When the rotator cuff is intact, 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].
Four comparisons help inform the choice: — patient-reported outcomes, motion, complications, and revision, each of which has its own nuances.
1. Patient-reported outcomes
On commonly used scores, the two types of arthroplasty are 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 RSA patients had an “unsatisfactory” Oxford Shoulder Score (<29) among 21,918 patients [6]. Single-center series using patient-acceptable-symptom-state (PASS) thresholds put the figure higher — 25–40% of RSA patients fail to reach PASS for ASES or SANE at two years [7], and 34–35% still fail at minimum five years [8], with pain the primary factor in these adverse outcomes.
With the Shoulder Arthroplasty Smart score, aTSA achieves higher absolute postoperative scores even when the improvement from baseline is similar [9]; and among patients who reach a “new normal” (SANE ≥95), aTSA significantly outperforms RSA on higher-demand tasks, motion, and return to sport and work [10].
2. Motion — a consistent aTSA advantage
Across the comparative meta-analyses, aTSA delivers better external and (where measured) internal rotation, with differences that exceed the MCID and reach roughly 10–11° of external rotation in pooled estimates [2,3,4] and better overall motion in matched cohorts [5]. Rotation affects the patient's ability to perform basic activities of daily living: dressing, toileting, perineal care, and reaching behind the back.
3. Complications
In pooled comparative data, RSA carries a lower overall complication rate than aTSA in the cuff-intact population [2,3]. But the complications the two implants produce differ 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]. So aTSA has failed at the glenoid and the cuff; RSA has failed through instability, infection, acromial or scapular-spine fracture and pain (as demonstrated above) [11].
4. Revision
At ten years, cumulative percent revision (all diagnoses, modern prostheses) is 5.5% for stemmed reverse, 5.2% for stemless anatomic, and 7.9% for stemmed anatomic with a polyethylene glenoid (Figure 2) [1].
It is worthwhile noting that the 7.9% ten-year revision rate for stemmed aTSA performed with a polyethylene glenoid 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 carried more than double the revision risk of crosslinked after 18 months (HR 2.3; 95% CI 1.6–3.1), with 12-year cumulative revision of 9% versus 5% [22]. Restricted to crosslinked glenoids, revisions for stemmed aTSAs (5%) is comparable to stemmed reverse (5.5%) and stemless anatomic (5.2%). A word on terminology: vitamin E–stabilized polyethylene is a subtype of crosslinked polyethylene, and registries pool the two — so the revision benefit shown here is the benefit of crosslinking. Vitamin E specifically reduces wear and osteolytic particle debris on the bench [23], but no study has yet isolated a vitamin-E–versus–plain-crosslinked revision difference in the shoulder.
The revision rates for modern stemless aTSA match those for the reverse; the reasons for this are not clear - perhaps more experienced surgeons, a higher use of modern glenoid components, or selection of healthier shoulders with better quality bone.
The pooled literature suggests that RSA is revised less often. Comparative meta-analyses report lower RSA revision (about four-fold lower 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 aTSA’s lower early rate narrows by midterm while later on it accrued more revisions and radiographic radiolucencies (note the latter a radiographic finding, not a revision) [12].
Revision is a poor proxy for failure in RSA. A National Joint Registry analysis tested exactly this and concluded that low RSA revision rates may not signify implant success, because patient and surgeon reluctance to undertake complex reverse revisions raise the threshold for revision [6]. The point is sharpest for the most common mode of RSA failure — a painful, poorly functioning but radiographically intact shoulder, which affects on the order of a quarter of patients [6,7,8] yet is revised rarely, because there is often nothing straightforward to revise. A failure that never becomes a revision never appears in the revision rate.
The most common failure — a painful, poorly functioning but radiographically intact shoulder — is the least likely to be revised, so the revision rate captures only a fraction of true RSA failure. Complication frequencies are 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 failures are not equally salvageable. When an anatomic shoulder does fail, it can be converted to a RSA with outcomes that are usually favorable. 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] (stemless anatomic converts to reverse in 93.8% of revisions). One the other hand, revision of a failed reverse to another reverse often fails to yield the desired result.
5. Durability and the younger, active patient
Durability matters most for the patient with decades of load ahead. At minimum ten-year follow-up, aTSA and the ream-and-run sustain their functional gains for primary osteoarthritis [13], the large concurrent ream-and-run/aTSA experience supports the anatomic options in the high-demand patient who wishes to avoid a reverse construct under load [14], and 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]. But “return” is not performance. A recent large weightlifting series reported high self-rated comfort, yet its endpoint is a single ordinal “difficulty” item — capturing neither load nor performance — with hemiarthroplasty folded into the anatomic arm and no radiographs to detect asymptomatic loosening under load [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 anatomic volume falls, fewer surgeons will be able to reliably deliver a good aTSA. While some hold that navigation, patient-specific instrumentation, and robotics may improve component positioning but none has been shown to improve patient-reported outcomes or reduce complications [19,20,21]
So, at a glance:
Bottom line:
Because mean patient-reported outcomes are similar, the patient's choice should turn on the factors that differ.
An aTSA may be attractive when function, the highest functional ceiling, and long-term salvageability matter most — typically the younger, more active patient with a reconstructable glenoid and a surgeon who can deliver a reliable anatomic construct; modern stemless anatomic now matches the reverse on ten-year revision.
A RSA may be attractive especially in the older patient, or one whose glenoid morphology or bone quality makes a durable anatomic reconstruction less certain.
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].
Both implants can, on average, relieve pain and restore function. The best choice for each patient depends on a good discussion of the pros and cons of each option.
A choice
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
Registry and comparative references verified against PubMed / the registry record (PMIDs and DOIs below). Figures 1–2 are the author’s own redrawings of AOANJRR data, used under attribution.
[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.
