Abstract
Background:
Unconstrained shoulder arthroplasty is one of several methods for treatment of proximal humeral fracture nonunions. The goal of this study was to define the results and complications of this procedure.
Methods:
From 1976 to 2007, sixty-seven patients underwent unconstrained shoulder arthroplasty for proximal humeral nonunion and were followed for more than two years. There were forty-nine women and eighteen men with a mean age of sixty-four years and a mean duration of follow-up of nine years (range, two to thirty years). The fracture type according to the Neer classification was two-part in thirty-six patients, three-part in sixteen, and four-part in fifteen. Hemiarthroplasty was performed in fifty-four patients and total shoulder arthroplasty was done in the remaining thirteen.
Results:
There were thirty-three excellent or satisfactory results according to the modified Neer rating. Tuberosity healing about the prosthesis occurred in thirty-five shoulders. The mean pain score improved from 8.3 preoperatively to 4.1 at the time of follow-up (p < 0.001). The average active shoulder elevation and external rotation improved from 46° and 26° to 104° and 50° (p < 0.001). Shoulders with anatomic or nearly anatomic healing of the tuberosities had greater active elevation at the time of final follow-up (p = 0.02). There were fourteen complications in twelve patients, with twelve reoperations including five revisions. Kaplan-Meier survivorship with revision as the end point was 97% (95% confidence interval [CI]: 94.3, 100) at one year and 93% (95% CI: 88.0, 99.2) at five, ten, and twenty years.
Conclusions:
Shoulder arthroplasty decreases pain and improves function in patients with a proximal humeral nonunion. However, the overall results are satisfactory in less than half of the patients. Tuberosity healing is inconsistent and influences the functional outcome.
Level of Evidence:
Therapeutic Level IV. See Instructions for Authors for a complete description of levels of evidence.
Fractures of the proximal part of the humerus are common injuries, especially in the elderly1. The management of these fractures can be challenging, and treatment options include nonoperative modalities, osteosynthesis, and arthroplasty. Nonunion is a well-recognized complication of both nonoperative and operative treatment, and it typically results in substantial pain and functional impairment2,3.
These nonunions are often associated with bone loss, poor bone quality, and soft-tissue contractures4,5. Prior reports have highlighted the difficulty in achieving fracture union with attempts at internal fixation and bone-grafting2,5. As an alternative, surgeons have advocated shoulder arthroplasty for the treatment of proximal humeral nonunions4,6-8. However, high rates of unsatisfactory results, especially with regard to functional outcomes, have been reported with arthroplasty9.
The goal of this study was to review the results and complications of conventional anatomic shoulder arthroplasty for the treatment of proximal humeral nonunions and to identify factors associated with success or failure.
This study was approved by the institutional review board of our institution. The Department of Orthopedic Surgery joint registry database was searched for shoulder arthroplasty procedures performed for proximal humeral nonunion from 1976 to 2007. We included patients treated with either conventional anatomic hemiarthroplasty or total shoulder arthroplasty who had been followed for at least two years. A nonunion was defined as a fracture that had failed to unite or show radiographic signs of progressive callus formation or bridging bone at three months following the injury despite operative or nonoperative treatment. Arthroplasties performed less than three months after the fracture, patients with inadequate follow-up, and those treated with constrained or reverse shoulder arthroplasty were excluded.
From 1976 to 2007, eighty-nine shoulder arthroplasties were performed for proximal humeral nonunions. Thirteen patients who had died of reasons unrelated to the shoulder arthroplasty less than two years postoperatively, five patients who had been lost to follow-up, and four patients treated with custom constrained implants were excluded. The remaining group included sixty-seven patients with a mean duration of clinical follow-up of nine years (range, two to thirty years). There were forty-nine women and eighteen men with an average age of sixty-four years (range, thirty-one to eighty-five years). The fracture was the result of a low-energy fall in forty-three patients and the result of a higher-energy mechanism, including motor-vehicle accidents, falls from heights, and assaults, in the remaining twenty-four. The initial fracture type according to the Neer classification was two-part in thirty-six patients, three-part in sixteen, and four-part in fifteen, with thirteen shoulders sustaining a fracture-dislocation. Surgical fixation of the fracture was attempted in thirty-eight patients, and the remaining twenty-nine underwent initial nonoperative treatment. Additional surgical procedures were required in seven patients; these included revision of fixation in three, removal of metallic internal fixation in three, and rotator cuff repair in one. Thirty-eight patients had a two-part surgical neck fracture nonunion; one, a two-part greater tuberosity nonunion; nine, a three-part nonunion involving the greater tuberosity and surgical neck; nine, a four-part nonunion; and the remaining ten, a tuberosity malunion with a nonunion of the humeral articular surface. The indication for arthroplasty was pain associated with the fracture nonunion in sixty-four patients, and all patients reported limited function of the affected shoulder.
Surgical Technique
Shoulder arthroplasty was performed at a median of twelve months (range, three to 110 months) after the initial fracture, with thirty-two performed at less than one year, twenty-six done between one and two years, and the remaining nine carried out at greater than two years. The arthroplasties were performed by seventeen orthopaedic surgeons, with thirty-five performed by the senior author (R.H.C.) (Figs. 1-A and 1-B). A deltopectoral approach was used in thirty-five patients. Thirty-two patients required an anteromedial approach that involved detachment of the anterior aspect of the deltoid from the clavicle and anterior aspect of the acromion, which is useful for gaining exposure if there is scarring or contracture. The management of the subscapularis tendon was selected by the operating surgeon and included tenotomy in sixteen patients, release of the insertion from the lesser tuberosity in twenty-one, osteotomy of the lesser tuberosity in thirteen, and exposure through the tuberosity fracture in seventeen. Hemiarthroplasty was performed in fifty-four patients, and total shoulder arthroplasty was done in the remaining thirteen. The decision to perform total shoulder arthroplasty was made by the operating surgeon intraoperatively if there was evidence of glenoid arthritis. The implants used in this series included thirty-two Neer II (3M, St. Paul, Minnesota), twenty-eight Cofield (Smith & Nephew, Memphis, Tennessee), and seven Aequalis (Tornier, Edina, Minnesota) prostheses. In twelve patients with an isolated nonunion of the surgical neck, osteotomy of the articular surface was performed and the remaining bone was skewered with the prosthesis and then inserted distally to stabilize the fracture nonunion. Additional procedures at the time of arthroplasty were performed in sixty-three patients; these included tuberosity osteotomy (sixteen), rotator cuff repair (eleven), humeral bone-grafting (forty), glenoid bone-grafting (one), pectoralis major tendon lengthening (one), and allograft reconstruction of the rotator cuff (one). At the time of surgery, twelve patients had a rotator cuff tear; eleven of them underwent repair and one had a reconstruction with a tendon allograft. Bone-grafting was performed in forty patients, with use of local autograft bone in twenty-three, iliac crest autograft in seven, and allograft bone in ten.
Clinical and Radiographic Evaluation
Clinical notes, operative reports, and radiographs were reviewed. Additional follow-up was conducted with use of joint registry questionnaires10. Follow-up examinations were performed by the consultant or his staff. Patients who were unable to return for physical examination completed questionnaires and indicated their shoulder motion on diagrams. Clinical outcomes were measured with use of a 10-point scale for pain, a 10-point scale for satisfaction, and physical examination findings. A modified Neer result-rating system was used. With this system, the result is considered excellent if the patient has no or slight pain, has forward elevation of the arm of at least 140° and shoulder external rotation of at least 45°, and reports satisfaction with the procedure. The result is considered satisfactory if the patient has no more than moderate pain with vigorous activity; is able to forward elevate the arm and externally rotate the shoulder 90° and 20°, respectively; and reports satisfaction with the surgery. Those who did not meet these criteria were considered to have an unsatisfactory result.
Preoperative radiographs were available for sixty-two patients and fifty-six patients were followed radiographically for six months or more. The remaining eleven patients had less than six months of radiographic follow-up. Preoperative radiographs were assessed for fracture and nonunion type. Radiographs made immediately postoperatively and at the time of the most recent follow-up (mean, 5.4 years; range, six months to thirty years) were evaluated for the position of the humeral component and tuberosities as well as for evidence of loosening with consensus agreement of three authors. Radiographs were obtained with use of standard techniques for anteroposterior, axillary, and scapular Y views of the shoulder, with magnification not controlled. The component was classified as being in a varus or valgus position if there was >10° of angulation from the axis of the humeral shaft. The tuberosity position was classified as anatomic (5 mm below the superior aspect of the humeral head), proximal, or distal on the anteroposterior radiograph, with the degree of displacement classified as 5 to 10 mm or as >10 mm. The humeral component was classified as being at risk for loosening if there were lucent lines of ≥2 mm in three or more zones or if there was evidence of subsidence. The position of the humeral component in relation to the glenoid was assessed on anteroposterior and axillary radiographs, with <25% translation of the center of the humeral head from the center of the glenoid classified as mild subluxation, 25% to 50% classified as moderate, and >50% translation or dislocation classified as severe.
Statistical Analysis
Kaplan-Meier survivorship analysis was performed with revision or removal of the prosthesis as the end point. T tests were used to compare preoperative and postoperative outcomes for shoulder pain and motion. The significance of age, sex, initial fracture type, initial treatment of the fracture, timing of the arthroplasty, use of a cemented or uncemented humeral component, use of a hemiarthroplasty or total shoulder arthroplasty, the presence or absence of a rotator cuff tear, and tuberosity healing or nonunion on shoulder pain and motion was evaluated with t tests. The Fisher exact test was used to evaluate the Neer ratings (excellent and satisfactory vs. unsatisfactory). Significance was defined as a p value of <0.05.
Source of Funding
There were no outside sources of funding for this study.
Clinical Outcome
The mean pain score improved significantly from 8.3 before the shoulder arthroplasty to 4.1 at the time of the most recent postoperative follow-up (p < 0.001) (Fig. 2). The patients with severe pain at the time of follow-up all had reported severe pain preoperatively. Hemiarthroplasty had been performed in all six of these patients; five developed tuberosity nonunion and two had anterosuperior dislocation (Fig. 3). Complex regional pain syndrome was diagnosed postoperatively in two of the patients with severe pain. One patient with severe pain at the time of the most recent follow-up had done well for eight years but developed substantial pain and limited motion because of a late rotator cuff tear.
Patient satisfaction data were available for fifty patients, who had a mean satisfaction score of 5.7 on a 10-point scale ranging from 1 (poor) to 10 (excellent) at the time of the most recent follow-up (Fig. 4). The pain score was moderate or severe for six of the eight patients with a poor satisfaction score. Only one of the eight with a poor satisfaction score had anatomic healing of the tuberosities (three had a nonunion; three, resorption; and one, malunion), and two had anterosuperior dislocation.
Active elevation and external rotation averaged 46° (range, 0° to 130°) and 26° (range, −15° to 100°), respectively, preoperatively and improved to 104° (range, 0° to 180°) and 50° (range, −20° to 105°), respectively, at the time of the most recent follow-up (Figs. 5-A and 5-B). Active elevation was ≥90° in forty-six patients, and active external rotation was ≥20° in fifty-three patients. The median shoulder internal rotation at the time of the most recent follow-up was to the fifth lumbar vertebra and did not differ from the preoperative motion.
Thirty-three patients had an excellent or satisfactory result according to the Neer result rating (Fig. 6). The factors associated with an unsatisfactory Neer result or a reoperation in the remaining thirty-four patients included patient-reported dissatisfaction (twenty-seven), limited shoulder motion (twenty-three), and moderate or severe pain (nineteen).
Radiographic Outcomes
Initial postoperative radiographs showed eight humeral components to be in varus and three, in valgus. The tuberosity position was anatomic (<5 mm of displacement) in thirty-two patients and malpositioned (≥5 mm of displacement) in twenty-nine; in the remaining six, the tuberosities were resorbed or resected at the time of arthroplasty. The degree of malpositioning was 5 to 10 mm in twenty-one patients and >10 mm in eight. The malpositioning was distal in eighteen patients and proximal in eleven.
At the time of the final radiographic follow-up, seventeen tuberosities had healed in an anatomic position (<5 mm of displacement), eighteen were malunited (≥5 mm of displacement, with seven having proximal displacement and eleven having distal displacement), eighteen were not united (two had been followed for between three and six months and could be classified as delayed unions), and fourteen had resorbed or had been previously resected. Of the thirty-two patients with initial anatomic tuberosity fixation, seventeen had anatomic healing, eight had nonunion, three had malunion, and four had resorption at the time of the most recent follow-up.
The radiographic alignment of the humeral head in relation to the glenoid on the most recent anteroposterior and axillary radiographs was rated as anatomic or mild subluxation in fifty-five patients and as severe subluxation or dislocation in eleven; it was not classifiable in one. The most common direction of displacement was anterosuperior (seven), followed by anterior (three) and superior (one). The assessment of humeral stem fixation at the time of final follow-up revealed thirty-five stable components with no lucent lines, twenty-one stable components with lucent lines, three components with between 5 and 10 mm of subsidence as compared with the position on the initial postoperative radiograph, and one component that required revision for aseptic loosening; seven patients had radiographs that were not clear enough to evaluate.
Complications/Reoperations
There were fourteen complications in twelve patients, with twelve reoperations including five revisions. Anterosuperior instability developed in seven patients and was recognized within the first six months in five. The instability was associated with tuberosity nonunion (four), tuberosity resorption (one), or a rotator cuff tear (two). Repeat surgery for tuberosity nonunion was performed on two patients; one of them subsequently had anatomic healing and a satisfactory outcome whereas the other developed a fibrous union and had an unsatisfactory outcome. Rotator cuff repair was attempted in two patients, and it was augmented with a fascia autograft in one. Both patients had an unsatisfactory outcome with persistent instability, pain, and limited motion.
There were two deep infections; one was treated with component removal and staged reimplantation, and the other was managed with resection arthroplasty. Additional complications included a hematoma and wound necrosis; both were treated with surgical debridement and healing was achieved. There was one late periprosthetic fracture treated with internal fixation and bone-grafting.
Survivorship
The Kaplan-Meier survival free of revision was 97% (95% confidence interval [CI]: 94.3, 100) at one year and 93% (95% CI: 88.0, 99.2) at five, ten, and twenty years. There were five revisions, which were performed at a mean of thirteen months (range, three to thirty-one months) after the primary arthroplasty. The indications for revision included deep infection (two), instability (two), and aseptic humeral stem loosening three years after a hemiarthroplasty with cement (one). Both cases of instability were associated with tuberosity nonunion; one of these patients had a revision to a reverse total shoulder arthroplasty, and the other was treated with a resection arthroplasty at an outside facility. The Neer rating following the revision was unsatisfactory for all five patients.
Factors Affecting Outcome
Comparison of the patients who were sixty-five years old or younger with those who were older than sixty-five at the time of the arthroplasty did not reveal any differences between the groups in terms of the pain scores, active arm elevation, or Neer rating at the time of the most recent follow-up (see Appendix). There were also no differences based on sex. In addition, the outcomes of the two-part surgical neck nonunions without tuberosity involvement were not significantly different from the nonunions that involved the tuberosities. Comparison of the patients who underwent initial surgical management of the proximal humeral fracture with those who had nonoperative management also did not reveal any differences.
There were also no differences in outcome between the patients in whom the arthroplasty was performed at six months or less after the proximal humeral fracture and those in whom it was done more than six months after the fracture. Similarly, there was no difference in the pain score (p = 0.17), active elevation (102° and 101°), or the modified Neer rating between those treated less than one year and those treated more than one year after the proximal humeral fracture.
Comparison of the patients treated with a cemented and those treated with a press-fit humeral component revealed no differences in the outcomes at the time of the most recent follow-up. Comparison of the hemiarthroplasty and total shoulder arthroplasty groups also revealed no differences. There were no glenoid component failures after the total shoulder arthroplasties or revisions of any of the hemiarthroplasties because of the development of symptomatic glenoid arthritis.
The mean active arm elevation of patients with an intact rotator cuff on intraoperative inspection at the time of the arthroplasty was better than that in the patients with a torn rotator cuff (p = 0.04). However, there was no difference in the pain score or the Neer rating (p = 0.24) at the time of the most recent follow-up.
Patients with tuberosity healing with <10 mm of displacement had better active elevation at the time of the most recent follow-up than those with tuberosity nonunion (p = 0.02). However, there was no difference in the pain score or the Neer rating (p = 0.77) at the time of the most recent follow-up. There was no difference in the tuberosity healing rates between patients who had had bone-grafting at the time of arthroplasty and those who had not (55% versus 54%; p = 1.00). Tuberosity healing occurred in eight (67%) of twelve patients treated with the “shish kabob” technique compared with twenty-seven (49%) of fifty-five patients treated with other tuberosity fixation methods; this difference was not significant (p = 0.19). The arthroplasties performed from 1976 to 1999 resulted in pain scores, active elevation, and modified Neer ratings (p = 0.39) that were similar to those resulting from arthroplasties performed from 2000 to 2007.
In this series of sixty-seven proximal humeral nonunions treated with conventional anatomic shoulder arthroplasty, there was a significant reduction of pain and improvement in shoulder motion compared with the preoperative state, but more than half of the patients had an unsatisfactory result because of persistent pain or limited function of the extremity.
Shoulder arthroplasty for the treatment of complex proximal humeral fractures was described by Neer, who reported results that were “satisfactory but imperfect.”11 A number of subsequent reports substantiated his experience12-14. The treatment of proximal humeral nonunion, as compared with that of acute fractures, poses additional challenges due to bone loss, compromised bone quality, associated malunion, and soft-tissue contractures. The limited available evidence indicates that the results of late prosthetic placement are inferior to those of acute replacement for the treatment of proximal humeral fractures8,15-17.
The initial management of proximal humeral fractures has been implicated in the ultimate outcome of arthroplasty for the late sequelae of these fractures. It has been reported that the results in patients treated initially with surgical treatment are inferior to those in whom initial nonoperative management has failed8,18. In our series, there was no difference in the outcome of unconstrained arthroplasty based on the initial management of the fracture or the number of surgical procedures performed prior to the arthroplasty done to address the nonunion. However, this comparison did not address the issue of the status of the tuberosities.
In a recent series in which Boileau et al. evaluated the results of the management of the sequelae of proximal humeral fractures, the treatment of surgical neck nonunions with unconstrained arthroplasty had the worst functional outcomes18. The poor functional results were attributed to the inability to achieve healing of the tuberosity fragments. Our study included isolated nonunions of the surgical neck (thirty-eight) as well as those with associated tuberosity nonunion (nineteen) or malunion (ten). The importance of tuberosity healing with regard to the functional outcome was confirmed in our series. The ability to achieve anatomic healing of the tuberosity fragments continues to be a substantial challenge. Healing of the tuberosities was achieved in approximately half of our patients, and half of the tuberosities that healed did so with ≥5 mm of displacement.
In a prior report on twenty-seven patients from our institution by Antuña et al., the treatment of proximal humeral nonunions with arthroplasty resulted in a decrease in pain and high satisfaction9. However, there was a high rate of persistent functional impairment, with an average active elevation of only 88°. In an attempt to assess whether our ability to manage these complicated cases had improved, we compared the results of our cases performed from 2000 to 2007 with those carried out prior to that time. There was no improvement in pain, shoulder motion, satisfaction, tuberosity healing, or the modified Neer rating following arthroplasties performed after 2000 compared with those performed prior to that time. Given the high rate of tuberosity nonunion or resorption and the associated limited functional outcome, alternatives to conventional anatomic shoulder arthroplasty for proximal humeral nonunions should be considered. Open reduction and internal fixation with bone-grafting of proximal humeral nonunions has had variable results, but recent reports on intramedullary bone-grafting and rigid fixation have demonstrated union rates of >90%5,19. The current literature suggests that, if there is sufficient bone stock, open reduction and internal fixation with bone-grafting outperforms arthroplasty for the treatment of proximal humeral nonunions. The use of a reverse shoulder arthroplasty may result in better outcomes for some patients with proximal humeral nonunion. There are limited reports of the use of the reverse prosthesis for the management of acute fractures and late sequelae of fractures of the proximal part of the humerus20-23. Further investigation of the indications for and long-term outcomes of reverse shoulder arthroplasty for proximal humeral fractures is required.
Although this study represents the largest review of arthroplasty for the management of posttraumatic proximal humeral conditions, it has limitations. It is a retrospective study of a series spanning three decades with several different surgeons, prostheses, and surgical techniques employed during the study period. Follow-up of many of these patients was performed not only with clinical examination but also with joint registry questionnaires and telephone interviews as many of the patients did not live in close proximity to our institution. The heterogeneous nature of the patient population and treatments makes analysis of factors that influence outcome difficult, and, although there were no significant differences in outcome based on factors such as age, sex, type of fracture, and surgical techniques, a power analysis was not performed and therefore there is a risk of a type-II error. Multiple other outcome scores have been created since the Neer rating; however, it was the only rating collected across the entire cohort studied. The range of motion was assessed as a part of the study, but there was no true measure of the functional outcome.
The survivorship analysis did reveal excellent longevity of the implants free of revision. However, the relevance of this analysis is limited as twenty-four patients with unsatisfactory outcomes did not undergo revision. The reasons why revision was not performed despite an unsatisfactory outcome are not known, but they likely include both patient factors and a lack of a reliable surgical option to improve the outcome. Therefore, caution should be exercised when using survivorship analysis as a prognostic indicator of outcome in this series.
This review of proximal humeral fracture nonunions treated with anatomic conventional shoulder replacement showed a significant decrease in shoulder pain and improvement in motion. However, there was a >50% rate of unsatisfactory outcomes at the time of follow-up averaging nine years. The unsatisfactory outcomes were largely due to limited shoulder motion. The presence of a rotator cuff tear at the time of the arthroplasty and nonunion of the tuberosities at the time of follow-up were associated with poor results. No patient factors or surgical techniques were identified as positive prognostic indicators in this series. The role of conventional arthroplasty, open reduction and internal fixation, or reverse shoulder arthroplasty for the treatment of proximal humeral nonunions requires further investigation.
A table showing factors affecting outcome is available with the online version of this article as a data supplement at jbjs.org.
Palvanen
M;
Kannus
P;
Niemi
S;
Parkkari
J. Update in the epidemiology of proximal humeral fractures. Clin Orthop Relat Res.
2006;442:87-92.[PubMed][CrossRef]
Neer
CS
II. Nonunion of the surgical neck of the humerus. Orthop Trans.
1983;7:389.
Court-Brown
CM;
McQueen
MM. Nonunions of the proximal humerus: their prevalence and functional outcome. J Trauma.
2008;64(
6):1517-21.
Boileau
P;
Trojani
C;
Walch
G;
Krishnan
SG;
Romeo
A;
Sinnerton
R. Shoulder arthroplasty for the treatment of the sequelae of fractures of the proximal humerus. J Shoulder Elbow Surg.
2001;10(
4):299-308.
Walch
G;
Badet
R;
Nové-Josserand
L;
Levigne
C. Nonunions of the surgical neck of the humerus: surgical treatment with an intramedullary bone peg, internal fixation, and cancellous bone grafting. J Shoulder Elbow Surg.
1996;5(
3):161-8.
Dines
DM;
Warren
RF;
Altchek
DW;
Moeckel
B. Posttraumatic changes of the proximal humerus: malunion, nonunion, and osteonecrosis. Treatment with modular hemiarthroplasty or total shoulder arthroplasty. J Shoulder Elbow Surg.
1993;2:11-21.
Mansat
P;
Guity
MR;
Bellumore
Y;
Mansat
M. Shoulder arthroplasty for late sequelae of proximal humeral fractures. J Shoulder Elbow Surg.
2004;13(
3):305-12.
Norris
TR;
Green
A;
McGuigan
FX. Late prosthetic shoulder arthroplasty for displaced proximal humerus fractures. J Shoulder Elbow Surg.
1995;4(
4):271-80.
Antuña
SA;
Sperling
JW;
Sánchez-Sotelo
J;
Cofield
RH. Shoulder arthroplasty for proximal humeral nonunions. J Shoulder Elbow Surg.
2002;11(
2):114-21.
Smith
AM;
Barnes
SA;
Sperling
JW;
Farrell
CM;
Cummings
JD;
Cofield
RH. Patient and physician-assessed shoulder function after arthroplasty. J Bone Joint Surg Am.
2006;88(
3):508-13.
Neer
CS
2nd. Displaced proximal humeral fractures. II. Treatment of three-part and four-part displacement. J Bone Joint Surg Am.
1970;52(
6):1090-103.
Antuña
SA;
Sperling
JW;
Cofield
RH. Shoulder hemiarthroplasty for acute fractures of the proximal humerus: a minimum five-year follow-up. J Shoulder Elbow Surg.
2008;17(
2):202-9. .
Grönhagen
CM;
Abbaszadegan
H;
Révay
SA;
Adolphson
PY. Medium-term results after primary hemiarthroplasty for comminute proximal humerus fractures: a study of 46 patients followed up for an average of 4.4 years. J Shoulder Elbow Surg.
2007;16(
6):766-73. .
Mighell
MA;
Kolm
GP;
Collinge
CA;
Frankle
MA. Outcomes of hemiarthroplasty for fractures of the proximal humerus. J Shoulder Elbow Surg.
2003;12(
6):569-77.
Tanner
MW;
Cofield
RH. Prosthetic arthroplasty for fractures and fracture-dislocations of the proximal humerus. Clin Orthop Relat Res.
1983;(179):116-28.
Frich
LH;
Søjbjerg
JO;
Sneppen
O. Shoulder arthroplasty in complex acute and chronic proximal humeral fractures. Orthopedics.
1991;14(
9):949-54.
Bosch
U;
Skutek
M;
Fremerey
RW;
Tscherne
H. Outcome after primary and secondary hemiarthroplasty in elderly patients with fractures of the proximal humerus. J Shoulder Elbow Surg.
1998;7(
5):479-84.
Boileau
P;
Chuinard
C;
Le Huec
JC;
Walch
G;
Trojani
C. Proximal humerus fracture sequelae: impact of a new radiographic classification on arthroplasty. Clin Orthop Relat Res.
2006;442:121-30.
Cheung
EV;
Sperling
JW. Management of proximal humeral nonunions and malunions. Orthop Clin North Am.
2008;39(
4):475-82, .
Bufquin
T;
Hersan
A;
Hubert
L;
Massin
P. Reverse shoulder arthroplasty for the treatment of three-and four-part fractures of the proximal humerus in the elderly: a prospective review of 43 cases with a short-term follow-up. J Bone Joint Surg Br.
2007;89(
4):516-20.
Boileau
P;
Watkinson
D;
Hatzidakis
AM;
Hovorka
I. The Grammont reverse shoulder prosthesis: results in cuff tear arthritis, fracture sequelae, and revision arthroplasty. J Shoulder Elbow Surg.
2006;15(
5):527-40.
Martin
TG;
Iannotti
JP. Reverse total shoulder arthroplasty for acute fractures and failed management after proximal humeral fractures. Orthop Clin North Am.
2008;39(
4):451-7, .
Gallinet
D;
Clappaz
P;
Garbuio
P;
Tropet
Y;
Obert
L. Three or four parts complex proximal humerus fractures: hemiarthroplasty versus reverse prosthesis: a comparative study of 40 cases. Orthop Traumatol Surg Res.
2009;95(
1):48-55. .