- Email: [email protected]
Radiographic changes differ between two different short press-fit humeral stem designs in total shoulder arthroplasty
ARTICLE IN PRESS J Shoulder Elbow Surg (2017) ■■, ■■–■■
www.elsevier.com/locate/ymse
ORIGINAL ARTICLE
Radiographic changes differ between two different short press-fit humeral stem designs in total shoulder arthroplasty Patrick J. Denard, MDa,b,*, Matthew P. Noyes, MDa, J. Brock Walker, MDc, Yousef Shishani, MDd, Reuben Gobezie, MDd, Anthony A. Romeo, MDe, Evan Lederman, MDc,f a
Southern Oregon Orthopedics, Medford, OR, USA Department of Orthopaedics and Rehabilitation, Oregon Health & Science University, Portland, OR, USA c University of Arizona College of Medicine–Phoenix, Phoenix, AZ, USA d The Cleveland Shoulder Institute, Beachwood, OH, USA e Rush University Medical Center, Chicago, IL, USA f The Orthopedic Clinic Association, Phoenix, AZ, USA b
Background: The purpose of this study was to compare the radiographic changes of the humerus in the short term after total shoulder arthroplasty with two different short-stem humeral components. The hypothesis was that there would be no difference in radiographic changes or functional outcome based on component type. Methods: A retrospective review was conducted of primary total shoulder arthroplasties performed with a short press-fit humeral component. Group A included a collarless humeral stem with an oval geometry and curved stem (Ascend or Ascend Flex; Wright Medical, Memphis, TN, USA). Group B included a humeral stem with a metaphyseal collar, rectangular geometry, and straight stem (Apex; Arthrex, Inc., Naples, FL, USA). Radiographic changes and functional outcome were evaluated at a minimum of 2 years postoperatively. Results: There were 42 patients in group A and 35 patients in group B available for analysis. There was no difference in functional outcome between the groups. In group A, the mean total radiographic change score of the humerus was 3.9, with changes classified as low in 38% and high in 62%. In group B, the mean total radiographic change score of the humerus was 2.5, with changes classified as low in 77% and high in 23% (P < .001). Medial calcar osteolysis was present in 71% of group A compared with 28.5% of group B (P < .001). Conclusion: At short-term follow-up, there is no difference in functional outcome or revision between 2 different humeral stem designs. However, bone adaptive changes and the rate of medial calcar osteolysis are significantly different. Level of evidence: Level III; Retrospective Cohort Design; Treatment Study © 2017 Journal of Shoulder and Elbow Surgery Board of Trustees. All rights reserved. Keywords: Total shoulder arthroplasty; short stem; humeral loosening; stress shielding; press fit; bone adaptations; metaphyseal fixation
Southern Oregon Institutional Review Board approved this study. *Reprint requests: Patrick J. Denard, MD, Southern Oregon Orthopedics, 2780 E Barnett Rd, Suite 200, Medford, OR 97530, USA. E-mail address: [email protected] (P.J. Denard).
Anatomic total shoulder arthroplasty (TSA) has proved to be a viable treatment for glenohumeral arthritis and leads to improvement in function and relief of pain in the majority of cases. However, revision is necessary in approximately 10%
1058-2746/$ - see front matter © 2017 Journal of Shoulder and Elbow Surgery Board of Trustees. All rights reserved. https://doi.org/10.1016/j.jse.2017.08.010
ARTICLE IN PRESS 2 of cases within 10 years and 20%-30% of cases within 20 years after the index procedure.4,8 The potential need for revision as well as the complication rate after revision (ie, greater tuberosity fracture from removal of a finned stemmed component) has led to development of shorter humeral stems that are potentially bone preserving and more easily revised in the event of the need for revision. Recent studies have reported on the clinical results and radiographic findings after TSA with a short-stem component. In 2 of these reports, the rate of radiographic changes was very high with a collarless stem with a curved shape.1,7 However, a variety of stem designs exist with variation in features, such as the presence of a metaphyseal collar or stem shape (curved circular vs. straight rectangular). To date, no studies have compared the radiographic findings with different types of short-stem components. The purpose of this study was to compare the radiographic changes of the humerus in the short term after placement of 2 different short-stem humeral components. The hypothesis was that there would be no difference in radiographic changes of the humerus based on component type.
Methods A retrospective review of prospectively collected data was conducted of anatomic TSAs performed at 2 institutions between July 2012 and June 2014. Inclusion criteria included a primary TSA with a short humeral stem placed with a press-fit technique and a minimum follow-up of 2 years. Exclusion criteria included revision arthroplasty and concomitant glenoid bone grafting.
P.J. Denard et al.
Surgical technique TSAs were performed by 2 different surgeons using a consistent technique with the type of humeral stem as the only intrasurgeon variation during the study period. A deltopectoral approach was used to gain access to the glenohumeral joint. A freehand anatomic cut of the humeral head (respecting native inclination and retroversion) was performed after removal of osteophytes. The humeral canal was broached to accept a press-fit humeral component using a compaction technique to minimize bone removal. A cemented all-polyethylene glenoid was used in all cases, and attention was turned to placement of the final humeral component. Group A was composed of a collarless stem with a circular geometry and curved shape (Ascend or Ascend Flex; Wright Medical, Memphis, TN, USA). This stem design was the short stem of choice for the surgeons in the first 18 months of the study period. The stem ranges in length from 66 to 98 mm and is available in 3 different humeral inclinations (127.5°, 132.5°, and 137.5°) (Fig. 1, A). Group B was another press-fit short humeral stem (Apex; Arthrex, Inc., Naples, FL, USA) that was the primary stem of choice for all TSAs performed by the surgeons during the last 6 months of the study period. This stem ranges in length from 60 to 65 mm based on the appropriate size. This stem has a metaphyseal collar, rectangular geometry, and adaptability in humeral inclination between 125° and 140° (Fig. 1, B).
Clinical evaluation Function and range of motion were assessed preoperatively and postoperatively at the final follow-up. Function was determined with the Simple Shoulder Test, Single Assessment Numeric Evaluation score,
Figure 1 (A) Photograph of Ascend Flex stem (Wright Medical). (B) Photograph of Apex stem (Arthrex, Inc.). (A modified with permission from Morwood MP, Johnston PS, Garrigues GE. Proximal ingrowth coating decreases risk of loosening following uncemented shoulder arthroplasty using mini-stem humeral components and lesser tuberosity osteotomy. J Shoulder Elbow Surg 2017;26:1246-52.)
ARTICLE IN PRESS Short stem radiographic changes
3
American Shoulder and Elbow Surgeons score, and visual analog scale pain score. Range of motion was assessed using a goniometer to determine forward flexion and external rotation with the arm at the side. Internal rotation was estimated to the nearest spinal level. Patient satisfaction (yes or no) and return to activity (yes or no) were recorded at the final follow-up. Complications and reoperations were determined by chart review.
Radiographic evaluation Radiographs were independently reviewed by 2 orthopedic shoulder surgeons (P.J.D. and M.P.N.). Disagreements were rectified by repeated evaluation with mutual agreement. Grashey (true glenohumeral anteroposterior view) and axillary radiographs were obtained immediately after surgery and at a final follow-up of a minimum of 2 years postoperatively and reviewed in Digital Imaging and Communications in Medicine (DICOM) format using OsiriX Lite (Pixmeo; Bernex, Switzerland). Immediate postoperative films were reviewed to determine alignment and filling ratio of the humeral component in the coronal plane. Anatomic alignment was defined as <5° of deviation of stem alignment from the anatomic axis of the humeral shaft. Nonanatomic varus or valgus placement was defined as alignment of the humeral stem ≥5° of deviation from the anatomic axis. Filling ratio of the metaphysis (FRmet) was determined at the level of the inferior head and of the outer diaphysis (FRdia) at the distal third of the prosthesis as described by Schnetzke et al.7 Radiographic changes were recorded by comparing the immediate postoperative and final postoperative radiographs. Radiographic changes in 5 humeral zones were recorded as described by Nagels et al and adapted to shorter stem by Schnetzke et al3,7 (Fig. 2). Each zone was evaluated for the presence or absence of bone remodeling features, including condensation lines, cortical thinning and osteopenia, and spot welds. The radiographic changes from each zone were summed for an overall grade of none (0-1 change), mild (2-3 changes), moderate (4-6 changes), or advanced (>6 changes). These changes were also categorized as low (no or mild changes) or high (moderate to advanced). Medial calcar osteolysis was noted as none, partial, or complete.1 In addition, humeral component subsidence and varus or valgus shift were recorded.
Statistical analysis Continuous data were described by mean and standard deviations. Paired t-test, sign test, and Wilcoxon rank sum were performed (depending on variable distribution) to analyze the difference in preoperative and postoperative range of motion and functional outcome scores. χ2 tests were conducted for the radiographic variables and for return to activity and satisfaction. A Spearman test was used to correlate radiographic findings with range of motion and functional outcome. Statistical analyses were conducted using SAS 9.4 (SAS Institute, Cary, NC, USA) and performed by a trained statistician. Two-tailed P values of < .05 were considered significant.
Figure 2 Zones for analysis of radiographic adaptations. L1, lateral metaphysis; L2, lateral diaphysis; US, under stem; M1, medial metaphysis; M2, medial diaphysis; G, glenoid. (Reproduced with permission from Schnetzke M, Coda S, Raiss P, Walch G, Loew M. Radiologic bone adaptations on a cementless short-stem shoulder prosthesis. J Shoulder Elbow Surg 2016;25:650-7.)
the 2 groups with the exception that follow-up was longer in group A and group B had higher preoperative forward flexion (Table I).
Clinical results There was no difference in range of motion or functional outcome scores between the 2 groups (Table II). In group A, there were 3 complications in 42 patients (7.1%), including 1 subscapularis repair failure that was revised, 1 revision to a reverse TSA for stem loosening at 3 years postoperatively, and 1 postoperative rotator cuff failure that was managed conservatively. In group B, there were no complications or revision operations.
Results Radiographic results There were 93 cases that met the study criteria. Complete 2-year follow-up was available in 77 cases (82.8%), which included 42 patients in group A and 35 patients in group B. Baseline demographics and follow-up were similar between
On implantation, the stems in group A were placed in anatomic alignment in 90% of cases, whereas the stems in group B were placed anatomically in 86% of cases (P = .724). There
ARTICLE IN PRESS 4
P.J. Denard et al. Table I
Baseline characteristics
Age (y) Follow-up (mo) Sex Male Female Dominant arm Forward flexion External rotation Internal rotation VAS pain ASES SST SANE
Table III
Group A Ascend (n = 42)
Group B Apex (n = 35)
P value
67.7 (36-84) 35.0 (24-46)
66.5 (52-79) 25.0 (24-29)
.565 <.001
28 (67) 14 (33) 22 (55) 103° (± 28°) 21° (± 15°) S1 6.9 (± 1.8) 34.5 (± 15.2) 3.8 (± 2.5) 29.8 (± 15.1)
22 (63) 13 (37) 23 (66) 119° (± 27°) 26° (± 11°) L5 6.1 (± 2.2) 36.6 (± 16.1) 4.2 (± 2.6) 32.0 (± 24.2)
.812 .345 .009 .052 .067 .199 .708 .492 .838
ASES, American Shoulder and Elbow Surgeons; SANE, Single Assessment Numeric Evaluation; SST, Simple Shoulder Test; VAS, visual analog scale. Categorical variables are presented as number (%). Continuous variables are presented as mean (± standard deviation) or mean (range). Bold values are statistically significant.
Table II
Postoperative outcome
Forward flexion External rotation Internal rotation VAS pain ASES SST SANE
Group A Ascend (n = 42)
Group B Apex (n = 35)
P value*
138° (± 27°) 49° (± 22°) L2 1.2 (± 2.2) 82.7 (± 20.3) 9.6 (± 2.8) 84.5 (± 18.5)
147° (± 19°) 55° (± 12°) L2 1.0 (± 1.3) 86.5 (± 11.7) 10.0 (± 2.4) 90.7 (± 10.3)
.135 .715 .479 .490 .611 .740 .282
ASES, American Shoulder and Elbow Surgeons; SANE, Single Assessment Numeric Evaluation; SST, Simple Shoulder Test; VAS, visual analog scale. * P values are based on comparing gains between preoperative and postoperative range of motion or functional scores.
was no difference in any functional outcome scores or range of motion between stems that anatomically and nonanatomically aligned (P > .05). Filling ratios were lower in group A compared with group B for both FRmet (55.0 vs. 61.6; P < .001) and FRdia (56.1 vs. 51.5; P = .006). In group A, the mean total radiographic change score of the humerus was 3.9, with 2 (4.8%) having no change, 14 (33.3%) having mild changes, 25 (59.5%) having moderate changes, and 1 (2.4%) having advanced changes. Therefore, changes were classified as low in 38% and high in 62%. In group B, the mean total radiographic change score of the humerus was 2.5, with 7 (20.0%) having no change, 20 (57.1%) having mild changes, 8 (22.9%) having moderate changes, and none having advanced changes. Therefore, in this group, changes were low in 77% and high in 23% (P < .001). There were multiple differences in the incidence
Radiographic adaptations by humeral stem
Zone/adaptation
Ascend (n = 42)
Apex (n = 35)
P value
Total score L1 No change CL CNO SW L2 No change CL CNO SW US No change CL CNO SW M1 No change CL CNO SW M2 No change CL CNO SW
3.9
2.5
9 (21.4) 2 (4.8) 30 (71.4) 0 (0)
15 (42.9) 10 (28.6) 7 (20.0) 0 (0)
.043 .009 <.001 NA
6 (14.3) 1 (2.4) 27 (64.3) 13 (30.9)
22 (62.9) 2 (5.7) 4 (11.4) 5 (14.3)
<.001 .588 <.001 .109
16 (38.1) 5 (11.9) 0 (0) 16 (38.1)
21 (60.0) 10 (28.6) 0 (0) 6 (17.6)
.055 .086 NA .051
2 (4.8) 0 (0) 39 (92.9) 0 (0)
10 (28.6) 6 (17.1) 22 (62.9) 0 (0)
.009 .007 .002 NA
12 (28.6) 2 (4.8) 21 (50.0) 8 (19.0)
20 (57.1) 4 (11.4) 5 (14.3) 6 (17.1)
.011 .402 .001 .829
.002
L1, lateral metaphysis; L2, lateral diaphysis; US, under stem; M1, medial metaphysis; M2, medial diaphysis; CL, condensation line; CNO, cortical thinning or osteopenia; SW, spot weld; NA, not applicable. Data are expressed as number (%). Bold values are statistically significant.
of postoperative radiographic changes, including condensation lines, cortical thinning, and spot welds, that are summarized in Table III. Subjective shoulder value (SSV) improvement was lower in patients with cortical thinning in the lateral diaphysis (L2) compared with those without cortical thinning (SSV improvement 46 vs. 60; P = .038). Otherwise, condensation lines, cortical thinning, and spot welds were not correlated with functional outcome. Medial calcar osteolysis was partial in 64% and complete in 7% of group A (Fig. 3). In group B, partial calcar osteolysis was seen in 28.5% of cases, and there were no cases of complete osteolysis (P < .001) (Fig. 4). Postoperative external rotation was lower in the patients with complete calcar osteolysis (53° none, 53° partial, 20° complete; P = .033). Six stems (14.2%) in group A demonstrated postoperative shift or subsidence compared with 3 stems (8.6%) in group B (P = .499). Given that group A contained prostheses with 2 different proximal stem coatings, separate analyses were performed. First, an analysis was performed to compare radiographic changes between the 2 different variations within group A
ARTICLE IN PRESS Short stem radiographic changes
5 was no statistically significant difference in loosening or shift/ subsidence between the 2 (13% and 16%, respectively; P > .05) (Fig. 5). Finally, the subset of stems in group A with proximal coating (Ascend Flex) were compared with group B separately. As in the overall comparison, the total adaptive change score (P = .019) and rate of medial calcar osteolysis (P = .011) were higher in this subset of group A compared with group B. Cortical thinning was more common in this subset of group A compared with group B for the lateral metaphysis (79% vs. 20%; P < .001), the lateral diaphysis (53% vs. 11%; P = .002), the medial metaphysis (95% vs. 63%; P = .011), and the medial diaphysis (50% vs. 14%; P = .009). There was no difference between these 2 subsets in the rate of loosening or shift/subsidence.
Discussion
Figure 3 The 2-year postoperative radiograph of an indivudal in groupAwith anAscend Flex stem shows partial medial calcar osteolysis.
Figure 4 The 2-year postoperative radiograph of an indivudal in group B with an Apex stem shows partial medial calcar osteolysis.
(Ascend and Ascend Flex). There was no difference between the 2 proximal coatings with regard to medial calcar osteolysis, total adaptive change score, and condensation lines, cortical thinning, or spot welds in any zone (P > .05). There
In this study, we sought to compare the short-term functional outcome and radiographic adaptations of 2 different short press-fit humeral stems commonly used in TSA. Whereas functional outcome was comparable between the 2 stems, our findings refute the hypothesis that radiographic adaptations would be the same. These findings may have implications for the design of short press-fit humeral stems. Stress shielding is a common phenomenon after TSA with a traditional-length stem. Raiss et al analyzed 395 humeral stems, of which 67 were press-fit TSAs, at a mean followup of 8.2 years.5 In the press-fit TSA group, 42% had osteolysis of the greater tuberosity, and 39% had osteolysis of the medial calcar (21% complete and 18% partial). The concern for stress shielding as well as the potential ease of revision has led to the development of shorter humeral stems. In a finite element analysis, Razfar et al examined humeral bone stresses with a stemless component, short stem, or traditional-length stem.6 The traditional-length stem and short stem had a rectangular geometry similar to the stem in group B in our study. Native proximal humeral cortical bone stresses were more closely mimicked with a short stem (78% of native) compared with the traditional-length stem (58% of native). They concluded that a short stem could be used to reduce stress shielding, assuming adequate stability. However, few clinical studies have evaluated short stems to date, and no studies to our knowledge have compared different short-stem designs. Schnetzke et al evaluated bone adaptations in 52 short stems (Ascend) at a mean of 32 months postoperatively.7 They used the same radiographic analysis as in this study and noted high adaptations in 52% of cases, which was similar to the 62% rate seen in this study in group A. They did not report the rate of medial calcar osteolysis but noted medial cortical thinning or osteopenia in 83%, which was similar to the 93% observed in this study. Casagrande et al evaluated 73 shortstem TSAs (Ascend) for radiolucent lines, loosening, and calcar osteolysis at a minimum of 2 years postoperatively.1 Nine (12%) were revised, 6 (8.2%) of which were for humeral loosening. They did not specifically report the rate of subsidence
ARTICLE IN PRESS 6
P.J. Denard et al.
Figure 5 (A) Immediate postoperative radiograph of an individual in group A with an Ascend Flex stem. (B) The 2-year postoperative radiograph demonstrates cortical thinning/osteopenia in the lateral metaphysis and diaphysis, a spot weld in the lateral diaphysis, and slight varus shift of the humeral stem. The postoperative subjective shoulder value was 90.
or shift, but these findings were used to categorize “at risk for loosening,” which they reported in an additional 6 stems (8.7%). In this study, our rate of revision for loosening was lower at 2.4% (1 of 42), but shift or subsidence was seen in 14%. Their rate of overall loosening therefore appears similar to that in our study. With regard to medial calcar osteolysis, Casagrande et al noted partial osteolysis in 14.5% and complete osteolysis in 3%, whereas our rates were much higher at 64% and 7%, respectively. Given that we had a 9-fold difference between partial and complete and they had a 5-fold difference, we may have had a lower threshold for classifying osteolysis. In an effort to reduce the risk of humeral loosening, the first-generation Ascend stem was modified in the secondgeneration Ascend Flex to include a proximal titanium plasma spray. Morwood et al compared 2-year outcomes in 34 firstgeneration Ascend stems without proximal porous coating with 34 Ascend Flex stems with proximal porous coating.2 They had 55% follow-up at a mean of 30 months postoperatively. Compared with the proximally coated design, the firstgeneration stem had a significantly higher risk of loosening (3% vs. 21%; P = .03) and radiolucencies (21% vs. 44%; P = .04). They concluded that proximal ingrowth coating on short stems may decrease the risk of loosening. In this study, we did not observe any difference in the rate of adaptive changes or subsidence/shift between the Ascend and Ascend Flex stems. This difference could be due to the different radiographic analyses between our studies, the small numbers in our study, or the follow-up differences (83% at 35 months in this study vs. 55% at 30 months in their study). Whereas proximal coating may influence fixation of a short stem, this may be design specific, and multiple design factors likely influence radiographic adaptations and loosening. In this study, the mean radiographic adaptation change score was higher in group A compared with group B (3.9 vs. 2.5). High
adaptations were seen in 61.9% of group A compared with 22.9% in group B. In addition, the rate of medial calcar osteolysis was 71% in group A vs. 28.5% in group B. Although overall functional outcome was not different between the 2 groups, outcome was influenced by radiographic changes. SSV and postoperative external rotation were lower in the setting of lateral cortical thinning and medial calcar osteolysis, respectively, and both radiographic findings were more common in group A. Longer follow-up is needed to evaluate the impact of these differences on implant survival and function. Whereas subsidence or shift was higher in group A (14.2% vs. 8.6%), the difference did not reach statistical difference with the numbers studied. At these values, it would require 505 patients in each group to have 80% power to detect a difference between the 2 groups; if the values were 20% and 10%, it would require 199 patients in each group. The humeral stems in this study have substantial differences in shape and design. The group A stem (Ascend) is oval, curved, and collarless (Ascend and Ascend Flex), whereas the group B stem (Apex) is rectangular and straight and has a collar. These design variables may lead to different proximal humeral bone stresses that influence radiographic adaptations. The curved oval stem in group A tapers in width substantially distally. Despite having a lower FRmet on 2-dimensional images with an anteroposterior radiograph, the oval shape may actually result in greater 3-dimensional (3D) proximal fill, which creates stress shielding at the periphery and could account for the greater calcar osteolysis in this group. Alternatively, it is possible that the curved stem alters proximal stresses unevenly to a greater extent than a straight stem and thus leads to greater adaptive changes. The collar feature is also a substantial difference in the designs. A collarless design may benefit from proximal porous coating to achieve fixation, whereas a collared design may have less need for proximal coating as suggested by the low rate of loosening in group B
ARTICLE IN PRESS Short stem radiographic changes in our study. It is notable that in the subanalysis of the stems in group A with proximal coating to group B, which does not have proximal coating, the overall differences were unchanged. Radiographic adaptations remained higher in this subset of group A compared with group B. This provides further evidence that the requirement for proximal coating is design specific. In other words, a collarless stem may benefit from proximal coating, but based on the results of this study, there is not a need for proximal coating with a collared stem. Filling ratio has also been reported to affect radiographic changes. Schnetzke et al observed that higher filling ratios of the implant relative to the proximal humerus were associated with higher radiographic adaptations.7 This finding was also supported by Nagels et al in an evaluation of traditionallength stems.3 Interestingly, in this study, filling ratios were higher in group B, which had fewer radiographic adaptations. This suggests that implant design is more important than filling ratio. However, the filling ratio is difficult to compare between the 2 stems because the designs are different in shape. The evaluation we performed on plain anteroposterior radiographs is insufficient to compare the difference in volume of bone that is occupied between the 2 stems; such evaluation would require 3D analysis. Nonetheless, it seems desirable on the basis of the previous studies both to implant and to design press-fit stems with the minimum fill required to achieve fixation. Preoperative templating or 3D planning of the humerus preoperatively may be useful for limiting the filling ratio. The major strength of this study is that it provides a comparison between 2 different humeral stems with a consistent surgical technique. However, there are several limitations. The follow-up is short term, and results may change at midterm to long term. Follow-up was longer in group A, which may have affected the results. However, previous study has suggested that stress shielding was detected in all cases by 24 months postoperatively without a difference in 24-month compared with 5-year follow-up; this suggests that our minimum follow-up was sufficient to detect radiographic changes of the humerus.9 The study design was retrospective and therefore introduces the potential for selection bias. We believe this plays a minor role, given that both surgeons adopted the group B stem at the same time. The radiographic review was performed by 2 surgeons, 1 of whom performed several of the operations. Although we attempted to eliminate bias by having a separate review, it is possible that this may have introduced bias. Finally, given that there are multiple differences in the designs of the 2 stems, it is unclear which factors (ie, collar vs. collarless or oval vs. rectangular) have the most impact on stress shielding. A finite element analysis examining these factors may be useful.
Conclusions At short-term follow-up, there is no difference in functional outcome or revision between 2 different humeral
7 stem designs. However, adaptive changes and the rate of medial calcar osteolysis varied substantially between the 2 short press-fit humeral components. Further study is needed to evaluate the long-term impact of these differences and to elucidate the design factors of short stems that are associated with stress shielding.
Disclaimer This study was supported by a grant from Arthrex, Inc. Patrick J. Denard, Reuben Gobezie, Anthony A. Romeo, and Evan Lederman are consultants for and receive royalties from Arthrex, Inc. The other authors, their immediate families, and any research foundations with which they are affiliated have not received any financial payments or other benefits from any commercial entity related to the subject of this article.
References 1. Casagrande DJ, Parks DL, Torngren T, Schrumpf MA, Harmsen SM, Norris TR, et al. Radiographic evaluation of short-stem press-fit total shoulder arthroplasty: short-term follow-up. J Shoulder Elbow Surg 2016;25:1163-9. http://dx.doi.org/10.1016/j.jse.2015.11.067 2. Morwood MP, Johnston PS, Garrigues GE. Proximal ingrowth coating decreases risk of loosening following uncemented shoulder arthroplasty using mini-stem humeral components and lesser tuberosity osteotomy. J Shoulder Elbow Surg 2017;26:1246-52. http://dx.doi.org/10.1016/ j.jse.2016.11.041 3. Nagels J, Stokdijk M, Rozing PM. Stress shielding and bone resorption in shoulder arthroplasty. J Shoulder Elbow Surg 2003;12:35-9. http://dx.doi.org/10.1067/mse.2003.22 4. Raiss P, Bruckner T, Rickert M, Walch G. Longitudinal observational study of total shoulder replacements with cement: fifteen to twenty-year follow-up. J Bone Joint Surg Am 2014;96:198-205. http://dx.doi.org/10 .2106/JBJS.M.00079 5. Raiss P, Edwards TB, Deutsch A, Shah A, Bruckner T, Loew M, et al. Radiographic changes around humeral components in shoulder arthroplasty. J Bone Joint Surg Am 2014;96:e54. http://dx.doi.org/ 10.2106/JBJS.M.00378 6. Razfar N, Reeves JM, Langohr DG, Willing R, Athwal GS, Johnson JA. Comparison of proximal humeral bone stresses between stemless, short stem, and standard stem length: a finite element analysis. J Shoulder Elbow Surg 2016;25:1076-83. http://dx.doi.org/10.1016/ j.jse.2015.11.011 7. Schnetzke M, Coda S, Raiss P, Walch G, Loew M. Radiologic bone adaptations on a cementless short-stem shoulder prosthesis. J Shoulder Elbow Surg 2016;25:650-7. http://dx.doi.org/10.1016/j.jse.2015.08 .044 8. Singh JA, Sperling JW, Cofield RH. Revision surgery following total shoulder arthroplasty: analysis of 2588 shoulders over three decades (1976 to 2008). J Bone Joint Surg Br 2011;93:1513-7. http://dx.doi.org/ 10.1302/0301-620X.93B11.26938 9. Spormann C, Durchholz H, Audigé L, Flury M, Schwyzer HK, Simmen BR, et al. Patterns of proximal humeral bone resorption after total shoulder arthroplasty with an uncemented rectangular stem. J Shoulder Elbow Surg 2014;23:1028-35. http://dx.doi.org/10.1016/j.jse.2014.02.024
www.elsevier.com/locate/ymse
ORIGINAL ARTICLE
Radiographic changes differ between two different short press-fit humeral stem designs in total shoulder arthroplasty Patrick J. Denard, MDa,b,*, Matthew P. Noyes, MDa, J. Brock Walker, MDc, Yousef Shishani, MDd, Reuben Gobezie, MDd, Anthony A. Romeo, MDe, Evan Lederman, MDc,f a
Southern Oregon Orthopedics, Medford, OR, USA Department of Orthopaedics and Rehabilitation, Oregon Health & Science University, Portland, OR, USA c University of Arizona College of Medicine–Phoenix, Phoenix, AZ, USA d The Cleveland Shoulder Institute, Beachwood, OH, USA e Rush University Medical Center, Chicago, IL, USA f The Orthopedic Clinic Association, Phoenix, AZ, USA b
Background: The purpose of this study was to compare the radiographic changes of the humerus in the short term after total shoulder arthroplasty with two different short-stem humeral components. The hypothesis was that there would be no difference in radiographic changes or functional outcome based on component type. Methods: A retrospective review was conducted of primary total shoulder arthroplasties performed with a short press-fit humeral component. Group A included a collarless humeral stem with an oval geometry and curved stem (Ascend or Ascend Flex; Wright Medical, Memphis, TN, USA). Group B included a humeral stem with a metaphyseal collar, rectangular geometry, and straight stem (Apex; Arthrex, Inc., Naples, FL, USA). Radiographic changes and functional outcome were evaluated at a minimum of 2 years postoperatively. Results: There were 42 patients in group A and 35 patients in group B available for analysis. There was no difference in functional outcome between the groups. In group A, the mean total radiographic change score of the humerus was 3.9, with changes classified as low in 38% and high in 62%. In group B, the mean total radiographic change score of the humerus was 2.5, with changes classified as low in 77% and high in 23% (P < .001). Medial calcar osteolysis was present in 71% of group A compared with 28.5% of group B (P < .001). Conclusion: At short-term follow-up, there is no difference in functional outcome or revision between 2 different humeral stem designs. However, bone adaptive changes and the rate of medial calcar osteolysis are significantly different. Level of evidence: Level III; Retrospective Cohort Design; Treatment Study © 2017 Journal of Shoulder and Elbow Surgery Board of Trustees. All rights reserved. Keywords: Total shoulder arthroplasty; short stem; humeral loosening; stress shielding; press fit; bone adaptations; metaphyseal fixation
Southern Oregon Institutional Review Board approved this study. *Reprint requests: Patrick J. Denard, MD, Southern Oregon Orthopedics, 2780 E Barnett Rd, Suite 200, Medford, OR 97530, USA. E-mail address: [email protected] (P.J. Denard).
Anatomic total shoulder arthroplasty (TSA) has proved to be a viable treatment for glenohumeral arthritis and leads to improvement in function and relief of pain in the majority of cases. However, revision is necessary in approximately 10%
1058-2746/$ - see front matter © 2017 Journal of Shoulder and Elbow Surgery Board of Trustees. All rights reserved. https://doi.org/10.1016/j.jse.2017.08.010
ARTICLE IN PRESS 2 of cases within 10 years and 20%-30% of cases within 20 years after the index procedure.4,8 The potential need for revision as well as the complication rate after revision (ie, greater tuberosity fracture from removal of a finned stemmed component) has led to development of shorter humeral stems that are potentially bone preserving and more easily revised in the event of the need for revision. Recent studies have reported on the clinical results and radiographic findings after TSA with a short-stem component. In 2 of these reports, the rate of radiographic changes was very high with a collarless stem with a curved shape.1,7 However, a variety of stem designs exist with variation in features, such as the presence of a metaphyseal collar or stem shape (curved circular vs. straight rectangular). To date, no studies have compared the radiographic findings with different types of short-stem components. The purpose of this study was to compare the radiographic changes of the humerus in the short term after placement of 2 different short-stem humeral components. The hypothesis was that there would be no difference in radiographic changes of the humerus based on component type.
Methods A retrospective review of prospectively collected data was conducted of anatomic TSAs performed at 2 institutions between July 2012 and June 2014. Inclusion criteria included a primary TSA with a short humeral stem placed with a press-fit technique and a minimum follow-up of 2 years. Exclusion criteria included revision arthroplasty and concomitant glenoid bone grafting.
P.J. Denard et al.
Surgical technique TSAs were performed by 2 different surgeons using a consistent technique with the type of humeral stem as the only intrasurgeon variation during the study period. A deltopectoral approach was used to gain access to the glenohumeral joint. A freehand anatomic cut of the humeral head (respecting native inclination and retroversion) was performed after removal of osteophytes. The humeral canal was broached to accept a press-fit humeral component using a compaction technique to minimize bone removal. A cemented all-polyethylene glenoid was used in all cases, and attention was turned to placement of the final humeral component. Group A was composed of a collarless stem with a circular geometry and curved shape (Ascend or Ascend Flex; Wright Medical, Memphis, TN, USA). This stem design was the short stem of choice for the surgeons in the first 18 months of the study period. The stem ranges in length from 66 to 98 mm and is available in 3 different humeral inclinations (127.5°, 132.5°, and 137.5°) (Fig. 1, A). Group B was another press-fit short humeral stem (Apex; Arthrex, Inc., Naples, FL, USA) that was the primary stem of choice for all TSAs performed by the surgeons during the last 6 months of the study period. This stem ranges in length from 60 to 65 mm based on the appropriate size. This stem has a metaphyseal collar, rectangular geometry, and adaptability in humeral inclination between 125° and 140° (Fig. 1, B).
Clinical evaluation Function and range of motion were assessed preoperatively and postoperatively at the final follow-up. Function was determined with the Simple Shoulder Test, Single Assessment Numeric Evaluation score,
Figure 1 (A) Photograph of Ascend Flex stem (Wright Medical). (B) Photograph of Apex stem (Arthrex, Inc.). (A modified with permission from Morwood MP, Johnston PS, Garrigues GE. Proximal ingrowth coating decreases risk of loosening following uncemented shoulder arthroplasty using mini-stem humeral components and lesser tuberosity osteotomy. J Shoulder Elbow Surg 2017;26:1246-52.)
ARTICLE IN PRESS Short stem radiographic changes
3
American Shoulder and Elbow Surgeons score, and visual analog scale pain score. Range of motion was assessed using a goniometer to determine forward flexion and external rotation with the arm at the side. Internal rotation was estimated to the nearest spinal level. Patient satisfaction (yes or no) and return to activity (yes or no) were recorded at the final follow-up. Complications and reoperations were determined by chart review.
Radiographic evaluation Radiographs were independently reviewed by 2 orthopedic shoulder surgeons (P.J.D. and M.P.N.). Disagreements were rectified by repeated evaluation with mutual agreement. Grashey (true glenohumeral anteroposterior view) and axillary radiographs were obtained immediately after surgery and at a final follow-up of a minimum of 2 years postoperatively and reviewed in Digital Imaging and Communications in Medicine (DICOM) format using OsiriX Lite (Pixmeo; Bernex, Switzerland). Immediate postoperative films were reviewed to determine alignment and filling ratio of the humeral component in the coronal plane. Anatomic alignment was defined as <5° of deviation of stem alignment from the anatomic axis of the humeral shaft. Nonanatomic varus or valgus placement was defined as alignment of the humeral stem ≥5° of deviation from the anatomic axis. Filling ratio of the metaphysis (FRmet) was determined at the level of the inferior head and of the outer diaphysis (FRdia) at the distal third of the prosthesis as described by Schnetzke et al.7 Radiographic changes were recorded by comparing the immediate postoperative and final postoperative radiographs. Radiographic changes in 5 humeral zones were recorded as described by Nagels et al and adapted to shorter stem by Schnetzke et al3,7 (Fig. 2). Each zone was evaluated for the presence or absence of bone remodeling features, including condensation lines, cortical thinning and osteopenia, and spot welds. The radiographic changes from each zone were summed for an overall grade of none (0-1 change), mild (2-3 changes), moderate (4-6 changes), or advanced (>6 changes). These changes were also categorized as low (no or mild changes) or high (moderate to advanced). Medial calcar osteolysis was noted as none, partial, or complete.1 In addition, humeral component subsidence and varus or valgus shift were recorded.
Statistical analysis Continuous data were described by mean and standard deviations. Paired t-test, sign test, and Wilcoxon rank sum were performed (depending on variable distribution) to analyze the difference in preoperative and postoperative range of motion and functional outcome scores. χ2 tests were conducted for the radiographic variables and for return to activity and satisfaction. A Spearman test was used to correlate radiographic findings with range of motion and functional outcome. Statistical analyses were conducted using SAS 9.4 (SAS Institute, Cary, NC, USA) and performed by a trained statistician. Two-tailed P values of < .05 were considered significant.
Figure 2 Zones for analysis of radiographic adaptations. L1, lateral metaphysis; L2, lateral diaphysis; US, under stem; M1, medial metaphysis; M2, medial diaphysis; G, glenoid. (Reproduced with permission from Schnetzke M, Coda S, Raiss P, Walch G, Loew M. Radiologic bone adaptations on a cementless short-stem shoulder prosthesis. J Shoulder Elbow Surg 2016;25:650-7.)
the 2 groups with the exception that follow-up was longer in group A and group B had higher preoperative forward flexion (Table I).
Clinical results There was no difference in range of motion or functional outcome scores between the 2 groups (Table II). In group A, there were 3 complications in 42 patients (7.1%), including 1 subscapularis repair failure that was revised, 1 revision to a reverse TSA for stem loosening at 3 years postoperatively, and 1 postoperative rotator cuff failure that was managed conservatively. In group B, there were no complications or revision operations.
Results Radiographic results There were 93 cases that met the study criteria. Complete 2-year follow-up was available in 77 cases (82.8%), which included 42 patients in group A and 35 patients in group B. Baseline demographics and follow-up were similar between
On implantation, the stems in group A were placed in anatomic alignment in 90% of cases, whereas the stems in group B were placed anatomically in 86% of cases (P = .724). There
ARTICLE IN PRESS 4
P.J. Denard et al. Table I
Baseline characteristics
Age (y) Follow-up (mo) Sex Male Female Dominant arm Forward flexion External rotation Internal rotation VAS pain ASES SST SANE
Table III
Group A Ascend (n = 42)
Group B Apex (n = 35)
P value
67.7 (36-84) 35.0 (24-46)
66.5 (52-79) 25.0 (24-29)
.565 <.001
28 (67) 14 (33) 22 (55) 103° (± 28°) 21° (± 15°) S1 6.9 (± 1.8) 34.5 (± 15.2) 3.8 (± 2.5) 29.8 (± 15.1)
22 (63) 13 (37) 23 (66) 119° (± 27°) 26° (± 11°) L5 6.1 (± 2.2) 36.6 (± 16.1) 4.2 (± 2.6) 32.0 (± 24.2)
.812 .345 .009 .052 .067 .199 .708 .492 .838
ASES, American Shoulder and Elbow Surgeons; SANE, Single Assessment Numeric Evaluation; SST, Simple Shoulder Test; VAS, visual analog scale. Categorical variables are presented as number (%). Continuous variables are presented as mean (± standard deviation) or mean (range). Bold values are statistically significant.
Table II
Postoperative outcome
Forward flexion External rotation Internal rotation VAS pain ASES SST SANE
Group A Ascend (n = 42)
Group B Apex (n = 35)
P value*
138° (± 27°) 49° (± 22°) L2 1.2 (± 2.2) 82.7 (± 20.3) 9.6 (± 2.8) 84.5 (± 18.5)
147° (± 19°) 55° (± 12°) L2 1.0 (± 1.3) 86.5 (± 11.7) 10.0 (± 2.4) 90.7 (± 10.3)
.135 .715 .479 .490 .611 .740 .282
ASES, American Shoulder and Elbow Surgeons; SANE, Single Assessment Numeric Evaluation; SST, Simple Shoulder Test; VAS, visual analog scale. * P values are based on comparing gains between preoperative and postoperative range of motion or functional scores.
was no difference in any functional outcome scores or range of motion between stems that anatomically and nonanatomically aligned (P > .05). Filling ratios were lower in group A compared with group B for both FRmet (55.0 vs. 61.6; P < .001) and FRdia (56.1 vs. 51.5; P = .006). In group A, the mean total radiographic change score of the humerus was 3.9, with 2 (4.8%) having no change, 14 (33.3%) having mild changes, 25 (59.5%) having moderate changes, and 1 (2.4%) having advanced changes. Therefore, changes were classified as low in 38% and high in 62%. In group B, the mean total radiographic change score of the humerus was 2.5, with 7 (20.0%) having no change, 20 (57.1%) having mild changes, 8 (22.9%) having moderate changes, and none having advanced changes. Therefore, in this group, changes were low in 77% and high in 23% (P < .001). There were multiple differences in the incidence
Radiographic adaptations by humeral stem
Zone/adaptation
Ascend (n = 42)
Apex (n = 35)
P value
Total score L1 No change CL CNO SW L2 No change CL CNO SW US No change CL CNO SW M1 No change CL CNO SW M2 No change CL CNO SW
3.9
2.5
9 (21.4) 2 (4.8) 30 (71.4) 0 (0)
15 (42.9) 10 (28.6) 7 (20.0) 0 (0)
.043 .009 <.001 NA
6 (14.3) 1 (2.4) 27 (64.3) 13 (30.9)
22 (62.9) 2 (5.7) 4 (11.4) 5 (14.3)
<.001 .588 <.001 .109
16 (38.1) 5 (11.9) 0 (0) 16 (38.1)
21 (60.0) 10 (28.6) 0 (0) 6 (17.6)
.055 .086 NA .051
2 (4.8) 0 (0) 39 (92.9) 0 (0)
10 (28.6) 6 (17.1) 22 (62.9) 0 (0)
.009 .007 .002 NA
12 (28.6) 2 (4.8) 21 (50.0) 8 (19.0)
20 (57.1) 4 (11.4) 5 (14.3) 6 (17.1)
.011 .402 .001 .829
.002
L1, lateral metaphysis; L2, lateral diaphysis; US, under stem; M1, medial metaphysis; M2, medial diaphysis; CL, condensation line; CNO, cortical thinning or osteopenia; SW, spot weld; NA, not applicable. Data are expressed as number (%). Bold values are statistically significant.
of postoperative radiographic changes, including condensation lines, cortical thinning, and spot welds, that are summarized in Table III. Subjective shoulder value (SSV) improvement was lower in patients with cortical thinning in the lateral diaphysis (L2) compared with those without cortical thinning (SSV improvement 46 vs. 60; P = .038). Otherwise, condensation lines, cortical thinning, and spot welds were not correlated with functional outcome. Medial calcar osteolysis was partial in 64% and complete in 7% of group A (Fig. 3). In group B, partial calcar osteolysis was seen in 28.5% of cases, and there were no cases of complete osteolysis (P < .001) (Fig. 4). Postoperative external rotation was lower in the patients with complete calcar osteolysis (53° none, 53° partial, 20° complete; P = .033). Six stems (14.2%) in group A demonstrated postoperative shift or subsidence compared with 3 stems (8.6%) in group B (P = .499). Given that group A contained prostheses with 2 different proximal stem coatings, separate analyses were performed. First, an analysis was performed to compare radiographic changes between the 2 different variations within group A
ARTICLE IN PRESS Short stem radiographic changes
5 was no statistically significant difference in loosening or shift/ subsidence between the 2 (13% and 16%, respectively; P > .05) (Fig. 5). Finally, the subset of stems in group A with proximal coating (Ascend Flex) were compared with group B separately. As in the overall comparison, the total adaptive change score (P = .019) and rate of medial calcar osteolysis (P = .011) were higher in this subset of group A compared with group B. Cortical thinning was more common in this subset of group A compared with group B for the lateral metaphysis (79% vs. 20%; P < .001), the lateral diaphysis (53% vs. 11%; P = .002), the medial metaphysis (95% vs. 63%; P = .011), and the medial diaphysis (50% vs. 14%; P = .009). There was no difference between these 2 subsets in the rate of loosening or shift/subsidence.
Discussion
Figure 3 The 2-year postoperative radiograph of an indivudal in groupAwith anAscend Flex stem shows partial medial calcar osteolysis.
Figure 4 The 2-year postoperative radiograph of an indivudal in group B with an Apex stem shows partial medial calcar osteolysis.
(Ascend and Ascend Flex). There was no difference between the 2 proximal coatings with regard to medial calcar osteolysis, total adaptive change score, and condensation lines, cortical thinning, or spot welds in any zone (P > .05). There
In this study, we sought to compare the short-term functional outcome and radiographic adaptations of 2 different short press-fit humeral stems commonly used in TSA. Whereas functional outcome was comparable between the 2 stems, our findings refute the hypothesis that radiographic adaptations would be the same. These findings may have implications for the design of short press-fit humeral stems. Stress shielding is a common phenomenon after TSA with a traditional-length stem. Raiss et al analyzed 395 humeral stems, of which 67 were press-fit TSAs, at a mean followup of 8.2 years.5 In the press-fit TSA group, 42% had osteolysis of the greater tuberosity, and 39% had osteolysis of the medial calcar (21% complete and 18% partial). The concern for stress shielding as well as the potential ease of revision has led to the development of shorter humeral stems. In a finite element analysis, Razfar et al examined humeral bone stresses with a stemless component, short stem, or traditional-length stem.6 The traditional-length stem and short stem had a rectangular geometry similar to the stem in group B in our study. Native proximal humeral cortical bone stresses were more closely mimicked with a short stem (78% of native) compared with the traditional-length stem (58% of native). They concluded that a short stem could be used to reduce stress shielding, assuming adequate stability. However, few clinical studies have evaluated short stems to date, and no studies to our knowledge have compared different short-stem designs. Schnetzke et al evaluated bone adaptations in 52 short stems (Ascend) at a mean of 32 months postoperatively.7 They used the same radiographic analysis as in this study and noted high adaptations in 52% of cases, which was similar to the 62% rate seen in this study in group A. They did not report the rate of medial calcar osteolysis but noted medial cortical thinning or osteopenia in 83%, which was similar to the 93% observed in this study. Casagrande et al evaluated 73 shortstem TSAs (Ascend) for radiolucent lines, loosening, and calcar osteolysis at a minimum of 2 years postoperatively.1 Nine (12%) were revised, 6 (8.2%) of which were for humeral loosening. They did not specifically report the rate of subsidence
ARTICLE IN PRESS 6
P.J. Denard et al.
Figure 5 (A) Immediate postoperative radiograph of an individual in group A with an Ascend Flex stem. (B) The 2-year postoperative radiograph demonstrates cortical thinning/osteopenia in the lateral metaphysis and diaphysis, a spot weld in the lateral diaphysis, and slight varus shift of the humeral stem. The postoperative subjective shoulder value was 90.
or shift, but these findings were used to categorize “at risk for loosening,” which they reported in an additional 6 stems (8.7%). In this study, our rate of revision for loosening was lower at 2.4% (1 of 42), but shift or subsidence was seen in 14%. Their rate of overall loosening therefore appears similar to that in our study. With regard to medial calcar osteolysis, Casagrande et al noted partial osteolysis in 14.5% and complete osteolysis in 3%, whereas our rates were much higher at 64% and 7%, respectively. Given that we had a 9-fold difference between partial and complete and they had a 5-fold difference, we may have had a lower threshold for classifying osteolysis. In an effort to reduce the risk of humeral loosening, the first-generation Ascend stem was modified in the secondgeneration Ascend Flex to include a proximal titanium plasma spray. Morwood et al compared 2-year outcomes in 34 firstgeneration Ascend stems without proximal porous coating with 34 Ascend Flex stems with proximal porous coating.2 They had 55% follow-up at a mean of 30 months postoperatively. Compared with the proximally coated design, the firstgeneration stem had a significantly higher risk of loosening (3% vs. 21%; P = .03) and radiolucencies (21% vs. 44%; P = .04). They concluded that proximal ingrowth coating on short stems may decrease the risk of loosening. In this study, we did not observe any difference in the rate of adaptive changes or subsidence/shift between the Ascend and Ascend Flex stems. This difference could be due to the different radiographic analyses between our studies, the small numbers in our study, or the follow-up differences (83% at 35 months in this study vs. 55% at 30 months in their study). Whereas proximal coating may influence fixation of a short stem, this may be design specific, and multiple design factors likely influence radiographic adaptations and loosening. In this study, the mean radiographic adaptation change score was higher in group A compared with group B (3.9 vs. 2.5). High
adaptations were seen in 61.9% of group A compared with 22.9% in group B. In addition, the rate of medial calcar osteolysis was 71% in group A vs. 28.5% in group B. Although overall functional outcome was not different between the 2 groups, outcome was influenced by radiographic changes. SSV and postoperative external rotation were lower in the setting of lateral cortical thinning and medial calcar osteolysis, respectively, and both radiographic findings were more common in group A. Longer follow-up is needed to evaluate the impact of these differences on implant survival and function. Whereas subsidence or shift was higher in group A (14.2% vs. 8.6%), the difference did not reach statistical difference with the numbers studied. At these values, it would require 505 patients in each group to have 80% power to detect a difference between the 2 groups; if the values were 20% and 10%, it would require 199 patients in each group. The humeral stems in this study have substantial differences in shape and design. The group A stem (Ascend) is oval, curved, and collarless (Ascend and Ascend Flex), whereas the group B stem (Apex) is rectangular and straight and has a collar. These design variables may lead to different proximal humeral bone stresses that influence radiographic adaptations. The curved oval stem in group A tapers in width substantially distally. Despite having a lower FRmet on 2-dimensional images with an anteroposterior radiograph, the oval shape may actually result in greater 3-dimensional (3D) proximal fill, which creates stress shielding at the periphery and could account for the greater calcar osteolysis in this group. Alternatively, it is possible that the curved stem alters proximal stresses unevenly to a greater extent than a straight stem and thus leads to greater adaptive changes. The collar feature is also a substantial difference in the designs. A collarless design may benefit from proximal porous coating to achieve fixation, whereas a collared design may have less need for proximal coating as suggested by the low rate of loosening in group B
ARTICLE IN PRESS Short stem radiographic changes in our study. It is notable that in the subanalysis of the stems in group A with proximal coating to group B, which does not have proximal coating, the overall differences were unchanged. Radiographic adaptations remained higher in this subset of group A compared with group B. This provides further evidence that the requirement for proximal coating is design specific. In other words, a collarless stem may benefit from proximal coating, but based on the results of this study, there is not a need for proximal coating with a collared stem. Filling ratio has also been reported to affect radiographic changes. Schnetzke et al observed that higher filling ratios of the implant relative to the proximal humerus were associated with higher radiographic adaptations.7 This finding was also supported by Nagels et al in an evaluation of traditionallength stems.3 Interestingly, in this study, filling ratios were higher in group B, which had fewer radiographic adaptations. This suggests that implant design is more important than filling ratio. However, the filling ratio is difficult to compare between the 2 stems because the designs are different in shape. The evaluation we performed on plain anteroposterior radiographs is insufficient to compare the difference in volume of bone that is occupied between the 2 stems; such evaluation would require 3D analysis. Nonetheless, it seems desirable on the basis of the previous studies both to implant and to design press-fit stems with the minimum fill required to achieve fixation. Preoperative templating or 3D planning of the humerus preoperatively may be useful for limiting the filling ratio. The major strength of this study is that it provides a comparison between 2 different humeral stems with a consistent surgical technique. However, there are several limitations. The follow-up is short term, and results may change at midterm to long term. Follow-up was longer in group A, which may have affected the results. However, previous study has suggested that stress shielding was detected in all cases by 24 months postoperatively without a difference in 24-month compared with 5-year follow-up; this suggests that our minimum follow-up was sufficient to detect radiographic changes of the humerus.9 The study design was retrospective and therefore introduces the potential for selection bias. We believe this plays a minor role, given that both surgeons adopted the group B stem at the same time. The radiographic review was performed by 2 surgeons, 1 of whom performed several of the operations. Although we attempted to eliminate bias by having a separate review, it is possible that this may have introduced bias. Finally, given that there are multiple differences in the designs of the 2 stems, it is unclear which factors (ie, collar vs. collarless or oval vs. rectangular) have the most impact on stress shielding. A finite element analysis examining these factors may be useful.
Conclusions At short-term follow-up, there is no difference in functional outcome or revision between 2 different humeral
7 stem designs. However, adaptive changes and the rate of medial calcar osteolysis varied substantially between the 2 short press-fit humeral components. Further study is needed to evaluate the long-term impact of these differences and to elucidate the design factors of short stems that are associated with stress shielding.
Disclaimer This study was supported by a grant from Arthrex, Inc. Patrick J. Denard, Reuben Gobezie, Anthony A. Romeo, and Evan Lederman are consultants for and receive royalties from Arthrex, Inc. The other authors, their immediate families, and any research foundations with which they are affiliated have not received any financial payments or other benefits from any commercial entity related to the subject of this article.
References 1. Casagrande DJ, Parks DL, Torngren T, Schrumpf MA, Harmsen SM, Norris TR, et al. Radiographic evaluation of short-stem press-fit total shoulder arthroplasty: short-term follow-up. J Shoulder Elbow Surg 2016;25:1163-9. http://dx.doi.org/10.1016/j.jse.2015.11.067 2. Morwood MP, Johnston PS, Garrigues GE. Proximal ingrowth coating decreases risk of loosening following uncemented shoulder arthroplasty using mini-stem humeral components and lesser tuberosity osteotomy. J Shoulder Elbow Surg 2017;26:1246-52. http://dx.doi.org/10.1016/ j.jse.2016.11.041 3. Nagels J, Stokdijk M, Rozing PM. Stress shielding and bone resorption in shoulder arthroplasty. J Shoulder Elbow Surg 2003;12:35-9. http://dx.doi.org/10.1067/mse.2003.22 4. Raiss P, Bruckner T, Rickert M, Walch G. Longitudinal observational study of total shoulder replacements with cement: fifteen to twenty-year follow-up. J Bone Joint Surg Am 2014;96:198-205. http://dx.doi.org/10 .2106/JBJS.M.00079 5. Raiss P, Edwards TB, Deutsch A, Shah A, Bruckner T, Loew M, et al. Radiographic changes around humeral components in shoulder arthroplasty. J Bone Joint Surg Am 2014;96:e54. http://dx.doi.org/ 10.2106/JBJS.M.00378 6. Razfar N, Reeves JM, Langohr DG, Willing R, Athwal GS, Johnson JA. Comparison of proximal humeral bone stresses between stemless, short stem, and standard stem length: a finite element analysis. J Shoulder Elbow Surg 2016;25:1076-83. http://dx.doi.org/10.1016/ j.jse.2015.11.011 7. Schnetzke M, Coda S, Raiss P, Walch G, Loew M. Radiologic bone adaptations on a cementless short-stem shoulder prosthesis. J Shoulder Elbow Surg 2016;25:650-7. http://dx.doi.org/10.1016/j.jse.2015.08 .044 8. Singh JA, Sperling JW, Cofield RH. Revision surgery following total shoulder arthroplasty: analysis of 2588 shoulders over three decades (1976 to 2008). J Bone Joint Surg Br 2011;93:1513-7. http://dx.doi.org/ 10.1302/0301-620X.93B11.26938 9. Spormann C, Durchholz H, Audigé L, Flury M, Schwyzer HK, Simmen BR, et al. Patterns of proximal humeral bone resorption after total shoulder arthroplasty with an uncemented rectangular stem. J Shoulder Elbow Surg 2014;23:1028-35. http://dx.doi.org/10.1016/j.jse.2014.02.024