To Patch or Not To Patch in Cuff Repairs

The Summary

A patch may make the cuff-repaired shoulder tendon look better on MRI. However, robust evidence is lacking that patient-reported outcomes are clinically significantly better for patients having cuff repairs with a patch than those with unpatched repairs.

Rotator cuff repair is one of the most common operations in orthopaedics. However, the repaired tendon does not always stay attached to the bone: on MRI, roughly a quarter of repairs show a retear within a year or two. This is especially the case for the most common type of cuff pathology: the degenerative tear, the repair of which is at risk for failure because of the limited quantity and quality of the muscle, tendon, bone unit required for cuff function.

Biologic patches are commonly used with the goal of improving the outcomes of cuff repair. For example, the bovine bioinductive patch (Regeneten) is a thin sheet of purified collagen from cow tendon. The surgeon sutures it over the repaired tendon, and the body is meant to grow new tendon-like tissue into it. In August 2025 the American Academy of Orthopaedic Surgeons gave the patch its highest grade of endorsement, a “strong” recommendation, saying it can lead to lower retear rates and better patient-reported outcomes.

The first half of that statement is supported. The second half appears to lack solid evidence.

Two trials, two different questions

The AAOS recommendation rests on two randomized trials of the same implant. The larger one, from Spain and supported in part by the manufacturer, randomized 124 patients with fully repairable tears 1 to 4 cm across (on average about 2 cm) to a standard repair with or without the patch. At two years, 12% of patched repairs had retorn on MRI, compared with 35% of unpatched ones. But the patients’ ratings of pain and function were the same in both groups at every visit. The authors titled their two-year report “a lower retear rate but similar outcomes.”

The smaller trial, of 60 patients at one center, asked a different question: could the patch replace a repair altogether in small tears (2.5 cm or less, with the rotator cable intact)? Those patients did score better: 8 points higher on the ASES shoulder score, but this difference is not clinically significant (less than the minimal clinically important difference).

In other words, neither trial showed an important difference in patient comfort and function.

Why a healed tendon does not necessarily lead to a better shoulder

A meta-analysis of 20 randomized trials of biologic augmentation helps explain the paradox. Shoulders whose repairs stayed intact scored better than those that retore, by about 14 points on the ASES and 10 on the Constant. But about three of four repairs heal anyway without augmentation, so augmentation prevented a retear in only about 12 of every 100 shoulders. If augmentation were applied to all cuff repairs, the average gain would be 1 to 2 points per patient on the 100-point ASES and Constant shoulder scores, far too small a benefit for the average patient to notice.

Other studies suggest that a retear has less impact on patient comfort and function than the appearance on MRI might imply. Pooled data from more than 3,000 repairs found that patients with retorn tendons scored only 6 to 7 points lower than those whose tendons healed, and most were satisfied.

The cost

The patch adds time and expense to surgery: estimates range from $5,000 to $10,000 per case. Non industry sponsored studies have also reported a downside that the AAOS guideline does not mention: stiffness. In one matched comparison, 9% of patched shoulders needed a second operation for inflammation and stiffness, compared with none of the unpatched ones. In another, six of 32 patched shoulders and none of 32 controls were reoperated for stiffness.

“What’s New in Sports Medicine” and “What’s New in Shoulder and Elbow Surgery”

The 2026 “What’s New” reviews in the Journal of Bone & Joint Surgery report the same pattern for other add-ons to cuff repair. Injected skin cells, drilling the bone and an osteoporosis drug each improved the tendon’s appearance on MRI without a clinically meaningful change in patient reported outcomes. Neither review mentions patches.

Readers of patch reports should be wary of conclusions from reviews. A 2026 study found that 13 of 16 published reviews of patch augmentation contained “spin” in their abstracts, presenting results more favorably than the data supported.

What patients and surgeons can ask

None of this means the patch never helps. For massive tears or poor-quality tendon, where retears are most likely, patches may prove worthwhile over the long haul, but most trials have followed patients for only one to two years. Until a trial shows that patients have better long-term comfort and function, the burden of proof falls on the proponents of the device. We can reasonably ask: Will it make the shoulder feel and work better? What will it add to the cost? What is the risk that it will make the shoulder stiff?

The details

Many shoulder surgeons are using bioinductive patches in rotator cuff repair. The evidence offered for patches is based mostly on MRI images, histology and retear rates. Robust evidence of a clinically significant improvement in patient comfort and function with their use is lacking. Recent JBJS “What’s New” articles review a series of randomized trials in which adjuncts to cuff repair improved healing on MRI without a clinically significant improvement in patient-reported outcomes [1,2]. One states that optimizing the biological environment after cuff repair “remains a challenge” [2]. Neither article mentions patch augmentation.

What the 2026 “What’s New” articles report

Both updates review a year of randomized trials in rotator cuff surgery [1,2]. The cuff trials they chose share a pattern: the adjunct improves the tendon on MRI; the patient outcomes are not different.

•    Autologous dermal fibroblast injection. Retear at 6 months fell from 24.3% to 5.6%, with no difference in ASES, Constant or SST scores [3].

•    Bone marrow stimulation. Incomplete healing at 2 years fell from 54.7% to 33.9%, but ASES scores were 89.2 and 87.8 (p = 0.498) [4].

•    Zoledronic acid. Retear at 6 months fell from 35% to 15%, with what the authors of the update called “modest” clinical improvement [1,5].

If a lower retear rate reliably meant a better shoulder, at least one of these trials should have shown a difference in patient-reported outcome at the minimal clinically important difference (MCID). None did.

What the pooled randomized trials show

Our recent meta-analysis of 20 Level I randomized trials (1,327 shoulders) of biologic augmentation confirms this pattern [6]. Augmentation lowered the retear rate from 26.7% to 14.8% (risk ratio 0.62). For patch augmentation alone (153 shoulders), the retear rate was 17.6%, a risk ratio of 0.55 [6]. Yet no patient-reported outcome improved by the MCID. The pooled differences were 0.9 points for the ASES and 2.2 points for the Constant score. In the patch subgroup the Constant score was only 3.9 points higher, and the ASES was not different [6].

Why does a healed tendon not translate into a better result? Intact repairs did score 14.3 ASES points and 10.5 Constant points higher than retorn repairs [6]. But because about 73% of unaugmented repairs heal, augmentation prevents a retear in only about 12 of every 100 shoulders. Spread across all the patients who receive it, the expected gain in outcome score per patient is only 1.3 to 1.7 points [6]. 

Other meta-analyses suggest that even this gap may overstate what a retear costs the patient. Across 43 study arms (3,350 repairs, 25% retorn at a median of 18 months), healed repairs scored only 6.1 Constant points and 7.1 ASES points higher than retorn repairs, both values are below the commonly used MCIDs. The authors concluded that most patients can expect satisfactory outcomes despite a retear [7]. An earlier meta-analysis of Level I and II studies found no difference in ASES between intact and retorn repairs; strength was the clearest difference [8].

“Healed” on MRI may also not mean healed where the tendon was fixed. In a study using tantalum markers in the repaired tendon, all 13 repairs retracted from their point of fixation during the first year, by 16 mm on average, yet only 4 showed a defect on MRI [9]. The MRI retear rate is thus an imperfect measure of the tendon’s condition, and a weak predictor of how the patient will do.

A second 2026 meta-analysis, of 9 randomized trials (556 patients) of biologic grafts, found a similar structural effect from augmentation: a retear risk ratio of 0.31 and tendons 1.2 mm thicker [10]. Its abstract reports the patient-reported outcomes only as improvement from before to after surgery, which occurs with or without a graft, and concludes that greater clinical improvement with augmentation was “less consistent” [10].

The AAOS strong recommendation for patches

The 2025 AAOS guideline states that “the use of bioinductive tendon implants to augment rotator cuff repair or as an alternative to non-augmented repair can lead to lower retear rates and better patient-reported outcomes” [11]. It grades the evidence High and the recommendation Strong. By the guideline’s own definition, Strong requires two or more high-quality studies “with consistent findings” [11]. The two studies are randomized trials of the same bovine collagen implant (Regeneten); the larger one received financial support from the manufacturer [12,13].

Trial	Patients and comparison	Imaging	Patient-reported outcomes
Ruiz Ibán et al. [12,13]	124 posterosuperior tears 1–4 cm across (mean about 2 cm), retraction ≤3 cm, fully repairable; transosseous-equivalent repair with vs. without patch	Retear at 2 years 12.3% vs. 35.1%	No difference in ASES, Constant, pain or EQ-5D; ASES MCID reached by 91% vs. 88%
Camacho Chacón et al. [14]	60 supraspinatus tears ≤2.5 cm with intact rotator cable; patch instead of repair vs. sutured repair	Better tendon histology at 6 months	ASES 88 vs. 80 at 2 years, a difference below the MCID

 

The guideline’s rationale acknowledges that in the larger trial “functional outcomes did not demonstrate any differences” [11]. That trial’s 2-year report, which the guideline does not cite, is titled “a lower retear rate but similar outcomes” [13]. The smaller trial asked a different question: can a patch replace a repair in a small tear with an intact rotator cable? Its 8-point ASES difference is below the MCID [14].  The two trials do not provide strong evidence of improved clinical outcomes for patients having patches used in their cuff repair

Neither trial studied the tears for which augmentation is most often proposed. The larger trial enrolled only fully repairable tears 1 to 4 cm across, with tears averaging about 2 cm; it excluded tendons retracted more than 3 cm and muscles with Goutallier grade 3 or 4 fatty infiltration [12,13]. The smaller trial enrolled only tears of 2.5 cm or less with an intact rotator cable [14]. Massive tears and tears with poor-quality tendon, where retears are most common, were not tested.

Figure 1. ASES and Constant scores over time, with and without the patch, in the two trials behind the AAOS strong recommendation. When the patch was added to a repair, the two groups tracked each other at every time point [12,13]. When the patch was used instead of a repair, the groups separated [14]. Ruiz Ibán values were read from the published figure; Camacho Chacón values are medians from the published tables.

The guideline lists the risks of these implants as longer operating time, possible infection and added cost. It does not mention stiffness or reoperation. It closes by calling for high-quality studies “to definitively establish the benefits of these devices” [11].

What else we know about patches

•    Partial-thickness tears. In a manufacturer-funded randomized trial of 41 patients, a patch without repair produced better WORC scores at 6 weeks and 3 months than takedown and repair. There was no difference in any patient-reported outcome at 6 or 12 months. The patch group was told to wear the sling for 1 week, the repair group for 4 to 6 weeks [15].

•    Meta-analysis. Improvement and the proportion of patients reaching the MCID after patch-augmented repair were similar to historical standard repair. The overall complication rate was 15.5% after full-thickness and 16.2% after partial-thickness repairs. There were no concurrent controls [16].

•    Matched cohort. At a minimum of 2 years, 47 patched repairs and 94 matched unpatched repairs had similar ASES (81 vs. 85), SSV (79 vs. 85) and pain scores. Reoperation for inflammation and stiffness was 9% (4/47) with the patch and 0% (0/94) without it. The authors stopped using the implant [17]. In the accompanying commentary, McAllister Nolan noted that these grafts are “heavily promoted to surgeons by industry” and recommended avoiding them for this indication [18].

•    Partial-thickness repairs. In a propensity-matched comparison, stiffness in the first 12 weeks occurred in 8 of 32 patched shoulders and 1 of 32 controls. Six patched shoulders, and no controls, required reoperation for stiffness [19]. This happened even though the controls more often had a tear takedown and repair (18 vs. 5 of 32), an operation that could be expected to cause more stiffness, not less [20].

•    Human dermal allograft in large tears. One randomized trial reports a longer follow-up. Sixty patients with 3 to 5 cm tears had repair with or without a human dermal allograft; 43 were seen at a mean of 5.7 years [21]. Retear was 9.1% with the allograft and 38.1% without it. The trial’s primary outcome, the Constant score, did not improve more with the allograft (p = 0.053). The ASES improved 13 points more with the allograft (p = 0.047), but the allograft group started 5 points lower [21]. This is a different product from the bovine implant in the AAOS strong recommendation, and the guideline rates the dermal allograft evidence as moderate [11].

A "strong" recommendation carries weight beyond the operating room [6]. A patient with a retear after an unpatched repair could point to it. A surgeon who left the patch out because the patient-reported data did not support it would have to defend that choice. Meanwhile the patient who receives a patch, becomes stiff and needs a second operation may ask why the implant was used. 

The cost is real. An economic analysis co-authored by an employee of the manufacturer assumed that the implant adds $3,641 to each repair; our meta-analysis cites $5,000 to $10,000 per case [6,22]. That analysis estimated an added $232,468 per 100 patients to obtain 18 more MRI-healed tendons [22]. That is a cost per healed tendon, not per patient with greater improvement.

A 2026 analysis, labeled "Level I evidence", concluded that the patch is “the dominant, cost-effective treatment strategy” over 10 years, saving money and adding 1.3 quality-adjusted life years per patient [23]. This is not an outcome study: no patient was followed for 10 years. It is a computer model, and its conclusion follows from its assumptions rather than from observed results. The study received grant funding from the manufacturer [23].

	What the model assumed	What the trial on which it was based actually found [12,13]
New retears	Each year, 26% of intact unpatched repairs and 8% of intact patched repairs retear. Over 10 years, that adds up to about 95% of unpatched and 57% of patched repairs.	35% of unpatched and 12% of patched repairs had retorn at 2 years. But the retear rate drops off over the first year. Only about 9% more control repairs retore in the second year. In contrast to the assumption, the retear rate is no the same "each year".
Quality of life	Assigned from MRI status: 0.7 if intact, 0.6 if torn, values taken from an earlier publication.	Quality of life was measured directly (EQ-5D) and did not differ between groups at any time.
Reverse arthroplasty	Each year, 16% of patients with a retorn tendon go on to reverse arthroplasty, at about $34,500 each.	No patient had a reverse arthroplasty. Two had early reoperations, both in the control group, and none occurred in the second year.

Each assumption favors the patch. The model turns a modest difference in retears at 2 years into a large one at 10, gives that difference a quality-of-life cost the trial did not find, and converts it into expensive operations that did not occur. Even its own quality-of-life values cannot produce the reported gain: if every unpatched tendon were torn and every patched tendon intact for all 10 years, the difference would be less than 0.9 quality-adjusted life years, not 1.3.

Who pays also matters. In one ambulatory surgery center, use of a biologic (a dermal allograft or bioinductive implant) was the largest single driver of charges for cuff repair, adding $17,791 to a risk-adjusted $19,728 [24]. Ambulatory cuff repair is increasingly paid under fixed-rate, episode-based bundles, in which the surgeon and facility bear the financial risk of added costs [25]. When the payment for the repair is fixed, the patch comes out of the margin, and a surgeon who owns a share of the center faces a three-way pull. On one side are a strong guideline recommendation and its medicolegal weight; on another, the surgeon’s own finances; and in between, trials showing no difference that patients are likely to notice. Contracts that pay for implants separately, or consulting relationships with manufacturers, can tilt the balance the other way. None of these is a clinical reason to use a patch, or to withhold one.

Questions to ask of any patch study

Abstracts deserve particular care. Among 16 systematic reviews comparing patch-augmented with unaugmented repair, 13 (81.3%) had spin in the abstract [26]. Spin was not associated with funding (P = .36), but each methodological flaw on the AMSTAR-2 checklist was associated with 61% more types of spin [26]. One example: a 2024 meta-analysis of 5 randomized trials of dermal patches concluded in its abstract that patches “improve outcomes.” Its discussion acknowledges that the differences, 2.8 Constant points and 5.6 ASES points, did not reach the MCIDs of 4.6 and 11.1 [27].

When the results of a well-controlled study on the effectiveness of patches in cuff repair are presented, we should ask:

1.   What was the difference in patient-reported outcome between groups, and did it reach the MCID?

2.   What proportion of patients in each group reached the MCID or the patient acceptable symptom state?

3.   How many had stiffness, capsulitis or reoperation, compared with controls?

4.   Was the control a repair alone, something less (debridement), or something more (takedown and repair)?

5.   Were the rehabilitation protocols the same in both groups?

6.   Who paid for the study?

Bottom line

What is demonstrated: (1) a bovine patch placed over a sound repair lowers the MRI retear rate; (2) a patch used instead of a repair combined with faster rehabilitation may return some patients to daily activities sooner; and (3) patches are associated with more stiffness and reoperation in unconflicted series.

What is not demonstrated: that adding a patch improves comfort or function by an amount that is clinically significant. 

Most trials follow patients for only one to two years; the mean follow-up in our meta-analysis was 18 months [6]. Whether fewer retears translate into better function later, or into less cuff tear arthropathy, is not known. 

Considering..

The Northern Shrike

AKA the "Butcher Bird" because he hangs his catch on tree thorns.

The author of this post has no financial relationship with any orthopaedic company.

Follow on twitter: https://twitter.com/shoulderarth

Follow on facebook: https://www.facebook.com/frederick.matsen

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

References

[1]        Rao AJ, Saltzman BM, Harrison AK. What’s new in shoulder and elbow surgery. J Bone Joint Surg Am. 2026;108(20):1574-7. doi:10.2106/JBJS.26.00744

[2]        Al-Hourani K, Khan S, Boufadel P, et al. What’s new in sports medicine. J Bone Joint Surg Am. 2026;108(11):763-70. doi:10.2106/JBJS.25.01633

[3]        Kim YK, Kim YT, Won YS, et al. Efficacy of an autologous dermal fibroblast injection in reducing the retear rate after arthroscopic rotator cuff repair: a prospective randomized controlled trial. Am J Sports Med. 2025;53(3):592-9.

[4]        Ye Z, Chen C, Chen H, et al. Bone marrow stimulation yielded similar clinical outcomes with improved tendon healing and abduction strength compared with decortication for arthroscopic rotator cuff repair in shoulder active patients. Arthroscopy. 2025;41(8):2740-9.e3.

[5]        Schanda JE, Ullrich R, Boesmueller S, et al. Intraoperative zoledronic acid for arthroscopic rotator cuff repair: short-term results from a prospective, randomized, placebo-controlled phase II trial. Am J Sports Med. 2026;54(2):342-52.

[6]        Checketts JX, Daniels SD, Schiffman CJ, Matsen FA III, Hsu JE. Biologic augmentation improves rotator cuff integrity but not clinically significant patient outcomes: a meta-analysis of randomized controlled trials. JBJS Open Access. 2026;11:e26.00248. doi:10.2106/JBJS.OA.26.00248

[7]        Holtedahl R, Bøe B, Brox JI. The clinical impact of retears after repair of posterosuperior rotator cuff tears: a systematic review and meta-analysis. J Shoulder Elbow Surg. 2023;32:1333-46. doi:10.1016/j.jse.2023.01.014

[8]        Russell RD, Knight JR, Mulligan E, Khazzam MS. Structural integrity after rotator cuff repair does not correlate with patient function and pain: a meta-analysis. J Bone Joint Surg Am. 2014;96:265-71. doi:10.2106/JBJS.M.00265

[9]        McCarron JA, Derwin KA, Bey MJ, et al. Failure with continuity in rotator cuff repair “healing.” Am J Sports Med. 2013;41(1):134-41. doi:10.1177/0363546512459477

[10]      Amin V, Ghayyad K, Oshikoya O, Goltz D, Huffman GR. Biologic graft augmentation is associated with lower retear rates and greater tendon thickness compared with arthroscopic rotator cuff repair without augmentation: a systematic review and meta-analysis of randomized controlled trials. Arthroscopy. 2026. doi:10.1002/arj.70619

[11]      American Academy of Orthopaedic Surgeons. Management of Rotator Cuff Injuries: Evidence-Based Clinical Practice Guideline. Published August 18, 2025. https://www.aaos.org/rccpg2025

[12]      Ruiz Ibán MA, Garcia Navlet M, Moros Marco S, et al. Augmentation of a transosseous-equivalent repair in posterosuperior nonacute rotator cuff tears with a bioinductive collagen implant decreases the retear rate at 1 year: a randomized controlled trial. Arthroscopy. 2024;40(6):1760-73.

[13]      Ruiz Ibán MA, García Navlet M, Moros Marco S, et al. Augmentation of a posterosuperior cuff repair with a bovine bioinductive collagen implant shows a lower retear rate but similar outcomes compared with no augmentation: 2-year results of a randomized controlled trial. Arthroscopy. 2025;41(10):3869-79. doi:10.1016/j.arthro.2025.03.057

[14]      Camacho Chacón JA, Roda Rojo V, Martin Martinez A, et al. An isolated bioinductive repair vs sutured repair for full-thickness rotator cuff tears: 2-year results of a double blinded, randomized controlled trial. J Shoulder Elbow Surg. 2024;33(9):1894-904. doi:10.1016/j.jse.2024.03.043

[15]      Wang A, Breidahl W, Ek ET, et al. Early functional recovery is improved in patients treated with bioinductive collagen implant augmentation compared with standard arthroscopic repair of high-grade partial-thickness rotator cuff tears: a prospective randomized trial. Orthop J Sports Med. 2026;14(3):23259671261418675. doi:10.1177/23259671261418675

[16]      Warren JR, Domingo-Johnson ER, Sorensen AA, et al. Bioinductive patch as an augmentation for rotator cuff repair, a systematic review and meta-analysis. J Shoulder Elbow Surg. 2024;33(11):2515-29. doi:10.1016/j.jse.2024.05.002

[17]      Haft M, Li SS, Pearson ZC, et al. No short-term clinical benefit to bovine collagen implant augmentation in primary rotator cuff repair: a matched retrospective study. Clin Orthop Relat Res. 2025;483:442-52. doi:10.1097/CORR.0000000000003247

[18]      McAllister Nolan B. CORR Insights®: No short-term clinical benefit to bovine collagen implant augmentation in primary rotator cuff repair: a matched retrospective study. Clin Orthop Relat Res. 2025;483:453-4. doi:10.1097/CORR.0000000000003281

[19]      Yeazell S, Lutz A, Bohon H, et al. Increased stiffness and reoperation rate in partial rotator cuff repairs treated with a bovine patch: a propensity-matched trial. J Shoulder Elbow Surg. 2022;31(6 Suppl):S131-5. doi:10.1016/j.jse.2022.02.003

[20]      Yeazell S, Lutz A, Bohon H, et al. Response to Bushnell et al regarding: "Increased stiffness and reoperation rate in partial rotator cuff repairs treated with a bovine patch: a propensity-matched trial." J Shoulder Elbow Surg. 2022;31:e572-3. doi:10.1016/j.jse.2022.06.028

[21]      Lee GW, Kim JY, Lee HW, Yoon JH, Noh KC. Clinical and anatomical outcomes of arthroscopic repair of large rotator cuff tears with allograft patch augmentation: a prospective, single-blinded, randomized controlled trial with a long-term follow-up. Clin Orthop Surg. 2022;14(2):263-71. doi:10.4055/cios21135

[22]      McIntyre LF, Nherera LM, Schlegel TF. Resorbable bioinductive collagen implant is cost effective in the treatment of rotator cuff tears. Arthrosc Sports Med Rehabil. 2023;5(2):e367-74. doi:10.1016/j.asmr.2023.01.002

[23]      Hurley ET, Ruiz Ibán MÁ, Oeding JF, et al. Bio-inductive collagen implant augmentation for arthroscopic rotator cuff repair is cost-effective in medium to large tears for reducing retears: a secondary analysis of a randomized controlled trial. Arthroscopy. 2026;42:73-82. doi:10.1002/arj.70000

[24]      Morris JH, Malik AT, Hatef S, Neviaser AS, Bishop JY, Cvetanovich GL. Cost of arthroscopic rotator cuff repairs is primarily driven by procedure-level factors: a single-institution analysis of an ambulatory surgery center. Arthroscopy. 2021;37(4):1075-83. doi:10.1016/j.arthro.2020.11.033

[25]      Lu Y, Labott JR, Salmons HI IV, et al. Identifying modifiable and nonmodifiable cost drivers of ambulatory rotator cuff repair: a machine learning analysis. J Shoulder Elbow Surg. 2022;31:2262-73. doi:10.1016/j.jse.2022.04.008

[26]      Al-Badaineh M, Josan H, Imran B, et al. Spin in systematic reviews on patch-augmented rotator cuff repair: prevalence, predictors, and methodological implications. J Shoulder Elbow Surg. 2026 [Epub ahead of print]. doi:10.1016/j.jse.2026.07.030

[27]      Hurley ET, Crook BS, Danilkowicz RM, et al. Acellular collagen matrix patch augmentation of arthroscopic rotator cuff repair reduces re-tear rates: a meta-analysis of randomized control trials. Arthroscopy. 2024;40(3):941-6. doi:10.1016/j.arthro.2023.09.025

What reverse total shoulder geometry yields the best clinical outcomes for the patient?

 

Summary

Surgeons and manufacturers can control many things about the reverse total shoulder (RSA), including component design, size and placement.

Computer models and laboratory tests attempt to predict the particular RSA geometries that restore muscle lengths and avoid unwanted contact between bone and implant. 

This post asks a narrower question: do any of these choices make the shoulder work better or reduce complications for patients receiving an RSA?

Published trials found little effect on how patients did. Five randomized trials each tested a single aspect of geometry: 6 mm of lateralization on the glenoid side, a baseplate tilted 10° downward, a humeral component with a 135° rather than 155° neck-shaft angle, and, in two trials, a humeral component in neutral version rather than 30° of retroversion. Four found no difference in function scores, pain or motion at one to two years. The fifth found that 30° of humeral component retroversion gave more external rotation with the arm at the side, while neutral version gave slightly better scores on two questionnaires; the ASES score did not differ. 

Observational comparisons of a more lateral glenosphere or a lower neck-shaft angle showed more motion or less notching on X-ray, but generally not better scores. A baseplate inclined slightly upward or a baseplate in marked anteversion or retroversion did not change the scores. Where lateralized designs did score better, the difference was small: about 3 to 6 points on the 100-point Subjective Shoulder Value. 

Two studies that did report better scores compared a surgeon's historical outcomes with later ones, after a switch to a lateralized design or to matching humeral component retroversion to each patient's own anatomy. In such comparisons, later patients may do better for other reasons, such as greater surgeon experience and better patient selection. In one study of shoulders placed with the same retroversion, the more recent patients did better than the earlier ones.

Implant labels such as "lateralized" or "135° inlay" describe the implant, but not where the parts end up in the patient. Reaming and seating change that.  

When surgeons measured the final position on X-rays in 630 shoulders, patients with similar geometry had widely different outcomes. 

Reproducing a preoperative plan more precisely has not been shown to improve outcomes. Patient-specific guides made from each patient's' CT scans put the glenoid baseplate closer to the surgeon's preoperative plan; robotics and augmented or virtual reality aim to do the same. However, in the one study that measured function, patients whose baseplate was placed with patient-specific guides did no better at two years than those whose baseplate was placed without them. 

Moreover, the scapula itself sits at a different angle in each person: in 93 patients whose baseplate was placed with patient-specific guides at a planned 0° inclination, the inclination relative to the floor ranged from 1° to 53° downward. Thus the functionally important position of the arm with respect to the body varied while the glenohumeral anatomy was held constant.

Notching. The humeral component can notch the scapular neck below the glenoid. Lateralized designs and a glenosphere placed low, overhanging the inferior glenoid, were associated with less notching on X-ray. In a randomized trial, a 135° humeral component reduced notching from 58% to 21% but there was no difference in patient-reported outcomes. How much notching harms the patient is unclear: one series found no effect in the first few years; others found slightly lower scores and less motion.

Fractures of the acromion and scapular spine. Stress fractures of the acromion and scapular spine are the complication most often blamed on lateralization. The largest study covered 6,320 reverses at 15 U.S. centers. There, the strongest predictors were not RSA geometry but patient factors: inflammatory arthritis, a massive cuff tear, osteoporosis, prior surgery and female sex. Findings on glenoid-side lateralization and arm lengthening were conflicting. One surgeon who began medializing the glenoid in high-risk patients saw fractures fall from 9% to 2%, but this study compared earlier patients with later ones. At another center, fractures were rarer when the glenosphere center was farther from the acromion than from the greater tuberosity (0.4% vs 5.0%). Fractures matter to patients: in one series, those who had them had less forward elevation and were less likely to return to their activities.

Instability. In one surgeon's series, a greater change toward superior  baseplate inclination from before to after surgery was associated with instability, and patients reporting instability had lower ASES scores. 

Few studies had enough dislocations to test the effect of implant position.

How often RSAs go wrong varies with the report. Pooling 52 published series gave a complication rate of 9.4% and a revision rate of 2.6%. A Swiss center that followed 854 reverses prospectively found complications in 18% and further surgery in about 10%. The most common complications were acromial fracture and glenoid component loosening. Patients with a complication scored well below those without.

A 2026 review concedes that whether a particular geometry helps patients "is unclear," and that design changes "should ultimately be evaluated by clinical outcomes."

What we know and what we don't. In randomized trials, single changes in baseplate tilt, glenoid lateralization, neck-shaft angle and humeral component version produced no or only small differences in patient-reported results. 

It has not been demonstrated that any particular geometry gives consistently better patient-reported outcomes.

The evidence in detail

A recent post covered the factors a surgeon could consider in choosing between an anatomic and a reverse total shoulder in the treatment of cuff-intact arthritis.

When a reverse is selected, the surgeon would like to know which geometry leads to optimal comfort, function and satisfaction for the patient, and to fewer complications that need further care.

That is the focus of this post.

A recent post, examined how each element of reverse geometry can be measured on postoperative radiographs: baseplate inclination, inferior glenosphere overhang, glenoid and humeral lateralization, distalization, humeral component version, and neck-shaft angle. It showed that many measures in this literature do not measure what they are named for, that angles often stand in for millimeters, and that values shown on the box are often reported in place of the geometry actually achieved in the shoulder. 

This post asks the question: does the geometry change what the patient experiences? 

Only clinical studies are considered: studies of patients, with outcomes reported by patients or measured on them. 

Computer models, cadaver studies and finite-element analyses are not considered.  

A 2026 review of the reverse in cuff-intact shoulders makes the same point. It describes the case for a 135° humeral component and 5–10 mm of glenoid lateralization, which rests on computational and biomechanical studies, then concedes that whether these translate into better motion, strength or patient-reported outcomes "is unclear," and that "design changes should ultimately be evaluated by clinical outcomes and PROs" (Wong et al., Curr Rev Musculoskelet Med 2026 [1]).

A 2022 systematic review of humeral and glenoid version in the reverse recommends 0° to 20° of retroversion for both components. But nine of its 11 included studies were biomechanical, one was clinical-radiographic, and one described implant design (Berton et al., J Clin Med 2022 [2]). Its recommendation is therefore not used here.

Geometry and clinical function

In studies of patients having RSA, no position or geometry has been shown to consistently improve patient-reported function by an amount the patient would notice. Of five randomized trials, four found no difference in patient-reported function and the fifth found small, mixed differences. The observational associations are weak.

Geometry	Best clinical evidence	Effect on patient-reported function
Glenoid lateralization, 6 mm vs nonlateralized-distalized	RCT, one surgeon, 53 of 66 enrolled analyzed, any indication including revision (Southam et al., JSES 2023 [3])	No difference in ASES, PROMIS upper extremity or rotation-dependent daily activities, notching or acromial fracture at 2 years
Glenosphere +4 mm vs standard	Matched cohort, 102 pairs (King et al., OTSR 2023 [4])	No difference in pain, motion or scores
Lateralized vs nonlateralized glenospheres, bony or metallic	Meta-analysis of 16 comparisons, 865 shoulders; low levels of evidence (Nunes et al., JSES 2021 [5])	No difference in ASES, Constant or SST; Subjective Shoulder Value 6 points higher and pain 1 point lower with lateralization
Glenosphere 0 vs 4 mm with a 135° humeral component	Cohort, 82 patients (Harmsen et al., JSES Int 2022 [6])	No difference reported
Humeral version, neutral vs 30° retroversion	Double-blinded RCT, 66 patients (Wiater et al., JBJS 2024 [7])	No difference in rotation, elevation, strength, pain, PROMIS or ASES
Humeral version, neutral vs 30° retroversion (second trial)	Single-blinded RCT, one center, 60 randomized and 57 analyzed; 87% massive cuff tears or cuff tear arthropathy; one lateralized system (Ventura-Parellada et al., JSES 2026 [8])	30° gave more external rotation at the side (50° vs 38°). Neutral gave better SST (76 vs 67 of 100, about one item of 12) and SF-12 physical scores (45 vs 39). No difference in ASES, Constant or pain at 2 years
Humeral component neck-shaft angle, 135° vs 155°	Single-surgeon RCT, 100 randomized, 68 analyzed at 2 years or more (9 revised shoulders excluded); irreparable cuff tear or osteoarthritis; neutral glenosphere; one inlay system (Gobezie et al., JSES 2019 [9])	No difference in ASES (74 vs 78), SANE (74 vs 76), SST (8 vs 7), pain, forward flexion (132° vs 135°) or external rotation (29° vs 30°). Powered for a 10° difference in flexion, not for patient-reported outcomes
Humeral component neck-shaft angle, 135° vs 145°, one onlay system	Retrospective, 103 shoulders, 30 at 135°; glenoid lateralization and inferior eccentricity also differed between groups (Neyton et al., JSES 2023 [10])	No difference in Constant (66 vs 68), Subjective Shoulder Value (77% vs 83%) or motion
Humeral version matched to each patient's CT (10°–40°) vs fixed 20°	Retrospective, one surgeon, 80 patients, nearly all cuff tear arthropathy; fixed 20° in 2007–2012, individualized afterward (Oh et al., JSES 2020 [11])	Individualized group better: ASES 80 vs 73, SST 7.3 vs 5.9, ER 49° vs 39°, flexion 142° vs 129°. But 21 later patients whose individualized angle was 20° or less (mean 19°) also did better than the earlier fixed-20° group, so the difference cannot be attributed to version
Glenoid tilt, 10° inferior vs neutral	RCT, 42 analyzed (Edwards et al., JSES 2012 [12])	No difference in clinical outcome
Glenoid inclination as measured, superior (median 6°) vs neutral or inferior (median −4°)	Retrospective, one surgeon, 154 patients, one lateralized system (Mahendraraj et al., BJJ 2021[13])	No difference in ASES, SANE, pain or motion at 2 years
Glenoid baseplate version measured on the axillary radiograph: 10° or more of anteversion, neutral, 10° to 20° of retroversion, or 20° or more of retroversion	Retrospective, one institution, 115 shoulders with osteoarthritis or cuff tear arthropathy; one system whose glenosphere is two-thirds of a sphere; from our group (Sheth et al., Int Orthop 2025 [14])	No difference in SST, ASES or change in SST, and no linear relation between version and SST; no difference in complications or revisions. Patients with 10° or more of anteversion more often reported difficulty putting on a coat (86% vs 42% with neutral version)
Inferior glenosphere overhang, measured, by tertile (<7.1, 7.1–9.9, >9.9 mm)	Retrospective, one center, 284 shoulders, one medialized-glenoid, lateralized-humerus system (Dean et al., JSES 2022[15])	High overhang gained 12° more active elevation than low (49° vs 37°). No differences in ASES, Constant, SPADI or SST changes reached the MCID
Measured global lateralization and distalization (LSA, DSA)	Two registries, 630 reverses for cuff tear arthropathy, 5 implant systems (Imiolczyk et al., JSES 2024 [16])	No relation of either angle to Constant score, Subjective Shoulder Value or motion at 2 years
Measured lateralization and distalization angles, split into glenoid and humeral parts	Retrospective, 217 patients (Kirkham et al., JSES 2026 [17])	A greater increase in the lateralization shoulder angle and a smaller distalization angle correlated with a slightly better Subjective Shoulder Value (a ratio of 1.01 per degree, not related to the MCID); no association with ASES
Three designs: 155° inlay (Grammont), 135° inlay with 4 mm lateralized glenosphere, 145° onlay with lateralized baseplate	Two-center registry, 226 patients with cuff tear arthropathy; not randomized (Freislederer et al., BMC Musculoskelet Disord 2024[18])	No difference in Constant score through 2 years. Flexion and abduction best with the 145° onlay; rotation better with both lateralized designs
The same three designs in primary osteoarthritis	Two-center registry, 97 patients; groups differed before surgery: supraspinatus intact in 56% of the onlay group and 11–12% of the others, 98% women in the 135° group (Imiolczyk et al., BMC Musculoskelet Disord 2025 [19])	Subjective Shoulder Value 90% with the 155° design vs 95% and 93% with the lateralized designs (P = .046); no difference in Constant score; more motion with the lateralized designs
155° humeral component with 2.5 mm lateral offset vs 135° humeral component with 6–10 mm lateralized glenosphere	Retrospective, one surgeon, 107 consecutive patients; the surgeon switched designs in mid-series (Zitkovsky et al., JAAOS 2020[20])	Lateralized 135° group better: ASES 87 vs 77, SANE 95 vs 85, ER 51° vs 40°; flexion did not differ. Lateralization, neck-shaft angle, implant system and era all changed together
Design category: medial glenoid–medial humerus, medial glenoid–lateral humerus, lateral glenoid–medial humerus	Meta-analysis of 42 studies (Burden et al., BJJ 2021 [21])	No difference in pooled Constant or ASES scores among the three
Global lateral humeral offset and humeral version on CT, relative to each patient's preoperative anatomy	30 patients, CT before and after surgery, isokinetic dynamometer at a mean of 2.4 years (Charles et al., J Biomech 2024 [22])	Greater offset predicted greater strength, but offset more than 5 mm beyond the patient's own was associated with less strength; external rotation was greatest with version within 10° of native and offset at or medial to native. No patient-reported outcome

Two other findings deserve mention. When the subscapularis was repaired and the glenosphere was lateralized, ASES improvement was smaller than with either alone (Werner et al., JAAOS 2018 [23]). A meta-analysis of nine nonrandomized studies of lateralized humeral designs (lower neck-shaft angle or onlay tray) found 7° more external rotation, but no difference in Constant or ASES scores (Dhillon et al., Cureus 2021 [24]).

Before-and-after comparisons. The two studies in the table with clearly better scores, Oh [11] and Zitkovsky [20], each compared a surgeon's earlier patients with later ones. Experience, technique, rehabilitation and patient selection all change over time, and none of these can be separated from the implant change. The Oh study tests this directly. Later patients who happened to receive about the same 20° as the earlier group also did better, which points to the era in which the surgery was done rather than the version. In the Zitkovsky study, the later group also had shorter follow-up.

Implant type is not placement. Most of these studies classify shoulders by what the manufacturer's catalog says the implant does: a "lateralized" glenosphere, a "135° inlay" humeral component. A deeply seated onlay behaves as an inlay; an incompletely seated inlay behaves as an onlay. The lateralization a patient receives depends on the glenoid reaming, the seating and the humeral cut, not only the label on the box. A study that measured LSA and DSA found that patients with identical angles had widely different outcomes. [16]

Putting the baseplate where it was planned. A 2026 meta-analysis of 16 studies (532 shoulders) found that patient-specific guides brought the glenoid component closer to the plan for inclination and central pin position, but less so for version. The mean errors were 3.0° for inclination, 1.7 mm for pin position and 3.8° for version. Only one included study measured function, and it found no difference at two years. The authors conclude that better accuracy "has not been shown to improve clinical outcomes and PROMS" (Vicary-Watts et al., JSEA 2026 [25]).

Even a baseplate placed exactly as planned on the scapula does not sit the same way in every patient. Ninety-three patients had their baseplate placed with patient-specific guides at a planned 0° inclination. Each patient's resting scapular rotation differed, so inclination relative to gravity averaged 18° downward, ranging from 1° to 53°. At 9 weeks, a more open resting angle between the arm and the scapula predicted better ASES scores. By 26 weeks, no radiographic measure, including global inclination, predicted outcome (Shekhbihi et al., JSES 2026 [26].

Geometry and complications

The clinical evidence links geometry to notching and to acromial and scapular spine fractures, but the studies point in different directions on the key variable, glenoid lateralization.

How often do complications arise? A meta-analysis of 52 studies of primary reverses (5,824 shoulders, mean follow-up 3.9 years) found an overall complication rate of 9.4% and a revision rate of 2.6%. Notching was reported in 14.4% (Galvin et al., JSES 2022[27]). A single center that collected data prospectively on 854 primary reverses, 99% with one 155° onlay system, found higher rates. Complications occurred in 156 shoulders (18%), and 82 (about 10%) needed further surgery. The most common complications were acromial fracture (5.3%), glenoid loosening (4.3%), instability (2.7%) and humeral fracture or loosening (2.5%). Patients with a complication had a mean relative Constant score of 57% vs 81% and a Subjective Shoulder Value of 53% vs 80% (Kriechling et al., BJJ 2022 [28]). The authors suggest their onlay design may have added acromial stress, but the series had no comparison group. The difference between these two sources is itself a finding: complication rates may depend in part on how carefully patients are followed. Across design categories, pooled glenoid loosening was 6% with medial glenoid–medial humerus designs vs 0% with medial glenoid–lateral humerus designs, a finding the reviewers rated as low-strength evidence (Burden et al. [21]).

Notching. In a randomized trial, 10° of inferior tilt did not reduce notching: 75% with tilt vs 86% without, all placed 3 mm inferior (Edwards et al. [12]). In a randomized neck-shaft angle trial, with a neutral glenosphere, notching was 21% with a 135° humeral component and 58% with 155°, while patient-reported outcomes did not differ (Gobezie et al. [9]). Within one onlay system, notching was 53% at 145° and 30% at 135°, though glenoid lateralization also differed (Neyton et al. [10]). In primary osteoarthritis, notching was 22% with a 155° design, 16% with a 135° lateralized design and 9% with a 145° onlay (Imiolczyk et al. [19]). A meta-analysis of lateralized glenospheres found lower odds of notching (odds ratio 0.14; Nunes et al. [5]). A +4 mm glenosphere was associated with less notching in one matched cohort (King et al. [4]) but not in another that used a 135° humeral component: 24.4% vs 19.5% (Harmsen et al. [6]). Lateralized humeral designs were associated with less notching in pooled nonrandomized studies (odds ratio 0.17; Dhillon et al. [24]). In the randomized lateralization trial, notching did not differ (Southam et al. [3]). In a registry comparison of three designs, notching was 50% with the Grammont design, 24% with the 135° inlay and 14% with the 145° onlay (Freislederer et al. [18]). Pooled across 42 studies, notching was 52% with medial glenoid–medial humerus designs, 18% with medial glenoid–lateral humerus designs and 12% with lateral glenoid–medial humerus designs (Burden et al. [21]). When one surgeon switched from a 155° design with 2.5 mm of offset to a 135° design with 6–10 mm, notching fell from 41% to 15% and heterotopic ossification from 72% to 35% (Zitkovsky et al. [20]). Greater measured glenosphere overhang was associated with less notching: notching was 18%, 11% and 3% in the low, middle and high overhang tertiles (Dean et al. [15]). Superior inclination up to about 6° on a lateralized design did not increase notching (Mahendraraj et al. [13]). In 517 shoulders with one 135° inlay system, 6 mm or more of glenoid lateralization was associated with less notching in women (5.3% vs 15.6%), and glenosphere overhang was associated with less notching; lateralization was the implant's nominal value (Pak et al., JSES 2026 [29]). 

Whether notching matters to the patient is less clear. In one series it had no effect on outcomes at short-term follow-up (Harmsen et al. [6]). In the 517-shoulder series, notched shoulders had 13° less forward flexion and ASES scores 7 points lower, below the MCID (Pak et al. [29]). In another, notching rose from 11.9% at 2 years to 19.5% at 5 years and was associated with worse scores, motion and complications (Shelley et al., J Orthop 2020 [30]).

Acromial and scapular spine fractures. These are the complications most often blamed on lateralization.

Study	Patients	Associated with more fractures	Associated with fewer fractures or no effect
Moverman et al., ASES multicenter group, JBJS 2024 [31]	6,320 primary and revision reverses, 15 US institutions; fracture rate 3.8%	Strongest: patient factors, including inflammatory arthritis (OR 2.29), massive cuff tear (2.05), osteoporosis (2.00), prior surgery (1.82) and female sex (1.74). Geometry: total glenoid lateral offset (OR 1.06), greater increase in LSA (1.42), greater postoperative LSA (1.76); the LSA ratios appear to be across the interquartile range, not per degree	More lateral humeral offset (OR 0.74, across the interquartile range; see below). Distalization: no association
Pak et al., JSES 2024[32]	470 primary reverses, all with one 135° inlay humeral component; fracture rate 5.5%	Greater increase in acromiohumeral distance (OR 2.21 per cm); each mm of inferior glenosphere overhang (+19%); dominant arm (OR 2.21)	Glenoid lateralization 0–8 mm: no association (7.4%, 5.6%, 4.4%, 6.0%)
Schenk et al., Orthopedics 2020 [33]	1,146 reverses, one institution; 21 fractures (1.8%) vs 84 matched controls	Less humeral lateralization from before to after surgery (4.1 vs 8.4 mm), higher glenoid inclination, osteoporosis; less distalization with more medialization of the center of rotation	—
Werthel et al., JSEA 2018[34]	Fracture rate 1.1%	Less lateral greater-tuberosity offset, more arm lengthening, thinner acromion	—
Eisenberg et al., JSES 2026[35]	282 primary reverses, one surgeon; 13 fractures (4.6%); before vs after the surgeon began medializing the glenoid in higher-risk patients	Glenoid lateralization over 5 mm by implant specifications (OR 14.5, 95% CI 1.75 to 119.6); 12 of 13 fractures in lateralized glenoids	Fracture rate fell from 9.2% to 2.3%, and from 10.7% to 3.1% in patients with two or more risk factors
Kawashima et al., JSES 2026[36]	526 reverses, one medialized-glenoid, lateralized-humerus onlay system; 360 after propensity matching	Acromion closer to the glenosphere center than the greater tuberosity: fractures 5.0%, subacromial notching 10.8%	Acromion as far or farther: fractures 0.4%, notching 0%. ASES, SPADI and SST did not differ
Haidamous et al., JSES 2020[37]	426 primary reverses, three systems; 26 scapular spine fractures	Onlay humeral component (11.9% vs 4.7% inlay); greater postoperative acromiohumeral distance (37.5 vs 33.7 mm)	Global lateralization (52.8 vs 53.9 mm): no association
Kriechling et al., JSES Int 2023 [38]	860 reverses, one 155° onlay system; 45 fractures (5%)	Scapular spine (Levy III) fractures only: larger postoperative LSA (89° vs 83°), less distalization (acromiohumeral distance 33 vs 38 mm)	Acromial (Levy I and II) fractures: no radiographic predictor
Polisetty et al., JSES 2023[39]	920 reverses, one surgeon, one lateralized-glenoid 135° inlay system; 47 fractures (5.1%) vs 141 matched controls	—	No difference in global lateralization, acromiohumeral interval, delta angle, humeral seating or acromioclavicular joint morphology

The patient factors are more consistent across these studies: bone quality, inflammatory arthritis, female sex and cuff deficiency. 

The geometric findings are not. 

Greater glenoid lateralization was associated with features in two studies, [31] [35] and not in two others [32] [39]. In one of the two that found it, lateralization came from implant specifications not from measured postoperative geometry, earlier patients were compared with later ones, and there were 13 fractures [35]. 

Greater glenoid distalization was associated with more fracture in three studies [32] [34] [37], with fewer fractures in two [33] [38] (scapular spine fractures only in [38]) and with neither in the other two [31] [29].

Only one study measured the humeral side alone: more lateral humeral offset, from a line down the center of the humeral shaft to a parallel line through the glenosphere–humeral interface, went with fewer fractures [31]. The measures called humeral lateralization by Schenk, about 50 mm from glenoid to humerus [33], by Werthel [34] and by Haidamous [37] span the whole reconstruction and include the glenoid contribution. Moverman's odds ratios for lateral humeral offset (0.74) and for the lateralization shoulder angle (1.42 and 1.76) are labeled per millimeter or per degree, but the paper's Figure 2 does not fit that reading. Across the interquartile range of lateral humeral offset (6.7 to 11.9 mm), the odds of fracture fall to about 0.74 of their starting value or about 6% per millimeter. Across the plotted range of change in the lateralization shoulder angle (about 37°), the odds only double, or about 2% per degree. 

One geometric measure combines the two directions: whether the acromion sits farther from the glenosphere center than the greater tuberosity does. In one large series that position was associated with fewer fractures [36]. 

All of these studies in this table are observational. The one randomized lateralization trial found no difference in acromial fractures, but with only 53 patients it was too small to settle the question [3].

Fractures matter to patients. Those with scapular spine fractures had less forward flexion (120° vs 135°) and were less likely to return to their activities (71% vs 92%) [37]. In another series, patients with fractures were less satisfied despite equivalent ASES, pain and motion (Werthel et al. [34]), a reminder that a score can miss what a patient feels.

Instability. In 97 reverses from one surgeon, 13% reported some instability and 4% had a dislocation that needed reduction. A greater change toward superior baseplate inclination from before to after surgery was associated with instability (odds ratio 1.08 per degree). No subscapularis was repaired and lateralized glenospheres were excluded. Patients reporting any instability had lower ASES scores, 61 vs 72 (Tashjian et al., CORR 2018 [40]). A systematic review of ten retrospective studies pooled an odds ratio of 2.48 (95% CI 1.22 to 5.07) for instability when the subscapularis was not repaired; the authors call their findings hypothesis-generating rather than causal (Ulenberg et al., JSES 2026 [41]). Whether the tendon can be repaired depends partly on its quality, which that comparison cannot account for. In the 217-patient series, a larger postoperative lateralization shoulder angle was associated with reoperation for baseplate failure or instability (odds ratio 1.11 per degree), but on only 5 events [17]. The double-blinded version trial had one revision for instability (Wiater et al. [7]), as did a matched series of reverses for cuff-intact osteoarthritis (Baek et al. [42]).

Where this leaves us

Demonstrated in patients: in randomized trials, glenoid tilt (0° vs 10° inferior), 6 mm of glenoid lateralization and a 135° vs 155° humeral component produced no detectable change in patient-reported function; the 135° component reduced notching from 58% to 21%. Of two trials of humeral version (neutral vs 30°), one found no clinically important difference; the other found more external rotation at 30° and slightly better SST and SF-12 physical scores at neutral, with no difference in ASES. Observational studies of measured inclination, baseplate version and overhang found no clinically important difference in patient-reported scores. Meta-analyses of design categories and of lateralized glenospheres found no difference in ASES or Constant; the lateralized glenospheres gave a Subjective Shoulder Value only 6 points higher.

Reported, but confounded: better scores after a surgeon changed to a lateralized 135° design or to individualized version, and fewer acromial fractures after a surgeon began medializing the glenoid in high-risk patients. However, in each, later patients were compared with earlier ones.

Associated with complications in a few retrospective series: superior baseplate inclination and an unrepaired subscapularis with instability; an acromion closer to the glenosphere center than the greater tuberosity with acromial and scapular spine fracture. The strongest predictors of fracture remain the patient's own: bone quality, inflammatory arthritis, female sex and cuff deficiency.

Inferred, not demonstrated: that any particular reverse geometry makes patients better. The observational associations with notching and fracture are real but conflicting, and they come mostly from implant box labeling categories rather than measured position. Placing the baseplate closer to a preoperative plan has not been shown to improve patient outcomes.

How important is RSA geometry to the outcome realized by the patient?

Cooper's hawk

Follow on twitter: https://twitter.com/shoulderarth

Follow on facebook: https://www.facebook.com/frederick.matsen

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

References

1.     Wong LH, Lurie BM, Davies MR. Reverse total shoulder arthroplasty in the setting of an intact rotator cuff: a contemporary review of current indications, outcomes, and surgical considerations. Curr Rev Musculoskelet Med. 2026;19:70. doi

2.     Berton A, Longo UG, Gulotta LV, De Salvatore S, Piergentili I, Calabrese G, Roberti F, Warren RF, Denaro V. Humeral and glenoid version in reverse total shoulder arthroplasty: a systematic review. J Clin Med. 2022;11:7416. doi

3.     Southam BR, Bedeir YH, Johnson BM, Hasselfeld KA, Kloby MA, Grawe BM. Clinical and radiological outcomes in lateralized versus nonlateralized and distalized glenospheres in reverse total shoulder arthroplasty: a randomized control trial. J Shoulder Elbow Surg. 2023;32:1420-1431. doi

4.     King JJ, Hones KM, Wright TW, Roche C, Zuckerman JD, Flurin PH, Schoch BS. Does isolated glenosphere lateralization affect outcomes in reverse shoulder arthroplasty? Orthop Traumatol Surg Res. 2023;109(4):103401. doi

5.     Nunes B, Linhares D, Costa F, Neves N, Claro R, Silva MR. Lateralized versus nonlateralized glenospheres in reverse shoulder arthroplasty: a systematic review with meta-analysis. J Shoulder Elbow Surg. 2021;30:1700-1713. doi

6.     Harmsen SM, Robaina J, Campbell D, Denard PJ, Gobezie R, Lederman ES. Does lateralizing the glenosphere center of rotation by 4 mm decrease scapular notching in reverse shoulder arthroplasty with a 135° humeral component? JSES Int. 2022;6(3):442-446. PMC

7.     Wiater JM, Lee JYJ, Shields EJW, Childers K, Dery L, Koueiter D. Humeral component version has no effect on outcomes following reverse total shoulder arthroplasty: a prospective, double-blinded, randomized controlled trial. J Bone Joint Surg Am. 2024;106(13):1181-1188. doi

8.     Ventura-Parellada C, Goñalons-Giol F, Alonso-Rodriguez Piedra J, Gamez-Baños F, Jovell-Fernandez E, Mora-Guix JM. Humeral component retroversion in the Comprehensive Reverse Shoulder System arthroplasty: rotations, clinical outcomes, and quality-of-life analysis in a prospective randomized study. J Shoulder Elbow Surg. 2026;35:2-10. doi

9.     Gobezie R, Shishani Y, Lederman E, Denard PJ. Can a functional difference be detected in reverse arthroplasty with 135° versus 155° prosthesis for the treatment of rotator cuff arthropathy: a prospective randomized study. J Shoulder Elbow Surg. 2019;28:813-818. doi

10.  Neyton L, Nigues A, McBride AP, Giovannetti de Sanctis E. Neck shaft angle in reverse shoulder arthroplasty: 135 vs. 145 degrees at minimum 2-year follow-up. J Shoulder Elbow Surg. 2023;32:1486-1493. doi

11.  Oh JH, Sharma N, Rhee SM, Park JH. Do individualized humeral retroversion and subscapularis repair affect the clinical outcomes of reverse total shoulder arthroplasty? J Shoulder Elbow Surg. 2020;29:821-829. doi

12.  Edwards TB, Trappey GJ, Riley C, O'Connor DP, Elkousy HA, Gartsman GM. Inferior tilt of the glenoid component does not decrease scapular notching in reverse shoulder arthroplasty: results of a prospective randomized study. J Shoulder Elbow Surg. 2012;21:641-646. doi:10.1016/j.jse.2011.08.057. PubMed

13.  Mahendraraj KA, Shields MV, Grubhofer F, Golenbock SW, Jawa A. Reassessing glenoid inclination in reverse total shoulder arthroplasty with glenosphere lateralization. Bone Joint J. 2021;103-B(2):360-365. doi

14.  Sheth MM, Schiffman CJ, Whitson AJ, Matsen FA III, Hsu JE. Baseplate version in reverse shoulder arthroplasty: does excessive retroversion or anteversion affect functional activities of daily living? Int Orthop. 2025;49:1119-1125. doi

15.  Dean EW, Dean NE, Wright TW, Hao KA, Schoch BS, Farmer KW, Struk AM, King JJ. Clinical outcomes related to glenosphere overhang in reverse shoulder arthroplasty using a lateralized humeral design. J Shoulder Elbow Surg. 2022;31:2106-2115. doi

16.  Imiolczyk JP, Imiolczyk T, Góralczyk A, Scheibel M, Freislederer F. Lateralization and distalization shoulder angles do not predict outcomes in reverse shoulder arthroplasty for cuff tear arthropathy. J Shoulder Elbow Surg. 2024;33:121-129. doi

17.  Kirkham MS, Guy CR, Puzzitiello RN, Staten TM, Christy KB, Zhang C, et al. Do radiographic measurements of glenoid or humeral lateralization and distalization affect post-operative outcomes for reverse total shoulder arthroplasty? J Shoulder Elbow Surg. 2026;35:2336-2345. doi

18.  Freislederer F, Moroder P, Audigé L, Schneller T, Ameziane Y, Trefzer R, Imiolczyk JP, Scheibel M. Analysis of three different reverse shoulder arthroplasty designs for cuff tear arthropathy – the combination of lateralization and distalization provides best mobility. BMC Musculoskelet Disord. 2024;25:204. doi

19.  Imiolczyk JP, Audigé L, Freislederer F, Schneller T, Ameziane Y, Touet A, Scheibel M. Clinical and radiological comparison of three different reverse shoulder arthroplasty designs for patients with primary osteoarthritis. BMC Musculoskelet Disord. 2025;26:577. doi

20.  Zitkovsky HS, Carducci MP, Mahendraraj KA, Grubhofer F, Jawa A. Lateralization and decreased neck-shaft angle reduces scapular notching and heterotopic ossification. J Am Acad Orthop Surg. 2020;28:e1073-e1080. doi

21.  Burden EG, Batten TJ, Smith CD, Evans JP. Reverse total shoulder arthroplasty: a systematic review and meta-analysis of complications and patient outcomes dependent on prosthesis design. Bone Joint J. 2021;103-B(5):813-821. doi

22.  Charles SJ-C, LeVasseur C, Rai A, Kane G, Munsch M, Hughes J, Anderst W, Lin A. Anatomic restoration of lateral humeral offset and humeral retroversion optimizes functional outcomes following reverse total shoulder arthroplasty. J Biomech. 2024;176:112372. doi

23.  Werner BC, Wong AC, Mahony GT, Craig EV, Dines DM, Warren RF, Gulotta LV. Clinical outcomes after reverse shoulder arthroplasty with and without subscapularis repair: the importance of considering glenosphere lateralization. J Am Acad Orthop Surg. 2018;26(5):e114-e119. Link

24.  Dhillon G, Warren M, Assiotis A, et al. Reverse shoulder arthroplasty humeral lateralisation: a systematic review. Cureus. 2021;13(11):e19845. doi

25.  Vicary-Watts RJ, Bayne JR, Dom K, Grant AL, Wilkinson MPR, Morse LP. Glenoid preparation in shoulder arthroplasty, is patient-specific instrumentation more accurate than standard instrumentation? A systematic review and meta-analysis. J Shoulder Elb Arthroplast. 2026;100116 (pre-proof). doi

26.  Shekhbihi A, Raiss P, Modelhart M, Randlkofer LM, Herbst EC, Moroder P. The global glenoid component inclination: why scapulothoracic orientation should be considered when defining glenoid component inclination in reverse total shoulder arthroplasty. J Shoulder Elbow Surg. 2026 (online ahead of print). doi

27.  Galvin JW, Kim R, Ment A, Durso J, Joslin PMN, Lemos JL, Novikov D, Curry EJ, Alley MC, Parada SA, Eichinger JK, Li X. Outcomes and complications of primary reverse shoulder arthroplasty with minimum of 2 years' follow-up: a systematic review and meta-analysis. J Shoulder Elbow Surg. 2022;31:e534-e544. doi

28.  Kriechling P, Zaleski M, Loucas R, Loucas M, Fleischmann M, Wieser K. Complications and further surgery after reverse total shoulder arthroplasty: report of 854 primary cases. Bone Joint J. 2022;104-B(3):401-407. doi

29.  Pak T, Kilic AI, Ardebol J, Gobezie R, Sears BW, Lederman E; ShARC; Werner BC, Denard PJ. Glenoid-sided lateralization decreases scapular notching with a 135° humeral component Arthrex reverse shoulder arthroplasty. J Shoulder Elbow Surg. 2026;35:e224-e232. doi

30.  Shelley RJ, DeFoor MT, Parada SA, Crosby LA. Clinical implications of scapular notching at 2 and 5-year follow-up after reverse total shoulder arthroplasty. J Orthop. 2020;21:384-389. Link

31.  Moverman MA, Puzzitiello RN, Glass EA, et al., Jawa A; ASES Complications of RSA Multicenter Research Group. Implant-positioning and patient factors associated with acromial and scapular spine fractures after reverse shoulder arthroplasty. J Bone Joint Surg Am. 2024;106:1384-1394. doi

32.  Pak T, Ardebol J, Menendez ME, Gobezie R, Sears BW, Lederman E; ShARC group; Werner BC, Denard PJ. Robert H. Cofield, MD, Award for Best Oral Presentation 2023: up to 8 mm of glenoid-sided lateralization does not increase the risk of acromial or scapular spine stress fracture following reverse shoulder arthroplasty with a 135° inlay humeral component. J Shoulder Elbow Surg. 2024;33:S1-S8. doi

33.  Schenk P, Aichmair A, Beeler S, Ernstbrunner L, Meyer DC, Gerber C. Acromial fractures following reverse total shoulder arthroplasty: a cohort controlled analysis. Orthopedics. 2020;43(1):15-22. PubMed

34.  Werthel JD, Schoch BS, van Veen SC, Elhassan BT, An KN, Cofield RH, Sperling JW. Acromial fractures in reverse shoulder arthroplasty: a clinical and radiographic analysis. J Shoulder Elb Arthroplast. 2018;2:2471549218777628. doi

35.  Eisenberg MT, Hui C, Schwartz K, Qubain L, Amini MH. In reverse shoulder arthroplasty, selectively medializing the glenoid in higher-risk patients is associated with reduced risk of acromial stress fracture. J Shoulder Elbow Surg. 2026;35:2354-2361. doi

36.  Kawashima I, King JJ, Wright JO, Struk AM, Banks SA, Wright TW. Shoulder geometry after reverse total shoulder arthroplasty with a medialized glenoid and a lateralized humerus predicts subacromial notching and acromial or scapular spine fractures. J Shoulder Elbow Surg. 2026;35:40-48. doi

37.  Haidamous G, Lädermann A, Frankle MA, Gorman RA II, Denard PJ. The risk of postoperative scapular spine fracture following reverse shoulder arthroplasty is increased with an onlay humeral stem. J Shoulder Elbow Surg. 2020;29:2556-2563. doi

38.  Kriechling P, Weber F, Karczewski D, Borbas P, Wieser K. Predictive factors of acromial fractures following reverse total shoulder arthroplasty: a subgroup analysis of 860 shoulders. JSES Int. 2023;7:812-818. doi

39.  Polisetty T, Cannon D, Grewal G, Vakharia R, Levy JC. Radiographic and anatomic variations on postoperative acromion fractures after inlay and lateralized reverse shoulder arthroplasty. J Shoulder Elbow Surg. 2023;32:76-81. doi

40.  Tashjian RZ, Martin BI, Ricketts CA, Henninger HB, Granger EK, Chalmers PN. Superior baseplate inclination is associated with instability after reverse total shoulder arthroplasty. Clin Orthop Relat Res. 2018;476:1622-1629. doi

41.  Ulenberg N, Herbstreit S, Dudda M, Vogel C, Mester B. Risk factors for instability and dislocation after reverse total shoulder arthroplasty: a systematic review. J Shoulder Elbow Surg. 2026 (in press). doi

42.  Baek CH, Lim C, Kim JG, Kim BT, Kim SJ. Comparative outcomes of anatomic and reverse total shoulder arthroplasty for rotator cuff-intact glenohumeral osteoarthritis: a propensity score-matched, age-stratified exploratory analysis. J Shoulder Elb Arthroplast. 2026;10:100050. doi