This study was

Brigham and W

*Reprint req

Medicine Servi

Women’s Hospi

E-mail addre

J Shoulder Elbow Surg (2014) 23, 1318-1326

1058-2746/$ - s

http://dx.doi.org

www.elsevier.com/locate/ymse

Imaging characteristics of lesser tuberosity

osteotomy after total shoulder replacement: astudy of 220 patients

Kirstin M. Small, MD, MBAa, Elana J. Siegel, BAb, Lindsay R. Miller, MPHb,Laurence D. Higgins, MDb,*

aDepartment of Radiology, Brigham and Women’s Hospital, Boston, MA, USAbBoston Shoulder Institute, Brigham and Women’s Hospital, Boston, MA, USA

Background: The lesser tuberosity osteotomy (LTO) has been gaining popularity as a method of exposingthe glenohumeral joint during total shoulder replacement, whereby a small fragment of bone is removedfrom the lesser tuberosity, thus preserving the subscapularis tendon. To date, no large, randomized studieshave reported evaluations of LTO healing and healing rates.Methods: We reviewed the radiographs and available computed tomography images of 362 patients whounderwent a total shoulder arthroplasty by the same surgeon between 2006 and 2012. The integrity of theLTO site was graded as not seen, bony union, nondisplaced nonunion, and displaced nonunion. The smok-ing status of patients was also assessed.Results: Of 362 patients investigated, 220 had a minimum of 6 months of radiographic follow-up. The LTOsite was not seen in 37 patients; of the remaining 183, 159 patients (86.89%) demonstrated bony union,8.80% of whom were smokers; 16 patients (8.74%) demonstrated nondisplaced nonunion, 6.3% of whomwere smokers; and 8 patients (4.3%) demonstrated displaced nonunion, 25.0% of whom were smokers. Over-all, 19 of the 24 nonunions were in male patients (79.1%) and 5 were in female patients (20.8%).Conclusions: This is the first large-scale study to report the healing rate of LTOs. LTO healing is bestassessed on radiographs; if nondisplaced or displaced nonunions are suspected, computed tomography canbe a helpful additional examination. The number of radiographs where there is a lack of adequate visualizationof the LTO site raises important questions about definitive radiographic evaluation using current techniques.Level of evidence: Level IV, Case Series, Treatment Study.� 2014 Journal of Shoulder and Elbow Surgery Board of Trustees.

Keywords: Total shoulder arthroplasty; lesser tuberosity osteotomy healing; joint replacement

approved by the Partners Human Research Committee,

omen’s Hospital (Protocol #: 2012-P-000631/1).

uests: Laurence D. Higgins, MD, Shoulder & Sports

ce, Department of Orthopedic Surgery, Brigham and

tal, 75 Francis St, Boston, MA 02115, USA.

ss: ldhiggins@partners.org (L.D. Higgins).

ee front matter � 2014 Journal of Shoulder and Elbow Surgery

/10.1016/j.jse.2013.12.010

Surgical management of the subscapularis tendon duringtotal shoulder replacement remains a topic of debate. Whenthe deltopectoral approach is used, the traditional sub-scapularis tenotomy provides direct access to the gleno-humeral joint and relies on primary tendon-to-tendonhealing to restore subscapularis integrity. With concernsabout consistent tendon-to-tendon healing and the histologicdata that the tendon-to-tendon healing does not heal with

Board of Trustees.

Characteristics of LTO after shoulder replacement 1319

new tendon, but rather disorganized scar,21 lesser tuberosityosteotomy (LTO) has recently gained popularity. Severalclinical series have documented improved subscapularisfunction compared with the standard transtendinousapproach in postoperative patients.4,7,19 By removing asmall fragment of bone from the lesser tuberosity with thesubscapularis tendon still attached, exposure to the gleno-humeral joint is similarly possible. The osteotomy may bereattached with transosseous sutures once the jointreplacement is complete. This technique evolved from thetheory that the healing potential of 2 bone surfaces is greaterthan that of tendon-to-tendon repairs traditionally done.9

Jandhyala et al7 recorded superior subscapularis functionwith LTO over tenotomy in a single-surgeon series with bothtechniques. As reported by Gerber et al,4 the healing of theLTO may be defined as cortical continuity between thefragment of the lesser tuberosity and the cortex of theproximal humerus onto which it has been repositioned.

Most published series using the LTO technique have re-ported osteotomy unions on radiographs to be 100%.4,18,19 Aseries published by Lapner et al9 reported a 95% healing rateof LTO when evaluated by computed tomography (CT) ex-amination. To date, 1 published series of 4 patients hasdescribed lesser tuberosity detachment during the earlypostoperative period.11

The principal objective of this study was to retrospec-tively assess LTO healing from routine postoperative ra-diographs, and CT scans when available, at a minimum of 6months of follow-up. We assess healing at the osteotomysite and describe the radiographic and CT appearance,including a characteristic but not previously described‘‘half-moon’’ sign, as seen in displaced nonunion cases.Healing trends for LTOs were then related to individualsmoking status. In reviewing all postoperative radiographsof a single surgeon, we also obtained the percentage ofpostoperative cases where the radiographic technique wasnot adequate to answer our specific clinical question ofLTO healing.

Materials and methods

Between January 2006 and June 2012, 362 patients were treatedwith a primary total shoulder arthroplasty using the LTO tech-nique at a single tertiary referral center, performed by the seniorauthor (L.D.H.). The study excluded 142 patients secondary to alack of postoperative imaging at a minimum of 6 months; thus,220 patients were included in this study. Of those 220, 183 hadLTOs that could be visualized on radiograph, and in 26 patients,follow-up CT examinations were reviewed in conjunction withradiographs.

Postoperative radiographs consisted of Grashey (true gleno-humeral anteroposterior [AP] with 30�-40� posterior oblique),axillary (patient supine with arm abducted 90� and raised 60�-70�

anteriorly), and transscapular (60� anterior) views. No fluoro-scopic control was used when obtaining radiographs. When therewas clinical concern about underlying subscapularis function, a

CT arthrogram was performed to further evaluate healing status ofthe LTO while simultaneously assessing the integrity of the sub-scapularis tendon. If the patient reported intravenous contrast al-lergy, CT air arthrogram or routine CT was performed.

CT examinations (arthrogram and nonarthrogram) wereobtained at 2-mm slice thickness with coronal and sagittal refor-mations on a 128-slice Siemens ASþ scanner (Siemens Health-care, Erlangen, Germany). A retrospective review of postoperativeradiographs and CTs was performed by consensus between boththe operating surgeon and an experienced fellowship-trainedmusculoskeletal radiologist.

The integrity of the lesser tuberosity osteotomy site was gradedas not seen, bony union, nondisplaced nonunion, and displacednonunion:

� The osteotomy site was defined as ‘‘not seen’’ when thereaders could not confidently see the osteotomy site to gradeits integrity.

� ‘‘Bony union’’ was defined when cortical continuity andbridging trabeculae were seen between the LTO fragment andthe proximal humerus.

� ‘‘Nondisplaced nonunion’’ was defined when the lesser tu-berosity was located in its expected anatomic position butthere was a lack of detectable cortical continuity or bridgingtrabeculae between the lesser tuberosity fragment and theproximal humerus. The term ‘‘fibrous union’’ was not used,because the presence of fibrous tissue at the interface betweenthe LTO and the donor site had not been seen clinically underdirect visualization. In addition, CT arthrography routinelyshowed intra-articular contrast undermining the LTO site.

� ‘‘Displaced nonunion’’ was defined when the lesser tuberosityfragment was not identified at its expected anatomic position.In these cases, the lesser tuberosity fragment was detachedfrom its osteotomy site and subsequently displaced.

The records of patients who were categorized as bony union,nondisplaced nonunion, and displaced nonunion were reviewed toevaluate smoking status.

Surgical technique

A standard deltopectoral approach was performed with the patientin a modified beach chair position with a pneumatic arm holder(Smith & Nephew, Andover, MA, USA). The subscapularis wasexposed and the proximal biceps removed to expose the inter-tubercular groove. The rotator interval was widely opened to exposethe proximal subscapularis, and an osteotomy of approximately5 mm of the lesser tuberosity was performed so that it exited lateralto the articular surface. The osteotomy varies by individual becausethe size of the lesser tuberosity varies by individual. Ideally, theosteotomy is 3- to 4-mm thick and approximately 15-mm wide (themean size of the lesser tuberosity lateral to the articular surface). Inessence, the entire lesser tuberosity in thickness was not harvested,just a thin segment robust enough to reattach. The subscapulariswas taken down in a subperiosteal fashion below the osteotomy andreflected medially. No attempt to ligate the anterior humeralcircumflex artery was made during exposure. A standard totalshoulder arthroplasty was then performed.

The lesser tuberosity was repaired with alternating #5 and #2FiberWire sutures (Arthrex, Naples, FL, USA) passed throughtransosseous tunnels and cerclaging the prosthetic stem (Fig. 1).

Figure 1 Schematic illustration of high tensile strength, non-resorbable sutures passed around the humeral component. Foursutures are passed in transosseous fashion: the 2 sutures goingthrough the 2 superior drill holes (red) are passed inferiorlyaround the stem, and the 2 sutures going through the 2 inferiordrill holes (green) are passed superiorly around the stem.

Table I Multivariable regression models

Model Variables in model

Unadjusted SexModel 1 Sex, surgical sideModel 2 Sex, current smoking statusModel 3 Sex, ageModel 5 Sex, smoking status, surgical sideModel 5 Sex, smoking status, age, surgical side

1320 K.M. Small et al.

Alternating simple and Mason Allen sutures were then placed atthe subscapularis tendon and lesser tuberosity junction. The del-topectoral interval was closed in routine fashion.

The patient engaged in a traditional postoperative therapy pro-tocol with passive range of motion for 4 weeks. Active assistiverange of motion began at 4 weeks and strengthening at 12 weeks. Ifa nonunion was visualized, the physical therapy was modified toavoid excessive external rotation and resisted internal rotation.

Statistical analysis

Variables of interest included the patient’s age at surgery, sex,surgical side, and smoking status. A binary variable was createdfor the outcome of interest, which was bony union of the LTOcompared with nonunion of the LTO (displaced and non-displaced). A binary variable was also created for the patient’ssmoking status: current smokers vs all other (never smokers andformer smokers). In bivariate and multivariable regression, agewas assessed using 10-year increments.

Analysis of radiograph visualization of the LTO according topatient positioning included all 220 patients. All advanced dataanalysis was done excluding those eligible individuals whose LTOcould not be visualized on the radiograph, resulting in 183 in-dividuals included in further analysis. Univariate descriptive sta-tistics were calculated for all variables of interest, with mean,

standard deviation (SD), and range calculated for continuousvariables, and number (%) calculated for categoric variables.Bivariate statistics were calculated for a comparison of explana-tory variables of interest to the binary union variable. The sub-sequent analysis was performed using logistic regression.Bivariate statistics were calculated for patient sex compared withthe other explanatory variables of interest (age, current smokingstatus, and surgical side) also using logistic regression.

Five adjusted multivariable logistic regression models werecreated (Table I). Multivariable models were compared with un-adjusted models as well as with each other by difference in the c2

likelihood ratio to assess for best fit model, with a threshold c2

value determined using a distribution table. Statistical significancewas attained with P < .05. Statistical analysis was performedusing SAS 9.3 software (SAS Institute Inc, Cary, NC, USA).

Results

Of 362 consecutive patients with total shoulder replacementperformed between January 2006 and June 2012, 220 metour follow-up imaging criteria of a minimum of 6 months(range, 3 months-6.5 years). Our criterion of a 6-monthminimum follow-up was instituted with consideration ofthe accepted definition of a ‘‘nonunion’’ as a lack of bonyconsolidation or bridging trabeculae across a fracture orsurgical site after 6 to 9 months.16

Evaluation of radiographic and CT findings

Evaluation of LTOOf the 220 patients with 6 or more months of follow-up, theLTO site was ‘‘not seen’’ in 37 patients (16.8%), leaving 183patients who had a visualized LTO. This was most oftensecondary to suboptimal positioning and nonorthogonalradiographic views. Although portions of the LTO could beseen on these radiographs, the actual osteotomy site was notvisualized in its entirety to confidently assess for the pres-ence of cortical continuity or bridging trabeculae, or both.

Overall, patients whose LTO was visualized were anaverage age of 62.46 years (SD, 11.44; range, 27.26-85.10years), 54.1% were male, 59.9% had surgery on their rightside, and 9.8% were current smokers at the time of surgicalconsent (Table II).

The LTO site demonstrated ‘‘bony union’’ (Fig. 2) in159 patients (86.9%), of whom 80 (50.3%) were male and79 (49.7%) were female. The median age of these patientswas 62.76 years (SD, 11.55; range, 27.26-85.10 years), and

Table II Univariate descriptive statistics of categoricvariables

Variable No. (%) [SD] (Mean or Range)

Age, y 62.5 [11.4] (27.3-85.1)Sex

Male 99 (54.1)Female 84 (45.9)

Union type)

Bony 159 (86.9)Nondisplaced nonunion 16 (8.7)Displaced nonunion 8 (4.4)

SideLeft 69 (40.1)Right 103 (59.9)

Smoking statusNever smoker 88 (50.9)Former smoker 68 (39.3)Current smoker 17 (9.8)

SD, standard deviation.) Excluding not seen.

Figure 2 Bony union at the lesser tuberosity osteotomy site.Axillary radiograph demonstrates cortical continuity and bridgingtrabeculae across the osteotomy site.

Table III Predictor variables of interest by odds of nonunion(displaced and non-displaced) compared with bony union

Variable No. (%)) Nonunion Py

Unadjusted OR (95% CI)y

Age increase10 years

0.846 (0.587-1.218) .3676

SexMale 19 (19.2) 3.752 (1.335-10.539) .0121z

Female 5 (6.0) 1.00Surgical sideLeft 5 (7.3) 0.369 (0.130-1.046) .0608Right 18 (17.5) 1.00

SmokingCurrent 3 (17.7) 1.457 (0.384-5.523) .5797Never þformer

20 (12.8) 1.00

CI, confidence interval; OR, odds ratio.) Number (%) of patients in that category who had any type of

nonunion.y Logistic regression.z Statistically significant.

Figure 3 Nondisplaced nonunion healing of the lesser tuber-osity osteotomy. An axillary radiograph demonstrates the lessertuberosity osteotomy fragment in the correct anatomic position.No cortical continuity or bridging trabeculae is seen between theosteotomy fragment and the underlying humerus.

Characteristics of LTO after shoulder replacement 1321

14 of the 159 patients (8.8%) who had a documentedsmoking status and a bony union were documented ascurrent smokers (Table III). The axillary and transscapularY views were the best radiographic views to visualize bonyunion at the LTO.

‘‘Nondisplaced nonunion’’ of the LTO was noted in 16of the 183 patients (8.7%) whose LTO was visualized. Inthis category, 13 of the 16 patients (81.3%) were male and 3(18.8%) were female. These patients were a median age of61.98 years (SD, 11.22; range, 42.06-82.58 years). One ofthe 16 patients (6.3%) with nondisplaced nonunion wasdocumented as a current smoker.

In 13 of the 16 patients, the nondisplaced nonunion wasbest demonstrated on axillary views (Fig. 3). This finding inmany of these patients was confirmed on the transscapular

views. In 3 of the 16 patients, the transscapular view wasthe best diagnostic view in the radiographic series toidentify the nondisplaced nonunion (Fig. 4, A). In thesecases, it was important to note that the nondisplacednonunion was not well seen on the additional radiographicprojections (Fig. 4, B). These results emphasize theimportance in obtaining true orthogonal views to make anaccurate diagnosis on radiographs. On CT arthrogram im-aging, nondisplaced nonunion demonstrated interposition

Figure 4 Nondisplaced nonunion of the lesser tuberosity osteotomy. (A) A transscapular radiograph demonstrates the lesser tuberosityosteotomy located at the correct anatomic location without cortical continuity or bridging trabeculae. (B) Nonorthogonal axillary radio-graph in the same patient where the arm is slightly twisted does not provide an optimal view of the osteotomy site. As a result, thenondisplaced nonunion is not seen on this view.

Figure 5 Nondisplaced nonunion healing of the lesser tuberosity osteotomy. (A) Axial and (B) sagittal computed tomography arthrogramimages demonstrate interposition of intra-articular contrast (arrows) between the lesser tuberosity osteotomy and the underlying humerus.

1322 K.M. Small et al.

of arthrographic contrast between the osteotomy fragmentand the humerus (Fig. 5). This finding confirms the lack ofintervening tissue between the osteotomy fragment and theunderlying humerus.

Of the 183 patients whose LTO was visualized, ‘‘dis-placed nonunion’’ at the LTO site (Fig. 6) was noted in8 patients (4.3%), of whom 6 (75%) were male and 2(25%) were female. These patients were a median age of57.55 years (SD, 9.57; range, 46.91-73.97 years). Of the8 patients with displaced nonunions, 2 (25.0%) weredocumented as current smokers. In contrast to the other

entities, the AP view proved to be an important source ofinformation, allowing visualization of the displaced lessertuberosity fragment (Fig. 7). A CT examination in 5 of the 8patients confirmed the displaced nonunion.

In patients with displaced nonunion at the LTO site,the displaced lesser tuberosity fragment has a characteristicappearance of a ‘‘half-moon’’: one border demonstratesthe outer curved border of the lesser tuberosity andthe opposing border exhibits a surgically formed straightmargin (Fig. 8). The images of all of our patientswith displaced nonunion demonstrated the ‘‘half-moon’’

Figure 6 Displaced nonunion of the lesser tuberosity osteot-omy. An axillary radiograph demonstrates a displaced lesser tu-berosity fragment.

Characteristics of LTO after shoulder replacement 1323

appearance of the displaced LTO fragment. Finding a dis-placed intra-articular body of this configuration in a patientwith total shoulder arthroplasty, with a history of priorLTO, greatly raises the suspicion for a displaced osteotomyfragment. Most naturally formed intra-articular bodies donot demonstrate such a characteristic shape.

Men had an increased odds of any type of nonunion(displaced and nondisplaced) compared with women,whereas there was no difference in odds of nonunion for ageor current smoking status. There was a nonsignificant lowerrisk for nonunion on the left shoulder, perhaps because it ismost often the nondominant shoulder (Table III). When theassociation between patient sex and other predictor variablesof interest was assessed, a significant association was foundbetween sex and age, so that was taken into account in thebuilding of the final multivariable models (Table IV). Table Vpresents the odds ratio for male sex when adjusted for thecovariates described in the Methods. Although all modelswere a better fit for the data than a model with no covariates,none of the adjusted models were a better fit than the un-adjusted model that included sex as the only covariate.However, the inclusion of a patient’s current smoking status(model 2) influenced the relationship between sex andnonunion by 32.73%, indicating that smoking has somehidden effect that is not adequately visualized by smoking’sindependent association with the union outcome or a patient’ssex (Tables II and III).

Smoking status was documented in the medical recordin 189 of our 220 patients, 20 of whom were documentedas current smokers. Of those whose unions were visualizedon radiograph, 17 (9.83%) were current smokers, and 68(39.31%) were former smokers. There was no difference inodds of nonunion in current smokers compared with allother patients (Table III).

Discussion

Subscapularis dysfunction after total shoulder arthroplastyhas been widely reported and may yield significant pain and

debilitation for patients.2,12,13 Furthermore, with enhancedphysical diagnosis of subtle subscapularis dysfunction,several publications have documented such deficiencies inup to 67% of patients.12 Alternative operative techniques,such as the subscapularis peel and LTO, have evolved as anattempt to optimize clinical function of the subscapularisafter total shoulder arthroplasty. Although Lapner et al9,10

reported no significant difference between LTO and sub-scapularis peel techniques in subscapularis strength, severalbiomechanical studies have demonstrated biomechanicaland clinical improvement using the lesser tuberosity tech-nique.4,8,18,19 To improve the quality of subscapularisrepair, the senior author began routinely using the LTOtechnique in total shoulder arthroplasties starting in theyear 2000.

Another touted advantage of this technique is the abilityto monitor its healing on radiographic follow-up.4,8,18,19

Most published studies have reported 100% healing ratesas determined by radiographs.4,18,19 To date, only one pub-lished series of 4 patients has described lesser tuberositydetachment during the early postoperative period.11 Radio-graphs may also be diagnostic of an osteotomy displacementpostoperatively if a patient experiences a trauma to theoperative shoulder, because it is the likely point of failure inthe early postoperative course.

Our study describes the variable imaging (radiographicand CT) findings of postoperative LTOs when criticallyreviewed; our associated rates are 37 of the 220 patientswere not seen; of the remaining 183, 159 patients presentedwith bony union, 16 with nondisplaced nonunion, and 8with displaced nonunion. In our retrospective study ofpostoperative radiographs, we found that the efficacy ofpostoperative radiographs is highly dependent on patientpositioning and projection of the radiographs. Although intheory radiographic technique should be able to identifyhealing at the osteotomy site, a nonorthogonal view is oftennondiagnostic. Small variations in patient positioning or inthe alignment of the x-ray beam greatly affect accurateassessment of a single small osteotomy site. Our resultsshow that the optimal orthogonal views to evaluate the LTOsite are the axillary and the transscapular views.

Complete detachment of the LTO is the functionalequivalent of complete subscapularis rupture and is associ-ated with a displaced LTO. This complication, althoughinfrequent, is devastating after total shoulder arthroplasty andwill lead to poor clinical outcome with anterior instability.Moeckel et al13 reviewed 236 shoulder arthroplasties andreported 7 complete subscapularis ruptures (3%) that requiredrevision surgery.13 Several authors have detailed factors otherthan the rotator cuff which play a role in glenohumeral sta-bility.3,5,6,20 Not only do component position and versionhave an effect upon stability, but native glenoid version andversion correction also have significant effects on the ultimaterelationship between the humeral head and the glenoid.

In optimized radiographic views, displaced nonunionLTOs can often be seen on all projections. The axillary and

Figure 7 A displaced nonunion lesser tuberosity osteotomy fragment is seen on (A) anteroposterior, (B) axillary, and (C) transscapularradiographs.

Figure 8 The characteristic ‘‘half-moon’’ sign of the displaced, nonunion lesser tuberosity osteotomy fragment is demonstrated on (A) acomputed tomography scan axial image and (B) a computed tomography air arthrogram sagittal reformation, showing the straight surgicalborder (arrow) and outer curved border.

1324 K.M. Small et al.

transscapular views are the best views to see the osteotomysite, but AP views can be used to identify the displacedlesser osteotomy fragment. Our study also describes thecharacteristic but not previously described ‘‘half-moon’’sign. Our findings demonstrate that this novel sign isdiagnostic of a displaced nonunion at the LTO site. Thisfinding was seen on radiographs and CT examinations inour patients with nonunion.

The correlation between smoking and poor fracturehealing is well detailed in previous reports.1,14,15,17 Thesearticles report delayed, incomplete, and nonunion of frac-tures in clinical outcomes studies and experimental animalmodels. There was no significant association in our databetween current smoking at the time of surgery and havingan outcome of nonunion, but smoking did appear to indi-rectly influence the relationship between a patient’s sex and

their union outcome. In a systematic review of 11 studiescontaining 1194 patients, Thomsen et al22 evaluated pre-operative smoking cessation programs and concluded thatpatients who began a program 4 to 8 weeks ahead of sur-gery were most likely to minimize the risk of postoperativecomplications. In fact, tobacco avoidance for fewer than 4weeks before surgery claimed no benefits.22 Yen et al23

found that in an experimental animal model, abstinencefrom smoking effectively combated the adverse effects ofsmoking when this equaled the overall smoking time. Thesenior author of this study routinely counsels his patientsto cease smoking a minimum of 4 weeks before theirreplacement surgery, a practice that should be strictlyenforced to maximize LTO healing potential.

The group whose LTO did not heal is overwhelminglypopulated by younger patients (median age, 57.5 years;

Table IV Variables of interest by patient sex

Variable Male sex P

Unadjusted OR (95% CI))

Age increase 10 years 0.737 (0.563-0.964) .0258y

SmokingCurrent 1.289 (0.467-3.560) .6240Never þ former 1.00

Surgical sideLeft 0.562 (0.303-1.040) .0665Right 1.00

CI, confidence interval; OR, odds ratio.) Logistic regression.y Statistically significant.

Table V Multivariable regression of bony union comparedwith nonunion (displaced and nondisplaced)

Variable Sex) Likelihoodratioy

Pz

Adjusted OR (95% CI) c2 df

Unadjusted 3.752 (1.335-10.539) 7.4829 1 .0062Model 1x 3.334 (1.165-9.535) 9.8266 2 .0073Model 2x 4.980 (1.616-15.346) 10.2250 3 .0060Model 3x 3.622 (1.279-10.256) 7.7282 2 .0210Model 4x 4.446 (1.418-13.935) 11.9804 3 .0075Model 5x 4.312 (1.365-13.624) 12.1539 4 .0162

CI, confidence interval; OR, odds ratio.) Odds of nonunion for men compared with women. OR values in

boldface are statistically significant.y Compared with a model with no covariates.z All P values are statistically significant.x Model 1: adjusted for side of surgery; model 2: adjusted for current

smoking status; model 3: adjusted for age; model 4: adjusted for

smoking status and surgical side; and model 5: adjusted for smoking

status, age, and surgical side.

Characteristics of LTO after shoulder replacement 1325

median age of all total shoulder replacements, 62.8 years),and favors men approximately 4:1; indeed, of a combined24 patients who displayed nondisplaced nonunion oftheir LTO or complete nonunion, 19 (79.2%) were male,whereas only 5 (20.8%) were female. We hypothesize thatyounger, fitter men are returning to active range of motionand lifting ahead of their rehabilitation schedule and thatthe subscapularis may be under increased tension withmore muscular patients, thus affecting the healing of theirLTO. This hypothesis is supported by our analysis, which inTable V shows that men have an increased odds ofnonunion of any type compared with union, both in unad-justed models and when adjusting for a combination ofpatient’s age, surgical side, and smoking status. In addition,patient age and surgical side are important factors toconsider in LTO healing.

Our study has several limitations, including the retro-spective nature of the study. In addition, when reviewingthe radiographs, we were not blind to the study objectives,

possibly increasing our sensitivity in the evaluation of theLTO site. An additional limitation was that consensus wasachieved by the reviewing orthopedic surgeon and muscu-loskeletal radiologist at a single interpretive session ratherthan at separate independent sessions. Finally, all of thepatients were derived from one surgeon’s practice, andalthough they were selected consecutively for this study,only those patients with appropriate postoperative imagingwere included, possibly introducing some bias betweenthose patients who were and were not included in the study.

Although 183 of 220 of radiographs did provideadequate LTO visualization, the potential information fromthe other 37 cases would have enhanced the power of ourstudy. Secondary to radiographic technique, nearly one-fifthof the obtained radiographs failed to help answer ourclinical question. In cases where there is a concern for LTOhealing and initial radiographs do not adequately show theLTO site, clinicians may consider CT for further evaluationof healing and possible displacement.

Conclusions

Our study is the first retrospective study to evaluate anddescribe LTO healing from routine postoperative radio-graphs on a large number of total shoulder arthroplastypatients. In the postoperative evaluation of LTOs,obtaining appropriately positioned and high-quality ra-diographs to accurately assess the osteotomy site is ofparamount importance. We believe that if initial radio-graphs are unrevealing and there is persistent suspicionof a nonhealed osteotomy, repeat radiographs or CTexamination can be considered. If a patient cannotadequately tolerate the axillary position for plain ra-diographs, a CT examination might be the next best test.Our novel ‘‘half-moon’’ sign should add to confidence inthe diagnosis of nonunion.

Finally, smoking cessation is a practice that should bediscussed preoperatively with patients and strictlyenforced to maximize LTO healing potential. In addi-tion, younger and fitter patients, especially men, shouldbe strongly advised to follow the rehabilitation protocolsaccording to schedule and not return to active range ofmotion ahead of schedule.

Further clinical studies regarding the significance of anondisplaced non union and a displaced nonunion areneeded in the future to define the outcomes associatedwith these radiographic findings.

Disclaimer

The authors, their immediate families and anyresearch foundation, with which they are affiliated,

1326 K.M. Small et al.

received no financial payments or other benefits fromany commercial entity related to the subject of thisarticle.

References

1. Castillo RC, Bosse MJ, MacKenzie EJ, Patterson BM. Impact of

smoking on fracture healing and risk of complications in limb-

threatening open tibia fractures. J Orthop Trauma 2005;19:151-7.

2. Edwards B, Boulahia A, Kempf JF, Boileau P, Nemoz C, Walch G.

The influence of rotator cuff disease on the results of shoulder

arthroplasty for primary osteoarthritis: results of a multicenter study. J

Bone Joint Surg Am 2002;84:2240-8.

3. Gerber C, Costouros JG, Sukthankar A, Fucentese SF. Statis posterior

humeral head subluxation and total shoulder arthroplasty. J Shoulder

Elbow Surg 2009;18:505-10. http://dx.doi.org/10.1016/j.jse.2009.03.003

4. Gerber C, Yian EH, Pfirrmann CA, Zumstein MA, Werner CM.

Subscapularis muscle function and structure after total shoulder

replacement with lesser tuberosity osteotomy and repair. J Bone Joint

Surg Am 2005;87:1739-45. http://dx.doi.org/10.2106/JBJS.D.02788

5. Habermeyer P, Magosch P, Lichtenberg S. Recentering the humeral

head for glenoid deficiency in total shoulder arthroplasty. Clin Orthop

Relat Res 2007;457:124-32.

6. Hill JM, Norris TR. Long-term results of total shoulder arthroplasty

following bone-grafting of the glenoid. J Bone Joint Surg Am 2001;

83:877-83.

7. Jandhyala S, Unnithan A, Hughes S, Hong T. Subscapularis tenotomy

versus lesser tuberosity osteotomy during total shoulder replacement:

a comparison of patient outcomes. J Shoulder Elbow Surg 2011;20:

1102-7. http://dx.doi.org/10.1016/j.jse.2011.03.019

8. Krishnan SG, Stewart DG, Reineck JR, Lin KC, Buzzell JE,

Burkhead WZ. Subscapularis repair after shoulder arthroplasty:

biomechanical and clinical validation of a novel technique. J Shoulder

Elbow Surg 2009;18:184-92. http://dx.doi.org/10.1016/j.jse.2008.09.

009. discussion 197-8.

9. Lapner PL, Sabri E, Rakhra K, Bell K, Athwal GS. Healing rates

and subscapularis fatty infiltration after lesser tuberosity osteotomy

versus subscapularis peel for exposure during shoulder arthroplasty.

J Shoulder Elbow Surg 2013;22:396-402. http://dx.doi.org/10.1016/j.

jse.2012.05.031

10. Lapner PL, Sabri E, Rakhra K, Bell K, Athwal GS. Comparison of

lesser tuberosity osteotomy to subscapularis peel in shoulder arthro-

plasty: a randomized controlled trial. J Bone Joint Surg Am 2012;94:

2239-46. http://dx.doi.org/10.2106/JBJS.K.01365

11. Levy JC, Kalandiak S. Radiographic appearance of displaced lesser

tuberosity: a case series. J Shoulder Elbow Surg 2011;20:395-400.

http://dx.doi.org/10.1016/j.jse.2010.07.015

12. Miller SL, Hazrati Y, Klepps S, Chiang A, Flatow EL. Loss of sub-

scapularis function after total shoulder replacement: a seldom recog-

nized problem. J Shoulder Elbow Surg 2003;12:29-34. http://dx.doi.

org/10.1067/mse.2003.128195

13. Moeckel BH, Altchek DW, Warren RF, Wickiewicz TL, Dines DM.

Instability of the shoulder after arthroplasty. J Bone Joint Surg Am

1993;75:492-7.

14. Moghaddam A, Weiss S, Wolfl CG, Schmeckenbecher K,

Wentzensen A, Grutzner PA, et al. Cigarette smoking decreases TGF-

b1 serum concentrations after long bone fracture. Injury 2010;41:

1020-5. http://dx.doi.org/10.1016/j.injury.2010.03.014

15. Moghaddam A, Zimmermann G, Hammer K, Bruckner T,

Grutzner PA, von Recum J. Cigarette smoking influences the clinical

and occupational outcome of patients with tibial shaft fractures. Injury

2011;42:1435-42. http://dx.doi.org/10.1016/j.injury.2011.05.011

16. Murphy WA Jr, Kaplan PA, Resnick D. Temporomandibular mani-

festations of articular disease. In: Resnick D, Kransdorf MJ, editors.

Bone and Joint Imaging. 3rd edition. Philadelphia: Elsevier Saunders,

ISBN 0-7216-0270-3; 2005. p. 5005-518.

17. N�asell H, Ottosson C, Tornqvist H, Linde J, Ponzer S. The impact of

smoking on complications after operatively treated ankle fracturesda

follow-up study of 906 patients. J Orthop Trauma 2011;25:748-55.

http://dx.doi.org/10.1097/BOT.0b013e318213f217

18. Ponce BA, Ahluwalia RS, Mazzocca AD, Gobezie RG, Warner JJ,

Millett PJ. Biomechanical and clinical evaluation of a novel lesser

tuberosity repair technique in total shoulder arthroplasty. J Bone

Joint Surg Am 2005;87(Suppl 2):1-8. http://dx.doi.org/10.2106/JBJS.

E.00441

19. Qureshi S, Hsiao A, Klug RA, Lee E, Braman J, Flatow EL. Sub-

scapularis function after total shoulder replacement: results with lesser

tuberosity osteotomy. J Shoulder Elbow Surg 2008;17:68-72. http://dx.

doi.org/10.1016/j.jse.2007.04.018

20. Spencer EE Jr, Valdevit A, Kambic H, Brems JJ, Iannotti JP. The effect

of humeral component anteversion on shoulder stability with glenoid

component retroversion. J Bone Joint Surg Am 2005;87:808-14. http://

dx.doi.org/10.2106/JBJS.C.00770

21. Strickland JW. Flexor tendons: acute injuries. In: Green D, Hotchkiss R,

Pedersen W, editors. Green’s operative hand surgery. New York:

Churchill Livingstone, ISBN 1416052798; 1999. p. 1851-97.

22. Thomsen T, Villebro N, Møller AM. Interventions for preoperative

smoking cessation. Cochrane Database Syst Rev 2010;7:CD002294.

http://dx.doi.org/10.1002/14651858.CD002294.pub3

23. Yen CY, Tu YK, Ma CH, Yeh JH, Kao FC, Yu SW, et al. Measurement

of tibial endothelial cell function after cigarette smoking, cessation of

smoking and hyperbaric oxygen therapy. Injury 2008;39(Suppl 4):

40-6. http://dx.doi.org/10.1016/j.injury.2008.08.030