Radiographic Study of Humeral Stem in Shoulder Arthroplasty After Lesser Tuberosity Osteotomy or Subscapularis Tenotomy
Authors’ Disclosure Statement: Dr. Voloshin reports that he is a paid consultant for Zimmer Biomet, Smith & Nephew, Arthrex, and Arthrosurface. The other authors report no actual or potential conflict of interest in relation to this article.
Dr. Mason and Dr. Southgate are Orthopaedic Residents, and Dr. Nicandri and Dr. Miller are Associate Professors, Department of Orthopaedics and Rehabilitation, University of Rochester, Rochester, New York. Dr. Buckley is an Orthopedic Surgeon, Northern Rockies Orthopedics; and an Orthopedic Surgeon, Saint Patrick’s Hospital, Community Medical Center, Missoula, Montana. Dr. Voloshin is Professor of Orthopedics; Chief, Shoulder and Elbow Division; and Team Physician, Rochester Knighthawks NLL Lacrosse, Rochester Rattlers MLL Lacrosse Shoulder, Elbow, Knee Surgery-Sports Medicine, Department of Orthopaedics and Rehabilitation, University of Rochester, Rochester, New York.
Address correspondence to: Ilya Voloshin, MD, University of Rochester Medical Center, 601 Elmwood Avenue, Box 665, Rochester, NY 14625 (tel, 585-276-3106; fax, 585-276-2550; email, Ilya_Voloshin@urmc.rochester.edu).
Am J Orthop. 2018;47(5). Copyright Frontline Medical Communications Inc. 2018. All rights reserved.
Lesser tuberosity osteotomy (LTO) and subscapularis tenotomy (ST) are used for takedown of the subscapularis during shoulder arthroplasty. LTO offers the theoretical but unproven benefit of improved healing and function of the subscapularis. However, humeral stem subsidence and loosening may be greater when osteotomy is performed, which may compromise functional outcomes. Our hypothesis is that no difference in proximal collar press-fit humeral stem subsidence or loosening exists, with no impairment of functional outcomes using the LTO technique.
Radiographs of 39 shoulders from 35 patients who underwent shoulder arthroplasty with a minimum of 1 year of radiographic follow-up were included in the study cohort. All patients received the same press-fit implant (Bigliani-Flatow; Zimmer Biomet). We collected data including demographic information; radiographic measurements, including humeral-acromial distance (HAD); subsidence; subluxation index; the presence of lucent lines >2 mm; and functional outcome scores using the Western Ontario Osteoarthritis of the Shoulder Index, the Disabilities of the Arm, Shoulder and Hand (DASH) questionnaire, and the Constant score.
Subsidence was 2.8 ± 3.1 mm for LTO vs 2.5 ± 3.1 mm for ST (P = .72). HAD did not differ between the LTO and ST groups preoperatively (9.5 ± 2.4 mm vs 10.9 ± 2.7 mm, P = .11). The first postoperative and final follow-up films for HAD for the LTO and ST groups showed a statistically significant difference (first postoperative film, 11.9 ± 3.7 mm vs 15.9 ± 4.5 mm, P = .005; final follow-up film, 11.8 ± 3.2 mm vs 14.5 +3.9 mm, P = .03).
We identified no differences in subsidence, lucent lines >2 mm, posterior subluxation, and Constant, and DASH functional outcome scores for patients undergoing total shoulder arthroplasty via the LTO vs ST techniques with the same proximal collar press-fit humeral stem at short-term follow-up.
- LTO and ST remain viable options for takedown of the subscapularis.
- No difference exists in subsidence, lucent lines, and posterior subluxation on radiographic evaluation between LTO and ST.
- No clinically significant difference exists between outcome scores of patients with either technique.
- HAD was statistically significant but not clinically relevant between the 2 techniques.
- Results from the study do not apply to metaphyseal fitting stems, only diaphyseal fitting stems.
During the surgical approach for total shoulder arthroplasty (TSA), the subscapularis is taken down for adequate exposure to the glenohumeral joint. Various methods are available for taking down the subscapularis, including lesser tuberosity osteotomy (LTO) and a subscapularis tenotomy (ST). LTO offers the theoretical but unproven benefit of improved healing and function of the subscapularis secondary to bone-to-bone healing. One concern, however, is that humeral stem subsidence may be greater when an osteotomy is performed owing to compromise of metaphyseal cortical bone, which may compromise functional outcomes. The humeral stem design may also influence subsidence when metaphyseal bone proximally is compromised. This is a concern in both metaphyseal and diaphyseal fitting stems. Metaphyseal collars on diaphyseal fitting stems rely on adequate bone stock in the metaphysis to provide the additional support needed. Also, posterior subluxation remains a challenge in shoulder arthroplasty. The integrity of the subscapularis is important in prevention of posterior subluxation.1 To our knowledge, no study to date has directly compared differences in humeral stem subsidence, loosening, or posterior subluxation between LTO and ST techniques with any humeral stem design. Our hypothesis is that no difference in proximal collar press-fit humeral stem subsidence or loosening exists, with no impairment of functional outcomes using the LTO technique. We also hypothesize that no difference in posterior subluxation exists between LTO and ST techniques.
Materials and Methods
Consecutive patients with a minimum of 12 months of radiographic follow-up were selected from 2007 to 2010 after TSA was performed by 1 of the senior authors (Dr. Miller and Dr. Voloshin). Study patients underwent primary TSA for primary osteoarthritis or rheumatoid arthritis.
Patients were excluded if they underwent TSA for posttraumatic glenohumeral arthritis, hemiarthroplasty, or osteonecrosis. Patients were also excluded if a rotator cuff tear was discovered intraoperatively or if they had a history of a rotator cuff repair. Additional exclusion criteria included postoperative trauma to the operative shoulder, postoperative infection, extensive documentation of chronic pain, and underlying neurologic disorder (eg, Parkinson disease, dystonia). Patients with a history of diabetes mellitus were not excluded.
All patients underwent TSA via a deltopectoral approach in a modified beach chair position. Biceps tendons were tenodesed at the level of the pectoralis major. All patients received the same proximal collar press-fit implant (Bigliani-Flatow; Zimmer Biomet). These stems provide rotational stability in the metaphyseal segment via fins, vertical stability with the proximal collar, and distal fixation via an interference fit. All parts of the procedure were performed in similar fashion with the exception of ST vs LTO (Figures 1A-1D). All patients followed the same postoperative rehabilitation protocol.
Lesser Tuberosity Osteotomy
LTO was performed as the primary or preferred technique of 1 surgeon. After completion of the biceps tenodesis, the lesser tuberosity is reflected off with the subscapularis intact using an osteotome. After placement of the press-fit humeral stem, the LTO is repaired using No. 5 Ethibond Excel sutures (Ethicon) passed through previously created bone tunnels in the greater tuberosity. These sutures are tied over metal buttons over the lateral cortex of the greater tuberosity. Last, the lateral corner of the rotator interval is repaired using a single No. 2 FiberWire (Arthrex).2
ST is the preferred surgical technique of the second surgeon. After a biceps tenodesis, the subscapularis tendon is released from the lesser tuberosity at the margin of the bicipital groove. Through careful dissection, a single flap including the underlying capsule is created and reflected medially to the level of the coracoid. After placement of the press-fit humeral stem and humeral head, the subscapularis is repaired back in place through previous bone tunnels and with a No. 5 Ethibond Excel suture under the appropriate tension. Then, the lateral corner of the rotator interval is closed using a single No. 2 Ethibond Excel suture in a figure-of-eight fashion.2
The primary variables analyzed were subsidence and loosening. Additional variables, including humeral-acromial distance (HAD) and subluxation index, were also analyzed to assess for any additional impact caused by subsidence or loosening.3 All radiographic measurements were taken from the Grashey (true anteroposterior) view, except subluxation index, which was calculated using the axillary view. All radiographic measurements were completed by 3 independent reviewers. All radiographs were completed in a consistent manner according to postoperative protocols.
HAD was measured preoperatively, immediately postoperatively, and at final follow-up at a minimum of 1 year. The HAD was measured from the lowest point on the acromion to the humerus using a perpendicular line (Figure 2).
Subsidence of the prosthesis was calculated by determining the difference between immediate postoperative heights of the prosthesis in comparison to the value of the final follow-up films. To calculate the height, 2 lines were drawn, 1 line was drawn perpendicular to the top of the prosthetic head and 1 perpendicular to the top of the greater tuberosity (Figure 3). This distance was measured both immediately postoperatively and at final follow-up. The height difference between the 2 time points was used to calculate the subsidence.
Posterior subluxation is indicated by a value >65%, a centered head is between 35% and 65%, and anterior subluxation is indicated by a value <35% (Figure 4).3 The subluxation index was calculated using the method outlined by Gerber and colleagues3 using the axillary view radiograph (Figure 4). The midpoint of the glenoid was identified. Then, a perpendicular line (line z) was drawn projecting from the midpoint of the glenoid through the humeral head. Two lines were then drawn in an anterior-to-posterior direction on the humeral head. The first line (measurement A) was drawn from the posterior cortex to the previous perpendicular line (line z). The second line (measurement B) was drawn from the posterior to the anterior cortex. Measurements A and B should be calculated at the same height on the humeral head, roughly at the anatomic neck. The subluxation index is then calculated as A/B × 100. This was calculated both preoperatively and postoperatively.
The humeral stems were evaluated for loosening by assessing for lucency on final radiographic follow-up films. These were evaluated in a zonal fashion as demonstrated by Sanchez-Sotelo and colleagues4 and in Figure 5. The humeral stem-bone interface was evaluated in 8 radiographic zones. A lucency was not considered significant unless >2 mm. The zone of lucency was then determined for any significant lucencies. Zones 1 and 7 were at the level of the neck component, zones 2 and 6 were at the proximal half of the stem component, and zones 3 and 5 were at the distal half of the stem component. Zone 8 was noted directly inferior to the humeral head prosthesis.4
Functional Outcome Evaluation
Before clinical evaluation, each study patient completed the Western Ontario Osteoarthritis of the Shoulder (WOOS) index; the Disabilities of the Hand, Arm and Shoulder (DASH) questionnaire, and the pain and function sections of the Constant score. The functional outcomes scores were captured postoperatively from October to November 2011. The WOOS is a validated outcome measure specific to osteoarthritis of the shoulder and has been used in prior studies evaluating outcomes of TSA.5-7 Previous studies have determined that the minimal clinically important difference for the WOOS score is 15 on a normalized 0 to 100 scale (100 being the best). The DASH score is a validated outcome measure for disorders of the upper extremity but is not specific to osteoarthritis of the shoulder.8 The Constant score is a validated outcome measure for a number of shoulder disorders, including TSA.9,10
Statistical analyses were completed by a trained biostatistician. A power analysis was calculated using the noninferiority test to determine if adequate data had been obtained for this study. This was calculated by using previously accepted data demonstrating a statistically significant difference for subsidence and HAD. The data from these studies were used to make assumptions regarding accepted standard deviations and noninferiority margins, as calculated from the mean values of the 2 groups analyzed in each respective study.4,11 This analysis demonstrated power of 0.97 and 0.85 for the subsidence and HAD, respectively, given the current sample sizes. Intraclass coefficients were calculated to evaluate the measurements obtained during the radiographic analysis to determine the interrater agreement. Two samples’ t tests were calculated for the variables analyzed, along with P values and means.
A total of 51 consecutive patients were retrospectively selected for analysis. Of these, 16 patients were excluded from the study because they had <9 months of radiographic follow-up and were unavailable for further follow-up evaluation. Of the remaining 35 patients available for analysis, 4 patients had bilateral TSA, providing 39 shoulders for evaluation. Demographic characteristics of the study cohort are reported in Table 1. Fifteen patients underwent LTO, and 24 underwent ST. One patient underwent a tenotomy of the right shoulder and LTO of the left shoulder. Three LTOs were performed by the surgeon who primarily performed ST, owing to potential benefits of LTO. He eventually returned to his preferred technique of ST because of surgeon preference. Three ST procedures were completed by the surgeon who typically performed LTO at the start of the series prior to establishing LTO as his preferred technique. There was no significant difference between the study populations in terms of age, follow-up, male-to-female ratio, hand dominance, and primary diagnosis of osteoarthritis vs rheumatoid arthritis.
There was no significant difference in preoperative HAD between the LTO and ST groups (9.5 ± 2.4 mm vs 10.9 ± 2.7 mm, P = .11). The immediate postoperative HAD was statistically significant between the LTO and ST groups (11.9 ± 3.7 mm vs 15.9 ± 4.5 mm, P = .005). There was as statistically significant difference noted in the final follow-up films between the LTO and ST groups (11.8 ± 3.2 mm vs 14.5 ± 3.9 mm, P = .025) (Table 2). There were no statistically significant differences found in subsidence between LTO and ST groups at final follow-up (2.8 mm ± 3.1 mm vs 2.5 mm ± 3.1 mm, P = .72) (Table 2). No statistically significant difference was noted in the subluxation index between the LTO and ST groups (0.55% ± .06% vs 0.54% ± 0.07%, P = .45), but there was a statistically significant difference noted postoperatively between the LTO and ST groups (0.55% ± 0.09% vs .48% ± 0.05%, P = .015) (Table 2).
Two stems were noted to have lucent lines >2 mm, both within the ST cohort. Each had 1 stem zone >2 mm, 1 in zone 7, and 1 in zone 4. No statistically significant difference was identified between the LTO and ST groups (0/15 vs 2/24, P = .51) (Table 2).
Study patients were evaluated using functional outcome scores, including the Constant, WOOS, and DASH scores (Table 3). No statistically significant differences were noted in the DASH scores (8.4 ± 6.6 vs 13.8 ± 4.9, P = .13) or Constant scores (83.3 ± 9.1 vs 81.8 ± 10.1, P = .64) between the LTO and ST cohorts. There was a statistically significant difference between the WOOS scores (93.3 ± 5.3 vs 81.5 ± 20.8, P = .013). Because separate radiographic reviews were done by 3 independent personnel at 3 different times, it was important to ensure agreement among the reviewers. This was compared using the intraclass correlation coefficients. In the statistical analysis completed, the intraclass coefficients showed the 3 reviewers agreed with each other throughout the radiographic analysis (Table 4).
At final follow-up, we identified no statistically significant difference between the LTO and ST patients in subsidence, lucent lines >2 mm, or functional outcomes (Constant and DASH scores) in patients who underwent TSA with the same proximal collar press-fit humeral stem. In regard to the functional outcome scores, although the WOOS score was statistically significant (P = .013) between the LTO and ST cohorts, we do not feel that this is clinically relevant because it does not reach the minimal clinically important difference threshold of 15 points.8
A statistically significant difference was noted in postoperative subluxation index but was not clinically relevant, because the values between the LTO and ST groups (0.55 vs 0.48) still showed a centered humeral head. Gerber and colleagues3 discussed using a value of 0.65 as a measure of posterior humeral head subluxation, whereas Walch and colleagues12 defined posterior humeral head subluxation as a value >0.55. On the basis of these numbers, the values obtained in this study demonstrated that the postoperative values were still centered on the glenoid, and therefore were not clinically significant.3,12
In regard to HAD, there was a statistically significant difference noted postoperatively (P = .005) and at final follow-up (P = .025) between the LTO and ST cohorts. Saupe and colleagues13 demonstrated that a HAD <7 mm was considered abnormal and reflected subacromial space narrowing. The values noted in the LTO and ST patients on postoperative and final follow-up radiographs were statistically significant (Table 2), but not clinically relevant because both were >7 mm. A potential source for the variation in HAD may be due to X-ray position and angle.
Studies have shown a concern regarding the integrity of the subscapularis after tenotomy or peel used in TSA with abnormal subscapularis function.14,15 Miller and colleagues15 reported 41 patients, nearly two-thirds, of whom described subscapularis dysfunction. Those authors’ response to the poor clinical outcomes was to remove a fleck of bone with the tendon to achieve “bone-to-bone” healing.14 Gerber and colleagues16 reported on a series of patients using LTO and repair in TSA with 75% and 89% intact subscapularis function on clinical testing.16 Studies by Qureshi and colleagues17 and Scalise and colleagues18 showed similar results after LTO. Biomechanical studies have shown mixed results. Ponce and colleagues19 showed biomechanically superior results for LTO in comparison to the various repair techniques for ST. In another study, Giuseffi and colleagues20 showed no difference in LTO vs ST during biomechanical testing. In response to the increased concern regarding subscapularis integrity, Caplan and colleagues21 reported on 45 arthroplasties in 43 patients with improved postoperative testing with intact subscapularis testing in 90% to 100% of patients. A level 1 randomized control trial conducted by Lapner and colleagues22 did not demonstrate any clear clinical advantage of LTO vs ST. Controversy still exists regarding which is the preferred technique for TSA.
Sanchez-Sotelo and colleagues4 evaluated uncemented humeral components in 72 patients who underwent TSA. They found a humeral component was at risk for loosening if a radiolucent line ≥2 mm was present in at least 3 radiographic zones. They also evaluated tilt or subsidence by measurement and whether the components were observed to have changed. Their measured values correlated with their observed values. That study provided a benchmark for evaluation of loosening and subsidence used during this study.4 Although radiographic follow-up is limited in this study, we feel that any potential subsidence secondary to use of the LTO technique would be radiographically apparent at 1 year. There were 16 patients without adequate radiographic follow-up included in the study. However, we feel that this was not a large concern, because the study was adequately powered with the patients available to determine a difference based on subsidence.
We found no difference in subsidence, lucent lines >2 mm, posterior subluxation, and the Constant and DASH functional outcome scores when we compared TSA performed by a LTO with an ST technique with proximal collar press-fit humeral stem. These data cannot be extrapolated to metaphyseal fit stems, which may exhibit different settling characteristics in the setting of the LTO technique.
This paper will be judged for the Resident Writer’s Award.