ORIGINAL ARTICLE

Operative versus nonoperative treatment of displaced midshaftclavicle fractures in adults: a systematic review

Carl-Henrik Rehn • Martin Kirkegaard •

Bjarke Viberg • Morten Schultz Larsen

Received: 29 August 2013 / Accepted: 16 November 2013

� Springer-Verlag France 2013

Abstract

Objectives Intervention studies of clavicle fracture treat-

ment are numerous, but only a few high quality studies

prospectively compare operative and nonoperative treat-

ment. The objective of this study was to review evidence

from randomized controlled trials on operative versus

nonoperative treatment of displaced midshaft clavicle

fractures in adults with focus on fracture healing, compli-

cations and functional outcome.

Data sources An electronic search was performed of

PubMed, Embase and Cochrane databases which revealed

559 articles. All articles published before October 18th

2012 and written in English, Danish or Swedish were

eligible.

Study selection Articles were excluded if they included

children under the age of 16, did not include acute midshaft

fractures, included concomitant fractures, did not meet the

requirements of Level I evidence according to Centre for

evidence based medicine 2009 guidelines, were systematic

reviews or meta-analyses, or did not compare operative and

nonoperative treatment.

Data extraction Articles were parsed for relevance by

two reviewers independently regarding title, abstract, and

full text. Extraction of data was done by both reviewers in

collaboration and sorted according to the aims of the study.

Complications were grouped according to additional sur-

gery required. The quality of studies was assessed by both

reviewers in unison using Critical Appraisal Skills Pro-

gramme 2010 checklists.

Conclusions It seems like operative intervention leads to

fewer nonunions at the cost of an increase in minor com-

plications compared to nonoperative treatment. However,

the effects of operation on functional outcome remains

controversial. High quality evidence is currently sparse

supporting either operative or nonoperative treatment on

displaced midshaft clavicle fractures in adults.

Keywords Displaced � Midshaft � Clavicle �Fracture � Systematic review � Treatment

Introduction

Patients presenting with clavicle fractures are not an

uncommon occurrence in the emergency department as

they represent about 2.5 % of all fractures [1, 2]. The

majority of these fractures occur in the middle third of the

clavicle and about three quarters of these are displaced

[2, 3].

Traditionally, displaced clavicle fractures have been

treated nonoperatively, usually with a sling or figure-of-

eight bandage [4]. This form of treatment was based on two

large studies conducted in the 1960s that reported a non-

union rate of 0.1–0.8 % amongst nonoperatively treated

patients and 3.7–4.6 % in operatively treated patients [5,

6]. These studies had a large number of patients, but the

inclusion of children in the studies makes it difficult to

consider them applicable to adults.

This paper was previously presented at Danish Orthopaedic Society

conference, Copenhagen, Denmark, 2012.

Carl-Henrik Rehn and Martin Kirkegaard have contributed equally to

this project and should be considered co-first authors.

C.-H. Rehn (&) � M. Kirkegaard � B. Viberg � M. S. Larsen

Department of Orthopaedic Surgery and Traumatology,

Odense University Hospital, Sdr. Boulevard 29,

5000 Odense C, Denmark

e-mail: carlhenrik@gmail.com

123

Eur J Orthop Surg Traumatol

DOI 10.1007/s00590-013-1370-3

More recent studies show a nonunion rate in displaced

clavicles amongst nonoperatively treated adult patients of

15–18 % [7, 8]. Studies using new surgical methods and

materials have shown that the same types of fractures heal

with a 97.3–100 % union rate in operatively treated

patients [9, 10]. Furthermore, studies show that patients

with displaced midshaft clavicle fractures treated nonop-

eratively have a lower functional ability than the back-

ground population, as well as a thirty to fifty percent

dissatisfaction rate with the appearance or function of

patient’s shoulders [7, 11].

There are many publications regarding the treatment of

clavicle fractures, but they are of varying quality and there

are few studies that prospectively compare operative and

nonoperative treatment. To critically appraise and sum-

marize the available comparative studies, a systematic

review is warranted. The quality of a systematic review is

determined by the quality of the included studies [12]. This

study was exclusively based on Level I research according

to the most recent Centre for Evidence Based Medicine

(CEBM) guidelines from 2009 [13].

The aim of this systematic review was to appraise the

available Level I evidence on operative versus nonopera-

tive treatment of displaced midshaft clavicle fractures in

adults with focus on fracture healing, functional outcome,

and resulting complications.

Materials and method

Database search

Electronic searches were made of PubMed, Embase and

Cochrane Databases on the 18th of October, 2012. The

following search string was used: ‘‘[(‘‘middle third’’ OR

displaced OR midshaft OR mid-shaft OR midclavicular)

AND clavicular OR ‘‘collar bone’’ fractures] AND (oste-

osynthesis OR bone screws OR pinning OR plating OR

‘‘plate fixation’’ OR nail OR fixation OR surgery OR

treatment OR therapy)’’. No limits were applied to the

search. Articles were excluded if they:

1. were not written in English, Danish or Swedish,

2. included children under the age of 16,

3. did not include acute midshaft fractures, or included

concomitant fractures,

4. did not meet the requirements of Level I evidence

according to CEBM 2009 guidelines [13],

5. were systematic reviews or meta-analyses,

6. did not compare operative and nonoperative treatment.

A detailed flowchart of the selection process can be seen

in Fig. 1. The search yielded a total of 559 articles. After

removal of duplicates, the titles of the remaining 376

articles were independently parsed for relevance by both

main authors to determine which articles qualified for

abstract reading. Abstract screening was performed, inde-

pendently, and articles not deemed eligible by either of the

authors were excluded, narrowing it down to 26 articles.

The articles were then read independently, with specific

focus on the method sections, and any differences were

resolved through discussion and consensus. Ultimately,

five randomized controlled trials (RCT) fulfilled the criteria

of Level I evidence that compared nonoperative and

operative treatment [14–18].

Data extraction

The five articles were reviewed independently by both

authors. All relevant information presented within the

articles, such as study information, demographics, fracture

data and classification, outcome scores, complications and

complication rates were collected. This resulted in a large

amount of information that was entered into a custom

spreadsheet. The extracted data was then divided according

to relevance of the study aim.

For demographics, numerous variables were extracted in

their original form from all articles: number of patients in

study, mean age and range, sex distribution, country and

setting of study. Study time was given in intervals and con-

verted to months where needed, counting the first and not the

last month in the given interval. Operative techniques were

plating [14, 16, 18], pinning [15] and nailing [17]. Com-

pression plate, 3.5 mm reconstruction plate, precontoured

plate and other plates were all combined as ‘‘Plate’’. Modi-

fied Hagie pin and elastic stable intramedullary nail were

Fig. 1 Flow chart representation of the article selection process

Eur J Orthop Surg Traumatol

123

labeled as ‘‘Pin’’ and ‘‘Nail’’, respectively. Data on nonop-

erative treatment was extracted.

Data on mean fracture shortening, fracture side, domi-

nant arm and mechanism of injury were extracted. All

articles defined fracture union as bony or callous healing

over fracture site seen on radiography. The variable ‘‘Mean

time to union’’ was combined from time to union and mean

time to union. Fractures were classified according to either

Robinson or OTA classification. In order to best compare

the fractures, all fracture types were converted to the

Robinson classification [3]. Fractures reported as OTA type

15B1 and 15B2 were pooled as Robinson 2B1, and frac-

tures reported as OTA type 15B3 were considered equal to

Robinson group 2B2. If no specification was given, data

was considered as Robinson group 2B, i.e. displaced

midshaft clavicle fracture.

Outcome scores were presented as Constant Shoulder

Score, L’insalata score, DASH score, or SANE score.

Since DASH scores were not reported consistently and

SANE differed a lot from the other scoring systems, these

were not used. Constant Shoulder Score and L’insalata

scores were reported most consistently at 3, 6 and

12 months across the articles and these values were used in

order to offer better comparability. One article had a fol-

low-up at 24 weeks, which was pooled into the 6 month

category.

Because of the difference in time intervals used to define

‘delayed union’ (3 months to 1 year), these could not be

compared and were listed but ignored as a complication in

this study. Nonunion and symptomatic nonunion were

grouped as ‘‘Nonunion’’. Malunion, non-anatomical union

and shortening were defined similarly across all articles as

healing with shortening, angulation, or displaced position

and all three were pooled as ‘‘Malunion’’. Symptomatic

malunions were extracted in original form. To compare

operative and nonoperative treatment the complications

were grouped according to whether or not they required

additional surgery. Complications that warranted ‘‘Major

surgery’’ were deep wound infection, osteomyelitis, non-

unions, symptomatic malunions, implant failure and early

mechanical failure that occurred before fractures had

healed, and persistent brachial plexus irritation. Lesser

complications treated with minor operative intervention

were grouped as ‘‘Minor surgery’’. This group included

superficial wound infection, pin fracture after osseous

healing, hardware irritation and telescoping. Complications

that did not require additional surgery were grouped as

‘‘No surgery’’. These were nonunions, malunions and

implant related complications that were not symptomatic

enough to warrant further surgery, as well as refractures

that healed without further intervention. Number needed to

treat (NNT) was calculated as the reciprocal value of the

absolute risk reduction.

Study quality

Study quality was assessed by the two main authors in

unison using the Critical Appraisal Skills Programme

(CASP) 2010 study quality checklist for RCT [19]. The

checklist contains 11 questions regarding randomization,

inclusion/exclusion criteria, blinding, follow-up and

results. A simple ‘‘yes’’, ‘‘no’’ or ‘‘can’t tell’’ answer was

sufficient for the majority of the questions and only a few

required a more detailed answer. For question 7 the focus

was on presentation method of results, i.e. tables or

graphs, and the accessibility of the article’s final conclu-

sion. Question 8 required special attention to measure-

ments of significance i.e. confidence intervals and

standard. Question 11 asked for the authors’ opinions on

cost-benefit within the articles. This question was not

included in our assessment as none of the included arti-

cles provided enough information to answer the question

satisfactorily.

Results

Demographics

The number of patients in the studies ranged from 57 in

Judd et al. to 132 in the study by the Canadian Orthopaedic

Trauma Society (COTS) (Table 1) [14, 15]. Judd et al. [15]

was the only study that had a mean age of less than

30 years in both groups. There was a higher frequency of

clavicle fractures in males in all articles. Studies were

performed in Austria, Canada, Finland, Iran and USA.

Study time of the articles range from 19 to 45 months

where Mirzatolooei had the shortest follow-up [16]. All

nonoperative treatment consisted of a simple sling whereas

the operative method differed from study to study. Study

settings varied and different level trauma centres were used

for data collection and treatment. Judd et al. [15] included

only military personnel.

Fracture information and outcome scores

There was no statistically significant difference in mean

initial fracture shortening, fracture side, dominant arm or

mechanism of injury (Table 2). Mean time to union was

reported by two studies and was shorter in the operative

groups than in the nonoperative groups [14, 17]. Robinson

fracture classification showed dissimilar distribution of

fractures between the articles. COTS did not classify

fractures at all [14], Virtanen et al. did not include any 2B2

fracture whereas the population in the study by Mir-

zatolooei consisted exclusively of type 2B2 fractures [16,

18]. Constant Shoulder Scores were reported by all articles

Eur J Orthop Surg Traumatol

123

except Judd et al. which reported L’insalata scores [15].

Scores were most consistently reported at 24 and 52 week

intervals and the Constant Shoulder Scores were higher in

the operative groups at these time points. Judd et al. [15]

was the only study that had a better score for the nonop-

erative group at the 1 year follow-up.

Complications

The prevalence of complications that required additional

major surgery was 3.3 % (6/183) in the operative groups

and 8.6 % (16/186) in the nonoperative groups (Table 3).

To prevent one patient from acquiring a complication that

required additional major surgery the number of patients

that needed to be treated (NNT) operatively was 18.5

(n = 369 total patients). Complications that were treated

with minor surgery were exclusive to the operative group

(n = 64). Two studies routinely offered implant removal to

their patients after fracture healing [15, 17]. This accounted

for 27 of 29 removals in Judd et al. [15] and 23 of 25 in

Smekal et al. [17]. Complications that did not require

surgery were more prevalent in the nonoperative groups.

Only one article reported a neurological complication so

severe that it required surgery [18]. Neurological compli-

cations in the other articles were transient and resolved

without further intervention.

Table 1 Study demographics

Article Patients in study Ages included Patients in

group

Mean age Sex Study time (months) Surgical method

? – ? – Male Female

COTS [14] 132 16-60 67 65 33.5 33.5 87 24 45 Plate

Judd et al. [15] 57 17–40 29 28 28 25 52 5 29 Pin

Mirzatolooei [16] 60 18-65 29 31 36 35.3 41 9 19 Plate

Smekal et al. [17] 60 18-65 30 30 35.5 39.8 52 8 32 Nail

Virtanen et al. [18] 60 18-70 28 32 33 41 52 8 35 Plate

? Operative group

- Nonoperative group

Table 2 Fracture data and outcome scores

Article Mean time to union (weeks) Robinson 2B Functional score

2B1 2B2 3 months 6 months 12 months

COTS [14]

Operative group 16.4* 62R 91 95* 96*

Nonoperative group 28.4* 49R 83 88* 91*

Judd et al. [15]

Operative group – 14 15 73.5 ± 14.3L 87.5 ± 11.2L* 95.5 ± 7.3L*

Nonoperative group – 11 17 66.4 ± 16.2L 85.3 ± 9.1L* 97.9 ± 2.4L*

Mirzatolooei [16]

Operative group – 0 26 – – 89.8*

Nonoperative group – 0 24 – – 78.8*

Smekal et al. [17]

Operative group 12.1 ± 8.6* 12 18 – 96 ± 3.9* –

Nonoperative group 17.6 ± 10.7* 15 15 – 87 ± 11.6* –

Virtanen et al. [18]

Operative group – 28 0 78 ± 10.9 – 86.5 ± 11.5

Nonoperative group – 32 0 80 ± 10.2 – 86.1 ± 8.9

±SD standard deviation

* p \ 0.05L L’insalata scoreR Robinson 2B, further classification not reported in study

Eur J Orthop Surg Traumatol

123

Quality assessment

The result of the quality assessment is presented in Table 4.

All five studies had a clearly focused research question and

the use of a randomized trial was appropriate. Judd et al.

[15] did not adequately describe their method of random-

ization. Mirzatolooei showed inconsistencies in group

allocation [16]. They had thirty sealed envelopes prepared

for each group but included thirty-one patients in the

operative group and twenty-nine in the other. No expla-

nation was offered for this discrepancy. All studies lacked

observer blinding at follow-up and observer bias can not be

excluded. All patients were accounted for at the end of the

studies. The statistical power was low throughout. Three

studies made power analyses a priori, but all failed to

achieve the preset number of patients due to people not

completing the trial [14, 16, 18]. COTS presented their

outcome scores as graphs without values or any explana-

tion of what the intervals in the graph represented [14].

Discussion

Time to union was shorter in the operative groups and

functional scores were better in the operatively treated

patients than in those treated nonoperatively, especially at

the early follow-up stages. There was a higher incidence of

severe complications that required additional major surgery

in the nonoperative groups than in the operative groups.

Minor complications were exclusive and numerous in the

operative intervention groups, but the majority of these

were planned removals.

Only articles that compared operative versus nonoper-

ative treatment in midshaft clavicle fractures in adults were

included. Despite this, the fractures in the included articles

are not entirely comparable. Furthermore, there is an ele-

ment of uncertainty when converting fractures from one

classification system to another as the groups of one system

does not perfectly fit into the other. In addition to displaced

fractures, Judd et al. [15] included an unknown number of

angulated, and thus technically non displaced, fractures.

Virtanen et al. [18] included exclusively type 2B1 fractures

which are less severe fractures. In contrast, Mirzatolooei

differ in two ways: firstly because they only include

comminuted fractures, and secondly because they did not

exclude open fractures [16]. This might skew the

Table 3 Complications

Article Major surgery Minor surgery No surgery

COTS [14]

Operative group 1 8 2

Nonoperative group 9 0 7

Judd et al. [15]

Operative group 3 29 1

Nonoperative group 1 0 1

Mirzatolooei [16]

Operative group 0 2 4

Nonoperative group 0 0 20

Smekal et al. [17]

Operative group 2 25 0

Nonoperative group 5 0 0

Virtanen et al. [18]

Operative group 0 0 4

Nonoperative group 1 0 10

Table 4 Study quality

Article Q1 Q2 Q3 Q4 Q5 Q6 Q7 Q8 Q9 Q10

COTS [14] Yes Yes Can’t

tell

No No No CSS, DASH score,

nonunion, malunion.

Results not clearly presented, CSS and DASH

only in graphs. No SD. Precision of results is

uncertain.

No Yes

Judd et al.

[15]

Yes Yes Yes No Yes Yes L’Insalata score, SANE

score, nonunions,

malunions.

Large SD before 6 months. Acceptably

narrow at 6 months and 1 year.

No No

Mirzatolooei

[16]

Yes Yes Yes No Yes Can’t

tell

CSS, DASH score,

nonunion, malunion,

satisfaction

No CI or SD presented. Precision of results is

uncertain.

No No

Smekal et al.

[17]

Yes Yes Yes No Yes No CSS, DASH score,

nonunion, malunion.

Narrow SD at 6 months and 2 years. Better in

operative group.

Yes Yes

Virtanen

et al. [18]

Yes Yes Yes No Yes Yes CSS, DASH score,

nonunion, malunion.

Large SD for CSS and DASH at 3 months and

1 year. Most likely due to small study

population

Yes No

Q1–10, Questions 1–10 from critical appraisal skills programme (CASP) RCT checklist 2010

CSS constant shoulder score, DASH disabilities of arm, shoulder and hand, SANE single assessment numeric evaluation, SD standard deviation

Eur J Orthop Surg Traumatol

123

complication rates and outcome scores of the respective

studies and may make it difficult to apply the results to

displaced midshaft clavicle fractures in general.

Where reported, mean time to union was shorter for the

operative groups [14, 17]. However, a large interval

between radiographs makes for uncertainty when estab-

lishing the actual time to union. Smekal et al. [17] mini-

mizes this imprecision by regular radiographs at 4 week

intervals. The wide range of intervals across studies may

not only affect the correct approximation of actual union

but also makes time to unions difficult to compare. Fur-

thermore, only two out of the five studies use time to union

as an end-point [14, 17]. In order to more accurately

measure the effectiveness of operative intervention, mean

time to union should be reported more consistently by

studies as this may have an impact on when the patient can

start with weight bearing exercise. A global standardization

of follow-up intervals would also be preferred in order to

better compare the studies.

Constant Shoulder Scores are generally higher in the

operative groups. Despite the statistically significant dif-

ference in scores, the differences are small and perhaps not

clinically significant. COTS mention that a 10 point dif-

ference in Constant Shoulder Score is clinically significant

[14]. We were unable to find a source to confirm this claim

and a systematic review from 2010 concludes that there is

no consensus in minimally clinically important difference

[20]. Furthermore, COTS only report a difference of [10

points at 6 weeks, but not at any later follow-up [14].

Mirzatolooei shows a difference of [10 points at 1 year

[16]. This may be attributed to the fact that they only

include comminuted fractures and do not exclude open

fractures. With this in mind, it seems reasonable to assume

that the more severe the fracture, the less likely it will heal

well without operative intervention. This hypothesis may

be further supported by the fact that Judd et al. and Vir-

tanen et al. [15, 18], include the least severe fractures of all

the studies, and are the ones to show the smallest differ-

ences in scores. One might be tempted to conclude that the

more severe fractures should therefore be treated opera-

tively and that the more simple fractures have less to gain

from surgery. However, there is not enough evidence to

support this claim at present.

Statistical power is low throughout the included articles.

Perhaps the drop-outs were more numerous than antici-

pated, but it leaves the studies without sufficient power to

find differences between the intervention groups. A number

of promising study protocols have been published, such as

a multicentre RCT by Stegeman et al. [21] that will include

350 patients and takes into account many of the issues that

have been raised in this systematic review. The results of

this study may provide powerful Level 1 evidence

regarding treatment of clavicle fractures.

A point to consider in this study is that our results rely

heavily on whether or not the investigators decided to

operate their patients after a complication was detected.

This is exemplified by the fact that Mirzatolooei and Vir-

tanen et al. [16, 18] do not surgically correct many of the

complications that they encounter. Had Mirzatolooei cho-

sen to operate on the 19 malunions that they found, like

COTS did, the NNT of this study would have been 6.4

instead of 18.5; a considerable improvement [14, 16].

Virtanen et al. [18] did offer surgical treatment to their

patients with nonunions but they all declined. One could

argue, like Ring, that cultural differences play a role and

that the Finns are more resilient than the patient in COTS

study, but this is always an issue when comparing inter-

national literature [14, 22]. On top of this, the five articles

do not define nonunion alike. Mirzatolooei and Smekal

et al. defined nonunion as no healing after 6 months, COTS

and Virtanen et al. after 1 year while Judd et al. does not

define nonunion at all. This poses a problem since a frac-

ture that has not healed within 6 months would count as a

nonunion in Mirzatolooei and Smekal et al. but not in

COTS or Virtanen et al. This is not a problem when

comparing operative and nonoperative groups within the

individual article, but is important to consider when com-

paring numbers between the articles.

This study focused on difference in number of compli-

cations between the intervention groups and whether or not

the complications required additional surgery and, if so, the

degree of surgery. Based on the number of complications

reported in the five articles, the number of patients needed

to treat operatively to prevent one patient from acquiring a

complication that requires additional major surgery is 18.5.

A recently published meta-analysis that includes four of the

articles also included in this systematic review, shows a

comparable rate of nonunions and symptomatic malunions

in operatively treated patients as well as a shorter time to

fracture union [23]. They report an NNT of 4.6 which is

based solely on preventing nonunions and symptomatic

malunions whereas we have included a wider range of

complications. If the same calculation with focus on non-

unions and malunions was made in this systematic review,

the NNT would be 4.2. This is very different from the NNT

of 18.5 based on our definition of major surgery. Therefore

it should be remembered that the value of NNT cannot be

compared directly because it is highly affected by the event

sought avoided in the given article.

Conclusion

Further studies, with a high level of evidence, that compare

operative and nonoperative treatment of displaced midshaft

clavicle fractures in adults are required. Internationally

Eur J Orthop Surg Traumatol

123

agreed upon fracture classifications, definitions and func-

tional, and preferably objective, scoring systems would be

extremely useful in order to generate convincing evidence.

Additionally, it is important for future studies to include

enough patients in order to obtain sufficient statistical

power. In conclusion, there currently is not enough Level I

evidence to support either operative or nonoperative

treatment of displaced midshaft clavicle fractures in adults.

Conflict of interest No conflicts of interest related to present study.

References

1. Postacchini F, Gumina S, De Santis P, Albo F (2002) Epidemi-

ology of clavicle fractures. J Shoulder Elbow Surg 11(5):452–456

2. Postacchini R, Gumina S, Farsetti P, Postacchini F (2010) Long-

term results of conservative management of midshaft clavicle

fracture. Int Orthop 34(5):731–736. doi:10.1007/s00264-009-

0850-x

3. Robinson CM (1998) Fractures of the clavicle in the adult. Epi-

demiology and classification. J Bone Joint Surg Br 80(3):476–484

4. Jeray KJ (2007) Acute midshaft clavicular fracture. J Am Acad

Orthop Surg 15(4):239–248

5. Neer CS 2nd (1960) Nonunion of the clavicle. J Am Med Assoc

172:1006–1011

6. Rowe CR (1968) An atlas of anatomy and treatment of midcla-

vicular fractures. Clin Orthop Relat Res 58:29–42

7. Hill JM, McGuire MH, Crosby LA (1997) Closed treatment of

displaced middle-third fractures of the clavicle gives poor results.

J Bone Joint Surg Br 79(4):537–539

8. McKee MD, Seiler JG, Jupiter JB (1995) The application of the

limited contact dynamic compression plate in the upper extrem-

ity: an analysis of 114 consecutive cases. Injury 26(10):661–666

9. Thyagarajan DS, Day M, Dent C, Williams R, Evans R (2009)

Treatment of mid-shaft clavicle fractures: a comparative study.

Int J Shoulder Surg 3(2):23–27. doi:10.4103/0973-6042.57895

10. Sohn HS, Shin SJ, Kim BY (2012) Minimally invasive plate

osteosynthesis using anterior-inferior plating of clavicular mid-

shaft fractures. Arch Orthop Trauma Surg 132(2):239–244.

doi:10.1007/s00402-011-1410-6

11. McKee MD, Pedersen EM, Jones C, Stephen DJ, Kreder HJ,

Schemitsch EH, Wild LM, Potter J (2006) Deficits following

nonoperative treatment of displaced midshaft clavicular fractures.

J Bone Joint Surg Am 88(1):35–40. doi:10.2106/jbjs.d.02795

12. Wright RW, Brand RA, Dunn W, Spindler KP (2007) How to

write a systematic review. Clin Orthop Relat Res 455:23–29.

doi:10.1097/BLO.0b013e31802c9098

13. Medicine CfEB (2009) Levels of Evidence. Centre for Evidence

Based Medicine. http://www.cebm.net/index.aspx?o=1025

14. Canadian Orthopaedic Trauma Society (2007) Nonoperative

treatment compared with plate fixation of displaced midshaft

clavicular fractures. A multicenter, randomized clinical trial.

J Bone Joint Surg Am 89(1):1–10

15. Judd DB, Pallis MP, Smith E, Bottoni CR (2009) Acute operative

stabilization versus nonoperative management of clavicle frac-

tures. Am J Orthop (Belle Mead NJ) 38(7):341–345

16. Mirzatolooei F (2011) Comparison between operative and non-

operative treatment methods in the management of comminuted

fractures of the clavicle. Acta Orthop Traumatol Turc

45(1):34–40. doi:10.3944/aott.2011.2431

17. Smekal V, Irenberger A, Struve P, Wambacher M, Krappinger D,

Kralinger FS (2009) Elastic stable intramedullary nailing versus

nonoperative treatment of displaced midshaft clavicular fractures-

a randomized, controlled, clinical trial. J Orthop Trauma

23(2):106–112. doi:10.1097/BOT.0b013e318190cf88

18. Virtanen KJ, Remes V, Pajarinen J, Savolainen V, Bjorkenheim

JM, Paavola M (2012) Sling compared with plate osteosynthesis

for treatment of displaced midshaft clavicular fractures: a ran-

domized clinical trial. J Bone Joint Surg Am. doi:10.2106/jbjs.j.

01999

19. Checklist for Randomised Controlled Trials. (2010) Critical

Appraisal Skills Programme. http://www.casp-uk.net/wp-content/

uploads/2011/11/CASP_RCT_Appraisal_Checklist_14oct10.pdf

20. Roy JS, MacDermid JC, Woodhouse LJ (2010) A systematic

review of the psychometric properties of the Constant-Murley

score. J Shoulder Elbow Surg 19(1):157–164. doi:10.1016/j.jse.

2009.04.008

21. Stegeman SA, de Jong M, Sier CF, Krijnen P, Duijff JW, van

Thiel TP, de Rijcke PA, Soesman NM, Hagenaars T, Boekhoudt

FD, de Vries MR, Roukema GR, Tanka AF, van den Bremer J,

van der Meulen HG, Bronkhorst MW, van Dijkman BA, van

Zutphen SW, Vos DI, Schep NW, Eversdijk MG, van Olden GD,

van den Brand JG, Hillen RJ, Frolke JP, Schipper IB (2011)

Displaced midshaft fractures of the clavicle: non-operative

treatment versus plate fixation (Sleutel-TRIAL). A multicentre

randomised controlled trial. BMC Musculoskelet Disord 12:196.

doi:10.1186/1471-2474-12-196

22. Ring D (2012) Disease and Illness: Commentary on an article by

Kaisa J. Virtanen, MD, et al., Sling compared with plate osteo-

synthesis for treatment of displaced midshaft clavicular fractures.

A randomized clinical trial. J Bone Joint Surg Am. doi:10.2106/

jbjs.l.00728

23. McKee RC, Whelan DB, Schemitsch EH, McKee MD (2012)

Operative versus nonoperative care of displaced midshaft cla-

vicular fractures: a meta-analysis of randomized clinical trials.

J Bone Joint Surg Am 94(8):675–684. doi:10.2106/jbjs.j.01364

Eur J Orthop Surg Traumatol

123