Injury, Int. J. Care Injured 41 (2010) 857–861

A computed tomography-based analysis of proximal femoral geometry for lateralimpingement with two types of proximal femoral nail anterotation insubtrochanteric fractures

Vineet Tyagi a, Jae Hyuk Yang b, Kwang Jun Oh a,*a Joint Replacement and Trauma Service, Department of Orthopaedic Surgery, Konkuk University Medical Center, 4-12 Hwayang-dong Gwangjin-gu, Seoul 143-729, Republic of Koreab Department of Orthopaedic Surgery, Seoul Veterans Hospital, Seoul, Republic of Korea

A R T I C L E I N F O

Article history:

Accepted 19 April 2010

Keywords:

Proximal femur

Lateral impingement

Proximal femoral nail anterotation

Subtrochanteric fractures

A B S T R A C T

Objective: To evaluate and analyse the geometrical discrepancies between the proximal femur and two

types of AO/Association for the Study of Internal Fixation (AO/ASIF) Proximal Femoral Nail Anterotation

(PFNA/PFNA-II) using computed tomography (CT)-based analysis in Asian patients, and its implication in

lateral cortical impingement during reduction intra-operatively in subtrochanteric fractures.

Materials and methods: Coronal CT images of hips in 50 randomly selected healthy cases were analysed

using a unique measurement method with respect to the height, diameter, bending angle and inclination

angle of lateral cortex of proximal femur. The data were then compared with dimensions of PFNA and

PFNA-II.

Results: The average height of proximal femur was 61.1 � 5.2 mm, diameter 18.1 � 1.5 mm, bending angle

8.4 � 2.28 and inclination angle of lateral cortex 11.9 � 1.18. The average impingement length of the lateral

cortex was 54.2 � 4.7 mm (range 41.4–64.2 mm), which was shorter than the height of the proximal femur.

On comparison with dimensions of PFNA and PFNA-II, the lateral inclination angle and impingement length

were found to be discrepant in PFNA; however, in the latter the flat lateral surface helps avoiding

impingement with the lateral femoral cortex.

Conclusion: Our study provides clear evidence that the flat lateral shape of PFNA-II is better suited for the

femur of Asian patients by reducing the chances of impingement with the lateral proximal femoral

cortex during intra-operative reduction in subtrochanteric fractures.

� 2010 Elsevier Ltd. All rights reserved.

Contents lists available at ScienceDirect

Injury

journa l homepage: www.e lsevier .com/ locate / in jury

Introduction

The treatment of subtrochanteric fractures has always been achallenging task and since the introduction of the proximalfemoral nails (PFN) for their treatment, peri- or postoperativecomplications related to the surgical implant were continuouslyreported, which led to numerous efforts to improve themorphological design of the nail. PFN’s design has evolved duringthe past decades either to reduce the proximal portion diameter10

or to improve the stress concentration and the rotational stabilityof the distal portion of the nail, thus reducing complications in thefemoral diaphysis.1,2,6,17 The PFNA and PFNA-II (Synthes,Solothurn, Switzerland), recent AO/Association for the Study ofInternal Fixation (AO/ASIF) intramedullary fixation devices,represent a new generation of nails aimed at treating stable andunstable fractures of proximal femur, and have a helical blade

* Corresponding author. Tel.: +82 10 6489 6748; fax: +82 2 2030 7369.

E-mail address: orthokonkuk@gmail.com (K.J. Oh).

0020–1383/$ – see front matter � 2010 Elsevier Ltd. All rights reserved.

doi:10.1016/j.injury.2010.04.018

rather than a screw for better purchase in the femoral head.However, problems with them were also recognised related eitherto mismatch between femoral bowing and the nail geometryduring intra-operative reduction manoeuvres or to the penetrationof helical blade through the femoral head into the hip joint.7,9

There is no evidence-based report available studying the mismatchbetween the proximal femur and nail geometry. In the presentstudy, authors evaluated computed tomography (CT) images fromhealthy cases of Asian ethnicity and retrospectively compared theanatomical measurements of the proximal femur with dimensionsof the currently used PFNA and PFNA-II. In addition, thegeometrical discrepancies were analysed in relation to the shapeof nail and their implication in lateral cortical impingement duringreduction intra-operatively in cases of subtrochanteric fractures.

Materials and methods

The present study was approved by our institution’s scientificresearch board and a written consent was obtained from eachpatient.

V. Tyagi et al. / Injury, Int. J. Care Injured 41 (2010) 857–861858

Patient selection

For establishing the anatomical measurements of the proximalfemur, a randomised selection of 50 cases, who had undergone hipjoint CT examination for pain without any obvious pathology, wasdone. The average age of the cases was 68.5 years (range 60–70)and included 27 males and 23 females. We then compared thesemeasurements with the dimensions of PFNA and PFNA-II.

Unique CT-based measurement method

A coronal CT section showing the largest diameter of theproximal femur was used for making the anatomical measure-ments. Two lines1 extending from the inner side of each medial andlateral cortex of the proximal femur were drawn and another line2

was drawn bisecting the area between these two previous lines;this line represents the femoral shaft axis (Fig. 1(A)). Then, a fourthline was drawn6 at 1308 to line2 that passed through the mostdistal part (or the end) of the cervix of the femur and coincides withthe insertion angle of the lag screw (Fig. 1(B)). A line7 was obtainedby connecting the insertion point of the proximal femur nail, thatis, the tip of the proximal greater trochanter with the crossingpoint made between lines2,6 (Fig. 1(C)). This line acts as a base linefor the centre of the proximal femur while inserting the proximal

Fig. 1. (A) The bisecting line between inner cortical extension lines of medial and lateral

bisecting line make the crossing point. (C) The perpendicular line to the central line betwe

lateral cortex. (D) The points of triangle is composed of greater trochanteric tip which is en

of proximal femur crosses the inner cortical line of medial and lateral cortex. Height of p

plane (a), Inclination angle of lateral cortex (b). (F) The length between inclination point

meets the line perpendicular to the central line passing through the tip of the greater

femoral nail. From the crossing point, a line3 was drawn that wasperpendicular to this base line. At this time, two new crossingpoints are created where this perpendicular line meets the linesdrawn from medial and lateral cortex of the proximal femur(Fig. 1(C)). Thus, we obtained a triangle4 that connects these twonew crossing points and the tip of the greater trochanter (Fig. 1(D)).The base (b) of this triangle corresponds with the diameter of theproximal femur and the height (a) of this triangle corresponds withthe height of the proximal femur. The angle formed by a base line ofthe proximal femoral medullary cavity or the femoral shaft axisand the mid-line of this triangle, was defined as the bending angle(a) for the proximal femur on the coronal section, and the anglemade between the line extending from lateral cortex5 and that ofthe slope of lateral cortex of the proximal femur,8 was defined asthe angle of inclination (b) for the lateral cortex of the proximalfemur and the point where these two extended line meet wasdefined as the inclination/impingement point (Fig. 1(E)). Whenrelated to the implant, the a angle represents the antero-posteriorbending angle of the nail, while the b angle represents the anglebetween the proximal portion and narrow portion of the nail. Aline9 which is perpendicular to the mid-line of the triangle wasdrawn from to the tip of the greater trochanter. Then, a line (c)which is parallel to the mid-line of the triangle was drawn from theimpingement point. This length that causes collision between the

cortex. (B) 1308 neck shaft angle line from inferior margin of femoral neck and the

en greater trochanteric tip and crossing point meets inner cortical line of medial and

try point of proximal nail, the points which the perpendicular line to the central line

roximal femur (a) and diameter of proximal femur (b). (E) Bending angle in coronal

and the point that the line parallel to the central line through the inclination point

trochanter (c).

Table 1Comparison results of morphological dimensions of proximal femur anatomy with PFNA and PFNA-II.

Dimensions Proximal femur

Mean� SD

PFNA PFNA-II

Proximal length (mm) 61.1 mm�5.2 56/61/65 65.1 mm (medial), 79.1 mm (lateral)

Proximal diameter (mm) 18.1 mm�1.5 16.5 mm 16.5 mm

AP bending angle (8) 8.4�2.2 68 58Lateral inclination angle (8) 11.9�1.1 68 Flat lateral shape, gradual reduction in diameter

Lateral impingement length (mm) 54.2 mm�4.7 56/61/65 No impingement (flat lateral surface)

V. Tyagi et al. / Injury, Int. J. Care Injured 41 (2010) 857–861 859

nail and the lateral cortex of femur is defined as the impingementlength (Fig. 1(F)).

The dimensions of PFNA and PFNA-II were then compared withthe anatomical morphology of the proximal femur (Table 1).Diameter of the nails were compared with the base of the trianglethat was obtained by CT, height of the nails were compared withthe height of the triangle, and each of the antero-posterior (AP)bending angle and the inclination angle on the lateral side of thenails were subsequently compared with the bending angle and theinclination angle of the proximal femur on the coronal plane. Allmeasurements were made by one independent observer who didnot participate in any of the operations performed. The femoralneck shaft angles of 1308 for PFNA and PFNA-II were used for thecalculating the average value of measured data.

Results

The average height of proximal femur was 61.1 � 5.2 mm (range51.4–72.5 mm), average diameter was 18.1 � 1.5 mm (range 15.3–21.8 mm), average bending angle (a) in the coronal plane was8.4 � 2.28 (range 4.9–12.58) and the average inclination angle (b) ofthe lateral cortex was 11.9 � 1.18 (range 10.2–13.68) (Table 1). Thedimensions of the PFNA and PFNA-II are summarised in Table 1.Proximal portion diameters were similar for both nails (16.5 mm)while the AP bending angle was 68 for PFNA and 58 for PFNA-II.

A total of 22 cases (44%) had height of their proximal femurshorter than that of the shortest (56 mm) proximal PFNAdimension, and there were eight patients (16%) whose proximalfemur were longer than the longest PFNA dimension (65 mm). Theaverage impingement length of the lateral cortex (c) was54.2 � 4.7 mm (range 41.4–64.2 mm), which was shorter than theproximal dimensions of proximal femur and PFNA both. The PFNA-IIhas different proximal portion lengths on medial and lateral sides,and the lateral side is much longer and has a flat shape. None of thepatients had proximal femur dimension longer than the longestproximal dimensions of PFNA-II (79.1 mm). As for the diameter, threepatients (6%) had smaller proximal femoral diameter than that ofPFNA and PFNA-II (16.5 mm). For the bending angle of the proximalfemur on the coronal plane, 40 patients (80%) had larger angle than 68,the largest AP bending angle of both the nails. Only two patientsshowed smaller angle values. Lastly, for the angle of inclination of thelateral cortex, all the patients (100%) had the angle larger than 68, thelargest angle for the PFNA. The bending angle and angle of inclinationof the lateral cortex of proximal femur were found to have an averagemeasured value of 8.48 and 11.98, respectively, which were greaterthan the corresponding measurements of both the femoral nails.

Discussion

Morphological studies for proximal femur have been done inthe past using three-dimensional methods14; further, measure-ments methods for proximal femur of Asian patients using simpleX-ray or CT scan have been reported.11,16 However, these studiesconsidered proximal femoral geometry as related to hip prosthesisdesign and did not try to relate the intramedullary (IM) implantdimensions with the anatomical measurements. In the present

study, the authors designed a unique CT-based measurementmethod to study the geometry of the proximal femur and thenmatched these measurements with the dimensions of PFNA andPFNA-II, which are the new generation IM implants for treatmentof subtrochanteric fractures.

Complex fractures of the proximal femur involving thesubtrochanteric region are challenging injuries for orthopaedicsurgeons. The difficulties faced in the treatment of these fracturesare related to the anatomic and biomechanical features unique tothis area. Anatomically, the subtrochanteric area consists of mostlycortical bone, which often is comminuted and tends to heal moreslowly than metaphyseal bone. In the proximal part, the canalwidens in the intertrochanteric area, which leads to less optimalfixation with IM devices because of the wide canal and shortsegment proximally. Biomechanically, the subtrochanteric area isan area of high stress concentration, and the muscle attachmentslead to strong deforming forces that can make fracture reductionand maintenance difficult, thus increasing the chances of delayedunion/nonunion.8,19 IM fixation offers mechanical, technical andbiologic advantages over other forms of fixation. Statically lockedIM nails are probably the most commonly used implant in thetreatment of subtrochanteric femoral fractures and considered bymany as the implant of choice.18 Their use has led to acceptableunion rates with decreased blood loss and short operative times.3–

5,8,15

Gamma nail, one of the first generation of the PFNs, has beenreported to cause an impact on the anterolateral cortex of thefemoral diaphysis, resulting in a difficult insertion of the lag screwor a fracture on the lateral cortex due to the morphologicalincompatibility with the diameter and the curvature of the femoraldiaphysis; and thus the nail was modified to the present design.12–

14 However, observations of proximal protrusion above the greattrochanter are often made while using proximal femoral nailwhere a deep enough insertion could not be achieved due to thelocation of the lag screw (Fig. 2). During closed reduction, using C-Arm image intensifier, we observed displacement of proximalfracture fragment as a result of collision between the proximalfemoral nail and the lateral cortex without the nail insertion pointbeing lateralised (Fig. 2) and despite the fact that pointed trocharswere used to maintain anatomic reduction during nail insertion.This phenomenon was especially observed in case of Russell–Taylor type I or simple subtrochanteric fractures. This finding couldnot be explained only by the influence of the strong abductor andflexor acting on the fracture fragments. Thus, there must be someother force such as morphological incompatibility of the nail to theanatomical geometry of the proximal femur, which would have ledto this loss of reduction. However, such displacement of fracturewas not noticed while inserting PFNA-II (Fig. 3), as there occurredno impingement between the lateral femoral cortex and nail due tothe flat lateral shape of PFNA-II.

In our study, there were significant differences between thebending angle of the proximal femur and that of the nail as per thecoronal view, and also between the inclination angle of the lateralcortex to that of the nails. The average bending angle exceeded themaximum bending angle of the nail by 2.40, which might beaccommodated by reaming the cortices. However, the average

Fig. 3. (A) Intra-operative C-arm radiograph shows AO type 31A3 subtrochanteric fracture in a 61-year-old female treated with PFNA-II, a pointed trochar is used to push the

lateral cortex. (B and C) Intra-operative C-arm radiograph shows no impingement of nail with lateral cortex due to flat lateral shape of the nail. (D) Postoperative radiograph

showing satisfactory alignment with no geometrical discrepancies between the nail and proximal femur.

Fig. 2. (A) Preoperative radiograph shows AO type 31A3 subtrochanteric fracture in a 69-year-old male. (B and C) Intra-operative C-arm radiograph shows the impingement of

proximal femoral nail with lateral cortex even though pointed trochar is used to push the lateral cortex, maintaining the reduction. (D) Postoperative radiograph shows

geometrical mismatch between proximal femoral nail and proximal femur with regards to length of nail (a) and impingement of nail (b).

V. Tyagi et al. / Injury, Int. J. Care Injured 41 (2010) 857–861860

lateral cortex inclination angle which is 11.90, is 5.90 greater thanthe maximum inclination angle of the PFNA and contributes todifficulties during insertion and impingement between the nailand the lateral cortex, which could potentially precipitate intra-operative reduction difficulties. However, PFNA-II has a flat lateralsurface, which helps in comfortable sliding of the nail past theimpingement point and thus, no impingement to lateral cortex wasobserved.

Our study showed no significant difference between thediameter of the proximal femur, which presents proper matchwith the proximal diameter of nails. In addition, the PFNA andPFNA-II length was found to be compatible with the averageproximal femoral length, with only 6% patients showing length ofthe proximal femur shorter than that of the nail. There were alsosome patients with greater diameter or height of the proximalfemur compared with the proximal femoral nail, which causeddifficulties in acquiring stable reduction and fixation and also sometroubles in removing the internal fixation materials. However, onthe other hand, this same condition may bring advantages duringthe insertion procedure, and may lessen the impact on the lateralcortex due to impingement. In such cases of shorter length of thenail with deep insertion, various lengths of end caps cancomplement the defect.

There exist difficulties in manufacturing an ideal nail that iscompatible to the femoral geometry of all patients with widerrange of anterolateral curvatures and lateral cortex inclinationangles, and such designs may cause a possibility of weakening thefixation by the decrease in the diameter caused by greater lateralinclination. However, the flat lateral shape of PFNA-II with gradualreduction of diameter distally lessens the impingement chanceswith the lateral femoral cortex and thus solves some of theproblems associated with both the differences in the angle ofinclinations and other geometrical discrepancies. Further, this

modification widens the permissible range of the insertion area atthe apex of the greater trochanter both medially and laterally andthus again decreases the chances of fracture malalignment thatmay be associated with an incorrect entry point.

We used two-dimensional coronal CT images to identify theanatomical mismatch between the proximal femur and nails;however, three-dimensional (3D) CT reconstructions could haveprovided with a better representation of the actual anatomicalconfigurations of proximal femur. Analysing the differencebetween the proximal femoral nail and the proximal femur ontheir length, bending angles and angles of inclination of the lateralcortex was meaningful. Another limitation of this study wasinclusion of a small number of cases. The authors also think thatsome differences in the anatomical configuration definitely existwith respect to gender, age and the race of the population studied.In future multicentre clinical studies, using 3D morphologicalanalysis and multiple demographic variables are required to clarifythe clinical influence of problems associated with morphologicalincompatibility of proximal femoral nails in subtrochantericfractures.

Conclusion

The authors identified morphological incompatibility betweenthe proximal femur and PFNA originating from the difference in thebending angle on the coronal section and the angle of inclination ofthe lateral cortex, which is specially of interest in subtrochantericfractures, and suggested that the flat lateral shape of PFNA-IIlessens the phenomenon of impingement collision between thelateral side of the proximal femoral nail and the lateral cortex ofthe proximal femur occurring as a result of differences between thebending curvature of the proximal femur and the angle ofinclination of the lateral cortex.

V. Tyagi et al. / Injury, Int. J. Care Injured 41 (2010) 857–861 861

Conflict of interest

None declared.

References

1. Al-yassri G, Langstaff RJ, Jones JW, et al. The AO/ASIF proximal femoral nail(PFN) for the treatment of unstable trochanteric femoral fracture. Injury2002;33(5):395–9.

2. Banan H, Al-Sabti A, Jimulia T, et al. The treatment of unstable, extracapsular hipfractures with the AO/ASIF proximal femoral nail(PFN)-our first 60 cases. Injury2002;33:401–5.

3. Barquet A, Francescoli L, Rienzi D, et al. Intertrochanteric–subtrochanteric frac-tures: treatment with the long Gamma nail. J Orthop Trauma 2000;14:324–8.

4. Bellabarba C, Herscovici Jr D, Ricci WM. Percutaneous treatment of peritro-chanteric fractures using the Gamma nail. Clin Orthop 2000;375:30–42.

5. Bess RJ, Jolly SA. Comparison of compression hip screw and gamma nail fortreatment of peritrochanteric fractures. J South Orthop Assoc 1997;6:173–9.

6. Boldin C, Seibert FJ, Fankhauser F, et al. The proximal femoral nail (PFN) – aminimal invasive treatment of unstable proximal femoral fractures: a prospec-tive study of 55 patients with a follow-up of 15 months. Acta Orthop Scand2003;74(1):53–8.

7. Brunner A, Jockel JA, Babst R. The PFNA proximal femur nail in treatment ofunstable proximal femur fractures—3. Cases of postoperative perforation of thehelical blade into the hip joint. J Orthop Trauma 2008;22:731–6.

8. French BG, Tornetta 3rd P. Use of an interlocked cephalomedullary nail forsubtrochanteric fracture stabilization. Clin Orthop 1998;348:95–100.

9. Hwang JH, Oh JK, Han SH, et al. Mismatch between PFNa and medullary canalcausing difficulty in nailing of the pertrochanteric fractures. Arch OrthopTrauma Surg 2008;128:1443–6.

10. Kamboj P, Siwach RC, Kundu ZS, et al. Results of modified proximal femoral nailin peritrochanteric fractures in adult. Internet J Orthop Surg 2007;6(2).

11. Khang G, Choi KW, Kim CS, et al. A study of Korean femoral geometry. ClinOrthop 2003;406:116–22.

12. Leung KS, Chen CM, So WS, et al. Multicenter trial of modified gamma nail ineast Asia. Clin Orthop 1996;323:146–54.

13. Lorich DG, Geller DS, Nielson JH. Osteoporotic pertrochanteric hip fractures:management and current controversies. ICL 2004;53:441–54.

14. Mahaisavariya B, Laupattarakasem W. Cracking of the femoral shaft by thegamma nail. Injury 1992;23:493–500.

15. Sailer R, Ulmer H, Hrubesch R, et al. Surgical stabilization of per- and sub-trochanteric femoral fractures with the gamma nail. Chirurg 2000;71:1380–4[German].

16. Shon WY, Lee SH, Chae DJ, et al. Radiological analysis of the proximal femoralgeometry in Korean adult and the design of the cementless femoral stem. JKorean Orthop 1994;29:577–84.

17. Simmermacher RKJ, Bosch AM, van der Werken C. The AO/ASIF proximalfemoral nail (PFN) a new device for the treatment of unstable proximal femoralfractures. Injury 1999;30:327–32.

18. Sims SH. Subtrochanteric femoral fractures. Orthop Clin N Am 2002;33:113–26.19. Toridis TG. Stress analysis of the femur. J Biomech 1969;2:163–74.