Abstract
Background:
We reviewed a consecutive series of patients with a humeral fracture around either an anatomic or a reverse shoulder prosthesis treated with either open reduction and internal fixation (ORIF) or revision shoulder arthroplasty. The purposes of the study were to (1) describe the treatment of these fractures by either method, (2) report the outcomes, and (3) assess the validity of a current classification system.
Methods:
Indications for surgery were a displaced unstable fracture, a fracture around a loose humeral stem, or a patient who was unable to tolerate conservative treatment. Outcomes were reported for two groups (patients treated with revision arthroplasty and those treated only with ORIF) and included American Shoulder and Elbow Surgeons (ASES) scores, radiographic evidence of fracture union, and complications.
Results:
The mean ASES score for the entire cohort was 50.3 (95% confidence interval: 41.2 to 59.5). Thirty-five of the thirty-six fractures healed, in a mean of 7.2 months (range, 3.25 to 13.5 months). Complications occurred in fourteen (39%) of the thirty-six patients. Our ability to classify these fractures with a previously defined system had a low interobserver reliability (mean kappa, 0.37; range, 0.24 to 0.50) and a high intraobserver reliability (mean kappa, 0.69; range, 0.52 to 0.89).
Conclusions:
Periprosthetic fracture around a humeral stem implant is a difficult clinical problem involving complex decision-making. Fracture union occurred in 97% of our patients. Complications were frequent, and a reoperation was required in 19% of the patients. More than half of the patients in our study had a loose humeral component that required revision.
Level of Evidence:
Therapeutic Level IV. See Instructions for Authors for a complete description of levels of evidence.
Information regarding the treatment and outcomes of periprosthetic fractures of the humerus remains limited to a few small case series, some of which included fractures treated nonoperatively (see Appendix)1-8. No study has included fractures that occurred around a reverse-style shoulder prosthesis, to our knowledge. In 1995, Wright and Cofield6 proposed a classification system based on the location of the fracture in relation to the humeral stem tip. Their study was based on nine patients, only four of whom had undergone surgical treatment. The validity and reliability of this system to guide treatment and predict outcomes have not been established. The low prevalence of humeral shaft periprosthetic fractures has made validation of a classification system challenging. The prevalence of these fractures has been reported to be 0.6% in patients treated with anatomic shoulder arthroplasty3. Matsen warned of the potential for an increase in the risk of fracture and loosening with reverse-geometry designs, stating that, because the prosthesis is a fixed-fulcrum device, “forces applied to the humerus … may cause humeral or scapular fracture and prosthesis failure by fatigue.”9 We therefore undertook an investigation to review our experience with the surgical treatment of periprosthetic humeral fractures occurring around both anatomic and reverse shoulder prostheses. The purposes of this investigation were to (1) describe the surgical treatment of these fractures, including the approach and type of implant; (2) report the functional outcomes, radiographic findings, and complications; and (3) assess the validity of a current classification system.
Patients
Thirty-six patients with a previous shoulder arthroplasty who were treated surgically for a periprosthetic fracture of the humerus between 1998 and 2010 were identified from our electronic database; twenty-four were referred to our institution, and twelve had had the primary arthroplasty performed at our institution. Seventeen of the thirty-six patients sustained a fracture around a reverse-geometry implant. During this time frame the senior authors (M.F. and M.M.) performed more than 2600 shoulder arthroplasties (1286 reverse, 1236 total, and ninety-nine hemiarthroplasties).
The patients were divided into two groups. Group 1 included seventeen patients with a well-fixed prosthesis who were treated with open reduction and internal fixation (ORIF) (Figs. 1-A and 1-B). Group 2 consisted of nineteen patients with a loose prosthesis and/or bone loss who required revision shoulder arthroplasty, with or without allograft (Figs. 2-A through 3-B).
The indications for surgery were (1) a displaced unstable fracture (fifteen patients), (2) a fracture around a loose stem (fourteen), or (3) a patient who was unable to tolerate conservative treatment (seven). The decision whether to perform an operation was made by the treating surgeon. Seven patients underwent an initial attempt at nonoperative treatment, but as a result of continued pain or their dissatisfaction with the fracture brace they opted to undergo operative treatment. The average time from the fracture to surgical treatment was 216 days for these seven patients and sixty-three days (range, one to 709 days) overall. (See Appendix for patient demographics.)
Outcome Analysis
Inclusion for outcome analysis in Group 1 (seventeen patients) required at least six months of follow-up (twelve patients, followed for a mean of 24.8 months; range, six to eighty-four months) or fracture union seen on radiographs (three patients, followed for mean of 3.5 months; range, 1.5 to five months). Two patients were lost to follow-up. The primary clinical outcome for the remaining fifteen patients was the American Shoulder and Elbow Surgeons (ASES) scores10 at the time of latest follow-up. The secondary outcome was the shoulder range of motion, including active forward flexion, abduction, external rotation, and internal rotation. Range-of-motion measurements were performed via video analysis postoperatively by an independent observer for nine of the fifteen patients, whereas the treating physician performed the measurements clinically for the remaining patients. Additionally, six of the fifteen patients had the index procedure performed at our institution so prefracture ASES scores and shoulder motion measurements were available for them. We compared these results with the ASES scores at the time of final follow-up after the ORIF.
Inclusion for outcome analysis in Group 2 (nineteen patients) required a minimum of two years of follow-up (eleven patients, followed for a mean of 64.9 months; range, twenty-six to 152 months). The mean duration of follow-up for the remaining eight patients, who were not included in our primary outcome analysis, was 8.1 months (range, 2.5 to 13.5 months). One patient was followed for only 2.5 months and subsequently died. No other patients in this group were lost to follow-up. The latest follow-up ASES scores were reported as the primary outcome measurements for the eleven patients with at least two years of follow-up. The secondary outcome was the range of motion, including active forward flexion, abduction, external rotation, and internal rotation. The range-of-motion measurements were performed with postoperative video analysis by an independent observer for nine of the eleven patients and clinically by the treating physician for the other two. Preoperative ASES scores and range-of-motion values were obtained for five of the eleven patients in Group 2 and were compared with the final ASES scores and postoperative range of motion.
All thirty-six patients were included for the purpose of reporting perioperative complications (during the period of hospitalization) or postoperative complications as well as for describing the surgical treatment. A table in the Appendix summarizes the surgical indications, durations of follow-up, and which patients were included in the outcome analyses.
Radiographic Analysis
Preoperative Radiographic Evaluation
A surgeon blinded to patient names and treatment type conducted an independent review of the injury radiographs of thirty of the thirty-six patients (Table I). The degrees of fracture displacement and angulation were measured. Orthogonal radiographs were reviewed, and the maximum value on either projection was recorded. Displacement was classified as mild (less than one-third of the humeral shaft diameter), moderate (one-third to two-thirds of the humeral shaft diameter), or severe (more than two-thirds of the humeral shaft diameter). Angulation was classified as mild (≤15°), moderate (16° to 30°), or severe (>31°)3.
The degree of osteopenia was measured according to the system of Campbell et al.11, in which the combined width of the mid-diaphyseal cortices is divided by the total diameter of the midpart of the diaphysis. The radiographs used for interpretation included the available thirty injury radiographs, and four preinjury radiographs that were made at a mean of 23.6 days (range, three to fifty-six days) before the fracture. Two patients had no radiographs within one year before the injury and were excluded from this analysis. The bone quality was graded as normal if the ratio was >50%, mildly osteopenic if the ratio was 25% to 50%, and severely osteopenic if it was <25%11.
Finally, the thirty-four available injury or prefracture radiographs were evaluated for stem loosening according to the criteria described by Sperling et al.12 and Sanchez-Sotelo et al.13. In their description, either a cemented or an uncemented stem is considered to be at risk for loosening if ≥2 mm of lucency is seen in three or more humeral zones, or if there is subsidence or tilt (Table I). Radiographs were also evaluated for bone loss, osseous ectasia, osteolysis, or endosteal scalloping. Endosteal scalloping occurring in isolation was not considered to be a sign of loosening.
Postoperative Radiographic Evaluation
An independent observer reviewed each patient’s latest postoperative radiographic studies, obtained at the time of the latest clinical follow-up. Four views were used, including standard anteroposterior, Grashey anteroposterior, scapular-Y, and axillary views. The primary outcome was fracture union and allograft incorporation (when allograft had been used). Fracture union in both groups was defined as the presence of bridging bone or the disappearance of the fracture line on two orthogonal views, without evidence of implant failure2,3. Allograft incorporation was defined as the disappearance of the host bone-allograft interface on two orthogonal views. Clinical union was defined as a painless fracture site on examination3. The secondary outcome of interest was radiographic stem loosening12,13. Radiographs were also evaluated for implant breakage and dislocation.
Validation of Classification System
An attempt was made to classify the fractures according to the criteria of Wright and Cofield6. Injury radiographs were reviewed by three separate fellowship-trained shoulder and elbow surgeons, who were blinded to patient names and surgical treatment type. The observers classified each fracture twice, with one week separating each attempt, after a tutorial that explained that a type-A fracture begins around the humeral stem tip and extends proximally, a type-B fracture begins around the stem tip and extends distally, and a type-C fracture both begins and remains distal to the stem tip6. Interobserver and intraobserver agreements were then calculated14.
Surgical Treatment
Group 1: ORIF
In Group 1 (ORIF), nine of the seventeen patients were treated with a posterior surgical approach. The triceps muscle was elevated from lateral to medial, and the radial nerve was identified in the spiral groove. All of these patients had had a previous deltopectoral approach to place the original implant. In the remaining eight patients, who were treated with an anterior surgical approach, the original deltopectoral incision from the prior arthroplasty was opened and extended to an anterolateral humeral approach. The brachialis was split, and the radial nerve was not routinely identified between the brachialis and brachioradialis muscles. When cables were used, they were always passed from lateral to medial, to lessen the risk of radial nerve damage. Implant loosening was assessed by applying a direct force to the humeral stem tip after exposure of the fracture site. In one case, stability was unclear, so proximal humeral exposure with take-down of the subscapularis tendon and capsule was performed. The proximal bone-implant interface was then observed as direct pressure was applied to the stem tip through the fracture site. In all of these cases, the stem was stable and we proceeded with ORIF. Hybrid fixation was employed in eight of the seventeen patients in the ORIF group by using locking plates with cables placed at the level of the stem to improve stability. Allograft struts were selected intraoperatively for eight of the seventeen patients on the basis of the surgeon’s impression of the need to enhance plate fixation because of osteopenia. The struts bridged the fracture site and were fixed with cables. Screws were often angled in nonlocking mode just away from the stem to achieve transcortical purchase where the plate overlapped with the stem more proximally (six of the seventeen patients). (See Appendix for additional details of fixation.)
Group 2: Revision Arthroplasty
The nineteen patients in Group 2 underwent revision arthroplasty. The original deltopectoral approach was used and extended to the anterolateral humeral approach. Two patients had a combined anterior/posterior approach; the posterior approach was used to facilitate distal plate fixation (Fig. 2-B). All nineteen patients had a loose stem at the time of surgery. Revision to a long-stem prosthesis was performed in fourteen of the nineteen patients whereas five of the nineteen underwent revision to a short stem with an ORIF plate. Bulk allograft and cables were used in seventeen of the nineteen patients, and eleven of the seventeen had allograft struts. The decision to use allograft was determined intraoperatively. It was utilized in patients with substantive proximal humeral bone loss and a patulous soft-tissue envelope to increase prosthetic humeral stability and glenohumeral stability. A proximal humeral-anatomic allograft composite was used in six of the seventeen patients as previously described by Chacon et al.15.
Statistical Analysis
Interobserver and intraobserver reliability kappa coefficients were calculated on the basis of radiographic measurements performed by three surgeons on two occasions as discussed previously for the validation of the classification system16.
Source of Funding
There was no external funding for this study.
Primary Clinical Outcome Measures
Group 1: ORIF
Eight of the seventeen fractures treated with ORIF and retention of the original prosthesis had occurred around a reverse prosthesis, and nine of the seventeen had occurred around an anatomic prosthesis. The average ASES score at the time of latest follow-up was 45.1 (95% confidence interval [CI]: 26.8 to 63.5). All fractures healed, at an average of 6.8 months (range, 3.25 to twelve months); the average duration of follow-up was 21.0 months (range, four to eighty-four months). Allograft strut was used in eight patients. Of the six patients who had prefracture ASES scores available, five had a return to their prefracture score.
Group 2: Revision Arthroplasty
Among the nineteen fractures that were treated with revision arthroplasty (sixteen reverse and three anatomic), nine had occurred around a reverse prosthesis and ten had occurred around an anatomic prosthesis. The average time to union was 7.7 months (range, 3.5 to 13.5 months), with an average duration of follow-up of 64.9 months (range, twenty-six to 152 months). Allograft was used in seventeen of the nineteen patients. There was one nonunion, and eighteen of the nineteen fractures healed.
Preoperative ASES scores had been obtained for five of the eleven patients who had a minimum of two years of follow-up. These scores were obtained at a mean 6.6 days (range, three to ten days) prior to surgical revision. In all five cases, the ASES score and range of motion at the time of follow-up exceeded the prefracture ASES score and range of motion. The mean prefracture ASES score was 32.0 (95% CI: 14.0 to 50.0), and the mean postoperative ASES score was 54.4 (95% CI: 44.6 to 64.2).
Secondary Clinical Outcome Measures
The Appendix shows the ranges of motion in both groups.
Radiographic Outcome Measures
Preoperative Radiographic Evaluation
Seventeen of the thirty-four patients with available injury or prefracture radiographs had humeral stem loosening (Table I). In Group 1, only one of the seventeen patients appeared to have a loose stem, and that stem was found to be stable at surgery. In Group 2, sixteen of the seventeen patients appeared to have a loose stem.
Postoperative Radiographic Evaluation
Postoperative radiographic analysis was performed at the latest clinical follow-up visit for each patient, at a mean of 31.9 months (range, 1.5 to 152 months). Union occurred in thirty-five of the thirty-six patients. The average time to union was 7.2 months (range, 3.25 to 13.5 months) in the entire group, 6.8 months (range, 3.25 to twelve months) in the ORIF group, and 7.7 months (range, 3.5 to 13.5 months) for the patients with revision arthroplasty. The single nonunion occurred in the revision arthroplasty group. Allograft incorporation occurred in twenty-three patients (Table I).
Validation of Classification System
The attempts, by the three surgeons, to classify the injury radiographs (available for thirty of the thirty-six patients) according to the system of Wright and Cofield6 demonstrated poor interobserver reliability but good intraobserver reliability. The kappa coefficients for the interobserver reliability of the first and second fracture classification attempts were 0.24 and 0.50 (mean, 0.37), respectively. The mean kappa coefficient for intraobserver reliability was 0.69 (range, 0.52 to 0.89).
Complications
Complications occurred in fourteen of the thirty-six patients (Table II). Seven were in Group 1 (ORIF), and four of them required a reoperation. One patient who required a reoperation sustained a broken humeral stem (of an anatomic prosthesis) five years postoperatively, after low-energy trauma. This appeared to be unrelated to the original periprosthetic fracture treatment. The patient required revision to a reverse-geometry prosthesis. The original periprosthetic fracture remained healed. Another patient had failure of the glenosphere baseplate fixation and the humeral socket at thirteen months postoperatively. She required revision to a reverse-geometry prosthesis. The original periprosthetic fracture remained healed. A third patient sustained distal plate fixation failure two weeks postsurgery, requiring immediate extension of the fixation with an additional locking plate. She also had postoperative radial and musculocutaneous nerve palsies, both of which resolved by her latest follow-up visit (at 1.5 months). She died of other causes shortly after. The fourth patient had distal fracture-line extension observed at one month postoperatively, and she required immediate extension of the fixation with an additional locking plate. She had union six months later. The three patients who had complications that did not require a reoperation had radial nerve palsies, which all recovered by six months.
There were seven patients with complications in Group 2 (revision arthroplasty), three of whom required a reoperation. One sustained another periprosthetic fracture, which was treated nonoperatively. The new fracture site became infected and developed a nonunion that ultimately required a resection arthroplasty at two years. At the time of final follow-up, the patient was infection-free with minimal pain but poor function. Another patient underwent multiple failed ORIF procedures. A revision to a short-stem prosthesis with femoral strut allograft was eventually performed; however, the allograft was considered nonviable and was removed in a subsequent ORIF procedure. The nonunion persisted and, at the time of writing, was being treated nonoperatively. A third patient had a dissociation of the humeral socket-stem Morse taper one month postoperatively. She underwent immediate surgery (within two days) to reattach the humeral socket to the stem. The implant was stable at the time of her latest follow-up (at twenty-six months). Of the patients who had complications that did not require a reoperation, one had a radial nerve palsy, which resolved. One had a loose stem noted five years postoperatively. She declined additional surgery and was being observed at the time of writing. One patient had a postoperative hematoma, which was successfully managed conservatively, and one had a dislocation, which was treated nonoperatively because of underlying dementia.
As the population continues to age and shoulder arthroplasty becomes more common, there will likely be an increasing frequency of periprosthetic humeral fractures. This review of thirty-six patients who underwent surgical treatment of a periprosthetic fracture of the humerus is, to our knowledge, the first to include patients with a fracture around a reverse prosthesis as well as patients treated with revision to a reverse-style implant. Our study suggests that successful outcomes in terms of osseous healing and return of prefracture motion can be achieved in many patients with a periprosthetic humeral fracture. Thirty-five of thirty-six fractures healed, and shoulder motion at the time of the latest follow-up paralleled prefracture motion. The average ASES score at the latest follow-up examination was 50.3 (95% CI: 41.2 to 59.5), which is in the range of functional outcomes seen in prior series1-3,5-7.
In Group 1 (ORIF), one of the seventeen humeral stems had radiographic evidence of loosening, but all of the stems were stable at surgery. We were able to treat Group-1 patients through either an anterior or a posterior approach. The anterior approach can be considered for more proximal fractures and is beneficial because it utilizes the deltopectoral incision used at the original arthroplasty. It also allows a revision arthroplasty to be performed if there is uncertainty regarding stem stability. If there is still uncertainty intraoperatively, proximal exposure of the implant-bone interface can provide additional information. Disadvantages of the anterior approach include possible difficulty in identifying the radial nerve distally in a previously utilized surgical field as well as difficulty achieving adequate humeral exposure for more distal fractures. The posterior approach (triceps-reflecting or triceps-splitting approach17) can be considered for any fracture associated with a well-fixed stem, especially those near the stem tip or more distally. This approach is beneficial because it facilitates distal humeral exposure for plate fixation as well as protects and allows for easier identification of the radial nerve. The disadvantages include more difficulty with patient positioning (lateral decubitus versus prone) and the need for a secondary incision. Furthermore, if the stem is loose, the patient may require repositioning for a separate approach. Fixation was done with hybrid plates and additional cerclage cables, with allograft struts added (in eight of seventeen patients) at the surgeon’s discretion, depending on the quality of the bone. All patients with adequate follow-up (fourteen of fourteen) had healing of the fracture without evidence of stem loosening.
In Group 2 (revision arthroplasty), sixteen of seventeen patients with available radiographs had evidence of humeral stem loosening. However, all nineteen patients had intraoperative confirmation of loosening before they underwent humeral implant removal and revision arthroplasty. An extended deltopectoral approach was used in all nineteen patients, but an additional posterior approach was necessary in two of the nineteen to access a more distal shaft fracture. A long-stem cemented implant was used in all but five cases to bypass the fracture. The decision to use allograft was made much more frequently in Group 2 (seventeen of nineteen patients). Ten of eleven patients with adequate follow-up had a healed fracture. The majority were able to return to their preinjury level of function or better. One patient showed radiographic evidence of progressive stem loosening at five years.
The radiographic signs of loosening described by Sperling et al.12 and Sanchez-Sotelo et al.13 were extremely accurate in predicting whether the stem was loose at the time of surgery. There was only one false-positive and one false-negative determination of humeral stem loosening in the assessment of the thirty-four radiographs. We attempted to validate the most recognized classification system for periprosthetic humeral fractures (described by Wright and Cofield6) by reviewing injury radiographs, as previously described. Although the intraobserver agreement was substantial (mean kappa, 0.69), the interobserver agreement was poor (mean kappa, 0.37). The results of the reliability analysis indicate that, although an individual rater may consistently classify the same patients in the same way, there is frequent disagreement between raters. This disagreement can be the result of a lack of training in the proper use of the system, cases that are ambiguous or difficult to classify, or a classification system that is inadequate. A valid classification system should demonstrate both substantial intraobserver and substantial interobserver reliability and still be highly indicative of treatment type. We observed that stem loosening in conjunction with acute or chronic bone changes was a better determinant of the treatment selected than the aforementioned classification. The results of our study suggest that the location of the fracture in relation to the stem tip (A, B, or C) is not important in determining the surgical approach to the fracture.
The rate of preinjury radiographic changes, including bone loss and humeral stem loosening, may be underrecognized. In our study, seventeen of thirty-four patients had evidence of humeral loosening on either injury or preinjury radiographs. Therefore, patients who present with an acute periprosthetic injury on the humeral side may, in fact, be experiencing the sudden progression of a pathologic fracture. For this reason, our surgeons were prepared with more than traditional plating systems. It is important to take a careful history, assessing for prodromal symptoms of pain and stiffness, and to review the patient’s injury radiographs so that the surgeon can be prepared to treat a potentially ongoing process of bone loss and stem loosening.
Our study is limited by its retrospective nature and by the fact that the operations were performed by only two surgeons at a single institution (Florida Orthopaedic Institute). Also, ASES scores and range-of-motion values were not available for many patients both preoperatively and postoperatively. Additionally, the indications for surgery were not uniform, and the nature of the surgical treatment was not predetermined or standardized. The treatment approach was quite uniform, with ORIF reserved for well-fixed stems and revision arthroplasty reserved for loose stems. Also, there was no comparison group such as a nonoperative treatment group. Previous studies have included some patients who were successfully treated (achieved union) with conservative measures1-3,5-7.
In summary, our data indicate that periprosthetic fractures around a humeral implant were either acute traumatic fractures with a stable implant that required internal fixation or chronic fractures around a loose, unstable implant that required revision arthroplasty. Detailed preoperative planning with readily available revision components is vital for success in treating these complex fractures. Finally, the current classification scheme has a low interobserver reliability and may not reliably guide treatment. In this series, the union rate after surgical treatment was 97% and the complication rate was 39%.
Tables providing a review of the current literature on periprosthetic fractures of the humerus treated surgically as well as the surgical indications and the analyses performed in the present study, and text describing patient demographics, details of the fixation, and range-of-motion values are available with the online version of this article as a data supplement at jbjs.org.
Boyd
AD
Jr;
Thornhill
TS;
Barnes
CL. Fractures adjacent to humeral prostheses. J Bone Joint Surg Am.
1992;74(
10):1498-504.[PubMed]
Corrales
LA;
Morshed
S;
Bhandari
M;
Miclau
T
3rd. Variability in the assessment of fracture-healing in orthopaedic trauma studies. J Bone Joint Surg Am.
2008;90(
9):1862-8.[CrossRef][PubMed]
Kumar
S;
Sperling
JW;
Haidukewych
GH;
Cofield
RH. Periprosthetic humeral fractures after shoulder arthroplasty. J Bone Joint Surg Am.
2004;86(
4):680-9.[PubMed]
Tammachote
N;
Sperling
JW;
Vathana
T;
Cofield
RH;
Harmsen
WS;
Schleck
CD. Long-term results of cemented metal-backed glenoid components for osteoarthritis of the shoulder. J Bone Joint Surg Am.
2009;91(
1):160-6.[CrossRef][PubMed]
Worland
RL;
Kim
DY;
Arredondo
J. Periprosthetic humeral fractures: management and classification. J Shoulder Elbow Surg.
1999;8(
6):590-4.[CrossRef][PubMed]
Wright
TW;
Cofield
RH. Humeral fractures after shoulder arthroplasty. J Bone Joint Surg Am.
1995 Sep;77(
9):1340-6.
Wutzler
S;
Laurer
HL;
Huhnstock
S;
Geiger
EV;
Buehren
V;
Marzi
I. Periprosthetic humeral fractures after shoulder arthroplasty: operative management and functional outcome. Arch Orthop Trauma Surg.
2009;129(
2):237-43. .[CrossRef][PubMed]
Steinmann
SP;
Cheung
EV. Treatment of periprosthetic humerus fractures associated with shoulder arthroplasty. J Am Acad Orthop Surg.
2008;16(
4):199-207.[PubMed]
Matsen
FA. Commentary and perspective on “The reverse shoulder prosthesis for glenohumeral arthritis associated with severe rotator cuff deficiency. A minimum two-year follow-up study of sixty patients.” J Bone Joint Surg Am.
2005;87(
8):1697.
Richards
RR;
An
KN;
Bigliani
LU;
Friedman
RJ;
Gartsman
GM;
Gristina
AG;
Iannotti
JP;
Mow
VC;
Sidles
JA;
Zukerman
JD. A standardized method for the assessment of shoulder function. J Shoulder Elbow Surg.
1994;3(
6):347-52.[CrossRef][PubMed]
Campbell
JT;
Moore
RS;
Iannotti
JP;
Norris
TR;
Williams
GR. Periprosthetic humeral fractures: mechanisms of fracture and treatment options. J Shoulder Elbow Surg.
1998;7(
4):406-13.[CrossRef][PubMed]
Sperling
JW;
Cofield
RH;
O’Driscoll
SW;
Torchia
ME;
Rowland
CM. Radiographic assessment of ingrowth total shoulder arthroplasty. J Shoulder Elbow Surg.
2000;9(
6):507-13.[CrossRef][PubMed]
Sanchez-Sotelo
J;
O’Driscoll
SW;
Torchia
ME;
Cofield
RH;
Rowland
CM. Radiographic assessment of cemented humeral components in shoulder arthroplasty. J Shoulder Elbow Surg.
2001;10(
6):526-31.[CrossRef][PubMed]
Sidor
ML;
Zuckerman
JD;
Lyon
T;
Koval
K;
Cuomo
F;
Schoenberg
N. The Neer classification system for proximal humeral fractures. An assessment of interobserver reliability and intraobserver reproducibility. J Bone Joint Surg Am.
1993;75(
12):1745-50.[PubMed]
Chacon
A;
Virani
N;
Shannon
R;
Levy
JC;
Pupello
D;
Frankle
M. Revision arthroplasty with use of a reverse shoulder prosthesis-allograft composite. J Bone Joint Surg Am.
2009;91(
1):119-27.[CrossRef][PubMed]
Landis
JR;
Koch
GG. The measurement of observer agreement for categorical data. Biometrics.
1977;33(
1):159-74.[CrossRef][PubMed]
Gerwin
M;
Hotchkiss
RN;
Weiland
AJ. Alternative operative exposures of the posterior aspect of the humeral diaphysis with reference to the radial nerve. J Bone Joint Surg Am.
1996;78(
11):1690-5.[PubMed]
Disclosure: None 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 any aspect of this work. 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. Also, one or more of the authors has had another relationship, or has engaged in another activity, 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.