This annual update on shoulder and elbow surgery is a review of the most relevant studies from July 2010 through June 2011. It includes clinical and basic-science articles from The Journal of Bone and Joint Surgery (American Volume), The Journal of Bone and Joint Surgery (British Volume), the Journal of Shoulder and Elbow Surgery, and Arthroscopy: The Journal of Arthroscopic and Related Surgery. Relevant level-I and level-II studies from other medical journals are included where appropriate.
The level of evidence is indicated at the end of each review when it is known. Particular attention should be paid to the level-I and level-II studies as they represent randomized controlled studies. Additional level-III and level-IV studies representing important topics in shoulder and elbow surgery are also included in the review.
General
Wiater and colleagues1 performed a review of 375 arthroscopic shoulder procedures performed in a community orthopaedic practice to determine the factors that might be associated with chondrolysis. The authors reported that the prevalence of postoperative chondrolysis was 13% (forty-nine of 375). Each case of chondrolysis was associated with the intra-articular post-arthroscopic infusion of a local anesthetic (either Marcaine [bupivacaine] or lidocaine). Subgroup analysis of patients who received an intra-articular infusion of a local anesthetic demonstrated that the risk of chondrolysis was greater in younger patients and in those with one or more suture anchors placed in the glenoid. The authors concluded that the risk of chondrolysis could be reduced by avoiding postoperative intra-articular injection of local anesthetic. (Level II)
Rotator Cuff
Pain Management
Early postoperative pain control following arthroscopic rotator cuff repair can be challenging. Two randomized, prospective studies investigated postoperative pain management following arthroscopic rotator cuff repair. DeMarco et al.2 performed a prospective, randomized, controlled study of the benefit of an interscalene block combined with a subacromial catheter infusing bupivacaine compared with a subacromial catheter infusing bupivacaine alone. All patients had the subacromial catheter placed arthroscopically. Bupivacaine was administered via an infusion pump catheter for seventy-two hours, with oral narcotics as needed. Visual analog scale (VAS) pain scores were significantly better in the block group, and narcotic requirements were less in the block group when recorded six hours after discharge. However, by twelve hours and at all time points up to eighty hours, pain scores were statistically the same in both treatment groups. Interestingly, the authors observed a rebound spike in pain scores in the patients in the block group at twenty hours. (Level I)
Bang et al.3 evaluated the effect of low-dose gabapentin on postoperative pain in patients undergoing arthroscopic rotator cuff repair. The patients were randomized to one of two treatment groups. One group of patients received 300 mg of gabapentin two hours prior to surgery, and the other group of patients received a placebo. The primary outcome measure was the VAS pain score, and the secondary outcome measures were fentanyl consumption and side effects. The patients who received gabapentin had significantly lower VAS scores than did the controls at two, six, and twelve hours after surgery, but there was no difference at twenty-four hours. There was also no difference between the two groups in terms of early side effects or fentanyl consumption. (Level I, randomized controlled trial)
Basic Science
Fatty atrophy of the rotator cuff can be clinically associated with chronic tears and denervation of the rotator cuff. Rowshan and colleagues4 studied the effect of subscapularis tendon tear and transection of the subscapular nerve on muscle atrophy and fat content of the subscapularis muscle in a rabbit rotator cuff model. The animals were divided into three groups: (1) partial rotator cuff tear without retraction of the muscle, (2) complete rotator cuff tear with retraction of the muscle, and (3) nerve transection of the subscapular nerve. The authors demonstrated that either subscapularis tenotomy or subscapular nerve transection yielded decreased mass and increased fat content (identified histologically) at six weeks after the injury as compared with the findings in a partial tear control group. Interestingly, in animals with complete tenotomy, the subscapular nerve showed gross evidence of neural injury at two weeks, suggesting that chronic tear may yield neurologic changes as well. These data suggest that the muscular changes associated with complete tenotomy are comparable with those seen with denervation of the muscle and suggest that chronic rotator cuff tears may induce an associated neurologic injury.
Patients with diabetes mellitus are thought to have an impaired ability to heal following rotator cuff repair. The effects of elevated blood sugar on tendon healing were recently investigated by Bedi et al5. Forty-eight male Lewis rats that underwent supraspinatus tendon release and immediate transosseous repair were divided into two groups. Half served as a control, and the other half had diabetes mellitus induced by an injection of streptozotocin. The animals were killed and analyzed at one and two weeks after surgery with biomechanical testing as well as histomorphometric and immunohistochemical analysis. Histological examination of the tendons revealed less fibrocartilage and organization in the experimental group at both time points. The load to failure was significantly decreased in the diabetes mellitus group at all time periods. These findings support the clinical observation of impaired healing for patients with diabetes mellitus who undergo rotator cuff repair. Because the analysis was not performed beyond two weeks, the authors were unable to make definitive comments about the effect of hyperglycemia at the time of ultimate healing.
Full-Thickness Tears
The factors that contribute to the progression of an asymptomatic rotator cuff tear to a symptomatic tear are poorly understood. Mall et al.6 prospectively monitored 195 patients with asymptomatic rotator cuff tears for the development of pain and for changes in various parameters such as tear size, fatty degeneration of the rotator cuff muscle on magnetic resonance imaging (MRI), glenohumeral kinematics, and shoulder function. Forty-four patients (23%) developed a new onset of shoulder pain. When compared with a similar subset of fifty-five patients who remained asymptomatic, pain was associated with an increase in tear size and/or progression from a partial tear to a complete tear. Compared with assessment when the patient was asymptomatic, the American Shoulder and Elbow Surgeons (ASES) scores for shoulder function were significantly decreased and all measures of shoulder motion were decreased except for external rotation at 90° of abduction. There was an increase in compensatory scapulothoracic motion in relation to glenohumeral motion during early shoulder abduction with pain development. Furthermore, it was observed that larger tears at the time of the initial evaluation were more likely to be associated with the development of pain than smaller tears. (Level III)
The type of suture anchor utilized and the suture anchor configuration remain areas of research interest and were the subject of two studies on the technical aspects of arthroscopic rotator cuff repair. Milano et al.7 performed a prospective randomized study of 110 patients (with fifty-five patients per group) to analyze the effect of metal versus biodegradable sutures anchors on the outcome of arthroscopic rotator cuff repair. The results were evaluated at two years of follow-up with respect to several variables, including baseline scores; age; sex; arm dominance; tear location, configuration, and retraction; fatty degeneration on MRI; treatment of biceps tendon; rotator cuff repair technique (anchors or anchors and side-to-side repair); and number of anchors used. The authors found no significant differences with regard to any of the variables that were considered. Final outcomes measured on the basis of the Disabilities of the Arm, Shoulder and Hand (DASH), Work-DASH, and Constant scores at two years of follow-up showed no significant difference. Only the baseline score, age, tear location, and fatty degeneration significantly and independently influenced the outcome. A major weakness of this study was that no postoperative imaging studies were performed. Many studies have showed that the outcome of rotator cuff repair can be satisfactory in the absence of rotator cuff healing. One of the concerns about biodegradable anchors is that they may adversely affect rotator cuff healing rates. Unfortunately, this study did not answer this important question. (Level I)
Koh et al.8 explored the effect of suture anchor configuration during arthroscopic rotator cuff repair on rotator cuff healing in a prospective study of seventy-one patients with rotator cuff tears measuring 2 to 4 cm who were randomized intraoperatively to either a three-anchor single-row repair or a four-anchor double-row repair. Of the seventy-one patients enrolled, sixty-two (thirty-one in each group) were available for follow-up evaluation. Retear rates and clinical scores were also analyzed for forty-seven patients who underwent a follow-up MRI evaluation. Preoperative clinical scores were significantly improved postoperatively in both groups. No significant differences in pain, clinical outcome scores, or retear rates were noted between the two groups. (Level I)
While the technical aspects of rotator cuff repair are important considerations, many surgeons believe that the biologic process of healing represents a critical and poorly understood process. Recently, attempts to manipulate the biology of healing have grown in popularity, with little science backing this clinical decision. Randelli et al.9 performed a prospective, randomized, controlled study of fifty-three patients. The treatment group received an intraoperative application of platelet-rich plasma (PRP) in combination with an autologous thrombin component. The control group did not receive PRP. Patients were evaluated with validated outcome scores. Magnetic resonance imaging was performed for all patients at more than one year postoperatively. The pain scores, external rotation strength, and outcome measures (University of California Los Angeles [UCLA], Simple Shoulder Test [SST], and Constant scores) were improved at three months or less in the PRP group as compared with the control group. There was no difference between the groups at six, twelve and twenty-four months, although when the results were stratified according to tear size, smaller tears were associated with improved external rotation strength in the treatment group at twenty-four months. The follow-up MRI evaluation showed no significant difference between the groups in terms of the rate of healing of the rotator cuff tear. (Level I)
Glenohumeral Instability
Shoulder instability has been described in a number of ways without a universally accepted or validated classification system. Kuhn et al.10 developed a classification system by first conducting a systematic literature review that identified eighteen systems for the classification of glenohumeral instability. In addition, thirty-one members of the ASES responded to a survey to identify features important for the characterization of glenohumeral instability. With use of this information, a four-part classification of instability on the basis of frequency, etiology, direction, and severity (the FEDS system) was developed. The interobserver and intraobserver reliability of the classification system was then tested in fifty patients with glenohumeral instability. Intraobserver and interobserver reliability were high. However, there was wide variation and lower agreement between the physician and patient scores, particularly with regard to the frequency of instability. The FEDS system demonstrated content validity and is highly reliable for classifying glenohumeral instability. Physical examination with use of provocative testing to determine the primary direction of instability produces very high levels of inter-rater and intra-rater agreement. (Level II)
Initial immobilization of the shoulder following primary traumatic shoulder dislocations varies widely, but its effect is largely unproven. Liavaag et al.11 prospectively evaluated the optimal position of immobilization following first-time glenohumeral dislocation. Ninety-five patients were managed with immobilization in internal rotation for three weeks, and ninety-three were managed with immobilization in external rotation for three weeks. The rate of follow-up was 98% in both groups. The rate of compliance with the method of immobilization was moderate for both the internal and external rotation groups (47.4% and 67.7%, respectively). At twenty-four months of follow-up, there was no difference between the two groups in terms of the rate of recurrent dislocation. The lack of patient compliance limits the strength of the conclusions. However, it appears that immobilization in external rotation does not reduce the rate of recurrence for patients with first-time traumatic anterior shoulder dislocation. (Level I)
The effect of immobilization on recurrence was also investigated by Paterson et al.12. Those authors performed a review of 2083 published studies on glenohumeral instability and identified nine level-I and level-II studies that they then systematically reviewed. The authors determined that the length of time that a shoulder is immobilized following traumatic anterior dislocation does not impact likelihood of recurrence. In their review of six level-I or II series, immobilization of the shoulder in internal rotation (in a sling) yielded a 37% to 41% rate of recurrence in patients younger than thirty years of age, with no significant difference in the rate of recurrence being identified between one week (41%) and three weeks (37%) of immobilization. An age of less than thirty years at the time of the index dislocation was significantly predictive of recurrence in most studies. Three studies evaluating the role of immobilization in external rotation demonstrated a recurrence rate of 25%. When the findings of these three studies were compared with the findings of the studies involving immobilization in internal rotation, there was no significant difference in terms of the rate of redislocation, with the numbers available. (Level II)
Glenohumeral Arthritis
Basic Science
Reverse shoulder arthroplasty is used in many clinical scenarios, including those associated with deficiency of the proximal part of the humerus. Clinical reports have documented dissociation of the humeral body from the stem with deficiency of the proximal of the humerus. Cuff et al.13 studied the effect of proximal humeral bone loss on rotational stability of the humeral implant following reverse total shoulder replacement. Modular and monoblock designs were implanted into humeral Sawbones (Pacific Research Laboratories, Vashon, Washington) to test an intact and a deficient proximal humeral state. Cyclic torsional loading of 2.5 to 25 N-m was applied to the humerus for 1000 cycles. The results of testing demonstrated lower rotational micromotion in the intact model as compared with bone-loss model. In the intact model, ten of twelve implant constructs survived testing. In the bone-loss model, seven of twelve implant constructs survived testing. All of the implant failures that occurred were modular humeral stems.
General
It is important to understand trends and projections of procedure volumes to plan for the future. Day et al.14 studied the Medicare Nationwide Inpatient Sample and United States census data from 1993 to 2007 to determine procedure rates and to determine year-to-year trends in primary and revision shoulder and elbow arthroplasty. On the basis of these data, shoulder and elbow arthroplasty rates in 2015 were projected. During the study period, procedural volumes and rates increased at annual rates of 6% to 13% from 1993 to 2007. Compared with 2007 levels, primary arthroplasty volume was projected to increase 192% by 2015 and revision rates were projected to increase 322%. Of particular concern was the increased revision burden. The rising number of arthroplasty procedures combined with increased cost has the potential to place a financial strain on the health-care system.
Outcomes of Arthroplasty
Cemented humeral stems for shoulder arthroplasty are the current standard of care. However, many surgeons recognize the potential advantages of press-fit fixation without cement. Litchfield et al.15 performed a multicenter prospective, randomized, double-blind clinical trial comparing humeral fixation with and without cement in patients managed with total shoulder arthroplasty for the treatment of primary shoulder osteoarthritis. Eighty patients were randomized to receive a cemented stem, and eighty were randomized to receive an uncemented stem. All patients received a standard shoulder implant system. The twelve, eighteen, and twenty-four-month outcomes showed a significant difference in favor of the cemented-stem group. The cemented-stem group also had better strength and forward flexion. For the implant utilized in this study, the results support the use of cement fixation for the humeral component. (Level I)
Fractures
Clavicular Fractures
The surgical treatment of displaced midshaft clavicular fractures has gained popularity recently. However, the advantages of one method of fixation over another are not clearly established. Ferran et al.16 performed a randomized clinical trial in which intramedullary fixation was compared with plate fixation. Locked intramedullary fixation was used for seventeen patients, and plate fixation was used for fifteen patients. The patients were followed for an average of 12.4 months. The primary outcome measure was the Constant score; secondary outcome measures included the Oxford shoulder score, the union rate, and complication rates. All patients in each group had healing of the fracture. No significant difference was found in terms of Constant or Oxford scores. Complication rates were similar between groups, but all pins required removal whereas eight (53%) of fifteen plates were removed. (Level I)
Proximal Humeral Fractures
The treatment of proximal humeral fractures in elderly patients remains controversial. The introduction of locking-plate technology has expanded the indications for plate fixation of these fractures. Bell et al.17 performed a Medicare database analysis to characterize the incidence, treatment, and revision surgical treatment of proximal humeral fractures according to geographic region both before and after the introduction of locking plates. By comparing data from 1999 to 2000 with data from 2004 to 2005, the authors were able to determine the total number of proximal humeral fractures and the number of fractures that were treated with hemiarthroplasty, plate fixation, or closed treatment. The authors found the total number of fractures remained relatively the same during the study periods. However, the overall rate of surgical treatment of proximal humeral fractures significantly increased by 25.6%. Specifically, locking-plate fixation increased by 28.5% and hemiarthroplasty increased by 19.6%. In addition, there was large regional variation in the rates of the different treatment types for proximal humeral fractures. The authors concluded that the variation signified a need to establish consensus regarding treatment. (Level II)
Zhu et al.18 performed a prospective randomized study of fifty-one patients in which humeral intramedullary nailing was compared with locking-plate fixation for the treatment of two-part surgical neck fractures. Clinical and radiographic assessments were conducted at one year and three years postoperatively. At one year, all fractures were healed. The outcomes that were evaluated were significantly better in the plate fixation group as compared with the intramedullary nail group. However, the complication rate was 31% in the plate fixation group and only 4% for the intramedullary nail group. Interestingly, at three years postoperatively, no differences were observed between the groups, leading the authors to conclude that both fixation methods yield satisfactory outcomes. (Level I)
Failure following open reduction and internal fixation of proximal humeral fractures can result in implant cutout from the humeral head, varus malunion, or nonunion. Support of the medial calcar with an intramedullary fibular allograft has been advocated as a way to minimize failure. Osterhoff et al.19 performed a biomechanical evaluation of a surgical neck fracture model in a study involving twenty synthetic bones. In one test group (n = 10), the fractures were fixed with a locking plate system without an intramedullary fibular graft. In the other group (n = 10), the same fixation system was used with the addition of an intramedullary graft. Cyclic loading of the model was performed. The authors demonstrated that the incorporation of an intramedullary fibular strut graft into locking plate fixation increased the overall stiffness of the construct and reduced humeral head migration as compared with plating alone. This technique might provide a useful tool for the treatment of displaced proximal humeral fractures, especially when there is medial comminution.