Introduction

The degree of glenoid bone loss associated with primary glenohumeral osteoarthritis can influence the type of glenoid implant selected and its placement in total shoulder arthroplasty (TSA). The literature has demonstrated inaccurate glenoid component placement when using standard instruments and two-dimensional (2D) imaging without templating, particularly as the degree of glenoid deformity or bone loss worsens. Published results have demonstrated improved accuracy of implant placement when using three-dimensional (3D) computed tomography (CT) imaging with implant templating and patient specific instrumentation (PSI). Accurate placement of the glenoid component in TSA is expected to decrease component malposition and better correct pathologic deformity in order to decrease the risk of component loosening and failure over time. Different types of PSI have been described. Some PSI use 3D printed single use disposable instrumentation, while others use adjustable and reusable-patient specific instrumentation (R-PSI). However, no studies have directly compared the accuracy of different types of PSI in shoulder arthroplasty. We combined our clinical experience and compare the accuracy of glenoid implant placement with five different types of instrumentation when using 3D CT imaging, preoperative planning and implant templating in a series of 173 patients undergoing primary TSA. Our hypothesis was that all PSI technologies would demonstrate equivalent accuracy of implant placement and that PSI would show the most benefit with more severe glenoid deformity.

The number of shoulder arthroplasty procedures performed in the United States continues to rise. Currently, the number of procedures performed per year ranges from 55,000–80,000 and is expected to increase more than 300% in the coming years. Periprosthetic joint infection (PJI) is one of the most serious complications associated with arthroplasty surgery, leading to poor outcomes, increased cost, and technically difficult revision surgery. The incidence of infection following primary shoulder arthroplasty has been reported between 0.7% and 4%, representing 2.9–4.6% of all complications.

Prosthetic shoulder joint infections are unlike prosthetic joint infections of the hip and knee. Shoulder PJIs are primarily indolent in nature and difficult to diagnose using traditional methods that have been shown to be accurate for periprosthetic infections of the hip and knee. The majority of infected revision shoulder arthroplasties are associated with growth of Propionibacterium acnes (P. Acnes). This slow-growing, anaerobic organism requires longer than normal incubation times for culture (7–21 days), and typically demonstrates a subtle, non-specific clinical presentation that can make the presence of infection difficult to identify. In the reported literature, P. Acnes accounts for about 70% of cases with positive cultures associated with revision for treatment of a painful shoulder arthroplasty and due to the bacteria's slow growing nature and virulence profile, the rate of infection following shoulder arthroplasty may often be underestimated.

A more recent and promising tool for evaluation of periprosthetic infection has been analysis of synovial fluid. Synovial fluid biomarkers have been identified as part of the innate response to pathogens, and include pro-inflammatory cytokines and anti-microbial peptides, and marker levels have shown promise for improved diagnostic efficacy in hip and knee PJI. Currently, no highly predictive clinical test for diagnosis of PJI in the shoulder exists, however, several of these synovial biomarkers have recently been analyzed for their diagnostic capacity in the setting of periprosthetic shoulder infection.

Synovial fluid cytokine analysis shows the potential to improve diagnosis of infection in revision shoulder arthroplasty. This information can help to guide decision-making in the management of PJI of the shoulder, including the decision to perform a single- vs. two-stage revision surgery, and the need for post-operative antibiotics following an unexpected positive culture result after revision surgery. However, there are still challenges to broader use of these synovial biomarkers. Synovial α-defensin (Synovsure, CD Diagnostic) is the only marker currently available as a commercial test, and no point-of-care test is currently available for any of the biomarkers to allow for intraoperative decision-making. While a preoperative synovial aspirate is possible to send for α-defensin analysis currently, with results back in approximately 24 hours, dry fluid aspirations are frequent in the shoulder because of the predominance of indolent pathogens and may limit utility of the test.

In summary, indolent infection associated with P. acnes is a common cause for the painful total shoulder arthroplasty. Pre-operative diagnosis of infection is difficult as a result of the poor diagnostic accuracy of traditional methods of testing. Synovial biomarker testing may ultimately improve our ability to more accurately diagnosis and treat prosthetic shoulder joint infections.

Humeral head size is defined by the radius of curvature and the thickness of the articular segment. This ratio of radius to thickness is within a narrow range with an average of 0.71. The articular surface of the normal humeral head measured within the AP plane is defined by three landmarks on the non-articular surface of the proximal humerus. The perfect circle concept can be applied for assessment of the anatomic reconstruction of the post-operative x-rays and more importantly can be used intra-operatively as a guide when choosing the proper prosthetic humeral head component. The humeral head is an elliptical shape with its AP dimension being approximately 2 mm less than the SI dimension. This shape contributes to the roll and translation of the normal shoulder but is not replicated by the spherical shape of the prosthetic humeral head.

The glenoid vault has a consistent 3D shape and use of the vault model within 3D planning software can define the patient's pre-morbid anatomy, specifically the location of the joint line and patient specific version and inclination. Use of this tool can assist the surgeon in defining the optimal implant and its location. In patients with little or no bone loss, a symmetric glenoid implant is often ideal for resurfacing. When there is asymmetric bone loss, often seen posteriorly with osteoarthritis, an asymmetric posteriorly augmented component can improve the ability to correct the deformity while maintaining the native joint line. It is suggested that these augmented implants in selected patients will help restore and maintain humeral alignment and lessen the risk for residual posterior humeral head subluxation and eccentric loading of the glenoid component.

CT-based three-dimensional (3D) pre-operative imaging along with 2D orthogonal sections defined by the plane of the scapula (axial, sagittal and coronal planes) has been demonstrated by many research groups to be a very accurate way to define the bone pathology and alignment/subluxation of the humeral head in relationship to the center line of the scapula or the center of the glenoid fossa. When 3D CT imaging is combined with 3D implant templating the surgeon is best able to define the optimal implant and its location for the desired correction of the bone abnormalities. The use and value of 3D imaging is best when the there is more severe bone pathology and deformity.

Transferring the computer-based information of implant location to the surgical site can involve multiple methods. The three methods discussed in the literature to date including use of standard instrumentation in a manner specified by the pre-operative planning, use of single-use patient specific instrumentation and use of reusable patient specific instrumentation. Several cadaver and sawbone studies have demonstrated significant improvement in placement of the glenoid implant with both single use and reusable patient specific instrumentation when compared to use of 2D imaging and standard instrumentation. Randomised clinical trials have also shown that 3D planning and implant templating is very effective in accurate placement of the implant in the desired location using all three types of instrumentation.

The optimal use of this technology is dependent upon the severity of the pathology and the experience and preference of the surgeon. With more severe pathology and less surgeon experience 3D pre-operative imaging and templating and use of some level of patient specific instrumentation provides more accurate placement of the glenoid implant.

Peri-prosthetic joint infection (PJI) can be both a diagnostic and therapeutic challenge in shoulder arthroplasty, due to the indolent nature of the common infecting organisms. Proprionobacterium acnes (P. acnes) is the most common pathogen cultured in revision shoulder arthroplasty. It is a slow growing, anaerobic organism – requires longer incubation period (7–21 days). Coagulase-negative Staphylococcus species (CNSS) is also a common organism responsible for PJI.

Established diagnostic tests for hip and knee PJI are often negative in the shoulder despite post-operative growth of intra-operative cultures. Pre-operative synovial aspiration often low volume due to indolent pathogens and successful aspiration is often reported to be 50% or less with Dilisio et al, JBJS 2014: reporting 16.7% sensitivity, 100% specificity. Variable culture length for P. acnes culture protocols are reported from 7–28 days with most groups recommending 14 days.

From our research, we demonstrated time to culture growth was significantly shorter in probable true positive culture group (median, 5 vs. 9 days, p=0.002). Frozen section analysis may help intra-operative decision-making (one- vs. two-stage reimplantation) yet the reported sensitivity and specificity in shoulder arthroplasty is far less than in hip and knee arthroplasty. Synovial fluid biomarkers have been identified as part of the innate response to pathogens include pro-inflammatory cytokines and antimicrobial peptides. In a series of prospective studies of revision shoulder arthroplasty, synovial fluid analysis reported by Frangiamore et al, JBJS 2015: IL-6, Frangiamore et al, JSES 2015: α-defensin (Synovasure^TM), Frangiamore et al, AAOS 2015: Broader cytokine analysis it was demonstrated that these markers are much more predictive of infection than synovial fluid cultures, frozen section or serum markers.

Background:

Glenoid component loosening remains as an unsolved clinical problem in total shoulder arthroplasty. Current clinical assessment relies on subjective quantification using a two-dimensional plane X-ray image with arbitrarily defined criteria. There is a need to develop a readily usable clinical tool to accurately and reliably quantify the glenoid component motion over time after surgery. A high-resolution clinical CT has the potential to quantify the glenoid motion, but is challenged by metal artifact from the prosthetic humeral components. The objective of this study is to demonstrate the feasibility of using a clinical CT reconstruction to quantify the glenoid implant motion with the aid of tantalum markers.

Introduction:

Patient medical comorbidities are well-established risk modifiers of THA patient outcomes. Patient's mental state preoperatively may influence postoperative functional outcomes though just like any medical comorbidity. This study sought to determine if patient confidence in attaining post-operative functional goals was associated with objective and subjective outcomes following THA.

Introduction:

Total knee arthroplasty (TKA) outcomes drive assessment of quality and reinvestment; therefore a risk stratified assessment is paramount for fair evaluation. Stratification can be affected by multiple factors including patient motivation. This study attempted to identify the correlation of patient's preoperative confidence in their ability to return to desired activity level after TKA and improved function and outcomes.

Accurate implant placement is important to the success of joint replacement surgery. Three dimensional pre-operative planning optimizing the ability to define the anatomy and select the desired implant and its location. Linking this information into implant and patient specific instrumentation has been termed smart instrumentation. Single use instruments contain the patient's topographical boney anatomy and implant information. This same information can be placed within a bone model and a reusable instrument placed onto the bone model can be adjusted to capture the information originally in the planning software. This instrument can then be placed into the surgical site, registered to the bone surfaces and used to modify the bone surfaces to replicate the surgical plan. These concepts, technology and devices have been developed and clinically tested in randomized clinical trials for shoulder arthroplasty.

Advanced 3D CT imaging of the shoulder using metal artifact reduction techniques have been developed and validated in pre-clinical and patient population using microscribe and RSA techniques. These studies demonstrate these methods to be highly accurate and reproducible in the measurement of implant position and implant migrations. 3D CT imaging is being used to study the anatomic and clinical factors associated with glenoid implant loosening. This imaging modality does not require marker based image registration and utilizes imaging equipment that is readily available in the clinical setting.

Background:

Acetabular component malpositioning in total hip arthroplasty increases the risk of dislocations, impingement, and long-term component wear. The purpose of this Sawbones study was to define the efficacy of a novel acetabular imprinting device (AID) with 3D preoperative planning in accurately placing the acetabular component.

Background: Time dependent, quantitative studies of muscle atrophy or passive muscle mechanics following chronic muscle detachment have not been previously reported. We developed a chronic tear of the rotator cuff tendon in a canine model to investigate and quantify the time related changes in the passive mechanics, volume, and fat of the infraspinatus muscle. We hypothesize that infraspinatus muscle stiffness will increase, volume will decrease, and fat content will increase at 12-weeks following tendon detachment.

Methods: The right infraspinatus tendons of eight adult mongrel dogs were surgically detached from the proximal humerus. The non-operated left shoulder served as a control. Muscle volume changes were quantified using MRI scans. At 12 weeks the passive mechanical properties of the chronically detached and control muscles were determined intraoperatively using a custom device. Intramuscular fat was evaluated histologically at sacrifice.

Results: After 12 weeks of detachment, the stiffness and modulus were significantly increased in the detached infraspinatus muscles relative to controls. MRI analysis demonstrated that the detached muscle volumes decreased by 33 percent in the first 6 weeks and remained constant thereafter. Intramuscular fat increased significantly in the detached muscles, and to a greater extent in the lateral regions.

Conclusions: The chronically detached muscle is not merely a smaller version of the original muscle but rather a “different” muscle. The detached muscle becomes stiffer and the passive loads required to repair it can become excessive. A significant reduction in muscle volume occurs within days to weeks following tendon detachment. The non-uniformity of muscle fat changes suggests that fat content should be used cautiously as an indicator of muscle quality.

Clinical Relevance: Clinically, chronic, large rotator cuff tendon tears are observed to have a qualitatively shorter and stiffer muscle-tendon unit than normal. We have developed a chronic rotator cuff model to quantitatively investigate changes in the detached infraspinatus muscle. The passive mechanical properties of a chronically torn rotator cuff muscle-tendon unit may be a useful predictor of reparability and clinical outcome.

Periprosthetic fractures occur in approximately 1–3% of case series. Periprosthetic fractures are associated with revision surgery with difficult exposure, osteoporosis, large canal filling non-cemented stem design, overreaming of the medullary canal, and excessive external rotation with inadequate exposure. Periprosthetic fractures can be intentional when removing a well-fixed humeral stem. In this circumstance a longitudinal unicortical osteotomy along the anterior length of the stem will allow for stem and cement removal without fragmentation of the humeral bone.

Periprosthetic fractures are classified as occurring intraoperative versus postoperative as well as the location of the fracture in relation to the stem. Most intraoperative humeral fractures and all diaphyseal fractures should be x-rayed at the time of their occurrence to determine the fracture configuration, the best exposure for repair, and the length of the stem required to internally fix the fracture. Under ideal circumstances the stem should be of sufficient length to extend two cortical widths past the distal most extent of the fracture site. For fractures limited to one or both of the tuberosities, the surgical neck, or metaphyseal-diaphyseal junction, a standard length prosthetic is sufficient. For diaphyseal fractures a long stem prosthetic is necessary. In the vast majority of fractures in which the fracture fragment is displaced, open reduction and cerclage fixation with heavy suture or wire is needed. For fractures in which the proximal bone is intact and of good quality thereby providing good prosthetic fixation and rotational stability, the diaphyseal fracture can be anatomically reduced and secured with two or three cerclage wires (Dall Meyers cables or the equivalent). In this case a non-cemented long stem prosthetic is preferred. When a cemented stem is used, it is necessary to insure that cement is not extruded from the fracture site. This is accomplished by having adequate surgical exposure of the fracture, an anatomic reduction, and secure fixation before you place the cement and stem. Extruded cement may result in nerve injury or nonunion.

Intentional longitudinal fractures require direct exposure of the length of the osteotomy to control its length and displacement. It is advised to pass the cerclage wires prior to making the osteotomy. In the humerus, the osteotomy is best made just lateral to the biceps groove with an osteotome. The osteotome is placed to the depth of the stem and through the cement mantle when this is present. When the osteotomy is nearly to the length of the stem the osteotome is placed at the proximal extent of the osteotomy at approximately the mid-level of the biceps groove to a depth of the stem and then turned. This will crack the cement mantle of the opposite side of the medullary canal and open the anterior cortex. It results in a stable fracture of the humeral shaft but allows easy removal of the stem and facilitates removal of the stem from both the proximal aspect of the medullary canal and from the osteotomy site. After completion of the stem and cement removal the cerclage wires are tightened and the new stem is inserted. When secure fixation is achieved with a periprosthetic fracture, regardless of the type of fracture, the postoperative rehabilitation is the same as a routine arthroplasty and the results and time for recovery is unchanged.

Nonoperative treatment of periprosthetic fractures are reserved for the postoperative fracture occurring below the stem in a patient with a well-fixed and a functioning prosthetic, or in patients that have medical contraindication to revision surgery. A functional hinged brace can be used to help in reduction of these fractures and immobilisation of the fracture site. The braces are difficult to use and are less effective in patients with a large soft tissue envelope. Skin problems and nonunions or malunions can occur. In most cases when there is an inadequate reduction, difficult immobilisation, or stem involvement, it is best to operate soon after the fracture as late revisions in the setting of a nonunion or malunion are difficult surgical challenges.

The humeral offset has a medial-lateral and anteriorposterior dimension and can be defined as the distance of the central axis of the humeral shaft and the center of rotation of the humeral head. When using a canal filling prosthetic stem, inserted in a collinear alignment with the long axis of the humeral shaft, the placement of the humeral head within the anatomic boundaries of the humeral osteotomy surface will be dictated by how closely the prosthetic stem-humeral head offsets match that of the patient’s natural anatomic offsets. Given the fact that there are several millimetres of variation in the medial-lateral and anterior-posterior humeral offsets among the normal patient population, it follows that in many cases the prosthetic offset will not precisely match that of each patient when a canal filling medullary component is properly inserted. This mismatch in the humeral offsets can result in malposition of the humeral head within the confines of the humeral osteotomy surface when using a centred Morse taper humeral head.

Iannotti and Williams have studied, using the Polhemus tracking system, the biomechanical consequences and tolerances for a malpositioned humeral head using a prosthetic reconstructed human cadaver model (JOR in press). A 4 mm or greater malposition of the humeral head, particularly in an inferior or anterior direction (most common malposition) will result in a measurable decrease in range of motion, abnormal humeral kinematics and subacromial impingement. In addition, humeral head malposition (anterior and inferior) can result in impingement of the non-articular portion of the humeral component, rotator cuff or proximal humeral metaphysis on the glenoid component. This impingement can result in abnormal glenoid component wear (type 2 wear) or premature glenoid loosening.

Solutions for management of humeral head malposition include: 1.) variable stem offsets (not currently available), 2.) humeral component with an eccentric Morse taper (DePuy – Global Advantage, Tomier – Aequalis, Zimmer – Bigliani/Flatow) or an equivalent design (Stryker – Anatomica), and 3.) under-sizing and shifting the humeral stem to a centred position. The last option, best for monoblock humeral components (original Neer design), requires impaction grafting or PMMA cement to stabilise the under-sized stem in the centred position. Whatever solution is used it is important, in most cases, to select a head size which matches the normal anatomy and center it within the confines of the normal humeral osteotomy surface area.

The indications for use of a glenoid component are: 1.) sufficient degenerative changes on the glenoid to expose the subchondral bone 2.) the glenoid should have sufficient glenoid bone stock to allow for secure and longterm fixation of the component, and 3.) the rotator cuff should be intact or repairable and the humeral head should be centred within the glenoid component. Other factors that secondarily affect the decision to use a glenoid component, include the patient’s age and activity level, which should be such that they are not likely to result in multiple revisions for glenoid wear or loosening.

Given these factors most patients with osteoarthritis, the leading indication for prosthetic replacement for arthritis should undergo a total shoulder replacement. Patients with acute proximal humeral fractures, the overall leading indication for prosthetic arthroplasty, should have a hemiarthroplasty. Patients with rotator cuff tear arthropathy or crystalline arthropathy are indicated for hemiarthroplasty due to the massive irreparable cuff tears present in these conditions. Patients with rheumatoid arthritis have variable diseases affecting the rotator cuff and variable degrees of bone loss resulting in the need to individualise the indications for the use of a glenoid to the patient’s pathoanatomy. The reason for use of a glenoid component, when indicated, is the fact that pain relief and function is predictably better when compared to hemiarthroplasty for the same indication and pathoanatomy. Proper insertion of a glenoid component requires wide exposure of the glenoid fossa and bone preparation, which for most general orthopaedic surgeons is difficult and not reproducible. This is, in my opinion, the primary reason that hemiarthroplasty or bipolar arthroplasty is used for treatment of many patients with primary osteoarthritis. Both of these procedures result, on average, in a less favourable outcome than non-constrained total shoulder arthroplasty.

Osteoarthritis is characterised by flattening and enlargement of the humeral head and is associated with peripheral osteophyte formation. Loss of articular cartilage results in eburnated bone and on the glenoid side posterior bone loss. Capsular contracture results in loss of passive arcs of motion, particularly anteriorly with loss of external rotation. Posterior subluxation of the humeral head can occur, associated with anterior soft tissue contracture and/or posterior glenoid bone loss. The severity of this pathoanatomy is variable among patients with primary osteoarthritis and each of these factors will have a variable effect on outcome of shoulder arthroplasty as well as the indication for hemiarthroplasty versus total shoulder arthroplasty.

In a 2–7 year follow-up multicentre study using the DePuy Global Shoulder in 127 patients, those cases with osteoarthritis without humeral head subluxation, severe glenoid bone loss, or rotator cuff tears had the best results, for pain relief and function, with total shoulder arthroplasty. In patients with severe glenoid bone loss total shoulder has improved function when compared to hemiarthroplasty. This finding supports the data of others that demonstrate less favourable results of hemiarthroplasty for treatment of osteoarthritis in cases with eccentric glenoid wear. Patients with humeral head subluxation have less favourable results regardless of the use of a hemiarthroplasty or total shoulder arthroplasty. The presence of a full thickness reparable rotator cuff tear limited to the supraspinatus tendon does not adversely affect outcome or the ability to use a glenoid component. Patients with less than 10° of external rotation achieve statistically less postoperative forward flexion and external rotation than those patients with greater degrees of preoperative external rotation.