ARTICLE IN PRESS

J Shoulder Elbow Surg (2017) ■■, ■■–■■

www.elsevier.com/locate/ymse

ORIGINAL ARTICLE

Surgery for the fractured clavicle: factors

predicting nonunion
Neil E. Jarvis, FRCS(Tr &Orth)*, Lucy Halliday, MB, ChB, Matthew Sinnott, MB, ChB,
Tanya Mackenzie, MSc(Trauma, Orthopaedics), Lennard Funk, FRCS(Tr&Orth),
Puneet Monga, MD, FRCS(Tr&Orth)

Wrightington Hospital, Wigan and Leigh NHS Foundation Trust, Appley Bridge, Wigan, UK

Background: This study identifies the reasons for failure after plate osteosynthesis of midshaft clavicle
fractures, complication rates, and time to radiographic union.
Methods: A retrospective review of 84 consecutive patients who had undergone surgical fixation for a
midshaft clavicle fracture was performed.
Results: There were 82 patients who were included for analysis and operated on by 11 different sur-
geons using a mixture of locking (63%) and nonlocking (37%) plates. The rate of osteosynthesis failure
was 12.2%. A logistical regression analysis found that failure of osteosynthesis had no relationship to type
of plate used (P = .82), gender (P = .42), number of proximal (P = .96) or distal (P = .63) screws to the
fracture, or length of plate (P = .42). Smoking was found to be the only risk factor (P = .02) that in-
creased failure rates after midshaft clavicle osteosynthesis.
Conclusion: Smoking was the only identifiable risk factor to increase failure rates in clavicle osteosyn-
thesis. Preoperative counseling can identify those at increased risk of implant failure and can help improve
clinical results by implementing a smoking cessation plan.
Level of evidence: Level IV; Case Series; Treatment Study
© 2017 Journal of Shoulder and Elbow Surgery Board of Trustees. All rights reserved.
Keywords: Clavicle; fracture; midshaft; osteosynthesis; plate; failure; smoking

Clavicle fractures are common, accounting for 5%-10% in malunion improves functional outcomes and patient
of all fractures, at an incidence of 50 per 100,000 people per satisfaction.8,20 The restoration of clavicle length is thought
year.15 Following early reports by Rowe17 of conservative man- to improve glenoid orientation and scapular position, reduc-
agement, indicating good range of motion and high rates of ing abnormal forces through the sternoclavicular and
union, treatment has previously been mostly conservative. acromioclavicular joints, and to improve muscle strength
Studies then began filtering through suggesting that patients around the shoulder.11 In active or semiprofessional ath-
complained of pain, loss of strength, distal paresthesia, and letes, osteosynthesis has been shown to allow quick return
issues of cosmesis after shortened malunion of midshaft clav- to sports and early pain relief.21 A randomized controlled trial
icle fractures.7 More recent reports have shown that restoration published by the Canadian Orthopaedic Trauma Society in
of clavicle length after failed conservative management ending 20073 changed surgeons’ outlook regarding clavicular osteo-
synthesis. It reported on functional and radiologic results of
*Reprint requests: Neil Jarvis, FRCS(Tr &Orth), Jan van Rijswijcklaan
midshaft clavicle fractures managed conservatively or with
141, B-2018 Antwerp, Belgium. osteosynthesis; the investigators recommended osteosynthe-
E-mail address: [email protected] (N. Jarvis). sis in displaced or shortened midshaft clavicle fractures.

1058-2746/$ - see front matter © 2017 Journal of Shoulder and Elbow Surgery Board of Trustees. All rights reserved.
https://doi.org/10.1016/j.jse.2017.10.010
 ARTICLE IN PRESS
2 N.E. Jarvis et al.

An improvement in our understanding of the mechanics ascertain the effects of number of proximal screws, number of distal
of the clavicle and our ability to identify fracture patterns such screws, size of plate, age at fracture, gender, smoking status, use
as shortened, severely comminuted, or displaced that will lead of fixation device, and time between injury and fixation on the like-
to an inferior clinical result have led to an increase in the lihood of clavicular union after osteosynthesis.
A power analysis for multiple regression was conducted using
number of clavicle osteosyntheses. The aim of our study was
the equation N = [(Zα + Zβ)/C] 2 + 3,9 where N is the total sample
to assess the complication rates, the time to union, and the
size, Zα is the standard normal deviate for α = .05, Zβ is the stan-
reasons for failure of osteosynthesis in midshaft clavicle dard normal deviate for β = .20, and C = 0.5 * ln[(1 + r)/(1 −
fractures. r)] = 0.310 in which r is the expected correlation coefficient. An α
of .05 was set as the threshold probability for rejecting the null hy-
pothesis (type I error); β was set at .02 as the probability of failing
Materials and methods to reject the null hypothesis under the alternative hypothesis (type
II error). R value was set at 0.30 as the expected correlation coef-
A retrospective review of 84 patients who had undergone surgery ficient. The result of the calculation is
for a shortened or displaced fracture of the midshaft clavicle (AO
N = [( Zα + Zβ ) C ] 2 + 3 = 85,
type 15-B) at the Wrightington, Wigan and Leigh NHS Founda-
tion Trust and The Arm Clinic, Manchester, in the period of January showing that the study was adequately powered.
2008–February 2012 was conducted. Patients were identified on the
hospital’s electronic patient records system. For each case, the age
of the patient at time of injury, gender, smoking status, and number
Results
of days from injury to primary operation (and, if applicable, sec-
ondary operation) were recorded. The mechanism of injury was also A total of 84 midshaft clavicle fractures were identified at the
noted. The time until radiologic union was assessed by calculating institutions. Two were lost to follow-up or had an incom-
the number of days from the date of operation to the date that union plete data set, and 82 were included for analysis. All were
was confirmed either in the notes or as observed on radiographs closed injuries, and none had distal neurovascular deficit.
through the hospital picture archiving and communication system. Surgery was undertaken by 11 different surgeons, with junior
Operation notes were accessed through electronic patient records surgeons being supervised by consultants. Fixation device was
and used to record the American Society of Anesthesiologists grade decided by surgeon. Three different fixation types were used:
of each patient, type of plate used, quality of reduction, and number
Dynamic Compression Plate (Synthes, Oberdorf, Switzer-
of proximal and distal screws.
Complications were defined as any adverse event in the pa-
land), Acumed clavicle locking plate (Hillsboro, OR, USA),
tient’s notes and compared according to comorbidity status of the and AO Synthes Recon plate. A total of 52 locking plates
patient, age of the patient, time between injury and osteosynthesis, (63%) and 30 nonlocking plates (Dynamic Compression Plate
smoking status, and number of proximal and distal screws. and Recon plates, 37%) were used. Patient demographics are
Reoperation was defined as any operation as a consequence of the shown in Table I, and the plate lengths and numbers of screws
initial osteosynthesis. Major complications were classified as any proximal and distal to the fracture are shown in Figures 1 and
complication that required further operation, and minor complica- 2. The mechanism of injury is displayed as a pie chart in
tions were classified as any deviation from a normal postoperative Figure 3. The overall complication rate was 20.7% (10 major
course not requiring further operation or altering the final outcome. and 7 minor complications), giving a major complication
The primary outcome was failure of osteosynthesis for any reason rate of 12.2% and a minor complication rate of 8.5% as de-
(plate fracture, loosening, and atrophic nonunion). Secondary out-
tailed in Tables II and III. Of the major complications, which
comes were complications (infection, numbness, secondary adhesive
capsulitis) and removal of symptomatic hardware.
all required reoperation, 2 plates fractured after a fall and 1
All operations were performed in the following manner. The while BMXing, and 4 became loose as a result of non-
patient was placed in the beach chair position, and preoperative in- union. One patient had a confirmed surgical site infection that
travenous antibiotic prophylaxis was administered. The unit’s standard
fixation of a midshaft clavicle fracture begins with an incision in-
ferior to the clavicle, dissection of the clavipectoral fascia, reduction Table I Patient demographics, gender, age at time of injury,
of the fracture, and temporary fixation with Kirschner pins or re- and time from injury to operation.
duction clamps before definitive fixation. The construct is then washed Total: 82
out with saline, hemostasis is obtained, and the fascia and skin are Male: 62
closed separately. A sling is applied for comfort up to 6 weeks post- Female: 20
operatively, and active and passive physiotherapy is begun when Mean age at time of injury (years): 31.02
comfort allows. For the first 2 weeks, movement is restricted to below Range: 12-58
90° of abduction to reduce rotatory forces through the construct. Pa- Mean time until primary operation (days): 12
tients are then followed up at regular intervals postoperatively until Range: 0-32
radiologic union. Mean time until secondary operation (days): 141
SPSS Statistics for Windows version 20.0 (IBM, Armonk, NY, Range: 13-360
USA) was used for statistical analysis. Non-normally distributed data Mean in-patient stay (nights): 1
were presented as median values and ranges, and binomial data were Range: 0-5
presented as proportions. A logistic regression was performed to
 ARTICLE IN PRESS
Surgery fractured clavicle: factors predicting nonunion 3

Figure 1 Number and length of osteosynthesis plates.

70

53
screws distal to fracture
screws proximal to fracture

35

18

0
2 3 4

Figure 2 Number of screws proximal and distal to fracture.

Table II Major complications arising after osteosynthesis of

midshaft clavicle fractures
Complication Number
Plate loosening (non-union) 4 (2 non locking, 2 locking)
Plate fracture 2
Wound dehiscence 1
Atrophic non-union 1
Periprosthetic fracture 2

the plate and scar, and 3 developed a secondary frozen shoul-

der. One patient received oral antibiotics for a superficial
Figure 3 The mechanism of injury as displayed in a pie chart. wound infection. Our overall infection rate was 2.4%. A total
RTA, road traffic accident. of 16 (21.6%) patients had their osteosynthesis material
removed because of discomfort and 1 early (after 13 days)
because of infection that could not be brought under control
responded to a course of antibiotics, both intravenously and with intravenous antibiotics. Of those having osteosynthe-
orally, but required plate removal and washout. One became sis material removed after radiologic union for comfort, there
an established atrophic nonunion and required reoperation with was a mean time to reoperation of 192 days (120-360 days).
pelvic autograft. The 2 periprosthetic fractures occurred after Of the 82 fractures, all were followed up to union. The
falls. Of the minor complications, 3 had problems of pain from mean time to union was 120 days, and union was confirmed
 ARTICLE IN PRESS
4 N.E. Jarvis et al.

cohort. Infection has been reported at a rate between 1%21

Table III Minor complications arising after osteosynthesis of
midshaft clavicle fractures and 7.4%2 after clavicle osteosynthesis. Ashman et al,1 in a
cohort similarly sized to ours, looked at the risk of reoperation
Complication Number
after osteosynthesis of midshaft clavicle fractures and failed
Pain from plate 1 to find statistically significant association between female
Mild infection 1 gender, plate type, plate position, age, or fracture class and
Pain from scar 1 risk of reoperation. Shin et al,19 in their series of 135 osteo-
Adhesive capsulitis 3
syntheses after displaced midshaft clavicle fractures, found
Plate prominence 1
a similar failure of osteosynthesis rate of 12%, with frac-
ture configuration (P = .012) and use of cerclage wire to
strengthen the surgical construct (P = .009) found to be sta-
radiologically. In total, there were 10 failures of fracture fix-
tistically significant risk factors for failure. Our series reports
ation (12.2%), and of them, a total of 5 nonunions were observed
that the type of plate used (locking or nonlocking) and number
(6.1%). Nagelkerke R2 = 0.30 indicated a relationship of 30%
of screws proximal and distal to the fracture also have no as-
between the predictors and the prediction in our model and
sociation with failure of osteosynthesis. It would appear that
correctly classified 89.2% of cases. A logistic regression was
biologic processes such as interference in blood supply after
performed to ascertain the effects of number of proximal screws
comminution and patient-specific factors interfering with
(P = .96), number of distal screws (P = .63), size of plate
healing, such as smoking, have a greater role to play in frac-
(P = .42), age at time of fracture (P = .96), gender (P = .42),
ture union after osteosynthesis.
smoking status (P = .014), type of fixation device (locking
Smoking has been shown to reduce fracture healing rates
or nonlocking plate; P = .82), and time between injury and
in other fractures. In a series of 105 open tibia fractures,
fixation (P = .06) on the likelihood of clavicular union after
smokers were 37% less likely to achieve union (P = .01).4 In
open reduction–internal fixation (ORIF). This demonstrated
conservatively managed clavicle fractures, it has also been
that only smoking status made a significant contribution to
shown to increase rates of nonunion (OR, 4.16; 95% confi-
fracture union after ORIF. With a P value of .02, smokers were
dence interval, 1.01-14.16)13 along with fracture comminution
14.28 times more likely to have nonunion of clavicular frac-
and displacement, and at 6 weeks after fracture in conjunc-
tures after ORIF than nonsmokers were. No other variables
tion with a Disabilities of the Arm, Shoulder, and Hand score
were significant predictors of clavicle union after ORIF for
of >35, the nonunion rate was 44%.5 It has also been shown
clavicle fracture. No other variables were significant predic-
to be associated with established nonunions of the clavicle
tors of clavicle union after ORIF for clavicle fracture.
treated with ultrasound14 and, in a large meta-analysis of more
than 2100 conservatively managed clavicle fractures.10 To date,
Discussion no study has shown smoking to be a risk factor for failure
of osteosynthesis of the clavicle.
Our complication rate of 20.7% and removal of hardware rate Our study shows that high rates of union after osteosyn-
of 21.6% are broadly the same as in other published series. thesis can be achieved with a mixed group of senior and junior
Fridberg6 et al in their series of 105 locking plate clavicle os- surgeons, using different plates, and appears to contradict the
teosyntheses found an overall complication rate of 23% and long held belief that clavicle fixation requires at least 3 screws
a reoperation rate of 34% for removal of hardware. However, proximal and distal to the fracture. This may be due to the
the Canadian trauma group in their multicenter randomized fact that patients are often young and healthy and have good
controlled trial reported a complication rate of 34% and healing potential. Their fractures would probably unite anyway
reoperation rate of 21%.3 The majority of reoperations in our after fixation of differing stability and technical quality. The
series were due to removal of symptomatic hardware after quest for improvement in results after clavicular osteosyn-
radiographic union. In the United Kingdom, routine removal thesis should now include patient-specific factors, and smoking
of hardware is not the norm, unlike in other countries, where cessation advice should be given to the patients perioperatively.
reoperation rates for hardware removal in other series reach
68%-90%.2,16,18
Leroux et al12 looked at reoperations after clavicle osteo-
synthesis in a large cohort of 1350 patients identified through
Conclusion
an administrative database and found a reoperation rate for
Osteosynthesis of midshaft clavicle fractures results in high
hardware removal of 18.8% and reoperation rate for non-
rates of union and can be performed with good results by
union, infection, and malunion of 6.3%. They found higher
surgeons of differing grades and with locking and
rates of reoperation in women (odds ratio [OR], 2.2; P = .04)
nonlocking plates. To further reduce the risk of osteosyn-
and in those with higher rates of comorbidities (OR, 2.8;
thesis failure, time should be spent counseling the patient
P = .009). Our smaller cohort found no link between gender
perioperatively on patient-specific factors, including
and risk of implant failure; however, our infection rate of 2.4%
smoking cessation.
is broadly comparable with the 1.1% identified in the larger
 ARTICLE IN PRESS
Surgery fractured clavicle: factors predicting nonunion 5

Reconstr 2007;2:59-61. http://dx.doi.org/10.1007/s11751-007-0024

Disclaimer -6
9. Hulley SB, Cummings SR, Browner WS, Grady D, Newman TB.
The authors, their immediate families, and any research Designing clinical research: an epidemiologic approach. 4th ed.
Philadelphia: Lippincott Williams & Wilkins; 2013. p. 79 Appendix 6C.
foundation with which they are affiliated have not re- 10. Jørgensen A, Troelsen A, Ban I. Predictors associated with nonunion
ceived any financial payments or other benefits from any and symptomatic malunion following non-operative treatment of
commercial entity related to the subject of this article. displaced midshaft clavicle fractures—a systematic review of the
literature. Int Orthop 2014;38:2543-9. http://dx.doi.org/10.1007/s00264
-014-2450-7
11. Ledger M, Leeks N, Ackland T, Wang A. Short malunions of the clavicle:
an anatomic and functional study. J Shoulder Elbow Surg 2005;14:349-
References 54. http://dx.doi.org/10.1016/j.jse.2004.09.011
12. Leroux T, Wasserstein D, Henry P, Khoshbin A, Dwyer T, Ogilvie-Harris
1. Ashman BD, Slobogean GP, Stone TB, Viskontas DG, Moola FO, Perey D, et al. Rate of and risk factors for reoperations after open reduction
BH, et al. Reoperation following open reduction and plate fixation of and internal fixation of midshaft clavicle fractures: a population-based
displaced mid-shaft clavicle fractures. Injury 2014;45:1549-53. study in Ontario, Canada. J Bone Joint Surg Am 2014;96:1119-25.
http://dx.doi.org/10.1016/j.injury.2014.04.032 http://dx.doi.org/10.2106/JBJS.M.00607
2. Böstman O, Manninen M, Pihlajamäki H. Complications of plate fixation 13. Liu W, Xiao J, Ji F, Xie Y, Hao Y. Intrinsic and extrinsic risk factors
in fresh displaced midclavicular fractures. J Trauma 1997;43:778-83. for nonunion after nonoperative treatment of midshaft clavicle fractures.
3. Canadian Orthopaedic Trauma Society. Nonoperative treatment compared Orthop Traumatol Surg Res 2015;101:197-200. http://dx.doi.org/
with plate fixation of displaced midshaft clavicular fractures. A 10.1016/j.otsr.2014.11.018
multicenter, randomized clinical trial. J Bone Joint Surg Am 2007;89:1- 14. Nolte PA, van der Krans A, Patka P, Janssen IM, Ryaby JP, Albers GH.
10. http://dx.doi.org/10.2106/JBJS.F.00020 Low-intensity pulsed ultrasound in the treatment of nonunions. J Trauma
4. Castillo RC, Bosse MJ, MacKenzie EJ, Patterson BM. Impact of smoking 2001;51:693-702.
on fracture healing and risk of complications in limb-threatening open 15. Nordqvist A, Petersson C. The incidence of fractures of the clavicle.
tibia fractures. J Orthop Trauma 2005;19:151-7. Clin Orthop Relat Res 1994;300:127-32.
5. Clement ND, Goudie EB, Brooksbank AJ, Chesser TJS, Robinson CM. 16. Poigenfürst J, Rappold G, Fischer W. Plating of fresh clavicular fractures:
Smoking status and the Disabilities of the Arm Shoulder and Hand score results of 122 operations. Injury 1992;23:237-41.
are early predictors of symptomatic nonunion of displaced midshaft 17. Rowe CR. An atlas of anatomy and treatment of midclavicular fractures.
fractures of the clavicle. Bone Joint J 2016;98-B:125-30. http:// Clin Orthop Relat Res 1968;58:29-42.
dx.doi.org/10.1302/0301-620X.98B1.36260 18. Shen WJ, Liu TJ, Shen YS. Plate fixation of fresh displaced midshaft
6. Fridberg M, Ban I, Issa Z, Krasheninnikoff M, Troelsen A. Locking plate clavicle fractures. Injury 1999;30:497-500.
osteosynthesis of clavicle fractures: complication and reoperation rates 19. Shin SJ, Do NH, Jang KY. Risk factors for postoperative complications
in one hundred and five consecutive cases. Int Orthop 2013;37:689-92. of displaced clavicular midshaft fractures. J Trauma Acute Care Surg
http://dx.doi.org/10.1007/s00264-013-1793-9 2012;72:1046-50. http://dx.doi.org/10.1097/TA.0b013e31823efe8a
7. Hill JM, McGuire MH, Crosby LA. Closed treatment of displaced middle 20. Simpson NS, Jupiter JB. Clavicular nonunion and malunion: evaluation
third fractures of the clavicle gives poor results. J Bone Joint Surg Br and surgical management. J Am Acad Orthop Surg 1996;4:1-8.
1997;79:537-9. 21. Verborgt O, Pittoors K, Van Glabbeek F, Declercq G, Nuyts R,
8. Hillen RJ, Eygendaal D. Corrective osteotomy after malunion Somville J. Plate fixation of middle-third fractures of the clavicle
of mid shaft fractures of the clavicle. Strategies Trauma Limb in the semi-professional athlete. Acta Orthop Belg 2005;71:17-21.