A subacromial injection is commonly performed for diagnostic and therapeutic purposes. The effectiveness of the injection hinges on the precise delivery of the therapeutic medication into the area of abnormality.1,2 Three recent studies have reported accuracy rates ranging from 29% to 83%.1,3,4 Although some authors have recommended the use of ultrasound as an adjunctive guide to increase the success of subacromial injections,2,5 a recent study showed no added benefit of ultrasound guidance over an injection based on anatomical landmarks.6
Health care practitioners use different routes for bursal injections with variable accuracy.7 Strizak et al8 found that the bursa is located more anteriorly in the coracoacromial arch with minimal posterior extension. Consequently, the midlateral route may be closer and more direct to the subacromial bursa. The current research was designed to compare the accuracy and effectiveness of the midlateral and posterior routes of subacromial injection.
Materials and Methods
This prospective, randomized clinical trial received institutional review board approval, and all of the participants gave written informed consent.
This multicenter study was performed initially from September 2015 to March 2016 at the Police General Hospital, Bangkok, Thailand, and subsequently from March 2017 to May 2017 at the Suranaree University of Technology Hospital, Nakhon Ratchasima, Thailand. Adequate sample size was calculated using the formula for comparing binary outcomes (n4Studies version 1.4.1; Prince of Songkla University, Songkla, Thailand)9–11 with a power of 80% to detect a 20% difference; a total of 46 patients were required.
The inclusion criteria were as follows: (1) a history of shoulder pain, especially with overhead activities and night pain; (2) symptoms having persisted for at least 2 months; (3) a positive impingement sign; (4) no major weakness of the rotator cuff with muscle testing; and (5) skeletal maturity. The exclusion criteria were as follows: (1) subacromial injections within the previous 6 months; (2) an os acromiale; (3) primary adhesive capsulitis; (4) previous shoulder surgery; (5) a history of fracture or dislocation; (6) pregnancy; (7) an allergy to any of the injection agents; and (8) a possible full-thickness tear of the rotator cuff tendon (supraspinatus, infraspinatus, subscapularis, and teres minor) that manifested as less than a grade 3 level (Medical Research Council muscle scale) of shoulder external rotation or internal rotation or a positive lag sign on the rotator cuff examination.
Age, sex, body mass index, shoulder involvement, circumference of the proximal humerus, and acromial anatomy were assessed. Diagnosis was based on clinical examination, including medical history and physical examination. The patients were divided into the following 3 groups: (1) impingement syndrome; (2) rotator cuff tear that manifested with some degree of weakness of the shoulder external or internal rotator; and (3) others (calcific tendinitis or biceps tendinitis). Two perpendicular radiographs of the shoulder (anteroposterior and outlet views) were obtained. Outlet-view radiographs of the shoulder were used to describe the acromial type on the basis of the acromial angle in the sagittal plane— type I (0° to 12°), type II (13° to 27°), or type III (>27°).12 Fifty shoulders met these criteria. Patients received a subacromial injection, which was randomized by selection of an opaque envelope.
The injection was performed while patients sat with their arm at their side. One physician (P.G.) administered all of the injections. After sterile preparation of the injection site, a 5-cm, 25-gauge needle was connected to a 10-mL syringe containing 5 mL of 1% lidocaine, 2 mL of iopamidol injection contrast medium, and 1 mL (40 mg) of triamcinolone.
A midlateral injection (Figure 1) was performed directly underneath the midlateral aspect of the acromion, 2 cm below the midlateral aspect, and was directed anteriorly and cephalad, with the arm being distracted by the assistant. In contrast, the posterior injection portal (Figure 2) was located 1 cm posterior and inferior to the posterolateral border of the acromion. The needle was directed cephalad, anteriorly, and medially toward the subacromial space. The medication was delivered when the needle was clinically considered to be in the subacromial space. For both routes of injection, if bone was contacted, the needle was withdrawn slightly to achieve injection without resistance.
A midlateral route of injection was performed directly underneath the midlateral aspect and 2 cm below the lateral border of the acromion. The needle was pointed anteriorly and cephalad.
A posterior route of injection was located 1 cm posterior and inferior to the posterolateral border of the acromion. The needle was directed cephalad, anteriorly, and medially toward the subacromial space.
All patients were evaluated using the modified University of California Los Angeles (UCLA) shoulder score and a 10-point visual analog scale (VAS) pain score (from 0 [no pain] to 10 [maximum pain]) just before the injection process.
The primary outcome of interest was the accuracy of the injection into the bursa. The patients were instructed to perform normal shoulder movements approximately 10 minutes after injection, and 2 perpendicular radiographs of the shoulder (anteroposterior and outlet views) were obtained. A single musculoskeletal radiologist (P.L.) who was blinded to the injection route and other patient-related data subsequently interpreted all of the radiographs 3 times, at 1-week intervals, to determine the intraobserver reliability. The injections were classified as either intrabursal (Figure 3) or extrabursal (Figure 4).
Anteroposterior (A) and outlet (B) radiographs showing accurate intrabursal injection via the lateral route.
Anteroposterior (A) and outlet (B) radiographs showing extrabursal injection via the posterior route.
The secondary outcome of interest was the change in the functional outcomes: the modified UCLA shoulder and VAS pain scores at 30 minutes after the subacromial injection. All patients' pain levels at the injection site were assessed using the VAS, and any side effects or allergic reactions associated with the contrast dye contained in the injections were identified.
Descriptive statistics were used to summarize the demographic data and clinical outcomes. Continuous data were presented as mean with standard deviation or median with range, whereas frequency and percentage were reported for categorical data. For comparisons of the demographic data and clinical outcomes of the 2 groups, the chi-square test, Fisher's exact test, the independent t test, or the Mann–Whitney U test was used, as appropriate. The Shapiro–Wilk test was also used to evaluate the normality of the data.
Fisher's exact test was used to compare the accuracies of the injections via each route. Changes in the modified UCLA shoulder and VAS pain scores were compared using the independent t test or the Mann–Whitney U test.
Potential factors associated with accuracy—the route of injection, sex, circumference of the proximal humerus, body mass index, and age—were determined using simple logistic regression. The crude odds ratio and 95% confidence interval for odds ratio were presented for each variable.
The intraobserver reliability of the radiographic interpretation was measured using Cohen's kappa coefficient. The threshold for significance was considered to be P<.05. All statistical analyses were performed using SPSS Statistics for Windows version 18.0 software (SPSS Inc, Chicago, Illinois) and STATA/SE version 14 software (StataCorp LP, College Station, Texas).
Fifty shoulders (48 patients, with 2 patients having symptoms on both sides) were included in the study. The patients had a mean age of 55.7 years. The patients' demographic data are listed in Table 1. The preinjection mean values of the modified UCLA shoulder and VAS pain scores were 21.66 (range, 8–32) and 6.5 (range, 3–9), respectively. There were more male patients in the posterior route group and more female patients in the midlateral route group (P=.010). The patients in the posterior route group weighed more and were taller than the patients in the midlateral route group (P<.05). Age, body mass index, side of involvement, circumference of the proximal humerus, and preinjection modified UCLA shoulder and VAS pain scores were not statistically different between the 2 groups (P>.05) (Tables 1–2). Analysis of the acromial type in the sagittal plane revealed a type I acromion in 13 shoulders (26%), a type II acromion in 33 shoulders (66%), and a type III acromion in 4 shoulders (8%).
Patient Demographic Data
Injection Accuracy and Clinical Outcomes
The overall accuracy of the injections was approximately 80% (Table 2). There was a significant difference in the accuracy rates of the 2 routes, with the midlateral route achieving 92% accuracy and the posterior route achieving 68% accuracy.
The mean modified UCLA shoulder score increased from 20.40 preinjection to 31.44 postinjection for the midlateral route and from 22.92 preinjection to 31.84 postinjection for the posterior route (Table 2). Improvements in VAS pain scores, from 6.8 preinjection to 2.2 postinjection for the midlateral route and from 6.2 preinjection to 2.0 postinjection for the posterior route, were also observed. The improvements in the postinjection mean modified UCLA shoulder and VAS pain scores of both groups were also statistically significant (P<.001).
However, the postinjection mean modified UCLA shoulder and VAS pain scores of the 2 groups were not statistically different (P>.05). The delta (difference between preinjection and postinjection) modified UCLA shoulder and VAS pain scores for the midlateral and posterior injection routes showed no statistical difference (P>.05).
The intraobserver reliability was found to be kappa=0.761, with P<.001.
The crude odds ratios and 95% confidence intervals for odds ratios of all of the factors potentially associated with route accuracy are listed in Table 3. Injection route had some influence on accuracy, with a crude odds ratio of 5.41 (95% confidence interval, 1.017–28.791; P=.048). Sex, circumference of the proximal humerus, body mass index, and age appeared not to have had an influence on accuracy (P>.05).
Univariate Analysis (Logistic Regression) of Potential Factors Associated With Accuracy of Injections
Relief of pain with provocative maneuvers after the injection of a local anesthetic can confirm a diagnosis of subacromial bursitis. In addition, the effectiveness of an injection hinges on the precise delivery of the therapeutic medication into the pathological location.1,2 In this study, the authors found that the accuracy of the midlateral route (92%) was significantly higher than that of the posterior route (68%). Marder et al7 also found that the accuracy of the posterior route was significantly lower than that of either the anterior route or the midlateral route, with a rate of 56% for the posterior route, 84% for the anterior route, and 92% for the midlateral route. However, Kang et al13 found that the overall accuracy of the 3 routes of subacromial injection (anterolateral, midlateral, and posterior) was 70%, with no difference between the 3 portals. In addition, Mathews and Glousman14 found comparable degrees of accuracy of subacromial injections via the anterolateral and the posterior approaches in a cadaveric model.
The authors believe that the anatomy of the subacromial bursa and the intervention techniques are the main foci of this study. Strizak et al8 found that the subacromial bursa is located predominantly beneath the anteriorly situated coracoacromial arch, with a major portion of the bursa extending laterally beyond the margin of the acromion and a minimal portion extending posteriorly. Beals et al15 confirmed that the subacromial bursa occupied the anterior portion of the acromion without extension to the posterior aspect. In view of the anatomical and pathological aspects, the midlateral route may be closer and more direct to the subacromial bursa than the posterior route.
For the midlateral approach in this study, the authors designed the position to be in the same location as that used for creating the midlateral portal while performing a rotator cuff repair. The patients' arms were distracted during injection. The shoulder surgeons were familiar with this approach.
Taking into consideration the distance from the skin to the subacromial bursa and the directional aspects, the authors believe that employing the midlateral approach makes it easier to aim for the target, the subacromial bursa. Sardelli and Burks16 found that the mean distances to the subacromial bursa were 2.9±0.6 cm with an anterior needle placement, 2.9±0.7 cm with a lateral placement, and 5.2±1.1 cm with a posterior placement. In the current study, a 5-cm, 25-gauge standard needle was used. This may not have been long enough to reach the subacromial bursa in some patients, especially those with larger shoulders or severe obesity, when using the posterior route of injection.
In addition, 2 patients in this study who had pain in both shoulders received subacromial injections through both the lateral and the posterior routes. The authors found 100% accuracy for the lateral route (2 of 2 injections) and 0% accuracy for the posterior route (0 of 2). Therefore, the authors advocate that the lateral route is the preferred choice for high-accuracy subacromial injections.
After injection, there were significantly improved modified UCLA shoulder and VAS pain scores for both routes (P<.001). Nevertheless, this study did not reveal any statistically significant difference in the functional outcomes of the 2 routes (P>.05). It is possible that the local anesthetic, whether intrabursal or extrabursal, blocks the sensory innervation to the subacromial bursa. Kang et al13 also reported no significant improvement in patient function, pain, or satisfaction with an accurate injection compared with an inaccurate injection at the 3-month follow-up. However, Marder et al7 found greater initial pain relief among patients who had intrabursal subacromial steroid injection. Eustace et al1 found greater clinical improvement at the 2-week follow-up evaluation among patients who had accurate steroid placement on radiographs. The authors believe that the midlateral route of subacromial injection offers greater accuracy than the posterior route, and it may help to diagnose patients with impingement syndrome, who would be provided with early pain relief after the injection.
In this study, on logistic regression with P>.05, it was found that age, sex, body mass index, and circumference of the proximal humerus did not influence the accuracy of the injections. Kang et al13 also found no differences in accuracy with various body mass indexes; similar findings were reported by Henkus et al.17 However, the current authors found that the route of the injection had some influence on the accuracy (P=.048).
Additionally, drawing on the authors' direct experience, there are some limitations to directing the needle in obese patients. Cook et al18 found that a longer needle should be considered for patients with a body mass index of greater than 35 kg/m2 because of the thickness of the subcutaneous tissues. Sardelli and Burks16 also found that there was a longer mean distance for placement of the posterior needle (5.2±1.1 cm) than for placement of the lateral needle (2.9±0.7 cm). Consequently, for obese patients or patients who have large proximal humeri, the authors recommend the use of the midlateral route for subacromial injections. Moreover, for surgeons who prefer the posterior route for subacromial injections, the authors recommend that an adequate needle length (ie, >5 cm) be employed with good techniques.
The major strength of this study was that it was a true randomized study with sufficient statistical power to answer the research question. Moreover, 1 physician performed all of the injections, which reduced bias.
Nevertheless, there were some limitations. Because of the research design, the physician could not be blinded. However, the VAS pain score and functional outcomes were measured by others who were blinded to the type of subacromial injection. Moreover, the radiographs were evaluated by 1 musculoskeletal radiologist who was blinded to the study. To compensate, the authors assessed the intraobserver reliability to increase the strength of the results. Finally, the authors did not record the functional outcomes or VAS pain scores at mid-term or long-term follow-up; rather, these items were only recorded shortly after the administration of the injections. Future studies with longer follow-up may be necessary.
This study found that the midlateral route for subacromial injection was more accurate than the posterior route. Although the accuracy of the injections was not related to factors such as age, sex, body mass index, or circumference of the proximal humerus, injection route had an influence.
- Eustace JA, Brophy DP, Gibney RP, Bresnihan B, FitzGerald O. Comparison of the accuracy of steroid placement with clinical outcome in patients with shoulder symptoms. Ann Rheum Dis. 1997;56(1):59–63. doi:10.1136/ard.56.1.59 [CrossRef]
- Naredo E, Cabero F, Beneyto P, et al. A randomized comparative study of short term response to blind injection versus sonographic-guided injection of local corticosteroids in patients with painful shoulder. J Rheumatol. 2004;31(2):308–314.
- Partington PF, Broome GH. Diagnostic injection around the shoulder: hit and miss? A cadaveric study of injection accuracy. J Shoulder Elbow Surg. 1998;7(2):147–150. doi:10.1016/S1058-2746(98)90226-9 [CrossRef]
- Yamakado K. The targeting accuracy of subacromial injection to the shoulder: an arthrographic evaluation. Arthroscopy. 2002;18(8):887–891. doi:10.1053/jars.2002.35263 [CrossRef]
- Chen MJ, Lew HL, Hsu TC, et al. Ultrasound-guided shoulder injections in the treatment of subacromial bursitis. Am J Phys Med Rehabil. 2006;85(1):31–35. doi:10.1097/01.phm.0000184158.85689.5e [CrossRef]
- Rutten MJ, Maresch BJ, Jager GJ, de Waal Malefijt MC. Injection of the subacromialsubdeltoid bursa: blind or ultrasound-guided?Acta Orthop. 2007;78(2):254–257. doi:10.1080/17453670710013762 [CrossRef]
- Marder RA, Kim SH, Labson JD, Hunter JC. Injection of the subacromial bursa in patients with rotator cuff syndrome: a prospective, randomized study comparing the effectiveness of different routes. J Bone Joint Surg Am. 2012;94(16):1442–1447. doi:10.2106/JBJS.K.00534 [CrossRef]
- Strizak AM, Danzig L, Jackson DW, Resnick D, Staple T. Subacromial bursography: an anatomical and clinical study. J Bone Joint Surg Am. 1982;64(2):196–201. doi:10.2106/00004623-198264020-00008 [CrossRef]
- Rosner B. Fundamentals of Biostatistics. 5th ed. Pacific Grove, CA: Duxbury-Thomson Learning; 2000:384–385.
- Fleiss JL, Levin B, Paik MC. Statistical Methods for Rates and Proportions. 3rd ed. Hoboken, NJ: John Wiley & Sons; 2003:76.
- Ngamjarus C. n4Studies: sample size calculation for an epidemiological study on a smart device. Siriraj Medical Journal, 2016;68(3):160–170.
- Toivonen DA, Tuite MJ, Orwin JF. Acromial structure and tears of the rotator cuff. J Shoulder Elbow Surg. 1995;4(5):376–383. doi:10.1016/S1058-2746(95)80022-0 [CrossRef]
- Kang MN, Rizio L, Prybicien M, Middlemas DA, Blacksin MF. The accuracy of subacromial corticosteroid injections: a comparison of multiple methods. J Shoulder Elbow Surg. 2008;17(1)(suppl):61S–66S. doi:10.1016/j.jse.2007.07.010 [CrossRef]
- Mathews PV, Glousman RE. Accuracy of subacromial injection: anterolateral versus posterior approach. J Shoulder Elbow Surg. 2005;14(2):145–148. doi:10.1016/j.jse.2004.06.012 [CrossRef]
- Beals TC, Harryman DT II, Lazarus MD. Useful boundaries of the subacromial bursa. Arthroscopy. 1998;14(5):465–470. doi:10.1016/S0749-8063(98)70073-8 [CrossRef]
- Sardelli M, Burks RT. Distances to the subacromial bursa from 3 different injection sites as measured arthroscopically. Arthroscopy. 2008;24(9):992–996. doi:10.1016/j.arthro.2008.04.070 [CrossRef]
- Henkus HE, Cobben LP, Coerkamp EG, Nelissen RG, van Arkel ER. The accuracy of subacromial injections: a prospective randomized magnetic resonance imaging study. Arthroscopy. 2006;22(3):277–282. doi:10.1016/j.arthro.2005.12.019 [CrossRef]
- Cook IF, Williamson M, Pond D. Definition of needle length required for intramuscular deltoid injection in elderly adults: an ultrasonographic study. Vaccine. 2006;24(7):937–940. doi:10.1016/j.vaccine.2005.08.098 [CrossRef]
Patient Demographic Data
|Characteristic||All Patients (N=50)||Midlateral Route Patients (n=25)||Posterior Route Patients (n=25)||Pa|
|Sex, No. (%)||.010b|
| Male||22 (44)||6 (24)||16 (64)|
| Female||28 (56)||19 (76)||9 (36)|
|Age, mean±SD (range), y||55.7±7.5 (39–75)||55.3±6.7 (41–68)||56.1±8.4 (39–75)||.730c|
|Weight, mean±SD; median (range), kg||64.8±13.7; 61.5 (42–105)||61.1±13.9; 58.0 (42–105)||68.4±12.8; 68.6 (45–93)||.014d|
|Height, mean±SD (range), cm||160.6±7.8 (143–175)||157.4±6.0 (147–170)||163.9±8.2 (143–175)||.002c|
|Body mass index, mean±SD; median (range), kg/m2||24.94±4.01; 24.97 (19.09–38.56)||24.57±4.61; 24.32 (19.09–38.56)||25.31±3.36; 25.50 (19.81–32.29)||.260d|
|Circumference of proximal humerus, mean±SD (range), cm||35.7±4.3 (28–45)||35.0±4.1 (28–45)||36.3±4.6 (29–45)||.288c|
|Side, No. (%)||.396e|
| Right||26 (52)||11 (44)||15 (60)|
| Left||24 (48)||14 (56)||10 (40)|
|Diagnosis, No. (%)||NA|
| Impingement syndrome||34 (68)||19 (76)||15 (60)|
| Rotator cuff tear||8 (16)||3 (12)||5 (20)|
| Others||8 (16)||3 (12)||5 (20)|
Injection Accuracy and Clinical Outcomes
|Variable||All Patients (N=50)||Midlateral Route Patients (n=25)||Posterior Route Patients (n=25)||P|
|Injection type, No. (%)||.034a|
| Intrabursal||40 (80)||23 (92)||17 (68)|
| Extrabursal||10 (20)||2 (8)||8 (32)|
|Modified UCLA shoulder score|
| Sum preinjection, mean±SD; median (range)||21.66±5.08; 21.00 (8–32)||20.40±4.96; 21.00 (8–32)||22.92±4.97; 21.00 (11–30)||.079b|
| Sum postinjection, mean±SD; median (range)||31.64±1.94; 31.00 (27–35)||31.44±1.80; 31.00 (28–35)||31.84±2.09; 31.00 (27–35)||.473b|
| Delta sum preinjection–postinjection, mean±SD (range)||9.98±4.69 (0–20)||11.04±4.61 (3–20)||8.92±4.61 (0–19)||.223c|
|VAS pain score|
| Preinjection, mean±SD; median (range)||6.5±1.7; 6.5 (3–9)||6.8±1.8; 7.0 (3–9)||6.2±1.7; 6.0 (3–9)||.296b|
| Postinjection, mean±SD; median (range)||2.1±1.3; 2.0 (0–5)||2.2±1.5; 2.0 (0–5)||2.0±1.2; 2.0 (0–5)||.604b|
| Delta preinjection–postin-jection, mean±SD (range)||4.4±1.5 (2–8)||4.6±1.5 (2–8)||4.2±1.6 (2–7)||.540c|
|Acromial type, No. (%)||.202a|
| I||13 (26)||4 (16)||9 (36)|
| II||33 (66)||18 (72)||15 (60)|
| II||4 (8)||3 (12)||1 (4)|
Univariate Analysis (Logistic Regression) of Potential Factors Associated With Accuracy of Injections
|Variable||Crude Odds Ratio||95% Confidence Interval for Odds Ratio||P|
|Circumference of proximal humerus||0.85||0.725–1.017||.077|
|Body mass index||1.05||0.875–1.273||.571|