VOL. 93-B, No. 8, SEPTEMBER 2011 1247

UPPER LIMB

A radiological study to define safe zones for drilling during plating of clavicle fractures

A. Sinha, J. Edwin, B. Sreeharsha, V. Bhalaik, P. Brownson

From Royal Liverpool and Broadgreen University Teaching Hospitals NHS Trust, Liverpool, United Kingdom

A. Sinha, MBBS, MRCS(Ed), Specialist Registrar in Trauma and Orthopaedics J. Edwin, MRCS, DNB Orth, Dip. Orth, Specialist Registrar in Trauma and OrthopaedicsQueen Elizabeth Hospital, Stadium Road, Woolwich, London SE18 4QH, UK.

B. Sreeharsha, MBBS, FRCR, Specialist Registrar in Radiology V. Bhalaik, MBBS, MRCS, FRCS(Tr & Orth), Consultant Hand and Upper Limb SurgeonWirral University Teaching Hospital NHS Foundation Trust, Upton, Wirral CH49 5PE, UK.

P. Brownson, DM, FRCS(Ed), FRCS(Tr & Orth), Consultant Orthopaedic Shoulder SurgeonRoyal Liverpool and Broadgreen University Teaching Hospitals NHS Trust, Prescot Street, Liverpool L7 8XP, UK.

Correspondence should be sent to Mr A. Sinha; e-mail: apurv16@gmail.com

©2011 British Editorial Society of Bone and Joint Surgerydoi:10.1302/0301-620X.93B9.25739 $2.00

J Bone Joint Surg Br 2011;93-B:1247-52.Received 20 December 2010; Accepted after revision 31 May 2011

This study investigated the anatomical relationship between the clavicle and its adjacent vascular structures, in order to define safe zones, in terms of distance and direction, for drilling of the clavicle during osteosynthesis using a plate and screws following a fracture. We used reconstructed three-dimensional CT arteriograms of the head, neck and shoulder region. The results have enabled us to divide the clavicle into three zones based on the proximity and relationship of the vascular structures adjacent to it. The results show that at the medial end of the clavicle the subclavian vessels are situated behind it, with the vein intimately related to it. In some scans the vein was opposed to the posterior cortex of the clavicle. At the middle one-third of the clavicle the artery and vein are a mean of 17.02 mm (5.4 to 26.8) and 12.45 mm (5 to 26.1) from the clavicle, respectively, and at a mean angle of 50° (12 to 80) and 70° (38 to 100), respectively, to the horizontal. At the lateral end of the clavicle the artery and vein are at mean distances of 63.4 mm (46.8 to 96.5) and 75.67 mm (50 to 109), respectively.

An appreciation of the information gathered from this study will help minimise the risk of inadvertent iatrogenic vascular injury during plating of the clavicle.

Fractures of the clavicle typically occur inyoung, active individuals and account forbetween 2.6% and 4% of all fractures, with upto 80% occurring in the middle third.1-3 Tradi-tionally treatment has been non-operative2,3

and there is a general consensus thatundisplaced fractures should continue to bemanaged in this manner, as the majority healwithout sequelae.4 On the other hand, out-comes following non-operative managementof displaced midclavicular fractures are lessfavourable, with several studies showingincreased rates of nonunion and symptomaticmalunion.5-8 Displaced fractures of the claviclethat have been managed non-operatively havebeen reported to show nonunion rates of up to15%, compared with 2% for internally fixedfractures.8 Consequently, over the past decadethere has been an increased tendency tooperate on displaced fractures of the clavicle.

The aim of this study was to establish safezones for drilling screw-holes during osteo-synthesis of the clavicle, by defining the proxim-ity of the adjacent major blood vessels usingreconstructed three-dimensional (3D) CTangiograms. The precision of CT angiographyin accurately demonstrating even microvascularstructures has already proved useful in studiesrequiring detailed anatomical visualisation ofthe vascular system.9 We believe that this work

represents the first study to use CT angiographyto determine safe drilling trajectories and depthsfor the clavicular plating.

Patients and MethodsInstitutional approval was acquired to obtain26 normal CT arteriograms of the neck andshoulder region, which had been requested byvascular surgeons to evaluate the subclavianvessels for non-trauma reasons. In theselected scans the most common reason forreferral was ischaemic symptoms in the upperlimb in 23 patients, thrombus in the sub-clavian or brachial artery in two and intra-operative injury to the subclavian vessel inone. All scans were performed with thepatient in the supine position with arm by theside of the patient in neutral rotation. Imageswere reconstructed in 3D and analysed by asenior radiologist (BS) using the PictureArchiving and Communication System(PACS) on radiologically diagnostic screenswith minimum screen resolution of 1280 ×1024 pixels. Three locations, representing themedial (A), middle (B) and lateral (C) zones ofthe clavicle, were defined (Fig. 1). Point A waslocated 1 cm lateral to the medial end of theclavicle, point B in the middle, and point C1 cm medial to the lateral end. Theparameters measured included the length of

NOSNWORB.P ,KIALAHB.V ,AHSRAHEERS.B ,NIWDE.J ,AHNIS.A8421

THE JOURNAL OF BONE AND JOINT SURGERY

the clavicle (as measured on 3D CT reconstructionimages); its thickness at points A, B and C; the minimumdistance of the artery and vein from points A, B and C;and the direction and location of the vessels with respectto the clavicle in each of the three zones. Measurementsfor minimum distances were taken from the cortex closestto the vessel to the tunica adventitia of the vesselconcerned. The trajectories were measured at Point ‘B’.The angle between a line joining the centre of clavicle andthe centre of the vessel in relation to the horizontal planewas noted in sagittal 3D CT reconstruction images.

There were 11 men and 15 women in the study, with anaverage age of 51 years (20 to 85), and all were consideredto be skeletally mature. The subclavian arteriograms werefrom the right shoulder in 11 and the left in 15. No differ-ence in the bone-vessel relationship was noted betweenthe sides.Statistical analysis. Data were analysed using MicrosoftExcel (Microsoft, Redmond, Washington). Mean values, SD

and range were calculated in each category. An unpairedtwo-tailed t-test was used to test statistical significancewhen comparing males with females.

ResultsAt the medial end of the clavicle (point A) the subclavianvessels lay posteriorly (Fig. 2). The mean distance of the sub-clavian vein from the clavicle was 4.77 mm (0.00 to 15.90),the mean distance of the subclavian artery was 25.93 mm(5.00 to 44.70) and the mean thickness of the clavicle was24.84 mm (13.32 to 34.16) (Table I). There were no signifi-cant gender differences in the measurements taken at thispoint (unpaired t-tests, p ? 0.45) (Fig 3; Table II).

In the middle of the clavicle (point B) the vessels layposteriorly and inferiorly (Fig. 4). The artery and vein wereat a mean angle of 50° (12° to 80°) and 70° (38° to 100°) tothe horizontal plane of the clavicle, respectively (Figs 4and 5; Table III). Here, the mean distance of the artery andvein from the clavicle was 17.02 mm (5.40 to 26.80) and12.45 mm (5.00 to 26.10), respectively. The mean thick-ness of the clavicle was 15.30 mm (11.47 to 28.35)(Table I). Once again there were no significant genderdifferences in the measurements taken at this point(unpaired t-tests, p ? 0.09) (Figs 3 and 6; Table II).

At the lateral end of the clavicle (point C) the vessels layinferiorly (Fig. 7). Here, the mean distance of the artery and

Fig. 1

Photograph of a non-anatomical model showing the three points A, B andC on the clavicle representing the three zones: medial (Red), middle(Amber) and lateral (Green), respectively.

Fig. 2

Sagittal section at the medial one-third (point A) of the clavicle ina three-dimensional reconstructed CT arteriogram (SA, sub-clavian artery; SV, subclavian vein).

Table I. The mean (range) distance in mm of the artery and vein from the clavicle, the thickness of the clavicle in the three zones, andthe direction of the vessels

Parameter Mean (mm) (range) Direction from clavicle

Point A (medial one-third) Thickness 24.84 (13.32 to 34.16) PosteriorDistance of artery 25.93 (5.00 to 44.70)Distance of vein 4.77 (0.00 to 15.90)

Point B (middle one-third) Thickness 15.30 (11.47 to 28.35) Posteroinferior (See Table III and Fig. 5 for angles)Distance of artery 17.02 (5.40 to 26.80)Distance of vein 12.45 (5.00 to 26.10)

Point C (lateral one-third) Thickness 13.60 (9.50 to 17.99) InferiorDistance of artery 63.40 (46.80 to 96.50)Distance of vein 75.67 (50.00 to 109.40)

A RADIOLOGICAL STUDY TO DEFINE SAFE ZONES FOR DRILLING DURING PLATING OF CLAVICLE FRACTURES 1249

VOL. 93-B, No. 8, SEPTEMBER 2011

vein from the clavicle measured 63.40 mm (46.8 to 96.5)and 75.67 mm (50.0 to 109.4), respectively. The meanthickness of the clavicle was 13.60 mm (9.50 to 17.90)(Table I). At this point there were significant gender differ-ences in the distances measured between the vessels and theclavicle (artery, p = 0.02; vein, p = 0.03; unpaired two-tailed t-tests) (Figs 3 and 6). The distances of the artery andvein from the clavicle were a mean of 11.54 mm and13.85 mm greater in men than in women, respectively(Table II).

The mean length of the clavicles measured on 3D CTreconstruction images was 147 mm (120 to 165); the maleclavicle was longer than the female clavicle, with a meanlength of 152.84 mm (128 to 165) compared with139.52 mm (120 to 164) respectively, but this was notstatistically significant (p = 0.49). There were no significantgender differences in the thickness of the clavicles with themean thickness at points A, B and C being 24.84 mm(13.32 to 34.16), 15.30 mm (11.47 to 28.35) and13.60 mm (9.50 to 17.99), respectively (Table I).

DiscussionThree-dimensional CT angiography enables detailed andprecise in vivo visualisation of the peri-clavicular bloodvessels and is capable of reliably measuring microvascularstructures with an internal diameter as small as 0.4 mm.9

The risk of iatrogenic injury to adjacent vital structuressuch as the subclavian vessels during the use of a plate andscrews during osteosynthesis of the clavicle is wellknown,10-13 particularly during drilling and tapping.Inappropriately directed or excessively long screws havealso been implicated in the development of late complica-tions such as thoracic outlet syndrome.14-16 The majorstructures at risk of injury include the subclavian vein, thesubclavian artery, the brachial plexus and the cervicalpleura. The brachial plexus, although at risk of injury, isfurther away from clavicle than the vessels in relation tothe medial two-thirds of the bone. A CT arteriogram

demonstrates the brachial plexus less precisely and clearlythan the vessels, and thus was not considered in this study.

A wide variety of plates are currently available whichmay be applied either superiorly or anteriorly for thefixation of clavicular fractures. They pose a varying degreeof risk with regard to inadvertent vascular injury duringplating, depending on where the clavicle is drilled.

The subclavian vessels follow an oblique course alongthe length of the clavicle, beginning posterosuperiorly to itat its medial end and eventually dipping down into theaxilla inferior to it at the lateral end. Upon crossing thelateral border of the first rib the vessels continue onward asthe axillary vessels (Fig. 1).17

Based on the results of this study a simple colour-coded‘traffic light’ system was devised which subdivides the clav-icle into three zones: Red (medial), Amber (middle) andGreen (lateral) to reflect the potential risk for inadvertentvascular injury during screw placement in each zoneaccording to the proximity of the blood vessels.

In the Red (medial) zone the subclavian vessels are situ-ated behind the clavicle. The vein is intimately related to theclavicle, whereas the artery is somewhat protected by theintervening scalenus anterior (Fig. 2). The mean distanceof the vein from the clavicle was 4.77 mm and the artery25.93 mm; in some cases the subclavian vein was in immedi-ate contact with the posterior cortex of the clavicle (Table I).An anteroposterior screw trajectory poses the greatest risk ofinjury to the subclavian vein in this zone (Figs 2 and 8),whereas a craniocaudal screw trajectory, as performed insuperior plating, appears to be a safe option (Fig. 8).

In the Amber (middle) zone the subclavian vessels passlaterally and begin their descent towards the axilla, takingup a position posteroinferior to the clavicle. The vesselseventually cross the lateral border of the first rib, continu-ing onward as the axillary vessels.17 The vessels are nowmuch closer together, with the mean distance of the veinand artery from the clavicle measuring 12.45 mm and17.02 mm, respectively (Table I). Although the rate ofdescent of the vessels gradually increases as they courselaterally, they continue to remain within a posteroinferiorquadrant relative to the clavicle. The vein and artery are ata mean angle of 70° and 50°, respectively, to the horizontalplane of the clavicle (Fig. 5, Table III). In the Amber zone aposteroinferior screw trajectory poses the greatest risk ofinjury (Fig. 9), with the subclavian vein being at highestrisk. Anteroposterior screw placement, as performed inanterior or anterior-inferior plating, appears to be the safestoption for this zone (Fig. 10).

The subclavian vein remains in closest proximity to theclavicle in the Red and Amber zones and is the vitalstructure most likely to be inadvertently injured. At thelateral end of the clavicle (point C) it descends moreacutely within the axilla, thereby placing the artery incloser proximity to the clavicle. The mean distance of theartery and vein to the clavicle at this point was 63.40 mmand 75.67 mm, respectively (Table I). There were also

Thickness at “A” medial

Dis

tan

ce (

mm

)

Thickness at “B” middle

FemaleMale

Thickness at “C” lateral

Length of clavicle “L”

0

2040

60

80

100

120

140160180

Fig. 3

Histogram showing the mean length and thickness of the clavicle atpoints A, B and C between men and women.

NOSNWORB.P ,KIALAHB.V ,AHSRAHEERS.B ,NIWDE.J ,AHNIS.A0521

THE JOURNAL OF BONE AND JOINT SURGERY

significant gender-based differences in measurementstaken at this point, with the mean distances from clavicleto artery and vein being 11.54 mm and 13.85 mm greater,respectively, in men than in women (Table II, Fig. 6). Asboth the artery and vein are at a considerable distancefrom the clavicle at this point, this is referred to as theGreen zone (Fig. 7).

It is important to note that the distances measured areonly accurate when the arm is placed in the neutral oranatomical position. Abduction of the shoulder or the useof reduction forceps are manoeuvres sometimes used dur-ing fracture reduction and can alter the anatomy consid-erably. It is therefore advised that caution be exercisedwhile performing these manoeuvres and drilling or tap-ping the clavicle. Alternatively, the arm should bereturned to its neutral position following fracture reduc-tion prior to drilling.

The images obtained for this study were taken insupine subjects, whereas most operations on the clavicleare performed with the patient in the ‘beach chair’position and sometimes with a sandbag under the scap-ula depending on surgeon preference. This may alter theanatomical relationships in this area, although we arenot aware of any evidence to suggest that this would besignificant.

Most fractures are in the middle and lateral one-third ofthe clavicle. The conclusion from our study is that in thesefractures anterior and anterior-inferior plating is safer thandorsal (Fig. 10).

Table II. Comparison between the mean parameters in men and women. A negative difference means a higher valuein women

ecnereffiDselameFselaMretemaraPp-value (unpaired t-test)

Length of clavicle (mm) 152.84 139.52 13.32 0.49Lateral end of the clavicle

34.040.0-26.3185.31)mm( ssenkcihT30.045.1115.8550.07)mm( yretra fo ecnatsiD20.048.3118.9656.38)mm( niev fo ecnatsiD

Middle of the clavicle90.014.031.5145.51)mm( ssenkcihT16.044.0-02.7167.61)mm( yretra fo ecnatsiD62.045.1-33.3197.11)mm( niev fo ecnatsiD

Medial end of the clavicle47.084.1-74.5299.32)mm( ssenkcihT29.019.0-13.6204.52)mm( yretra fo ecnatsiD54.054.085.430.5)mm( niev fo ecnatsiD

Fig. 4

Sagittal section at the middle (point B) of the clavicle in a three-dimensional reconstructed CT arteriogram (SA, subclavian artery; SV,subclavian vein).

Fig. 5

Illustration showing the subclavian artery (SA) and vein (SV) making anmean angle of 50° and 70° with the horizontal, respectively, in relationto the clavicle at point B (middle one-third).

A RADIOLOGICAL STUDY TO DEFINE SAFE ZONES FOR DRILLING DURING PLATING OF CLAVICLE FRACTURES 1251

VOL. 93-B, No. 8, SEPTEMBER 2011

In fractures of the medial third, the subclavian vein is veryclose to the clavicle, and in some scans was even apposed toposterior cortex (Fig. 2, Table I). Therefore, in this zoneextreme caution should be exercised in dissecting around,manipulating with forceps, drilling or tapping the clavicle.

A craniocaudal direction for drills, taps and screws is lesslikely to damage the vein and therefore favours superior plat-ing, although access is more difficult owing to the patient’shead and neck getting in the way. Anterior plating, which

appears technically easier, will clearly pose a significant riskof damage to the subclavian vein (Figs 2 and 8). If an antero-posterior trajectory cannot be avoided, direct visualisation ofthe vein or using a metallic restraint beyond the posterior cor-tex of the clavicle while drilling or tapping is recommended.

Using reconstructed 3D CT arteriograms, we were able todefine the in vivo anatomical relationship of the clavicle withits adjacent vessels. In the Red zone (medial one-third),anteroposterior screw trajectories are to be avoided becauseof the proximity of the subclavian vein; in the Amber zone(middle one-third) posteroinferior screw trajectories are tobe avoided because of the proximity of both the vein and the

Table III. The mean (range) angles of the artery and vein inrelation to the clavicle at point B

Angle of artery (°) Angle of vein (°)

Mean (range) 49.55 (12 to 80) 70.14 (38 to 100)

Point “A”medial

Vein from “A”

Artery from“A”

Vein from “B”

Artery from“B”

Vein from “C”

Artery from “C”

Dis

tan

ce (

mm

)

Point “B”middle

FemaleMale

Point “C”lateral

0

1020

30

40

5060

70

80

90

Fig. 6

Histogram showing the mean distances of the subclavian artery andvein from the clavicle at points A, B and C between men and women.

Fig. 7

Sagittal section at the lateral end (point C) of the clavicle in a three-dimensional reconstructed CT arteriogram (SA, subclavian artery; SV,subclavian vein).

Fig. 8

Photograph of a model showing the medial end (point A) of the clavicleand the subclavian vessels, with two drills showing anteroposterior andcraniocaudal trajectories (SV, subclavian vein, SA, subclavian artery; BP,brachial plexus).

Fig. 9

Photograph of a model showing a dangerous trajectory represented bya drill at point B (middle one-third). Inset image shows the mean anglesof the subclavian artery (SA) and vein (SV) with the horizontal in relationto the clavicle.

NOSNWORB.P ,KIALAHB.V ,AHSRAHEERS.B ,NIWDE.J ,AHNIS.A2521

THE JOURNAL OF BONE AND JOINT SURGERY

artery; and in the Green zone caution should be taken whenusing craniocaudal trajectories. In the middle and lateralthirds of the clavicle, where the majority of fractures occur,anterior and anterior-inferior plating appears to be safer. Anunderstanding and appreciation of the proximity and rela-tionship of the vessels with respect to the clavicle will hope-fully enable surgeons to further minimise the risk of vascularinjury during plating of the clavicle.

Listen liveListen to the abstract of this article at www.jbjs.org.uk/interactive/audio

The authors are grateful to the Radiology Department and Consultant Radio-logist Dr Hifz-ur-Rahman Aniq at Royal Liverpool and Broadgreen UniversityTeaching Hospitals NHS Trust, and Mr P. Ralte, for their support and help withthis research project.

No benefits in any form have been received or will be received from a com-mercial party related directly or indirectly to the subject of this article.

References1. Nordqvist A, Petersson C. The incidence of fractures of the clavicle. Clin Orthop

1994;300:127–132.2. Crenshaw AH Jr. Fractures of the shoulder, arm and forearm. In: Canale ST, ed.

Campbell's operative orthopaedics. Vol. 3. Tenth ed. St Louis: Mosby, 2003:2985-3069.

3. Craig EV. Fractures of the clavicle. In: Rockwood CA Jr, Buchholz RW, Green DP,Heckman JD, eds. Rockwood and Green's fractures in adults. Vol. 1. Fourth ed. Phila-delphia: Lippincott Williams and Wilkins, 1996:1109-61.

4. Khan LA, Bradnock TJ, Scott C, Robinson CM. Fractures of the clavicle. J BoneJoint Surg [Am] 2009;91-A:447–460.

5. Robinson CM, Court-Brown CM, McQueen MM, Wakefield AE. Estimating therisk of nonunion following nonoperative treatment of a clavicular fracture. J BoneJoint Surg [Am] 2004;86-A:1359–1365.

6. Hill JM, McGuire MH, Crosby LA. Closed treatment of displaced middle-third frac-tures of the clavicle gives poor results. J Bone Joint Surg [Br] 1997;79-B:537–539.

7. McKee MD, Wild LM, Schemitsch EH. Midshaft malunions of the clavicle. J BoneJoint Surg [Am] 2003;85-A:790–797.

8. Zlowodzki M, Zelle BA, Cole PA, Jeray K, McKee MD. Treatment of acute mid-shaft clavicle fractures: systematic review of 2144 fractures: on behalf of the Evi-dence-Based Orthopaedic Trauma Working Group. J Orthop Trauma 2005;19:504–507.

9. Tregaskiss AP, Goodwin AN, Bright LD, Ziegler CH, Acland RD. Three-dimen-sional CT angiography: a new technique for imaging microvascular anatomy. ClinAnat 2007;20:116–123.

10. Lyons FA, Rockwood CA Jr. Migration of pins used in operations on the shoulder.J Bone Joint Surg [Am] 1990;72-A:1262–1267.

11. Norrell H Jr, Llewellyn RC. Migration of a threaded Steinmann pin from an acro-mioclavicular joint into the spinal canal: a case report. J Bone Joint Surg [Am]1965;47-A:1024–1026.

12. Rey-Baltar E, Errazu D. Unusual outcome of Steinman wire: case of fractured clav-icle. Arch Surg 1964;89:1024–1025

13. Glauser F, Kremens V. Unusual sequela following pinning of medial clavicular frac-ture. Am J Roentgenol Radium Ther Nucl Med 1956;76:1066–1069.

14. Johnson B, Thursby P. Subclavian artery injury caused by a screw in a clavicularcompression plate. Cardiovasc Surg 1996;4:414–415.

15. Casselman F, Vanslembroek K, Verougstraete L. An unusual cause of thoracicoutlet syndrome. J Trauma 1997;43:142–143.

16. Shackford SR, Connolly JF. Taming of the screw: a case report and literaturereview of limb-threatening complications after plate osteosynthesis of a clavicularnonunion. J Trauma 2003;55:840–843.

17. Warwick R, Williams PL. The subclavian system of arteries. In: Gray's anatomy.35th ed. Edinburgh: Longmans Green and Co., 1973:1551.

Fig. 10

Illustration showing an anterior plate with safe trajectory and screwplacement at the middle one-third of the clavicle (point B) with the meandistance of the subclavicular vein (SV; 12 mm) and artery (SA; 17 mm).