Are patients having shoulder arthroplasty doing better over the last decade, if so, why?

This post examines two questions of importance to the surgeons and patients interested in shoulder arthroplasty:

(1) Over the past decade, is there evidence that patient-reported outcomes (PROs) after shoulder arthroplasty have gotten better by a clinically significant amount — that is, by at least the minimal clinically important difference (MCID)?

(2) If they have, is that improvement associated with the surgeon, the technology, the shoulder, or the patient?

Based on the available evidence, the answers are: (1) yes, modestly; and (2) the gain is not attributable to new technology or to any change in the shoulders we treat. The largest lever on the outcome is the patient, while the decade’s trend itself most plausibly reflects accumulated surgical experience and care.

A word on the measure. Surrogates — such as radiographic component position, glenoid version, three-dimensional planning accuracy, and data on implant survivorship drawn from registries — are not the outcomes of primary interest. The important outcome is what the patients report. A technology that improves a surrogate but does not move a PRO beyond its MCID has not improved the outcome for the patient.

Now the detail.

1. Are patient-reported outcomes improving over time?

Anatomic TSA (aTSA). In a single-institution series of 1,899 arthroplasties (2008–2018), the 5-year ASES improvement rose by 1.65 points per year of index surgery on univariable analysis and by 2.20 points per year after adjustment for patient factors [1]. The per-year increment is small — well under one MCID — and reaches clinical significance only when accumulated across the decade. The authors could not attribute the gain to any single factor [1].

Reverse TSA (RSA). Some of the most relevant data come from the U.K. National Joint Registry (24,411 RSAs, 2013–2021): mean 6-month Oxford Shoulder Score (OSS) improvement rose from 15.84 to 20.29 — about half a point per year, roughly one MCID across the whole period [2].

In the Danish Shoulder Arthroplasty Registry (2,867 arthroplasties for osteoarthritis, 2006–2015), the adjusted WOOS score rose by about 10 points [3] — short of the WOOS MCID of 12.3 [16].

So.....both aTSA and RSA show measured, year-over-year PRO improvement. The increments are small per year and reach clinically meaningful magnitude only when accumulated across the decade — about one MCID for RSA across the period, and a larger accumulated difference for aTSA, whose adjusted per-year gain compounds over ten years. These are population- and cohort-level shifts: a patient operated late in the decade does modestly better, on average, than one operated early.

Survivorship bias. A patient-reported outcome plotted against time after surgery is available only for survivors: to contribute a 6-month OSS or a 5-year ASES, a patient must be alive, unrevised, and reachable. Those who die, are revised, or move their care to another surgeon before the minimum follow-up never enter the average — which is therefore selected for better outcomes.

Which way this biases the trend is not obvious. Perioperative mortality and early reoperation  fell across this same decade [2], so fewer patients were removed by death in the later years; if anything the later cohorts are less selected, and across-year survivorship is unlikely to be what produced the rising trend. The selection that matters is not across calendar years but across follow-up length — the late-timepoint averages rest on progressively smaller, more selected subsets of the original cohorts.

Implant durability is vulnerable to the same problem in a more specific way. When revision is estimated with a standard Kaplan–Meier survival curve, a patient who dies is treated as still at risk of a revision that can no longer occur. This overestimates the cumulative revision rate, and it overestimates most where mortality is highest — the older patients who receive reverse  replacements [15]. Comparisons of durability across groups with different mortality are confounded for the same reason: part of any apparent difference reflects who lived long enough to ever reach a revision, not how the implant performed. Revision rates in older, higher-mortality groups should accordingly be read as upper bounds rather than firm figures [15].

The usual adjustments for age, comorbidity, and surgeon volume operate only among the patients who actually reported a score; they cannot recover those missing because they died or never returned. So even where adjustment strengthens the trend [1], survivorship bias remains in the data. It matters most where follow-up completion is low and uneven, as it was in the National Joint Registry [2].

So two cautions in reading these data. The within-period findings — smoking [9], unemployment and education [10], resilience and mental health [11], and the absence of any effect of glenoid shape [5] or technology [4] — compare patients operated on in the same period, so the across-year attrition that inflates the temporal trend cannot account for them. The year-over-year trend is the vulnerable part: surgeon experience, technique, rehabilitation, and perioperative care all improved over the same years, so a dataset of this kind cannot apportion the gain among them; the measured improvement is best taken as the most that can be claimed, not a precise figure.

2. The surgeon, the technology, the shoulder, or the patient — what is the change associated with?

The surgeon. Surgeon volume and fellowship training have a well-documented effect on revision, reoperation, complications, and length of stay — and a weak-to-absent effect on the size of the PRO improvement patients realize. In the National Joint Registry, surgeons averaging at least 10.4 shoulder replacements a year had lower revision risk (the hazard of revision roughly doubled below that threshold) along with fewer reoperations, fewer serious adverse events, and shorter stays, but no patient-reported outcome was assessed [12]. Fellowship training shows the same pattern: shoulder-and-elbow and sports-medicine fellowship-trained surgeons have significantly lower complication rates at 90 days, 1 year, and 5 years, again with no PRO assessed [14]. The most direct test comes from the Australian registry: a surgeon’s two-year revision rate showed no clinically relevant correlation with PROMs, and all surgeons produced similar postoperative Oxford scores [13]. Revision rates and PROs measure two different facets of the result — surgeon factors move the first; patient factors move the second.

The technology. Advanced technology does not appear to be related to patient-reported outcomes. A systematic assessment of three-dimensional planning, patient-specific instrumentation, navigation, stemless humeral components, and augmented glenoid components found that no individual technology was associated with a statistically or clinically significant improvement in PROs [4]. A single-surgeon consecutive series of 389 anatomic arthroplasties reaches the same conclusion from the other side: implant choice and surgical technology do not predict outcome [5]. The durable result there came from fundamentals — subscapularis management, a conservative neck cut, a securely seated all-polyethylene glenoid, soft-tissue balancing — not new technology.

Cross-linked polyethylene is the one material change with an apparent benefit: in the Australian registry it had a lower anatomic revision rate than conventional polyethylene [6]. But the benefit is on durability — reduced wear-particle osteolysis and loosening — not on comfort and function; no study has shown that it improves PROs. And the advantage appears smaller in contemporary practice, where glenoid component design and fixation, not polyethylene type, emerge as the dominant predictor of revision [7].

The shoulder. Glenoid morphology — the usual proxy for a “worse” arthritic shoulder — does not predict outcome: in a 389-case series, B2 and B3 glenoids met or exceeded the outcomes of other glenoid types, and symptomatic glenoid loosening occurred in 1 of 389 shoulders [5]. Nor has the case mix drifted toward milder disease. In the Australian registry, the reverse diagnosis mix barely changed across the decade — osteoarthritis fell only slightly as a share of reverse procedures (48.2% to 44.7%) while cuff arthropathy rose by almost the same margin (36.2% to 39.7%) — if anything a drift toward classic cuff-deficient disease, the opposite of an “easier shoulders” explanation [8]. A shift in shoulder characteristics does not appear to be driving the PRO gain.

The patient. The data point to the patient — and to the judgment applied in selecting and preparing the patient. Three patient-side domains stand out. Modifiable risk: in an analysis of 14,465 arthroplasties, smoking was independently associated with increased surgical complications [9]; cessation, along with glycemic, nutritional, and opioid optimization, appears to improve the result. Socioeconomic context: in a nationwide cohort of 2,292 arthroplasties, after adjustment for age, sex, diagnosis, implant, and comorbidity, unemployment was associated with a WOOS deficit of roughly 14 to 19 points at one year and lower educational attainment with a further 5 to 8 points [10] — differences that exceed the MCID. Resilience and mental health: in 399 anatomic arthroplasties, greater resilience and better mental health were associated with better outcomes [11].

So among the candidate factors, the patient is the strongest predictor of the outcome a given arthroplasty achieves — selection, optimization, socioeconomic context, and resilience each move the PRO, several beyond the MCID. Whether the patient mix shifted across the decade in a way that would explain the time trend is a separate question, and these data do not show that it did.

Conclusion

The evidence answers both opening questions.

First, patient-reported outcomes after shoulder arthroplasty have improved over the past decade — modestly, and by an amount that reaches the MCID only when accumulated across the period, not within any single year.

Second, two different questions hide inside "what is the change associated with," and they have different answers.

(A) The decade-long trend is real but cannot be apportioned by the available data; it is most plausibly the product of accumulated surgeon experience, refinements in technique, rehabilitation, and perioperative care. The case mix did not drift toward easier shoulders [8], and no individual technology was associated with the improved outcomes over the decade [4].

(B) The other question — what determines the result for an individual patient — has a clearer answer. The available evidence has not shown a patient-reported benefit from any technology assessed, nor from any change in the shoulders we treat. Surgeon volume and training affect revision and complication rates, but not, as far as the data show, the patient-reported result. That leaves the patient: selection, optimization of modifiable risk, socioeconomic context, and resilience.

The greatest demonstrated effect on the outcome therefore lies in choosing and preparing the patient; the experience the surgeon brings to the procedure most likely matters too, though these data cannot show by how much.

Market forces are clearly driving orthopaedic companies to develop new implants and technologies that strengthen their competitive position, even though evidence that these have a major impact on patient outcomes is lacking.

What this analysis does offer, though, is actionable intelligence: the surest routes to improving outcomes for our patients are (1) thoughtful patient selection and optimization and (2) education that enables surgeons to improve their craft of shoulder arthroplasty.

Will the future be better?

Broad-billed Hummingbird

Tucson

Follow on twitter/X: https://x.com/RickMatsen
Follow on facebook: https://www.facebook.com/shoulder.arthritis

Follow on LinkedIn: https://www.linkedin.com/in/rick-matsen-88b1a8133//

References

[1] Mathew JI, Nicholson AD, Finocchiaro A, Okeke L, Dines DM, Dines JS, Taylor SA, Warren RF, Gulotta LV. Outcomes of shoulder arthroplasty by year of index procedure: are we getting better? J Shoulder Elbow Surg. 2022;31(2):245-251. doi:10.1016/j.jse.2021.08.024. PMID 34592407.

[2] O’Malley O, Davies A, Taghavi Azar Sharabiani M, Rangan A, Sabharwal S, Reilly P. Are we getting better over time? Clinical and patient-reported outcomes for reverse shoulder arthroplasty: a National Joint Registry cohort study. BMJ Open. 2025;15(9):e096084. doi:10.1136/bmjopen-2024-096084. PMID 40953878.

[3] Rasmussen JV, Amundsen A, Sørensen AKB, Klausen TW, Jakobsen J, Jensen SL, Olsen BS. Increased use of total shoulder arthroplasty for osteoarthritis and improved patient-reported outcome in Denmark, 2006-2015: a nationwide cohort study from the Danish Shoulder Arthroplasty Registry. Acta Orthop. 2019;90(5):489-494. doi:10.1080/17453674.2019.1633759. PMID 31240980.

[4] Schiffman CJ, Prabhakar P, Hsu JE, Shaffer ML, Miljacic L, Matsen FA 3rd. Assessing the value to the patient of new technologies in anatomic total shoulder arthroplasty. J Bone Joint Surg Am. 2021;103(9):761-770. doi:10.2106/JBJS.20.01853. PMID 33587515.

[5] Schiffman CJ, Chin, Whitson AJ, Hsu JE, Matsen FA 3rd. Anatomic shoulder arthroplasty: a single surgeon’s consecutive series of 458 patients (389 anatomic TSAs, mean 9.2-year follow-up). University of Washington SHORE Registry. [Manuscript under review.]

[6] Page RS, Alder-Price AC, Rainbird S, Graves SE, de Steiger RN, Peng Y, Holder C, Lorimer MF, Gill SD. Reduced revision rates in total shoulder arthroplasty with crosslinked polyethylene: results from the Australian Orthopaedic Association National Joint Replacement Registry. Clin Orthop Relat Res. 2022;480(10):1940-1949. doi:10.1097/CORR.0000000000002293. PMID 35901440.

[7] Gill DRJ, Corfield S, Harries D, Page RS. Modeling highly crosslinked polyethylene vs. non-highly crosslinked polyethylene glenoid revision rates for anatomic shoulder arthroplasty in osteoarthritis including differing polyethylene glenoid fixation designs. Semin Arthroplasty JSES. 2024;34(4):843-853. doi:10.1053/j.sart.2024.06.003.

[8] Australian Orthopaedic Association National Joint Replacement Registry (AOANJRR). Hip, Knee & Shoulder Arthroplasty: 2025 Annual Report. Adelaide: Australian Orthopaedic Association; 2025.

[9] Althoff AD, Reeves RA, Traven SA, Wilson JM, Woolf SK, Slone HS. Smoking is associated with increased surgical complications following total shoulder arthroplasty: an analysis of 14,465 patients. J Shoulder Elbow Surg. 2020;29(3):491-496. doi:10.1016/j.jse.2019.07.012. PMID 31519425.

[10] Jensen ML, Valsamis EM, Madrid AS, Olsen BS, Rasmussen JV. Association between socioeconomic status and patient-reported outcome at 1 year after shoulder arthroplasty for osteoarthritis or cuff-tear arthropathy: a nationwide cohort study of 2,292 arthroplasties. Acta Orthop. 2025;96:45-51. doi:10.2340/17453674.2024.42700.

[11] Levins JG, Dasari SP, Quinlan NJ, Whitson AJ, Matsen FA 3rd, Hsu JE. Anatomic shoulder arthroplasty: the correlation between patient resilience, mental health, and outcome. J Shoulder Elbow Surg. 2024;33(6S):S9-S15. doi:10.1016/j.jse.2024.03.008. PMID 38548096.

[12] Valsamis EM, Collins GS, Pinedo-Villanueva R, Whitehouse MR, Rangan A, Sayers A, Rees JL. Association between surgeon volume and patient outcomes after elective shoulder replacement surgery using data from the National Joint Registry and Hospital Episode Statistics for England: population based cohort study. BMJ. 2023;381:e075355. doi:10.1136/bmj-2023-075355. PMID 37343999.

[13] Hoskins W, Bingham R, Corfield S, Harries D, Harris IA, Vince KG. Do the revision rates of arthroplasty surgeons correlate with postoperative patient-reported outcome measure scores? A study from the Australian Orthopaedic Association National Joint Replacement Registry. Clin Orthop Relat Res. 2024;482(1):98-112. doi:10.1097/CORR.0000000000002737. PMID 37339166.

[14] Harkin WE, Saad Berreta R, Turkmani A, Williams T, Scanaliato JP, McCormick JR, Nicholson GP, Garrigues GE. How fellowship training affects complication rate after shoulder arthroplasty: a nationwide assessment. J Shoulder Elbow Surg. 2025;34(2):499-506. doi:10.1016/j.jse.2024.05.014. PMID 38944372.

[15] Lacny S, Wilson T, Clement F, Roberts DJ, Faris PD, Ghali WA, Marshall DA. Kaplan-Meier survival analysis overestimates the risk of revision arthroplasty: a meta-analysis. Clin Orthop Relat Res. 2015;473(11):3431-3442. doi:10.1007/s11999-015-4235-8. PMID 25804881.

[16] Nyring MRK, Olsen BS, Amundsen A, Rasmussen JV. Minimal clinically important differences (MCID) for the Western Ontario Osteoarthritis of the Shoulder Index (WOOS) and the Oxford Shoulder Score (OSS). Patient Relat Outcome Meas. 2021;12:299-306. doi:10.2147/PROM.S316920. PMID 34588833.

Why are surgeons performing reverse total shoulders rather than anatomic arthroplasties for cuff intact arthritis?

 In the prior post we reviewed the evidence that - in contrast to reverse arthroplasty (RSA) - anatomic total shoulder arthroplasty (aTSA) is less expensive and provides patients with better comfort and function, fewer serious complications, and safer revision options for complications should they occur.

Paradoxically, however, the proportion of aTSAs being performed for cuff-intact arthritis is dropping precipitously.  

We can speculate on possible reasons for this paradox.

(1) Surgeons may perceive that a lower revision rate for RSA is a positive factor for the patient, when in fact the lower RSA revision rate is due in large part to the fact that some of the most common and serious RSA complications are often not revisable (e.g. pain and poor function, displaced acromial/spine fractures). 

(2) Surgeons may perceive that the RSA is easier to perform. This is, of course, due to the fact that few surgeons have training/experience in performing a basic aTSA, not that the operation is of itself more challenging.

(3) Industry influence and conflicts of interest preferentially motivate the more expensive/profitable RSA option.

(4) Recent "innovations" targeting the use of preoperative planning to achieve high levels of "accuracy and precision" - that may be clinically irrelevant - can make the aTSA unnecessarily complex, expensive and daunting. 

The solution may lie in assuring that shoulder surgeons are well trained in both aTSA and RSA.  This requires that organizations such as AAOS and ASES provide hands-on educational opportunities and that training programs assure that their fellows and residents have a meaningful experience in both.  Interestingly in our most recent round of interviews for our fellowship, a number of applicants reported they had never seen, much less performed, an aTSA.

Below is the basic approach I use for anatomic arthroplasty presented at the amazing Nice Shoulder Course of Pascal Boileau.  These steps may be helpful for surgeons wishing to build their aTSA skills.

Keeping it simple

House Finch

Matsen yard

Follow on twitter/X: https://x.com/RickMatsen
Follow on facebook: https://www.facebook.com/shoulder.arthritis
Follow on LinkedIn: https://www.linkedin.com/in/rick-matsen-88b1a8133/

aTSA vs RSA for cuff intact arthritis: what is the evidence that informs the choice for each patient?

Both anatomic total shoulder arthroplasty (aTSA) and reverse shoulder arthroplasty (RSA) are considerations for patients with cuff-intact glenohumeral osteoarthritis. Currently many, if not most shoulder surgeons are trending toward RSA. In fact many surgeons have little working experience in performing aTSA. For example, per the Australian registry, the share of primary total shoulder replacements that is anatomic has collapsed from roughly 57% in 2008 to about 4% by 2024, while stemmed reverse has risen to nearly 90% (see figure below) [1]. In some circles the pro RSA argument is based on the contentions that (1) the RSA is easier to perform by less experienced-as well as experienced-surgeons and (2) the RSA as a lower rate of revision.

A look at the published evidence may inform the choice for patients and surgeons:

1. Patient-reported outcomes

With commonly used scores, the two types of arthroplasty seem similar for cuff-intact arthritis. A 2026 meta-analysis of 1,716 patients aged ≥70 with a competent cuff found no significant differences in ASES, Constant, or SST scores [2]. A meta-analysis of 14 studies (4,819 cases) found similar ASES, Constant, SST, SSV, and VAS pain scores [3], as did an earlier systematic review [4] and a propensity score–matched JBJS analysis [5]. 

However, in the UK National Joint Registry, roughly a quarter of 21,918 RSA patients had an “unsatisfactory” Oxford Shoulder Score (<29) [6]. Single-center series using patient-acceptable-symptom-state (PASS) thresholds put the figure higher — 25–40% of RSA patients failed to reach PASS for ASES or SANE at two years [7], and 34–35% still failed at minimum five years [8], with pain the primary factor in these adverse outcomes. 

With the Shoulder Arthroplasty Smart score, aTSA achieved higher absolute postoperative scores even though the improvement from baseline was similar [9]. Among patients who reached a “new normal” (defined as a SANE score  ≥95), aTSA significantly outperformed RSA on higher-demand tasks, motion, and return to sport and work [10]. 

2. Motion 

Across the comparative meta-analyses, aTSA delivered better external and internal rotation, with differences that exceed the MCID and reached 10–11° of external rotation in pooled estimates [2,3,4] as well as  better overall motion in matched cohorts [5]. Rotation enables the patient to perform basic activities of daily living: dressing, toileting, perineal care, and reaching behind the back.

3. Complications

In pooled comparative data, RSA carried a lower overall complication rate than aTSA in the cuff-intact population [2,3]. But the complications the two implants produced differed in kind. Historically, primary stemmed anatomic shoulders done for osteoarthritis using legacy techniques and implants have been revised chiefly for glenoid component loosening (29.1%), rotator cuff insufficiency (27.6%), and instability/dislocation (23.1%), with loosening being predominant [1]. Stemmed reverse shoulders have been revised chiefly for instability/dislocation, infection, loosening, and fracture [1]. 

4. Revision 

At ten years, cumulative percent revision (all diagnoses, modern prostheses) was 5.5% for stemmed reverse, 5.2% for stemless aTSA, and 7.9% for stemmed aTSA (Figure 2 below) [1]. 

It is worthwhile noting that the 7.9% ten-year revision rate for stemmed aTSA includes decades of older, non-crosslinked polyethylene. When the glenoid is crosslinked, aTSA durability improves markedly: in a dedicated AOANJRR study of 10,102 stemmed aTSAs done for osteoarthritis, non-crosslinked polyethylene had more than double the revision risk of crosslinked polyethylene after 18 months (HR 2.3; 95% CI 1.6–3.1), with 12-year cumulative revision of 9% versus 5% [22]. Considering only crosslinked anatomic glenoids, the revision rate for stemmed aTSAs (5%) was comparable to stemmed RSA (5.5%) and stemless aTSA (5.2%). 

Vitamin E–stabilized polyethylene is a type of crosslinked polyethylene, and registries tend to pool the two.  Vitamin E reduces wear and osteolytic particle debris on the bench [23], but no study has yet demonstrated a vitamin-E–versus–plain-crosslinked revision difference in shoulder arthroplasty.

The revision rates for stemless aTSA match those for the RSA; the reasons for this are not clear - perhaps more experienced surgeons, greater ability to achieve the desired humeral component position,  a higher rate of use of modern glenoid components, and/or preferential selection of healthier shoulders with better quality bone. 

Comparative meta-analyses report RSA revision rates about four-fold lower than aTSA in the cuff-intact analysis, OR 0.43; 95% CI 0.29–0.65; p<0.001) [3], although an earlier meta-analysis found no mid-term difference (OR 0.33; p=0.16) [4]. A 2026 propensity-matched study showed a lower early revision rate for aTSAs, but at midterm followup the revision rate increased [12].

However, it is critical to recognize that revision is a poor proxy for clinical failure in RSA. A National Joint Registry analysis concluded that low RSA revision rates may not signify implant success. Instead, patients with poor outcomes and their surgeons may be reluctant to undertake complex RSA revisions which have unpredictable results.[6]. The point is apparent for the most common mode of RSA failure — a painful, poorly functioning but radiographically satisfactory RSA. Such an outcome is experienced by about a quarter of RSA patients [6,7,8], yet RSAs are rarely revised for this indication. A failure that is not revised never appears in the revision rate.

[Complication frequencies were drawn from indexed systematic reviews [26–29] (e.g., PJI 2.4%, acromial/scapular fracture 2.5%, primary-RSA instability 2.5%); the revisability column reflects their reported management — acromial fractures are predominantly treated non-operatively, instability usually presents within 90 days and is treated by component revision, and infection is nearly always surgical. The registry anchor for pain/PROM failure is O’Malley [6].]

The different failure types are not equally salvageable. If an anatomic shoulder fails, it can usually be converted to a RSA with outcomes that approach those of primary RSA.  Primary stemmed anatomic shoulders done for osteoarthritis are revised to a reverse in 89% of cases, keeping the original humeral stem 58% of the time [1]; 93.8% of failed stemless aTSAs are converted to RSAs. On the other hand, revision of a failed reverse to another reverse often fails to yield the desired improvement in comfort and function.

5. Durability

Durability matters most for the patient with decades of active use ahead. At minimum ten-year follow-up, aTSAs sustain their functional gains for primary osteoarthritis [13], the large concurrent aTSA experience supports this option in the high-demand patient who wishes to avoid a RSA [14]. The Australian registry shows modern stemless anatomic matching reverse on revision out to ten years, and crosslinked stemmed-anatomic glenoids more than halve the revision risk of older non-crosslinked ones [1,22].

6. Return to sport 

Return-to-sport rates are high after both implants and highest after aTSA in pooled data [15]. A recent large weightlifting series reported high self-rated comfort, yet its endpoint is a single ordinal “difficulty” item — capturing neither the amount of load nor performance. [16]. When actual one-repetition-max recovery is measured, returners perform below their presymptomatic level, with the largest decrement in bench press [17]. 

7. Surgeon capability. It is often said that a good aTSA outperforms a good RSA, which outperforms a bad aTSA, which outperforms a bad RSA [18].  Some say it is technically easier to do a good RSA than an aTSA (not my view). However it is for sure that as aTSA volume falls, fewer surgeons will be able to reliably provide a good anatomic to their patients with cuff intact arthritis. While some hold that navigation, patient-specific instrumentation, and robotics may improve component positioning; none has been shown to improve patient-reported outcomes or reduce complications for any type of arthroplasty [19,20,21]. It appears that the surgeon is sill the method.

So, in rough summary

Bottom line: 

The patient and the surgeon considering arthroplasty for cuff intact shoulder arthritis should discuss the available evidence on aTSA and RSA.

An aTSA - when performed by a surgeon who can deliver a reliable aTSA -  may be attractive when function and salvageability matter most —  particularly in the more active patient with a reconstructable glenoid.

A RSA may be more attractive in the less active patient, or one whose glenoid morphology or bone quality makes a durable aTSA less certain or when the shoulder surgeon is not comfortable performing an anatomic shoulder arthroplasty.

Two cautions bear on the consideration: the revision rate understates RSA failure, because its most common failure — a painful but intact shoulder — is rarely revised [6]; and roughly a quarter of RSA patients do not reach a satisfactory outcome at all [6,7,8]. 

A choice

Pileated Woodpeckers

Seattle

Follow on twitter/X: https://x.com/RickMatsen
Follow on facebook: https://www.facebook.com/shoulder.arthritis
Follow on LinkedIn: https://www.linkedin.com/in/rick-matsen-88b1a8133/

References

[1]Lewis PL, Gill DR, McAuliffe MJ, et al. Hip, Knee and Shoulder Arthroplasty: 2025 Annual Report. Australian Orthopaedic Association National Joint Replacement Registry. AOA: Adelaide; 2025. doi:10.25310/MXFR3061. (Figures ST1, ST2; Tables ST6, ST39, ST46–47, ST76.)

[2]Gupta MS, Krishan A, Rashid A, et al. Reverse versus anatomic total shoulder arthroplasty in patients over 70 with a competent rotator cuff and glenohumeral osteoarthritis: a meta-analysis. J Shoulder Elbow Surg. 2026 (online 2025). PMID: 41276069.

[3]Thamrongskulsiri N, Limskul D, Tanpowpong T, et al. Comparison of revision rates and clinical outcomes between anatomic and reverse total shoulder arthroplasty for rotator cuff-intact osteoarthritis: a systematic review and meta-analysis. Clin Orthop Surg. 2025;17(6):907–921. doi:10.4055/cios25012.

[4]Kim H, Kim CH, Kim M, et al. Is reverse total shoulder arthroplasty more advantageous than anatomic TSA for osteoarthritis with intact cuff tendon? A systematic review and meta-analysis. J Orthop Traumatol. 2022;23(1):3. PMID: 34993646.

[5]Kirsch JM, Puzzitiello RN, Swanson D, et al. Outcomes after anatomic and reverse shoulder arthroplasty for glenohumeral osteoarthritis: a propensity score-matched analysis. J Bone Joint Surg Am. 2022. PMID: 35867705.

[6]O’Malley O, Davies A, Sabharwal S, et al. Is there a difference in thresholds for revision between shoulder arthroplasty types? A National Joint Registry study. PLoS One. 2025. doi:10.1371/journal.pone.0330975.

[7]Werner BC, Lederman E, Gobezie R, et al. Understanding the variables associated with failure to achieve an acceptable symptom state after reverse shoulder arthroplasty. Semin Arthroplasty JSES. 2021.

[8]Ardebol J, et al. Defining the MCID and PASS following reverse shoulder arthroplasty for glenohumeral arthritis or cuff tear arthropathy at minimum 5-year follow-up. JSES Int. 2025.

[9]Marigi EM, Hao KA, Friedman RJ, et al. Exactech Equinoxe anatomic versus reverse total shoulder arthroplasty for primary osteoarthritis: case-controlled comparisons using the machine learning-derived Shoulder Arthroplasty Smart score. J Shoulder Elbow Surg. 2023. PMID: 39292145.

[10]Beleckas CM, Schodlbauer DF, Mousad AD, et al. Evaluation of new normal after shoulder arthroplasty: comparison of anatomic vs. reverse total shoulder arthroplasty. J Shoulder Elbow Surg. 2025;34:S43–S49. doi:10.1016/j.jse.2025.02.010. PMID: 40074195.

[11]Barco R, Savvidou OD, Sperling JW, et al. Complications in reverse shoulder arthroplasty. EFORT Open Rev. 2016;1:72–80. doi:10.1302/2058-5241.1.160003.

[12]Leinweber KA, Bowler AR, Diestel DR, et al. Reverse and anatomic total shoulder arthroplasty for glenohumeral osteoarthritis: a propensity-matched comparison at early and midterm follow-up. J Shoulder Elbow Surg. 2026. PMID: 41564999.

[13]Sharareh B, Whitson AJ, Matsen FA III, et al. Minimum 10-year follow-up of anatomic total shoulder arthroplasty and ream-and-run arthroplasty for primary glenohumeral osteoarthritis. J Shoulder Elbow Surg. 2024;33(6):1276–1284. PMID: 37777045.

[14]Matsen FA III, Whitson A, Jackins SE, et al. Ream and run and total shoulder: patient and shoulder characteristics in five hundred forty-four concurrent cases. Int Orthop. 2019;43(9):2105–2115. PMID: 31240359.

[15]Liu JN, Steinhaus ME, Garcia GH, et al. Return to sport after shoulder arthroplasty: a systematic review and meta-analysis. Knee Surg Sports Traumatol Arthrosc. 2018;26(1):100–112. PMID: 28409200.

[16]Abdelshaheed J, Chatterji R, Levy J, et al. Return to weightlifting following anatomic and reverse shoulder arthroplasty. J Shoulder Elbow Surg. 2026;35:1660–1666. doi:10.1016/j.jse.2026.02.002.

[17]Ames A, Shah SS, Pettit R, et al. Against surgeons’ advice: the return to sport in high-demand weightlifters following anatomic total shoulder arthroplasty at average 3.6 years’ follow-up. J Shoulder Elbow Surg. 2023;32(4):e153–e159. doi:10.1016/j.jse.2022.09.027.

[18]Menendez ME, Garrigues GE, Jawa A. Clinical Faceoff: anatomic versus reverse total shoulder arthroplasty for primary glenohumeral osteoarthritis. Clin Orthop Relat Res. 2022;480(11):2095–2100.

[19]Daher M, Fares MY, Boufadel P, et al. Patient-specific instrumentation in primary total shoulder arthroplasty: a meta-analysis of clinical outcomes. Clin Shoulder Elb. 2025;28(2):129–136. doi:10.5397/cise.2024.01095.

[20]Patient-specific instrumentation in shoulder arthroplasty: high tech, low yield? [editorial]. Clin Shoulder Elb. 2025;28(2). doi:10.5397/cise.2025.00423.

[21]Gaj E, Pagnotta SM, Berlinberg EJ, et al. Intraoperative navigation system use increases accuracy of glenoid component inclination but not functional outcomes in reverse total shoulder arthroplasty. Arch Orthop Trauma Surg. 2024;144(1):91–102. doi:10.1007/s00402-023-05038-y.

[22]Page RS, Alder-Price AC, Rainbird S, et al. Reduced revision rates in total shoulder arthroplasty with crosslinked polyethylene: results from the Australian Orthopaedic Association National Joint Replacement Registry. Clin Orthop Relat Res. 2022;480(10):1940–1949. doi:10.1097/CORR.0000000000002293. PMID: 35901440.

[23]Khan AZ, Maxwell MJ, Parrott RM, et al. Effect of vitamin E–enhanced highly cross-linked polyethylene on wear rate and particle debris in anatomic total shoulder arthroplasty: a biomechanical comparison to ultrahigh-molecular-weight polyethylene. J Shoulder Elbow Surg. 2024. PMID: 38182025.

[24]Gowd AK, Liu JN, Cabarcas BC, et al. Single Assessment Numeric Evaluation and patient acceptable symptom state thresholds following shoulder arthroplasty. J Shoulder Elbow Surg. 2021. (PASS: ASES 81.9, SANE 75.5; n=207, mixed TSA/RSA.)

[25]DeVito P, Damodar D, Berglund DD, et al. Predicting outstanding results after reverse shoulder arthroplasty using percentage of maximal outcome improvement. J Shoulder Elbow Surg. 2019;28(6):1223–1231. PMID: 30910258. (SST threshold 61.3% MPI; n=198.)

[26]Shah SS, Gaal BT, Roche AM, et al. The modern reverse shoulder arthroplasty and an updated systematic review for each complication: part I. JSES Int. 2020;4(4):929–943. (Periprosthetic joint infection 2.4% for primary RSA.)

[27]Shah SS, Roche AM, Sullivan SW, et al. The modern reverse shoulder arthroplasty and an updated systematic review for each complication: part II. JSES Int. 2020;5(1):121–137. doi:10.1016/j.jseint.2020.07.018. PMID: 33554177. (Instability, humerus/glenoid fracture, acromial/scapular-spine fracture.)

[28]Zumstein MA, Pinedo M, Old J, et al. Problems, complications, reoperations, and revisions in reverse total shoulder arthroplasty: a systematic review. J Shoulder Elbow Surg. 2011;20(1):146–157.

[29]Lau SC, Large R. Acromial fracture after reverse total shoulder arthroplasty: a systematic review. Shoulder Elbow. 2020;12(6):375–389. doi:10.1177/1758573219876486. PMID: 33281942.

Pyrocarbon Section 4: clinical data

In this post we take a bit of a deeper dive into the clinical information available regarding pyrocarbon humeral arthroplasty - looking at evidence from the Australian and New Zealand registries alongside the currently available reports on specific cohorts.

Please note that this is a "best effort" attempt to put together what's out there. If the reader finds any of this to be in error, please do let me know.

Newspaper-style, the post starts with a summary

and then presents data that may support the points presented in the summary.

1. There is no single “pyrocarbon shoulder” — there are at least three. Resurfacing (PyroTITAN, not FDA-cleared in the U.S.), the U.S. FDA-cleared stemmed pyrocarbon hemiarthroplasty, and the European stemmed hemiarthroplasty with systematic head-downsizing. Each uses different implants or different techniques; a finding for one may not transfer to another, and pooling them is a common error in this literature.

2. Registries present data on revision rate, not function. McBride 2026, the AOANJRR 2025 special clinical assessment, and the New Zealand registry agree that pyrocarbon is revised about as often as a good anatomic TSA. However revision is a surrogate endpoint: the observation that a patient's implant was not revised does not necessarily mean that the patient had good shoulder comfort and function (we know this from reverse total shoulders where patients living with poor postoperative comfort and function often do not have revision surgery).  What matters most to the patient and to us is the degree to which the arthroplasty improved the patient's quality of life.

3. The currently available data often do not relate to implants and techniques currently in use. Most series are based heavily on approaches no longer in play today.

4. Where patient-reported outcomes exist, they show - like just about every other type of shoulder arthroplasty - clinically significant improvement over the preoperative state that lasts 5–10 years. Appropriately controlled studies that compare clinical outcomes for pyrocarbon humeral arthroplasty to other surgical options for managing glenohumeral arthritis are, however, uncommon.

5. The one head-to-head functional comparison still favors total over hemiarthroplasty early on. In the New Zealand registry, anatomic TSA surpassed pyrocarbon hemi on the Oxford score at both 6 months and 5 years. Notably, the functional advantage of total over hemi is not eliminated by changing the use of pyrocarbon bearing surface.

6. Royalties or research funding from pyrocarbon implant manufacturers are disclosed in many of the cohort series.  That does not invalidate the data, but it is recognized that industrial support can affect study design, data analysis and conclusions.

7. The bearing surface is not the most important aspect of the reconstruction. As is the case for all other types of shoulder arthroplasty, the outcome of pyrocarbon arthroplasty is critically dependent on patient selection, component size, component positioning, glenoid management, soft tissue balancing, and rehabilitation.

Future clinical research will hopefully define the indications, the appropriate implants, the surgical technique, and the clinical outcomes for pyrocarbon humeral arthroplasty.

Now for the details

Two Australian Orthopaedic Association (AOANJRR) analyses: two different pyrocarbon devices. Both draw on the AOA registry, but they study different pyrocarbon configurations. Both presen revision/survivorship endpoints with no patient-reported outcomes.

1. McBride presents the pyrocarbon humeral hemi-resurfacing (PHR/PyroTITAN — a resurfacing cap, no head excision, the Australian research-restricted device that is not FDA-cleared for use in the U.S.), comparing PHR in patients <65 versus the five lowest-CPR (cumulative percent revision) aTSA combinations.

2. The AOA Annual Report also presents data on a hemi stemmed pyrocarbon implant. That stemmed configuration is basically the same as the U.S.-cleared device and the European (Boileau) stemmed construct.

To complement these registry studies, cohort reports concern the three pyrocarbon devices in current use: (1) resurfacing (PHR/PyroTITAN); (2) the U.S. stemmed hemiarthroplasty (FDA-cleared Tornier/Stryker); and (3) the Boileau/European stemmed hemiarthroplasty. Devices 2 and 3 share the same hardware (Aequalis Ascend Flex stem + pyrocarbon head) but have separate literature, study designs, and technique: the Boileau approach systematically downsized the component.

In the table above, each cell carries a generation/technique caveat: the long-term numbers may not describe the devices and techniques commonly used today.

AOANJRR 2025

The AOANJRR 2025 “hemi stemmed anatomic — pyrocarbon head” class does not differentiate U.S. vs Boileau technique.  This Annual Report compares four shoulder arthroplasty classes in patients under 60 with OA, restricted to prostheses still implanted in 2024 (the “modern prostheses” filter), with data to 31 December 2024. 

Note that the confidence intervals for all four classes overlap. Adjusted for age and sex, no comparison to the pyrocarbon hemi reached significance. The executive summary states it plainly: “a hemi stemmed anatomic with a pyrocarbon head was not different to traditional total shoulder replacement options in this age group.” The report contains no patient reported outcome data.

New Zealand National Joint Registry

Gao and colleagues reported the 159 stemmed pyrocarbon hemiarthroplasties (PyCHAs (Tornier Flex stem, pyrocarbon head) against 1,280 conventional metal HAs and 4,285 aTSAs. Importantly, average follow-up was shorter for PyCHA (3.3 yr vs 12.7 and 8.3). With this caveat, PyCHA retention (96.9%) was comparable to aTSA in patients under 60 and better than conventional metal HA on both retention and Oxford Shoulder Score. aTSA had numerically better Oxford scores than PyCHA at 6 months and 5 years.

Clinical outcomes and revision rates by device — the cohort series

The cohort series add what the registries omit — but they are single-arm, small, and almost all industry-linked. The registry analyses above report revision only; none measures a patient-reported outcome. The published cohort series below supply PROs, range of motion, and graded glenoid erosion — but each one is a single-arm case series (Level IV, no comparator), the largest is ~100 shoulders; follow-up is short-to-intermediate. They establish that each device improves shoulder scores against the patient’s own baseline; however, they do not establish superiority over aTSA or any other glenoid-sparing alternative. 

Resurfacing (PHR):  the contemporary third-generation resurfacing implant specifically has neither long-term revision nor any PRO follow-up. 

Hemiarthroplasty — US: the U.S. experience with the currently available implant is anchored by Griswold 2025 (JBJS), drawn from the Stryker pyrocarbon IDE cohort plus a subsequent prospective follow-up: 45 patients at a mean 73 months. Every PRO improved past MCID (ASES 47->96, Constant 48->88, SANE 39->94, VAS pain 5.0->0.2), satisfaction was 97.8%, and 7-year revision-free survival was 95.7% — both revisions for infection, none for glenoid erosion or breakage. 

Hemiarthroplasty — Boileau / European: the largest, longest evidence. The European stemmed device is the same Ascend Flex + pyrocarbon-head hardware as the U.S. implant, but placed with systematic head downsizing, anchored by the Boileau (Nice) and Garret/Godenèche (Lyon) groups, and corroborated by two independent European centers.

Cointat 2022 (64 shoulders, 92% survival at 3 yr), Garret 2024 (45 patients, 4.4% revision, scores maintained at 5–9 yr), and Boileau 2026 (103 shoulders, revision-free 94% at 5 yr and 89% at 10 yr) consistently show Constant rising from the mid-30s to ~80, SSV from ~35 to ~84, and return to work and sport above 90%. Mathon 2023 (Marseille; 41 shoulders, Constant 34->80 at 3 yr, 100% return to work, glenoid wear <0.6 mm on CT 3D modeling, no revisions) and Kleim 2024 (Munich/Agatharied; 31 shoulders, Constant 45->79 sustained to 5.5 yr, MCID surpassed in every diagnosis, medial glenoid erosion only ~0.3 mm/yr after a biphasic first year, 100% survival).  Kleim independently reproduced Boileau’s finding that glenoid reaming does not drive more erosion. 

The dominant, reproducible failure mechanism across the series was humeral-head oversizing: the pyrocarbon head sits ~2 mm proud of a metal head of equal diameter (a 1.5-mm support tray plus a 0.5-mm taper gap), nonanatomic reconstruction occurred in 24–29% of cases, and in Boileau’s series nonanatomic reconstruction carried a roughly 19-fold higher revision rate (25% vs 1.3%) along with worse erosion and function — which is why the group now downsizes the head by one size as routine. 

What the cohort series add — and still cannot settle. 

These series add to the registries: all three devices produce within-patient PRO gains exceeding MCID, durable to 5–10 years, with infrequent revision. However, there is still no well-controlled comparison against aTSA or any other glenoid-sparing alternative.

Summary

What the evidence supports:

• Resurfacing PHR (McBride): no detectable difference in revision risk versus best-in-class aTSA in patients <65 with OA — on a research-restricted device that is not available in the U.S., with shorter followup and thinner data, adjusted for age and sex only.

• AOA Annual Report: Stemmed pyrocarbon hemiarthroplasty not different from anatomic TSA with respect to revision rate at up to 7 years in patients <60 with OA. 

• New Zealand registry: In comparison to conventional metal hemiarthroplasty, both pyrocarbon configurations are at least non-inferior; pyrocarbon may outperform metal hemi with respect to implant retention.

• Cohort series: For each device-and-technique combination studied, patient-reported and clinician scores rise from baseline by margins exceeding MCID and hold to 5–10 years — across the resurfacing device (Caughey), the U.S. stemmed implant (Griswold), and the European stemmed implant at five centers (Cointat, Garret, Boileau, Mathon, Kleim). 

What the current evidence does not support:

• Any comparative effectiveness with respect to patient reported outcomes. The registries do not report PROs; the cohort series report PROs with clinically significant preoperative to postoperative gains, but lack well-controlled studies comparing clinical benefit relative of pyrocarbon to aTSA or other glenoid-sparing alternatives.

Conclusion

Future clinical research is needed define the indications, the appropriate implants, the surgical technique, and the clinical outcomes for pyrocarbon humeral arthroplasty relative to other methods for managing glenohumeral arthritis.

While the gold standard for comparing surgical treatments is a prospective randomized controlled trial, this is difficult to accomplish with meaningful numbers in a procedure performed on relatively few, highly selected patients. Authors therefore turn to propensity matching to compare separately collected series. The approach carries two challenges. The first is deciding which characteristics to match on — diagnosis? Walch type? how the glenoid was managed? length of follow-up? age? sex? The variables that most influence the result, such as Walch type and glenoid management, are often the very ones not recorded in the comparison series, so they cannot be matched even in principle. The second is attrition: cases are lost both because stricter matching discards more unmatched cases and because missing data remove others, so the compared groups end up a fraction of the original cohorts — sometimes a small one — and no longer fully representative of them. The commentary by Sanchez-Sotelo is required reading on this point (Pyrocarbon Shoulder Hemiarthroplasty Seems to Outperform Metallic Hemiarthroplasty at a Short-Term Follow-up. JBJS Am2026;108(8):529–530. DOI: 10.2106/JBJS.25.01142)

There is work to be done

Pileated woodpecker

Follow on twitter/X: https://x.com/RickMatsen
Follow on facebook: https://www.facebook.com/shoulder.arthritis
Follow on LinkedIn: https://www.linkedin.com/in/rick-matsen-88b1a8133/

References

1. McBride A, Hurley R, Gill D, Du P, Duke P, Taylor F, Hoy G, Page R, Ross M. Outcomes of pyrolytic carbon humeral resurfacing hemiarthroplasty compared to best-in-class total shoulder arthroplasty in young patients with osteoarthritis: analysis from the Australian Orthopaedic Association National Joint Replacement Registry. J Shoulder Elbow Surg. 2026;35(5):1209–1218. doi:10.1016/j.jse.2025.09.007.

2. Australian Orthopaedic Association National Joint Replacement Registry. Hip, Knee and Shoulder Arthroplasty: 2025 Annual Report. Adelaide: AOA; 2025. Special Clinical Assessment: Shoulder Implant Choice — patients aged <60 years with OA (Table ST110, Figure ST77); data to 31 December 2024.

3. Gao R, Viswanath A, Frampton CM, Poon PC. Short-term outcomes following 159 stemmed pyrolytic carbon shoulder hemiarthroplasties and how they compare with conventional hemiarthroplasties and total shoulder arthroplasties in patients younger than 60 years with osteoarthritis: results from the New Zealand National Joint Registry. J Shoulder Elbow Surg. 2023;32(8):1594–1600. doi:10.1016/j.jse.2023.01.020.

4. Caughey MA, Penny I, Frampton CM. Medium-term results of the Ascension Pyrotitan surface replacement and Pyrocarbon hemiarthroplasty in the shoulder. Semin Arthroplasty JSES. 2024;34(1):1–10. doi:10.1053/j.sart.2023.01.005.

5. Griswold BG, Berger JM, Davis BP, Mauter L, Boyd M, Schuette HB, Johnston PS, Sears BW, Hatzidakis AM. Five-year radiographic and clinical outcomes of pyrocarbon hemiarthroplasty for glenohumeral arthritis and osteonecrosis. J Bone Joint Surg Am. 2025;107(24):2751–2762. doi:10.2106/JBJS.25.00163.

6. Cointat C, Raynier JL, Vasseur H, Lareyre F, Raffort J, Gauci MO, Boileau P. Short-term outcomes and survival of pyrocarbon hemiarthroplasty in the young arthritic shoulder. J Shoulder Elbow Surg. 2022;31(1):113–122. doi:10.1016/j.jse.2021.06.002.

7. Garret J, Cuinet T, Ducharne L, ReSurg, Godenèche A. Pyrocarbon humeral heads for hemishoulder arthroplasty grant satisfactory clinical scores with minimal glenoid erosion at 5-9 years of follow-up. J Shoulder Elbow Surg. 2024;33(2):328–334. doi:10.1016/j.jse.2023.06.021.

8. Boileau P, Cointat C, Raynier JL, Schippers P, Ranieri R. Pyrocarbon hemiarthroplasty for the treatment of shoulder osteoarthritis in young, active patients: survival and risk factors for revision. J Shoulder Elbow Surg. 2026;35(2):421–437. doi:10.1016/j.jse.2025.06.021.

9. Mathon P, Chivot M, Galland A, Airaudi S, Gravier R. Pyrolytic carbon head shoulder arthroplasty: CT scan glenoid bone modeling assessment and clinical results at 3-year follow-up. JSES Int. 2023;7:2476–2485. doi:10.1016/j.jseint.2023.06.028.

10. Kleim BD, Zolotar A, Hinz M, Nadjar R, Siebenlist S, Brunner UH. Pyrocarbon hemiprostheses show little glenoid erosion and good clinical function at 5.5 years of follow-up. J Shoulder Elbow Surg. 2024;33(1):55–64. doi:10.1016/j.jse.2023.05.027.

11. Lajoinie L, Garret J, van Rooij F, Saffarini M, Godenèche A. Pyrocarbon hemi-shoulder arthroplasty provides satisfactory outcomes following prior open Latarjet. J Shoulder Elb Arthroplast. 2024;8:24715492241292857. doi:10.1177/24715492241292857.

12. Barret H, Garret J, Favard L, Bonnevialle N, Collin P, Gauci MO, Boileau P. Long-term (minimum 10 years) survival and outcomes of pyrocarbon interposition shoulder arthroplasty. J Shoulder Elbow Surg. 2025;34:739–749. doi:10.1016/j.jse.2024.05.026.

13. U.S. Food and Drug Administration, Center for Devices and Radiological Health. De Novo Classification Request for Tornier Pyrocarbon Humeral Head — Decision Summary. DEN220012. Silver Spring, MD: FDA; granted December 16, 2022. (Regulatory decision file for IDE G140202; documents the propensity-subclassified historical cobalt-chrome control [Tornier Flex CoCr, n=169] from the Aequalis Post-Market Outcomes Registry. The composite-clinical-success analysis was subsequently published as Hatzidakis 2026, ref 14.)

14. Hatzidakis AM, Garrigues GE, Mauter LA, de Gast A, Venegoni MR, Yang Y, Johnston PS. Clinical Outcomes of Pyrocarbon Hemiarthroplasty: A Short-Term, Multicenter Study. J Bone Joint Surg Am. 2026;108(8):572–583. doi:10.2106/JBJS.25.00054. 

15. Sanchez-Sotelo J. Pyrocarbon Shoulder Hemiarthroplasty Seems to Outperform Metallic Hemiarthroplasty at a Short-Term Follow-up. Commentary on Hatzidakis et al. J Bone Joint Surg Am. 2026;108(8):529–530. doi:10.2106/JBJS.25.01142. (Independent JBJS commentary flagging the 43% missing control data as substantially weakening the study, the 2-year follow-up as very limited, the absence of radiographic assessment of humeral-head reconstruction, and the need to compare pyrocarbon HA with contemporary anatomic TSA using modern polyethylene.)