The B2 glenoid.
A B2 glenoid is a biconcave glenoid: posterior wear has formed a second concavity (the neoglenoid) behind the original surface (the paleoglenoid), with a ridge between them, and the humeral head is decentered posteriorly onto the neoglenoid. The defining feature is the biconcavity, not a degree of retroversion [1, 2].
Axillary "truth" view. Note biconcavity and posterior decentering
seen when arm elevated in plane of scapula
In brief
In a B2 shoulder the humeral head sits back on a glenoid it has worn into two concavities. Many surgeons focus on addressing glenoid version. Let's put that into perspective:
• The goal of an anatomic arthroplasty, whether a ream-and-run or a total shoulder, is to restore the mobility and stability of the articulation, not to recreate premorbid anatomy. Rebuilding the premorbid anatomy by "restoring the native joint line" in a tight joint can leave the shoulder stiff.
• Correcting version has not been shown to improve what the patient reports. Studies relating the version of the implanted component to the result have not found that less retroversion gives a better shoulder. In a consecutive series of 458 anatomic total shoulders with standard components and no attempt to correct version, B2 and B3 shoulders did as well as type A shoulders [3].
• The evidence points to the importance of complete seating of the component and centering of the humeral head on the glenoid, rather than to the angle of version [10, 11, 20]. Centering the head on a well-seated glenoid component can avoid the rocking-horse loading that can lead to component loosening [1, 12, 20].
• "Correction" of version has costs. In pooled data, eccentric "high side" reaming carried more complications and revisions than noncorrective reaming [18]. Augmented components are intended to correct retroversion; however none has yet been shown to give the patient a better functioning shoulder [3, 15, 34, 37, 38].
• The reverse can be used in this setting [30, 36, 39, 46], but the differences reported between it and the anatomic total shoulder are small, below the minimal clinically important difference for the SST and the ASES scores [21, 22, 30, 36, 41]; the patients who received the reverse were older [30, 36].
• Preoperative CT and patient-specific instrumentation have not been shown to reduce revision rates [47].
• A standardized axillary "truth" view, taken with the arm elevated in the plane of the scapula, shows version, decentering, and biconcavity. The same view can be obtained after surgery to examine the pre to post operative anatomy [6, 7, 8, 9].
Two questions
The B2 glenoid has attracted more new technology than almost any other problem in shoulder arthroplasty: three-dimensional CT, planning software, patient-specific instrumentation, navigation, robotic assistance, posteriorly augmented glenoid components, augmented baseplates, and a growing use of the reverse in shoulders with an intact rotator cuff. Each is offered as a better way to manage posterior glenoid wear.
Two questions are worth asking about each of them.
• What problem are we trying to solve?
• What evidence shows that the proposed solution improves the patient’s comfort, function, and the durability of the arthroplasty — rather than just improving a measurement on an image?
What a B2 glenoid is
In advanced glenohumeral arthritis the humeral head often sits back on the glenoid. Over time it wears a second socket into the back of the glenoid. What remains is a biconcave surface: the original anterior socket that has not been worn (the paleoglenoid) and the newly worn posterior socket (the neoglenoid), with a ridge between them. The face of the glenoid is often retroverted, and the humeral head is decentered posteriorly onto the neoglenoid. Walch, who defined the B2, observed that "the value of the retroversion does not explain the biconcavity of the glenoid" [1].
A B3 is further along the same path. The ridge is gone and the surface is a single concavity again, but it is more retroverted, more medialized, and there is less bone left to support a component.
Type B glenoids are not rare problems. In the series from our practice they are among the most common glenoid types encountered in shoulder arthroplasty: 100 of 349 classified anatomic total shoulders (29%) were B2, and 152 (44%) were B2 or B3 [3]; 47 of 113 ream-and-runs in a recent study (42%) were B2 [4]; and in 544 concurrent cases, 46% of the ream-and-runs and 28% of the total shoulders were B2 [5].
Figure 1. The B glenoids. Left: B1 (posterior decentering without biconcavity). Center: B2 (posterior decentering with biconcavity). Right: B3 (centering on severely retroverted and medialized glenoid).
Seeing the B2: the standardized axillary view
Glenoid pathoantomy does not fall in to discrete categories, rather there are many intermediate types--fifty shades of gray.
Figure 2. Five type arthritic glenoids on standardized axillaryviews, showing a range of retroversion, decentering, and biconcavity.
The standardized axillary view, which we call the "truth" view, is a practical way of characterizing the glenoid articular surface and humeral centering on it. Importantly, this view can be used before and sequentially after shoulder arthroplasty without the cost and radiation of a CT scan. It is taken with the patient supine and the arm in neutral rotation, elevated 60 degrees in the plane of the scapula. The beam is directed from the axilla, parallel to the blade of the scapula, toward a cassette held over the top of the shoulder. The film is adequate when the spinoglenoid notch is clearly seen [6].
Figure 3. The axillary “truth” view.From Rockwood and Matsen’s The Shoulder, 6th edition, Figure 16-32.
The position of the arm matters because where the head sits on the glenoid depends on it. In an open MRI study of 20 arthritic shoulders, 10 with primary osteoarthritis and 10 with cuff tear arthropathy, imaged in different positions of elevation in the scapular plane, 5 showed posterior decentering at 30 degrees of abduction and 13 at 90 degrees. The authors called this "functional decentering" and pointed out that it “can often not be detected by standard radiologic workup with the arm in the neutral position.” The densest subchondral bone on the glenoid was where the head sat [7]. A CT or MRI taken with the arm at the side shows only the decentering that persists in that position. The two images below were taken on the same patient the same day. MRI to the right shows a centered head and the axillary "truth" view shows dramatic decentering [6].
The axillary view shows the head on the glenoid
with the arm elevated in a position of function.
From this one view the surgeon can read three things:
• Glenoid version: the angle between a line along the body of the scapula and a line joining the anterior and posterior rims of the glenoid. Retroversion is 90 degrees minus that angle.
• Biconcavity: whether posterior wear has formed a second socket, and how much of the glenoid face lies on each side of the ridge between them.
• Decentering: the percentage of the humeral head lying behind a line perpendicular to the midpoint of the glenoid face, as Walch defined it (see below) [1].
These readings are reproducible. In 127 shoulders, two independent observers, one of them a resident without special training in shoulder surgery, agreed closely on glenoid version (r = 0.81) and glenoid type (weighted kappa 0.859). Each reading took less than two minutes and needed no CT, reformatting or software. The proportions of type A and type B glenoids fell within the ranges reported in CT series [6].
What the view offers above all is the same view before and after surgery. Notably it does not give the same number for version as a CT. In one comparison of 50 shoulders imaged both ways, retroversion read greater on the axillary view in 86%, by a mean of 6.5 degrees. While CT is widely used to assess the shoulder before surgery, postoperative imaging is most often done with plain radiographs [8].
A systematic review found that 35% of studies measuring postoperative version used a different imaging method after surgery than before making comparison difficult [9].
Standardized axillary views taken the same way before and after surgery measure the same thing on both occasions, at a fraction of the cost and radiation of CT [6].
Figure 4. Standardized axillary views of the same shoulder before (left) and after (right) a ream-and-run with an anteriorly eccentric humeral head. Before surgery the head is severely decentered posteriorly on a biconcave glenoid; afterward it is recentered on the glenoid. The same value applies to the anatomic total shoulder.
The classification that gave us the B2 was built on CT, but CT is not necessarily the "gold standard," with variations related to patient positioning, imaging technique and slice selection.
Walch and colleagues read 113 osteoarthritic shoulders on axial slices taken with the patient supine and the arm at the side in neutral rotation. Fifteen percent were B2, with a mean retroversion of 23 degrees. The authors observed that the study showed "a wide range of glenoid polymorphism" [1]. Surgeons reading CT scans have not always agreed on which type a glenoid is. In later studies using the full classification, agreement between observers ranged from fair (kappa 0.37) to moderate (kappa 0.60); Walch's original figure of 0.65 to 0.70 was for the three main types alone. In one comparison, MRI identified fewer B2 glenoids than CT did, and its authors suggested that MRI may not show biconcavity reliably [2].
Decentering and humeroscapular alignment are different measurements
Unfortunately, the B2 literature uses one word, "subluxation", for two different measurements of where the humeral head sits. We prefer to avoid the word "subluxation" and identify each measurement for what it refers the head to.
Humeroscapular alignment refers the humeral head position to the plane of the scapular body. A line is drawn along the scapular body to the midpoint of the glenoid face; the position of the center of the humeral head is expressed as the percentage of the head lying posterior to that line. Fifty percent means the head lies squarely on the scapular axis; higher values mean it sits posteriorly. Confusingly, this is what most current reports mean by "subluxation", but it does not indicate the degree of decentering. For example, in a B3 glenoid the head will lie behind the plane of the scapula, but the head is not decentered on the face of the glenoid.
Decentering refers the head to the glenoid. A line is drawn perpendicular to the midpoint of the glenoid face, and the percentage of the humeral head lying behind that line is measured; 50% means the head is centered on the glenoid. This was in fact Walch's original measure (see figure above) [1]. Some reports express decentering relative to zero instead: the distance of the center of the head from the same line as a percentage of the head's diameter, so that 0% is centered. It is the same measurement as the Walch percentage less 50; a head 64% behind the line is decentered 14% [10, 11]. Papers also differ in the sign they give to posterior decentering: some report it as a positive number, others as a negative one.
Figure 5. In the retroverted B3 glenoid, the humeral head center (yellow) commonly lies centered on the perpendicular bisector of the glenoid face (green line). As a result of the glenoid retroversion, the head center lies substantially posterior to the plane of the scapula (blue line).
Humeroscapular alignment combines two things: the version of the glenoid and the position of the head on it. Decentering isolates the second, which is the relationship between the two joint surfaces and the thing the rocking-horse concern is about.
Version and centering are different. In 33 arthroplasties requiring an anteriorly eccentric head, preoperative decentering bore no relationship to preoperative glenoid version (R² = 0.0009) [12]. In 40 B2 glenoids treated without changing version, decentering was corrected from −14% to −1% while version stayed at −21 to −19 degrees [13]. In 30 shoulders treated with the ream-and-run for the combination of retroversion, posterior decentering, and biconcavity, glenoid version did not change, while the centering of the humeral head on the glenoid was largely corrected [14]. The head can be recentered on the glenoid without changing version. On the other hand, one augmented-component series reported humeroscapular alignment fell from 69.1% to 53.5%, and decentering from 5.8% to −3.0% [15].
None of this would matter much if the reports kept the two apart. Many do not. Most augmented-component and reverse series report humeroscapular alignment and call it subluxation; the series from our group report decentering; and some report a quantity called subluxation on yet another scale, such as the series in which it fell from 66% to 12% [16] — a value that cannot be on the scapular scale, where 50% is centered.
Numbers using the same word cannot be compared across papers without checking which reference line was used. Neither measure has been shown to predict the result. In a 5-year series with non-corrective reaming, the final SST score and its change were not correlated with preoperative decentering (P = .398) or with humeroscapular alignment (P = .718). In patients with B1 and B2 glenoids, there were no differences in outcomes between patients with postoperative retroversion measurements of more or less than 15 degrees [11].
Where the concern about the B2 came from
A retroverted, biconcave socket with the head sitting back on it could load a glenoid component eccentrically — the rocking-horse mechanism — and could lead to early loosening and to continued posterior instability [10].
Figure 6. Left: eccentric loading of the glenoid exposes it to risk through the "rocking horse" mechanism (from Matsen FA III, Lippitt SB. Shoulder Surgery: Principles and Procedures. Philadelphia: Saunders; 2004:508). Right: a clinical example of rocking-horse loosening of a glenoid component.
That concern has clinical support. Walch and colleagues reported on 92 total shoulder arthroplasties performed for biconcave glenoids, most treated with asymmetric reaming, and found 15 revisions (16%) at a mean of 77 months, 6 of them for glenoid loosening [17]. Findings like these suggested that the glenoid component should be placed in less than 15 degrees of retroversion and center the head.
The rocking-horse mechanism concerns the second part of that target. By centering the humeral head on a well-seated glenoid component, the risk of rocking-horse loosening can be minimized whether glenoid version is changed or not.
Does correcting version improve what the patient reports?
Here is what has been found in patients in whom no specific effort was made to change glenoid version.
• At two years, glenoid components implanted in 15 degrees or more of retroversion did not produce lower Simple Shoulder Test (SST) scores, less improvement, more radiolucency, or more revision than components implanted in less retroversion. Among the 39 type B glenoids in that series, the retroverted components improved by 7.8 SST points against 5.3 for the non-retroverted — a difference in the direction opposite to the one expected by the advocates of version correction [10].
• At a minimum of five years (mean, eight years) in 210 shoulders treated with conservative reaming and standard components, the final SST score, the change in SST score, and the percentage of maximum possible improvement were not correlated with preoperative or postoperative humeral head centering, with Walch classification, or with glenoid version. Among the 110 B1 and B2 shoulders, outcomes did not differ above and below 15 degrees of postoperative retroversion [11].
• A systematic review and meta-analysis of 15 studies and 1,190 shoulders did not find published evidence that a glenoid component placed in less than 15 degrees of retroversion produced better patient-reported outcome scores, better motion, fewer complications, or fewer revisions than one placed in 15 degrees or more [18].
• A second systematic review, from a different group, reached the same conclusion by a different route. Rutledge and colleagues identified 16 studies and 1,211 shoulders; 13 of the studies were Level IV. Of the 11 that analyzed clinical outcomes, 8 found no association between preoperative or postoperative retroversion and the result. The authors concluded that the evidence is insufficient to show that version affects outcomes independently of other features of the deformity and that it does not establish that retroversion routinely needs to be corrected [19].
These studies have not shown that correcting version improves the comfort or function the patient reports. That is different from showing that version does not matter. What can be said is that the benefit assumed for version correction has not yet been demonstrated in patient-reported outcomes.
The same meta-analysis reports a radiographic finding in the direction opposite to what some may have suspected. Shoulders with 15 degrees or more of postoperative retroversion were more likely to have no radiolucency at all around the component — a Lazarus grade of 0 — than shoulders with less than 15 degrees (odds ratio 0.38 for the group under 15 degrees, P = .0002) [18]. These are pooled retrospective cohorts that differ in surgical technique, so the comparison cannot show that retroversion protects the component. What it shows is that the expected association was not found.
The 5-year series points to something that may matter. Among the 51 shoulders with radiographs at a minimum of five years, 15 (29%) had components that met the radiographic definition of loose. On multivariable analysis the two variables most strongly associated with that finding were incomplete seating of the component (odds ratio 3.3, P = .082) and preoperative humeral head decentering (odds ratio 2.8, P = .127), rather than postoperative version (odds ratio 1.1, P = .311); none of the three reached statistical significance [11].
A group that does correct version has reported the same pattern. Harold and colleagues followed B2 glenoids treated with partially corrective high-side reaming for a mean of 8.6 years. Three of 59 shoulders (5.1%) had been revised, and 3 of the 30 with final radiographs had components that had failed radiographically without revision. Failure was associated with how well the component was seated at surgery and with posterior displacement of the head that recurred afterward (21% against 7.0% at final follow-up). It was not associated with preoperative retroversion, inclination, or posterior displacement of the head. In this series the radiographic finding did associate with patient outcome: the shoulders with failed components had a mean ASES score of 73 against 88, and more pain (2.1 against 0.8) [20]. That 15-point ASES difference exceeds the 13.6-point minimal clinically important difference derived in one registry [21], though not the 20.9 points derived in another [22]. It points the same way as the 5-year series: seating and centering seem to loom large, rather than the angle.
The 2-year series reports one consequence of retroversion. The fluted central peg perforated the anterior glenoid vault in 52% of the retroverted shoulders against 18% of the others, and in the type B subgroup in 53% against 17%. That series did not test whether perforation affected the result; it points instead to earlier reports in which perforation was not associated with inferior clinical or radiographic outcomes [10]. Since we intentionally perforate the glenoid vault in performing a reverse total shoulder, perhaps we should be less concerned about perforation in an anatomic.
A commentary published alongside that series sets up the comparison worth having. Two surgeons used the same glenoid implant. Ho and colleagues, aiming for a component perpendicular to the plane of the scapula, reported central peg osteolysis in 20 of 66 shoulders (30%) at a mean of 3.8 years; the series above, in which no attempt was made to change version, reported 10 of 71 (14%) at a mean of 2.5 years [10]. The commentator asks whether the additional reaming accounts for the difference, notes that the two groups also differed in how the peg holes were prepared and in the use of an anteriorly eccentric humeral head, and observes that single-surgeon series of this size will differ for reasons that are difficult to separate [23]. The follow-up is not the same in the two reports and the comparison is not controlled. However, the question is the right one.
What has been reported for different approaches to glenoid version.
Conservative, non- corrective glenoid reaming with a standard all-polyethylene component
The residual cartilage over the anterior glenoid is removed, the ridge between the paleoglenoid and the neoglenoid is taken down with a burr, and the reamer is then oriented equidistant from the anterior and posterior edges of the glenoid so as to remove the least possible bone. Version change is not the target; a single, supported concavity that matches the back of the component is.
In 66 shoulders with B2 (n = 40) and B3 (n = 26) glenoids treated this way, mean version was essentially unchanged (−21 to −19 degrees for the B2 glenoids, P = .267; −26 to −23 degrees for the B3 glenoids, P = .129). Humeral head decentering was corrected all the same, from −14% to −1% in the B2 glenoids and from −4% to −1% in the B3 glenoids. The SST improved from 3.2 to 9.9 and from 3.0 to 9.4 points. Osseous ingrowth between the flanges of the central peg was seen in 83% of the B2 and 81% of the B3 glenoids. None of these results was inferior to those for the other glenoid types treated by the same surgeon over the same period [13].
A 2026 systematic review of anatomic arthroplasty in B2 and B3 glenoids, pooling six studies and 443 patients treated with a mixture of techniques, found that both groups improved in patient-reported outcomes and pain. Radiolucency was reported more often in the B2 glenoids (31%) than in the B3 (22%), and the B3 glenoids remained more medialized after surgery [24]. In these pooled series, the more advanced deformity was not the one with more radiolucency.
A second group has reported the same approach from a different starting point. In 79 shoulders with mild to moderate B2 or B3 deformity — defined as the ability to template a component within 10 degrees of premorbid version — in which the surgeon had an augmented component available and chose a standard one, the median Penn Shoulder Score at a median of 4.3 years was 96, posterior displacement of the humeral head fell from 66% to 12%, and grade-1 anchor peg osteolysis was present in 8 of 79 (10%). Walch classification, preoperative version, inclination, and joint-line medialization were not significantly associated with the Penn score or with osteolysis, and osteolysis was not associated with a lower Penn score [16].
When the head decenters posteriorly by more than 50% of the width of the glenoid with trial components in place, an anteriorly eccentric humeral head can recenter it: in 33 such arthroplasties, radiographic decentering fell from 10.4% to 0.9% and the SST improved from 4.8 to 10.0, with no postoperative instability [12]. Two details from that series are worth taking to the operating room. Preoperative decentering did not correlate with preoperative glenoid version (R² = 0.0009), and a third of the shoulders that needed an eccentric head had shown little or no decentering on preoperative imaging — the instability appeared only after exposure, capsular release, and osteophyte resection. The decision belongs at the trial, not on the scan.
Figure 7. Posterior translation of the humeral head by more than 50% of the width of the glenoid with the trial components in place at surgery. This is the finding that may call for an anteriorly eccentric humeral head.
A second group has reported the same use of the eccentric head in B2 glenoids. In 20 patients in whom the head was rotated anteriorly for excessive posterior decentering found at surgery, radiographic decentering fell from 9.9% to 0.5%, pain, ASES, and SST scores improved significantly at a mean of 48 months, and no shoulder was revised [25].
Asymmetric (high-side) reaming
In a single-surgeon series of 85 shoulders followed for a mean of 9.4 years, survivorship of B2 glenoids treated with asymmetric reaming was 97.6% at 5 years and 94.1% at 10 years, compared with 97.6% and 91.9% for type A glenoids, with no significant difference in revision rate [26]. That series also reports 15-year figures of 73.3% and 83.5%. Those two numbers cannot support a comparison. Only two shoulders in each group remained at risk at 15 years, and the authors state that the entire B2 decline is produced by a single revision at 12.3 years: immediately before it, survivorship was 85.5% for the B2 glenoids against 83.5% for the type A.
Series from other centers with similar follow-up report a wider range. Gerber's group reported 32 shoulders with milder B2 glenoids (mean version −14 degrees) treated with asymmetric reaming limited to 10 degrees of correction and followed for a mean of 9.2 years. Clinical scores improved substantially. The complication rate was 15.6% and the revision rate 12.5%; survivorship free of revision was 94% at 5 years and 85% at 10, and survivorship free of advanced glenoid loosening was 91% and 84% [27]. The Harold series described above reported 5.1% revision at a mean of 8.6 years [20]. The series are small, and none has a comparison group treated without correction.
The cost of high-side reaming is glenoid bone loss. A computational study found that correcting version by eccentric reaming compromises the quality of the bone that remains in B2 glenoids [28], and a simulation study found a high risk of vault perforation when correction is carried far [29]. The meta-analysis above bears on this as well: across the pooled cohorts, shoulders treated with eccentric reaming had a higher complication rate (odds ratio 3.2) and a higher revision rate (odds ratio 6.2) than those treated with noncorrective reaming [18]. In the 5-year series, among the B1 and B2 shoulders, the 12 in which version was changed by more than 10 degrees improved by 4.6 SST points, against 6.0 points in the 69 changed by 10 degrees or less; that difference did not reach significance (P = .134) [11].
A systematic review of 16 studies covering 414 anatomic arthroplasties and 78 reverses performed for B2 glenoids reports better figures for eccentric reaming: a complication rate of 5.6% (5 of 89 shoulders) and a revision rate of 0.9% (1 of 116) among the shoulders in which version was corrected to within 10 degrees of neutral. That second number carries a caveat: the 92 shoulders of the Walch series — the largest eccentric-reaming cohort in the review, and the one reporting a 16.3% revision rate — could not be placed in the eccentric-reaming subgroup, because that report did not separate out the shoulders that also received a bone graft. They were assigned instead to a subgroup labeled unspecified, whose revision rate is 16.3%. The 0.9% figure describes eccentric reaming with its largest and least favorable series removed from the numerator and the denominator both [30].
A 2024 systematic review of 779 B2 shoulders reports that complications after eccentric reaming fell from 13.2% in series published before 2012 to 3.3% in those published after [31]. The later series were followed for less time (a mean of 51 against 59 months), and techniques, implants, and patient selection all changed over the same period.
Whenever correction is undertaken, a modest one costs less bone.
Posteriorly augmented glenoid components
These components strive to correct retroversion and re-center the head while removing less anterior bone. In pooled data, shoulders treated with posterior augments were more likely to have no radiolucency at all than those treated with either eccentric or noncorrective reaming [18].
The clinical evidence is harder to read. An early systematic review of 239 B2 shoulders reported revision rates of 15.6% for asymmetric reaming, 9.5% for posterior bone grafting, and 0% for posteriorly augmented components. Two figures from the same review qualify the third: mean follow-up in the augmented group was 42 months, and that group had the highest rate of radiolucent lines of the three techniques, at 64.7%, against 28% for asymmetric reaming and 45% for bone grafting [32].
Two series report what augments do over the first few years. With a stepped augment, Ho and colleagues reported 71 shoulders at a median of 2.4 years: version, humeral head centering, motion, and the Penn Shoulder Score all improved. Persistent posterior displacement of the head was associated with greater postoperative retroversion and with fatty infiltration of the teres minor; central peg osteolysis was associated with preoperative joint-line medialization and posterior bone loss; and, after adjustment, a lower Penn score was associated with greater preoperative posterior displacement of the head [33]. With a biconvex augment, a seven-center series reported 86 of 110 consecutive patients at a mean of 35 months. Version was corrected from 19.3 to 7.4 degrees, humeroscapular alignment fell from 69.1% to 53.5%, the ASES score rose from 49.8 to 86.7, 79 of the 86 components showed no radiolucency, and one shoulder was revised, for infection with loosening [15]. Both series show the measurement corrected and the patient improved. Neither has a comparison group, so neither can say whether the patient improved more than with a standard component. One finding in Ho's series concerns something the augment cannot change: the lower Penn scores varied with the posterior displacement of the head that the patient had preoperatively.
Two later reports carry longer follow-up. In B3 glenoids at mid- to long-term follow-up, augmented and standard components did not differ in mean ASES score (93.3 vs. 85.7, P = .217), satisfaction (95.6% vs. 96.8%), forward elevation, or external rotation, and no patient in either group underwent revision [34]. In a prospective series of 10 patients followed with CT at 5 years, augmented components held their correction of both version and head position, the mean Constant score was 78.2, and all 10 patients were satisfied — while 5 of the 10 had CT signs of aseptic glenoid loosening [35].
The B2 systematic review cited in the previous section is worth reading for what it reports about improvement rather than final score. Its augmented subgroup reached good final values — a mean Constant score of 77.4 and a mean ASES score of 90.3 — while recording the lowest mean improvement of any group in the review on both instruments, 36.1 and 47.6 points. Its complication rate was 10.5%, most often posterior displacement of the humeral head [30]. Patients who start higher finish higher.
The most recent pooled analysis, of 36 studies of anatomic and reverse arthroplasty in B2 and B3 glenoids, reports that among the anatomic arthroplasties, posteriorly augmented components were associated with lower complication and revision rates than eccentric or noncorrective reaming [36]. It compares revision rates across series with different lengths of follow-up, and the augmented series have been the shortest: a mean of 42 months in one review and 32.7 months in another, which reported a 2.8% complication rate for augments [32, 31]. Revisions after anatomic arthroplasty accumulate with time. Until the groups are followed for the same length of time, a lower revision rate is not yet evidence of greater durability.
Two more reports bear directly on the question. In a case-matched series from one institution, 37 shoulders with a full-wedge augment and 37 with a standard component were followed for a mean of 3.2 years. The augmented group had more B2 and B3 glenoids (86% against 51%). The two groups improved by similar amounts on every measure of motion and outcome — the ASES score by 41.5 points against 44.6 (P = .5) — and had similar rates of glenoid lucency (54% against 46%) and reoperation (5% against 3%). The 16-degree augment, used in 4 shoulders, had a higher mean Lazarus score than the 8-degree augment (4.2 against 1.1) and is no longer used at that institution; with it excluded, no augmented component required revision [37].
In the second, a stepped augment was used in 29 B2 and 21 B3 glenoids and compared with a standard component in 42 A1 glenoids, with CT before surgery, within 3 months, and at a mean of 2.3 years. Penn Shoulder Scores at follow-up did not differ (median 95.0 for A1, 97.4 for B2, and 96.7 for B3). Central peg osteolysis was found in 5% of the A1, 10% of the B2, and 29% of the B3 shoulders (P = .028). After surgery the head was as well centered on the glenoid in the B2 shoulders as in the A1, but sat further back relative to the scapula, because the B2 components were left in more retroversion [38]. The comparison is between an augment in a worn glenoid and a standard component in an unworn one. It does not show what a standard component would have done in the same B2 shoulders.
The comparison an augment has to win is against a standard component placed without correcting version. In the consecutive series of 458 anatomic total shoulders described below, done that way, the 152 B2 and B3 shoulders did as well as the 162 type A shoulders: final SST 9.5 against 9.1, percentage of maximal possible improvement 73% against 67%, and revision 1.3% against 4.3% [3]. An augment would have to improve on that.
Augmented components correct the measurement reliably. No report to date has shown that they give the patient better comfort or function than a standard component.
Augments, corrective reaming, or neither
No randomized trial and no same-surgeon cohort has compared posterior augments with corrective reaming for the B2 glenoid. The comparisons that exist are made across series, from different surgeons, implants, and eras.
The meta-analysis described above is the most direct. It could not compare the techniques on patient-reported outcomes, because fewer than three cohorts in each group reported the same score; the postoperative ASES of 86.7 for augments comes from a single series of 86 shoulders. Augments did not differ from eccentric reaming in complications (P = .322) or in revision (P = .225). On the radiograph, augments were more likely than eccentric reaming to show no radiolucency at all (odds ratio 15.7) and to be fully seated (odds ratio 3.0), and less likely to show bone growth between the flanges of the central peg (odds ratio 0.07) [18]. The pooled findings favoring augments on revision [36, 32] come from the shortest follow-up in the literature, as discussed above.
The one comparison between techniques in that meta-analysis to reach significance on complications and revision did not involve augments at all. Eccentric reaming had more complications (odds ratio 3.2) and more revisions (odds ratio 6.2) than noncorrective reaming [18]. On the evidence available, neither method of correcting version has been shown to give the patient a better result than the other. Neither has been shown to give a better result than leaving version alone. The trial that would settle the question would assign B2 shoulders to an augment, to eccentric reaming, or to noncorrective reaming, and would measure what the patients report at five years or more.
Reverse total shoulder arthroplasty
The strongest argument for the reverse in this setting is durability over time. In a single-surgeon comparison of eccentric wear patterns with an intact cuff, the anatomic group treated with high-side reaming did better at two years (ASES 85 vs. 80, SST 10 vs. 9.6, forward elevation 162 vs. 151 degrees, satisfaction 95% vs. 93%). By a minimum of seven years the picture had reversed on some measures: the reverse group had a higher ASES score (80 vs. 77), a higher SST score (9.4 vs. 8), and satisfaction of 95% against 82%, and 8% of the anatomic cohort had been revised against 2% of the reverse cohort. The anatomic group kept better external rotation (45 vs. 41 degrees) and internal rotation (74% vs. 54% with full internal rotation) [39].
The operation was chosen by the surgeon at the time of surgery rather than assigned, and the reverse cohort started with more deformity (mean retroversion 23 vs. 17 degrees; mean posterior displacement of the head 78% vs. 69%), which the authors identify as a possible selection bias. Of the original 242 shoulders, 194 were available at final follow-up. And the differences favoring the reverse at final follow-up were 3 ASES points and 1.4 SST points.
Whether differences of that size are large enough for a patient to notice is a question the study does not take up, but the published thresholds answer it. Using an anchor-based method in 326 shoulder arthroplasties, the minimal clinically important difference was 20.9 points for the ASES and 2.4 points for the SST [22]. An analysis of 1,856 arthroplasties from an industry-maintained multicenter registry put the thresholds lower, at 13.6 ASES points and 1.5 SST points [21], and set the substantial clinical benefit — the improvement that separates the patients who called the result much better from those who called it unchanged or worse — at 31.5 ASES points and 3.4 SST points [40]. A 3-point ASES difference falls below the lowest of these thresholds by a factor of four, below the lowest prosthesis-specific value published anywhere — 10.3 for the reverse in that same registry — by more than three, and below substantial clinical benefit by a factor of ten.
The SST difference is best read against a threshold specific to the operation, because the threshold is not the same for every operation. In 887 arthroplasties the SST minimal clinically important difference was 2.3 overall, but 1.6 for the anatomic total shoulder, 2.6 for the ream-and-run, and 3.7 for the reverse [41]. A 1.4-point difference is smaller than the threshold for either of the operations being compared, and it is less than half the threshold for the reverse — the operation the difference favors. One published value does match it: 1.4 for the reverse in that same registry [21], which is the lowest figure reported anywhere for that operation, against 2.9 and 3.7 elsewhere [22, 41]. Matching the lowest threshold published anywhere is not the same as showing a difference the patient would notice.
There is a further reason these thresholds understate what this comparison would have to exceed. The registry that produced the most permissive of them also reports them stratified by length of follow-up, and both rise as follow-up lengthens. Beyond 72 months the SST minimal clinically important difference is 3.3 and the substantial clinical benefit 5.7, against pooled values of 1.5 and 3.4; the ASES threshold is 21.7 against 13.6 [21, 40]. The comparison above is reported at a minimum of seven years. Measured against the values the same investigators derived from the same patients at that length of follow-up, a 1.4-point SST difference falls short by a factor of more than two and a 3-point ASES difference by a factor of seven.
These values were derived for improvement within a patient rather than for a difference between two groups, which is a distinct quantity. They are also less precise than their decimal places suggest. In the same registry table that gives 1.4 for the SST after the reverse, the Constant threshold for that operation is −0.3, and the thresholds for external rotation, abduction, and forward flexion are −5.3, −1.9, and −2.9 degrees [21]. A negative threshold means the patients who called themselves better had improved less than the patients who called themselves unchanged — the method returning noise in a subgroup rather than a measurement. Values carried to one decimal place and applied across studies deserve to be read with that in mind. But no reading of them turns 3 ASES points and 1.4 SST points into something a patient could feel.
A multicenter matched-cohort analysis of B2 and B3 shoulders puts the same comparison in terms of the minimal clinically important difference. Most patients reached the MCID with either operation (anatomic 95.0%, reverse 98.0%, P = .436), 82% and 90% respectively reached the threshold for substantial clinical benefit (P = .292), and overall complication rates did not differ (P = .781). Pain scores were lower after the reverse; internal rotation was better after the anatomic [42]. The difference between the two operations is small next to the difference either one makes for the patient.
The pooled evidence is thinner than the confidence with which the reverse is now offered for this problem. In that same systematic review, the entire published experience of the reverse for B2 shoulders with an intact cuff came to 78 shoulders across five studies at a mean follow-up of 44.8 months. The pooled complication rate was 6%, with a confidence interval running from 0% to 28%, and the pooled revision rate 1%, with an interval from 0% to 15% — both calculated from two studies and 43 shoulders. Improvement in the ASES score differed between the two operations by a single point, 49.2 for the reverse against 48.2 for the anatomic. The Constant score improved by 7.8 points more after the reverse, which does exceed the minimal clinically important difference for that instrument, 5.7 points, and falls well short of its substantial clinical benefit of 19.1 [21, 40]. And the patients receiving the reverse were nine years older, 72.8 against 63.7 [30].
A larger, updated review has since been published. Aleisawi and colleagues pooled 36 studies of B2 and B3 glenoids with an intact cuff: 1,349 anatomic and 478 reverse arthroplasties, followed for a mean of 4.9 and 4.2 years, in patients with a mean age of 68 and 72. Gains in motion were similar. Mean improvement in the ASES score was 44 points after the anatomic and 50 after the reverse, and in the Constant score 43 and 46. Pooled complication rates were 7.8% and 4.4%, and revision rates 5.2% and 1.6% [36]. The authors concluded that both operations are effective and that the anatomic carries a modestly higher risk of complication and revision.
A 6-point difference in ASES improvement is below every ASES threshold cited above. The difference in revision rate is the more important finding and deserves to be taken seriously. Two things belong beside it. The reverse groups were followed for less time, in patients four years older. And a revision rate counts operations performed, not shoulders that have failed. A failed anatomic arthroplasty can be converted to a reverse; the options after a failed reverse are fewer and less attractive. That revision rates may therefore understate failure of the reverse is an inference, not a finding, and it is one more reason to compare the two operations by what the patients report.
Figure 8. ASES points for B2 and B3 shoulders with an intact cuff. Each operation improves the score by more than the substantial clinical benefit; the reported differences between the two operations fall below every published minimal clinically important difference.
For an active patient in their sixties with an intact cuff, the reverse permanently changes the mechanics of the shoulder, gives up some internal rotation, and commits the patient to a construct whose behavior at 15 and 20 years is not yet known.
The more active patient
For patients who prefer to avoid a polyethylene glenoid component altogether, the ream-and-run — concentric reaming of the glenoid combined with a humeral hemiarthroplasty — remains an option for those who understand the longer and more demanding rehabilitation. The anteriorly eccentric humeral head applies here in the same way it does to the total shoulder; in the series above, 24 of the 33 arthroplasties were hemiarthroplasties with nonprosthetic glenoid arthroplasty [12].
Durability at ten years has now been reported. Of 127 consecutive patients eligible for a minimum 10-year survey, 63 (50%) responded: 34 ream-and-run and 29 anatomic total shoulders. In the ream-and-run group the SST improved from 5.4 to 10.3 and the pain score from 6.5 to 0.9, with a mean of 83% of the maximum possible improvement. Twenty-eight of the 34 (82%) exceeded the ream-and-run threshold of 2.6 SST points, and 27 (79%) exceeded the corresponding threshold of 38.2% of maximum possible improvement [41, 43]. Five patients (15%) were reoperated: four open revisions, at 6, 7, 27, and 72 months, each to another hemiarthroplasty, and one manipulation under anesthesia at 6 months. None had humeral loosening and none was converted to a total shoulder. In the total shoulder group the SST improved from 3.8 to 8.9 and the pain score from 6.6 to 1.2, and no patient was reoperated [43].
The most useful number in that report is one the series takes from itself. The mean SST at ten years, 10.3, is the same as the mean SST at two years in an earlier cohort of 101 ream-and-run patients, and the reoperation rate is essentially the same as well, 14% at two years against 15% at ten. Revision after this operation happens early; the shoulders that get past two years appear to hold what they gained. The proportion of patients exceeding the threshold holds as well: 84% of 330 ream-and-run patients exceeded it at two years, against 82% at ten [41, 43].
The percentage of maximum possible improvement is worth attention in this operation in particular. Ream-and-run patients come to surgery with more function than any other arthroplasty group — a mean preoperative SST of 5.0, against 1.7 for the reverse — so they have less room to gain in absolute points, and an absolute threshold understates what they get. The threshold for maximum possible improvement runs the other way for the same reason: 38.2% for the ream-and-run against 21.8% for the anatomic total shoulder [41].
What that report cannot do is compare the two operations. The ream-and-run patients were younger (60 vs. 68 years), 97% male against 41%, and healthier by ASA class. The comparison that would settle the question — patient-reported comfort and function against the alternatives, in patients alike enough to be compared — has not been made.
Two further reports bear on the B2 shoulder in particular. In 201 ream-and-run arthroplasties followed for a mean of 7.6 years, the most common glenoid morphology was B2, at 44%, with another 15% B1. The SST improved from 4.9 to 10.2, 82% of patients exceeded the threshold, and 11% underwent open revision. Glenoid morphology was not associated with whether a patient exceeded the threshold, and it was not associated with revision. What was associated with revision was youth: the patients who returned to the operating room averaged 54 years against 60 for those who did not [44].
The B2 shoulder in each of the three operations
No study has compared the ream-and-run, the anatomic total shoulder, and the reverse in B2 shoulders. What can be done is to set side by side what each operation has reported for B2 shoulders in particular, and who received it.
Ream-and-run. The B2 is the most common glenoid among the patients who choose this operation. In 544 concurrent ream-and-run and total shoulder arthroplasties at one institution, 46% of the ream-and-run shoulders were B2 against 28% of the total shoulders, and retroversion was greater (19 against 15 degrees). At two years the SST was 10.0 and 9.5 and the percentage of maximum possible improvement 72% and 73% [5]. The 30 shoulders with retroversion, posterior decentering, and biconcavity described above improved from an SST of 5 to 10 at a mean of 3 years, with no posterior instability; two were revised to a total shoulder within the first year because the patients were dissatisfied with the rehabilitation [14]. In 201 ream-and-run arthroplasties, 44% of them B2, glenoid morphology was not associated with outcome at a mean of 7.6 years [44].
Anatomic total shoulder. Without changing version: in a consecutive single-surgeon series of 458 anatomic total shoulders, the glenoid was reamed to a single concavity without effort to correct retroversion and a standard, non-augmented all-polyethylene component was used. At a mean follow-up of 6.7 years, the 100 B2 shoulders had the highest final SST (9.7) and percentage of maximal possible improvement (75%) of any glenoid type, and the B2 and B3 shoulders together were revised at 1.3% against 4.3% for type A (P = .307) [3]. In 40 B2 glenoids the SST improved from 3.2 to 9.9 [13], and among the B1 and B2 shoulders of the 5-year series, outcome did not differ above and below 15 degrees of postoperative retroversion [11]. A second group reported 51 B2 glenoids treated with noncorrective concentric reaming, 45 of them followed for a mean of 4.9 years: the ASES score improved from 52.5 to 79.6, survivorship was 95%, and reoperation is given as 4 patients in the abstract and 6 (11.8%) in the results [45]. With corrective reaming, the reports above range from 16% revision at a mean of 77 months [17] to survivorship of 85% to 94% at ten years [26, 27], and 5.1% revision at a mean of 8.6 years [20]. The augmented-component series do not yet extend beyond five years.
Reverse. The principal series confined to B2 shoulders is from Walch's group: 27 reverses in patients with a mean age of 74 years, mean preoperative retroversion of 32 degrees, and mean humeroscapular alignment of 87%. At a mean of 54 months the Constant score had improved from 31 to 76. Four patients (15%) had complications, one an early glenoid loosening and three neurologic; 37% had scapular notching; posterior instability did not recur [46]. The rest of the reverse evidence for this problem comes from comparisons with the anatomic total shoulder [39, 42]; in the two systematic reviews, the reverse patients were older than the anatomic patients by nine years and by four [30, 36].
Three things can be said. In every one of these series, the improvement the operation produced was several times the minimal clinically important difference. In the ream-and-run and noncorrective anatomic series, being a B2 did not predict a worse result. And the differences reported between operations are small next to the difference each operation makes.
One thing cannot be said: which of the three is best for a particular patient with a B2 glenoid. Each operation's B2 results come from a different population. In the concurrent series, the ream-and-run patients averaged 58 years and were 92% male, and the total shoulder patients averaged 67 [5]; the reverse patients in the B2 literature averaged 72 to 74 [30, 36, 46].
A word about planning technology
The number these technologies aim at is not measured the same way from one study to the next. A systematic review of 61 studies (17,070 arthroplasties) relating preoperative glenoid version to an outcome found that only 27 (44%) described how version was measured. Of the 43 that measured postoperative version, 35% used a different imaging modality after surgery than before. Three-dimensional CT was used in 12.5% of studies published in 2012–2014 and in 52% of those published in 2021–2023 [9]. A target measured this inconsistently is difficult to relate to outcomes, and studies that use it are difficult to pool.
Three-dimensional CT is useful for knowing how much bone is present and where the vault is likely to be perforated. That is a different use from setting a version target. Planning software, custom guides, navigation, and robotic assistance add cost, radiation, and operative time in the service of reaching a number that has not been shown to predict how the patient will do [10, 11, 18]. In 8,117 arthroplasties in a large integrated health system, neither preoperative CT (hazard ratio 1.22, 95% confidence interval 0.87 to 1.72) nor patient-specific instrumentation (hazard ratio 1.44, 0.71 to 2.92) reduced the risk of aseptic revision. Mean follow-up was 2.9 years, which the authors note is short against a 5-year glenoid failure rate of 7.4% for the anatomic total shoulder in the Australian registry — the interval over which these technologies would show their value has not yet been observed [47].
An approach to the B2 shoulder
The goal of an anatomic arthroplasty, whether a ream-and-run or a total shoulder, is an articulation that moves freely and is stable. It is not to restore the premorbid anatomy of the glenoid or the humerus. Each step below serves that goal.
• Assess the rotator cuff, the remaining glenoid bone stock, and the patient's age and activity.
• Resect the humeral head at the anatomic neck and remove all osteophytes.
• Limit the capsular release to the midsagittal plane so that posterior capsular support is preserved.
• Take the glenoid ridge down with a burr, then ream minimally and concentrically to a single concavity preserving subchondral bone. Accept the residual retroversion. Reaming is sufficient when a round-backed trial does not tip under eccentric load.
• If the patient desires an anatomic total shoulder, use a standard all-polyethylene component and seat it completely on well-prepared bone.
• Select a humeral head thickness that allows no more than 50% posterior translation and no more than 60 degrees of internal rotation with the arm abducted. If excessive posterior decentering (the humeral head can be displaced substantially more than 50% of the width of the glenoid) is found when the trial components are in place, consider an anteriorly eccentric humeral head; add a rotator interval plication if that is not enough.
• Reserve the reverse for rotator cuff deficiency, for bone that will not support a glenoid component, and for some older, lower-demand patients.
Where we are
The B2 glenoid is commonly found in shoulder arthritis. Patients with B2 glenoids vary widely: by age, sex, activity expectations, cuff integrity, range of motion, preoperative pathoanatomy (version, decentering, severity of biconcavity), bone quality and intraoperative findings. We can offer these patients a number of different procedures, including ream-and-run, pyrocarbon hemiarthroplasty, anatomic total shoulder (with or without corrective reaming and with or without augmented glenoid components), and reverse total shoulder.
The task for the surgeon is to match each patient with the procedure most likely to durably improve their comfort and function.
One thing for sure
The author has no financial relationships with any orthopaedic device company.
Follow on twitter/X: https://x.com/RickMatsen
Follow on facebook: https://www.facebook.com/shoulder.arthritis
Follow on LinkedIn: https://www.linkedin.com/in/rick-matsen-88b1a8133//
References
1. Walch G, Badet R, Boulahia A, Khoury A. Morphologic study of the glenoid in primary glenohumeral osteoarthritis. J Arthroplasty. 1999;14:756-760. doi:10.1016/S0883-5403(99)90232-2
2. Vo KV, Hackett DJ, Gee AO, Hsu JE. Classifications in brief: Walch classification of primary glenohumeral osteoarthritis. Clin Orthop Relat Res. 2017;475:2335-2340. doi:10.1007/s11999-017-5317-6
3. Schiffman C, Chin E, Whitson A, Hsu J, Matsen FA 3rd. Anatomic shoulder arthroplasty: a single surgeon’s consecutive series of four hundred and fifty eight patients. Int Orthop. 2026. doi:10.1007/s00264-026-07017-5
4. Collins AP, Sheth MM, Stenson JF, Kahsai EA, Khoo KJ, Ogunleye T, Whitson AJ, Matsen FA 3rd, Hsu JE. Characterizing glenoid wear after hemiarthroplasty with concentric glenoid reaming: a study of 113 arthroplasties at a mean of 6.7 years of follow-up. J Shoulder Elbow Surg. 2026;35:995-1002. doi:10.1016/j.jse.2025.08.014
5. Matsen FA 3rd, Whitson A, Jackins SE, et al. Ream and run and total shoulder: patient and shoulder characteristics in five hundred forty-four concurrent cases. Int Orthop. 2019;43:2105-2115. doi:10.1007/s00264-019-04352-8
6. Matsen FA 3rd, Gupta A. Axillary view: arthritic glenohumeral anatomy and changes after ream and run. Clin Orthop Relat Res. 2014;472:894-902. doi:10.1007/s11999-013-3327-6
7. von Eisenhart-Rothe R, Müller-Gerbl M, Wiedemann E, Englmeier KH, Graichen H. Functional malcentering of the humeral head and asymmetric long-term stress on the glenoid: potential reasons for glenoid loosening in total shoulder arthroplasty. J Shoulder Elbow Surg. 2008;17:695-702. doi:10.1016/j.jse.2008.02.008
8. Nyffeler RW, Jost B, Pfirrmann CW, Gerber C. Measurement of glenoid version: conventional radiographs versus computed tomography scans. J Shoulder Elbow Surg. 2003;12:493-496. doi:10.1016/S1058-2746(03)00181-2
9. Varady NH, Bram JT, Chow J, Taylor SA, Dines JS, Fu MC, Ode G, Dines DM, Gulotta LV, Brusalis C. Inconsistencies in measuring glenoid version in shoulder arthroplasty: a systematic review. J Shoulder Elbow Surg. 2025;34:639-649. doi:10.1016/j.jse.2024.08.020
10. Service BC, Hsu JE, Somerson JS, Russ SM, Matsen FA 3rd. Does postoperative glenoid retroversion affect the 2-year clinical and radiographic outcomes for total shoulder arthroplasty? Clin Orthop Relat Res. 2017;475:2726-2739. doi:10.1007/s11999-017-5433-3
11. Sheth MM, Mills ZD, Dasari SP, Whitson AJ, Matsen FA 3rd, Hsu JE. Anatomic total shoulder arthroplasty for posteriorly eccentric and concentric osteoarthritis: a comparison at a minimum 5-year follow-up. J Shoulder Elbow Surg. 2025;34:473-483. doi:10.1016/j.jse.2024.04.026
12. Hsu JE, Gee AO, Lucas RM, Somerson JS, Warme WJ, Matsen FA 3rd. Management of intraoperative posterior decentering in shoulder arthroplasty using anteriorly eccentric humeral head components. J Shoulder Elbow Surg. 2016;25:1980-1988. doi:10.1016/j.jse.2016.02.027
13. Matsen FA 3rd, Whitson AJ, Somerson JS, Hsu JE. Anatomic total shoulder arthroplasty with all-polyethylene glenoid component for primary osteoarthritis with glenoid deficiencies. JB JS Open Access. 2020;5(4):e20.00002. doi:10.2106/JBJS.OA.20.00002
14. Matsen FA 3rd, Warme WJ, Jackins SE. Can the ream and run procedure improve glenohumeral relationships and function for shoulders with the arthritic triad? Clin Orthop Relat Res. 2015;473:2088-2096. doi:10.1007/s11999-014-4095-7
15. Garrigues GE, Quigley RJ, Johnston PS, Spencer E, Walch G, Neyton L, Kelly J, Schrumpf M, Gillespie R, Sears BW, Hatzidakis AM, Lau B; PerFORM Investigators. Early clinical and radiographic outcomes of anatomic total shoulder arthroplasty with a biconvex posterior augmented glenoid for patients with posterior glenoid erosion: minimum 2-year follow-up. J Shoulder Elbow Surg. 2022;31:1729-1737. doi:10.1016/j.jse.2021.12.047
16. Cogan CJ, Patel M, Guevara J, Sahoo S, Jun BJ, Zhang C, Shemo C, Baker A, Qi A, Entezari V, Iannotti JP, Ho JC, Ricchetti ET. Short-term outcomes of anatomic total shoulder arthroplasty with nonaugmented glenoid component for Walch B2 and B3 glenoid morphology. J Shoulder Elbow Surg. 2025;34:S10-S21. doi:10.1016/j.jse.2025.01.023
17. Walch G, Moraga C, Young A, Castellanos-Rosas J. Results of anatomic nonconstrained prosthesis in primary osteoarthritis with biconcave glenoid. J Shoulder Elbow Surg. 2012;21:1526-1533. doi:10.1016/j.jse.2011.11.030
18. Checketts JX, Sanchez B, Norris G, Williamson TK, Hachadorian ME, Hsu JE, Schiffman CJ, Matsen FA 3rd. Does postoperative glenoid component retroversion following anatomic total shoulder arthroplasty affect clinical outcomes? A systematic review and meta-analysis. J Shoulder Elbow Surg. 2026;35:1103-1116. doi:10.1016/j.jse.2025.08.011
19. Rutledge JC, Dey Hazra RO, Geissbuhler AR, Yamaura K, Dey Hazra ME, Hanson JA, Rupp MC, Millett PJ. Does glenoid version and its correction affect outcomes in anatomic shoulder arthroplasty? A systematic review. J Shoulder Elbow Surg. 2024;33:e384-e399. doi:10.1016/j.jse.2023.10.031
20. Harold RE, Sweeney PT, Torchia MT, Chamberlain AM, Keener JD. Total shoulder arthroplasty in patients with a B2 glenoid addressed with corrective reaming: mean 8-year follow-up. J Shoulder Elbow Surg. 2023;32:S8-S16. doi:10.1016/j.jse.2022.12.019
21. Simovitch R, Flurin PH, Wright T, Zuckerman JD, Roche CP. Quantifying success after total shoulder arthroplasty: the minimal clinically important difference. J Shoulder Elbow Surg. 2018;27:298-305. doi:10.1016/j.jse.2017.09.013
22. Tashjian RZ, Hung M, Keener JD, Bowen RC, McAllister J, Chen W, Ebersole G, Granger EK, Chamberlain AM. Determining the minimal clinically important difference for the American Shoulder and Elbow Surgeons score, Simple Shoulder Test, and visual analog scale (VAS) measuring pain after shoulder arthroplasty. J Shoulder Elbow Surg. 2017;26:144-148. doi:10.1016/j.jse.2016.06.007
23. Pearl ML. CORR Insights®: Does postoperative glenoid retroversion affect the 2-year clinical and radiographic outcomes for total shoulder arthroplasty? Clin Orthop Relat Res. 2017;475:2740-2743. doi:10.1007/s11999-017-5457-8
24. Whong Z, Chow AK, Whitaker S, Smith MS, Cuthbert AS, Setliff J, Cyrus J, Vanderbeck J, Satalich J. Outcomes of anatomic total shoulder arthroplasty in B2 versus B3 glenoid morphology: a systematic review. J Orthop. 2026;76:230-238. doi:10.1016/j.jor.2026.04.003
25. Chamberlain AM, Orvets N, Patterson B, Chalmers P, Gosselin M, Salazar D, Keener JD. Total shoulder arthroplasty with an anterior-offset humeral head in patients with a B2 glenoid. JSES Int. 2020;4:638-643. doi:10.1016/j.jseint.2020.02.001
26. Cirino CM, Kantrowitz DE, Boucher T, Patel A, Duey AH, Cagle PJ, Flatow EL. Anatomic total shoulder arthroplasty for B2 glenoids treated with asymmetric reaming has equivalent survivorship and patient outcomes compared with type A glenoids at a mean 9-year follow-up. J Shoulder Elbow Surg. 2024;33:2392-2399. doi:10.1016/j.jse.2024.03.030
27. Hinse S, Pastor T, Hasler A, Ernstbrunner L, Wieser K, Gerber C. Mid- to long-term clinical and radiological results of anatomic total shoulder arthroplasty in patients with B2 glenoids. JSES Int. 2023;7:464-471. doi:10.1016/j.jseint.2023.01.006
28. Chen X, Reddy AS, Kontaxis A, Choi DS, Wright T, Dines DM, et al. Version correction via eccentric reaming compromises remaining bone quality in B2 glenoids: a computational study. Clin Orthop Relat Res. 2017;475:3090-3099. doi:10.1007/s11999-017-5510-7
29. Aleem AW, Orvets ND, Patterson BC, Chamberlain AM, Keener JD. Risk of perforation is high during corrective reaming of retroverted glenoids: a computer simulation study. Clin Orthop Relat Res. 2018;476:1612-1619. doi:10.1007/s11999.0000000000000302
30. Reahl GB, Abdul-Rassoul H, Kim RL, Ardavanis KS, Novikov D, Curry EJ, Galvin JW, Eichinger JK, Li X. Anatomic vs. reverse shoulder arthroplasty for the treatment of Walch B2 glenoid morphology: a systematic review and meta-analysis. JSES Rev Rep Tech. 2021;1:317-328. doi:10.1016/j.xrrt.2021.08.003
31. Hollo D, Raniga S, Cadosch D, Müller AM, Bokor DJ. Improved surgical techniques and outcomes with anatomic total shoulder arthroplasty managing B2 glenoids in midterm follow-up. Obere Extremität. 2024;19:44-53. doi:10.1007/s11678-023-00780-6
32. Luedke C, Kissenberth MJ, Tolan SJ, Hawkins RJ, Tokish JM. Outcomes of anatomic total shoulder arthroplasty with B2 glenoids: a systematic review. JBJS Rev. 2018;6(4):e7. doi:10.2106/JBJS.RVW.17.00112
33. Ho JC, Amini MH, Entezari V, Jun BJ, Alolabi B, Ricchetti ET, Iannotti JP. Clinical and radiographic outcomes of a posteriorly augmented glenoid component in anatomic total shoulder arthroplasty for primary osteoarthritis with posterior glenoid bone loss. J Bone Joint Surg Am. 2018;100:1934-1948. doi:10.2106/JBJS.17.01282
34. Kohan EM, Hendy BA, Kowal LL, Kirsch J, Gregori P, Williams GR, Namdari S. Mid- to long-term outcomes of augmented and nonaugmented anatomic shoulder arthroplasty in Walch B3 glenoids. J Shoulder Elbow Surg. 2022;31:S103-S109. doi:10.1016/j.jse.2021.12.016
35. Goetti P, Lannes X, Becce F, Farron A, Hoffmann L, Terrier A. Five-year outcomes of anatomic total shoulder arthroplasty with Aequalis Perform+ posterior augmented glenoid implants in Walch B2 and B3 glenoids. J Shoulder Elbow Surg. 2026;35:e13-e21. doi:10.1016/j.jse.2025.04.001
36. Aleisawi H, Kruse C, Nucci N, Alturki N, Abdel Khalik H, Athwal GS, Khan M. Outcomes of anatomic versus reverse shoulder arthroplasty for B2 and B3 glenoids with an intact rotator cuff: an updated systematic review and proportional meta-analysis. Shoulder Elbow. 2026;18(3). Published online July 17, 2025. doi:10.1177/17585732251359590
37. Priddy M, Zarezadeh A, Farmer KW, Struk AM, King JJ 3rd, Wright TW, Schoch BS. Early results of augmented anatomic glenoid components. J Shoulder Elbow Surg. 2019;28:S138-S145. doi:10.1016/j.jse.2019.04.014
38. Iannotti JP, Jun BJ, Derwin KA, Ricchetti ET. Stepped augmented glenoid component in anatomic total shoulder arthroplasty for B2 and B3 glenoid pathology: a study of early outcomes. J Bone Joint Surg Am. 2021;103:1798-1806. doi:10.2106/JBJS.20.01420
39. Cuff DJ, Simon P, Patel JS, Munassi SD. Anatomic shoulder arthroplasty with high side reaming versus reverse shoulder arthroplasty for eccentric glenoid wear patterns with an intact rotator cuff: comparing early versus midterm outcomes with minimum 7 years of follow-up. J Shoulder Elbow Surg. 2023;32:972-979. doi:10.1016/j.jse.2022.10.017
40. Simovitch R, Flurin PH, Wright T, Zuckerman JD, Roche CP. Quantifying success after total shoulder arthroplasty: the substantial clinical benefit. J Shoulder Elbow Surg. 2018;27:903-911. doi:10.1016/j.jse.2017.12.014
41. McLaughlin RJ, Whitson AJ, Panebianco A, Warme WJ, Matsen FA 3rd, Hsu JE. The minimal clinically important differences of the Simple Shoulder Test are different for different arthroplasty types. J Shoulder Elbow Surg. 2022;31:1640-1646. doi:10.1016/j.jse.2022.02.010
42. Polisetty TS, Swanson DP, Hart PJ, Cannon DJ, Glass EA, Jawa A, Levy JC, Kirsch JM. Anatomic and reverse shoulder arthroplasty for management of type B2 and B3 glenoids: a matched-cohort analysis. J Shoulder Elbow Surg. 2023;32:1629-1637. doi:10.1016/j.jse.2023.02.125
43. Sharareh B, Whitson AJ, Matsen FA 3rd, Hsu JE. Minimum 10-year follow-up of anatomic total shoulder arthroplasty and ream-and-run arthroplasty for primary glenohumeral osteoarthritis. J Shoulder Elbow Surg. 2024;33:1276-1284. doi:10.1016/j.jse.2023.08.028
44. Stenson JF, Collins AP, Yao JJ, Sharareh B, Whitson AJ, Matsen FA 3rd, Hsu JE. Factors associated with success of ream-and-run arthroplasty at a minimum of 5 years. J Shoulder Elbow Surg. 2023;32:S85-S91. doi:10.1016/j.jse.2023.01.024
45. Grantham WJ, Dekker TJ, Lacheta L, Horan MP, Goldenberg BT, Elrick BP, Millett PJ. Total shoulder arthroplasty outcomes after noncorrective, concentric reaming of B2 glenoids. JSES Int. 2020;4:644-648. doi:10.1016/j.jseint.2020.04.006
46. Mizuno N, Denard PJ, Raiss P, Walch G. Reverse total shoulder arthroplasty for primary glenohumeral osteoarthritis in patients with a biconcave glenoid. J Bone Joint Surg Am. 2013;95:1297-1304.
47. Navarro RA, Chan PH, Prentice HA, Pearl M, Matsen FA 3rd, McElvany MD. Use of preoperative CT scans and patient-specific instrumentation may not improve short-term adverse events after shoulder arthroplasty: results from a large integrated health-care system. JB JS Open Access. 2023;8:e22.00139. doi:10.2106/JBJS.OA.22.00139
