Clinical results and 10-year survivorship of the Synthes Epoca resurfacing total shoulder arthroplasty

Abstract

Aims

This prospective cohort study reports the 10-year survivorship, patient-reported outcome measures and radiographic outcomes of the first series of the Epoca resurfacing head total shoulder arthroplasty.

Methods

From July 2008 to July 2014, 59 Epoca resurfacing head total shoulder arthroplasties were implanted in 50 patients by a single surgeon. Minimum 4.8 year and maximum 11.3 year follow-up (mean: 7.9 years) was analysed using the latest available Oxford Shoulder Score. Kaplan–Meier survivorship analysis was performed with revision as the end point. Most recent radiographs were reviewed for component radiolucency, osteolysis and proximal humeral migration.

Results

Two shoulders underwent revision (3.4%); one for pain with posterior subluxation following a fall, the other for pain with failure of the glenoid. Kaplan–Meier analysis revealed 10-year survivorship of 98.2% (95% confidence intervals: 88.0%–99.8%). Mean Oxford Shoulder Score improved significantly compared to pre-operative values from 18.2 (5–45) to 46.6 (36–48) (p < 0.001). Fifty-three shoulders underwent radiographic analysis (89.8%). This revealed humeral radiolucency in two cases (3.8%), glenoid radiolucency in three cases (5.7%) and radiographic rotator cuff failure in eight cases (15.1%).

Conclusion

This prospective cohort study shows excellent 10-year survivorship, medium to long-term clinical and radiological results and for the Synthes Epoca resurfacing head total shoulder arthroplasty.

Evidence level

IV.

Introduction

Glenohumeral arthritis is associated with significant morbidity and loss of shoulder function. When conservative management has been exhausted, shoulder arthroplasty can provide pain relief and restoration of function.^1–5 With an intact rotator cuff, anatomical total shoulder arthroplasty (TSA) is the preferred surgical intervention for glenohumeral arthritis.^2,3,6,7

Compared to conventional-stemmed designs, humeral head resurfacing is bone preserving, facilitates more accurate restoration of centre of rotation, has a relatively low risk of peri-prosthetic fracture and allows for less complex revision.^8–10 Biomechanical studies have shown greater capacity to accommodate variation in proximal humeral anatomy with resurfacing arthroplasty.^9,11 When used as a hemiarthroplasty, humeral head resurfacing is associated with a high risk of revision due to excessive glenoid erosion. ¹² Simultaneous resurfacing of the glenoid addresses this issue, however humeral head resurfacing itself has been associated with ‘overstuffing’ of the glenohumeral joint and the additional thickness added by a glenoid component could exacerbate this. The Copeland resurfacing head (RH) is 6 mm thick at the apex, and has been shown to sometimes result in overstuffing even without use of a glenoid implant.^13,14 The Epoca RH TSA is designed with a thin humeral component (∼2 mm) in order to accommodate a glenoid component without increasing the offset and overstuffing the joint. Although newer short-stem prostheses help reduce the risk of overstuffing and improve glenoid access through resection of the humeral head, they require good metaphyseal bone stock for fixation and the amount of bone sacrificed is more than in resurfacing. ¹⁵ The Epoca RH TSA provides the benefits of TSA with minimal bone resection and a reduced risk of overstuffing due to minimal component thickness.

To date, there are no published studies reporting the outcomes of the Synthes Epoca RH TSA. The purpose of this prospective cohort study is to report the 10-year survivorship and medium to long-term patient-reported outcome measures (PROMs) and radiographic results of the first series of the Epoca RH TSA.

Patients and methods

All patients undergoing TSA using the Epoca RH TSA at our institution between July 2008 and July 2014 under the care of the senior author were included in this study. Patients in which the Epoca RH was used as a hemiarthroplasty were excluded. Institutional review board ethical approval was granted for this prospective cohort study.

Indications

Indication for TSA in this study was a painful shoulder joint refractory to conservative management with arthritic changes on plain film radiography. Conservative management prior to TSA included oral analgesia and physiotherapy. All patients had a clinically intact rotator cuff at the time of operation.

Implant

The Epoca RH TSA consists of an uncemented, anatomic, cobalt chromium humeral head resurfacing component lined internally with hydroxyapatite, and a size-matched, ultra-high-molecular-weight all-polyethylene, double-pegged glenoid implant. Fixation of the glenoid implant can be achieved in three ways. It can be used as a standalone implant and cemented directly into the prepared glenoid. It can be used in conjunction with hollow, titanium alloy (TAV) ‘shell screws’, which are screwed into glenoid bone and into which the pegged polyethylene glenoid is cemented. Finally, it can be used together with the cementless, titanium alloy (TAV) ‘metalback’ glenoid, through which cortical screws can be used to augment fixation, and into which the polyethylene glenoid is inserted with a press-fit (Figures 1–3).

Figure 1.

Epoca resurfacing head humeral component.

Figure 3.

(a) AP plain radiograph of the Epoca RH TSA. (b) Axillary plain radiograph of the Epoca RH TSA. AP: anterioposterior; Epoca RH TSA: Epoca resurfacing head total shoulder arthroplasty.

Surgical technique

All procedures were performed by the senior author. The deltopectoral approach was used to access the glenohumeral joint. All long heads of biceps tendons were tenotomised if not already ruptured. Subscapularis was mobilised by lesser tuberosity osteotomy in all cases. The humeral head was then reamed to achieve an anatomic fit of a trial head, which was then applied as a protector. Following capsular release, the humeral head was dislocated posteriorly to permit access to the glenoid. The glenoid was then prepared and a trial glenoid component was inserted. After trialling, the trial glenoid was removed and a definitive component implanted; in patients with soft or cystic bone, the pressfit metalback glenoid was used. In all others, the hybrid glenoid with shell screws or the cemented all polyethylene glenoid was used. Hybrid and cemented glenoids were cemented into position using Palacos (Heraeus Medical, Wehrheim, Germany) bone cement with digital pressurisation. The definitive humeral implant was then impacted onto the prepared humeral head until it was seated flushed to the reamed surface. After confirming soft tissue tension and joint stability, the lesser tuberosity osteotomy was repaired using non-absorbable transosseous sutures. Wound closure was performed and the arm was placed in a polysling. Post-operative rehabilitation was performed following the protocol in Table 1.

Table 1.

Post-operative rehabilitation protocol.

0–4 weeks post-op	Avoid abduction and ER. Avoid hand behind back. Exercises as inpatient until outpatient appointment: Active motion neck, hand and elbow Scapula setting and postural exercises Passive flexion (assisted by relative/carer/cradle/table slides from day 1) Passive ER table slide to the angle specified in op notes Exercises from initial outpatient appointment until 4 weeks post-op. Continue with exercises taught as an inpatient Continue with sling use and respect movement restrictions stated above Progress active assisted flexion exercises to stick assisted flexion and pulley as able
4–12 weeks post-op	Avoid weighted exercises/loading joint. Continue active assisted exercises into flexion and ER, aiming to increase range towards full range of motion Add active assisted abduction May begin active use of the arm in all directions Wean from sling Continue scapula setting and postural exercises as needed Exercise ideas: supine short lever elevation, supine ER, abduction with stick, abduction with pulley, wall slides Consider hydro if struggling Advice on pacing and analgesia May include isometric cuff exercises
12 weeks onwards	Continue with the range of motion work, increasing stretches if needed Add strengthening as needed for patient’s functional needs

ER: external rotation; post-op: post-operative; op: operative.

Outcome measures

The Oxford Shoulder Score (OSS) was used to assess clinical outcomes. Scores were recorded pre-operatively and post-operatively at three and six months, then one, two, five and ten years. Scores from the latest available follow up were compared with pre-operative scores.

The endpoint for survivorship was defined as revision surgery with the exchange of any component. Details of revision or any other procedure performed on included shoulders were collected prospectively and the information cross-checked against National Joint Registry (NJR).

Radiographic evaluation

Radiographs were taken at day one post-operatively, six months, and at each subsequent follow up (1, 2, 5 and 10 years). The most recently available radiograph for each patient was assessed by two independent orthopaedic surgeons (TK and RZ). For shoulders that underwent revision surgery, the latest radiograph prior to revision was reviewed. Both glenoid and humeral components were assessed for radiolucency/osteolysis and implant migration. Radiographic rotator cuff failure was defined as superior migration of the humeral head with acetabularisation of the acromion (Hamada Grade 3). ¹⁶

Statistical analysis

Statistical analysis was performed using the IBM Statistical Package for the Social Services (SPSS) v26. Continuous non-parametric dependant and independent variables were compared with the Wilcoxon signed-rank and Mann–Whitney U tests respectively. A p-value of <0.05 was considered statistically significant. Kaplan–Meier survivorship analysis was performed with joint revision as the endpoint, defined as an operation in which at least one of the components was changed.

Results

Patient demographics

Fifty-nine Epoca RH TSAs were implanted in 50 patients from July 2008 to July 2014 (Figure 2). The mean age of at time of operation was 69 years (range: 38–87). 45 patients (90.0%) were female and five (10.0%) were male. Nine patients (18.0%) had bilateral implants which in all cases were staged, with a mean time interval of 14 months (range: 2–58). Fifty-three shoulders (89.8%) received a ‘shell screw’ glenoid, five (8.5%) a ‘metalback’ glenoid and one (1.7%) a cemented all-polyethylene glenoid. Demographic data of patients undergoing Epoca RH TSA is displayed in Table 2; indications are shown in Table 3. At the time of writing, five patients (10.0%, six shoulders) had died, at a mean of six years (range: 0.33–9) following surgery. No deaths were directly attributed to shoulder arthroplasty. Two patients (4.0%, two shoulders) were still alive but lost to follow-up. Overall, follow up data was available for 56/59 (94.9%) shoulders of the original cohort. Paired pre-operative and post-operative OSSs were available for 48/59 (81.4%) shoulders.

Figure 2.

Epoca resurfacing head glenoid components. From left to right; all-polyethylene cemented glenoid, polyethylene cemented glenoid with titanium alloy (TAV) shell screws, metalback uncemented glenoid.

Table 2.

Patient characteristics of total shoulder arthroplasty (n = 59).

	N	%
Age, mean (range)	69 (38-87) years
Sex
Female	45	90.0
Male	5	10.0
Unilateral/bilateral
Unilateral	41	82.0
Bilateral	9	18.0
Laterality
Left	29	49.2
Right	30	50.8

Table 3.

Indications for total shoulder arthroplasty.

	N	%
Osteoarthritis	51	86.4
Inflammatory arthropathy	5	8.4
Post-traumatic osteoarthritis	3	5.1

Revision and survivorship

Two shoulders (3.4%) underwent revision surgery; one at 5.6 years post-operatively, the other at 10.6 years (mean: 8.1 years) (Table 4). The patient revised at 5.6 years was experiencing pain following a mechanical fall 3 years post-primary procedure, with clinical and radiographic evidence of posterior subluxation of the humeral head. No component loosening was noted intraoperatively at the time of revision. The patient revised at 10.6 years had pain with loosening and failure of the glenoid component and posterior subluxation of the glenohumeral joint on plain film radiography. Both shoulders had initially received a ‘shell screw’ glenoid implant; both were revised to reverse polarity TSA.

Table 4.

Characteristics of revision arthroplasty.

Age	Sex	Time (years)	Reason
55	Female (F)	5.6	Pain post-fall with posterior humeral head subluxation
60	F	10.6	Glenoid loosening and failure

A total of two revisions were therefore included in the survival analysis. Kaplan–Meier analysis revealed 10-year survivorship of 98.2% (95% confidence intervals, 88.0%–99.8%) (Figure 4).

Figure 4.

Kaplan–Meier implant survivorship plot with 95% upper and lower confidence interval.

Oxford shoulder scores

Paired pre-operative and post-operative OSSs were available for a total of 48 shoulders (81.4%), with the latest follow up at a mean duration of 7.9 years post-operatively (range: 4.8–11.3). Shoulders with less than 1-year post-operative OSSs were excluded (two lost to follow-up, one death). For shoulders undergoing revision surgery, the final OSS prior to revision was used. Mean pre-operative OSS was 18.2 (range: 5–45). Mean OSS at six months post-operative was 44.0 (range: 22–48). Mean post-operative OSS at latest follow up had improved significantly compared to pre-operative values by 28.4 points to 46.6 (range: 36–48, p < 0.001) (Figure 5).

Figure 5.

Box Whisker plot showing pre- and post-operative Oxford Shoulder Score with median and mean scores labelled. p-value displayed Wilcoxon signed-rank test.

Complications

Two patients (3.4% shoulders) suffered mechanical falls post-operatively. One fell at five months post-operative and sustained a periprosthetic fracture. This was treated conservatively and, subsequently, healed successfully without the need for further intervention. The other fell at three years post-operative and experienced ongoing pain following the fall. As described previously, this patient subsequently underwent revision at 5.6 years post-primary for pain with posterior subluxation of the humeral head. One shoulder (1.7%) underwent arthroscopic sub-acromial decompression at four years post-procedure due to symptoms of impingement; the symptoms settling following this and latest OSS was 42/48. No cases of infection, iatrogenic nerve injury, or dislocation were found in our cohort.

Radiographic analysis

Post-operative radiographs were available for 53/59 shoulders (89.8%). The mean time from the index procedure to the latest available radiograph was 76.7 months (23–132).

Two of 53 (3.8%) shoulders showed radiolucency of proximal humeral bone. In both cases, loss of bone at the calcar and greater tuberosity was evident, with migration of the humeral component. Neither case experienced any clinical symptoms and had OSSs of 46 and 48 at 10-year follow up. Both of these cases and one additional case (3/53, 5.7%) showed radiographic evidence of glenoid osteolysis. No radiolucent lines related specifically to the pegs of the glenoid components were noted, but loss of glenoid bone stock was evident. The case with isolated glenoid osteolysis was revised at 10.6-years post-primary procedure for symptomatic glenoid component loosening and failure. Radiographs of one of the cases (1.9%) showed migration of one of the glenoid shell screws suggesting a grossly loose glenoid component (Figure 6). The latest OSS of this patient was 46/48. ‘Shell screw’ glenoid components had been implanted in all cases with radiographic evidence of osteolysis.

Figure 6.

Anterioposterior (AP) radiograph showing proximal humerus and glenoid radiolucency with shell screw migration.

Hamada grade 3 proximal humeral migration representing radiographic rotator cuff failure was observed in 8/53 (15.1%) of the shoulders. No revisions were performed on any shoulders with proximal humeral migration. The average age of patients with Hamada grade 3 proximal migration was 65.8 years old at the time of surgery. The average OSS at the latest follow up of all patients with proximal humeral migration was 46/48.

Discussion

Our study has shown excellent long-term survivorship, clinical outcomes and radiographic results for the Epoca RH TSA. Kaplan–Meier analysis revealed survivorship of 100% at 5 years and 98.2% at 10 years (95% confidence intervals: 88.0%–99.8%). Overall revision rate was 3.4% (2/59 shoulders) with a mean time to revision of 8.1 years; one case was revised due to pain post-fall with posterior humeral head subluxation and the other for symptomatic glenoid component loosening and failure.

According to NJR data, the popularity of resurfacing TSA peaked in 2013 at 99 cases (2.3% total TSAs), and its use gradually declined to just 15 cases (0.2%) by 2019. ¹⁷ This decline has coincided with the rise in use of both stemless TSA and reverse polarity TSA, which constituted 12.2% and 52.2%, respectively of all primary shoulder arthroplasty in 2019. Perceived advantages of stemless TSA over resurfacing TSA include more predictable restoration of native offset with reduced chance of ‘overstuffing’ the joint, and better access to the glenoid. ¹⁵ The NJR reports revision rates at five years of 3.66% for resurfacing TSA and 3.54% for stemless TSA. With an overall revision rate of 3.4% in our study, our results compare favourably to those reported in the NJR. Improvement in OSSs also compare favourably to those of the NJR; stemless TSA and resurfacing TSA report an improvement of 20.8 and 20.5 points respectively at six months post-operatively. Our study has shown an improvement of 25 points at six months post-operatively, and of 28.6 points at latest follow up (mean, 7.9 years). ¹⁷

Although there are currently no studies reporting outcomes of the Epoca RH TSA, several have reported good results using other resurfacing TSA systems. Dattani et al. ¹⁸ presented results of resurfacing TSA at the American Academy of Orthopaedic Surgeons (AAOS 2014) annual meeting, showing a significant improvement in OSS from 18.9 pre-operative to 38.9 post-operative at 2–7 years (p < 0.01). Survivorship at 7 years was 100%. A series of Copeland shoulder resurfacings (Copeland Shoulder; Biomet, Swindon, UK) published by Levy et al. ¹⁹ included 14 Copeland TSAs. Constant-Murley Score (CMS) in these cases increased from a mean pre-operative score of 11.5% to 58.1% at latest follow up. Survivorship was 100% at 5 years, 71% at 11 years, 71% at 14 years and 61% at 22 years. Revisions were undertaken for five of these cases; four for component (glenoid and humeral) loosening, and one for rotator cuff failure. A comparison of resurfacing TSA versus stemmed TSA by Glanzmann et al. ²⁰ included 37 Promos resurfacing prostheses (Smith & Nephew Orthopaedics AG, Baar, Switzerland). CMS scores in this series increased from mean 33.4% pre-operatively to 70.6% at 24 months post-operative. One revision was undertaken due to subscapularis tear with anterior escape (2.3%). Our results are similar or better than those presented by both of these studies.

Our results also compare favourably to those of stemless TSA. A recent systematic review of 900 stemless TSAs reported revision rates of between 0% and 8.2%; the prosthesis with 0% revision had only been used in a handful of cases. ²¹ A study of 151 Mathys Affinis Short TSAs by Jordan et al. ²² revealed survivorship of around 95% at nine years with a revision rate of 5.4%. Our 10-year survivorship of 98.2% and revision rate of 3.3% are comparable to that of stemless TSA designs.

Radiographic follow up was available for 53/59 shoulders (89.8%). Humeral component radiolucency was observed in 2/53 shoulders (3.8%). In one experimental study, the Epoca RH has been shown to have satisfactory osseous integration, similar or better than other similarly designed humeral head resurfacings. ²³ The humeral radiolucency seen in both cases was associated with wear of the glenoid component and radiolucency of glenoid bone, suggesting glenoid component wear and osteolysis as the cause of humeral radiolucency. No stress shielding was seen in our study; a known phenomenon in stemmed TSA. A recent study of the Zimmer Biomet Comprehensive Micro System (manufacturer), which uses an uncemented ‘mini’ stem, revealed medial calcar resorption in 23% of cases and greater tuberosity stress shielding in 14% of cases. ²⁴ Our findings suggest resurfacing implants compare favourably to stemmed implants in this regard.

Glenoid loosening is one of the primary modes of failure in TSA and has been reported to occur between 0% and 20% at medium-term follow-up and 39% mid-term to long-term follow-up.^25–30 A study of Copeland resurfacing TSA by Levy et al. ¹⁹ reported radiographic loosening in four of nine (44%) glenoid components with minimum 10-year follow up. In our study, glenoid radiolucency was seen in a total of 3/53 cases (5.7%), with one revision undertaken due to pain with loosening and failure of the glenoid. In this series, choice of glenoid implant was made intraoperatively dependent on findings; where glenoid bone was poor, the uncemented, metalback glenoid was used. In other cases, the hybrid ‘shell screws’ were used in all but one case where the cemented, all-polyethylene glenoid was used. The ability to tailor the method of fixation dependence on glenoid morphology may go some way to explaining the rate of loosening compared to other studies. Additionally, despite the polyethylene glenoid components being size-matched to the humeral component, an over-sized glenoid was frequently used, resulting in a controlled glenohumeral radial mismatch and reduced implant conformity. With modern materials, a theoretical result of this is reduced volumetric polyethylene wear and therefore reduced likelihood of osteolysis and loosening. Due to the larger surface area of the glenoid, the potential for edge-loading and the rocking-horse phenomenon with humeral head translation are also reduced.

Hamada grade 3 proximal humeral migration representing radiographic superior rotator cuff failure was seen in 8/53 (15.1%) shoulders. No shoulders underwent revision surgery directly as a result of proximal migration. There was no significant difference in mean age or mean latest OSS between the groups with and without proximal migration. A systematic review by Levy et al. ³¹ of the outcomes of ten studies reporting proximal humeral migration after TSA or hemiarthroplasty revealed proximal humeral migration in 29.9% of shoulders at mean 6.6 years follow-up, with severe proximal migration reported in 18% cases. The rate reported in our study is comparable to this.

90% of the patients in our study were female. The explanation for this is that access to the glenoid in TSA can be technically difficult; especially so in the case of resurfacing and in the presence of significant muscle bulk or stiff soft tissues. ³² In cases where glenoid access was insufficient, the decision was made intraoperatively to perform a stemmed Epoca TSA instead, allowing access to the glenoid following resection of the humeral head. Hemiarthroplasty was not performed because TSA has been shown to have comparatively superior outcomes. ³³ Although resurfacing TSA has now been largely superseded by short-stem TSA, our study has shown survivorship and outcomes similar to those of short-stem TSA. The availability of resurfacing TSA provides another option in the treatment of glenohumeral osteoarthritis and may prove useful in young patients where bone sparing procedures are of utmost importance.

There are a number of limitations associated with this study. It is a single surgeon series. Only one PROM score was used to evaluate clinical outcomes, and shoulder range of motion was not recorded. It is a cohort study with no control group available for direct comparison. The cohort was 90% female and although favourable results were seen in the five male patients, there may have been a difference in survivorship and outcomes were more males included in the study.

Conclusion

This prospective cohort study of 59 Synthes Epoca RH TSAs shows encouraging medium to long-term results in this predominantly female cohort. Kaplan–Meier survivorship analysis revealed five- and ten-year survivorship of 100% and 98.2%, respectively with a revision rate of 3.4%. There was a statistically significant increase in OSS, from a mean of 18.2 pre-operatively (5–45) to 46.6 at latest follow up (36–48), with a mean latest follow of 7.9 years.