These authors used their shoulder arthroplasty registry to identify 471 patients having reverse shoulder arthroplasty (RSA) prior to 12/31/2010 performed by an individual high volume surgeon. In this sample of 471 patients, only 25 had completed the 10-year follow-up in the clinic. The authors searched national, state, and local mortality indexes and identified 225 patients as deceased, resulting in 221 patients needing follow-up. Each patient was called at their last known phone number. Since this follow-up was at least 10 years after RSA, many patients had moved or changed contact information since their last clinical visit. They used local and state databases to try and locate patients for follow-up when the contact information was incorrect. At least 3 attempts were made to contact each patient before they were considered lost to follow-up. Upon successful contact with a patient, the patient was asked if they could return to the clinic. A standardized phone questionnaire was administered if the patient was unable to return to clinic.
With this extraordinary vigorous effort, 10 year followup was obtained for 94 patients (20%)
The patients were 63% female with an average age of 66±10 years at time of RSA. There were 71 (75%) primary RSAs and 23 (25%) revision RSAs.
The most common indication for primary RSA was cuff tear arthropathy (77%) followed by proximal humerus fracture (13%).
The prostheses used were Grammont-design with a medialized center of rotation and a 155 degree neck-shaft angle.
Patient-reported outcome measures at 10-years or more included a current pain rating averaging 2±3 on a 0 to 10 scale and Single Assessment Numeric Evaluation (SANE) of 73±28.
The SANE and pain scores changed by less than the MCID or improved in 87 and 90% of patients.
56% were very satisfied, 26% satisfied, 14%, dissatisfied, and 4% very dissatisfied.
All revisions after the initial procedure were included in the survivorship analysis for the first model and a second model excluded revisions due to trauma (i.e., fall or motor vehicle accident). Isolated polyethylene exchanges were not considered revisions.
There were 64 complications in 60 patients after at least 10-years of follow-up (complication rate of 64%), and 48 patients (51%) required either a reoperation or component revision.
In the primary RSA group, 41 of 71 patients (58%) had postoperative complications; 4 patients had multiple complications. They were, in order of frequency, dislocation (11), periprosthetic infection (PJI) (11), periprosthetic fracture (7), acromial stress fracture/reaction (5), aseptic baseplate loosening (4), subjective instability(defined as patients who felt that their shoulder was “clicking” or “near dislocating,” and whose symptoms improved after a polyethylene exchange) (2), traumatic glenoid loosening (1), early superficial wound infection (1), hematoma (1), symptomatic intra-articular loose body (1), and axillary nerve palsy (resolved at 6 months) (1).
The percentage of patients who experienced complications that occurred during the 0-6 month, 6 month to 2 year, 2-5 year, 5-10 year and after 10 year timepoints for primary RSA were 24%, 16%, 21%, 32%, and 8%, respectively.
There were 18 component revisions after primary RSA (25%).
To treat PJI, 4 patients underwent two-stage revision and 6 patients underwent resection arthroplasty. Four periprosthetic fractures required open reduction and internal fixation (ORIF) and revision to a long stem prosthesis.
Three cases of aseptic baseplate loosening underwent revision of the glenoid component with iliac crest autograft, and one patient refused further revision.
The case of traumatic glenoid loosening required baseplate revision.
There were 10 closed reductions and polyethylene/glenosphere exchanges after dislocations, 2 polyethylene exchanges for subjective instability, 1 removal of an intra-articular loose body and 1 bearing exchange for the wound infection.
In the revision RSA group, 19 of 23 (83%) patients had postoperative complications. These included PJI (6), dislocations (5), dislocations followed by diagnosis of PJI (3), periprosthetic humerus fracture (2), aseptic baseplate loosening (2), and aseptic humeral loosening (1).
Twelve (53%) patients in the revision RSA group required revision. Of the 9 patients with PJI, 6 underwent resection arthroplasty, 1 underwent two-stage revision and 2 were planned to undergo 2-stage revision but later elected to live with a spacer.
The two cases of aseptic baseplate loosening underwent baseplate revision with iliac crest autograft, and the one case of aseptic humeral loosening underwent revision with a proximal humerus replacement prosthesis.
In addition, there were 5 polyethylene exchanges for dislocation and 2 ORIF with stem revision of periprosthetic humeral shaft fractures.
The percentage of patients who experienced complications that occurred during the 0–6-month, 6 month to 2 year, 2-5 year, and 5-10 year periods after revision RSA were 47%, 32%, 11%, and 11%, respectively.
Among the 30 patients (32%) who required a revision or reoperation, 14 patients (47%) required another operation during our follow-up period. In the primary and revision RSA groups, multiple reoperations or revisions were required in 8 (11%) and 6 (26%) of patients, respectively.
The overall 5- and 10-year prosthesis survival rate was 88% and 81%, respectively.
Comment: This is a well-done and candid study. It demonstrates a progressive decrement in survival over the decade after the RSA procedure. It also demonstrates some of the key challenges in a 10 year followup study
(1) the surgeon's indications, patient selection, technique, implant selection, surgical skills and rehabilitation protocols prior to December 2010 are likely to be quite different than what they are today. Thus these outcomes, complication rates, and revision rates may not be directly applicable to the current practice.
(2) in spite of a most vigorous effort, a 80% of patients were not included in the followup due to death or inability to contact. The biasing effect of this loss to followup is difficult to assess. For more on non-response bias see this link.
(3) radiographic followup was lacking
(4) although cuff tear arthropathy and fracture were the most common diagnoses in this study, currently RSA is being applied to other diagnoses - most particularly to osteoarthritis with an intact cuff. This study does not provide data on which to base a comparison of anatomic TSA with RSA for this diagnosis.
(5) while newer implants and techniques may seem promising, it is apparent that this type of vigorous followup over a decade or longer is necessary to determine if they yield added value to the patient.
It seems likely that the survivorship of this surgeon's RSAs done after 12/31/2020 will be better - even at two years - than what is shown in the graph above. Perhaps his indications, technique and prosthesis choices may also be different.
Follow on twitter: https://twitter.com/shoulderarth
Follow on facebook: https://www.facebook.com/frederick.matsen
Follow on LinkedIn: https://www.linkedin.com/in/rick-matsen-88b1a8133/
How you can support research in shoulder surgery Click on this link.
Here are some videos that are of shoulder interest
