Anatomic versus Reverse, the Debate Continues

There can be no doubt that total shoulder arthroplasty (TSA) is a successful operation that improves the lives of patients. Shoulder arthroplasty is increasing in frequency due in part to its success as well as the introduction of the reverse shoulder replacement (rTSA), which offers solutions to problems without previous good options. In fact, rTSA is now performed more frequently than anatomic TSA (aTSA), and sometimes the indications for one versus the other (previously quite distinct) may overlap, leading to debates about optimal implant choice. Unfortunately, with any surgery comes complications which can certainly impact outcomes and negatively effect quality of life. The authors of this study, Drs. Parada et al, sought to determine the types of failure modes of shoulder replacement and the differences in these between aTSA and rTSA.

This study is a retrospective cohort comparison using an international database of patients who received a single-platform shoulder arthroplasty system. Patients were treated by 40 different fellowship-trained surgeons and 2224 aTSAs and 4158 rTSAs were included, while revisions were excluded. These were reviewed for the incidence, type, and timing of complications. Complications and revisions were separately analyzed. Overall, the mean age of aTSA patients was significantly younger than rTSA, as would be expected. For aTSA, the overall complication rate was 10.7% with a 5.6% revision rate. The most common complications were rotator cuff tear/subscapularis failure in 3.1%, aseptic glenoid loosening in 2.5%, infection in 1.3%, and pain in 1.1%. For aTSA, the most common reasons for revision were aseptic glenoid loosening, rotator cuff/subscapularis failure, and infection. For rTSA, the overall complication rate was 8.9% with a 2.5% revision rate. The most common complications were acromial/scapular stress fractures in 1.7%, instability in 1.4%, pain in 1.2%, and infection in 0.9%. The most common reasons for revision were instability, infection, and aseptic glenoid loosening.

When compared head to head, rTSA had significantly fewer complications (aTSA 10.7% v rTSA 8.9%, p=0.0434), due to significantly fewer rates of aseptic glenoid loosening despite higher rates of instability and humeral fracture. The overall revision rate was also significantly less in rTSA (aTSA 5.6% v rTSA 2.5%, p<0.0001). There was a significantly lower rate of aseptic glenoid loosening but a higher rate of instability.

The authors are to be commended for this study which adds greatly to the knowledge base on the incidence and types of shoulder arthroplasty complications. This is very helpful when discussing the potential risks and benefits of shoulder arthroplasty with patients. The authors note in their discussion that the complication rates in this paper compare very favorably with older reports of high complication rates with rTSA, in part due to increased surgeon experience and improved implant design.

Historically, rTSA was introduced as a treatment option for patients with rotator cuff tear arthropathy, or end-stage osteoarthritis combined with an irreparable rotator cuff tear as aTSA in these patients is contraindicated due to high failure rates. With its great success, indications for rTSA have expanded to include revision shoulder arthroplasty, proximal humeral fractures and fracture/dislocations, inflammatory arthropathies, and even patients with an intact rotator cuff but advanced glenoid wear or in elderly patients. It is generally held that a well-performed aTSA with an intact rotator cuff will outperform a rTSA, but subscapularis healing does not always occur and rotator cuff tears develop even without trauma as the shoulder (and patient) ages.

In some clinical settings, aTSA and rTSA are both a viable option for the patient and understanding the different complication rates and their frequencies will help surgeons better counsel their patients to determine the appropriate implant. Predictive analytics, utilizing machine learning and large patient databases such as this, may further help with this decision-making. Predict+ is such a software. Predict+ utilizes a few simple questions regarding the patient’s pain, their demographics, diagnoses, comorbidities, and range of motion to apply algorithms tested through machine learning to “predict” the patient’s expected improvement in range of motion and function with aTSA and rTSA and provides their predicted complication rates and types. This, in effect, personalizes the results of this paper to an individual patient. This information, combined with preoperative planning and intraoperative navigation to optimize implant placement, can lead to reproducibly improving outcomes and reducing complications. The future of shoulder arthroplasty may be upon us.