Abstract
Background:
The aim of this study was to radiographically analyze the long-term glenoid migration patterns following total shoulder arthroplasty to better understand the factors responsible for loosening.
Methods:
Complete radiographic follow-up of more than five years was available for 518 total shoulder arthroplasties performed for primary glenohumeral osteoarthritis with use of an anatomically designed prosthesis with a cemented, all-polyethylene, keeled glenoid component. Radiographs were assessed for humeral head subluxation, periprosthetic radiolucent lines, and shifting of the position of the glenoid component. The type of migration of the glenoid was defined according to the direction of tilt, or as subsidence in the case of medial migration.
Results:
Definite radiographic evidence of glenoid loosening was observed in 166 shoulders (32%) and was characterized by radiolucency of ≥2 mm over the entire bone-cement interface in thirty shoulders and by a migration of the glenoid component (shift or subsidence) in 136 shoulders. Three predominant patterns of migration of the glenoid component were observed: superior tilting in fifty-two shoulders (10%), subsidence in forty-one shoulders (7.9%), and posterior tilting in thirty-three shoulders (6.4%). Superior tilting of the glenoid was associated with three risk factors: low positioning of the glenoid component, superior tilt of the glenoid component on the immediate postoperative coronal plane radiographs, and superior subluxation of the humeral head (p < 0.05 for all). Subsidence of the glenoid component was associated with the use of reaming to optimize the seating and positioning of the glenoid component (p < 0.001). Posterior tilting of the glenoid component was associated with preoperative posterior subluxation (i.e., a Walch type-B glenoid) and with excessive reaming (p < 0.01 for both).
Conclusions:
The three patterns of migration observed in this study underscore the potential importance of the supporting bone beneath the glenoid component. In some shoulders, use of a keel or pegs to provide fixation of a polyethylene component in the absence of good support from subchondral bone may not be sufficient to resist compressive and eccentric forces, resulting in loosening. Preserving subchondral bone may be important for long-term longevity of the glenoid component.
Level of Evidence:
Therapeutic Level III. See Instructions for Authors for a complete description of levels of evidence.
Radiographic assessment of the results of shoulder arthroplasty with a cemented polyethylene glenoid component typically involves determining the presence or absence of radiolucency at the junction between the bone and cement1,2. The reported prevalence of radiolucent lines has ranged from 0% to 90% depending on the duration of follow-up, the etiology for which the arthroplasty was performed, the type of prosthesis used, the cementing technique, and the method of assessment (fluoroscopically positioned radiographs, standard radiographs, or computed tomography [CT] scans)3-9. These radiolucencies are associated with micromotion and glenoid loosening, but an accurate analysis of the causes has not been possible because of the variability in the reported rates and the short duration of follow-up and limited number of patients in previous series. Even if the radiolucent lines are not always clinically relevant, most researchers agree that they often progress to radiographic evidence of loosening (a radiolucent line >2 mm in width over the entire bone-cement interface) or migration of the glenoid component (tilt and/or subsidence)4,8,10-13.
The aim of this study was to analyze the type of migration of the glenoid component in a large number of patients who had undergone total shoulder arthroplasty for the same indication of primary osteoarthritis and had more than five years of follow-up. We hoped to develop a better understand of the factors responsible for glenoid loosening by analyzing the rates and patterns of migration and the influence of different surgical techniques.
Study Group
This retrospective multicenter study involved 704 total shoulder arthroplasties performed for primary glenohumeral osteoarthritis at ten European centers between 1991 and 2003. The same unconstrained, cemented, anatomic implant (Aequalis; Tornier, Edina, Minnesota), which has a cemented, keeled, all-polyethylene glenoid component, was used in all shoulders. Exclusion criteria included a diagnosis other than primary glenohumeral osteoarthritis, previous shoulder surgery, a rotator cuff tear involving more than an isolated tear of the supraspinatus, bone-grafting of the glenoid, and use of a metal-backed glenoid component. In addition, 108 of the 704 eligible total shoulder arthroplasties were excluded because the patients were lost to follow-up, did not want to participate in the study, or had incomplete preoperative records, leaving 596 shoulders in 542 patients who agreed to return for a clinical and radiographic evaluation. A complete series of radiographs was available for 518 shoulders (representing 73.6% of the arthroplasties performed) with a minimum follow-up of five years or until revision arthroplasty, and these formed the study group.
Radiographic Analysis
Anteroposterior radiographs were made under fluoroscopic control in neutral, internal rotation, and external rotation at the follow-up evaluation. An axillary radiograph was also made, but the technique used varied slightly between centers because this radiograph could not be made under fluoroscopic control. All radiographs were assessed several times and finally scored by consensus by three orthopaedic surgeons (A.A.Y., G.W., and D.G.) specialized in shoulder surgery and blinded to the origin of the radiographs.
Postoperative and follow-up radiographs were evaluated to determine the presence of humeral head subluxation, periprosthetic radiolucent lines, and/or a shift in the position of the glenoid component. Superior humeral head subluxation was evaluated according to the method described by Torchia et al.14. The presence of periprosthetic radiolucent lines around the glenoid component was scored according to the method described by Molé2 et al. (in which 0 points corresponds to no radiolucency and the maximum of 18 points corresponds to radiolucency of ≥2 mm in six zones). A score of 0 to 6 points was considered to indicate no loosening; 7 to 12 points, possible loosening; and 13 to 18 points, definite loosening. By convention, any migration or shift of the component received a score of 18 points and was also considered as definite radiographic evidence of loosening even in the absence of a periprosthetic radiolucent line. The type of migration was defined according to the direction of glenoid tilt, or as glenoid subsidence in the case of medial migration; two of the three examiners (who had been involved in the surgery in some cases) had to agree on the existence and type of migration. Since some of the findings (such as subsidence) had not been expected initially, repeated assessments of the radiographs were necessary to take into account new findings.
Analysis of the preoperative morphology of the glenoid in the horizontal plane was made with use of a computed tomography (CT) arthrogram, which was available for 371 of the 518 shoulders with complete clinical and radiographic data. The glenoid morphology was classified according to the system of Walch et al.15 and was type A1 in 101 shoulders, type A2 in 102, type B1 in seventy-two, type B2 in 92, type C in three, and not classifiable in one shoulder. The distribution of glenoid types was not homogeneous among the centers; however, although the difference was significant (p < 0.0001), we were unable to characterize specific, consistent variations according to the center.
Operative Technique
The surgical technique used for all shoulders has been described previously16; however, there were certain technical variations between centers. First, preparation of the slot in the glenoid to accommodate the keel of the component was performed with use of two different techniques, either the curettage technique described by Neer et al.6 or the bone compaction technique described by Gazielly17, according to surgeon preference. Second, some centers routinely performed no reaming of the glenoid surface and simply curetted any soft tissue from the glenoid face, whereas others used a motorized reamer to ream as much as necessary to achieve optimal seating and normal version of the glenoid component. Third, the back side of the glenoid component was changed from a flat to a convex geometry during the study period. A glenoid component with a flat back was used in 217 shoulders, and a component with a convex back was used in 301. Two small reamers were used to provide a flat surface for the glenoid component with a flat back. A convex-shaped reamer was used to prepare the glenoid for the component with a convex back if reaming was used; three reamers with the same radius of curvature but different size (small, medium, or large) were available, and the choice of size depended on the size of the glenoid and the glenoid component to be implanted. Fourth, the reaming technique was known to be more aggressive at one center, although it was not possible to objectively quantify this. The goal of the surgeon at this center was to obtain normal version and perfect seating even if this resulted in sacrifice of the subchondral bone; perforation of the anterior cortex resulting in leakage of cement was common at this center, presumably because of the decreased depth of bone in the keel slot (Fig. 1). Fifth, the cementing technique was not standardized among centers. In the majority of shoulders, the glenoid surface and the slot for the keel were irrigated with saline solution and dried with a sponge before injection of cement into the slot with a syringe. The glenoid component was then impacted and held with a pusher until the cement hardened.
Data Analysis
The chi-square test for independence was used for assessing the relationship between migration of the glenoid component and categorical variables (i.e., glenoid preparation with use of curettage or compaction, reaming of the glenoid surface, sex, and operative side). The Kruskal-Wallis test was used for assessing the relationship between migration and quantitative variables (i.e., age at the time of the operation and duration of follow-up). A p value of <0.05 was considered significant.
Source of Funding
There was no external funding for this study.
Five hundred and eighteen total shoulder arthroplasties were reviewed after a mean of 103.6 months (range, sixty to 219 months). Three hundred and eighty-three shoulders (74%) had five to ten years of follow-up, 109 shoulders (21%) had eleven to fifteen years of follow-up, and twenty-six shoulders (5%) had sixteen years or more of follow-up. The majority of patients were women (68.8%) and had had the arthroplasty performed in the dominant arm (60.7%). The average age at the time of surgery was 68.2 years (range, thirty-five to ninety years).
Definite radiographic evidence of loosening was observed in 166 shoulders (32.0%); 136 of these shoulders had migration of the glenoid component, and the evidence was limited to a radiolucency score of >12 in the other thirty shoulders. Evidence of possible loosening (a radiolucency score of 7 to 12) was observed in eighty shoulders (15.4%), and no evidence of loosening was observed in the remaining 272 shoulders (52.5%).
Migration of the glenoid component was associated with greater duration of follow-up (p < 0.0001), with a preoperative glenoid type of B1 or B2 compared with type A1 or A2 (p < 0.0001), with use of the curettage technique compared with bone compaction (p < 0.0001), with a glenoid back that was flat compared with convex (p < 0.0001), with arthroplasty involving the dominant side (p < 0.01), and with younger age at the time of the surgery (p < 0.01). With the numbers available, migration was not significantly associated with the sex of the patient (p = 0.34) or with the size of the glenoid component (p = 0.84). An a priori power analysis was not performed, and these nonsignificant findings are therefore potentially subject to Type-II statistical error.
Superior tilting of the glenoid component was observed in fifty-two shoulders (10.0%) and was associated with the duration of follow-up (p < 0.0001), superior subluxation of the humeral head on the anteroposterior radiograph at the time of the latest follow-up (p < 0.0001), lower positioning of the glenoid component (p < 0.01) (Fig. 2), and superior tilting of the glenoid component (p < 0.01) on the immediate postoperative coronal-plane radiograph.
Subsidence of the glenoid component was observed in forty-one shoulders (7.9%) and was associated with use of a glenoid component with a convex back and with the surgical center, with subsidence observed exclusively at the centers that performed motorized reaming (p < 0.0001) (Fig. 3).
Posterior tilting of the glenoid component was observed in thirty-three shoulders (6.4%) and was associated with a preoperative glenoid shape of Type B1 or B2 (p < 0.01) and with reaming of the glenoid (p < 0.01) (Fig. 4).
Other outcomes were observed in ten shoulders (1.9%) and involved dislocation of the glenoid component (four shoulders), anterior tilting of the glenoid component (three shoulders), or a glenoid component migration pattern that could not be classified (three shoulders).
At a mean of 9.6 years of follow-up, Torchia et al.14 reported radiographic evidence of loosening in 65% of thirty-four shoulders treated with the Neer total shoulder arthroplasty prosthesis for primary osteoarthritis. The evidence of loosening was definite in nineteen shoulders and probable in three. Furthermore, a shift in the position of the glenoid component was observed in 38% of the shoulders.
At a minimum of five years of follow-up, we observed definite radiographic evidence of loosening in 166 (32.0%) of the 518 shoulders in the present study, which involved implantation of a cemented, keeled, polyethylene glenoid component in patients with primary osteoarthritis of the shoulder. Furthermore, 136 (26.3%) of the glenoid components had migrated. Critical evaluation of these cases revealed three major patterns of migration of the glenoid component and provided data that shed light on possible underlying mechanisms responsible for loosening.
Superior tilting was observed in 10% of the shoulders, and especially in shoulders in which the glenoid component had been implanted in a low position; this pattern involved gradual sinking of the superior aspect of the glenoid component into the bone and superior translation of the humeral head. A similar pattern of glenoid component loosening with superior tilting (described as the “rocking horse” phenomenon) and pathologic ascension of the humeral head has been described previously in patients with a massive rotator cuff tear18. However, we observed this complication in patients with primary glenohumeral osteoarthritis who did not have a massive rotator cuff tear at the time of surgery. Greater eccentric loading of the humeral head on the glenoid due to the superior translation of the humeral head required at the beginning of arm elevation may account for the biomechanical forces responsible for this loosening pattern19. Such loosening with superior tilting of the glenoid component suggests the potential importance of maintaining the integrity of the subchondral bone, especially in the superior quadrant of the glenoid. It also appears advisable to avoid implantation of the glenoid component with any superior tilt, which could increase the superior migration of the head and superior eccentric loading.
Subsidence, which was observed in 7.9% of the shoulders, occurred exclusively at the centers that performed reaming during glenoid preparation. The use of reaming to obtain a perfect fit with the back side of the glenoid component and to optimize version can result in sacrificing much of the subchondral or hard sclerotic glenoid bone. It has recently been suggested that reaming of the subchondral bone may cause structural weakness20. Although there were other potential confounding variables that could not be excluded in our study, our findings appear to suggest that preservation of subchondral bone is important to achieve the solid osseous support necessary to resist compressive forces against the glenoid implant.
Posterior tilting, which was observed in 6.4% of the shoulders, was correlated with preoperative static posterior subluxation of the humeral head associated with glenoid types B1 and B2. This result is consistent with a previous report of an increased rate of early glenoid loosening in association with preoperative static posterior subluxation21. Unbalanced soft-tissue tension following total shoulder arthroplasty can create an uneven distribution of forces across the glenoid component that contributes to failure through posterior tilting. The exact cause of the static posterior subluxation of the humeral head observed in some patients with primary osteoarthritis is unknown22. The limit beyond which consideration should be given to using bone-grafting of the posterior defect during the total shoulder arthroplasty, or even using a reverse shoulder prosthesis with or without bone-grafting, remains to be defined. We excluded shoulders requiring bone-grafting of the glenoid from the present series, and we therefore likely excluded a number of severely abnormal type-B2 glenoids. Furthermore, we did not consider the potential influence of humeral component version on the development of posterior tilting.
Each of the patterns of glenoid migration that we identified emphasizes the importance of osseous support. Cemented, all-polyethylene, keeled glenoid components do not always prevent glenoid migration, which may be related to compressive forces and posterosuperior eccentric loading, after total shoulder arthroplasty for primary glenohumeral osteoarthritis. On the basis of our analysis of the findings in this study, we believe that it is important to maximize the osseous support of the glenoid component, particularly in the superior and posterior aspects where the greatest stresses are experienced during arm elevation, in order to optimize the survivorship of a cemented all-polyethylene glenoid component in total shoulder arthroplasty. Although we have emphasized the importance of osseous support, it must be recognized that the causes of glenoid component failure in total shoulder arthroplasty are multifactorial12,23 and cannot be completely determined by this retrospective study design. Rather, multiple potential causes that require further study have been identified.
Our study has several weaknesses, including its multicenter design involving multiple surgeons who used a variety of techniques. It is possible that the influence of these variations in surgical technique was overriding and resulted in the influence of other important factors remaining hidden. Since an a priori power analysis was not performed, caution is required in interpreting the meaning of statistically nonsignificant findings since these findings may have been influenced by a lack of adequate power. It was also difficult to assess the influence that an unequal distribution of preoperative glenoid shapes among centers may have had, and only 72% of the shoulders underwent preoperative assessment with use of CT or CT arthrography. This study did not investigate the potential roles of bone mineral density and patient activity level. On the other hand, the greater number of cases made available by inclusion of multiple centers allowed for assessment of the influence of specific technical factors such as reaming.
There are multiple potential confounding variables that could not be accounted for in this study and may have influenced the results. We did not assess the influence of glenoid and humeral component version, glenoid component seating and fit, or cementing technique on the radiographic outcomes. Glenoid component designs with both convex and flat backs were included in this study to increase the number of cases of glenoid failure available for analysis; however, we recognize that this also introduced another potential limitation. Although we attempted to review all of the eligible total shoulder arthroplasties, a large number of patients were lost to follow-up. Preoperative imaging was not standardized in this study, and consequently not all patients underwent preoperative CT scanning for assessment of the glenoid morphology. Another weakness was the very large variation that we observed in the quality of the axillary views. We discovered that it was extremely difficult to standardize the technique for obtaining this particular view because, unlike the anteroposterior radiographs, the axillary radiograph could not be made under fluoroscopic control. We propose that further efforts be made to standardize radiographic assessment of the glenoid component, especially in the axial plane, in studies that report the results after total shoulder arthroplasty.
In summary, we observed that just over 26% of the cemented glenoid components had migrated at a mean of 103.6 months of follow-up in a series of 518 total shoulder arthroplasties performed for primary glenohumeral osteoarthritis. Three main patterns of glenoid component migration were observed: superior tilting, subsidence, and posterior tilting. Although the causes of glenoid component failure in total shoulder arthroplasty are multifactorial, we believe that preservation of glenoid bone stock, and in particular the sclerotic subchondral bone, is important in providing long-term resistance to the stresses imposed by the prosthetic humeral head.
Note: The authors acknowledge the following surgeons who contributed cases for this study: Richard Aswad, Luc Favard, Frank Gohlke, Christophe Levigne, Jean-Charles Le Huec, and Francois Sirveaux. The authors also acknowledge Angelique Delarue for her assistance with the statistical analysis.
Lazarus
MD;
Jensen
KL;
Southworth
C;
Matsen
FA
3rd. The radiographic evaluation of keeled and pegged glenoid component insertion. J Bone Joint Surg Am.
2002;84:1174-82.
Molé
D;
Roche
O;
Riand
N;
Levigne
C;
Walch
G. Cemented glenoid components: results in osteoarthritis and rheumatoid arthritis. : Walch
G;
Boileau
P, . Shoulder arthroplasty. Berlin: Springer-Verlag; 1999. p 163-71.
Amstutz
HC;
Thomas
BJ;
Kabo
JM;
Jinnah
RH;
Dorey
FJ. The Dana total shoulder arthroplasty. J Bone Joint Surg Am.
1988;70:1174-82.
Cofield
RH. Total shoulder arthroplasty with the Neer prosthesis. J Bone Joint Surg Am.
1984;66:899-906.
Figgie
MP;
Inglis
AE;
Figgie
HE
3rd;
Sobel
M;
Burstein
AH;
Kraay
MJ. Custom total shoulder arthroplasty in inflammatory arthritis. Preliminary results. J Arthroplasty.
1992;7:1-6.
Neer
CS
2nd;
Watson
KC;
Stanton
FJ. Recent experience in total shoulder replacement. J Bone Joint Surg Am.
1982;64:319-37.
Raiss
P;
Aldinger
PR;
Kasten
P;
Rickert
M;
Loew
M. Total shoulder replacement in young and middle-aged patients with glenohumeral osteoarthritis. J Bone Joint Surg Br.
2008;90:764-9.
Yian
EH;
Werner
CM;
Nyffeler
RW;
Pfirrmann
CW;
Ramappa
A;
Sukthankar
A;
Gerber
C. Radiographic and computed tomography analysis of cemented pegged polyethylene glenoid components in total shoulder replacement. J Bone Joint Surg Am.
2005;87:1928-36.
Szabo
I;
Buscayret
F;
Edwards
TB;
Nemoz
C;
O'Connor
DP;
Boileau
P;
Walch
G. Radiographic comparison of two glenoid preparation techniques in total shoulder arthroplasty. Clin Orthop Relat Res.
2005;431:104-10.
Throckmorton
TW;
Zarkadas
PC;
Sperling
JW;
Cofield
RH. Pegged versus keeled glenoid components in total shoulder arthroplasty. J Shoulder Elbow Surg.
2010;19:726-33.
Szabo
I;
Buscayret
F;
Edwards
TB;
Nemoz
C;
Boileau
P;
Walch
G. Radiographic comparison of flat-back and convex-back glenoid components in total shoulder arthroplasty. J Shoulder Elbow Surg.
2005;14:636-42.
Wirth
MA;
Rockwood
CA
Jr.. Complications of total shoulder-replacement arthroplasty. J Bone Joint Surg Am.
1996;78:603-16.
Boyd
AD
Jr;
Thomas
WH;
Scott
RD;
Sledge
CB;
Thornhill
TS. Total shoulder arthroplasty versus hemiarthroplasty. Indications for glenoid resurfacing. J Arthroplasty.
1990;5:329-36.
Torchia
ME;
Cofield
RH;
Settergren
CR. Total shoulder arthroplasty with the Neer prosthesis: long-term results. J Shoulder Elbow Surg.
1997;6:495-505.
Walch
G;
Badet
R;
Boulahia
A;
Khoury
A. Morphologic study of the glenoid in primary glenohumeral osteoarthritis. J Arthroplasty.
1999;14:756-60.
Walch
G;
Boileau
P. The Aequalis Shoulder System. : Walch
G;
Boileau
P, . Shoulder arthroplasty. Berlin: Springer; 1999. p 185-90.
Gazielly
DF. Comment je scelle un implant glénoïdien en polyéthylène convexe à quille avec ma technique de compaction d'os spongieux. Maitrise Orthopédique.
2007;162:6-11.
Franklin
JL;
Barrett
WP;
Jackins
SE;
Matsen
FA
3rd. Glenoid loosening in total shoulder arthroplasty. Association with rotator cuff deficiency. J Arthroplasty.
1988;3:39-46.
Karduna
AR;
Williams
GR;
Williams
JL;
Iannotti
JP. Kinematics of the glenohumeral joint: influences of muscle forces, ligamentous constraints, and articular geometry. J Orthop Res.
1996;14:986-93.
Haines
JF;
Trail
IA;
Nuttall
D;
Birch
A;
Barrow
A. The results of arthroplasty in osteoarthritis of the shoulder. J Bone Joint Surg Br.
2006;88:496-501.
Boileau
P;
Avidor
C;
Krishnan
SG;
Walch
G;
Kempf
JF;
Molé
D. Cemented polyethylene versus uncemented metal-backed glenoid components in total shoulder arthroplasty: a prospective, double-blind, randomized study. J Shoulder Elbow Surg.
2002;11:351-9.
Gerber
C;
Costouros
JG;
Sukthankar
A;
Fucentese
SF. Static posterior humeral head subluxation and total shoulder arthroplasty. J Shoulder Elbow Surg.
2009;18:505-10.
Bohsali
KI;
Wirth
MA;
Rockwood
CA
Jr.. Complications of total shoulder arthroplasty. J Bone Joint Surg Am.
2006;88:2279-92.
Disclosure: One or more of the authors received payments or services, either directly or indirectly (i.e., via his or her institution), from a third party in support of an aspect of this work. In addition, one or more of the authors, or his or her institution, has had a financial relationship, in the thirty-six months prior to submission of this work, with an entity in the biomedical arena that could be perceived to influence or have the potential to influence what is written in this work. No author has had any other relationships, or has engaged in any other activities, that could be perceived to influence or have the potential to influence what is written in this work. The complete Disclosures of Potential Conflicts of Interest submitted by authors are always provided with the online version of the article.