Treatment of Grade III Acromioclavicular Joint.9 - Orthobullets

CLINICAL ORTHOPAEDICS AND RELATED RESEARCH

Number 455, pp. 38¨C44

? 2006 Lippincott Williams & Wilkins

Treatment of Grade III Acromioclavicular Joint Injuries

A Systematic Review

Edwin E. Spencer, Jr., MD

Downloaded from by BhDMf5ePHKav1zEoum1tQfN4a+kJLhEZgbsIHo4XMi0hCywCX1AWnYQp/IlQrHD3gy8G8VagVy3VhRs9XWxdMaC8r4WaEC1Tfz1UleLHqK/lQ4xR+A2oQw== on 05/08/2018

ter casts and straps.38 There is no evidence any one form

of immobilization is better than another and compliance is

poor among the arduous forms of immobilization.22,24,28,36,43 Many authors have reported good results

with nonoperative treatment, with 80% to 90% patient

satisfaction rates.6,14,33 Still other authors have reported up

to 50% of patients treated nonoperatively will have residual symptoms of pain and weakness.10,30,43

This has prompted several authors to consider operative

stabilization.2,5,7,8,11,12,16,18,20,21,26,27,34,35,39¨C41,44 The first

recorded surgical procedure for an AC joint injury was

reported by Cooper in 1861.9 When Urist38 published his

review of acromioclavicular joint injuries in 1959 there

were at least 30 different methods of surgical repair or

stabilization. These various operative stabilizations have

included everything from AC joint fixation with pins to

coracoclavicular fixation to fascial weaves. Most reports

are a series of patients treated with one form of stabilization and represent Level IV evidence. As we try to elucidate the best method to treat our patients, we look to

evidence-based medicine for answers. This requires prospective randomized controlled studies or metaanalyses of

the literature. Unfortunately, there is a paucity of studies

about the treatment of Grade III acromioclavicular joint

injuries.

Therefore, a systematic review of the literature was

performed to determine whether there was evidence to

suggest Grade III AC separations would best be treated

nonoperatively or operatively.

Although nonoperative treatment is considered the standard

of care for the treatment of Grade I and II acromioclavicular

(AC) joint injuries, the treatment of Grade III injuries is

controversial. There are as many methods of nonoperative

treatment as there are for operative stabilization. Most of the

literature represents Level IV evidence with very few Level

II and III studies upon which to base decisions. A systematic

review of the English-language literature was performed to

determine if Grade III AC joint separations are best treated

operatively or nonoperatively. Based on limited lowevidence, nonoperative treatment was deemed more appropriate than traditional nonoperative treatments because the

results of the latter were not clearly better and were associated with higher complication rates, longer convalescence,

and longer time away from work and sport.

Level of Evidence: Level III, therapeutic study. See the

Guidelines for Authors for a complete description of levels of

evidence.

Acromioclavicular joint injuries have been recognized

since the time of Hippocrates, and the treatment has

ranged from skillful neglect to various forms of operative

stabilization. It is well accepted the preferred treatment of

acute minor injuries and/or separations is nonoperative.1,4,37 Although surgical stabilization is generally

recommended for the more severe separations,1,31 the

treatment of intermediate separations or dislocations

(Grade III1,42 defined as separation of the acromioclavicular joint greater than half of its normal depth with widening of the coracoclavicular relationship) is controversial.

At least 50 forms of nonoperative treatment have been

advocated, from simple slings used for a few days to plas-

MATERIALS AND METHODS

A systematic review of the English-language literature was performed using the Medline and EMBASE databases. The key

words for the search included ¡°Grade III AC joint separations,¡±

¡°acromioclavicular joint injuries,¡± and ¡°acromioclavicular joint

dislocations.¡± Four hundred sixty-nine articles were found. Only

Level II or III studies in which a nonoperatively treated cohort of

patients was compared with a cohort of operatively treated patients were included in the review. The search and review were

conducted by one surgeon (EES).

From the Shoulder and Elbow Service, Knoxville Orthopaedic Clinic, Knoxville, TN.

The author certifies that he has no commercial associations (eg, consultancies, stock ownership, equity interest, patent/licensing arrangements, etc) that

might pose a conflict of interest in connection with the submitted article.

Correspondence to: Edwin E. Spencer, Jr., MD, Shoulder and Elbow

Service, Knoxville Orthopaedic Clinic, 260 Fort Sanders West Boulevard,

Knoxville, TN 37922. Phone: 865-450-1227; Fax: 865-769-4536; E-mail:

spencer9882@.

DOI: 10.1097/BLO.0b013e318030df83

38

Copyright ? Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

Number 455

February 2007

Grade III Acromioclavicular Joint Injury

The review was limited to those studies that used the Tossy37

and Allman1 classification of Grade III AC separations. This was

the most commonly used classification in the literature and defines a Grade III AC separation as a separation of the acromioclavicular joint greater than half of its normal depth with widening of the coracoclavicular relationship.37 This classification

was later expanded by Williams et al42 to include Grades IV, V,

and VI but the definition of Grade III remains the same. In the

Williams and Rockwood classification a Grade III separation is

described as a superior displacement of the clavicle and a 20% to

100% increase in the coracoclavicular interspace.42

Although the literature was replete with Level IV studies

describing various nonoperative and operative methods of treating Grade III AC separations, only those in which the two groups

were directly compared were included because they provide the

highest level of evidence on which to form an opinion. The

initial search with the aforementioned key words resulted in 469

references. Fifty-six of these references met the inclusion criteria

of pertaining to Grade III separations. Only nine studies met the

inclusion criteria of comparing the results of nonoperative treatment of Grade III AC joint separations to various forms of operative stabilization.4,13,19,23,25,28,29,32,36 Three of these studies

were prospective and randomized (Level II evidence), whereas

the others were retrospective analyses (Level III evidence).

These results must be interpreted carefully because they are spe-

TABLE 1.

39

cific to the particular type of operative stabilization and are not

necessarily applicable to other forms of operative stabilization.

Performing a metaanalysis of these studies was not possible

because they used different outcome measures and procedures,

and the patient populations varied.

The results are presented with a review of the prospective

studies followed by the retrospective studies in chronological

order to convey changing thoughts and surgical procedures over

time (Table 1). The correlating various outcome measures and

results are also listed in chronological order (Table 2).

Imatani et al19 conducted a prospective randomized study

comparing nonoperative treatment to two different forms of operative treatment and concluded surgical stabilization was no

better than nonoperative treatment. Inclusion criteria included a

vertical displacement of the clavicle by 50% or more on stress

radiographs37 in patients between the ages of 17 and 40 years.

Thirty patients were identified, but five were lost to followup and

two patients with surgical complications were excluded, leaving

23 patients in the study. There were 12 patients treated nonoperatively with a sling for 3 weeks and 11 treated operatively with

a 12-month minimum followup. Patients were randomized by

alternating treatment options based on time of presentation.

Open reduction and stabilization with Steinman pins placed

across the acromioclavicular joint or stabilization with a coracoclavicular screw were performed. The coracoclavicular liga-

Demographic Data and Description of Treatment

Study

Imatani et al19

Larsen et al23

Bannister et al4

Powers et al28

Number of

Patients

11 Operative

12 Nonoperative

41 Operative

43 Nonoperative

27 Operative

33 Nonoperative

19 Operative

Average Age

(years)

23.5

36

32.5

Not given

28 Nonoperative

Rosenorn et al32

11 Operative

13 Nonoperative

37 (operative)

41.5 (nonoperative)

Galpin et al13

16 Operative

21 Nonoperative

52 Operative

28.9 (operative)

36.7 (nonoperative)

93 were between

18 and 25, but no

average given

Taft et al36

63 Nonoperative

Description of Treatment

Operative: AC pinning or Bosworth

method

Nonoperative: sling ¡Á 3 weeks

Operative: AC pinning

Nonoperative: Sling ¡Á 4 weeks

Operative: Bosworth method

Nonoperative: sling ¡Á 2 weeks

Operative: AC pinning, DCE, and

fascial weave

Nonoperative: 20 via arm cast;

8 not given

Operative: Bosworth method

Nonoperative: bandage, PT,

no treatment

Operative: Bosworth method

Nonoperative: sling and early ROM

Operative: 26 AC pinning and

26 Bosworth method

Nonoperative: 43 sling, 11 splint,

9 taping

MacDonald et al25

12 No treatment

10 Operative

Press et al29

10 Nonoperative

16 Operative

31.7 (nonoperative)

30.7 (operative)

10 Nonoperative

49.6 (nonoperative)

25 (operative)

Operative: 5 AC pinning and

5 Bosworth method

Nonoperative: taping or sling

Operative: 9 Weaver-Dunn and

7 suture CC fixation

Nonoperative: Sling

Followup

12-month minimum

13 months

4 years

12 years

12 months (operative)

84 months (nonoperative)

35 months (operative)

33.7 months (nonoperative)

10.8 years (operative)

9.5 years (nonoperative)

13 years (no treatment)

5.7 months (operative)

6.3 months (nonoperative)

32.3 months (operative)

33.4 months (nonoperative)

AC = acromioclavicular; CC = coracoclavicular; DCE = distal clavicular excision; ROM = range of motion

Copyright ? Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

40

Clinical Orthopaedics

and Related Research

Spencer

TABLE 2.

Outcome Measures, Results, and Level of Evidence

Study

Imatani et al19

Larsen et al23

Bannister et al4

Powers et al28

Rosenorn et al32

Galpin et al13

Taft et al36

MacDonald et al25

Press et al29

Outcome Measure

Custom-designed 100-point scale:

40 points for pain

30 points for function

30 points for movement

Custom-designed 12-point scale:

4 points for pain

4 points for motion

4 points for strength

Same as Imatani et al19

Subjective rating of good, fair, or poor

based on pain, motion and ability to

carry heavy objects

Subjective rating of good, fair, or poor

based on pain and motion

Subjectively asked about pain, strength

and ability to throw

Strength measured by cable tensiometer

Custom-designed 12-point scale:

4 points for pain and stiffness

4 points for ROM and strength

4 points for radiographic appearance

Custom-designed 30-point scale:

24 points, answered by patient;

6 points, answered by examiner

Custom-designed 20-point scale based on

patients response to five questions

Results (number of

patients per rating)

Level of

Evidence

Operative: 5 good/excellent, 1 fair, 5 poor

Nonoperative: 7 good/excellent, 2 fair, 3

poor

II

Operative: 38 good/excellent, 1 poor

Nonoperative: 39 good/excellent, 1 fair

II

Operative: 23 good/excellent, 4 fair/poor

Nonoperative: 33 good/excellent

Operative: 12 good, 2 fair, 4 poor

II

III

Nonoperative: 24 good, 4 poor

Operative: 5 good, 4 fair, 2 poor

Nonoperative: 7 good, 5 fair, 1 poor

Operative: 12 pain free, 9 of 11 throwing =

normal

Nonoperative: 15 pain free, 8 of 12

throwing = normal; both groups had

85% of normal strength

Operative: 9.4

Nonoperative: 8.4

No treatment: 8.2

(average values)

Operative: 24.4

Nonoperative: 23.7

(average values)

Operative: 17

Nonoperative: 15.4 (average values)

III

III

III

III

III

ROM = range of motion

ments were not routinely repaired. Fixation devices were removed at 3 months after the surgical procedure. Patients were

assessed radiographically and clinically at 1 year with a custom

100-point outcome measure that included pain, function, and

motion. In this scoring system 40 points are allocated for pain,

30 points for function (weakness, use of shoulder, change in

occupation), and 30 points for movement (abduction, flexion,

and adduction). The results were classified as excellent (90¨C

100), good (80¨C89), fair (70¨C79), and poor (less than 70). In the

nonoperative group there were seven patients with satisfactory

(good and excellent) results and five with unsatisfactory (fair and

poor) results. In the operative group there were five patients with

satisfactory results and six with unsatisfactory results. Unsatisfactory results were mainly attributable to pain and weakness

with relatively normal range of motion (ROM). The authors

recommended all acute, complete acromioclavicular separations

be treated nonoperatively. Although this was a prospective randomized study, bias was introduced with the patients lost to

followup and those excluded because of failure. In addition, this

study had a relatively low number of patients with a short followup period and used a nonvalidated custom outcome measure.

Larsen et al23 prospectively randomized 84 patients via sealed

envelopes to nonoperative management or modified Phemister

procedure and concluded nonoperative management was better

because functional results were similar, but there were many

complications in the operative group. Nonoperative treatment

(43 patients) consisted of placement in a sling for 4 weeks and

physiotherapy; operative treatment (41 patients) consisted of

open reduction and repair of the acromioclavicular and coracoclavicular ligaments with temporary AC joint pinning (pins removed at 5 to 12 weeks after the procedure). Patients were

assessed at 3 and 13 months radiographically and with a 12-point

custom scoring system that included pain (4 points), motion (4

points) and strength measured via spring weight (4 points). Results were classified as excellent (11¨C12), good (9¨C10), fair (7¨C

8), and poor (less than 7). Results at 3 months were better in the

nonoperative group, but there was no difference at 13 months

with 38 good or excellent results and one poor result in the

nonoperative group, and 39 good or excellent results and one

poor result in the operative group. Radiographically, all but two

of the patients who had surgery maintained the reduction,

whereas all patients in the nonoperative group continued to have

residual displacement that did not increase with weighted views.

Return to work occurred at 8 weeks in the operative group and

6 weeks in the nonoperative group. Ossification of the coracoclavicular ligaments was common in both groups, but did not

Copyright ? Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

Number 455

February 2007

affect the results or ROM. There were six superficial infections

in the operative group and pin migration or breakage in 21 patients. There were two patients in the operative group who underwent a revision to a distal clavicular excision (DCE) because

of pain. There were three patients in the nonoperative group who

had an operative procedure for residual symptoms. This was a

DCE and coracoacromial ligament transfer (Weaver-Dunn).39

The authors concluded most patients can be treated nonoperatively with a shorter rehabilitation time. They also suggested thin

patients and those required to perform heavy or overhead work

be considered for operative stabilization based on subjective

complaints from these patients. They elected to perform a

Weaver-Dunn procedure for the three patients initially treated

nonoperatively instead of the original procedure. The strength of

this study is that it is prospective, randomized, and includes a

large group of patients; however, followup was short and it used

a nonvalidated outcome measure. Bias was introduced as the

three patients in the nonoperative group who were subsequently

treated operatively were excluded and the two patients in the

operative group who had revision surgery were also excluded. It

is worth noting the authors¡¯ operative procedure of choice

seemed to change from acromioclavicular pinning to a WeaverDunn procedure.

In the final prospective study, reported by Bannister et al,4 60

patients were randomized via drawn numbers to nonoperative

treatment (33 patients) or operative stabilization (27 patients)

with the Bosworth7 method. The study concluded nonoperative

treatment was superior except in cases of severe displacement.

Patients were treated nonoperatively with a sling for 2 weeks and

then participated in a rehabilitation program, or were treated

operatively with coracoclavicular screw fixation (coracoclavicular ligaments not repaired) and then engaged in the same rehabilitation program. The screws were removed at 6 weeks. Patients were evaluated at 4 years via the scoring system of Imatani

et al.19 Six patients were lost to followup. At 4 years the results

for the nonoperatively treated group were good or excellent in

100% of patients, whereas the operatively treated group had

good or excellent results in 84% of patients and fair results in

16%. However, in the nonoperative group four patients failed

treatment and underwent surgery for weakness or pain. It is

unclear how these patients could have been rated with a good or

excellent result. There were five complications in the operative

group including loss of reduction and hardware failure. The operative group had a slower recovery and a twofold increase in

time away from work or sport. The authors concluded nonoperative treatment was superior. The strength of the study is that

it is prospective and randomized with intermediate-term followup; however, there were 12 patients who had a displacement

of 2 cm or greater, which is better classified as Grade V separation according to Williams et al.42 The authors stated surgery

yielded a better result in these patients.

In 1974, Powers and Bach28 retrospectively reported on 42

patients with complete AC separations and concluded there was

no advantage to surgical stabilization. This study compared 28

patients treated nonoperatively (20 placed in a body arm cast and

eight placed in a sling) with 19 treated surgically with one of

three different procedures (ligament repair with AC joint pinning

Grade III Acromioclavicular Joint Injury

41

[14 patients], distal clavicle excision [four patients] and a fascial

weave [one patient]). The patients were evaluated clinically at an

average of 12 years with a three-tier outcome measure of good,

fair, or poor based on residual instability, tenderness, and subjective reports of pain and ADL (activities of daily living) function. Based on this simple subject outcome measure, 24 patients

who received nonoperative treatment had outcomes rated as

good, and four had fair outcomes. Two of the four patients with

fair outcomes were subsequently treated with DCE and had good

outcomes. In the operatively treated group, there were nine patients with good results, two with fair results, and three with poor

results. There seemed no difference in outcome between the

methods of nonoperative treatment. The conclusion was nonoperative treatment yielded better results. This is a weak study as

the number of patients was small, groups were not equal, and the

outcome measure was simple and nonvalidated. The authors also

sent a questionnaire in 1974 to the heads of each orthopaedic

program in the United States and found 29% of responders

treated all Grade III separations with surgery, whereas 47% reserved surgery for young active individuals.

In 1974, Rosenorn and Pedersen32 retrospectively compared

11 patients with acromioclavicular dislocations treated surgically

to 13 patients who were treated nonoperatively. They observed

no benefit from surgery. The nonoperative treatment ranged

from bandaging to physiotherapy to no treatment, whereas operative treatment was Bosworth7 coracoclavicular screw fixation. Followup averaged 12 months for the operative group and

84 months for the nonoperative group. The outcome measure

was a custom system rating the patient outcome as good, fair, or

poor based on pain and active range of motion. Radiographic

results were recorded as dislocation or subluxation. Five patients

in the operative group had good results, four had fair results, and

two had poor results, whereas in the nonoperative group there

were seven patients with good results, five with fair results, and

one with a poor result. Only three patients in the operative group

maintained radiographic reduction. Seven of nine manual laborers treated nonoperatively were able to return to heavy work.

Incapacitation lasted 1.5 times longer in the operative group and

the authors felt, overall, nonoperative treatment was superior to

this type of surgical fixation. This is also a weak study because

there no validated outcome measure was used, few patients were

enrolled, and followup was short term.

In 1985 Galpin et al13 retrospectively compared 21 patients

treated nonoperatively for Grade III AC separations with 16

patients treated with Bosworth coracoclavicular fixation and

concluded there was no benefit from surgical intervention. The

patients were followed for approximately 34 months and were

assessed subjectively (patient asked about pain, strength, and

ability to throw) and objectively with a cable tensiometer. Objectively, each group had 85% of their normal strength (using the

opposite side for comparison) except in external rotation, which

was weaker in both groups. Subjectively, there was no difference

in the two groups in pain, strength, and ability to throw (Table 1).

However, there was a threefold increase in time away from work

in the operative group. In addition, there were two patients with

septic complications in the operative group who were excluded

from the study. The authors concluded there was no benefit to

Copyright ? Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

42

Clinical Orthopaedics

and Related Research

Spencer

operative stabilization. Although the study is retrospective and

does not use a validated outcome measure it is one of the best

Level III studies because patients were approximately the same

age, the numbers of patients in each group were similar, and the

exact grade of the separation was accurately defined.

The largest retrospective review, reported by Taft et al,36

concluded nonoperative treatment was superior to two different

methods of operative stabilization. There were 63 patients in the

nonoperative group. Treatment included the use of a sling (43

patients), splint with closed reduction (11 patients), or taping or

casting with closed reduction (nine patients). The operative

group included 52 patients of whom half had an open reduction

and acromioclavicular joint pinning and the other half had an

open reduction and Bosworth7 coracoclavicular screw fixation.

Twelve patients received no treatment. Patients were followed

up for an average of 10 years and were assessed with a custom

12-point scoring system that included measures for pain (4

points), ROM and strength (4 points), and radiographic appearance (4 points). Twelve points was a perfect score. The average

score was 8.4 for the nonoperative treatment group, 9.4 for the

operative treatment group, and 8.2 for the group of patients who

received no treatment. The point values varied mostly in radiographic appearance with higher scores in the operative group

because most were reduced. When the radiographic appearance

was excluded, the results were approximately the same in all

three groups, indicating anatomic reduction was not necessary

for a good result. The complication rates were 52% in the nonoperative group, 100% in the operative group, and 42% and in

the untreated group. Complication rates must be interpreted carefully because the list of complications included arthritis, which

was very common and might have been a function of age given

the length of followup. Four patients treated with taping and

casting had skin necrosis develop. Two of those patients required

skin grafting. However, more complications were reported in the

operative group with four wound infections, 12 pin-related complications including migration and breakage, and bone erosion

around the screw in nine. Three of the 32 patients treated nonoperatively in whom a reduction was attempted had a reduced

joint at followup, thereby emphasizing reduction is not necessary

with nonoperative treatment. However, the development of arthritis seemed to correlate with a reduced joint. Forty-five percent of patients with a nonanatomic reduction developed arthritis, whereas 15% of those in whom an anatomic reduction was

maintained developed arthritis; however, the arthritis was mostly

asymptomatic, as only 10 of the arthritic patients in this study

required a DCE. Those patients in whom a DCE was subsequently performed had a substantial improvement in pain. Although the scoring system used in this study was not a validated

measure, the strength of this study is the long followup of a large

number of patients, making it one of the better Level III studies.

MacDonald et al25 reported on a group of 20 patients and

focused on strength measurements via isokinetic dynamometer.

They concluded nonoperative treatment was superior. Ten patients were treated nonoperatively with a sling or taping, and 10

were treated operatively with coracoclavicular fixation with a

Bosworth screw (five patients), or AC joint pinning (five patients). Patients were evaluated at an average of 6 months after

treatment with a custom-designed 30-point scoring system based

on a questionnaire completed by the patient and the examiner.

Patients were also objectively evaluated and with an isokinetic

dynamometer. There were no subjective differences between the

two groups except patients in the surgical group rated themselves

better in cosmetic appearance (Table 1). Objectively, there were

few differences in strength except in eccentric abduction at fast

speeds, concentric external rotation at slow speeds, and eccentric

external rotation at fast speeds in which the nonoperative group

performed better. The strength differences between the groups

were less pronounced when compared more than 6 months after

treatment, indicating operatively treated patients may have a

longer recovery period. This supports the findings of other authors who observed slower recovery times for operatively treated

patients.23,32 There were four complications in the operative

group, including hardware pullout and infections. The authors

concluded nonoperative treatment is superior to the operative

methods used in this study. The strengths of the study are that

both groups were similar and the objective strength testing was

the most extensive of all the studies; however, the number of

patients was small, the followup was extremely short, and a

nonvalidated outcome measure was used.

The most recent comparative study was published by Press et

al29 and is the only study to report the subjective outcome of

surgical intervention was better than that of nonoperative treatment. This study retrospectively compared 10 patients treated

nonoperatively (sling) with 16 patients treated surgically. There

were two surgical procedures performed: nine Weaver Dunn

procedures39 and six stabilizations with suture coracoclavicular

fixation. Patients were evaluated at 33 months with a customdesigned 20-point scoring system based on the patients¡¯ response

to five questions. In addition, 21 patients had strength assessed

with an isokinetic dynamometer. Although the operative group

took longer to return to work and sport, they subjectively rated

their pain, ROM, functional limitations, cosmesis, and overall

satisfaction as better than the nonoperative group. There were no

substantial differences in strength between the two groups when

expressed as a percentage of the uninvolved extremity, but abduction was 6% to 15% weaker than the uninvolved side in both

groups. The study concluded there was little difference between

operative and nonoperative treatment in terms of objective measures, but subjectively the surgical results are better. The authors

suggested this may be attributable to self-selection on the part of

the patients. This may also be secondary to selection bias on the

part of the surgeon, as the indications for surgery in this study

included patient activity level, requirements of manual labor, and

¡°failure of nonoperative treatment.¡±

RESULTS

The best evidence consists of prospective randomized controlled studies comparing two similar groups of patients

with the same pathology (Level II evidence). The three

such studies in this review all concluded nonoperative

treatment was superior.4,19,23 These conclusions were

based on the findings that the surgical results were no

better and were associated with more complications. In

Copyright ? Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

 ................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download