Ccj

E

TlwtmtsmgccsatsavNctsod

IcsmRtMio

F

R

C

1d

3

linical and radiologic outcomes of surgical andonservative treatment of type III acromioclavicularoint injury

milio Calvo, MD, PhD, Mariano López-Franco, MD, and Ignacio M. Arribas, MD, Madrid, Spain

ii

mcpHtpt

isi

M

AwwlrTtc

pmqt

fi4tdsr

awvdtt

(

he management of acute acromioclavicular joint dis-ocations is controversial. The purpose of this studyas to compare the incidence of posttraumatic ana-

omic alterations after surgical or conservative treat-ent of type III injuries and to analyze their effect on

he outcome. Forty-three patients were evaluated retro-pectively, clinically and radiographically, at a 12-onth minimum follow-up. Thirty-two were treated sur-ically, using the Phemister technique, and 11 hadonservative treatment. A comparison of the overalllinical results in both groups showed no statisticallyignificant differences. The acromioclavicular joint wasnatomically reduced in only half of the surgical pa-

ients. Those shoulders treated surgically showed aignificantly higher incidence of osteoarthritis and cor-coclavicular ligament ossification. Differences in cla-icular deformity or osteolysis were not significant.one of these abnormalities had any influence on the

linical result. Because operative and conservativereatments achieve equally good clinical results andurgery carries a higher risk of osteoarthritis, we rec-mmend managing this injury conservatively. (J Shoul-er Elbow Surg 2006;15:300-305.)

t is generally accepted that types I and II acromio-lavicular joint (ACJ) injuries should be treated con-ervatively, whereas operative treatment is recom-ended for type IV, V, and VI injuries, according toockwood et al’s classification.20 The management ofype III ACJ injuries is controversial, however.2,3,11,13

any authors advocate a surgical approach, especiallyn young active individuals, because of the potential riskf shoulder deformity and poor function due to instabil-

rom the Department of Orthopaedic Surgery. Fundación JiménezDíaz, Universidad Autónoma.

eprint requests: Dr Emilio Calvo, Department of OrthopaedicSurgery, Fundación Jiménez Díaz, Avda. Reyes Católicos, 2,28040 Madrid, Spain (E-mail: ecalvo@fjd.es).opyright © 2006 by Journal of Shoulder and Elbow SurgeryBoard of Trustees.

058-2746/2006/$32.00

4oi:10.1016/j.jse.2005.10.006

00

ty.1,9-11 Others propose nonsurgical treatment becauset yields excellent and painless shoulder function.6,13,17

The clinical results of conservative and surgical treat-ent of type III ACJ injuries are widely reported, butomparative studies have failed to demonstrate the su-eriority of one treatment over the others.1,6,7,9,14,17

owever, the incidence of complications of surgicalreatment that might compromise shoulder function com-ared with conservative treatment and the influence on

he outcome have not been addressed.The purposes of this study were to compare the

ncidence of posttraumatic anatomic alterations afterurgical or conservative treatment of type III ACJnjuries and analyze their effect on the clinical result.

ATERIALS AND METHODS

We report the results of a retrospective study of type IIICJ injuries treated acutely from 1983 to 2002. Patientsere included in the study if they were18 years or older,ere treated acutely for type III ACJ injuries, followed for at

east 12 months after the injury, completed the clinical andadiologic examination, and gave their informed consent.he exclusion criteria were previous shoulder symptoms,reatment for chronic type III ACJ injuries, or any pathologicondition except type III ACJ injury in the opposite shoulder.

The medical record was reviewed retrospectively for theresenting history and radiographs and to document treat-ent details. Follow-up was performed by means of auestionnaire and a physical and radiographic examina-

ion of the shoulders.Sixty-eight cases were retrieved from our institution’s

les. Fifty-seven patients could be contacted by phone, and8 agreed to be reviewed. Six patients were dropped from

he study because of insufficient follow-up or radiographicata. The final study group consisted of 42 patients (43houlders). All gave informed consent, and an institutionaleview board approved the study.

Thirty-two shoulders had undergone surgical treatment,nd 11 had been managed conservatively. In 1 patient,ho was treated conservatively, both shoulders were in-olved at different times. Surgical treatment was performeduring the initial part of the index period, and cases later in

he index period were treated non-surgically. There was,herefore, no preselection for each type of treatment.

The patients were evaluated at a mean 122.8 monthsrange, 12-228 months) in the surgical repair group and at

0.5 months (range, 12-108 months) in the conservative

ty6rmma

dtarwotrfitirm

pattAw

taradtada

orwssswtctp

atwsumh

ttn

tsrodbo

gcecww

Opthmnt

Mtpocttoptfc

R

euseTvclthwe

bpjsi

a

J Shoulder Elbow Surg Calvo, López-Franco, and Arribas 301Volume 15, Number 3

reatment group. The mean age of the patients was 39.6ears (range, 18-68 years) and 34.5 years (range, 18 to3) in the surgically and conservatively treated groups,espectively. The surgically treated group consisted of 27en and 5 women, whereas all nonsurgical patients wereen. Both groups had similar levels of physical or sportingctivity.

A modified Phemister procedure was done on the shoul-ers that underwent surgery.12 The ACJ, the lateral end of

he clavicle, and the coracoid were exposed through annterior approach. The articular disc was resected, the ACJeduced, and the bones were transfixed with two Kirschnerires that crossed the joint space. The ends of the acromi-clavicular and coracoclavicular ligaments were then su-

ured and any obvious muscle ruptures repaired. Once theeduction and the pin position were radiographically con-rmed to be accurate, the lateral end of the pins was bento prevent migration. Postoperatively, the shoulder wasmmobilized in a sling. After 6 to 8 weeks, the wires wereemoved, and progressive passive and active shoulderotion was initiated.In the conservatively treated group, the shoulder was

laced in a sling. During the first days, an ice pack waspplied, and some mild analgesics were given to the pa-

ient. We recommended that the arm should be used belowhe shoulder level as soon as the symptoms disappeared.fter 2 weeks, the sling was removed and shoulder motionas progressively instituted as symptoms permitted.During the postoperative evaluation, the clinical func-

ional result was determined by using the score of Imatani etl,8 which was specifically designed to evaluate ACJ inju-ies. This scale assigns 40 points for pain, 30 for functionnd strength, and 30 for motion. A score of 90 to 100 waseemed excellent; 80 to 89, good; 70 to 79, fair; and less

han 70, poor. All patients underwent a detailed postoper-tive physical exam for keloid or unsightly scars, shouldereformities, ACJ pain on palpation or with the cross-chestdduction maneuver, and clavicular instability.

For the postoperative radiographic study, radiographsf both shoulders were obtained in each patient, includingoutine anteroposterior and anteroposterior stress viewsith an 8-kg weight suspended from each arm with wrist

traps, as well as a 20° cephalic tilt view of the ACJ. Totudy the effect of each type of treatment (surgical or con-ervative) on ACJ anatomy, 4 pathologic characteristicsere analyzed: the degree of reduction finally obtained,

he presence of ACJ osteoarthritis, ossification of the cora-oclavicular ligaments, and osteolysis of the lateral end ofhe clavicle. Semiquantitative scales were used for thisurpose.

The degree of the reduction of the ACJ was evaluatedccording to a modified Rosenørm and Pedersen classifica-

ion.15 The joint was considered to be reduced when thereas no displacement compared with the uninvolved side,

ubluxed when there was a moderate (less than 50%)pward displacement of the clavicle relative to the acro-ion, and dislocated if the displacement was greater thanalf of the height of the ACJ.

The degree of osteoarthritis for each shoulder was de-ermined from conventional radiographic signs of osteoar-hritis. Shoulders were considered normal when there were

o signs of osteoarthritis. Mild osteoarthritis indicated that c

here was narrowing of the joint space and subchondralclerosis; moderate osteoarthritis, that the subchondral scle-osis was evident and osteophytes were present; and severesteoarthritis, that the joint was badly deformed. The inci-ence of osteoarthritis in the injured ACJ was comparedetween the two types of treatment and also with thepposite healthy shoulder in each group.

Ossification of the coracoclavicular ligaments was cate-orized into three degrees: mild when there were thinalcifications around the coracoclavicular ligaments, mod-rate if large amounts of radiopaque material around theoracoclavicular ligaments were appreciated, and severehen the calcification of the coracoclavicular ligamentsas complete.Three degrees of clavicular osteolysis were also defined.

steolysis was considered mild if the clavicular end wasoorly defined or showed scattered areas of demineraliza-

ion. If demineralization was evident, but extended less thatalf of the clavicular end, the osteolysis was rated asoderate. A severe osteolysis implied clearly visible demi-eralization involving more than half of the clavicular ex-remity.

Postoperative functional scores were compared with theann-Whitney test for statistical analysis. The Fisher exact

est was used to compare frequency distributions of theosttraumatic anatomic alterations between the two groupsf treatment. Differences with a two-tailed P � .05 wereonsidered significant. The strength of the association be-ween the degree of reduction obtained with surgery andhe incidence of osteoarthritis was estimated by calculationf the � concordance index. The Kruskal-Wallis test waserformed to evaluate the effect of the posttraumatic ana-

omic alterations on the clinical result. Analyses were per-ormed with SPSS 9.0 statistical software (SPSS Inc, Chi-ago, IL) for Windows (Microsoft, Redmond, WA).

ESULTS

We could not find statistically significant differ-nces between the two groups in the clinical results bysing the Imatani scale8 (P � .71) (Table I). Oneurgical patient was rated as poor because of mod-rate pain at the ACJ when the shoulder was used.he patient underwent a resection of the lateral cla-icular end, with surgical reconstruction of the cora-oclavicular ligaments by using the acromioclavicularigament, with a favorable outcome. Two patientsreated nonsurgically were rated fair because theyad moderate daily pain or thought their shoulderas unstable, but these conditions were not severenough to consider surgical repair.

Although all nonsurgical shoulders showed aump due to the dislocation of the ACJ, only twoatients were dissatisfied with the deformity. No ma-

or complications occurred, but the number of un-ightly scars, problems with the fixation device, ornfection was higher after surgical treatment.

Data on the results of the radiographic evaluationre presented in Table II. After surgery, the ACJ was

ompletely reduced to its anatomic position in only

1dttos(distpprocws(

aota0

ldcsgt1sw

b

sc

D

cstts

dtuAtgdoti

vmfdiswdhcthd

acptblaadrhtswjbnp

T

I

C

A*†

Dw

302 Calvo, López-Franco, and Arribas J Shoulder Elbow SurgMay/June 2006

6 shoulders (50%). Obviously, the ACJ remainedislocated in all shoulders treated conservatively. At

he latest follow-up, radiographic signs of osteoarthri-is of the ACJ could be detected in 26 cases treatedperatively (81.2%), whereas the opposite shoulderhowed signs of osteoarthritis in 13 cases (37.5%)Figure 1). Interestingly, statistical analysis of dataemonstrated significant differences both when the

ncidence of radiographic osteoarthritis in the injuredhoulder in the two groups was compared and whenhe incidence of osteoarthritis in both shoulders ofatients undergoing surgical treatment was com-ared (P � .002). Moreover, differences did noteach statistical significance when the incidence ofsteoarthritis in the healthy shoulder in surgically andonservatively managed patients was compared orhen differences between the injured and the oppo-

ite ACJ in the conservatively treated were calculatedP � 1.00).

Despite the high incidence of these radiographicbnormalities, none of these patients required anyther treatment for osteoarthritis. The association be-

ween the degree of reduction achieved with surgerynd osteoarthritis was poor (� concordance index �.022).Posttraumatic ossification of the coracoclavicular

igaments was evident in 19 surgically treated shoul-ers (59.4%) (Figure 2), but only 2 shoulders treatedonservatively showed this alteration. The statisticaltudy also found significant differences between the 2roups evaluated with respect to ligament ossifica-

ion. The lateral clavicular end showed osteolysis in8 shoulders (56.2%) in the surgical group and in 6houlders treated conservatively. These differencesere not statistically significant at comparison.We could not establish any statistical association

able I Clinical evaluation after acromioclavicular dislocation

Clinical results*Operated

groupNon-operated

group

matani score† 93.7 � 9.9 94.1 � 12Excellent 21 9Good 10 0Fair 0 2Poor 1 0omplications*ACJ deformity 3 11Unattractive scar or keloid 7 0Superficial infection 1 0Pin migration 7 0

CJ, Acromioclavicular joint.Data are expressed as number of cases.Data of Imatani score are expressed as mean � standard deviation.ifferences in the Imatani score between operated and nonoperated groupsere not significant (P � .71).

etween any of these conditions, ACJ dislocation or s

ubluxation, posttraumatic osteoarthritis, or ligamentalcification, and the functional result obtained.

ISCUSSION

The findings of this study, where operative andonservative treatment of type III ACJ injury yieldedimilar satisfactory clinical results, are consistent withhose reported by other comparative investiga-ions.1,6,9,14,17 Several points from our radiographictudy merit discussion, however.

The first important finding is that surgical treatmentoes not imply an anatomic and stable reduction of

he ACJ displacement. The Phemister technique wassed in this series, and a complete reduction of theCJ dislocation was finally achieved in only half of

he patients, although their functional result was veryood. If the rationale of surgical reconstruction of ACJislocations is to achieve an anatomic reconstructionf the shoulder that permits full functional recovery,

his may not always be achieved in type III ACJnjuries.

A few previously reported investigations on conser-ative treatment of ACJ dislocations determined thataintenance of anatomic reduction is not a prerequisite

or regaining adequate shoulder function, and that theegree of displacement in type III ACJ injuries does not

nfluence the result.6,9,13,21 In this study, we reached theame conclusion when the results of surgical treatmentere evaluated; thus, anatomic reconstruction aloneoes not seem to justify surgery. We have to admit,owever, that no procedure to reconstruct the coraco-lavicular ligaments, except its suture, was done. Thoseechniques that involve ligament reconstruction mightave a higher probability of achieving an anatomic andurable reconstruction.4,5,19

A second remarkable finding is that operative man-gement of this lesion is associated with a signifi-antly higher risk of ACJ osteoarthritis. Although theossibility of osteoarthritis after surgical reconstruc-

ion of ACJ injury has been pointed out, it has rarelyeen addressed, and the reports on this complication

ack a control group or statistical analysis.16,17 Frommethodological point of view, comparison of osteo-rthritis between surgically reduced and untreatedislocated joints might not be appropriate. For thiseason, the incidence of osteoarthritis in the oppositeealthy shoulder was also assessed. Because no pa-ient had previous shoulder symptoms, we hypothe-ized that the degree of osteoarthritis before injuryould be similar in both ACJs, and the noninjured

oint could be used as control. No differences coulde detected when the degree of osteoarthritis of theoninjured joint in both treatment groups was com-ared, which confirms our hypothesis.

Interestingly, the surgically treated joints showed a

ignificantly higher incidence of osteoarthritis when

cawjjmi

atrAfttmb

FasocbT

Ftv

T

R

A

C

O

*

J Shoulder Elbow Surg Calvo, López-Franco, and Arribas 303Volume 15, Number 3

ompared with both conservatively treated shouldersnd noninjured shoulders. Moreover, differencesere not significant when conservatively treated

oints were compared with their opposite healthyoints. These results support the idea that osteoarthritisight be the result of the manipulation and the inabil-

igure 1 Anteroposterior cephalic tilt view of a surgically treatedcromioclavicular joint (ACJ) dislocation. The ACJ clearly showsigns of osteoarthritis. The articular space is narrowed, and ansteophyte extending inferiorly from the acromion and the claviclean be observed. Note the sclerosis of the acromial subchondralone and the osteolysis of the upper portion of the clavicular end.he coracoclavicular ligaments display ossification.

able II Radiographic evaluation after acromioclavicular dislocation

Variable

Operated gr

Injured shoulder Opp

eductionComplete 16Subluxed 15Dislocated 1cromioclavicular joint osteoarthritisNo 6Yes 26

Mild 10Moderate 11Severe 5

oracoclavicular ligaments ossificationNo 13Yes 19

Mild 9Moderate 4Severe 6

steolysis of the lateral clavicleNo 18Yes 14

Mild 8Moderate 2Severe 4

Operated compared to nonoperated shoulders.

ty to achieve a functional reconstruction of the joint b

natomy with surgery. The presence of ACJ osteoar-hritis did not show any association with the clinicalesults, but it has to be considered that patients withCJ injuries are usually young, and no long-term

ollow-up studies have analyzed the effect of osteoar-hritis due to ACJ injury and shoulder function. Thus,he possibility of this complication should be kept inind when treatment for type III ACJ joint injuries iseing considered.

Ossification of the coracoclavicular ligaments has

igure 2 Anteroposterior shoulder radiograph of a surgicallyreated ACJ dislocation. The coracoclavicular ligaments are se-erely calcified.

Non-operated group

P*shoulder Injured shoulder Opposite shoulder

.0012 0 10

0 011 1

.0029 8 73 3 40 1 21 1 12 1 1

.04492011

.0625500

oup

osite

3

111

een demonstrated to be associated with ACJ injury

wtlidot

tcahwgscnswo

twautagiamtefAs

TwvfwtocAaastnwfa

wt

ctadwab

Etf

R

1

1

1

1

1

1

1

1

304 Calvo, López-Franco, and Arribas J Shoulder Elbow SurgMay/June 2006

ith ligament disruption,4 but our study demonstrateshat this finding is significantly more frequent whenigament tissue is manipulated in an attempt to repairt. However, other investigations have also failed toemonstrate any influence of the extent or location ofssification of the ligaments supporting the joint on

he outcome of ACJ injury.9Biomechanical factors have been incriminated in

he pathogenesis of primary osteolysis of the laterallavicle; thus, we hypothesized that the anatomiclteration resulting from ACJ injury could result in aigher incidence of this alteration. Although thereas a clear trend to more osteolysis developing in theroup treated surgically, the differences were nottatistically significant at comparison. So, no clearonclusions can be drawn from this study. Finally, theumber of other complications, such as unsightlycars, problems with the fixation device, or infectionas higher after surgical treatment of ACJ injuries, asthers have previously suggested.9,17

Several authors recommend surgical reconstruc-ion in athletes or patients who have to lift heavyeights at work.1,9,10,18 However, ACJ injuries usu-lly affect young active patients, in whom it is also oftmost importance to regain full shoulder function inhe shortest possible period of time. In addition to theforementioned risks of surgical treatment, and theood results obtained with conservative treatment, it

s clear that the recovery time is significantly lowernd that strength is preserved with conservative treat-ent.7,9,18,21 We therefore consider that an acute,

ype III ACJ injury should be managed conservatively,ven in this group of patients, because the excellentunctional results are obtained with a lower risk ofCJ osteoarthritis and coracoclavicular ligaments os-ification.

The present study has a number of weaknesses.he investigation was retrospective, and the patientsere not randomized to undergo surgery or conser-ative treatment; however, there was no preselectionor treatment. Moreover, the study group is small,hich limits the benefit of statistical analysis. To make

he patient series homogeneous, only cases operatedn with the modified Phemister technique were in-luded, and we cannot rule out that transfixion of theCJ could play a role in the development of osteo-rthritis. So, these conclusions are not necessarilypplicable to other techniques, but the trends pre-ented here are consistent with other investigations onype III ACJ injury operated on with other tech-iques.2,3,6,7,13,15 In addition, the time of follow-upas short in the conservatively treated group; thus, a

uture effect on the outcome of any of the variablesnalyzed cannot be ruled out.

In conclusion, type III ACJ injury constitutes a lesionith a high probability of good outcome regardless of

he type of treatment, surgical or conservative. Be-

ause surgical treatment does not achieve better func-ional results, implies a higher risk of osteoarthritis,nd does not guarantee a complete reduction of theislocation or an acceptable, good cosmetic result,e suggest conservative management. Patients mustlways be fully informed about the risk and benefits ofoth surgical and conservative treatments.

We acknowledge Dr Granizo, from the Department ofpidemiology (Fundación Jiménez Díaz), for his participa-ion in the statistical processing of data, and Dr Murcianoor her contribution to this study.

EFERENCES

1. Bannister GC, Wallace WA, Stableforth PG, Hutson MA. Themanagement of acute acromioclavicular dislocation. A random-ised prospective controlled trial. J Bone Joint Surg Br 1989;71:848-50.

2. Bathis H, Tingart M, Bouillon B, Tiling T. Conservative or surgicaltherapy of acromioclavicular joint injury: what is reliable? Asystematic analysis of the literature using “evidence-based medi-cine” criteria. Chirurg 2000;71:1082-9.

3. Bradley JP, Elkousy H. Decision making: operative versus nonop-erative treatment of acromioclavicular joint injuries. Clin SportsMed 2003;22:277-90.

4. Cook DA, Heiner JP. Acromioclavicular joint injuries. Orthop Rev1990;16:510-6.

5. Ferris BD, Bhambra M, Paton DF. Coracoid process transfer foracromioclavicular dislocations. Clin Orthop 1989;242:184-94.

6. Fremerey RW, Lobenhoffer P, Ramacker K, Gerich T, Skutek M,Bosch U. [Acute acromioclavicular joint dislocation—operativeor conservative therapy? ] Unfallchirurg 2001;104:294-9.

7. Galpin RD, Hawkins RJ, Grainger RW. A comparative analysis ofoperative vs. nonoperative treatment of grade III acromioclavicu-lar separations. Clin Orthop 1985;193:150-5.

8. Imatani RJ, Hanlon JJ, Cady GW. Acute, complete acromiocla-vicular separation. J Bone Joint Surg Am 1975;57:328-32.

9. Larsen E, Bjerg-Nielsen A, Christensen P. Conservative or surgicaltreatment of acromioclavicular dislocation: a prospective, con-trolled, randomized study. J Bone Joint Surg Am 1986;68:552-5.

0. Meislin RJ, Zucherman JD, Nainzadeh N. Type III acromioclavic-ular joint separation associated with late brachial-plexus neura-praxia. J Orthop Trauma 1992;6:370-2.

1. Nuber GW, Bowen MK. Acromioclavicular joint injuries anddistal clavicle fractures. J Am Acad Orthop Surg 1997;5:11-8.

2. Phemister PB. The treatment of dislocation of the acromioclavicu-lar joint by open reduction and threaded wire fixation. J Bone JointSurg 1942;24:166-8.

3. Philips AM, Smart C, Groom AF. Acromioclavicular disloca-tion. Conservative or surgical therapy. Clin Orthop 1988;353:10-7.

4. Press J, Zuckerman JD, Gallagher M, Cuomo F. Treatment ofgrade III acromioclavicular separations. Operative versus nonop-erative management. Bull Hosp Jt Dis 1997;56:77-83.

5. Rosenørm M, Pedersen B. A comparison between conservativeand operative treatment of acute acromioclavicular dislocation.Acta Orthop Scand 1974;45:50-9.

6. Smith MJ, Stewart MJ. Acute acromioclavicular separations. A20-year study. Am J Sports Med 1979;7:62-71.

7. Taft TN, Wilson FC, Oglesby JW. Dislocation of the acromiocla-vicular joint. An end-result study. J Bone Joint Surg Am 1991;69:

1045-51.

1

1

2

J Shoulder Elbow Surg Calvo, López-Franco, and Arribas 305Volume 15, Number 3

8. Tibone J, Sellers R, Tonino P. Strength testing after third-degree acro-mioclavicular dislocations. Am J Sports Med 1992;20:328-31.

9. Weinstein DM, McCann PD, McIlveen SJ, Flatow EL, Bigliani LU.

Surgical treatment of complete acromioclavicular dislocations.Am J Sports Med 1995;23:324-31.

2

0. Williams GR, Nguyen VD, Rockwood CA. Classification andradiographic analysis of acromioclavicular dislocations. ApplRadiol 1989;29-34.

1. Wojtys EM, Nelson G. Conservative treatment of grade III acro-mioclavicular dislocations. Clin Orthop 1991;268:112-9.