Opinion editorial—defining the ideal indications and surgical technique for an anatomic total shoulder arthroplasty for primary osteoarthritis of the shoulder

ABOUT THE EXPERT

Joseph Iannotti, MD, PhD, is Professor of Orthopaedic Surgery at the Cleveland Clinic School of Medicine at Case Western Reserve University; he is also Chief of Staff and Chief Academic and Innovations Officer at Cleveland Clinic Florida. Dr. Iannotti’s research interests are related to the study of shoulder biomechanics, prosthetic design, and clinical outcomes in shoulder arthroplasty and rotator cuff repair.

Eric Ricchetti, MD, is Associate Professor of Orthopaedic Surgery at the Cleveland Clinic School of Medicine at Case Western Reserve University; he also holds the Maynard Madden Chair for Arthritis Research. Dr. Ricchetti’s research includes the study of clinical outcomes of prosthetic shoulder arthroplasty and rotator cuff repair.

Introduction

Anatomic total shoulder arthroplasty (TSA) has been established to reliably achieve relief from pain and restoration of function in a vast majority of patients with primary osteoarthritis. In recent years the literature reports more than 90% of patients have marked improvement of pain and function, with less than 5% of patients requiring revision within 10 years of surgery. However, studies are still limited to defining the factors associated with failure after 10 or more years of follow-up, particularly with more modern implant designs. *

* References , , , , , , .

An excellent clinical result is defined by the absence of all pain and return of normal function. Although improvement in pain and function is reported, restoration of normal function and complete pain relief do not occur often, yet surgeons and patients try to achieve this type of result. Clinical outcome is defined by patient-reported scores and the patient’s physical function; however, there is not a universally accepted method to define an excellent outcome after shoulder arthroplasty. Clinical outcome is generally measured using assessment tools completed by the patient. These tools have limits because they are designed to be useful for the majority of patients and do not define the outcome of patients with very high functional demands.

One of the most important reasons that this chapter has greater relevance nowadays than a decade ago results from the widespread use of reverse TSA. There is an ever-increasing use of reverse TSA to treat the more complex pathology that in past decades was treated with hemiarthroplasty or anatomic TSA with highly variable outcomes and in some cases a high rate of poor outcomes. As a result, the primary indication for anatomic TSA currently is in the patients who are younger, with higher physical demands and expectations and lesser forms of bone pathology with an intact rotator cuff. That means that the principles of anatomic reconstruction become increasingly more important in this population of patients to help them achieve the long-term clinical and radiographic outcomes that they need and expect.

The anatomic result is defined by postoperative imaging, yet there is no agreement among surgeons on how to define and measure a perfect anatomic reconstruction. Postoperative radiographs are the current standard method of imaging. Routine radiographs provide two-dimension imaging that is not able to reliably define early signs of implant shift and have poor reproducibility in making the same measurements over time. Three-dimensional (3D) imaging is not routinely performed and currently requires repeated computed tomography (CT) scans with special techniques to reduce metal artifact that would otherwise degrade the ability to measure implant location and define the bone implant interface. In addition, 3D CT imaging techniques are expensive and carry higher radiation exposure than routine radiographs. These limitations in postoperative imaging have severely limited our ability to define the cause of implant failure over time.

The goal of this chapter is to define the current ideal state for anatomic TSA. It would be worth recognizing that neither the literature nor our clinical experience supports our ability to obtain a perfect anatomic result consistently, even if we are able to define what those criteria are for an individual patient. In fact, when we approach a perfect anatomic outcome, it is possible that we do not achieve a perfect or even a satisfactory clinical result. In fact, a good clinical result, defined by the patient, can occur with nonanatomic placement of the implants. This discordance, along with compromised postoperative imaging, creates even greater challenges in proving the relationship between anatomic reconstruction and clinical outcome. Coupled with the highly variable biologic factors of healing, variation in rehabilitation protocols and effort, and recognized and unrecognized complications, a number of postoperative events can influence the final result and shorten the duration of an excellent outcome. These challenges, coupled with the declining frequency of anatomic TSA, make rigorous clinical studies that define the correlation between anatomic arthroplasty (imaging) and clinical outcome difficult and rarely reported in the current literature.

These inconsistencies between anatomic and clinical outcomes make clinical research imprecise. These difficulties are made worse by the imprecision of routine two-dimensional imaging, the lack of longitudinal follow-up of the same cohort of patients, and the underpowered sample size of most retrospective studies. These facts make it difficult to precisely define an anatomic reconstruction. Moreover, there are many implant and surgical factors that can influence the variations between good to excellent clinical outcomes or the factors associated with excellent long-term survivorship of the implants. For the most part, our clinical studies have focused on defining the variations of pathology or surgical technique that are correlated with a poor result, rather than what factors define an excellent long-term outcome. Therefore what remains to be defined are the factors, implants, and surgical techniques that are reliably associated with an excellent long-term result.

Retrospective studies, some being underpowered and poorly controlled, have suggested several anatomic or surgical factors to be associated with less favorable outcomes of anatomic TSA, including Walch B2 or B3 glenoid morphology ^, and tears or fatty infiltration of the rotator cuff. ^, ^, This has resulted in increased use of reverse TSA by many surgeons, including an increasing trend to use a reverse TSA for mild pathology, such as partial-thickness rotator cuff tears seen on preoperative magnetic resonance imaging or mild glenoid bone loss noted on a preoperative CT scan. Less research has been focused on improving the results of anatomic TSA, in part due to the ability to default to a reverse TSA. It is our opinion that the overuse of reverse TSA in younger and more active patients with less severe pathology will result in clinical failures that will renew our interest in the future to define the factors that are associated with better outcomes following anatomic TSA.

It is also noteworthy to consider that as implant design, implant materials, and surgical techniques improve, the indications and clinical circumstances to achieve an excellent and durable result after anatomic TSA will broaden. Our opinions and position, as stated as follows, will evolve as we learn more about the factors associated with our current clinical and anatomic outcomes. The value of our position statement is therefore to identify what we believe to be current surgical techniques that have proven, in our hands, to yield the best and most consistent short- and mid-term results. We will present the data that we have developed in our clinical studies that support these opinions.

Ideal patient

We all learn early in our clinical training and practice experience that it is often who we treat surgically that most often influences the outcome of the surgery we perform. For example, we know that use of opioid narcotics before surgery or those with low mental health scores have less favorable outcomes. These patient factors are not consistently measured or taken into consideration when reporting clinical or anatomic outcomes. This decreases the ability to study the clinical outcome associated with the severity of the pathology, the implant, or surgical technique. The smaller the cohort or the less evenly distributed these patient factors are within the cohort, the more likely these patient factors compromise the ability to find differences in the factors we are trying to measure. Moreover, the less precise the imaging and less discriminating the patient-reported outcome tools used, the more the outcome of the surgery may be attributed to variation in patient factors.

All of these patient factors and limitations associated with not measuring them should be considered when reviewing the literature, particularly in retrospective studies with a small sample size in a heterogeneous population.

The ability of the patient to understand and perform the daily exercises that begin on postoperative day 1 is critical to achieving an excellent functional result. Multimodal pain management and having sufficient and dedicated protected time to perform the exercise program are also important. There are patients with ideal pathology or surgical technique who do not participate in the prescribed rehabilitation. Some patients have comorbidities related to age or movement disorders that limit their ability to perform these basic exercises. Chronic diseases that are not well managed or any acute event in the first few weeks after surgery can result in less than ideal performance of the exercises, which has a negative correlation with ultimate functional outcome. Progression to strengthening exercises after the first 8 weeks from surgery is equally important to achieving normal shoulder function. Once postoperative rehabilitation is complete, returning to functions that are very strenuous on a regular basis will shorten the duration of an excellent result.

Ideal pathology

The ideal pathologic condition with end-stage primary osteoarthritis will have a centered humeral head without significant glenoid bone loss (A1 Walch), an intact rotator cuff with no more than grade 2 fatty infiltration of any of the rotator cuff musculature, measured on CT or magnetic resonance imaging scan, no more than moderate joint stiffness, and no prior history of surgery. In our patient population and in most reported series, this occurs in approximately 30% of patients with primary osteoarthritis. ^, Commonly understood current clinical practice would indicate that almost all surgeons would recommend and perform an anatomic TSA on this type of patient. Most surgeons would offer this surgical option because the clinical data would support an anatomic TSA to be the best surgical option to get an excellent result.

It is the other 70% of patients with varying degrees of more severe pathology that we debate the role of anatomic TSA, the type of implant, its ideal position, and surgical technique. With moderate glenoid bone loss and some thinning of the rotator cuff, some surgeons will elect to perform a reverse TSA, and other surgeons will perform an anatomic TSA with modification of the implant or surgical technique. How should the surgical technique be performed, and which anatomic TSA implant should be used with this pathology? Will the clinical outcome be equivalent to the patient without this pathology? This debate remains active and controversial. It changes often and is dependent upon new data regarding new implants and new data on implants and surgical techniques used years ago.

Patients with an intact rotator cuff that is thin but intact and associated inflammatory arthritis of any type have a higher risk for later rotator cuff tears and ultimately loosening of the glenoid component resulting from eccentric loading of the rotator cuff. ^, ^, ^, ^, If this type of patient is disabled with other joint problems and has a lower activity level, then a reverse TSA is preferred over an anatomic replacement in our practice. If an anatomic TSA is performed with this pathology in a younger patient, then we would recommend use of a noncemented inlay platform humeral stem that can be converted to a reverse TSA without removal of the stem if a rotator cuff tear develops in the future.

Asymmetric bone loss in primary osteoarthritis can be classified by the shape of the glenoid and the position of the humeral head in relation to the plane of the glenoid fossa or scapula. Glenoid shape in primary osteoarthritis is commonly classified by the modified Walch classification: A1 or A2; B1, B2, or B3; C1 or C2; and D. ^, ^, The Walch classification does not define severity but rather the shape of the glenoid bone loss. Severity is best measured quantitatively in 3D using a method that compares the pathology with the premorbid bone anatomy. Software technology has used statistical modeling and machine learning techniques to estimate premorbid anatomy from pathologic anatomy. Since 2007 our group has developed, validated, and used the glenoid vault model as a reference defining the location of the premorbid joint line, as well as premorbid version and inclination ( Fig. 62.1 ). The vault model can be used in the preoperative 3D CT scan to define glenoid bone loss relative to the premorbid anatomy. It is also used in the postoperative 3D CT scan to define the location of the implant relative to the premorbid anatomy, thereby defining how “anatomic” the glenoid implant was placed at the time of surgery. A 3D determination of the center of the humeral head component in relation to the center of the glenoid component and the scapula plane defines this relationship postoperatively. Use of the perfect sphere concept ^, can be used to define anatomic sizing of the prosthetic humeral head relative to the anatomy for that patient.

Glenoid bone loss is common in patients with primary osteoarthritis. In our clinical experience, A1 glenoid morphology with a centered humeral head occurs in approximately 30% of patients undergoing anatomic TSA. An A2 glenoid with central glenoid bone loss occurs in an additional 20% of patients. Asymmetric posterior glenoid bone loss occurs with posterior humeral head translation in the remaining 50% of patients. ^, Very few patients undergoing anatomic TSA have a C1, C2, or D glenoid. Many of the patients with a C glenoid are indicated for a reverse replacement, hemiarthroplasty, or modified TSA.

Asymmetric glenoid bone loss in patients with primary osteoarthritis is managed in our practice with use of anatomic TSA and a posteriorly augmented glenoid component. Our treatment goal is to restore the patient-specific premorbid glenoid joint line. This would require augmentation of the bone deficiency using an augmented implant, with minimal reaming of the anterior glenoid. The surgical goal is to restore the normal location of the joint line in version, inclination, and lateral location of the implant surface ( Fig. 62.2 ). This is best defined by 3D preoperative CT imaging and templating using the vault model to define the premorbid location of the joint line. There are different shaped augmented components currently available in the United States. These include full wedge, half wedge, and stepped glenoid components. In our current practice, we use an implant shape that best matches the shape of the bone loss. In a B3 glenoid a full wedge provides the best match, whereas in a B2 glenoid a half wedge or stepped component provides the best match. The size of the augmentation is selected to best match the severity of the defect.