VOL. 83-B, NO. 1, JANUARY 2001 83

K.-H. Koo, MD, Associate ProfessorH.-R. Song, MD, Associate ProfessorJ.-W. Yang, MD, Clinical FellowP. Yang, BS, Research FellowDepartment of Orthopaedic SurgeryJ.-R. Kim, MD, Associate ProfessorDepartment of Preventive Medicine, Gyeong-Sang National UniversitySchool of Medicine, 90 Chilam-dong, Chinju 660-702, South Korea.

Y.-M. Kim, MD, ProfessorDepartment of Orthopaedic Surgery, Seoul National University School ofMedicine, 28 Yeon-Geon-dong, Seoul 110-799, South Korea.

Correspondence should be sent to Dr K.-H. Koo.

©2001 British Editorial Society of Bone and Joint Surgery0301-620X/01/110914 $2.00

Trochanteric rotational osteotomy forosteonecrosis of the femoral headTHE USE OF MRI IN THE SELECTION OF PATIENTS

K.-H. Koo, H.-R. Song, J.-W. Yang, P. Yang, J.-R. Kim, Y.-M. KimFrom the Gyeong-Sang National University School of Medicine, Chinju and SeoulNational University School of Medicine, Seoul, South Korea

The rate of success of transtrochanteric rotationalosteotomy for osteonecrosis of the femoral head

may be improved if patients are preselected usingMRI. We have used three criteria for selection forosteotomy (i) minor collapse of the infarcted area, lossof congruity or the crescent sign, without narrowingof the joint space or acetabular involvement (ii)patients younger than 55 years and with a painful hip(iii) patients with an intact area constituting an arc ofmore than 120° between the central vertical line of thefemoral head and the posterior or anterior margin ofthe necrotic portion as seen on a midsagittal MRI.Seventeen patients were selected, with a follow-up ofmore than 42 months. A bone scan four weeks afteroperation showed adequate perfusion of the proximalsegment in all hips. The hip score of Merle d’Aubignéet al improved from 13.5 points before operation to17.2 points at the latest follow-up. Further collapse ofthe femoral head did not occur.

The use of MRI instead of plain radiographs for theselection of patients has improved the success rate oftranstrochanteric rotational osteotomy.

J Bone Joint Surg [Br] 2001;83-B:83-9.Received 11 January 2000; Accepted 24 February 2000

Osteonecrosis may lead to progressive collapse of thefemoral head followed by degenerative arthritis of thehip.1-4 The condition occurs in young patients, with a meanage of less than 50 years. Replacement arthroplasty in thesecircumstances is associated with high rates of failure5-8 and

there has been a continued search for procedures whichpreserve the hip.9-16

In transtrochanteric rotational osteotomy the necrotic por-tion of the femoral head is displaced from the weight-bearing zone and replaced by healthy cartilage and livingbone. The osteotomy is indicated when at least one-third ofthe articular cartilage and underlying bone is intact. Accurateevaluation of the necrotic area is essential before this techni-cally demanding procedure is adopted. Previous studies haveused plain radiography to select patients and have reportedvariable rates of success,9-11,14-16 but this investigation is notsufficiently accurate to define the extent of infarction. Incor-rect selection of patients based on radiography and inap-propriate operative techniques may give poor results.

MRI can give an accurate determination of the extent ofosteonecrosis17-20 and the rate of success may be improvedby more precise selection of patients. To date, there hasbeen no study which has demonstrated the change in localanatomy and the extent of necrosis after osteotomy usingMRI. We have therefore used MRI to formulate a new setof criteria for selection by which suitable patients forosteotomy may be identified. MR scans have been obtainedafter three years to evaluate the changes in the necrotic areafollowing osteotomy.

Patients and Methods

Patients who had clinical evidence of osteonecrosis of thefemoral head were included in the study and the diagnosiswas established by plain radiography and/or MRI.21 Thehips were graded according to the system of Ficat andArlet.22 From this group, patients who met the followingcriteria were selected for surgery (i) those with minorcollapse of the infarcted area with loss of congruity orcrescent sign without narrowing of the joint space oracetabular involvement (Ficat and Arlet stage 2B) (ii) thoseaged younger than 55 years with a painful hip, and (iii)patients with a healthy area subtending an arc of more than120° between the central vertical line of the femoral headand the posterior or anterior margin of the necrotic portionas seen on the midsagittal MR scan (Fig. 1). The extent ofthe necrotic area affects the fate of the femoral head. Hipswith small areas of necrosis do not collapse even withoutsurgical intervention.1-4,18-20

Our study was not a randomised trial and it lacks acontrol group. It may be argued that some of the patientshave small areas of necrosis which do well without opera-tion. We therefore limited the subjects to those with definitecollapse on plain radiographs. We selected patients lessthan 55 years of age because in young patients total hiparthroplasty is associated with high rates of failure;5-8 inolder patients the results of the procedure are satisfactory.

Between January 1993 and December 1995, 17 patientsmet the predefined selection criteria and were treated bytranstrochanteric rotational osteotomy. There were 15 menand two women with a mean age 30 years (16 to 48).Associated conditions or factors contributing to osteonecro-sis included alcohol abuse in six patients, a history ofsteroid treatment in one and a traumatic dislocation of thehip in two patients. In the remaining eight, no risk factorswere identified.

Transtrochanteric rotational osteotomy is indicated whena segment of the femoral head is viable and intact. Thisarea must be of such a size that, after rotation, the intactportion becomes the weight-bearing surface.9-11 Thereforewe limited subjects to those with a healthy area subtendingan arc greater than 120° (one-third of a circle) on a mid-sagittal MR scan. Patients who participated in this studygave informed consent and the protocol was approved byan institutional review.

Before surgery, we performed MRI using a 1.5 Teslasuperconducting unit (Magnetom; Siemens, Erlangen, Ger-many). The T1-weighted spin-echo images (echo time(TE), 20 ms; repetition time (TR), 500 ms) were obtainedin the coronal and parasagittal planes. The section thicknesswas 0.4 cm with a gap of 0.1 cm. Images were recon-structed using a matrix of 128 � 256 or 256 � 256. Theexcitation number varied between 1 and 4.

When the infarcted area was located in the anteriorhemisphere of the femoral head, the angle of the healthyarea was defined as that which subtended the arc betweenthe central vertical line of the femoral head and the poster-ior margin of the infarct, measured on the midsagittal MRI.When the necrotic area was located in the posterior hemi-sphere, the angle was defined as that which subtended thearc between the central vertical line of the femoral head andthe anterior margin of the infarct (Fig. 1). If this subtendedangle of healthy area exceeded 120°, the patient was selec-ted for the osteotomy.

The volume of the infarct was measured by two inde-pendent radiologists who did not participate in this study.The Siemens Vision Evaluation Package was used to definethe necrotic area in each coronal image. The areas weremultiplied by the distance between the images (imagethickness (0.4 cm) + gap (0.1 cm) = 0.5 cm) to obtain theslice volume of each image. The total volume of thenecrotic area was determined as the sum of the slicevolumes (Fig. 2). The total necrotic volume (mm3) is thenecrotic area multipled by the interval.Operative technique. The procedure of transtrochanteric

rotational osteotomy was performed by one surgeon (K-HK) using a technique which has been described by Sugio-ka9 and by Sugioka et al.10,11

The direction and extent of rotation were determined bythe site of the infarct and the rotational angle on themidsagittal MR scans. If the necrosis was located ante-riorly, the femoral head was rotated anteriorly. The rota-tional angle was defined as the angle between the anteriormargin of the acetabulum and the posterior margin of thenecrotic area in the midsagittal image. If the necrosis layposteriorly, the femoral head was rotated posteriorly. Therotational angle was then defined as the angle between theposterior margin of the acetabulum and the anterior marginof the infarct (Fig. 3). Theoretically, if the proximal seg-

84 K.-H. KOO, H.-R. SONG, J.-W. YANG, P. YANG, J.-R. KIM, Y.-M. KIM

THE JOURNAL OF BONE AND JOINT SURGERY

Fig. 1

Diagram showing the method of measurement of the intact area in theposterior part of the femoral head on the midsagittal section of T1-weighted MR scans. The angle of intact area (I) is defined as the arcbetween the central vertical line of the femoral head and the posteriormargin of the necrotic area (left). In the case of a posterior infarct, theangle is defined as the arc between the central vertical line of the femoralhead and the anterior margin of the necrotic area (right).

Fig 2

Diagram showing the method of measurement of the area of necrosis(shaded) using a cursor to define the necrotic area in each coronal image.The areas are multiplied by the interval between the images (imagethickness (0.4 cm) + gap (0.1 cm) = 0.5 cm) to obtain the volume of eachslice (V1, V2, V3, . . .). The total volume of the necrotic area is the sumof volumes of all the individual slices.

ment is rotated as much as the angle determined on themidsagittal MR scans, the necrotic portion will move out ofthe acetabulum when the hip is extended. In addition torotation, the osteotomy line was inclined at 20° from thatperpendicular to the femoral neck in order to place the hipsin varus angulation as suggested by Sugioka9 and Sugiokaet al10,11 (Fig. 4). To avoid excessive stretching and damageto the medial femoral circumflex vessels, the rotation anglewas limited to less than 90°.

The osteotomy was stabilised using three or four 6.5 mmcancellous screws. Postoperative care consisted of skintraction for six weeks, and the use of crutches whilewalking with protected weight-bearing for another sixweeks.

In order to detect possible injury to the medial femoralcircumflex artery during the operation, we performed scin-

tigraphy at four weeks after the operation and evaluated theblood perfusion of the proximal segment. Patients receivedan injection of 20mCi of 99mTc methylene disphosphonate,and three hours later anterior and posterior radionucleidescans were obtained and the uptake in the proximal seg-ment measured.Clinical evaluation. Every three months, the patients wereassessed clinically using the scoring system of Merle d’Au-bigné et al1 modified by Charnley,23 by two independentobservers who did not participate in this study. Up to sixpoints each were awarded for pain, walking ability, andrange of movement. A rating of 17 to 18 was consideredexcellent, equivalent to a normal hip, 15 to 16 good, 13 to14 fair and less than 13 poor, or a failure. Standardanteroposterior and frog-leg lateral plain radiographs werealso obtained and assessed blindly by two independentobservers who did not know the clinical status of the hips.The contour of the femoral head was evaluated using plainradiographs with a template of concentric circles.24 Radio-logical failure of the operation was defined as subsidence ofthe surface measuring 2 mm or more compared with thecontour of the immediate postoperative radiograph.

The screws were removed three years after the osteot-omy, and follow-up MRI was performed to evaluate thetransposed necrotic area and the newly-formed weight-bearing area of the femoral head beneath the dome of theacetabulum. The location of the necrotic area was com-pared with that demonstrated on the preoperative MR scan.The total volume of the infarct in the follow-up MR scanwas determined as described above. In order to determinewhether the volume of the necrotic area changed after theosteotomy, the volumes in the preoperative MR scans werecompared with those of the follow-up images using theWilcoxon signed-rank test.

Results

Table I gives the details of the patients. The preoperativemean hip score was 13.5 points (11 to 15). The meanfollow-up period was 54 months (42 to 78).

In the 15 hips with an anterior lesion, the proximalsegment was rotated anteriorly (Figs 5a to 5i). In theremaining two hips, with a posterior lesion, the proximalsegment was rotated posteriorly (Figs 6a to 6k). The meanoperating time was 232 minutes (195 to 270), and a meanof 3.2 units of whole blood were transfused.MRI findings. Before operation, in the 15 hips with a non-traumatic aetiology, the necrotic area was located in theanterior aspect of the femoral head, whereas in the remain-ing two with traumatic osteonecrosis (cases 3 and 7), it wasin the posterior aspect. The mean angle of the intactarticular area was 139° (120 to 155). The mean rotationalangle determined on the midsagittal MR scan was 82° (65to 90). The mean volume of the necrotic area was 10.1 cm3

(6.9 to 12.2).At follow-up after three years, the area of necrosis, as

85TROCHANTERIC ROTATIONAL OSTEOTOMY FOR OSTEONECROSIS OF THE FEMORAL HEAD

VOL. 83-B, NO. 1, JANUARY 2001

Fig. 3

Diagram showing the method of measurement of the rotational angle(R).

Fig. 4

Diagram showing the line of the osteotomyat 20° of varus inclination (V).

identified by MRI, had been successfully moved from theweight-bearing region and the volume of the necrotic areain the weight-bearing zone had been significantly reduced.The mean necrotic volume judged by follow-up MRI was9.9 cm3 (5.9 to 12.9), which was similar to that measuredbefore operation. There was no statistically significant netchange.Postoperative scintigraphy. This showed adequate perfu-sion of the proximal segment indicating preservation of themedial femoral circumflex artery. Uptake of radionucleidewas increased at the site of the osteotomy (Figs 5f and6h).

At the most recent clinical follow-up, all patients hadeither excellent (13 patients) or good results (4 patients).The hip score improved to a mean of 17.2 points (16 to 18).There was no case of clinical failure; 12 patients had nopain, and five had mild pain but without limitation ofactivity. The mean arc of flexion at the hip was 113.5° (80to 130). In all patients, the operated limb was shortened bybetween 1 and 1.5 cm because of intentional varus angula-tion of the osteotomy. Five patients (cases 2, 4, 5, 9 and 13)had a slight limp but none required the use of a crutch orcane.Radiological results. Collapse of the femoral head wasarrested in all hips and no patient required replacement hiparthroplasty. There was no radiological evidence of failure.A circumferential osteophyte formed around the femoralhead in two patients (cases 8 and 15) but the joint spacewas maintained and both patients retained a satisfactoryrange of flexion and abduction.

Complications. Delayed union of the osteotomy withexcessive varus occurred in one patient (case 10) who wastreated with a hip spica cast for six weeks. The clinical andradiological result, however, was successful.

Discussion

Osteonecrosis of the femoral head is a destructive diseasethat usually leads to secondary arthritis.1-4 Most patientsare in the third to fifth decade. Although there have beenconsiderable improvements in the materials and techniquesused for total hip arthroplasty, a high rate of failure hasbeen reported in young and active patients.5-8 It is impor-tant therefore to preserve the hip and to obviate or at leastdelay the need for replacement arthroplasty.

Transtrochanteric rotational osteotomy is one of the sur-gical procedures used to preserve the hip.9-11 For thisoperation more than one-third of the femoral head shouldbe intact and healthy. MRI is more accurate than radio-graphy in measuring the extent of the infarct.17-20 We usedMRI instead of radiography for the selection of patients,with excellent results.

Sugioka9 and Sugioka et al10-11 have emphasised twoimportant concepts of the procedure. The first is that it is avascular procedure and involves a rotational flap of theproximal femur based on a vascular pedicle of the medialfemoral circumflex vessels. Osteotomy and rotation willfail if the proximal segment is avascular which is why werecommend that a bone scan be performed after theosteotomy.

86 K.-H. KOO, H.-R. SONG, J.-W. YANG, P. YANG, J.-R. KIM, Y.-M. KIM

THE JOURNAL OF BONE AND JOINT SURGERY

Table I. Details of the 17 patients with osteonecrosis of the femoral head

Necrotic Angle of NecroticHip score at volume on intact volume on Duration of Hip score

Age the initial initial MRI portion on MRI Degree of follow-up MRI follow-up at the latestCase (yr) Gender Hip Aetiology* evaluation† (cm3) (°) rotation (cm3) (mth) follow-up

1 38 M R Id 3/5/6(14) 10.6 140 90 11.2 78 6/6/6(18)

2 27 M L Al 2/5/6(13) 11.4 125 90 12.9 72 6/5/6(17)

3 29 F L Tr 3/5/6(14) 8.3 140 85 9.1 66 6/6/6(18)

4 25 M R Al 2/5/5(12) 11.5 135 90 9.0 60 5/5/6(16)

5 29 M R St 4/5/6(15) 10.9 135 85 12.7 60 6/5/6(17)

6 36 M L Al 3/5/6(14) 9.5 140 85 8.3 60 6/6/6(18)

7 16 M R Tr 2/5/6(13) 7.8 130 90 8.5 54 5/6/5(16)

8 22 M R Id 2/5/6(13) 10.9 155 65 10.2 54 6/6/6(18)

9 36 M R Id 3/5/6(14) 11.7 135 90 10.4 54 6/5/5(16)

10 48 M R Al 2/5/6(13) 9.3 155 60 10.0 48 5/6/6(17)

11 39 M L Al 3/5/6(14) 8.7 145 80 7.7 48 6/6/6(18)

12 24 M L Id 2/5/6(13) 10.3 150 65 12.0 48 5/6/6(17)

13 27 M R Id 3/5/6(14) 9.8 150 85 10.9 42 6/5/6(17)

14 38 F L Id 3/5/6(14) 6.9 120 90 5.9 42 6/6/6(18)

15 22 M R Id 2/5/4(11) 10.7 145 70 8.2 42 5/6/5(16)

16 22 M L Id 3/5/6(14) 12.2 130 90 11.6 42 6/6/6(18)

17 31 M R Al 3/5/6(14) 11.3 140 80 10.5 42 6/6/6(18)

* Id, idiopathic; Al, alcoholic; Tr, traumatic; St, steroid† according to the system of Merle D'Aubigné et al1 as modified by Charnley23; pain/walking ability/range of movement (total)

The second concept concerns the line of the osteotomy.In most patients with osteonecrosis, the infarct is extensiveand located at the anteromedial aspect of the femoral head.In patients with such a lesion, intentional varus positioningshould be considered in addition to adequate rotation inorder to remove the necrotic portion from the weight-bearing area and to achieve a greater area of healthy bonein the weight-bearing zone. Dean and Cabanela15 reportedunsatisfactory results after osteotomy in 21 hips. Theyshowed that the neck-shaft angle had increased in 18 hips.The high rate of failure may have been caused by valgus

rather than varus positioning.In our study, MRI was used to demonstrate changes in

the location and the volume of necrosis after osteotomy.The volume of the infarct did not change and the newlyformed weight-bearing area has remained viable. Thisobservation justifies the procedure.

The staging system of Ficat and Arlet,22 which is oftenused to decide treatment, is based on the radiological stageand clinical symptoms without reference to the extent ofnecrosis. A necrotic index was designed by Koo and Kim.20

Collapse occurred in none of the hips with necrosis of less

87TROCHANTERIC ROTATIONAL OSTEOTOMY FOR OSTEONECROSIS OF THE FEMORAL HEAD

VOL. 83-B, NO. 1, JANUARY 2001

Fig. 5a Fig. 5b Fig. 5c

Fig. 5d Fig. 5e Fig. 5f

Fig. 5g Fig. 5h Fig. 5i

Case 13. A 27-year-old man who had pain in the right hip for four months. Figures 5a and 5b – Anteroposterior (a) and frog-leg (b) radiographs showingosteonecrosis with slight collapse of the femoral head. Figures 5c and 5d – MR scans showing the necrotic area (arrowheads) located in the superioraspect of the femoral head on the midcoronal image (c) and the anterosuperior aspect of the femoral head on the midsagittal image (d). The angle ofintact area in the posterior aspect of the femoral head was 150°. Figure 5e – Postoperative radiograph showing that the proximal segment was rotatedanteriorly with a rotational angle of 85°. Figure 5f – Postoperative scintigraphy showing uptake of radionucleide in the proximal segment. Figures 5gand h – Follow-up midcoronal (g) and midsagittal (h) MR scans taken three years after the osteotomy showing that the necrotic area (arrowheads) hadmoved downwards away from the weight-bearing area. Figure 5i – Radiographs taken at 42 months showing no evidence of further collapse.

88 K.-H. KOO, H.-R. SONG, J.-W. YANG, P. YANG, J.-R. KIM, Y.-M. KIM

THE JOURNAL OF BONE AND JOINT SURGERY

Fig. 6a Fig. 6b Fig. 6c

Fig. 6d Fig. 6e Fig. 6f

Fig. 6g Fig. 6h Fig. 6i

Fig. 6j Fig. 6k

Case 7. A 15-year-old boy who had been injured in a road-traffic accident. Figures 6a to 6d – Radiographs showing a) fracture and dislocation b) afterclosed reduction had been attempted; the epiphysis was separated, c) open reduction and internal fixation and d) one year later when pain had developedwith osteonecrosis and collapse of the femoral head. Figures 6e and 6f – Midcoronal e) and midsagittal f) MR scans showing the necrotic area(arrowheads) located in the superior (e) and anterior (f) areas of the femoral head. The angle of the intact area in the anterior aspect of the femoral headwas 130°. Figure 6g – Postoperative radiograph showing that the proximal segment was rotated anteriorly with a rotational angle of 90°. Figure 6h –Postoperative scintigraph showing uptake of radionucleide in the proximal segment. Figures 6i and 6j– The mid-coronal (i) and midsagittal (j) MR scanstaken three years after the osteotomy (arrowheads) showing that the necrotic area has moved downwards from the weight-bearing area. Figure 6k –Radiograph taken after 54 months showing no evidence of further collapse.

than 30%, in 50% of those with necrosis of 30% to 40%,and in all with necrosis of more than 40%. In hips withnecrosis of less than 30%, surgical intervention is hardlyjustified. In hips with necrosis of 30% to 40%, periodicalobservation is necessary. If there is evidence of collapseduring follow-up, and the arc of intact articular area is morethan 120°, transtrochanteric rotational osteotomy is recom-mended. In hips with necrosis of more than 40%, werecommend transtrochanteric rotational osteotomy evenwithout evidence of collapse (Ficat and Arlet stage 1 and2A) if the arc of the intact articular portion is more than120° (Table II).

The results of transtrochanteric rotational osteotomy canbe improved by criteria based on the use of MRI.

No benefits in any form have been received or will be received from acommercial party related directly or indirectly to the subject of thisarticle.

References1. Merle D’Aubigne R, Postel M, Mazabraud A, Massias P, Gueguen

J. Idiopathic necrosis of the femoral head in adults. J Bone Joint Surg[Br] 1965;47-B:612-33.

2. Kerboul M, Thomine J, Postel M, Merle d’Aubigne R. Theconservative surgical treatment of idiopathic aseptic necrosis of thefemoral head. J Bone Joint Surg [Br] 1974;56-B:291-6.

3. Ohzono K, Saito M, Takaoka K, et al. Natural history of non-traumatic avascular necrosis of the femoral head. J Bone Joint Surg[Br] 1991;73-B:68-72.

4. Koo K-H, Kim R, Ko G-H, et al. Preventing collapse in earlyosteonecrosis of the femoral head: a randomised clinical trial of coredecompression. J Bone Joint Surg [Br] 1995;77-B:870-4.

5. Cornell CN, Salvati EA, Pellicci PM. Long-term follow-up of totalhip replacement in patients with osteonecrosis. Orthop Clin North Am1985;16:757-69.

6. Saito S, Saito M, Nishina T, Ohzono K, Ono K. Long-term results oftotal hip arthroplasty for osteonecrosis of the femoral head: a compar-ison with osteoarthritis. Clin Orthop 1989;244:198-207.

7. Katz RL, Bourne RB, Rorabeck CH, McGee H. Total hip arthro-plasty in patients with avascular necrosis of the hip: follow-up obser-vations on cementless and cemented operations. Clin Orthop1991;281:145-51.

8. Kim YH, Oh JH, Oh SH. Cementless total hip arthroplasty in patientswith osteonecrosis of the femoral head. Clin Orthop1995;320:73-84.

9. Sugioka Y. Transtrochanteric anterior rotational osteotomy of thefemoral head in the treatment of osteonecrosis affecting the hip: a newosteotomy operation. Clin Orthop 1978;130:191-201.

10. Sugioka Y, Katsuki I, Hotokebuchi T. Transtrochanteric rotationalosteotomy of the femoral head for the treatment of osteonecrosis:follow-up statistics. Clin Orthop 1982;169:115-26.

11. Sugioka Y, Hotokebuchi T, Tsutsui H. Transtrochanteric anteriorrotational osteotomy for idiopathic and steroid-induced necrosis of thefemoral head: indications and long-term results. Clin Orthop1992;277:111-20.

12. Yoo MC, Chung DW, Hahn CS. Free vascularised fibula grafting forthe treatment of osteonecrosis of the femoral head. Clin Orthop1992;277:128-38.

13. Urbaniak JR, Coogan PG, Gunneson EB, Nunley JA. Treatment ofosteonecrosis of the femoral head with free vascularized fibulargrafting: a long-term follow-up study of one hundred and three hips. JBone Joint Surg [Am] 1995;77-A:681-94.

14. Masuda T, Matsuno T, Hasegawa I, et al. Results of transtrochan-teric rotational osteotomy for nontraumatic osteonecrosis of the femo-ral head. Clin Orthop 1988;228:69-74.

15. Dean MT, Cabanela ME. Transtrochanteric anterior rotational osteot-omy for avascular necrosis of the femoral head: long-term results. JBone Joint Surg [Br] 1993;75-B:697-601.

16. Langlais F, Fourastier J. Rotational osteotomies for osteonecrosis ofthe femoral head. Clin Orthop 1997;343:110-23.

17. Beltran J, Knight CT, Zuelzer WA, et al. Core decompression foravascular necrosis of the femoral head: correlation between long-termresults and preoperative MR staging. Radiology 1990;175:533-6.

18. Lafforgue P, Dahan E, Chagnaud C, et al. Early-stage avascularnecrosis of the femoral head: MR imaging for prognosis in 31 caseswith at least 2 years of follow-up. Radiology 1993;187:199-204.

19. Sugano N, Ohzono K, Masuhara K, Takaoka K, Ono K. Prognosti-cation of osteonecrosis of the femoral head in patients with systemiclupus erythematosus by magnetic resonance imaging. Clin Orthop1994;305:190-9.

20. Koo K-H, Kim R. Quantifying the extent of osteonecrosis of thefemoral head: a new method using MRI. J Bone Joint Surg [Br]1995;77-B:875-80.

21. Totty WG, Murphy WA, Ganz WI, et al. Magnetic resonanceimaging of the normal and ischemic femoral head. AJR1984;143:1273-80.

22. Ficat RP, Arlet J. Necrosis of the femoral head. In: Hungerford DS,ed. Ischemia and necrosis of bone. Baltimore, etc: Williams & Wilk-ins, 1980:171-82.

23. Charnley J. The long-term results of low-friction arthroplasty of thehip performed as a primary intervention. J Bone Joint Surg [Br]1972;54-B:61-76.

24. Aaron RK, Lennox D, Bunce GE, Ebert T. The conservativetreatment of osteonecrosis of the femoral head: a comparison of coredecompression and pulsing electromagnetic fields. Clin Orthop1989;249:209-18.

89TROCHANTERIC ROTATIONAL OSTEOTOMY FOR OSTEONECROSIS OF THE FEMORAL HEAD

VOL. 83-B, NO. 1, JANUARY 2001

Table II. Treatment protocol according to the extent of necrosis and radiological stage

Index of necrosis (%)20 Ficat and Arlet stage22 Treatment

<30 1 and 2A No treatment

30 to 40 1 and 2A Periodical follow-upcollapse (-): no treatmentcollapse (+) and arc of intact area*>120° : osteotomy†

2B and 3 Arc of intact area* >120° : osteotomy†Arc of intact area* <120° : THA

4 Total hip arthroplasty

>40 1, 2A, 2B and 3 Arc of intact area* >120° : osteotomy†Arc of intact area* <120° : THA

4 THA

* arc of the intact portion on midsagittal MRI† transtrochanteric rotational osteotomy