and radiographic results forpatients with full-thickness cartilage lesions of the shoulderwho have been treated by osteochondral autologous trans-plantation It must also be remembered that not all of thepatients had the same aetiology for their osteochondraldefect Finally in three patients the associated underlyingpathology was also corrected which must be taken intoaccount when interpreting the clinical results

In summary osteochondral autologous transplantationin the shoulder appears to offer good clinical results for thetreatment of osteochondral lesions of the glenohumeraljoint However the results of our study suggest that thedevelopment of osteoarthritis and the progression of pre-existing osteoarthritic changes cannot be altered by thetechnique

No benefits in any form have been received or will be received from a commer-cial party related directly to the subject of this article

References1 Buckwalter JA Mankin HJ Articular cartilage Part I tissue design and chondro-

cyte-matrix interactions J Bone Joint Surg [Am] 199779-A600-11

2 Buckwalter JA Mankin HJ Articular cartilage II degeneration and osteoarthro-sis repair regeneration and transplantation J Bone Joint Surg [Am] 199779-A612-32

3 Mandelbaum BR Browne JE Fu F et al Articular cartilage lesions of the kneeAm J Sports Med 199826853-61

4 Newman AP Articular cartilage repair Am J Sports Med 199826309-24

5 Baumgaertner MR Cannon WD Jr Vittori JM Schmidt ES Maurer EC Arthro-scopic debridement of the arthritic knee Clin Orthop 1990253197-202

6 Jackson RW Marans HJ Silver RS Arthroscopic treatment of degenerativearthritis of the knee J Bone Joint Surg [Am] 198870-A332

7 Weinstein DM Bucchieris JS Pollock RG Flatow EL Bigliani LU Arthro-scopic debridement of the shoulder for osteoarthritis Arthroscopy 200016471-6

8 Moseley JB OrsquoMalley K Petersen NJ et al A controlled trial of arthroscopic sur-gery for osteoarthritis of the knee N Engl J Med 200234781-8

9 Childers JC Jr Ellwood SC Partial chondrectomy and subchondral bone drilling forchondramalacia Clin Orthop 1979144114-20

10 Insall J The Pridie debridement operation for osteoarthritis of the knee Clin Orthop197410161-7

11 Johnson LL Arthroscopic abrasion arthroplasty historical and pathological perspec-tive present status Arthroscopy 1986254-69

996 M SCHEIBEL C BARTL P MAGOSCH S LICHTENBERG P HABERMEYER

THE JOURNAL OF BONE AND JOINT SURGERY

pletely pain-free and able to perform all activities of dailyliving Two patients had even returned to overhead sportsactivities It is important to understand that the functionalresults which our patients obtained are not only due to theosteochondral transplantation procedure but also to thecorrection of the underlying and additional glenohumeralpathology In four patients with recurrent instability of theshoulder a capsular shift with labral augmentation was per-formed

It is significant that in all of our patients glenohumeralosteoarthritic changes were present at their latest follow-up Although the glenohumeral distance on the antero-posterior radiograph did not change significantly inferiorosteophyte formation was seen in all patients For thosewho had stage I osteoarthritis pre-operatively all but oneshowed a significant deterioration between the pre- andpost-operative assessments although this was never bymore than one stage The one patient who did not progresswas only available for short-term follow-up Four patientshad no signs of glenohumeral osteoarthritis pre-operativelyHowever by their latest follow-up all had developed newinferior humeral osteophyte formation Despite this theseosteoarthritic changes did not correlate with either post-operative pain or the final post-operative result Thisprogression in glenohumeral osteoarthritis still poses thequestion as to whether osteochondral transplantation mayslow down the development of symptomatic glenohumeralosteoarthritis at all In all but one patient MRI follow-uprevealed excellent graft viability and congruence of thechondral surfaces In one patient there were signs of graftinsufficiency probably due to avascular necrosis Howeverthis patient was pain-free and very satisfied with the result

The potential disadvantages of this technique includedonor site morbidity a limited supply of grafts dead spacebetween circular grafts graft integration and the differentmechanical properties and geometry between donor andrecipient hyaline cartilage2-4 Donor site morbidity is per-haps the most important risk factor when performing anosteochondral transplantation from the knee to the shoul-der Studies have shown that donor sites normally refill withcancellous bone and fibrocartilage and do not cause signif-icant problems131931 However persistent pain and recur-rent effusions of the donor knee were seen in one of ourpatients Although the shoulder was pain-free the patientrequired two revision arthroscopic procedures Donor sitemorbidity can clearly be a serious problem and must betaken into consideration when counselling a patient foran osteochondral transplantation from the knee to theshoulder

The ideal osteochondral defect for an osteochondralautologous transplantation in the shoulder is relativelysmall perhaps 10 to 20 mm in diameter13 In our studygroup the mean size of the affected area was 150 mm2 (105to 250) Large osteochondral defects are not suitable forthis technique as osteochondral grafts are limited and it isalso technically difficult to reconstruct a large subchondral

defect A large number of grafts may also lead to instabilityin the transplantation area Other problems with thistechnique are differences in the thickness biomechanicalcomposition and mechanical properties of the articularcartilage of the knee joint when compared with theshoulder

Although we experienced no difficulties in performingthis procedure the final clinical results may be verytechnique-dependent There are many peri-operative pit-falls that need to be considered In particular an appropri-ate length of graft is essential to a successful outcome If toolong or too short an osteochondral cylinder can lead toincongruity of the articular surface Also if the graft is notinserted in an orthograde fashion an adequate press-fitcannot be achieved which may lead to loosening or failedintegration

Our study has certain limitations It includes only a smallnumber of patients so that statistical analysis is restrictedDespite this to our knowledge it is the first study to docu-ment the medium-term clinical and radiographic results forpatients with full-thickness cartilage lesions of the shoulderwho have been treated by osteochondral autologous trans-plantation It must also be remembered that not all of thepatients had the same aetiology for their osteochondraldefect Finally in three patients the associated underlyingpathology was also corrected which must be taken intoaccount when interpreting the clinical results

In summary osteochondral autologous transplantationin the shoulder appears to offer good clinical results for thetreatment of osteochondral lesions of the glenohumeraljoint However the results of our study suggest that thedevelopment of osteoarthritis and the progression of pre-existing osteoarthritic changes cannot be altered by thetechnique

No benefits in any form have been received or will be received from a commer-cial party related directly to the subject of this article

References1 Buckwalter JA Mankin HJ Articular cartilage Part I tissue design and chondro-

cyte-matrix interactions J Bone Joint Surg [Am] 199779-A600-11

2 Buckwalter JA Mankin HJ Articular cartilage II degeneration and osteoarthro-sis repair regeneration and transplantation J Bone Joint Surg [Am] 199779-A612-32

3 Mandelbaum BR Browne JE Fu F et al Articular cartilage lesions of the kneeAm J Sports Med 199826853-61

4 Newman AP Articular cartilage repair Am J Sports Med 199826309-24

5 Baumgaertner MR Cannon WD Jr Vittori JM Schmidt ES Maurer EC Arthro-scopic debridement of the arthritic knee Clin Orthop 1990253197-202

6 Jackson RW Marans HJ Silver RS Arthroscopic treatment of degenerativearthritis of the knee J Bone Joint Surg [Am] 198870-A332

7 Weinstein DM Bucchieris JS Pollock RG Flatow EL Bigliani LU Arthro-scopic debridement of the shoulder for osteoarthritis Arthroscopy 200016471-6

8 Moseley JB OrsquoMalley K Petersen NJ et al A controlled trial of arthroscopic sur-gery for osteoarthritis of the knee N Engl J Med 200234781-8

9 Childers JC Jr Ellwood SC Partial chondrectomy and subchondral bone drilling forchondramalacia Clin Orthop 1979144114-20

10 Insall J The Pridie debridement operation for osteoarthritis of the knee Clin Orthop197410161-7

11 Johnson LL Arthroscopic abrasion arthroplasty historical and pathological perspec-tive present status Arthroscopy 1986254-69

OSTEOCHONDRAL AUTOLOGOUS TRANSPLANTATION 997

VOL 86-B No 7 SEPTEMBER 2004

12 Al-Shaikh RA Choul LB Mann JA Dreeben SM Prieksorn D Autologousosteochondral grafting in the knee indication results and reflections Foot Ankle Int2002401170-84

13 Bobic V Arthroscopic osteochondral autograft transplantation in anterior cruciateligament reconstruction a preliminary clinical study Knee Surg Sports TraumatolArthrosc 19963262-4

14 Gautier E Kolker D Jakob RP Treatment of cartilage defects of the talus by auto-logous osteochondral grafts J Bone Joint Surg [Br] 200284-B237-44

15 Hangody L Fules P Autologous osteochondral mosaicplasty for the treatment offull-thickness defects of weight-bearing joints ten years of experimental and clinicalexperience J Bone Joint Surg [Am] 200385-A(Suppl 2)25-32

16 Hangody L Kish G Karpati Z Eberhart R Treatment of osteochondritis dissecansof the talus use of a mosaicplasty technique a preliminary report Foot Ankle Int199718628-34

17 Hangody L Kish G Karpati Z Szerb I Udvarhelyi I Arthroscopic autogenousosteochondral mosaicplasty for the treatment of femoral condylar articular defectsKnee Surg Sports Traumatol Arthrosc 19975262-7

18 Jakob RP Franz T Gautier E Mainil-Varlet P Autologous osteochondral graftingin the knee indication results and reflections Clin Orthop 2002401170-84

19 Morelli M Nagamori J Miniaci A Management of chondral injuries of the kneeby osteochondral autogenous transfer (mosaicplasty) J Knee Surg 200215185-90

20 Sammarco GJ Makwana NK Treatment of talar osteochondral lesions using localosteochondral graft Foot Ankle Int 200223693-8

21 Outerbridge RE The etiology of chondromalacia patellae J Bone Joint Surg [Br]196143-B752-7

22 Samilson RL Prieto V Dislocation arthropathy of the shoulder J Bone Joint Surg[Am] 198365-A456-60

23 Constant CR Murley AH A clinical method of functional assessment of the shoul-der Clin Orthop 1987214160-4

24 Lysholm J Gillchist J Evaluation of knee ligament surgery with special emphasiseon use of a scoring scale Am J Sports Med 198210150-4

25 Simonian PT Sussmann PS Wickiewicz TL Paletta GA Warren RF Contactpressures at osteochondral donor sites in the knee Am J Sports Med 199826491-4

26 Harryman DT Ballmer FP Harris SL Sidles JA Arthroscopic labral repair to theglenoid rim Arthroscopy 19941020-30

27 Matsen FA III Lippitt SB Sidles JA Harryman II DT Practical evaluation andmanagement of the shoulder Philadelphia WB Saunders 1994

28 Bentley G Biant LC Carrington RW et al A prospective randomised comparisonof autologous chondrocyte implantation versis mosaicplasty for osteochondraldefects in the knee J Bone Joint Surg [Br] 200385-B223-30

29 Horas U Pelinkovic D Herr G Aigner T Schnettler R Autologous chondrocyteimplantation and osteochondral cylinder transplantation in cartilage repair of theknee joint a prospective comparative trial J Bone Joint Surg [Am] 200385-A185-92

30 Romeo AA Cole BJ Mazzocca AD et al Case report autologous chondrocyterepair of an articular defect in the humeral head Arthroscopy 200218925-9

31 Ahmad CS Guiney WB Drinkwater CJ Evaluation of donor site intrinsic healingresponse in autologous osteochondral grafting of the knee Arthroscopy 20021895-8