Prereduction radiographs in clinically evident anterior shoulder dislocation

Prereduction Radiographs in Clinically Evident Anterior Shoulder Dislocation MICHAEL SHUSTER, MD, FRCPC,* RIYAD B. ABU-LABAN, MD, FRCPC,*I" JEFF BOYD, MBBS, CCFP(EM)* The main study objective was to determine if experienced emergency physicians can accurately identify a subgroup of patients with anterior shoulder dislocation for whom prereduction radiographs do not alter patient management. Our prospective study evaluated 97 patients who presented to 2 ski-hill clinics and to our rural emergency department with possible shoulder dislocation between November 1996 and May 1997. Emergency physicians were certain of shoulder dislocation by clinical examination alone in 40 of 59 cases (67.8%) of possible dislocation. All 40 cases were found to have a dislocation (100%; 95% CI, 91.19% to 100%), and the prereduction radiograph did not affect management of the injury. Prereduction radiographs added 29.6 -+ 12.68 minutes to treatment. We conclude that shoulder dislocation is often readily apparent from history and physical examination. When the experienced emergency physician is certain of the diagnosis of anterior shoulder dislocation, prereduction radiography delays treatment and does not alter management. (Am J Emerg ied 1999;17:658-658. Copyright © 1999 by W.B. Saunders Comparty) Anterior shoulder dislocation is a relatively common injury. It is standard teaching that radiographs be performed both before and after reduction of the dislocation. 1-4 Recent retrospective studies have challenged the standard teaching by questioning the need for postreduction radiographs. 5-6 Our prospective study also addresses the issue of duplication of radiographs, but we question the need for prereduction radiographs. There is a rationale for the classic approach to shoulder dislocation. The prereduction radiograph confirms that the shoulder is dislocated (rather than fractured), shows the type of dislocation (anterior or posterior), and documents the presence or absence of a fracture prior to the manipulation of the shoulder. The postreduction radiograph confirms the reduction, and documents the presence or absence of a fracture which may have resulted from the reduction or which could not be seen on the prereduction radiograph. In their 1996 study, Hendey and Kindaw 5 conclude that routine postreduction radiographs are unnecessary because they only rarely influence management of the injury. Many physicians who are experienced in treating shoulder dislocations feel that the argument applies more strongly to

From the *Department of Emergency Medicine, Mineral Springs Hospital, Banff, Alberta, and the l'Department of Emergency Medicine, Vancouver General Hospital and the Division of Emergency Medicine, University of British Columbia, Canada. Manuscript received October 9, 1998, returned November 1, 1998; revision received November 29, 1998, accepted December 2, 1998. Address reprint requests to Dr Shuster, Mineral Springs Hospital, Box 1050, Banff, Alberta, Canada T0L 0C0. Key Words:Dislocation, shoulder dislocation, radiograph. Copyright © 1999 by W.B. Saunders Company 0735-6757/99/1707-0007510.0(3/0

prereduction radiographs. Moreover, they believe that there are good clinical reasons not to perform prereduction radiographs routinely, but rather to confirm the location of the humeral head and the presence or absence of a fracture with postreduction radiographs. Patients with shoulder dislocation are often in considerable pain when they present to a ski hill clinic or to the hospital emergency department. Delaying reduction until a radiograph can be performed prolongs the time that the patient must endure pain, and may worsen muscle spasm and result in a more difficult reduction. Consequently, in our area, physicians who see patients at a ski hill clinic and who are confident in the diagnosis of shoulder dislocation perform the reduction, then send the patient to hospital for a postreduction radiograph. The postreduction radiograph confirms the successful reduction and reveals a fracture, if any, so that appropriate information and advice for follow-up can be given to the patient. It is the primary hypothesis of our study that experienced emergency physicians can accurately identify a subgroup of patients with shoulder dislocation for whom prereduction radiographs do not alter patient management. It is our secondary hypothesis that prereduction radiographs, when obtained on the identified subgroup, significantly delay time to reduction. An additional secondary hypothesis is that experienced physicians (ski area physicians who encounter a high volume of shoulder dislocations) can identify and successfully reduce shoulder dislocation when prereduction radiographs are not immediately available. MATERIALS AND METHODS Mineral Springs Hospital is a rural community hospital with an active emergency department staffed by full-time specialty-trained emergency physicians who see 13,000 patients annually. The hospital is located in the town of Banff, a mountain community of 7,100 inside Canada's oldest national park. The park attracts people of all ages to the many recreational activities which are available. Recreational activities include skating, alpine and Nordic skiing, backcountry skiing, snowboarding, ice climbing, rock climbing, hiking, mountain biking, fishing, canoeing, and kayaking. The 3 ski hill areas which operate within the park report more than 750,000 skier/snowboarder days per year. Two ski hills have basic first aid clinics without radiographic capability and are approximately 60 minutes from the hospital. Injuries from the third ski hill, which is 20 minutes away, are sent directly to the hospital. In the winter season prior to our study, more than 75 shoulder dislocations were treated in the Mineral Springs Hospital emergency department. A number of the dislocations were diagnosed and reduced at the ski hill clinics, whereas others were sent from the ski hill to the hospital for reduction, and still others first presented at Mineral Springs Hospital. 653

654

AMERICAN JOURNAL OF EMERGENCY MEDICINE • Volume 17, Number 7 • November 1999

Time1

Time2

Patient to radiology, pre-reduction radiograph, patient returns to ED

MD exam

Triage nurse first encounters patient

Time3

MD orders radiograph

Time4

Shoulder reduced, post-reduction radiograph, post-analgesic/sedation recovery

MD ready to begin reduction

MD orders discharge

FIGURE. Definitions of Times 1-4. For our study, we enlisted the cooperation of the ski hill clinics as well as the hospital emergency department staff. Patients were considered eligible for our study if, after triage assessment, the triage nurse at the hospital or the nurse at the ski hill clinic had any suspicion of a shoulder dislocation. For patients presenting to the hospital, a data collection form was initiated by the triage nurse at the time of initial patient contact and completed by the attending physician at the time of treatment. In cases where the patient was examined and treated on the ski hill, the physician at the ski hill initiated a data collection form. Following ski hill treatment, patients were sent to the hospital for radiographs. The emergency physician then completed the postradiograph section of the data collection form. Nurses and physicians were fully informed of the nature of the study. All subjects gave valid informed consent prior to inclusion in the study. The study protocol required that patients who presented to the emergency department with possible dislocation be radiographed (unless refused), but that no patient was to have a radiograph until the attending physician had performed an examination and completed the prerednction portion of the study data collection form. The data collection form included questions on the patient's past history, circumstances of injury, discomfort, and range of motion. Prior to obtaining prereduction radiographs, physicians were asked to rate their confidence in the diagnosis of anterior shoulder dislocation on a 10 point visual analogue scale (VAS) anchored by the terms "none" at 0 and "certain" at 10. No guidance was provided to the physicians in rating their confidence. Reduction attempts, including technique, use of analgesia, difficulty, and whether a "clunk" was felt, were then documented, and physicians were asked to rate their confidence in a successful reduction on the same VAS prior to obtaining postreduction radiographs. Physicians were also asked if care could have been provided without a prereduction radiograph. Time 1 was defined as the time the triage nurse made contact with the patient. Time 2 is the time following MD examination when the radiograph was ordered (if applicable). Time 3 is the time following radiograph when the MD was ready to begin the reduction. In some cases an intravenous (IV) was established prior to the MD being ready to reduce the dislocation, and thus prior to Time 3; in other cases an IV was established after Time 3. Time 4 is the time the patient was judged ready to leave the department and the discharge order was written (Figure). All times were recorded by the treating nurse as they occurred. When patients were treated at the ski hill, the physician performed the reduction without taking a prereduction radiograph, and recorded Times 1, 3, and 4. Patients were then directed to the emergency department for postreduction radiographs. The data were entered into a Microsoft Excel (Version 5.0a for PC, Microsoft Corp., Redmond, WA) database and transferred to

SPSS (Version 6.1 for Macintosh, SPSS, Inc., Chicago, 1L) for analysis. SPSS was used for descriptive statistics (mean, standard deviations, medians, range). Exact 95% confidence intervals for proportions, tests of differences between proportions, and T-tests were calculated using Stata (Version 5.0 for PC, STATA Corp., College Station, TX) after normality distribution assessments were performed. P -< = 0.05 was considered statistically significant. The study was approved by the hospital committee responsible for research and ethics.

RESULTS The study comprised 97 patients (82 men and 15 women) who had a mean age of 33.49 years (SD + 13.95), a median age of 28 years, and an age range of 16 to 75 years (Table 1). The majority of the injuries occurred while downhill skiing (64%), but a substantial number occurred while snowboarding (26%). Of the 59 patients identified by the triage nurse as possibly having a dislocated shoulder, 49 patients (83.1%) actually had a dislocated shoulder (48 anterior, I posterior). Every patient with a discharge diagnosis of dislocated shoulder had been correctly identified by the triage nurse as a possible dislocation. Data form completion was 100%. The physician was certain (confidence rating of 10) that a patient had a dislocation in 69 of the 97 patients (71.1%). All 69 of the patients in whom the physician was certain of dislocation (100%; 95% CI, 94.79%-100%) were confirmed to have a dislocation, 36 by radiological confirmation and 33 by clinical confirmation (reduction "clunk," restored range of motion, and change in shoulder contour). In 9 o f the 97 patients (9.3%) who had been identified as having a possible dislocation, the physician was certain that the shoulder was not dislocated (confidence rating of 0): all 9 of these patients TABLE 1. Patient Characteristics

Total Patients Refused prereduction xray Male:Female Mean age (years) Median age (years) Downhill Skiing Snowboarding Other

Hospital

Ski Hill

Not Dislocated

49 4 40:9 35.3 31 30 10 9

36 N/A 30:6 31.4 27 24 12 0

12 N/A 10:2 32 26 8 3 1

SHUSTER ET AL • PREREDUCTION RADIOGRAPHS

655

TABLE 2. Confidence in Diagnosis (scale 0 to 10, 10 highest) Confidence

Hospital Dislocated

10 (certain) 9 8 7 6 5 (moderate) 4 3 2 1 0 (none) radiograph deemed necessary

Ski Hill

Not Dislocated

40 6 1 2

1

Dislocated

Not Dislocated

29 2 1 2

DISCUSSION

1

1

7

0/49

9/1 o

2

2/36

2/2

(100%; 95% CI, 66.38%-100%) were confirmed not to have a shoulder dislocation. In the remaining 19 patients (19.6%) the diagnosis of dislocation was not certain: 16 of these patients did have a dislocation, 3 patients did not (Table 2). Mean time to commencing reduction (Time 3 minus Time 1) was 26 minutes shorter when the reduction was performed at the ski hill clinic (no prereduction radiograph) than when it was performed in the emergency department (prereduction radiograph in 44 of 49 patients). Mean time in the treatment facility averaged 59.4 minutes less when the dislocation was reduced in the clinic: 69.5 minutes less when the patient received analgesic and 26.6 minutes less when no analgesic was given. Patients who received analgesics (n = 33) had a mean emergency department time 60.4 minutes longer than those who did not receive analgesics (n = 16) (Table 3). In 74 (87.1%) of the reduction attempts, a "clunk" was felt by the treating physician on reduction (Table 4). In 83 of the 85 (97.7%) reduction attempts, the physician was confident (VAS rating of 7 to 9) or certain (VAS rating of 10) that the reduction was successful, and all of these reduction attempts were proven successful on radiograph. In both cases where the confidence in reduction was rated as "none" (VAS rating of 0), low (VAS rating of 1 to 3), or unsure (VAS TABLE 3.

rating of 4 to 6), the shoulder was shown on radiograph not to have been reduced (Table 5). Fractures were found on 7 of 45 (15.6%) prereduction radiographs and 11 of 63 (17.5%) postreduction radiographs (Table 6). No fracture was identified on a prereduction radiograph that was not also clearly seen on the postreduction radiograph, and no fracture required surgical intervention.

Standard teaching regarding the need for postreduction radiographs was recently questioned in a retrospective study by Hendey and Kindaw. 5 They concluded that postreduction radiographical findings rarely changed treatment. Hendey and Kindaw's study followed one by Harvey et al 6 with a similar conclusion. Neither study questioned the need for prereduction radiographs. In our study, 17% of patients had a fracture in addition to shoulder dislocation. Because some fractures require operative intervention and others affect the course of rehabilitation, a radiograph is required sometime in the course of treatment, even if the radiograph is not required to assess dislocation or to document reduction. Despite the suggestion of previous studies that postreduction radiographs can be omitted, an informal survey of emergency physicians found that few are comfortable with omitting postreduction radiographs. The physicians we spoke with were concerned that a fracture, not evident on the prereduction radiograph, would be missed by not doing a postreduction radiograph. This, the physicians felt, would alter the advice they gave to patients on how to care for their shoulder and on what to expect in terms of rehabilitation and future function. The physicians were also concemed about the missed failed reductions that were identified only by postreduction radiographs in both of the previous studies. 5,6 Although classic teaching in the care of shoulder dislocation is to obtain a prereduction radiograph, 1-4 it is not uncommon in the setting of ski hills and other nonhospital locales for shoulder reductions to be performed without a prereduction radiograph. 7 In our own emergency department some physicians choose not to do a prereduction radiograph when the diagnosis of dislocation is certain, the patient is in pain, and a radiograph is not immediately available. One of the questions that we wished to answer was

Comparison of Treatment Delay Between Facilities and With and Without Premedication Hospital (minutes)

Time to commencing reduction Mean _+ SD Median Delay to reduction attributable to prereduction radiograph Mean _+ SD Median Time in treatment facility All patients--Mean _+ SD reduced with premedication Mean _+ SD reduced without premedication Mean _+ SD NOTE: *P < 0.0001; **P < 0.0127.

38.21 _+ 15.62 37.00 29,6 + 13.05 29.5 99,6 _+ 41.1

Ski Hill (minutes) 12.00 + 14,52 5.00

P <0.0001 n/a

n/a n/a 40.4 _+ 21.7

<0.0001

119.5 _+ 31.8" (n = 33)

50 _+ 18,4"* (n = 17)

<0.0001

59.1 _+ 25.3* (n = 16)

32.5 _+ 21.5"* (n = 19)

<0.0024

656

AMERICAN JOURNAL OF EMERGENCY MEDICINE • Volume 17, Number 7 • November 1999

TABLE 4.

Reduction Characteristics

Reduction Difficulty (Scale 0 to 10, 10 highest) All patients Mean + SD Median Patients who received analgesia Mean + SD Patients who received no analgesia Mean - SD Reduction "clunk" felt Reduction "clunk" not felt

Hospital

Ski Hill

P

4.1 _+ 2.34 3

4.4 __ 2.51 4

NS

4.58 _+ 2.39*

6.36 + 2.37**

0.0172

3.25 -+ 2.02* 46 3

3.32 _+ 2.16"* 28 8

NS 0.0289

NOTE: *P = 0.061; **P = <0.001.

whether experienced emergency physicians can reliably detect anterior shoulder dislocation. Although the attempt is unlikely to cause damage, we do not believe that it is it is optimal care to try to reduce a shoulder that is not in fact dislocated, but is fractured or merely contused. In our study, the physicians were able to ascertain reliably (by history and clinical exam alone) that 40 of 59 emergency department patients (67.8%; 95% CI, 54.4%-79.4%) had a shoulder dislocation. The physicians reported that they would be comfortable performing a reduction without radiograph in each of these 40 cases. In every case where the emergency physician rated the diagnosis of dislocation as certain, prereduction radiograph proved that the shoulder was indeed dislocated. In one of the 40 cases of dislocation, the physician was confident that the shoulder was posteriorly dislocated, and it was. When physician confidence in a diagnosis of shoulder dislocation was rated as high (VAS rating of 7 to 9), one of 10 patients (10%) did not have a shoulder dislocation. The erroneous diagnosis was made in a patient who had a completely displaced fracture of the surgical neck of the humerus. In this case, the initial examination was cursory, and the physician failed to view the radiograph prior to attempting reduction. The correct diagnosis was apparent to the physician as soon as he lifted the arm to begin the reduction. This anomalous case demonstrates the limitation of both radiographs and physical examination: whereas either may be sufficient to make a diagnosis, both require an attentive approach by the physician in order to be useful. Our study protocol required the treating physician to complete a section of the data form at each stage in the TABLE 5.

MD Conference in Reduction Success

Confidence in reduction (Scale 0 to 10, 10 highest) 10---certain 9 8 7 6 5 4 3 2 1 0

Hospital

Ski Hill

Not Reduced

39 9 1

25 5 3 1

0 0 0 0

1

1

1

1

treatment. Prior to obtaining a prereduction radiograph, the emergency physician was asked if a prereduction radiograph was necessary to the medical care of the patient. The emergency physicians reported that reduction could have been performed without the prereduction film in 45 cases of shoulder dislocation where radiograph was obtained and in 4 cases where the patient refused radiograph. Forty-four of 45 radiographs (97.8%; 95% CI, 88.2%-99.9%) would have been avoided had the emergency physicians relied on clinical judgment and their certainty or high degree of confidence in the diagnosis. One of the 45 prereduction radiographs would have been obtained in any case, as it was ordered by the ski hill physician who first saw the patient. There were a total of 40 cases (36 patients at the ski hill, 4 patients who refused a radiograph in the emergency department) where a shoulder reduction was attempted without a prereduction radiograph. In 2 of these cases, a prereduction radiograph was rated "desirable" by the treating physician: one because fracture was suspected (no fracture found on postreduction radiograph), and the second because the physician had a low confidence of dislocation. This second patient was sent from the ski hill clinic to the emergency department for radiograph, which revealed dislocation of the shoulder and fracture of the greater tuberosity. When confidence in the diagnosis of dislocation was not certain or high, treating physicians felt that a prereduction radiograph would be preferred. When the confidence in shoulder dislocation was none (rating of 0), low (rating of 1 to 3), or moderate (rating of 4 to 6), the shoulder was usually found on radiograph not to be dislocated (11 of 13 = 84.6%; 95% CI, 54.6%-98.1%). The patients without dislocations had acromioclavicular (AC) separation or fracture of the distal clavicle. When no prereduction radiograph was performed (ski hill clinic), mean time to reduction was 12 minutes after presentation. In the emergency department, there was a mean delay to reduction of 29.6 minutes between the time the radiograph was ordered (and the reduction could have been performed) and the time the reduction was actually begun (when the patient had returned from radiology and the physician was ready to begin the reduction). We did not determine the total delay to treatment as we could not reliably establish the time interval from ski hill (where most of the dislocations occurred) to hospital. Since transport from the ski hill to the hospital typically takes more than 60 minutes, time to treatment was more than 100 minutes for patients who did not receive definitive care at the ski hill.

SHUSTER ET AL • PREREDUCTION RADIOGRAPHS

657

TABLE6. Incidence of Fracture Seen on Radiograph

Pre-reduction fracture* Post-reduction fracture*

Hospital

Ski Hill

P

7/45 7/45

na 4/18

NS

NOTE: *Greater tuberosity fracture, Hills-Sachs deformity, Bankart deformity.

Physicians typically gave patients a choice of analgesia or no analgesia prior to a trial of reduction. In 41.2% of cases, reduction was accomplished without medication. Physicians reported that reduction was accomplished with less difficulty in patients who did not require medication (mean difficulty rated at 3.25 and 3.32 v 4.58 and 6.36 for emergency department and ski hills respectively by group). The physicians felt that the self-selected group of patients who did not receive medication were better able to relax, which facilitated the reduction. Also, the difficulty in reduction in the analgesic group may be overstated, as patients who had an initial attempt at reduction without analgesia, but who went on to receive analgesia, were recorded as having received analgesia. Patients requiring medication in the emergency department most commonly received a combination of midazolam and fentanyl, while the ski hill clinics most commonly administered either midazolam or diazepam. Patients who did not receive medication spent significantly less time in the facility than those who did (mean 60.4 minutes less at hospital, mean 17.5 minutes less at clinic). In our study, physicians were accurately able to determine clinically whether their treatment had been successful. We asked physicians to rate their confidence in the reduction, and in 64 of the 85 cases of dislocation (75.3%; 95% CI, 64.7%-84%), the treating physician was certain of successful reduction (rating of 10). In all cases where confidence in reduction was rated as high (VAS rating of 7 to 9), the shoulder was confirmed on radiograph to be reduced. In 2 cases, the confidence in reduction was not high (VAS rating less than 7), the shoulder was confirmed on radiograph still to be dislocated. On the basis of this study we believe that postreduction radiographs are not necessary to prove reduction of shoulder dislocation in patients in whom the experienced emergency physician has high confidence that the dislocation has been reduced. No postreduction radiograph changed treatment except, in the case of a fracture of the greater tuberosity, to confirm that the displaced fracture fragment had also been reduced by the joint reduction and thus did not require surgical intervention. While postreduction radiographs may not be necessary to prove reduction of dislocation when there is high confidence in reduction, postreduction radiographs do serve to reveal fractures, and thus to identify the need for additional treatment or changes to the rehabilitation regimen. Eighteen ski hill patients did not receive postreduction radiograph in Banff and were lost to follow-up. Some of those patients stated their intention not to seek follow-up, whereas others undoubtedly waited until their return home (Europe, Australia, other parts of North America) to seek further medical attention. The actual number of fractures may therefore be underreported. While we are not aware of any, adverse outcomes cannot be excluded.

The interpretation of our data is limited by the unavailability of prereduction radiography for the 36 ski hill reductions, and the lack of patient follow-up in 18 of those cases. With no prereduction radiograph it could be argued that some ski hill patients may not have actually had a shoulder dislocation. We feel it is exceedingly unlikely, however, that an experienced physician could be mistaken first in thinking that there is a shoulder dislocation based on initial examination, and then in finding a reduction "clunk," relief of pain, and normalization of anatomy and range of motion in a patient who did not truly have a dislocation. It is well established that Hill-Sachs lesions, Bankart lesions, and greater tuberosity fractures commonly occur in association with shoulder dislocation. 3,4 None of these fractures alters the need for prompt reduction as initial management. Humeral neck fracture in conjunction with shoulder dislocation is a rare combination which results from a high energy mechanism of injury. A published series of 7 cases cautions against closed reduction of the shoulder dislocation in such cases, citing fracture displacement and resulting avascular necrosis as a common complication. 8 Prereduction radiography may thus be prudent in patients who present with possible dislocation from a high-energy mechanism, where associated humeral neck fracture is considered a possibility. A theoretical disadvantage of performing reduction without radiography is legal rather than clinical: when a fracture is found on the postreduction radiograph, an allegation could be made that the fracture was caused by the reduction. However, in a Medline search of the literature from 1966 to present, we were unable to find a single report of a fracture caused by reduction of a shoulder dislocation. It may be that these complications are not being reported; however, we believe it is more likely that the most commonly used reduction techniques do not involve leverage and thus rarely, if ever, cause fractures. It remains necessary for physicians to individually determine how to apply our findings in the context of their own patient populations, level of experience, and circumstances. Our patient population was heavily weighted towards young, athletic individuals, and the physicians taking part in our study were highly experienced with shoulder dislocation. Shoulder dislocation is usually readily apparent from history and physical examination. We conclude that when an experienced emergency physician is certain of the diagnosis of anterior shoulder dislocation, prereduction radiographs delay treatment and do not alter management. Selective ordering of prereduction radiographs in anterior shoulder dislocation will improve patient care and will reduce costs. The authors would like to acknowledge Brian Rowe, MD, for his review of the initial study protocol; Charles Gauthier, MD, Ronald Oshry, MB, Lance Shepherd, MD, Chris Turner, MD, for their contributions to the design of the data collection instrument and for their clinical contribution to this study; Anita Battrum, RN and the rest of the emergency staff who made the study possible; Jeff Quon, DC, for sharing his statistical expertise.

REFERENCES 1. Roberts J, Hedges J: Clinical Procedures in Emergency Medicine, ed 3. Philadelphia, Saunders, 1998, pp 820-829

658

AMERICAN JOURNAL OF EMERGENCY MEDICINE • Volume 17, Number 7 • November 1999

2. Linscott MS: Joint dislocations, in Harwood-Nuss A, Linden C, Luten R, et al (eds): The Clinical Practice of Emergency Medicine, ed 1. Philadelphia, JB Lippincott, 1991, pp 382-390 3. Simon RR, Koenigsknecht SJ: Orthopedics in Emergency Medicine, ed 3. New York, Appleton-Century-Crofts, 1995, pp 392-400 4. Daya M: Shoulder, in Rosen P, Danzel DF, Hockberger RS, et al (eds): Emergency Medicine, ed 4. St. Louis, CV Mosby, 1998, pp 709-739 5. Hendey GW, Kinlaw K: Clinically significant abnormalities in post-reduction radiographs after anterior shoulder dislocation. Ann Emerg Med 1996;28:399-402

6. Harvey RA, Trabulsy ME, Roe L: Are postreduction anteroposterior and scapular Y views useful in anterior shoulder dislocations? Am J Emerg Med 1992;10:149-151 7. Macnab AJ, Cadman R, MacPhail I, et al: Ski-injury research and safety initiatives in British Columbia: A review. BC Medical Journal 1997;39:320-323 8. Hersche O, Gerber C: latrogenic displacement of fracturedislocations of the shoulder. A report of seven cases. J Bone Joint Surg Br 1994;76:30-33