➢ Reverse total shoulder arthroplasty can restore forward elevation in pseudoparalytic shoulders, can reduce pain, and can increase quality of life for patients; however, there are a number of issues with regard to the design and ideal implantation technique of reverse total shoulder arthroplasties that remain controversial.
➢ Glenospheres with a Grammont, medialized center of rotation and designs with a more lateralized center of rotation have specific pros and cons. Medialized designs have historically decreased glenosphere loosening but increase scapular notching, and lateralized designs decrease notching and improve range of motion but historically have resulted in increased baseplate failure. As such, a concrete conclusion with regard to medialization or lateralization of the glenosphere cannot be made at this time.
➢ Caudal, inferior placement of the glenosphere increases internal rotation, external rotation, abduction, and adduction while reducing notching. Superior baseplate positioning and tilt are correlated with baseplate failure and therefore should be avoided. It remains unclear whether neutral or slightly inferior tilt is preferred.
➢ The ideal humerosocket inclination angle remains unclear. The 135° designs appear to have a lower rate of scapular notching than the Grammont-style 155° humerosockets. The implications of this angle for instability remain unclear as, to our knowledge, there have been no randomized comparisons and the largest meta-analysis was confounded by other implant variables.
In the past decade, use of reverse total shoulder arthroplasty has risen dramatically because of the ability of this procedure to restore active anterior elevation in pseudoparalytic shoulders, to reduce pain, and to increase quality of life1-3. Although reverse total shoulder arthroplasty has been successful in improving patient function while decreasing pain, the reported complication rate remains high and controversies still exist. This review seeks to highlight the current state of the literature on key debates in reverse total shoulder arthroplasty implant positioning and design.
Glenosphere Placement: Medial, Lateral, and BIO-RSA
The results of early reverse total shoulder arthroplasty were poor, with many of the highly constrained systems failing4. Extensive glenosphere loosening and scapular fractures, despite emphasis on careful glenoid placement and robust scapular fixation, prevented the widespread adoption of the reverse total shoulder arthroplasty in the 1970s and 1980s4.
Paul Grammont proposed an alternative of minimizing the forces on the implant baseplate rather than continually attempting to further increase scapular fixation. In 1985, he debuted his Version 1 prosthesis, a two-thirds glenosphere that medialized the center of rotation but left it still lateral to the glenoid face. This design served to reduce shear forces on the baseplate as well as improve deltoid fiber recruitment, increasing the moment arm by 20% to 42%5. In 1991, Grammont released his Delta III implant, an exact hemisphere, which put the center of rotation directly on the glenoid face, further minimizing shear forces6. Both Grammont’s Version 1 and Delta III improved on previous rates of glenosphere loosening, with this complication observed in only 4.1% of Grammont’s prostheses6.
Despite the improved rates of glenosphere loosening, scapular notching was a heretofore unseen phenomenon created by this medialized, reverse ball-and-socket design (Fig. 1). A consequence of humerosocket adduction resulting in impingement on the lateral scapular pillar, scapular notching occurs with a rate of 44% to 96%, with most studies showing around 66%6-9. The clinical implications of scapular notching are unclear, with some studies noting no difference in clinical outcomes8,10 as other studies of similar methodology noted a negative influence on clinical parameters such as strength and anterior elevation correlated with the presence of a radiographic notch9,11.
Furthermore, there has been debate on whether the notch increases in size over time or whether the notching no longer progresses once space is made in the scapula to accommodate the humerosocket. Radiographic follow-up studies have been split among notching steadily progressing over time, notching occurring initially and then no longer progressing, and notching stabilizing in some patients and progressing over time in others8,9,11.
In an effort to eliminate the notching observed with the Grammont, medialized center of rotation design, surgeons continued to experiment with reverse total shoulder arthroplasty designs with a center of rotation lateral to the glenoid face. In a 2008 virtual modeling study, Gutiérrez et al. revealed a lateralized sphere to be the most important factor for overall impingement-free range of motion and an important factor to minimize adduction deficit12. This bench-top model was confirmed in a 2005 clinical series when Frankle et al. observed no notching in any patients with a minimum follow-up of five years using a lateralized glenosphere design13. Subsequent studies using a lateralized center of rotation design also found decreased scapular notching; however, these gains were not without complications14,15.
The lateralized design, where the center of rotation is lateral to the glenoid face, although potentially medial to the native center of rotation, leads to stress concentration at the bone-implant interface and an increase in shear forces. As discussed above, this has historically led to increased baseplate failure (Fig. 2)4,6. A 1-body weight compressive and shear load force analysis has shown a 69% increase in the baseplate-bone interface with the lateralized baseplate compared with the medialized Delta III16. The 2005 clinical series by Frankle et al. with a lateralized center of rotation proved no exception, with 12% of patients having baseplate failure at a mean time of twenty-one months13. In a 2008 clinical series again evaluating Dr. Frankle’s patients with an improved design incorporating stronger, fixed-angle fixation with four 5.0-mm locking screws instead of four 3.5-mm compression screws, Cuff et al. published excellent results with a low rate of baseplate failure17. It appears that locking-screw technology may allow implants with a lateralized center of rotation to withstand, at least in large part, the increased implant forces inherent in this design17.
In an attempt to avoid both the limited impingement-free range of motion and notching seen with medialized designs and the increased baseplate forces seen with lateralized designs, a third, hybrid, option is the bony increased-offset reversed shoulder arthroplasty (BIO-RSA) (Figs. 3-A and 3-B)14. This method involves placing an autologous bone graft harvested from the humeral head beneath the baseplate, effectively lengthening the glenoid14. This lateralizes the prosthesis with respect to the location of potential notching on the lateral border of the scapula, while keeping the center of rotation at the glenoid bone-implant interface14. The cancellous humeral head autograft is under even compression from the baseplate, and graft incorporation rates have been high, ranging from 98% to 100%14,18.
Although controversy remains, the lateralization of BIO-RSA may, indeed, lead to less impingement of the humerosocket on the scapula, in turn improving impingement-free range of motion and decreasing notching. Indeed, the series by Boileau et al. showed improvement in internal rotation up the back, external rotation with the arm at the side, and decreased notching14. Interestingly, a comparative cohort study of BIO-RSA and the Grammont design showed a decrease in notching with the BIO-RSA design but no significant change in outcome parameters or range of motion between the two designs18.
Overall, there has been no clear consensus with regard to glenosphere placement with a medialized center of rotation, a lateralized one, or BIO-RSA. To our knowledge, there have been no Level-I or II studies directly comparing these parameters and meta-analyses have been hampered because many large outcome studies using the Grammont design use the Constant score as an outcome measure and most large outcome studies evaluating a lateralized center of rotation design use the American Shoulder and Elbow Surgeons (ASES) score19.
Advocates for a medialized center of rotation glenosphere design point to the substantially decreased glenoid baseplate forces and concomitant decreased rates of glenosphere loosening5,6. They dismiss the notching as largely avoidable with baseplate placement low on the glenoid and note that the clinical importance of notching is unclear8-11. Advocates for a lateralized design have pointed to the potentially greater impingement-free range of motion and the lower baseplate loosening rates seen with current designs that include locking-screw fixation17,20. Advocates for the BIO-RSA have pointed to the high rate of graft incorporation, low rate of notching, and favorable bone-implant forces, but must concede that this approach is newer and has less supporting data14,18.
Baseplate Positioning: Craniocaudal and Tilted
Several surgeon teams have explored various baseplate positioning options in reverse total shoulder arthroplasty. Initially, many surgeons implanted the center of the baseplate in the center of the glenoid, but the humerosocket in adduction would impinge on the inferior aspect of the glenoid, causing levering and dislocation or severe notching. In 2008, Kelly et al. touted the success of inferior, or caudal, baseplate positioning21. In another cadaveric study, this caudal placement increased internal rotation, external rotation, abduction, and adduction while reducing notching22.
Because of difficulties with exposure or superior glenoid erosion, some early implants were placed with the baseplate tilted superiorly23. A validated computer model was used to show that this position leads to an uneven force distribution with increased shear and tensile forces24. Thus, perhaps it is not surprising that multiple clinical studies have shown superior baseplate tilt to correlate with baseplate failure (Figs. 4-A and 4-B)14,23,24.
Although it is clear that superior baseplate tilt is undesirable, there is an active debate around whether neutral or inferior baseplate tilt is preferred. One biomechanical virtual model revealed that inferior tilt offered a greater impingement-free arc; however, this result may be dependent on the type of sphere utilized24. On the basis of the biomechanical virtual model, inferior tilting of the baseplate is recommended for concentric and laterally offset glenospheres; however, this same recommendation may be detrimental to inferiorly eccentric glenospheres, for this combination yields an uneven distribution of forces24. Neutral baseplate tilt yielded an acceptable force distribution, regardless of the type of glenosphere utilized24.
In addition to the biomechanical model of Gutiérrez et al. showing that inferior tilt improves range of motion and reduces impingement that may lead to notching24, Nyffeler et al. used a cadaveric model and found significantly improved adduction and abduction angles with inferior placement and neutral tilt of the glenosphere when compared with a component placed with inferior tilt25. In a radiographic follow-up of a large clinical series, Simovitch et al. found that inferior tilt actually led to increased notching, potentially because reaming with inferior tilt removes more of the inferior aspect of the glenoid and may bring the center of rotation and humerosocket even closer to the lateral border of the scapula9. Kempton et al. performed a similar retrospective case-control study to assess the effect of neutral compared with 10° to 15° inferior baseplate tilt, and they found no correlation with grade or rate of notching (Fig. 5)26. Finally, in their prospective, randomized controlled trial, Edwards et al. compared neutral and 10° inferior tilt in implantation with intraoperative computer navigation and showed no effect on scapular notching severity ratings, rate, or clinical outcomes27. These conflicting assessments have yielded no clear consensus on the favorability of neutral tilt compared with that of slightly inferior baseplate tilt.
Caudal positioning of the baseplate, flush with the inferior aspect of the glenoid, can improve impingement-free range of motion, can reduce notching, and can reduce instability25. With regard to baseplate tilt, superior tilt introduces baseplate forces that lead to early implant failure and should be avoided. However, there is no clear consensus on whether neutral or inferior baseplate tilt is advantageous23.
135° and 155° Humerosocket Inclination
Debate has also centered around the angle of humerosocket inclination in reverse total shoulder arthroplasty. Grammont championed the 155° inclination over the anatomical 135° inclination, citing increased stability by placing the humerosocket further under the glenosphere to better resist the pull of the deltoid3. Further analysis revealed that this increased stability is dependent on arm positioning28. Additionally, the Grammont 155° inclination was shown, in a validated computer model, to contribute to scapular impingement, with this change in inclination having the largest effect on scapular impingement compared with other design variations of the reverse prosthesis28. The 155° design places the medial portion of the humerosocket closer to the lateral pillar of the scapula. It is important to note that, in the model, the increase in impingement from the 155° design could be ameliorated by inferior baseplate positioning, increasing the size of the glenosphere, or lateralizing the glenosphere12.
In an attempt to clarify the clinical effect of humerosocket inclination on reverse total shoulder arthroplasty, Kempton et al. performed a retrospective study evaluating the difference between a 143° inclination and the Grammont 155° inclination in a single-surgeon case series. The results revealed a 16.2% notching rate at the one-year follow-up for the 143° cohort and a 60.7% notching rate for the 155° cohort29. To achieve a broader picture of the effect of increased varus inclination compared with the Grammont design, Erickson et al. performed a meta-analysis of thirty-eight studies and 2222 shoulders, comparing 135° inclination with 155° inclination (Fig. 6)7. Their results corroborated those by Kempton et al.29, with the 135°-inclination shoulders having a mean notching rate of 2.8% compared with 16.8% exhibited by the 155° group7. Interestingly, there was no significant difference in instability between the two groups. Unfortunately, the study by Erickson et al. was significantly hampered by the fact that all shoulders in the 135° group had a lateralized glenosphere design, but the 155° group had a mixture of medialized and lateralized center of rotation—a confounding bias that has already been shown to significantly affect notching, irrespective of humerosocket inclination7,14.
It appears as if the 155° humerosocket inclination proposed by Grammont may contribute to impingement and notching; however, confounding study variables have prevented a reliable direct comparison of the 155° and 135° inclinations. Without a direct comparison, a conclusive recommendation regarding humerosocket inclination in reverse total shoulder arthroplasty remains premature at this time.
Two decades after Grammont first published the results of sixteen cases using his reverse total shoulder arthroplasty design1, tens of thousands of reverse shoulder replacements are implanted annually30. However, unsolved questions and controversy with regard to implant placement and design remain. Although the literature is clear that superior baseplate positioning and tilt are undesirable and should be avoided14,23,24, no concrete conclusion with regard to medialization or lateralization of the glenosphere, neutral or slightly inferior baseplate tilt, and humerosocket inclination can be made. Further research is necessary to determine optimal implant positioning and design in reverse total shoulder arthroplasty.
Investigation performed at the Division of Sports Medicine and Shoulder Surgery, Department of Orthopaedic Surgery, Duke University Medical Center, Durham, North Carolina
Disclosure: There was no external funding source for this study. On the Disclosure of Potential Conflicts of Interest forms, which are provided with the online version of the article, one or more of the authors checked “yes” to indicate that the author had a relevant financial relationship in the biomedical arena outside the submitted work.
- Copyright © 2016 by The Journal of Bone and Joint Surgery, Incorporated