Injury, Int. J. Care Injured (2007) 38, 206—211

www.elsevier.com/locate/injury

Percutaneous insertion of a proximal humerallocking plate: An anatomic study§

Jason Smith a, Greg Berry a,*, Yves Laflamme b, Etienne Blain-Pare c,Rudy Reindl a, Edward Harvey a

aDepartment of Orthopedic Surgery, McGill University, CanadabDepartment of Orthopedic Surgery, Universite de Montreal, CanadacMcGill University Medical School, Canada

Accepted 7 August 2006

KEYWORDSMinimally invasive;Proximal humerus;Fracture;Adult;Geriatric;Osteoporotic;Surgery;Axillary nerve

Summary This cadaveric study sought to evaluate the feasibility of applying alocking proximal humerus plate with a novel minimally invasive technique. A uniquepre-contoured locking plate was placed on cadaveric proximal humeri through adescribed minimally invasive approach. Proximity of the plate and screws to theaxillary nerve and their respective surgical tracks were quantified. Safe screw holeplacement with respect to the axillary nerve was determined. Risk of entrapment ofthe nerve beneath the plate was evaluated. Three of the holes near themiddle of thelocking plate consistently intersected the course of the axillary nerve and wereunsafe for percutaneous placement of the screws. The axillary nerve could bepalpated during the course of surgery and easily protected from injury. No entrap-ment of the axillary nerve occurred using this minimally invasive technique. Thescrew-in locking guide cannot be used with this technique as it caused tenting of theaxillary nerve.

Placement of a locking proximal humerus plate via a minimally invasive lateraltrans-deltoid approach is safe if the locking screws are limited to superior andinferior holes. This can be done safely without entrapment of the axillary beneaththe plate. Plate fixation of proximal humerus fractures may now be more desirablewith the use of this approach.# 2006 Elsevier Ltd. All rights reserved.

§ Research conducted at McGill University Health Centre andDepartment of Anatomy, McGill University, Montreal, Canada.* Correspondence to: Department of Orthopaedic Surgery,

Montreal General Hospital, Room B5.159.4, 1650 Cedar Ave.,Montreal, Quebec, Canada H3G 1A4.Tel.: +1 51 49 341 934x42735; fax: +1 51 49 348 453.

E-mail address: greg.berry@muhc.mcgill.ca (G. Berry).

0020–1383/$ — see front matter # 2006 Elsevier Ltd. All rights resedoi:10.1016/j.injury.2006.08.025

Introduction

Although proximal humeral fractures are common,their treatment remains a challenge and an issue ofdebate. These fractures frequently occur in the

rved.

Percutaneous humerus plate 207

elderly osteoporotic population and fixation meth-ods may be compromised by precarious fixation inthe humeral head.4,5,10,13 The SynthesTM (SynthesCanada, Mississauga, Ont., Canada) 3.5 mm LCPproximal humerus plate with locking screws(Fig. 1) has been shown to provide superior fixationin the humeral head compared to conventional plateand screw fixation 15 andmarks a true advance in thetreatment of these difficult fractures. Descriptivetechniques of this procedure use a standard delto-pectoral approach to achieve fracture reduction andapplication of the plate. However, open reductionand internal fixation using plates have a significantlyhigher rate of avascular necrosis.7,14 In particular,damage to the anterior circumflex humeral arteryand its branches will further compromise vascularityof the humeral head. Also, shoulder injuries andsurgery in the elderly have been associated withdecreased range of motion and lower functionalscores which may be due in part to the surgicaltrauma suffered by the soft tissues surroundingthe proximal humerus which occurs with thisapproach.6

In an attempt to address the issues of avascularnecrosis of the humeral head and peri-articular softtissue scarring, we developed a percutaneous

Figure 1 The alphanumerical designation of holes in the3.5 mm 5-hole proximal humerus locking plate. Note thatdue to the obliquity of the second row of holes (B1 and B2)their location on the humerus would differ depending onwhich arm to which the plate was fixed. Hence, they werelabeled in our results with R and L corresponding to rightand left arms.

operative technique to treat proximal humerus frac-tures amenable to indirect reduction. It involvesinserting a locking plate using two incisions throughthe lateral deltoid splitting approach; a 4 cm inci-sion from the lateral acromion to access the tuber-osities and the head and a second 2 cm incision atthe level of the deltoid insertion.

The hypothesis of using a minimally invasiveapproach is that it will reduce the soft tissue dis-section required for an anterior approach to theshoulder and thereby minimise the risk of vascularcompromise and soft tissue scarring associated withit. Prior to employing this technique clinically, how-ever, its feasibility and safety had to be clarified.Although small-incision surgery around the shoulderhas been associated with a better outcome,14 therisk of damaging the axillary nerve needs to becarefully considered as any damage to it may havea significant impact on post-operative shoulderfunction.

The purpose of this study was to define quantita-tively the risk to the axillary nerve using thisapproach by defining anatomically the course ofthe nerve relative to the plate and each of its screwholes. In addition, we sought to ensure that theplate could be consistently applied without entrap-ment of the nerve between the humerus and theplate.

Materials and methods

Twelve upper limbs from six embalmed cadaverswere procured for the study. The edge of the acro-mion was palpated and a direct lateral longitudinalincision was made extending 4 cm distal from thispoint. The skin, subcutaneous tissue, fascia and del-toid muscle were sharply dissected longitudinally toexpose the lateral surface of the proximal humerus.Blunt finger dissection was carried out to lift thedeltoid and axillary nerve off the humeral headandneck. Thiswas best done bybeginning posteriorlywhere the axillary nerve was more readily palpatedand then sweeping laterally and anteriorly. The 5-hole side-plate 3.5 mm Synthes LCP proximalhumerus plate was then inserted in line with thehumerus along its lateral surface and beneath theaxillary nerve through this incision. The nerve washeld raised off the humerus with a finger as the platewas slid beneath it. Two 3.5 mm locking handleswerescrewed into the two most proximal holes to aid inmanipulation of the plate during insertion. Initially,the screw-in guide block (Fig. 2) used in the delto-pectoral approach to ensure accurate screw place-ment was employed. However, it was immediatelyobvious that the bulk of this metal block caused it to

208 J. Smith et al.

Figure 2 (A) The Synthes 3.5 mm 5-hole locking proximal humerus plate with the screw-in screw guide block. (B)Lateral view of the plate with the guide block attached. The guide block must not be used with the percutaneoustechnique as the axillary nerve is placed under tension along its lower border, creating a risk of traction injury.

tent the axillary nerve at its distal edge. For thisreason the guide block was abandoned, to bereplaced with the individual screw-in guides andthese two specimens were eliminated from analysisand replaced. The final position of the plate on thebone was determined by palpation and visual inspec-tion. Fluoroscopy was not used in these cadavercases. The precontoured nature of the plate deter-mined the proximal-distal placement with respect tothe greater tuberosity. The antero-posterior place-ment was adjusted with respect to the bicipitalgroove (slightly lateral to the groove). As no osteot-omy had been created to simulate a fracture, theproximal humeral anatomywasnormal and thus plateplacement was simplified.

Following optimal proximal positioning, the platewas fixed proximally with a K-wire. Via a secondlongitudinal approach at the deltoid insertion (welldistal to the course of the axillary nerve), the platewas aligned with the humeral diaphysis and fixedwith two K-wires, locking it in place.

Once the plate had been definitively positionedand fixed proximally and distally, the skin and del-toid muscle were carefully dissected away from theplate exposing all the holes, the axillary nerve andthe edge of the acromion. Care was taken not todisrupt the soft tissue anterior or posterior to theplate that could potentially disturb the course of theaxillary nerve. The final position of the plate wasreassessed once the soft tissue dissection was com-pleted to ensure it was acceptable.

The holes on the plate were designated alphanu-merically from superior-to-inferior and anterior-to-posterior (Fig. 1). The anterior-posterior designa-tion of the holes allowed consistent numbering withthe designation 1 being anterior and 2 being poster-ior regardless of side of application (left or right).Two independent observers took the following mea-surements with the shoulder in neutral position interms of rotation, flexion and abduction: acromionto superior edge of plate; acromion to axillarynerve, the centre of each hole to closest edge ofnerve. The same research team carried out allinsertions and dissections to ensure consistent pla-cement and measurement.

Results

The average values of the two observers’ measure-ments were calculated and are displayed inTable 1.

The nine superior holes (A1—D) were all situ-ated superior to the axillary nerve with theiraverage distance being at least 10 mm from thenerve (lowest confidence interval 9—13 mm). Theinferior four holes (G to J) sat inferior to theaxillary nerve and never came into contact withit (95% confidence interval 3—7 mm). Of note,three holes in the middle of the plate were con-sistently intersected by the course of the axillarynerve (rows E and F).

Percutaneous humerus plate 209

Table 1 Anatomic measurements with data analysis

Measurement Mean(mm)

Samplesize

Standarddeviation (mm)

Maximum(mm)

Minimum(mm)

Range(mm)

95% Confidenceinterval (mm)

A1 32 12 4 39 27 12 29—34A2 34 12 4 40 28 13 32—36B1R 26 6 4 32 22 10 22—30B1L 21 6 4 27 15 12 16—25B2R 24 6 4 30 19 11 20—28B2L 27 6 4 34 21 14 22—32C1 13 12 4 20 8 12 11—16C2 17 12 4 23 13 11 15—19D 11 12 4 18.5 7 12 9—13E1 1 12 2 6.5 �1 8 0—3E2 2 12 4 10 �1 11 0—5F �5 12 3 0 �9 9 �7—�3G �22 12 4 �14.5 �27 12 �24—�19H �33 12 4 �26.5 �39 12 �36—�31I �45 12 4 �38.5 �51 13 �48—44J �59 12 4 �51.5 �64 13 �61—�56Acromion to nerve 72 12 8 85 62 23 67—78Acromion

to top of plate31 12 7 41 21 21 27—36

The average of the two observers’ measurements were used. Negative numbers represent distances inferior to the axillary nerve andpositive numbers superior.

The average distance from the lateral acromionto the top of the plate was 31 mm (range, 21—41 mm). The average distance from the lateral edgeof the acromion to the axillary nerve was 72 mm(range, 62—85 mm). The plate was situated safelyunder the axillary nerve in all specimens. The nerveappeared grossly normal (and undamaged by sur-gery) over its course around the humeral head andneck in all specimens.

The position of the plate on the proximal humeruswas judged to be acceptable (the contour of theplate matched that of the proximal humerus, allow-ing for good apposition of the plate on bone) in allcases, both prior to and following removal of theoverlying soft tissue.

Discussion

Surgeons have long been cautioned regarding theaxillary nerve when performing surgery around theshoulder and proximal humerus.1,2,11 Because thedeltoid muscle is the main flexor and abductor and amain extender of the arm, loss of innervation to thismuscle can significantly affect patient function. Asecond risk of surgery around the shoulder is damageto the blood supply of the humeral head. Traditionalmethods of plate fixation require extensive softtissue dissection either through a direct deltoidapproach or an anterior deltopectoral approach.According to Sturzenegger, the incidence of avas-cular necrosis was five times greater with plate and

screw fixation because of the stripping involved withthe deltopectoral approach and with the placementof the plates.14 The theoretical advantages of theminimally invasive deltoid-splitting approach forORIF described in this paper include avoidance ofdevascularisation of fracture fragments, minimisa-tion of the risk of injury to the anterior circumflexhumeral artery and reduction of operative morbid-ity by avoiding the extensive soft tissue dissectionentailed in the delto-pectoral approach, This morebiological approach, by limiting the surgical insult,may also allow accelerated rehabilitation, whichhas been shown to improve functional results.6

Twomain concerns of percutaneously inserting theproximal humerus locking plate were addressed inthis study: (1) can it be done without direct damageto the axillary nerve and (2) which holes are safe touse with a ‘‘percutaneous’’ insertion once the platehas been properly positioned. Firstly andmost impor-tantly, the plate was safely positioned beneath theaxillary nerve in all cases and no gross damage to thenerve was seen upon inspection. The nerve was read-ily palpable in all the specimens, ensuring safe andaccurate plate placement.

In terms of which plate holes can be considered‘safe’, measurements were made from the centre ofthe plate holes to the axillary nerve. Using anatomiclandmarks on the humerus to determine the ‘‘safezone’’ for percutaneous pinning has been advocatedin other studies.12 However, due to the possibility ofdeviant humeral anatomy secondary to fracturedisplacement if a less than perfect reduction is

210 J. Smith et al.

obtained, the ability to position the plate with itsgreater tuberosity contour corresponding to thepatient’s own (our endpoint) may be difficult. There-fore, based on this study, we recommend using thepercutaneous insertion of the locking plate andscrews only when anatomic or near anatomic reduc-tion of the fracture is obtained and theposition of theplate confirmed with intraoperative fluoroscopy orradiograph. In the case of non-anatomic reductions,it is crucial to palpate the axillary nerve as it travelsaround the neck of the humerus on the posteriorsurface of the deltoid in order to avoid damaging it.

Once theplatehasbeenproperly positionedon theproximal humerus, our results show that all the rowsexcept E and F can be used (see Fig. 1). Extending themargin of safety by an additional 1 cm still leaves thefirst three rows (representing six potential screws)and last four rows of holes (representing four possiblescrews) available for percutaneous insertion. Thus,only holes in the rows E and F are absolutely contra-indicated for screws inserted in a percutaneous man-ner. If use of these holes is necessary to achievestability of fixation, we suggest direct visualisationof the hole tominimise risk to the axillary nerve. Thiscan be achieved by first ensuring stable fixation usingthe ‘safe’ proximal and distal holes and then abduct-ing the shoulder to deliver these more distal holesinto thewound and performing fixation through themunder direct vision.

The extended anterolateral acromial approach ofthe proximal humerus was recently described byGardner et al. as a minimally invasive approachfor plate fixation.3 This anatomic study has demon-strated the predictability of the position of theaxillary nerve as it crosses the anterior deltoidraphe. Subsequently, their approach was used in16 patients with no axillary nerve injury. Anotherrecent study from Germany reports on locking plateuse for proximal humerus fractures using a smallanterior-lateral-deltoid splitting approach techni-que in 29 patients with a 3-month follow-up. Noapproach-related problems and good short-termresults were observed.8

Although these studies using an anterolateralapproach are different fromour direct lateral deltoidsplitting technique, they do support the concept ofpercutaneous plating of the proximal humerus andthepredictability of theaxillary nerveposition allow-ing it to be isolated and protected during insertion ofthe plate underneath the nerve.3,8

There are a number of limitations with this study.Firstly, twelve specimens may be considered a rela-tively small sample although similar studies haveemployed a similar number of specimens.9 Largeranatomic studies of the proximal humerus haveproduced ranges in anatomic variability in the

course of the axillary nerve which would supportthe use of an extra 1 cm margin of error which wehave employed in this study.1

A second criticism would be the use of an intactmodel rather than a fracture model with an osteot-omy. We sought to evaluate the relationship of theimplants to the axillary nerve in the normal situationas a standard reference point to be considered whenusing the technique in the clinical situation where afracture exists. The aim of achieving anatomic ornear anatomic reduction allows the direct applica-tion of the numbers published here. If this goalcannot be achieved, the surgeon must decide if itis prudent to employ this approach or revert to adeltopectoral approach. The permutations possiblewith regards to fracture fragments (2—4) anddegrees of malreduction made the testing of afracture model impracticable.

A last limitationwas theuseof palpationandvisualinspection to determine plate placement rather thanfluoroscopy as is done clinically. Although this wouldhave been optimal, we did not have access to afluoroscope in the department of anatomy wherethe study was carried out. However, the use of intacthumeri coupled with readily palpable hard and softtissue reference points allowed the accurate place-ment of these precontoured plates on the proximalhumerus in a consistent manner. This was confirmedby assessment of plate position both before andfollowing superficial soft tissue dissection.

Conclusion

This study suggests that proximal humerus fracturescan be safely fixed with the SynthesTM proximalhumerus plate using the percutaneous techniquedescribed, in particular with regards to the axillarynerve. The holes in the centre of the plate areunsafe to use percutaneously as they are consis-tently intersected by the axillary nerve. The screw-in guide block must not be used with this techniqueas it causes tenting of the axillary nerve, potentiallycausing a traction injury to it. Individual screw-inguides are used instead.

The exact clinical role of the approach as well asany associated improvement in outcomes whencompared to more traditional approaches remainsto be determinedwith clinical studies which are nowongoing.

Acknowledgment

Support: Implants were donated by Synthes Canada(Mississauga, Ont., Canada).

Percutaneous humerus plate 211

References

1. Bono CM, Grossman MG, Hochwald N, Tornetta 3rd P. Radialand axillary nerves. Anatomic considerations for humeralfixation. Clin Orthop Relat Res 2000;259—64.

2. Burkhead WZ, Scheinberg RR, Box G. Sugical anatomy of theaxillary nerve. J Shoulder Elbow Surg 1992;31—2.

3. Gardner MJ, Griffith MH, Dines JS, et al. The extendedanterolateral acromial approach allows minimally invasiveaccess to the proximal humerus. Clin Orthop Relat Res2005;123—9.

4. Kannus P, Palvanen M, Niemi S, et al. Osteoporotic fracturesof the proximal humerus in elderly Finnish persons: sharpincrease in 1970—1998 and alarming projections for the newmillennium. Acta Orthop Scand 2000;71:465—70.

5. Koval KJ, Blair B, Takei R, et al. Surgical neck fractures of theproximal humerus: a laboratory evaluation of ten fixationtechniques. J Trauma 1996;40:778—83.

6. Koval KJ, Gallagher MA, Marsicano JG, et al. Functionaloutcome after minimally displaced fractures of the proximalpart of the humerus. J Bone Joint Surg Am 1997;79: 203—7.

7. Kristiansen B, Christensen SW. Plate fixation of proximalhumeral fractures. Acta Orthop Scand 1986;57:320—3.

8. Lill H, Hepp P, Rose T, et al. The angle stable locking-proximal-humerus-plate (LPHP) for proximal humeral

fractures using a small anterior-lateral-deltoid-splitting-approach–—technique and first results. Zentralbl Chir2004;129:43—8.

9. Loomer R, Graham B. Anatomy of the axillary nerve and itsrelation to inferior capsular shift. Clin Orthop Relat Res1989;100—5.

10. Naidu SH, Bixler B, Capo JT, et al. Percutaneous pinning ofproximal humerus fractures: a biomechanical study. Ortho-pedics 1997;20:1073—6.

11. Riemer BL, D’Ambrosia R. The risk of injury to the axillarynerve, artery, and vein from proximal locking screws ofhumeral intramedullary nails. Orthopedics 1992;15: 697—9.

12. Rowles DJ, McGrory JE. Percutaneous pinning of the proximalpart of the humerus. An anatomic study. J Bone Joint Surg Am2001;83-A:1695—9.

13. Ruch DS, Glisson RR, Marr AW, et al. Fixation of three-partproximal humeral fractures: a biomechanical evaluation. JOrthop Trauma 2000;14:36—40.

14. Sturzenegger M, Fornaro E, Jakob RP. Results of surgicaltreatment of multifragmented fractures of the humeralhead. Arch Orthop Trauma Surg 1982;100:249—59.

15. Walsh S, Reindl R, Harvey E, Berry G, Beckman L, Steffen T.Biomechanical comparison of the unique plate versus stan-dard plate for internal fixation of proximal humerus fracturesin a cadaveric model. Podium presentation. Canadian Ortho-pedic Association Annual Meeting; 2004.