Patients who use their upper extremities for weight bearing (WB) place a significant stress on their rotator cuff and glenohumeral joint. These patients have been shown to have a significantly higher prevalence of rotator cuff tears compared with the rest of the population.1,2 The cumulative effect of wheelchair propulsion, use of the arms for transfers, and repetitive overhead activities from a seated position can result in degenerative cuff pathology and potential arthropathy of the shoulder. Shoulder pathology resulting in pain and dysfunction can markedly affect the quality of life of patients who rely on their upper extremities for WB and activities of daily living.3
Given their reliance on each upper extremity and unwillingness or inability to tolerate the protracted healing time and rehabilitation involved with a rotator cuff repair, these patients may tolerate years of progressive rotator cuff pathology in an effort to avoid surgery. Consequently, they may eventually present with severe rotator cuff deficiency and advanced arthritis of the glenohumeral joint, posing a formidable challenge to treating surgeons. The postoperative course restricting the use of the operative arm, which is typically relied on for daily activities and transfers, and the eventual stress these patients will place on the operative arm when cleared for use make this a unique patient population.
For upper extremity WB patients with advanced arthritis and rotator cuff deficiency, reverse shoulder arthroplasty (RSA) has been described as a treatment option.4 However, concerns exist about the long-term durability of this implant, given the demands placed on it by a patient who routinely bears weight using the shoulder. This retrospective review compared the mid-term outcomes of RSA in patients who used the shoulder for WB vs non–weight-bearing (NWB) patients. The authors' hypothesis was that RSA would produce similar favorable outcomes in these 2 groups of patients.
Materials and Methods
Inclusion Criteria and Patient Demographics
From September 2007 through March 2011, a total of 25 shoulders of 21 patients who used their upper extremity for WB (WB group) were treated with RSA because of degenerative rotator cuff deficiency and had a minimum of 5 years of postoperative follow-up (average, 73 months). Patients included in this study either were completely wheelchair bound due to lower extremity impairment or had lower extremity limitations that required them to use their arms for all transfers and loading via assistive aids for any attempt at ambulation. This study received institutional review board approval.
All of the patients included in this study had massive, nonrepairable rotator cuff tears, based on surgeon interpretation of magnetic resonance imaging findings. The patients included either had concomitant advanced degenerative changes of the glenohumeral joint in conjunction with their massive tear or had no glenohumeral arthritis but did have pseudoparalysis as the primary indication for arthroplasty. All patients included had undergone an extensive trial of nonoperative treatment with cortisone injections and physical therapy. Patients with previous attempts at an arthroplasty procedure were excluded from this study. Pertinent patient demographics are listed in Table 1.
Patient Demographics and Clinical Information
Seventy-five consecutive shoulders of 72 patients who were more traditional NWB patients (NWB group) treated with the same implant for the same indication during the same time period and matched for duration of follow-up (average, 72 months) served as a control group. As in the WB group, no patients with previous arthroplasty attempts were included in the NWB group. Pertinent patient demographics are included in Table 1.
Surgical Technique and Rehabilitation
All surgeries were done using a delto-pectoral approach. The implant used for each of these cases was the DJO Reverse Shoulder Prosthesis (DJO Surgical, Austin, Texas). The humeral component was cemented in 30° of retroversion for all cases. At the time of the study, the implant used was a modular humeral component designed for cement fixation on the humeral side. The glenosphere used in this implant incorporated some amount of lateral offset (between 6 and 10 mm, dependent on glenosphere selected). If viable subscapularis was present, could be mobilized, and could be repaired with low tension while the arm was in 30° of external rotation, it was repaired. In 3 (12%) of the WB group patients and 5 (7%) of the NWB group patients, a L'Episcopo transfer of the latissimus and teres major was done in the fashion described by Boileau et al5 in conjunction with the RSA because of severe preoperative external rotation lag signs.
The patients were placed in a shoulder immobilizer that was worn for 6 weeks postoperatively. No WB was allowed for the first 6 weeks postoperatively. On postoperative day 1, they began pendulum exercises and active elbow, wrist, and hand range of motion. At 6 weeks postoperatively, they began passive range of motion with forward elevation limited to 130° and 30° of external rotation. At 8 weeks postoperatively, the patients began passive range of motion to tolerance and active-assisted range of motion. Strengthening was started at 12 weeks postoperatively.
Nineteen of 21 patients in the WB group elected to be admitted to a rehabilitation facility and remained there for between 6 and 10 weeks postoperatively. Placement in a rehabilitation facility was recommended for the other 2 patients but refused. Patients in the WB group were allowed to begin transfers and to bear weight on the operative extremity at 6 weeks postoperatively.
Outcome Measures and Range of Motion Analysis
All patients were required to complete questionnaires to determine their American Shoulder and Elbow Surgeons (ASES) scores and Simple Shoulder Test (SST) scores. Patient-reported satisfaction was recorded at 1 year postoperatively and at all subsequent annual follow-up visits. Patients were asked to rate their outcome as satisfactory or not satisfactory.
Patients' range of motion was digitally recorded at 1 year postoperatively and at all subsequent annual follow-up visits. Patients performed a standardized range of motion examination evaluating forward elevation, external rotation, and internal rotation in a manner previously described.6 The digital camera was placed at shoulder height directly in line with the shoulder. Forward elevation was recorded in the sagittal plane with the patient oriented perpendicular to the camera. External rotation was recorded with the patient seated and the camera overhead so trunk rotation could be accounted for. Internal rotation was measured with the patient standing and the camera directly posterior to the patient. Patients were assigned a “yes” or a “no” in terms of their ability to completely internally rotate their arm and place the entire dorsum of their hand flat against their lumbar spine.
Three observers analyzed the patients' last available digital videos using a digital goniometer (Screen Protractor; Iconico, New York, New York) in a manner previously described.6 The 3 observers' measurements for each patient's range of motion were averaged and recorded.
Radiographic analysis was performed by an independent observer not involved with the surgical treatment of these patients. The radiographs were used to judge humeral-sided loosening or glenoid-sided loosening as well as scapular notching. Humeral loosening was measured using the grading system described by Sperling et al.7 In a manner previously described, glenosphere and baseplate fixation was graded as stable (no evidence of radiolucency around the baseplate bone interface or around any screw), at risk (>1 mm of circumferential radiolucency around the baseplate bone interface or around any screw), or loose (>1 mm of radiolucency around the baseplate and all screws or a shift in position of the implant).6 Scapular notching was measured using the grading system of Sirveaux et al.8
Intragroup comparisons of continuous data were performed with a paired t test. Between-group comparisons of continuous clinical outcomes and range of motion data were performed at baseline and at final follow-up with a Student's t test. Patient satisfaction and radiographic evidence of notching and humeral loosening at final follow-up were compared across the WB and NWB groups with a Fisher's exact test. All analyses were performed with SPSS version 22 (IBM Corp, Armonk, New York) statistical software. Significance was set at P=.05.
In the WB group, 16 patients were completely wheelchair bound due to lower extremity impairment, and 5 patients had lower extremity limitations that required them to use their arms for all transfers and loading via assistive aids for any attempt at ambulation. Thirteen had previous traumatic spinal cord injury, 3 had a history of transverse myelitis, and 5 had a history of post-polio syndrome. Most of the patients in the WB group (20 of 25 shoulders) had concomitant advanced degenerative changes of the glenohumeral joint. Five of the 25 shoulders did not have arthritis but rather had pseudoparalysis as the primary indication for arthroplasty. In the NWB group, in conjunction with their massive tear, 64 of 75 patients had concomitant advanced degenerative changes of the glenohumeral joint. Eleven of the 75 patients did not have arthritis but rather had pseudoparalysis as the primary indication for surgery.
Table 2 lists the results for both the WB group and the NWB group. Regarding outcome scores, the preoperative to postoperative ASES scores and SST scores were not significantly different between the WB group (34 to 78 [P<.001] and 2.4 to 8.0 [P<.001], respectively) and the NWB group (34 to 77 [P<.001] and 2.9 to 8.0 [P<.001], respectively). Regarding range of motion, the preoperative to postoperative forward elevation, external rotation, and internal rotation were not significantly different between the WB group (86° to 141° [P<.001], 22° to 39° [P<.001], and 76% full internal rotation to 76% full internal rotation [P=1.0], respectively) and the NWB group (88° to 138° [P<.001], 24° to 36° [P<.001], and 80% full internal rotation to 73% full internal rotation [P=.440], respectively).
Results for the 2 Groups of Patients
At final follow-up, 92% of patients in the WB group vs 94% in the NWB group were satisfied with their outcome. Radio-graphic analysis revealed that the scapular notching rate was higher in the WB group compared with the NWB group (20% vs 5%; P=.041). In the WB group, 4 patients had grade 1 notching and 1 patient had grade 2 notching, according to the grading system of Sirveaux et al.8 In the NWB group, each patient had grade 1 notching. The incidence of radiographic humeral loosening was also higher in the WB group than in the NWB group (12% vs 0%; P=.014).
The complication rate was higher in the WB group than in the NWB group (12% vs 4%). In the WB group, 1 patient had a nondisplaced type 2 acromial fracture, based on the classification described by Levy et al.9 This occurred 11 weeks postoperatively, was treated nonoperatively with 6 additional weeks of shoulder immobilization, and went on to heal with no further issues. Two other patients in the WB group fell during attempted transfers, sustaining a dislocation at weeks 4 and 5, respectively. They were the only patients in the WB group who refused to go to a rehabilitation facility after surgery and were discharged home from the hospital. Both ended up requiring an open reduction and had no further instability events or complications. In the NWB group, all 3 patients with complications had nondisplaced type 2 acromial stress fractures. These occurred at 12, 14, and 16 weeks, respectively. They were treated with 6 weeks of shoulder immobilization and went on to heal with no further complications. The underlying causes of the acromial fractures were unknown. There were no complications in the 8 patients who underwent a L'Episcopo transfer. No glenoid-sided baseplate failures were observed in the WB group or the NWB group.
Patients who rely on their upper extremity for WB with daily activities and present with severe rotator cuff deficiency in conjunction with advanced arthritis of the glenohumeral joint are a complex population to manage surgically. The authors hypothesized that RSA would be a viable option in upper extremity WB patients and could produce favorable outcomes similar to those of traditional NWB patients. The results favored this hypothesis, as the upper extremity WB patients had outcome scores, postoperative range of motion values, and patient satisfaction similar to those of the NWB patients.
A review of the literature reveals that there is limited information regarding the use of arthroplasty for the management of end-stage degenerative disease in patients with a history of paraplegia or neuromuscular conditions that cause them to rely on their upper extremities for transfers, propulsion, and activities of daily living. Garreau et al10 reported on 5 paraplegic patients undergoing standard total shoulder arthroplasty, noting improvements in their Constant and ASES scores with reasonable satisfaction at a mean follow-up of 30 months. Hattrup and Cofield11 reported on 6 patients (1 hemiarthroplasty, 5 total shoulder arthroplasties) who were followed for an average of 80 months, noting excellent results in 1 patient, satisfactory results in 4 patients, and unsatisfactory results in 1 patient. The authors did note a high complication rate in this cohort of patients.
A study by Kemp et al4 most closely mirrors the current study in that the authors used RSA to treat 19 shoulders in 16 patients who were wheelchair dependent. There were 3 early failures (15%) with a dislodged baseplate after a fall, 2 patients with dislocations who were not revised, and an overall complication rate of 25% in that study. There was some loss to follow-up, with 12 shoulders being available for full inclusion with average follow-up of 40 months, and the authors noted improvements in outcome scores and range of motion. The authors did note that the acute postoperative management of these patients can be difficult, given the limitations that they face with arm immobilization to protect the shoulder as they recover from surgery.
Compared with the study by Kemp et al,4 in the current study, the authors examined more patients with a longer follow-up (average, 73 months) and compared them with a control group of patients who did not use their upper extremity for WB and were matched for duration of follow-up using the same implant and surgical technique. It was promising that the outcome scores, range of motion, and satisfaction scores were similar between the 2 groups; however, one of the authors' primary concerns about the patients in the WB group was possible implant-related failures, as it was unknown if the RSA could tolerate daily WB activities during a long period.
To date, there have been no glenoid-sided or humeral-sided implant-related failures requiring revision surgery in this cohort. The implant used had a lateral center of rotation glenosphere with a center screw baseplate and peripheral locking screws. Compared with the more medialized Grammont design, this lateral center of rotation potentially increases the stresses seen on the glenoid side; however, to date, the implants used in this study continue to tolerate these stresses. The authors did clear all of these patients for normal use of the arm for WB and their activities of daily living at 6 weeks postoperatively, which the authors thought was somewhat aggressive. The authors did so knowing that the glenoid baseplate would not be fully ingrown at that point, but they found the patients typically unwilling to spend more than 6 to 7 weeks in a rehabilitation facility. The authors decided to clear them for use at that point in effort to restore their independence. Despite this aggressive approach, the authors saw no early glenoid failure.
The complication rate was higher in the WB group than in the NWB group (12% vs 4%). In the WB group, 2 of the complications were due to traumatic dislocations, resulted in revision surgery, and occurred as a result of falls during attempted transfers during the first 6 weeks postoperatively. These were the only 2 patients in the WB group who refused admission to an inpatient rehabilitation facility for at least the first 6 weeks postoperatively, instead attempting to return home from the hospital with family assistance. This points to some of the difficulties these patients can have while the arm is immobilized acutely after surgery. Because of these difficulties, the authors currently do not offer patients surgery unless they agree to spend at least the first 6 weeks postoperatively in an inpatient rehabilitation facility.
Some concerns exist regarding the radiographic findings in the WB group. Despite the same implant being used in the 2 groups, there was a significantly higher scapular notching rate in the WB group than the NWB group (20% vs 5%). The lateral center of rotation glenosphere and 135° neck shaft angle humeral component used in this study are designed to minimize scapular notching and the adduction deficit as described by Gutiérrez et al.12 It is possible that, while transferring, the patients in the WB group may be loading the arm in full adduction as they push off their elbows. The cumulative effect of this over time could lead to the notching rates seen in the WB group compared with the NWB group. On the humeral side, 12% of patients in the WB group had signs of humeral loosening compared with 0% in the NWB group. A modular cemented humeral component was used in this study. Reverse shoulder arthroplasty involves a constrained implant, and many of the forces of the articulation of the implant are transferred to the humeral side. Among the WB patients, their repetitive use of their arms for loading and repetitive activities of daily living may have contributed to this humeral loosening. Although the patients with scapular notching and humeral loosening had no clinical complaints, the authors continue to follow their radiographs closely, as long-term survivorship of this implant is a concern in this patient population.
The main weakness of this study stems from its retrospective design. The addition of a control group matched for duration of follow-up and implant does add some strength. However, the NWB group is not truly a totally matched cohort. The WB group was, on average, 4 years younger than the NWB group, and 20% of the WB group had an attempted previous cuff surgery vs 39% of the NWB group. Also, 76% of the patients were men and 24% were women in the WB group compared with 48% and 52%, respectively, in the NWB group, further indicating that it was not truly a matched cohort. Finally, all of the patients in the current study received a cemented RSA. Therefore, the results may not be applicable to a press-fit RSA. A study comparing cemented vs uncemented stems in upper extremity WB RSA patients may be worthwhile.
Patients who used their shoulder for WB had outcome scores, range of motion improvements, and patient satisfaction after RSA similar to those of NWB patients. At mid-term follow-up, WB patients had a significantly higher scapular notching rate, a higher humeral loosening rate, and a higher complication rate compared with NWB patients. Because the radiographic findings raised some concerns, long-term follow-up is necessary to monitor implant survivorship in the WB patients. Although upper extremity WB patients require significant assistance in the early postoperative period, RSA appears to be an effective treatment option for them.
- Akbar M, Balean G, Brunner M, et al. Prevalence of rotator cuff tear in paraplegic patients compared with controls. J Bone Joint Surg Am. 2010; 92(1):23–30. doi:10.2106/JBJS.H.01373 [CrossRef]
- Escobedo EM, Hunter JC, Hollister MC, Patten RM, Goldstein B. MR imaging of rotator cuff tears in individuals with paraplegia. AJR Am J Roentgenol. 1997; 168(4):919–923. doi:10.2214/ajr.168.4.9124140 [CrossRef]
- Wang JC, Chan RC, Tsai YA, et al. The influence of shoulder pain on functional limitation, perceived health, and depressive mood in patients with traumatic paraplegia. J Spinal Cord Med. 2015; 38(5):587–592. doi:10.1179/2045772314Y.0000000271 [CrossRef]
- Kemp AL, King JJ, Farmer KW, Wright TW. Reverse total shoulder arthroplasty in wheel-chair-dependent patients. J Shoulder Elbow Surg. 2016; 25(7):1138–1145. doi:10.1016/j.jse.2015.11.006 [CrossRef]
- Boileau P, Rumian AP, Zumstein MA. Reversed shoulder arthroplasty with modified L'Episcopo for combined loss of active elevation and external rotation. J Shoulder Elbow Surg. 2010; 19(2)(suppl):20–30. doi:10.1016/j.jse.2009.12.011 [CrossRef]
- Cuff D, Pupello D, Virani N, Levy J, Frankle M. Reverse shoulder arthroplasty for the treatment of rotator cuff deficiency. J Bone Joint Surg Am. 2008; 90(6):1244–1251. doi:10.2106/JBJS.G.00775 [CrossRef]
- 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):507–513. doi:10.1067/mse.2000.109384 [CrossRef]
- Sirveaux F, Favard L, Oudet D, Huquet D, Walch G, Molé D. Grammont inverted total shoulder arthroplasty in the treatment of glenohumeral osteoarthritis with massive rupture of the cuff: results of a multicentre study of 80 shoulders. J Bone Joint Surg Br. 2004; 86(3):388–395. doi:10.1302/0301-620X.86B3.14024 [CrossRef]
- Levy JC, Anderson C, Samson A. Classification of postoperative acromial fractures following reverse shoulder arthroplasty. J Bone Joint Surg Am. 2013; 95(15):e104. doi:10.2106/JBJS.K.01516 [CrossRef]
- Garreau De Loubresse C, Norton MR, Piriou P, Walch G. Replacement arthroplasty in the weight-bearing shoulder of paraplegic patients. J Shoulder Elbow Surg. 2004; 13(4):369–372. doi:10.1016/j.jse.2004.01.019 [CrossRef]
- Hattrup SJ, Cofield RH. Shoulder arthroplasty in the paraplegic patient. J Shoulder Elbow Surg. 2010; 19(3):434–438. doi:10.1016/j.jse.2009.07.003 [CrossRef]
- Gutiérrez S, Comiskey CA IV, Luo ZP, Pupello DR, Frankle MA. Range of impingement-free abduction and adduction deficit after reverse shoulder arthroplasty: hierarchy of surgical and implant-design-related factors. J Bone Joint Surg Am. 2008; 90(12):2606–2615. doi:10.2106/JBJS.H.00012 [CrossRef]
Patient Demographics and Clinical Information
|Characteristic||WB Group||NWB Group||P|
|Patients/shoulders (male/female), No.||21/25 (19/6)||72/75 (36/39)||<.001a|
|Age, average (range), y||68 (58–75)||72 (62–81)||<.001|
|Follow-up, average (range), mo||73 (62–98)||72 (63–72)||.474|
|Portion of cohort with prior cuff surgery||20%||39%||.142|
|Portion of cohort with diabetes mellitus||20%||24%||1.00|
|Portion of cohort smokers||20%||13%||.538|
|Portion of cohort with advanced OA + massive cuff tear||80%||85%||.538|
|Portion of cohort with no OA + massive cuff tear||20%||15%||.538|
Results for the 2 Groups of Patients
|Parameter||WB Group||NWB Group||P|
|Preoperative ASES score, average (range)||34 (26–45)||34 (24–42)||.765|
|Postoperative ASES score, average (range)||78 (70–84)||77 (60–80)||.094|
|Preoperative SST score, average (range)||2.4 (1–4)||2.9 (1–6)||.059|
|Postoperative SST score, average (range)||8.0 (5–10)||8.0 (6–11)||1.00|
|Preoperative forward elevation, average (range)||86° (36°–124°)||88° (31°–154°)||.691|
|Postoperative forward elevation, average (range)||141° (104°–170°)||138° (104°–160°)||.152|
|Preoperative external rotation, average (range)||22° (0°–46°)||24° (4°–50°)||.481|
|Postoperative external rotation, average (range)||39° (30°–50°)||36° (3°–48°)||.029|
|Patients with full internal rotation preoperatively||76%||80%||.778|
|Patients with full internal rotation postoperatively||76%||73%||1.00|
|Patients with scapular notching||20%||5%||.041|
|Patients with humeral loosening||12%||0%||.014|