Treatment of avascular necrosis of the hip by

a non-vascularised cortical graft

S. B. Keizer,
N. B. Kock,
P. D. S. Dijkstra,
A. H. M. Taminiau,
R. G. H. H. Nelissen
From Leiden
University Medical
Centre, Leiden, The
Netherlands

" S. B. Keizer, MD,

Orthopaedic Surgeon
" N. B. Kock, MD, Resident
" P. D. S. Dijkstra, MD, PhD,
Orthopaedic Surgeon
" A. H. M. Taminiau, MD,
PhD, Professor
" R. G. H. H. Nelissen, MD,
PhD, Professor
Department of Orthopaedic
Surgery
Leiden University Medical
Centre, P. O. Box 9600, 2300
RC Leiden, The Netherlands.
Correspondence should be
sent to Dr S. B. Keizer; e-mail:
[email protected]
2006 British Editorial
Society of Bone and
Joint Surgery
doi:10.1302/0301-620X.88B4.
16950 $2.00
J Bone Joint Surg [Br]
2006;88-B:460-6.
Received 11 July 2005;
Accepted after revision
26 October 2005
460

This retrospective study describes the long-term results of core decompression and
placement of a non-vascularised bone graft in the management of avascular necrosis of the
femoral head. We treated 80 hips in 65 patients, 18 by a cortical tibial autograft and 62 by a
fibular allograft. The mean age of the patients was 36 years (SD 13.2). A total of 78 hips were
available for evaluation of which pre-operatively six were Ficat-Arlet stage 0, three stage I,
31 stage IIA, 16 stage IIB, 13 stage III and nine stage IV.
A total of 34 hips (44%) were revised at a mean of four years (SD 3.8). Survivorship
analysis using a clinical end-point showed a survival rate of 59% five years after surgery. We
found a significant difference (p = 0.002) in survivorship, when using a clinical and
radiological end-point, between the two grafts, in favour of the tibial autograft. We
considered this difference to be the result of the better quality and increased volume of
tibial bone compared with that from the trochanteric region used with the fibular allograft.
This is a relatively simple, extra-articular and reproducible procedure. In our view core
decompression, removal of the necrotic tissue and packing of the cancellous grafts into the
core track are vital parts of the procedure.

Avascular necrosis of the femoral head commonly affects patients in the second, third and
fourth decades of life. Prosthetic replacement
of the hip in this group is complicated by the
relatively long expected life span of the patient
and finite life expectancy of the prosthesis.1-3
The optimal treatment requires preservation of
the femoral head or at least to delay its collapse
or the onset of degenerative changes. Avascular
necrosis accounts for between 5% and 12% of
total hip replacements.4
Many conditions, both traumatic or atraumatic, have been associated with avascular
necrosis.5 The natural history of the condition
is one of progression, with subchondral fractures leading to collapse and osteoarthritis.6-8
Mont and Hungerford4 in a meta-analysis
describing the natural history, found that the
femoral head was preserved in between 13%
and 35% of hips depending upon the stage of
the disease.
The non-prosthetic treatment of avascular
necrosis is still controversial. Many options
have been described, including core decompression, vascularised and non-vascularised
bone grafting and various osteotomies.7 The
use of a non-vascularised graft is more appealing than that of a vascularised graft because it
is less technically demanding and may reduce

donor-site morbidity. Non-vascularised autologous bone grafting has numerous other theoretical advantages. The procedure provides
decompression of the avascular lesion and
removal of the necrotic bone in order to interrupt the cycle of ischaemia and interosseous
hypertension. Grafting of the defect with fresh
cancellous bone and placement of a cortical
strut support the subchondral surface and
introduce a scaffold for repair and remodelling
of subchondral bone. It is a relatively simple
procedure which can be performed by one surgeon.
Our study describes the long-term results of
core decompression, popularised by Phemister,9 Boettcher, Bonfiglio and Smith10 and Bonfiglio and Bardenstein,11 combined with placement of an autologous cortical tibial graft or a
fibular allograft. The latter technique was
introduced in order to reduce donor-site morbidity from the tibial graft before we analysed
the different patients.

Patients and Methods

Between 1984 and 2000, we treated 80 hips in
65 patients with avascular necrosis using a cortical strut graft after core decompression. The
diagnosis was based on radiological analysis
and clinical history. From 1984 we used anteroTHE JOURNAL OF BONE AND JOINT SURGERY

TREATMENT OF AVASCULAR NECROSIS OF THE HIP BY A NON-VASCULARISED CORTICAL GRAFT

461

Table I. Details of the 80 hips (65 patients)

Number
of hips

Mean (SD) age of

patients in years

Gender*
(M:F)

Fibula

62

37 (13.1)

Tibia

18

80

Diagnosis

Ficat-Arlet stage

42:20

Steroids 39
Trauma 10
Alcohol 5
Other 8

33 (13.6)

14:4

Steroids 9
Trauma 5
Alcohol 2
Other 2

Mean 36 (13.7)

56:24

80

0
I
IIA
IIB
III
IV
0
I
IIA
IIB
III
IV
80

4
3
28
10
12
5
2
0
4
6
1
5

Mean Merle dAubign

score (SD)
8 (1.8)

7 (1.8)

Mean 8 (1.8)

* M, male; F, female
Other: anabolic steroid use (1), radiotherapy (1), idiopathic disorders with no associated clinical condition (6)
Other: idiopathic disorders with no associated clinical condition (2)

Table II. Modified Ficat-Arlet classification of avascular necrosis of the femoral head12
Stage

Findings on plain radiographs

0
I
IIA
IIB
III
IV

Silent hip, normal radiograph

Normal or, at most, minor changes (subtle loss of clarity, blurring of trabecular pattern, slight patchy osteoporosis)
Diffuse or localised osteoporosis, sclerosis, or cysts of the femoral head
Crescentic subchondral line, segmental flattening of the femoral head (asymmetrical appearance)
Sequestrum, break in articular cartilage from one end of the affected area to the other, normal or increased joint space
Decreased joint space, collapse of the femoral head, acetabular osteoarthritic changes

posterior and lateral radiographs and technetium-99 bone

scans; after 1990, MRI was used to evaluate the extent and
degree of avascular necrosis. In all cases the final diagnosis
was determined by histological examination on intra-operative biopsy. Of the 80 cortical strut grafts, the first 18 were
tibial autografts (1984 to 1990) and, thereafter, 62 fibular
allografts were used. The patients in both groups were similar in all characteristics (Table I). A total of 56 hips in 44
men and 24 hips in 21 women underwent surgery. Bilateral
operations were performed on 12 men and three women.
The mean age at the time of surgery was 36 years (SD 13.2).
Clinical conditions associated with avascular necrosis
included steroid use (48 hips), trauma (15), alcohol abuse
(7), anabolic steroid use (1), radiotherapy (1) and idiopathic disorders (8). Of the 65 patients, nine suffered from
systemic lupus erythematosus, nine had undergone an
organ transplant and two had sarcoidosis. The 15 patients
who underwent bilateral treatment suffered from avascular
necrosis because of steroid use.
All patients had a minimum follow-up of three years.
The mean follow-up was seven years (SD 4.8). All but two
patients were available for follow-up. Of the two who were
not available one died four years after surgery and one
could not be traced. Both patients had been treated by an
autogenous tibial strut graft.
Radiological evaluation. The most widely used classification systems were introduced by Ficat,12 Steinberg, Hayken
and Steinberg,13 and Gardeniers.14 The last two also used
MRI, which was not available for all our patients. ConseVOL. 88-B, No. 4, APRIL 2006

Table III. Clinical evaluation of the hip according to the Merle dAubign
et al classification17
Description

Points

Pain
Intense and permanent
0
Severe even at night
1
Severe when walking, prevents any activity
2
Tolerable with limited activity
3
Mild when walking, disappears with rest
4
Mild and inconstant; normal activity
5
No pain
6
Ability to walk
None
0
Few yards with crutches
1
Only with canes and crutches
2
With one cane less than one hour; very difficult without
3
cane; able to stand
Long time with cane; short time without can and with limp
4
Without cane but with slight limp
5
Normal
6
Functional grade (pain + ability to walk)
Very good
11 or 12
Good
10
Medium
9
Fair
8
Poor
7

quently, we used the Ficat-Arlet classification (Table II)

which uses the anteroposterior and lateral radiographic
appearances of the femoral head and a technetium-99 bone
scan. Pre-operatively, six hips were diagnosed as stage 0,
three as stage I, 32 as stage IIA, 16 as stage IIB, 13 as stage

462

S. B. KEIZER, N. B. KOCK, P. D. S. DIJKSTRA, A. H. M. TAMINIAU, R. G. H. H. NELISSEN

Fig. 1
Fluoroscopic view demonstrating the operative technique.

III and ten as stage IV. The combined necrotic angle measurement of Kerboul et al15 was used to determine the
extent of the avascular area on the pre-operative radiographs. This measurement evaluates the extent of the
necrosis by measuring the arc of involvement of the hip on
the anteroposterior and lateral radiographs. At follow-up
the Ficat-Arlet stage and the position of the autologous
graft or the allograft were judged with respect to the distance between the lesion and the subchondral bone. All
measurements were made by one of the authors (SBK).
Clinical evaluation. We used the two parts of the scale of
Merle dAubign and Postel16 and Merle dAubign et al17
as modified by Charnley,18 for clinical evaluation pre-operatively, one year post-operatively and at the latest followup (Table III).
Operative technique. The patient is placed in a supine position on a fracture table without traction. A straight lateral
approach is used in combination with an image intensifier
in order to locate the avascular lesion. After placing a
guide-wire in the necrotic area, which is located pre-operatively by MRI or radiographs, an 8 mm hole is drilled
through the base of the greater trochanter, along the femoral neck and into the necrotic area of the head (Fig. 1). A
core biopsy is then taken for histological examination.
Curettes are used under fluoroscopic control to remove
necrotic bone from the anterosuperior aspect of the femoral
head and autogenous cancellous bone is impacted into the
excavated area. When a tibial autograft is used, cancellous
bone is harvested from the proximal tibia and with a fibular
allograft the cancellous bone is harvested from the greater
trochanter. After these procedures the tibial or fibular graft

is inserted into the core and its position is checked by image

intensifier.
The position of the graft is considered to be best when its
proximal end is within the necrotic lesion and no more than
10 mm beneath the subchondral bone, as advised by Smith,
Bonfiglio and Montgomery.19 Post-operatively patients are
non-weight-bearing with crutches for six weeks.
All procedures were performed according to this specific
protocol by four different surgeons or residents supervised
by them.
Statistical analysis. For statistical evaluation the statistical
program SPSS 11.0.1 (SPSS Inc, Chicago, Illinois) was used.
A paired Student t-test was used to analyse differences
between groups. The level of significance was set at
p < 0.05. Univariate and multivariate Cox regression were
used to study the influence of various clinical and demographic parameters on the time to failure. A Kaplan-Meier
survival analysis, with a log-rank significance test for group
differences, was also performed. The Spearmans rank correlation coefficient was used to study the correlation
between various clinical and demographic parameters. The
end-point was clinical failure of the structural bone grafting, defined as the need for an operation, or a poor Merle
dAubign and Postel score (< 8 points).16 Clinical success
was defined as a good or excellent score (> 9 points) and no
revision surgery (Table III, functional grade). A second endpoint was radiological failure, as defined by a progression
in the Ficat-Arlet staging during follow-up compared with
pre-operatively. A radiological success was defined as no
radiological progression.

Results
Of the 78 treated hips available for evaluation 34 (44%)
were revised at a mean of four years (SD 3.8, 1 to 15). Revision consisted of total hip arthroplasty in 26 hips, Thomine-cup resurfacing arthroplasty (Protek GmbH, Switzerland) in seven and arthrodesis in one. Eight (10%) patients
reached their end-point because of a poor Merle dAubign
Postel score (< 8) after a mean period of 6.5 years (5 to 13).
Consequently, 42 patients (54%) had clinical failure (secondary surgery or a poor Merle dAubign and Postel score
(< 8 points) at a mean of 4.5 years (1 to 15).
There were complications in seven patients (9%). In the
group with a fibular graft two had a fracture through the
femoral drill hole, and in three perforation of the subchondral plate occurred during surgery. Two complications
occurred in the group receiving a tibial graft; one patient
had a tibial fracture at the donor site and one a lesion of the
peroneal nerve. There was no significant difference in the
complication rates for the two types of graft.
Clinical evaluation. The Merle dAubign and Postel score
improved significantly from a mean of 7.6 (SD 1.84) preoperatively to a mean of 9.4 (SD 2.12; p < 0.001) after one
year. A good or very good functional grade (i.e. > 9 points)
was seen in 41 hips (53%) after one year. At a mean followup of nine years (SD 4.4) the mean Merle dAubign and
THE JOURNAL OF BONE AND JOINT SURGERY

TREATMENT OF AVASCULAR NECROSIS OF THE HIP BY A NON-VASCULARISED CORTICAL GRAFT

463

Table IV. Clinical results according to the Merle dAubign and Postel score (mean SD)
Merle dAubign and Postel score

Ficat-Arlet stage

Pre-operative
number of hips (%)

Pre-op

At 1 year post-op

Number of hips undergoing

revision surgery (%)

Clinical success
at 9 years (%)*

0
I
IIA
IIB
III
IV
Total

6 (8)
3 (4)
32 (40)
16 (20)
13 (17)
10 (12)
80 (100)

9
6
7
7
8
8
8

10
10
10
9
8
9
9

1
3
8
9
7
6
34

4
0
10
4
4
2
24

(3.6)
(0.0)
(1.7)
(1.6)
(1.3)
(1.6)
(1.8)

(1.7)
(1.5)
(1.9)
(2.2)
(2.6)
(2.1)
(2.1)

(17)
(100)
(26)
(56)
(54)
(67)
(44)

(66)
(0)
(31)
(25)
(31)
(20)
(31)

* clinical success defined as the number of hips without revision surgery and a Merle dAubign score > 9 points

Table V. Relationship between clinical and radiological results

Ficat-Arlet
stage

Pre-operative
number of hips (%)

Number of clinical
failures (%)

Number of radiographic
failures* (%)

0
I
IIA
IIB
III
IV
Total

6 (8)
3 (4)
32 (40)
16 (20)
13 (17)
10 (12)
80 (100)

2
3
13
10
8
6
42

6 (100)
3 (100)
23 (74)
12 (75)
8 (67)
NA
52 (67)

(33)
(100)
(42)
(63)
(62)
(67)
(54)

* radiological failure defined as a progression in Ficat-Arlet stage

NA, not available

Postel score for the 44 non-revised hips was 9.2 (SD 2.22),
which was a significant improvement (p < 0.001) compared
with the pre-operative score. Of the non-revised hips 24
(55%) had a good or very good functional score at their
latest follow-up. There was no significant difference
(p = 0.238) between the improvement in the Merle
dAubign and Postel score for the two operative groups
(tibial autograft and fibular allograft).
The clinical success rate (no revision and Merle
dAubign and Postel score > 9 points) at a mean follow-up
of nine years (4 to 17) for Ficat-Arlet stage 0 was 66%, for
stage I it was 0%, for stage IIA 32%, for stage IIB 25%, for
stage III 31% and for stage IV 22% (Table IV).
Pre-operatively, there had been a significant negative correlation (Spearman rank correlation coefficient -0.3, p =
0.012) between the use of steroids and the Ficat-Arlet stage.
No significant correlation was found between the pre-operative Merle dAubign and Postel scores and the Ficat-Arlet
stage, operative technique (tibial autograft, fibular allograft), risk factors, age, gender or bilateral treatment.
Univariate Cox regression analysis showed a significantly higher (p = 0.026) chance of clinical failure in
women than in men (hazard ratio 2.04, 95% confidence
interval (CI) 1.09 to 3.81). Patients older than 30 years had
a significantly higher (p = 0.004) chance of clinical failure
compared with those younger than 30 years of age (multivariate Cox regression hazard ratio 2.83; 95% CI 1.32 to
6.05) with correction for gender.
Radiological evaluation. There was a significant correlation
(Spearman rank correlation 0.3; p = 0.006) between the
pre-operative Ficat-Arlet stage and the need for revision,
VOL. 88-B, No. 4, APRIL 2006

but no correlation between the Kerboul combined necrotic

angle measurement and revision. There was also no correlation between the distance of the graft to the subchondral
bone and revision. The paired t-test showed a significant
difference between the mean Ficat-Arlet stage of 2.6 (SD
0.97) measured pre-operatively and that post-operatively
3.7 (SD 0.77) (p = 0.001). After one year 24 patients (31%)
had an increased Ficat-Arlet stage. At the latest follow-up
25 (57%) of the 44 non-revised hips had an increased FicatArlet stage.
The radiological success rate (no progression in FicatArlet staging) at a mean follow-up of nine years (4 to 17)
for Ficat-Arlet stage 0 was 17.0%, for stage I 0%, for stage
IIA 24%, for stage IIB 25% and for stage III 26%. The hips
which showed a progression in their Ficat-Arlet stage did
not have a higher chance of subsequent clinical failure
(Table V).
Survivorship. A Kaplan-Meier survivorship analysis was
performed with a clinical end-point (subsequent operation
or a Merle dAubign and Postel score < 8 points), a radiological end-point (a progression in the Ficat-Arlet staging)
or revision (total hip arthroplasty, Thomine-cup resurfacing arthroplasty and arthrodesis) as an end-point.
Survivorship analysis with the clinical end-point showed
a mean survival of 59% at five years (95% CI 48 to 70) and
a mean of 44% at ten years (95% CI 31 to 56). There was
a difference in survival between the two operative techniques, although this was not significant (p = 0.114). Treatment by a tibial autograft showed a mean survival rate of
75% at six years (95% CI 54 to 96) and a mean of 63% at
ten years (95% CI 39 to 86). After 14 years, 56% of the

464

S. B. KEIZER, N. B. KOCK, P. D. S. DIJKSTRA, A. H. M. TAMINIAU, R. G. H. H. NELISSEN

Tibial autograft

100
90

Fibular allograft

90
80
Cumulative survival (%)

80
Cumulative survival (%)

Tibial autograft

100

Fibular allograft

70
60
50
40
30

70
60
50
40
30

20

20

10

10

0
1

11

13

15

Survival time (yrs)

100

Cumulative survival (%)

90
80
70
60
50
40
30
20
10
0
5

11

13

15

Fig. 3

Survivorship (clinical end-point) (whiskers denote the SEM).

Survival time (yrs)

Fig. 2

11

13

15

Survival time (yrs)

Fig. 4
Survivorship (revision for any reason as end-point) (whiskers denote the
SEM).

hips treated by a tibial autograft needed revision. Survivorship analysis of the fibular allograft showed a mean survival
rate of 49% at six years (95% CI 36 to 62) and a mean of
38% at ten years (95% CI 24 to 75) (Fig. 2).
Kaplan-Meier survivorship analysis with a clinical and
radiological end-point showed a mean survival rate of 55%
at five years (95% CI 44 to 66) and a mean of 33% at ten
years (95% CI 21 to 44). When a combination of the clinical and radiological end-points was used there was a significant difference (p = 0.002) between the two operative

Survivorship (clinical and radiological end-point) (whiskers denote the

SEM).

techniques. The tibial autograft showed a mean survival

rate of 75% at six years (95% CI 54 to 96) and a mean of
63% at ten years (95% CI 39 to 86). However, the fibular
allograft showed a mean survival rate of 42% at six years
(95% CI 30 to 55) (Fig. 3).
Survivorship analysis with revision surgery for any reason as an end-point showed a mean survival rate of 66%
(95% CI 55 to 77) at five years and of 52% (95% CI 39 to
65) at ten years (Fig. 4). The survivorship analysis of the hip
for the different Ficat-Arlet stages showed a mean survival
rate of 83% (95% CI 68 to 98) for stage 0, 0% for stage I,
80% (95% CI 66 to 94) for stage IIA, 63% (95% CI 39 to
87) for stage IIB, 54% (95% CI 27 to 81) for stage III and
56% (95% CI 23 to 86) for stage IV at five years.
Confounding factors considered as risk factors for avascular necrosis (e.g. alcohol intake, smoking), the pre-operative Merle dAubign scores and the pre-operative Ficat
classification were not related to post-operative collapse of
the femoral head and, consequently, were not related to the
mean survivorship.

Discussion
Our clinical findings differ little from those which have
been previously reported and our clinical data show acceptable results over a follow-up of nine years, although some
discrepancy still remains between various published
reports.20 It is difficult to compare different studies in the
literature because of the high variability in patient populations (e.g. stage of avascular necrosis, age, operative technique, etc). We did not exclude the bilaterally-treated
patients as we did not believe this led to a bias in treatment
THE JOURNAL OF BONE AND JOINT SURGERY

TREATMENT OF AVASCULAR NECROSIS OF THE HIP BY A NON-VASCULARISED CORTICAL GRAFT

since the end-point was revision surgery and a poor clinical

outcome.
The radiological findings were generally less favourable
and in contrast with the findings of Buckley, Gearen and
Petty21 we found a significant difference (p = 0.002) in survivorship (clinical and radiological end-point) between
treatments by tibial autograft and that by a fibular
allograft.
Our treatment was only symptomatic and was not
intended to revascularise the bone. Finite-element analyses
have shown that, ideally, the graft should be placed as close
as possible to the subchondral bone, and in a lateral position.19,21-23 In our study there was no significant correlation
between survival of the hip and the distance of the graft to
the subchondral bone.
In earlier studies non-vascularised bone-grafting procedures had encouraging clinical results.10,21,24,25 At a mean
of eight years (two to 19), Buckley et al21 reported excellent
results in 18 (90%) of 20 hips in which a Ficat and Arlet
stage-I or stage-II lesion had been treated by core decompression combined with tibial autografting and fibular
grafting (both autogenous and allogenic). Boettcher et al10
initially reported success in 27 (71%) of 38 hips six years
after the use of cortical tibial strut grafts. However, in a
long-term study by Smith et al,19 which included the original 38 patients evaluated by Boettcher et al,10 40 (71%) of
56 hips had a poor clinical result after a mean follow-up of
14 years (4 to 27). Later studies showed a high rate of
radiological progression even in hips with a pre-collapse
stage of avascular necrosis. Nelson and Clark26 evaluated
the results of 52 hips treated by Phemister bone grafting
and concluded that the technique was not effective in
arresting progression and that its role in the treatment of
avascular necrosis was uncertain. Dunn and Grow27
reported only four good results in 23 patients treated in this
way.
In a comprehensive review of the literature, Mont and
Hungerford4 reviewed 24 studies involving 1206 hips
which had been treated only by core decompression. They
included studies in which small cancellous grafts had been
packed at the apex of the core track (220 hips). Stratification of the hips which had undergone core decompression
according to the stage of avascular necrosis, gave better
results for the early stages. The mean rate of survival,
judged by clinical success, was 84% for Ficat-Arlet stage I,
65% for stage II and 47% for stage III after a mean followup of 2.5 years. After this, 33% of the hips required further
operation.
Currently, three distinct approaches can be used to insert
a cortical strut graft into the femoral head: 1) a core track;
2) a window in the femoral neck (a light-bulb procedure);
or 3) a trapdoor made through the articular cartilage in the
femoral head. For the past 15 years we have used the strutgrafting procedure through a core track, as first described
by Phemister,9 since it is a relatively simple, extra-articular,
reproducible procedure.
VOL. 88-B, No. 4, APRIL 2006

465

When we first undertook this technique we used a tibial

autograft, as originally described by Phemister.9 However,
of our 16 tibial autografts, two had donor-site complications. In order to reduce this risk we then used a fibular
allograft.
Several groups of surgeons have performed free vascularised fibular grafting procedures in large series of
patients, with high rates of satisfactory results.28-30 However, the complexity of the procedure, its complication rate
and the length of the operation have raised questions
regarding its efficacy. In a series of 103 hips Urbaniak et
al30 reported that 30% required a subsequent operation at
a median of seven years compared with 43% in our series.
Survivorship analysis in their series showed a conversion
rate to a total hip arthroplasty within five years of 89% for
stage IIA (our study 80%), 77% for stage IIB (our study
63%), 71% for stage III (our study 54%) and 73% for
stage IV (our study 56%). In a more recent study by Marciniak, Furey and Shaffer,31 58% of hips required a subsequent operation after eight years. It appears that patients
treated by vascularised fibular grafting have a higher complication rate. Vail and Urbaniak32 found in their study of
247 vascularised fibular grafts, a donor-site morbidity of
35%. Obviously, an advantage of a fibular allograft is an
absence of this.
There is no consensus regarding the indication for nonvascularised bone grafting. Proponents recommend it for
hips with depression of the femoral head of less than 2 mm.
Some investigators have reported good results in patients
with a collapsed femoral head, but only small numbers of
patients have been studied.
Recent advances in our understanding of the pathophysiology of avascular necrosis suggest that a decrease in
the mesenchymal stem-cell pool of the proximal femur may
provide insufficient osteoblasts to meet the needs of bone
remodelling in the early stages of the disease.33 This insufficiency may explain the inadequate repair mechanism,
which may then lead to collapse of the femoral head.
In our view core decompression, removal of the necrotic
tissue and packing of the cancellous grafts into the core
track are the most important parts of the procedure. The
strut graft has a minimal effect on the outcome of the procedure since we found no correlation between the relationship of the graft to the subchondral bone and the
subsequent revision rate. This hypothesis is also supported
by the good results of the light-bulb and bone impaction
grafting techniques.34-37 The effectiveness of these procedures may be related to the availability of stem cells
endowed with osteogenic properties, arising from an
increase in the supply of such cells to the femoral head
through bone-marrow implantation.38,39 Our hypothesis
for the difference in survival between the two grafts, despite
the fact that in both types cancellous autograft is used, is
that the autogenous cancellous bone of the proximal tibia is
not only of better quality, but also that a larger volume can
also be harvested in comparison with the trochanteric

466

S. B. KEIZER, N. B. KOCK, P. D. S. DIJKSTRA, A. H. M. TAMINIAU, R. G. H. H. NELISSEN

region. Furthermore, the large autologous tibial graft itself

has osteogenic properties while allograft only has osteoconductive properties.
This hypothesis is supported by the two studies by Hernigou et al33 and Hernigou and Beaujean.40 They demonstrated decreased activity of bone-marrow cells in the
intertrochanteric region and iliac crest in patients with
AVN related to steroid therapy or alcohol abuse. This could
explain the difference which we found between tibial
autogenous graft and fibular allograft with autologous trochanteric cancellous graft.
In summary, our study demonstrates that the Phemister
technique for the treatment of avascular necrosis is a relatively simple, extra-articular and reproducible procedure.
In our view, core decompression, removal of the necrotic
tissue and packing of the cancellous grafts into the core
track are vital parts of the procedure. However, further
studies are necessary to determine the best cancellous donor
site.
No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article.

References
1. Cornell CN, Salvati EA, Pellici PM. Long-term follow-up of total hip replacement
in patients with osteonecrosis. Orthop Clin North Am 1985;16:757-69.
2. Fyda TM, Callaghan JJ, Olejniczak J, et al. Minimum ten-year follow-up of
cemented total hip replacement in patients with osteonecrosis of the femoral head.
Iowa Orthop J 2002;22:8-19.
3. Steinberg ME, Mont MA. Osteonecrosis. In: Chapman MW, ed. Chapmans orthopedic surgery. Third ed. Philadelphia: Lippincott, Williams & Wilkins, 2001:3263-308.
4. Mont MA, Hungerford DS. Non-traumatic avascular necrosis of the femoral head.
J Bone Joint Surg [Am] 1995;77-A:459-74.
5. Jones JP Jr. Etiology and pathogenesis of ostenecrosis. Semin Arthroplasty 1991;2:
160-8.
6. Hungerford DS, Mont MA. The role of core decompression in the treatment of
osteonecrosis of the femoral head. In: Urbaniak JR, Jones JP, eds. Osteonecrosis: etiology, diagnosis and treatment. Rosemont, IL: Americam Academy of Orthopaedic
Surgeons, 1997:287-92.
7. Mont MA, Hungerford DS. Current concepts review: nontraumatic avascular necrosis of the femoral head. J Bone Joint Surg [Am] 1995;77-A:459-74.
8. Stulberg BN, Davis AW, Bauer TW, Levine M, Easley K. Osteonecrosis of the
femoral head: a prospective randomized treatment protocol. Clin Orthop 1991;268:
140-51.
9. Phemister DB. Treatment of the necrotic head of the femur in adults. J Bone Joint
Surg [Am] 1949;31-A:55-66.
10. Boettcher WG, Bonfiglio M, Smith K. Non-traumatic necrosis of the femoral head.
II: experience in treatment. J Bone Joint Surg [Am] 1970;52-A:322-9.
11. Bonfiglio M, Bardenstein MB. Treatment by bone-grafting of aseptic necrosis of
the femoral head and non-union of the femoral neck (Phemister technique). J Bone
Joint Surg [Am] 1958;40-A:1329-46.
12. Ficat RP. Idiopathic bone necrosis of the femoral head: early diagnosis and treatment. J Bone Joint Surg [Br] 1985;67-B:3-9.
13. Steinberg ME, Hayken GD, Steinberg DR. A new method for evaluation and staging of avascular necrosis of the femoral head. In: Arlet J, Naz, Hungerford DS, eds.
Bone circulation and bone necrosis. Baltimore: Williams & Wilkins, 1984:398-403.

14. Gardeniers JWM. ARCO (Association Research Circulation Osseous) Committee on

Terminology and Classification. ARCO News 1993;5:79-82.
15. Kerboul M, Thomine J, Postel M, Merle dAubign R. The conservative surgical
treatment of idiopathic aseptic necrosis of the femoral head. J Bone Joint Surg [Br]
1974;56-B:291-6.
16. Merle dAubign R, Postel M. Functional results of hip arthroplasty with acrylic
prosthesis. J Bone Joint Surg [Am] 1954;36-A:451-75.
17. Merle dAubign R, Postel M, Mazabraud A, et al. Idiopathic necrosis of the femoral head in adults. J Bone Joint Surg [Br] 1965;47-B:612-33.
18. Charnley J. The long term results of the low friction arthroplasty of the hip performed as a primary intervention. J Bone Joint Surg [Br] 1972;54-B:61-76.
19. Smith KR, Bonfiglio M, Montgomery WJ. Non-traumatic necrosis of the femoral
head treated with tibial bone-grafting: a follow-up note. J Bone Joint Surg [Am] 1980;
62-A:845-7.
20. Mont MA, Jones LC, Sotereanos DG, Amstutz HC, Hungerford DS. Understanding and treating osteonecrosis of the femoral head. Instr Course Lect 2000;49:169-85.
21. Buckley PD, Gearen PF, Petty RW. Structural bone-grafting for early atraumatic
avascular necrosis of the femoral head. J Bone Joint Surg [Am] 1991;73:1357-64.
22. Brown T, Pedersen D, Baker K, Brand R. Mechanical consequence of core drilling
and bone grafting on osteonecrosis of the femoral head. J Orthop Res 1993;11:1358-67.
23. Penix A, Cook S, Skinner H, Weinstein A, Haddad RJ Jr. Femoral head stresses
following cortical bone grafting for aseptic necrosis: a finite element study.
Clin Orthop 1983;173:159-65.
24. Chandler FA. Coronary disease of the hip. J Int Coll Surg 1948;11:34-6.
25. Marcus ND, Enneking WF, Massam RA. The silent hip in idiopathic aseptic necrosis: treatment by bone-grafting. J Bone Joint Surg [Am] 1973;55-A:1351-66.
26. Nelson LM, Clark CR. Efficacy of phemister bone grafting in nontraumatic aseptic
necrosis of the femoral head. J Arthroplasty 1993;8:253-8.
27. Dunn AW, Grow T. Aseptic necrosis of the femoral head: treatment with bone grafts
of doubtful value. Clin Orthop 1977;122:249-54.
28. Malizos KN, Soucacos PN, Beris AE. Osteonecrosis of the femoral head: hip salvaging with implantation of a vascularized fibular graft. Clin Orthop 1995;314:67-75.
29. Sotereanos DG, Plakseychuk AY, Rubash HE. Free vascularized fibula grafting for
the treatment of osteonecrosis of the femoral head. Clin Orthop 1997;344:243-56.
30. Urbaniak JR, Coogan PG, Gunneson EB, Nunley JA. Treatment of osteonecrosis
of the femoral head with free vascularized fibular grafting: a long-term follow-up
study of one hundred and three hips. J Bone Joint Surg [Am] 1995;77-A:681-94.
31. Marciniak D, Furey C, Shaffer JW. Osteonecrosis of the femoral head: a study of 101
hips treated with vascularized fibular grafting. J Bone Joint Surg [Am] 2005;87-A:742-7.
32. Vail TP, Urbaniak JR. Donor-site morbidity with use of vascularised autogenous fibular grafts. J Bone Joint Surg [Am] 1996;78-A:204-11.
33. Hernigou P, Beaujean F, Lambotte JC. Decrease in the mesenchymal stem-cell
pool in the proximal femur in corticosteroid-induced osteonecrosis. J Bone Joint Surg
[Br] 1999;81-B:349-55.
34. Mont MA, Einhorn TA, Sponseller PD, Hungerford DS. The trapdoor procedure
using autogenous cortical and cancellous bone grafts for osteonecrosis of the femoral
head. J Bone Joint Surg [Br] 1998;80-B:56-62.
35. Mont MA, Etienne G, Ragland PS. Outcome of nonvascularized bone grafting for
osteonecrosis of the femoral head. Clin Orthop 2003;417:84-92.
36. Rijnen WHC, Gardeniers JWM, Buma P, et al. Treatment of femoral head
osteonecrosis using bone impaction grafting. Clin Orthop 2003;417:74-83.
37. Rosenwasser MP, Garino JP, Kiernan HA, Michelsen CB. Long term followup of
thorough debridement and cancellous bone grafting of the femoral head for avascular
necrosis. Clin Orthop 1994;306:17-27.
38. Hauzeur JP, Pasteels JL. Pathology of bone marrow distant from the sequestrum in
nontraumatic aseptic necrosis of the femoral head. In: Arlet J, Mazires B, eds. Bone
circulation and bone necrosis. Berlin: Springer-Verlag, 1990:73-6.
39. Inoue A, Ono K. A histological study of idiopathic avascular necrosis of the head of
the femur. J Bone Joint Surg [Br] 1979;61-B:138-43.
40. Hernigou P, Beaujean F. Abnormalities in the bone marrow of the iliac crest in
patients who have osteonecrosis secondary to corticosteroid therapy or alcohol
abuse. J Bone Joint Surg [Am] 1997;79-A:1047-53.

THE JOURNAL OF BONE AND JOINT SURGERY