Long term outcomes in shoulder surgery are a challenge for some key reasons:
(1) the longer the minimum followup duration for the study, the greater the percentage of patients who will be lost to followup (because they have moved, no longer wish to provide followup, are grappling with other illnesses, are getting care elsewhere, or have passed away). So over time the final followup cohort becomes progressively less representative of the initial group of patients having the procedure (this is known as attrition bias).
(2) long term radiographic followup is even harder to get in high percentage because of the logistics in obtaining preoperative and followup images that are comparable.
(3) shoulder surgery has changed a lot over the past decade: improved patient selection, better techniques, newer implants, better educated surgeons, and more refined rehabilitation protocols. Many surgeons no longer apply the implants or techniques that they used a decade ago. Thus how indicative are the results from ten years ago of how today's patients will be doing a decade from now?
That having been said, let's look at some 10+ year followup studies on RSA.
There was no difference in the mean SANE score or VAS pain score for patients having midterm (2-5 yr) and those having long-term follow-up (>10 yr).
52 patients were very satisfied, 24 satisfied, 13 dissatisfied, and 4 very dissatisfied. Kaplan-Meier prosthesis survival rate for all 471 RSA patients was 88% at 5 years and 81% at 10 years (revision defined as removal or replacement of metal components).
There were 64 complications in 60 patients; 48 patients had repeat surgery. The figure below is particularly informative regarding the type and timing of complications, noting the diagnoses of component loosening, PJI, acromial stress fractures, and instability continue to be made more than five years after surgery.
In Functional and radiographic outcomes of reverse shoulder arthroplasty with a minimum follow-up of 10 years. the senior author performed 119 Delta Xtend reverse total shoulders between 2005 and 2012. Surgical technique included an inferior overhang of the glenosphere and leaving the subscapularis unrepared.
35 were deceased before reaching the 10-year follow-up and 23 could not be reached. 63 (out of 119) RSAs were included.
10 complications were identified, of which seven required a revision at a median of 3 years. There were no cases with acromial fractures. 4 shoulders were revised for instability, one for loosening, one for infection, and one for periprosthetic fracture.
At final follow-up, the median anterior elevation was 135°; the median level reached with internal rotation was L5. The median Constant score was 68.
Radiographs could be obtained in 25 patients. Among these, scapular notching occurred in 10 patients. Ossification occurred in 10 patients, and stress shielding in 2 patients. Radiolucencies were observed around the humeral component in 24 patients and around the glenoid component in 13 patients.
Here's an example of a periprosthetic fracture
ossification
and loosening
The authors of Clinical outcomes are unchanged after a mean of 12 years after reverse shoulder arthroplasty: a long-term re-evaluation previously reported their outcomes having Delta III (Grammont style) reverse total shoulder arthroplasty between 2003 and 2008. 109 patients fulfilled their inclusion criteria and had a minimum of 5 year followup. Of these, 5 refused to participate in the study, 15 patients were lost to follow-up, and 9 patients died of unrelated causes leaving 80 patients for study. They then attempted to re-evaluate the patients at a minimum of 10 years after surgery: 27 (out of 109) were available for follow-up at 10+ years. 9 refused to participate; 12 patients were lost to follow-up; 14 patients died from causes unrelated to the prosthetic implant. Neither the mean range of motion or the Constant score were different for the 27 at 10 years and the 80 at 5 years. No loosening of implants was noted, and the rate of scapular notching was 66%, mostly grade 1 or 2.
Primary reverse shoulder arthroplasty: how did medialized and glenoid-based lateralized style prostheses compare at 10 years? compared 10 year outcomes for 56 Grammont style medialized prosthesis and 44 glenoid lateralized prosthesis41 (out of 100) patients had an average of 10.2 years' follow-up showed clinical improvements without significant differences between the two groups.
There were 16 complications; reoperation was required in 6 shoulders. Notching rates were significantly higher in the medialized group (77% in M group vs. 47% in lateralized group). The reasons for reoperation were dislocation (2), polyethylene disassociation (1), glenosphere disassociation (1), infection (1), and acromial stress fracture (1). Other complications included intraoperative or post- operative periprosthetic fractures (6), acromial stress fractures (2), brachial plexopathy (1), and a distal clavicle insufficiency fracture.
One or more complications were recorded in 13 patients (59%): mechanical block (1), scapula fracture (3), humeral fracture (1), dislocation (3), glenoid component loosening (2), humeral component loosening (2), polyethylene wear (1), infection (6). 12 required reoperation, 6 RSAs failed.
Their original report included the outcomes for 186 patients (191 RSAs) who had been followed for a mean of 40 months. In the present study, in which the mean duration of follow-up was 150 months, follow-up clinical evaluations were available for 84 patients (87 (out of 191) prostheses) and radiographic assessments were available for 64 patients (67 prostheses).
Seventy-seven patients (79 prostheses) had died before the 10-year follow-up, and 17 patients (17 prostheses) had been lost to follow-up.
Constant scores decreased significantly compared with the scores at the medium-term follow-up evaluation (at a minimum of 2 years). Forty-nine shoulders (73%) exhibited scapular notching. Forty-seven complications (29%) were recorded, with 10 cases (10%) occurring after 2 years. Sixteen (12%) of the original patients underwent revision surgery. The 10-year overall prosthetic survival rate using revision as the end point was 93%.
Comment: These long-term followup studies are important. They document that clinically significant improvement in shoulder comfort and function can be sustained a decade after reverse total shoulder.
The authors of these studies endeavored to capture 10+ year outcomes for the highest possible percentage followup for their reverse total shoulders; yet data on only 25% to 50% of the cases were attainable. How representative is this sample of the overall results: are patients with followup likely to have better or worse outcomes than those lost to followup? Out of a cohort of patients having RSA, which patients would be most likely to make themselves available for 10 year clinical and radiographic followup?
The implants available for RSA a decade ago are no longer commonly used. Techniques have changed as has preoperative planning. How well do these 10+ outcomes from surgeries done a decade ago predict the 10+ year outcomes of the RSAs being done today? Will the complications be less frequent?
Many of the patients having 10+ follow up after RSA were older and had their procedure performed for the original indication: rotator cuff tear arthropathy. Now RSAs are being increasingly performed for younger patients and for those with other diagnoses, including osteoarthritis with an intact cuff - a diagnosis conventionally treated with an anatomic arthroplasty (aTSA). In determining the comparative value of RSA to aTSA, clinical investigators will need to stratify patients by diagnosis as well as age and sex.
These articles document that complications continued to accrue well after first two to five years post reverse arthroplasty. The complications include dislocation, glenoid loosening, humeral loosening, glenoid fracture, humeral fracture, inadequate seating, polyethylene wear, acromial stress fractures, infection, nerve injuries, and scapular notching. Thus long term studies are important in pointing out that patients remain at risk for complications a decade after surgery.
Authors have presented "survival rates" as a quality outcome metric. "Survival" is commonly defined as retention of the implant. With the RSA, implants may be retained even if there is a debilitating complication, such as dislocation, displaced acromial stress fracture, or neurological injury. Implant retention may not indicate clinical success.
The question about whether the comfort, function and radiographic appearance change over time cannot be addressed by comparing, for example, the mean Constant scores on the 50 patients that were available for followup at 2-5 years with the mean Constant scores for the 25 patients that were available for followup at 10 years. In this example it may be better to take the 25 patients with 10 year data and look at the data for each of these patients at 1, 2 and 5 years after surgery.
A last point is that because of the importance of including the maximum percentage of the original group of patients having RSA in the final followup, we need to make it as easy as possible for the patient to continue to participate in the clinical followup program. Thus the program needs to include (1) explaining the importance of followup to all patients having the procedure and consenting them prospectively for long term followup, (2) using an outcome metric that is short and equally valid whether it is completed in the surgeon's office, electronically or by postal mail (i.e. one that doesn't require travel for a physical examination, but rather one that is based on the patient's self-assessed comfort and function), (3) obtaining a preoperative baseline assessment, and (4) organizing a system for reminding patients to submit their assessments of their shoulder at the desired intervals after surgery. Our preference for the Simple Shoulder Test lies in the facts that it meets the above criteria and provides data on individual shoulder functions of importance to the patient, rather a single number composite score based on weighting of different elements that may be of different importance to different patients.
You can support cutting edge shoulder research and education that are leading to better care for patients with shoulder problems, click on this link.
