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

Abstract

Backgroud

This study aimed to compare the outcomes of reverse total shoulder arthroplasty (rTSA) and anatomic total shoulder arthroplasty (aTSA) in patients with rotator cuff-intact osteoarthritis, assessing revision and complication rates, patient-reported outcomes, and postoperative range of motion.

Methods

This systematic review and meta-analysis included comparative studies with levels I–III evidence that assessed rTSA and aTSA in patients with rotator cuff-intact osteoarthritis. Inclusion criteria required studies to report postoperative revision rates, complication rates, patient-reported outcomes, or range of motion with a minimum follow-up of 2 years. Studies focusing on noncomparative designs, biomechanical analyses, or case series were excluded. A comprehensive search of PubMed, Ovid Medline, and Scopus databases was conducted from their inception through December 2024. Odds ratios (ORs) with 95% CI were used for dichotomous outcomes, and mean differences (MDs) with 95% CI were used for continuous outcomes.

Results

A total of 14 studies, encompassing 4,819 cases, were included in the analysis. rTSA demonstrated a significantly lower revision rate compared to aTSA (OR, 0.43; 95% CI, 0.29 to 0.65; p < 0.001), while both procedures yielded similar Constant scores (MD, -2.23; 95% CI, -5.80 to 1.33; p = 0.22), simple shoulder test (MD, 0.11; 95% CI, -0.30 to 0.52; p = 0.59), American Shoulder and Elbow Surgeons scores (MD, -1.51; 95% CI, -4.91 to 1.90; p = 0.39), subjective shoulder values (MD, 2.16; 95% CI, -2.44 to 6.75; p = 0.36), visual analog scale for pain (MD, -0.25; 95% CI, -0.72 to 0.21; p = 0.29), and ranges of motion, except for external rotation, where aTSA demonstrated superiority (MD, -11.28; 95% CI, -14.95 to -7.61; p < 0.001).

Conclusions

In patients with rotator cuff-intact osteoarthritis, rTSA is associated with a lower revision rate compared to aTSA, while achieving comparable clinical outcomes and range of motion, with the exception of external rotation.

Rotator cuff-intact shoulder osteoarthritis is a degenerative condition leading to pain, stiffness, and functional limitations, commonly treated with anatomic total shoulder arthroplasty (aTSA). It can arise from various causes, including primary osteoarthritis, rheumatoid arthritis, osteonecrosis, and posttraumatic changes.¹⁾ This condition typically occurs in middle-aged or older patients and is associated with pain, stiffness, reduced range of motion, and functional limitations. While nonoperative management, including physical therapy and medications, is the first-line treatment, cases that do not improve after appropriate nonoperative care often require surgical intervention.²⁾ Shoulder arthroplasty is a treatment option for patients with severe shoulder osteoarthritis, aiming to restore function and relieve pain by removing the damaged articular cartilage and replacing it with an implant.³⁾

Unlike cuff tear arthropathy, shoulder osteoarthritis without a rotator cuff tear is not associated with rotator cuff insufficiency, thereby preserving the shoulder's native biomechanics.¹⁾ Consequently, aTSA is traditionally performed in patients with intact rotator cuff structures, as it relies on the rotator cuff tendons for stability and function.⁴⁾ However, reverse total shoulder arthroplasty (rTSA), originally designed for patients with rotator cuff deficiency, is increasingly being utilized in select cases of rotator cuff-intact osteoarthritis.⁵⁾ This shift was driven by the recognition that many elderly patients might have had underlying rotator cuff tendinopathy or deficiency and by rTSA's unique biomechanical advantages.⁶⁾ By reversing the shoulder joint's native mechanics, rTSA offers a potential solution for patients with compromised glenohumeral biomechanics or those at higher risk of developing rotator cuff pathology postoperatively.⁷⁾

The comparative effectiveness of rTSA versus aTSA in patients with rotator cuff-intact osteoarthritis remains a topic of significant clinical interest. While aTSA was traditionally favored for this population, the expanding indications for rTSA necessitate a deeper understanding of its outcomes, particularly regarding pain relief, functional restoration, and complication rates.⁸⁾ Some studies suggest that rTSA provides superior outcomes, while others report similar results between the 2 implants.⁹,¹⁰⁾ However, despite the increasing use of rTSA in this setting, the comparative effectiveness of rTSA and aTSA remains unclear, with conflicting findings regarding revision rates, complications, and functional outcomes.

This study aimed to address this research gap by systematically comparing the outcomes of rTSA and aTSA in patients with rotator cuff-intact osteoarthritis. Specifically, the authors evaluated their impact on revision rates, complication profiles, patient-reported outcomes, and postoperative range of motion. Given the biomechanical advantages of rTSA in addressing glenoid-related complications and instability, the authors hypothesized that rTSA would demonstrate a lower revision rate than aTSA while achieving comparable clinical outcomes.

METHODS

As this study is a meta-analysis, ethical approval and informed consent were not required. The study was registered in PROSPERO (CRD42024627960).

Database Search

Two authors (NT and DL) separately performed database searches using PubMed, Ovid Medline, and Scopus databases to locate studies comparing postoperative outcomes of rTSA and aTSA in patients with rotator cuff-intact osteoarthritis. The searches included all relevant publications from the establishment of the databases up to December 2024, when the search was conducted. This systematic review adhered to the 2020 Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines and was registered with PROSPERO under registration number CRD42024627960. The search strategy for PubMed and Scopus databases involved terms such as (“reverse total shoulder arthroplasty” OR “reverse shoulder arthroplasty” OR “rTSA”) AND (“anatomic total shoulder arthroplasty” OR “total shoulder arthroplasty” OR “aTSA” OR “TSA”) AND (“primary osteoarthritis” OR “rotator cuff intact”). Similarly, the Ovid Medline database utilized the same terms but applied additional filters to include studies in English, full-text availability, and research focused on human subjects.

Inclusion and Exclusion Criteria

To be included, studies needed to fulfill the following criteria: they had to be clinical comparative studies with evidence levels of I–III and a minimum follow-up period of 2 years, published in English, directly evaluating the outcomes of rTSA versus aTSA in patients with rotator cuff-intact osteoarthritis, reporting postoperative clinical results or complications, and have the full text accessible. Studies were excluded if they were basic science or biomechanics research, case series, case reports, or review articles or if they analyzed populations derived from the same institution, the same surgical team, or previously published cohorts.

Data Extraction

The eligibility of the identified studies was independently assessed by 2 authors (NT and DL), who reviewed their titles, abstracts, and full-texts. Any conflicts during this process were resolved by seeking input from a third author (TI). The data collected from the included studies covered various aspects, including study design, patient demographic information, details of the surgical techniques and implants used, and outcomes such as revision rates, complications, clinical outcome scores, and range of motion.

Bias Assessment

This study employed the Methodological Index for Non-Randomized Studies (MINORS) and the Modified Coleman Methodology Score (MCMS) to assess the quality of included studies. Two authors (NT and DL) independently performed the evaluations, with discrepancies resolved through consensus or consultation with a third author (TI). These tools provided a systematic approach to assess risk of bias and methodological quality, categorized as poor, fair, good, or excellent. The MINORS tool evaluates 8 criteria for noncomparative studies and 12 for comparative studies, scoring each from 0 (not addressed) to 2 (adequately addressed), with a maximum score of 24 for comparative designs. The MCMS assesses 10 domains, scoring each from 0 to 10 for a total of 100, focusing on validated outcomes, blinding, and follow-up completeness. Studies scoring 50–69 were rated fair, 70–84 good, and ≥ 85 excellent, ensuring a rigorous evaluation of study quality.

Statistical Analysis

Statistical analysis was performed using Review Manager version 5.4.1 for Windows (The Cochrane Collaboration). Odds ratios (ORs) with 95% CI were calculated for dichotomous outcomes, while mean differences (MDs) with their corresponding 95% CI were computed for continuous outcomes. To evaluate statistical heterogeneity, the chi-square test was employed, with a p-value less than 0.1 indicating significant heterogeneity. A fixed-effects model was utilized in the absence of statistical or graphical heterogeneity. Conversely, when heterogeneity was detected, either statistically or graphically, a random-effects model was implemented.

RESULTS

Included Studies

The systematic search of the databases identified 1,819 studies. After eliminating 191 duplicates, 1,628 articles were screened based on their titles and abstracts. During this process, 1,575 studies were excluded for not meeting the study's objectives. The full texts of the remaining 53 studies were carefully reviewed, resulting in the exclusion of 39 articles due to reasons such as reporting irrelevant outcomes or overlapping cohorts with other included studies. In the end, 14 studies¹¹,¹²,¹³,¹⁴,¹⁵,¹⁶,¹⁷,¹⁸,¹⁹,²⁰,²¹,²²,²³,²⁴⁾ met the inclusion criteria (Fig. 1).

Fig. 1. The 2020 Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flow for study selection.

The extracted data from the included studies were analyzed to compare rTSA and aTSA in patients with rotator cuff-intact osteoarthritis. A total of 14 studies, encompassing 4,819 cases, were included in the analysis. All studies were categorized as level III evidence, with mean patient ages ranging from 66 to 79.5 years. Follow-up periods varied between 24 months and 5 years, with sample sizes spanning from 8 to 1,307 per group. Commonly reported outcome measures included American Shoulder and Elbow Surgeons (ASES) score, Constant score, simple shoulder test (SST), range of motion, revision rates, and glenoid loosening, among others. The quality of the included studies was assessed using both criteria and the MCMS scoring system. The MCMS scores ranged from 67 to 83, while the MINORS scores fell between 15 and 21, reflecting good methodological quality (Tables 1 and 2).

Table 1. Characteristics of Included Studies: Encompassing LOE, Mean Age, Clinical Follow-up, Sample Size, Loss to Follow-up, Outcomes Measurement, MINORS, and MCMS.

First author (year)	LOE	Mean age (yr) (rTSA/aTSA)	Clinical FU (mo) (rTSA/aTSA)	Sample size (n, rTSA/aTSA)	Loss FU (n, rTSA/aTSA)	Outcome measure	MINORS	MCMS
Gallusser (2014)11)	3	79 (range, 73–85)/66 (range, 47–79)	43 (range, 24–69)/57 (range, 24–95)	8/19	NR	Constant, SST, Quick DASH, SSV, ROM, revision, glenoid loosening, scapular notching	17	67
Steen (2015)12)	3	77.7 ± 8.0/76.7 ± 8.0	42 (range, 24–92)/49 (range, 25–97)	24/96	NR	ASES, SST, ROM, revision, glenoid loosening, scapular notching	19	80
Alentorn-Geli (2018)13)	3	72.5 ± 5.4/70.5 ± 7.5	35.1 ± 14.2/42.7 ± 18.4	16/15	NR	ASES, SST, pain VAS, ROM, revision, glenoid loosening	16	73
Haritinian (2020)14)	3	71 ± 11/68 ± 7.5	24	12/39	18/14	Constant, SSV, ROM, revision	17	76
Merolla (2020)15)	3	74 (range, 69–75)/70 (range, 68-72)	28.5 ± 4.5/29 ± 1.3	36/47	NR	Constant, pain VAS, ROM, revision, glenoid loosening, scapular notching	17	78
Wright (2020)16)	3	77 (range, 70–90)	57 (range, 24–99)/80 (range, 27–148)	33/102 (chart analysis) 21/46 (patient-reported outcome analysis)	14	ASES, pain VAS, WOOS, ROM, revision	16	73
Polisetty (2021)17)	3	74 ± 7.1/73 ± 5.8	38 (range, 24–85)/54 (range, 24–122)	63/252	NR	ROM, revision, glenoid loosening	21	83
Friedman (2022)18)	3	73/73	42/41	370/370	NR	Constant, ASES, SST, UCLA, SPADI, ROM, revision	20	79
Kirsch (2022)19)	3	67 ± 3.5/67 ± 4.7	27.2 ± 6.0/32.8 ± 13.4	67/67	NR	ASES, pain VAS, SANE, ROM, revision, glenoid loosening	20	78
Mowbray (2022)20)	3	≥ 70	5 yr	1,183/1,307	NR	OSS, revision	17	76
Hao (2023)21)	3	72.7 ± 8.6/71.2 ± 7.1	41.8 ± 21.1/40.1 ± 22.2	87/87	NR	Constant, ASES, SST, UCLA, SPADI, SAS, ROM, revision, glenoid loosening	17	72
Ardebol (2024)22)	3	79.5 ± 3.9/79.0 ± 3.5	32.5 ± 13.8/43.2 ± 19.6	37/67	NR	ASES, pain VAS, SSV, ROM, revision	15	75
Kim (2024)23)	3	75 ± 5/76 ± 4	44.5 ± 16.3/38.9 ± 13.7	26/41	NR	Constant, ASES, pain VAS, SANE, ROM, revision	17	75
Mahylis (2024)24)	3	70.9 ± 7.0/66.3 ± 7.7	40.6 ± 22.9/62.0 ± 37.8	149/187	NR	Constant, ASES, SST, pain VAS, UCLA, SPADI, SAS, ROM, revision, glenoid loosening	17	80

Values are presented as mean ± standard deviation unless otherwise indicated.

LOE: level of evidence, MINORS: Methodological Index for Non-Randomized Studies, MCMS: modified Coleman methodology score, rTSA: reverse total shoulder arthroplasty, aTSA: anatomic total shoulder arthroplasty, FU: follow-up, NR: not reported, SST: simple shoulder test, Quick DASH: abbreviated version of the Disabilities of the Arm, Shoulder, and Hand questionnaire, SSV: subjective shoulder value, ROM: range of motion, ASES: American Shoulder and Elbow Surgeons, VAS: visual analog scale, WOOS: Western Ontario osteoarthritis of the shoulder index, UCLA: University of California Los Angeles, SPADI: shoulder pain and disability index, OSS: Oxford shoulder score, SAS: shoulder arthroplasty smart score.

Table 2. Characteristics of Included Studies: Encompassing Inclusions, Exclusions, and Surgical Technique.

First author (year)	Inclusion	Exclusion	Surgical technique (rTSA)	Surgical technique (aTSA)
Gallusser (2014)11)	· Primary shoulder OA with posterior glenoid wear of at least 20°	NR	Anesthesia: NR Approach: deltopectoral Biceps management: NR Subscapularis management: NR Implant: Aequalis reversed prosthesis (Tornier) Glenoid: cementless, standard baseplate with a 15-mm-long central peg, four 4.5-mm screws Humerus: cemented	Anesthesia: NR Approach: deltopectoral Biceps management: NR Subscapularis management: NR Implant: Aequalis Anatomic Prosthesis (Tornier) Glenoid: cemented Humerus: cemented
Steen (2015)12)	· Primary shoulder OA with intraoperative confirmation of an intact rotator cuff · Minimum follow-up 2 yr	· Previous arthroplasty · Proximal humerus fracture · Rotator cuff repair or full-thickness rotator cuff tear · Supraspinatus Goutallier grade > 1	Anesthesia: NR Approach: deltopectoral Biceps management: NR Subscapularis management: lesser tuberosity osteotomy or peel off Implant: NR Glenoid: NR Humerus: cementless or cemented	Anesthesia: NR Approach: deltopectoral Biceps management: NR Subscapularis management: lesser tuberosity osteotomy Implant: NR Glenoid: NR Humerus: cementless
Alentorn-Geli (2018)13)	· Aged > 18 yr · Primary shoulder OA with severe posterior subluxation and bone loss (Walch B2 glenoid)	· Secondary shoulder OA · Full-thickness rotator cuff tear · Superior humeral head subluxation	Anesthesia: NR Approach: deltopectoral Biceps management: NR Subscapularis management: NR Implant: Comprehensive Total Shoulder System (Biomet) Glenoid: cementless Humerus: cementless	Anesthesia: NR Approach: deltopectoral Biceps management: NR Subscapularis management: NR Implant: comprehensive Total Shoulder System (Biomet); ReUnion Total Shoulder System (Stryker); Cofield Total Shoulder System (Smith and Nephew) Glenoid: cemented, pegged all-polyethylene base Humerus: cementless
Haritinian (2020)14)	· Primary shoulder OA with bone loss (Walch B2 glenoid) · Intact rotator cuff · Goutallier grade ≤ 2	· Traumatic cases · Revision cases · Torn rotator cuff · Follow-up < 24 mo	Anesthesia: general anesthesia and interscalene block Approach: deltopectoral Biceps management: NR Subscapularis management: tenotomy Implant: Aramis Reversed Total Shoulder (3S Ortho) Glenoid: cementless, baseplate with 4 screws Humerus: cementless	Anesthesia: general anesthesia and interscalene block Approach: deltopectoral Biceps management: NR Subscapularis management: tenotomy Implant: Aramis Anatomical Total Shoulder (3S Ortho) Glenoid: cemented, pegged all-polyethylene base Humerus: cementless
Merolla (2020)15)	· Primary shoulder OA · Intact rotator cuff · Minimum follow-up 2 yr	· Humeral head necrosis · Post-capsulorrhaphy arthropathy · Sequelae of fractures · Previous debridement or subacromial decompression	Anesthesia: NR Approach: deltopectoral Biceps management: soft tissue tenodesis Subscapularis management: lesser tuberosity osteotomy or tenotomy Implant: ascend Flex (Wright Medical) Glenoid: cementless, 29-mm baseplate with central post and 4 screws Humerus: cementless, short stem	Anesthesia: NR Approach: deltopectoral Biceps management: soft tissue tenodesis Subscapularis management: lesser tuberosity osteotomy or tenotomy Implant: Ascend Flex (Wright Medical) Glenoid: cemented, keeled component (Perform) or 4 peg component (CortiLoc) Humerus: cementless, short stem
Wright (2020)16)	· Aged ≥ 70 yr · Primary shoulder OA with no full-thickness rotator cuff tear · Active elevation of less than 90°	· Inflammatory arthropathy · Proximal humerus fractures · Sequelae of fractures · Prior rotator cuff repair · Prior shoulder arthroplasty	Anesthesia: NR Approach: deltopectoral Biceps management: NR Subscapularis management: lesser tuberosity osteotomy Implant: Grammont-style implant Glenoid: cemented Humerus: cemented or cementless	Anesthesia: NR Approach: deltopectoral Biceps management: NR Subscapularis management: lesser tuberosity osteotomy Implant: NR Glenoid: cemented, pegged all-polyethylene base Humerus: cemented or cementless
Polisetty (2021)17)	· Primary shoulder OA · Intact rotator cuff · Minimum follow-up 2 yr	· Revision cases	Anesthesia: NR Approach: NR Biceps management: NR Subscapularis management: NR Implant: NR Glenoid: NR Humerus: NR	Anesthesia: NR Approach: NR Biceps management: NR Subscapularis management: NR Implant: NR Glenoid: NR Humerus: NR
Friedman (2022)18)	· Primary shoulder OA · Intact rotator cuff · No history of previous shoulder surgery	· Fractures · Non-OA diagnosis · Follow-up < 24 mo	Anesthesia: NR Approach: NR Biceps management: NR Subscapularis management: NR Implant: NR Glenoid: NR Humerus: NR	Anesthesia: NR Approach: NR Biceps management: NR Subscapularis management: NR Implant: NR Glenoid: NR Humerus: NR
Kirsch (2022)19)	· Primary shoulder OA · Intact rotator cuff · Minimum follow-up 2 yr · Complete preoperative and postoperative functional outcomes scores · Availability of preoperative CT and MRI	· Diagnosis other than primary osteoarthritis · Torn rotator cuff · Incomplete clinical follow-up · Ipsilateral shoulder surgery other than an arthroscopic debridement	Anesthesia: general anesthesia Approach: deltopectoral Biceps management: soft tissue tenodesis to the pectoralis major tendon Subscapularis management: peel off Implant: AltiVate Reverse (DJO Surgical) Glenoid: cementless Humerus: cementless, inlay standard length humeral stem	Anesthesia: general anesthesia Approach: deltopectoral Biceps management: soft tissue tenodesis to the pectoralis major tendon Subscapularis management: lesser tuberosity osteotomy Implant: AltiVate (DJO Surgical); Zimmer Sidus stemless implant (Biomet); Zimmer Anatomical implant (Biomet); Ascend Flex (Wright Medical) Glenoid: cemented, all-polyethylene base Humerus: NR
Mowbray (2022)20)	· New Zealand Joint Registry · Primary shoulder OA · Aged > 70 yr	· Revision cases	Anesthesia: NR Approach: NR Biceps management: NR Subscapularis management: NR Implant: NR Glenoid: NR Humerus: NR	Anesthesia: NR Approach: NR Biceps management: NR Subscapularis management: NR Implant: NR Glenoid: NR Humerus: NR
Hao (2023)21)	· Rotator cuff-intact shoulder OA · Minimum follow-up 2 yr	· Posttraumatic arthritis · Oncologic indication · Preoperative nerve injury	Anesthesia: NR Approach: NR Biceps management: NR Subscapularis management: NR Implant: Equinoxe (Exactech) Glenoid: NR Humerus: NR	Anesthesia: NR Approach: NR Biceps management: NR Subscapularis management: NR Implant: Equinoxe (Exactech) Glenoid: NR Humerus: NR
Ardebol (2024)22)	· Rotator cuff-intact shoulder OA · Aged > 75 yr · Complete preoperative and postoperative functional outcomes scores · Minimum follow-up 2 yr	· Secondary arthritis · Torn rotator cuff · Rotator cuff arthropathy · Fractures	Anesthesia: NR Approach: deltopectoral Biceps management: NR Subscapularis management: Peel off Implant: Univers Apex (Arthrex) Glenoid: Lateralized glenoid Humerus: 135° inlay	Anesthesia: NR Approach: deltopectoral Biceps management: NR Subscapularis management: lesser tuberosity osteotomy Implant: Univers Apex (Arthrex) or Eclipse (Arthrex) Glenoid: cemented, all-polyethylene base Humerus: cementless, short stem (Univers Apex) or stemless humeral component (Eclipse)
Kim (2024)23)	· Rotator cuff-intact shoulder OA · Aged > 70 yr · Partial articular side rotator cuff tears involving ≤ 50% of the cuff	· Follow-up < 2 yr · Full thickness rotator cuff tears · Inflammatory arthritis · Proximal humeral fractures · Sequelae of fractures · Prior rotator cuff repair	Anesthesia: NR Approach: deltopectoral Biceps management: NR Subscapularis management: peel off Implant: Reverse Shoulder Prosthesis (DJO Surgical); Aequalis Ascend Flex prosthesis (Tornier); Equinoxe (Exactech) Glenoid: cementless Humerus: cementless	Anesthesia: NR Approach: deltopectoral Biceps management: NR Subscapularis management: peel off Implant: Aequalis Anatomic Prosthesis (Tornier); Aequalis Ascend Flex prosthesis (Tornier); Equinoxe (Exactech) Glenoid: cemented, pegged all-polyethylene base Humerus: cemented or cementless
Mahylis (2024)24)	· Rotator cuff-intact shoulder OA with glenoid retroversion of at least 15° · Minimum follow-up 2 yr	· Diagnosis other than shoulder OA · Revision cases · Stemless aTSA	Anesthesia: NR Approach: NR Biceps management: NR Subscapularis management: NR Implant: Equinoxe (Exactech) Glenoid: NR Humerus: NR	Anesthesia: NR Approach: NR Biceps management: NR Subscapularis management: NR Implant: Equinoxe (Exactech) Glenoid: NR Humerus: NR

rTSA: reverse total shoulder arthroplasty, aTSA: anatomic total shoulder arthroplasty, OA: osteoarthritis, NR: not reported, CT: computed tomography, MRI: magnetic resonance imaging.

Revision Rates

A total of 14 studies¹¹,¹²,¹³,¹⁴,¹⁵,¹⁶,¹⁷,¹⁸,¹⁹,²⁰,²¹,²²,²³,²⁴⁾ reported revision rates for rTSA and aTSA. The pooled data included 1,783 rTSA cases and 2,367 aTSA cases. The overall revision rate was 1.7% (31 / 1,783) for the rTSA group and 3.9% (93 / 2,367) for the aTSA group. The I² test showed a p-value of 0.59 and an I² of 0%, indicating no statistically significant heterogeneity; therefore, a fixed-effect model was used. The meta-analysis demonstrated a statistically significantly lower revision rate for rTSA compared to aTSA, with an OR of 0.43 (95% CI, 0.29–0.65; p < 0.001) (Fig. 2).

Fig. 2. Forest plot comparing revision rates following reverse total shoulder arthroplasty (rTSA) and anatomic total shoulder arthroplasty (aTSA) for rotator cuff-intact osteoarthritis. M-H: Mantel-Haenszel.

Glenoid Loosening and Scapular Notching

Glenoid loosening is the primary known complication of aTSA, often attributed to cement fixation on the polyethylene peg. Conversely, scapular notching is a well-recognized complication of rTSA, primarily due to the medialized center of rotation. Eight studies¹¹,¹²,¹³,¹⁵,¹⁷,¹⁹,²¹,²⁴⁾ reported glenoid loosening, while 3 studies¹¹,¹²,¹⁵⁾ reported scapular notching. In this review, glenoid loosening occurred far more frequently in the aTSA group (12.3%) compared to the rTSA group (0.2%). Scapular notching was observed only in the rTSA group, occurring in 7.4% of cases. In contrast, no cases of scapular notching were reported in the aTSA group (0%) (Table 3). These findings underscore the distinct complication profiles of the 2 procedures, with glenoid loosening being predominantly associated with aTSA and scapular notching specific to rTSA.

Table 3. Glenoid Loosening and Scapular Notching Following rTSA and aTSA for Rotator Cuff-Intact Osteoarthritis.

First author (year)	Glenoid loosening				Scapular notching
	rTSA		aTSA		rTSA		aTSA
	Event	Total	Event	Total	Event	Total	Event	Total
Gallusser (2014)11)	0	8	2	19	1	8	0	19
Steen (2015)12)	0	24	5	96	2	24	0	96
Alentorn-Geli (2018)13)	0	16	4	15	NR	NR	NR	NR
Haritinian (2020)14)	NR	NR	NR	NR	NR	NR	NR	NR
Merolla (2020)15)	0	36	0	47	2	36	0	47
Wright (2020)16)	NR	NR	NR	NR	NR	NR	NR	NR
Polisetty (2021)17)	0	63	61	252	NR	NR	NR	NR
Friedman (2022)18)	NR	NR	NR	NR	NR	NR	NR	NR
Kirsch (2022)19)	0	67	10	67	NR	NR	NR	NR
Mowbray (2022)20)	NR	NR	NR	NR	NR	NR	NR	NR
Hao (2023)21)	0	87	2	87	NR	NR	NR	NR
Ardebol (2024)22)	NR	NR	NR	NR	NR	NR	NR	NR
Kim (2024)23)	NR	NR	NR	NR	NR	NR	NR	NR
Mahylis (2024)24)	1	149	11	187	NR	NR	NR	NR
Percentage (%)	0.2		12.3		7.4		0

rTSA: reverse total shoulder arthroplasty, aTSA: anatomic total shoulder arthroplasty, NR: not reported.

Both complications can affect clinical outcomes by contributing to pain, decreased shoulder function, and the potential need for revision surgery, underscoring the importance of careful patient selection and surgical planning.

Clinical Outcome Scores

Six studies¹¹,¹⁴,¹⁸,²¹,²³,²⁴⁾ reported the Constant score for rTSA and aTSA. The pooled meta-analysis included 324 rTSA cases and 414 aTSA cases. The I² test showed a p-value of 0.01 and an I² of 70%, indicating statistically significant heterogeneity; therefore, a random-effect model was used. The MD between the groups was -2.23 (95% CI, -5.80 to 1.33; p = 0.22), indicating no statistically significant difference between rTSA and aTSA in terms of the Constant score (Fig. 3A). For the SST, 6 studies¹¹,¹²,¹³,¹⁸,²¹,²⁴⁾ included 326 rTSA cases and 443 aTSA cases. The I² test showed a p-value of 0.63 and an I² of 0%, indicating no statistically significant heterogeneity; therefore, a fixed-effect model was used. The MD was 0.11 (95% CI, -0.30 to 0.52; p = 0.59), indicating no statistically significant difference between the 2 procedures (Fig. 3B).

Fig. 3. Forest plot comparing clinical outcome scores following reverse total shoulder arthroplasty (rTSA) and anatomic total shoulder arthroplasty (aTSA) for rotator cuff-intact osteoarthritis. (A) Constant score. (B) Simple shoulder test. SD: standard deviation, IV: inverse variance.

Nine studies¹²,¹³,¹⁶,¹⁸,¹⁹,²¹,²²,²³,²⁴⁾ reported the ASES score, including 469 patients in the rTSA group and 647 in the aTSA group. The I² test showed a p-value of 0.002 and an I² of 70%, indicating statistically significant heterogeneity; therefore, a random-effect model was used. The pooled MD was -1.51 (95% CI, -4.91 to 1.90; p = 0.39), indicating no statistically significant difference between rTSA and aTSA (Fig. 4A).

Fig. 4. Forest plot comparing clinical outcome scores following reverse total shoulder arthroplasty (rTSA) and anatomic total shoulder arthroplasty (aTSA) for rotator cuff-intact osteoarthritis. (A) American Shoulder and Elbow Surgeons score. (B) Subjective shoulder value. (C) Visual analog scale for pain. SD: standard deviation, IV: inverse variance.

Three studies¹¹,¹⁴,²²⁾ reported the subjective shoulder value (SSV) score, including 57 patients in the rTSA group and 123 patients in the aTSA group. The I² test showed a p-value of 0.36 and an I² of 2%, indicating no statistically significant heterogeneity; therefore, a fixed-effect model was used. The pooled MD was 2.16 (95% CI, -2.44 to 6.75; p = 0.36), demonstrating no statistically significant difference between rTSA and aTSA (Fig. 4B).

Five studies¹⁶,¹⁹,²²,²³,²⁴⁾ reported the visual analog scale (VAS) for pain, involving 300 patients in the rTSA group and 408 in the aTSA group. The I² test showed a p-value of 0.03 and an I² of 66%, indicating statistically significant heterogeneity; therefore, a random-effect model was used. The pooled MD was -0.25 (95% CI, -0.72 to 0.21; p = 0.29), showing no statistically significant difference between the 2 groups (Fig. 4C).

Range of Motion

Eleven studies¹¹,¹²,¹³,¹⁴,¹⁷,¹⁸,¹⁹,²¹,²²,²³,²⁴⁾ reported postoperative forward flexion, including 531 patients in the rTSA group and 911 in the aTSA group. The I² test showed a p-value of < 0.001 and an I² of 83%, indicating statistically significant heterogeneity; therefore, a random-effect model was used. The pooled MD was -5.01 (95% CI, -10.40 to 0.37; p = 0.07), showing no statistically significant difference between rTSA and aTSA (Fig. 5A). Seven studies¹¹,¹²,¹⁴,¹⁸,²¹,²³,²⁴⁾ evaluated postoperative abduction, including 348 patients in the rTSA group and 510 in the aTSA group. The I² test showed a p-value of 0.12 and an I² of 43%, indicating no statistically significant heterogeneity; therefore, a fixed-effect model was used. The pooled MD was -4.25 (95% CI, -8.05 to -0.45; p = 0.03), indicating no statistically significant difference between the 2 groups (Fig. 5B).

Fig. 5. Forest plot comparing clinical range of motion following reverse total shoulder arthroplasty (rTSA) and anatomic total shoulder arthroplasty (aTSA) for rotator cuff-intact osteoarthritis. (A) Forward flexion (B) Abduction. SD: standard deviation, IV: inverse variance.

Eleven studies¹¹,¹²,¹³,¹⁴,¹⁷,¹⁸,¹⁹,²¹,²²,²³,²⁴⁾ reported external rotation, with 531 patients in the rTSA group and 911 in the aTSA group. The I² test showed a p-value of 0.002 and an I² of 66%, indicating statistically significant heterogeneity; therefore, a random-effect model was used. The pooled MD was -11.28 (95% CI, -14.95 to -7.61; p < 0.001), showing a statistically significant advantage for aTSA over rTSA, suggesting that this difference may have important clinical implications. Improved external rotation could translate into an enhanced ability to perform daily activities, such as reaching behind the back or executing overhead tasks, thereby potentially improving overall shoulder function (Fig. 6A).

Fig. 6. Forest plot comparing clinical range of motion following reverse total shoulder arthroplasty (rTSA) and anatomic total shoulder arthroplasty (aTSA) for rotator cuff-intact osteoarthritis. (A) External rotation. (B) Internal rotation. SD: standard deviation, IV: inverse variance.

Five studies¹²,¹⁴,¹⁷,²³,²⁴⁾ evaluated internal rotation, involving 274 patients in the rTSA group and 615 in the aTSA group. The I² test showed a p-value of < 0.001 and an I² of 95%, indicating statistically significant heterogeneity; therefore, a random-effect model was used. The pooled MD was 0.01 (95% CI, -1.67 to 1.68; p = 1.00), indicating no statistically significant difference between rTSA and aTSA (Fig. 6B).

Publication Bias

A funnel plot was used to evaluate the potential for publication bias in the study. The plot appeared symmetrical (Fig. 7), suggesting that publication bias was unlikely.

Fig. 7. Funnel plot of publication bias regarding the odds ratio (OR) of revision rates following reverse total shoulder arthroplasty and anatomic total shoulder arthroplasty for rotator cuff-intact osteoarthritis. SE: standard error.

DISCUSSION

The primary findings of this study revealed that rTSA demonstrated a significantly lower revision rate compared to aTSA (OR, 0.46; 95% CI, 0.30–0.70; p = 0.0003), while both procedures yielded comparable clinical outcomes and range of motion, except for external rotation, where aTSA demonstrated superiority (MD, -11.28; 95% CI, -14.95 to -7.61; p < 0.001).

A prior systematic review by Kim et al.,¹⁰⁾ which examined 6 retrospective comparative studies, reported no significant difference in the midterm revision rates between rTSA and aTSA. The pooled OR for revision was 0.33 (95% CI, 0.07–1.57; p = 0.16), indicating statistical insignificance. However, the primary reasons for revision differed between the procedures: rotator cuff tears and posterior instability were common causes after aTSA, whereas revisions after rTSA were rare, primarily related to infection. However, in the present study, which included 14 retrospective cohort studies, the revision rate for aTSA was nearly twice as high as that for rTSA (3.9% vs. 1.7%), with a statistically significant difference (OR, 0.46; 95% CI, 0.30–0.70; p = 0.0003). Glenoid loosening, the most common complication associated with aTSA, accounted for a considerable proportion of the observed revisions.²⁵⁾ This is consistent with prior studies that highlight the challenges of achieving long-term stability of the cemented glenoid components in aTSA, especially in patients with preoperative glenoid retroversion or posterior wear patterns—features commonly associated with primary shoulder osteoarthritis.²⁶⁾ Additional complications necessitating revision included posterior humeral subluxation and rotator cuff dysfunction, both of which are design-specific issues inherent to aTSA.²⁷⁾ Notably, these complications are effectively addressed by rTSA, which not only contributes to its lower revision rates but also demonstrates a remarkably low incidence of glenoid-related complications. This advantage is likely attributable to the cementless fixation and the inherent mechanical benefits offered by medialized glenoid baseplates.⁸,²⁸⁾ However, scapular notching, a well-known complication of rTSA, was observed in approximately 7.4% of rTSA cases, although its clinical impact remains unclear in the short-term follow-up of the included studies.²⁹⁾

The findings from this meta-analysis revealed that patient-reported outcome measures, including the Constant, ASES, SST, SSV, and VAS scores, showed no statistically significant differences between rTSA and aTSA. This indicates that both procedures provide comparable pain relief and functional improvement for patients with rotator cuff-intact osteoarthritis. The previous systematic review by Kim et al.¹⁰⁾ also showed similar patient-reported outcome scores, consistent with the findings of the present study. These results support the expanding use of rTSA, even in the absence of rotator cuff deficiency, particularly in elderly patients or those with concomitant risk factors for postoperative cuff degeneration.

Postoperative external rotation was statistically significantly better in the aTSA group compared to rTSA (MD, 11.28; 95% CI, -14.95 to -7.61; p < 0.001), a finding that aligns with the biomechanical design of the implants. The aTSA preserves the native glenohumeral kinematics, relying on an intact rotator cuff for stabilization and mobility, thereby facilitating greater rotational range.⁴⁾ In contrast, rTSA, with its medialized center of rotation, inherently limits external rotation despite improvements in other planes, such as forward flexion and abduction.³⁰⁾ Nevertheless, the observed external rotation deficit in rTSA may be of minimal functional consequence, particularly in older, low-demand patients who benefit from the overall stability and reduced risk of revision.

Although aTSA provides better external rotation and rTSA has a lower revision rate, these findings should be interpreted cautiously due to differences in patient selection and follow-up duration. The long-term durability of both procedures remains uncertain, and rTSA's complications, such as scapular notching and deltoid fatigue, warrant consideration. Further prospective studies and randomized trials are needed to refine surgical indications and assess long-term outcomes.

The results of this study provide valuable guidance for surgeons when selecting between aTSA and rTSA for patients with rotator cuff-intact osteoarthritis. While aTSA remains a good option for younger, more active patients due to its superior external rotation and ability to preserve native biomechanics, rTSA offers distinct advantages, including lower revision rates and fewer glenoid-related complications. The growing use of rTSA shows its reliability, especially for older patients or those with risk factors for cuff degeneration. Future research should prioritize prospective randomized controlled trials with long-term follow-up to better clarify the comparative outcomes of rTSA and aTSA in patients with rotator cuff-intact osteoarthritis. Additionally, studies investigating patient-specific factors, such as preoperative glenoid erosion, activity level, age, and comorbidities, may help further refine the indications for each procedure.

This study has several limitations. First, all included studies were level III evidence, reflecting the lack of randomized controlled trials comparing rTSA and aTSA in this population. Second, significant heterogeneity in implant designs, surgical techniques, and outcome measures among the included studies may have influenced the pooled results. Third, the follow-up duration varied, with many studies reporting short- to midterm outcomes, limiting the evaluation of long-term implant survivorship and complications. Fourth, there is a risk of missing relevant articles during the search process, despite a comprehensive search strategy, which could introduce potential selection bias. Finally, as with any systematic review, the included studies exhibit heterogeneity in various aspects. This variability may influence the overall interpretation of the findings and should be considered when drawing conclusions. To further validate these findings and minimize bias, future research should prioritize well-designed randomized controlled trials to provide higher-level evidence on the comparative outcomes of aTSA and rTSA.

In patients with rotator cuff-intact osteoarthritis, rTSA is associated with a lower revision rate compared to aTSA while achieving comparable clinical outcomes and range of motion, with the exception of external rotation. The results suggest that rTSA may be a preferable option for patients at higher risk of revision, whereas aTSA remains beneficial for preserving external rotation and native joint mechanics.