Osteonecrosis following resurfacing arthroplasty.

Abstract

One of the main concerns regarding resurfacing arthroplasty is the viability of the remaining part of the femoral head, and the postoperative risk of a femoral neck fracture or collapse. In contrast to radiographic methods, positron emission tomography using the radiotracer [18F]-fluoride (Fluoride-PET) enables us to visualize the viability of bone in the remaining part of the head, despite the presence of the covering metal component.
This is preliminary prospective study of 14 patients who underwent an ASR resurfacing arthroplasty. Apart from clinical and radiographic analyses, all patients were analyzed by PET scan 1 week, 4 months, and 1 year after surgery.
1 patient had a minor region of osteonecrosis on PET scan at 1 week and at 4 months. After 1 year, the necrosis had increased to include most of the head. 2 other patients, normal at 4 months, had developed equally large osteonecrosis at 1 year. A fourth patient had a minor osteonecrosis at 1 year. None of the patients had clinical symptoms, and the necrotic areas were not visible on plain radiographs.
We found Fluoride PET to be a sensitive and useful method for evaluation of bone metabolism at resurfacing arthroplasty. 3 of the 14 patients had developed osteonecrosis, involving most of the head at 1 year. The late onset of the phenomenon does not support the hypothesis of surgically damaged vascularity. The presence of this complication together with the lack of visibility on plain radiographs gives reason for concern.

Full text

Page 1

670 Acta Orthopaedica 2009; 80 (6): 670–674
Osteonecrosis following resurfacing arthroplasty
A clinical positron emission tomography study of 14 cases
Gösta Ullmark1, Kent Sundgren1, Jan Milbrink2, Olle Nilsson2, and Jens Sörensen3
1Department of Orthopedics, Gävle Hospital and Center for Research and Development, Uppsala University/County Council of Gävleborg, Gävle;
Departments of 2Orthopedics and 3Nuclear Medicine, Uppsala University Hospital, Uppsala, Sweden
Correspondence: gosta.ullmark@lg.se
Submitted 08-10-22. Accepted 09-04-23
Open Access - This article is distributed under the terms of the Creative Commons Attribution Noncommercial License which permits any noncommercial use,
distribution, and reproduction in any medium, provided the source is credited.
DOI 10.3109/17453670903278258
Background and purpose One of the main concerns regarding
resurfacing arthroplasty is the viability of the remaining part of
the femoral head, and the postoperative risk of a femoral neck
fracture or collapse. In contrast to radiographic methods, posi-
tron emission tomography using the radiotracer [18F]-fluoride
(Fluoride-PET) enables us to visualize the viability of bone in the
remaining part of the head, despite the presence of the covering
metal component.
Patients and methods This is preliminary prospective study
of 14 patients who underwent an ASR resurfacing arthroplasty.
Apart from clinical and radiographic analyses, all patients were
analyzed by PET scan 1 week, 4 months, and 1 year after sur-
gery.
Results 1 patient had a minor region of osteonecrosis on PET
scan at 1 week and at 4 months. After 1 year, the necrosis had
increased to include most of the head. 2 other patients, normal
at 4 months, had developed equally large osteonecrosis at 1 year.
A fourth patient had a minor osteonecrosis at 1 year. None of the
patients had clinical symptoms, and the necrotic areas were not
visible on plain radiographs.
Conclusions We found Fluoride PET to be a sensitive and
useful method for evaluation of bone metabolism at resurfacing
arthroplasty. 3 of the 14 patients had developed osteonecrosis,
involving most of the head at 1 year. The late onset of the phe-
nomenon does not support the hypothesis of surgically damaged
vascularity. The presence of this complication together with the
lack of visibility on plain radiographs gives reason for concern.


Resurfacing arthroplasty of the hip is now widely used. The
historically inferior results of this concept are claimed to
be solved by more precise manufacturing methods of the
implants and more accurate surgical methods. One of the per-
sistent main concerns is the viability of the rest of the femoral
head. During the surgical insertion, capsulotomy, hip dislo-
cation, reaming of the head, and cement pressurization may
damage the blood supply to the remaining part of the head.
The vascular insufficiency might lead to osteonecrosis, which
many months or years later can result in a neck fracture, or
collapse of the construct. There have been some reports in
the literature of late failures with femoral neck fractures after
resurfacing arthroplasty that might support this theory. In a
histological study, Little et al. (2005) described necrosis after
resurfacing arthroplasty. Morlock et al. (2008) studied a large
material of resurfacing revisions by means of histology and
morphology. They found that fractures involving the implant
rim occurred during the first few months, whereas fractures
inside the femoral head often took place closer to a year after
surgery. Campbell et al. (2006) have also analyzed failed sur-
face arthroplasties. They found neck fracture to occur a few
months after surgery, and failure by femoral component loos-
ening often took place a year or two after surgery. They also
discussed the tendency of a thick cement layer to produce
thermal osteonecrosis to a depth of up to 2 mm.
In a peroperative study of resurfacing arthroplasty, Khan et
al. (2007) found diminished blood supply to the femoral head
when using a posterolateral approach rather than a transgluteal
approach. Steffen et al. (2005) analyzed the oxygen tension
inside the femoral head peroperatively in 10 patients having
resurfacing arthroplasty by a posterior approach. The oxygen
tension was found to decline by two-thirds during the proce-
dure and remained so after skin closure. For 3 of the patients,
the tension declined to zero. In another study published in
2007, the same author used this technique to analyze oxygen
tension during resurfacing arthroplasty in 12 patients using
an anterolateral approach. The result was a mean decline in
oxygen tension of 41%.
Fluoride-PET is a sensitive diagnostic method for analysis
of bone metabolism (Grant et al. 2008), such as new bone for-
mation (Sörensen et al. 2003) and bone viability (Ullmark et
al. 2007). Validation studies have been carried out to study the

Page 2

Acta Orthopaedica 2009; 80 (6): 670–674 671
correlation between Fluoride-PET and bone histomorphome-
try (Messa et al. 1993, Piert et al. 2001). In this pilot study, we
analyzed bone metabolism and viability during the first year
after resurfacing arthroplasty using Fluoride PET-CT scans.
Patients and methods
Patients
The study was approved by the local ethics committee (Uppsala
2006:056). The patients received radioprotection information
and gave their informed consent to participate in the study.
14 patients with a primary hip osteoarthritis of a grade and
anatomical shape suitable for resurfacing and without any sys-
temic disease, osteoporosis, cortisone medication, or alcoholic
abuse had an ASR resurfacing arthroplasty (DePuy Johnson &
Johnson, Warsaw, IN). MR was used preoperatively to rule
out any segmental osteonecrosis or cysts of the femoral heads.
One half of the patients had surgery at Gävle Hospital, and
the other half at Uppsala Hospital. Mean age was 52 (32–70)
years, and there were 12 males.
Surgical technique
The surgery was performed by two surgeons (KS and JM)
who were well-accustomed to the surgical method. By using
cement (Palacos cum Gentamycin), the femoral components
were gently cemented in place without any excessive impac-
tion force. The Gävle group of patients had surgery according
to the existing surgical routine, i.e. with a posterior approach,
and an anterolateral approach was used for the Uppsala group.
In both groups, the surgery was performed according to the
recommendations of the manufacturer. Mobilization on the
first day after surgery (walking with crutches) was used for
all patients.
PET analysis
We used a Siemens/CTI Exact HR+ scanner (Siemens/CTI,
Knoxville, TN) for the PET measurements and a hybrid PET
and computerized tomography (CT) device (Discovery ST;
General Electric, Milwaukee, TE). Patients were placed in the
supine position on the camera bed. The legs were stabilized
by a vacuum cushion to reduce motion. A venous catheter was
inserted in an antecubital or dorsal hand vein for injection of
tracer.
40 min after intravenous injection of 150 MBq [18F]-fluo-
ride, a 15-cm section of the body covering the acetabulum and
the intertrochanteric region was scanned in 2D whole-body
mode for 15 min. A 10-min transmission scan for attenua-
tion correction was performed after completing the emission
acquisition. The CT image was co-registered and fused with
the HR+ PET images to indicate exact anatomical locations in
the analysis.
Image processing and analysis
The quantitative emission scans were corrected for attenua-
tion, scatter, and decay and reconstructed by a process of itera-
tive reconstruction. Also non-attenuation corrected emission
scans were reconstructed.
Standardized uptake values (SUVs) from 4 regions of inter-
est (ROIs) were calculated by the formula: SUV of tissue =
activity in tissue (Bq/mL) × body weight (g) / total injected
dose (Bq). Setting average body density to 1 g/mL, this
expression gives a unitless value of the regional tissue activity
in proportion to the average activity per mL of the entire body.
Average values of the contralateral healthy femoral head are
presented as REF in the text.
The 4 ROIs analyzed were located according to Forrest et al.
(2006), except for 1 correction. In summary, the regions were
4 mm high and 10 mm wide: 1 located inside the lateral aspect
of the femoral neck (LFN), and 2 inside the head, lateral to
the stem (LFH) and medial to the stem (MFH). The latter
one was corrected 10 mm towards proximal, medial direc-
tion compared to Forrest et al. in order to be located under
the resurfacing shell inside the head. A fourth ROI (20-30 mm
wide) was located in the proximal femur on a level with the
lesser trochanter. All images were also examined visually for
any photopenic areas outside the ROIs that could represent
osteonecrosis. The non-attenuation corrected images were
evaluated qualitatively to rule out uptake artifacts related to
motion.
Results
The clinical result, including range of movement, was good in
all patients. No one suffered from hip pain.
Radiographic results
Results of preoperative MR analysis were normal in all
patients without any signs of segmental osteonecrosis. All
implants were well-placed and stable and no signs of osteo-
necrosis were seen on the plain radiographs (Figure 1). Het-
erotopic ossification (HO), Brooker group 1–2, was found in
7 of the patients.
PET results
All PET-CT images clearly showed the anatomy of the femoral
head and fluoride uptake (Figure 2). The HO formations were
clearly visible on the PET scans (Figure 3). We found patients
with minor reduction in uptake and 3 with major reduction
in uptake in their femoral heads (1 major reduction being
from the posterior approach group). 3 of the 14 patients (95%
CI: 1–7). 1 patient had a small segment of low uptake in the
medial head-neck region after 1 week. At 4 months, the region
of defective uptake was still present and had increased some-
what in size. After 1 year, the defect had increased further and
included most of the head (Figure 3). This phenomenon was

Page 3

672 Acta Orthopaedica 2009; 80 (6): 670–674
interpreted as progressive osteonecrosis. The other 2 cases
with major defects were normal at 1 week and 4 months but
at 1 year, they had a major region without uptake. The fourth
patient had a minor region of defective uptake at 4 months and
1 year. The intensity of the uptake in those patients had simul-
taneously risen at the edge of viable bone in the necks and
at the tips of the stems. The mean value of quantitative 18F
uptake in reference femoral heads (MFH and LFH) was 1.5
(SD 0.8). The mean uptake values for the operated side in the
11 patients without large defects was 4.2 (1.8), 4.0 (1.2), and
2.2 (0.9) after 1 week, 4 months, and 12 months, respectively.
The corresponding values for the 3 patients with a large defect
were 2.8 (1.6), 2.5 (1.7), and 0.4 (0.9).
Discussion
Until recently, it has not been possible to study bone metabo-
lism under the metal femoral component of a resurfacing
arthroplasty but in contrast to radiographic methods, PET
enables us to visualize the viability of bone in the remaining
part of the head.
The concern for bone viability was substantiated by our
finding that 4/14 patients had regions with no uptake in the
femoral head. The regions of no uptake in these cases did cor-
respond to an absence of viable bone. Subsequently, the bone
might either be resorbed or become necrotic. With bone resorp-
tion, the remaining support for the prosthetic femur compo-
nent is diminished and the construct is at risk of collapse. If,
on the other hand, the low uptake corresponds to absence of
metabolism in the remaining bone, which we believe to be
more likely, the bone is necrotic but may retain its strength for
some years—until the process of ageing of the bone mineral
will result in fragility and risk of fracture. If the necrotic bone
is revascularized, it is at risk of collapse during this process. In
a report by Nelson et al. (1997), osteonecrosis of the femoral
head was treated by resurfacing arthroplasty, and the clinical
results after more than 5 years were assessed as good.
Only 1 of our patients had signs of a minor necrosis as
early as 1 week after surgery, increasing slightly at 4 months
and increasing more obviously at 1 year. The second patient
had a minor necrosis only at 4 months, which was stable at 1
year. 2 more patients had normal metabolism at 1 week and
4 months, but they had developed a large necrosis at 1 year.
Figure 1. Osteonecrotic hip without necrotic signs on plain radiographs,
1 year after surgery.
Figure 2. Scans with normal viable bone 4 months after surgery. Top
panel: CT; middle panel: PET; bottom panel: combined PET-CT.
Figure 3. PET scans of osteonecrosis (red arrows). Top panel: 1 week;
middle panel: 4 months; bottom panel: 1 year after surgery.

Page 4

Acta Orthopaedica 2009; 80 (6): 670–674 673
This finding contrasts with the hypothesis of surgical distur-
bance of the vascular supply being the cause of osteonecrosis.
The mechanism of this course is unclear, but fatigue of the
femoral neck due to altered mechanical conditions from the
arthroplasty might be one cause. In contrast to our findings,
Stephen et al. (2006), using SPECT in 36 cases after resurfac-
ing arthroplasties, concluded that all femoral heads would be
assessed as vascularized 12–47 months after surgery. In a PET
study of 10 resurfacing cases, analyzed 10–33 months after
surgery (Forrest et al. 2006), the viability of the bone under
the femoral components was assessed as good (although the
metabolism in one case had declined to half of the normal in
part of the head).
As expected, we found that metabolism in the femoral head
was greatly increased both 1 week and 4 months after sur-
gery, when the reamed bone tissue of the heads was healing
and rebuilding. In addition, the region of the proximal femur
had a slight increase in metabolism during the postoperative
time period. The individual metabolic responses to the surgi-
cal procedures for the non-necrotic cases had wide variation.
We could not determine any individual patient-related factor
that would explain this variation. The variation in metabolic
response to resurfacing arthroplasty could have an effect on
both the ability of bone to survive surgical procedures, and
possibly also on the degree of postoperative bone ingrowth to
an implant surface, and hence on the survival of any implant.
This field of bone metabolism in response to implants requires
further study. Also, we were impressed by the clarity by which
heterotopic ossification was visualized by PET.
The ROI distribution model presented by Forrest et al.
(2006), and with our modification, has the advantage of pro-
ducing values from constant anatomical regions. There are,
however, regions inside the head not covered by these ROIs,
which must be manually observed during the analysis. In some
of the cases, the contralateral hip (evaluated as the reference
(REF)) was found to have slight osteoarthritis on plain radio-
graphs. To a minor degree, this may raise the SUV values of
the REF.
One matter of debate has been the possible influence of the
surgical approach on the risk of disruption of the circulation
to the femoral head. The posterior approach involves inward
rotation of the flexed hip and section of the entire capsule,
while the anterolateral approach involves an outward rotation
and section of the anterior capsule. There has been evidence
for severely compromised oxygen tension and blood supply to
the femoral head peroperatively using both approaches (Nötzli
et al. 2002, Steffen et al. 2007), with some support for more
profound effects on circulation by the anterolateral approach.
As the major necrotic process took place in the time interval
4–12 months after surgery, the idea of a surgically disturbed
vascularity is not supported by our findings. In addition, osteo-
necrosis occurred with both surgical approaches (2 major in
the anterolateral and 1 major in the posterior), indicating that
the surgical approach is not decisive for this complication.
Trabecular bone has the specific characteristic of withstand-
ing dynamic stress, and of continuously adapting and rebuild-
ing accordingly. This normal stress condition is altered both
through deep cement penetration and by the replacement of
the original corticochondral layer of the head by a rigid metal
layer. There may be a late mechanism resulting in necrosis,
related to the profoundly altered stress conditions for trabecu-
lar bone enclosed in a rigid metal layer.
The metal components preclude any radiological analysis of
the head viability, except for scintigraphic methods.
Since Fluoride-PET analyses are expensive, our study has
the shortcoming of having a small number of patients; thus, it
is hazardous to draw any conclusions regarding the frequency
or risk of osteonecrosis from the finding that 3 of 14 patients
had a major osteonecrosis. However, the fact that we found
3 clear cases is certainly a matter of concern, especially in
the light of previous historical experience. In modern studies,
the Australian Orthopaedic Association (2007) has reported a
4.4% revision rate for resurfacing arthroplasty, Amstutz and
Le Duff (2008) a revision rate of 4.8% at 5 years, and the
Oswestry registry a rate of 4.6% at 7 years (Kahn et al. 2008).
These results contrast with our findings, and may indicate
that the necrotic cases we found might heal—or several years
might elapse before possible clinical failures. The absence
of symptoms and radiographic signs of failure in our study
emphasizes the importance of long-term follow-up and further
Fluoride-PET studies. We will follow these cases and report
their eventual course.
The patients in this study got ASR resurfacing prostheses.
We have no reason to believe that this prosthetic model puts
the viability of bone at any more risk than other modern resur-
facing prostheses.
All authors took part in the planning and design of the study and all revised
the final manuscript. KS and JM performed the operations. GU performed
collection and interpretation of data, and writing of the manuscript. JS per-
formed the PET analyses, data collection, and interpretation. ON was involved
in interpretation of the results.
Amstutz H C, Le Duff M J. Eleven years of experience with Metal-on-metal
hybrid hip resurfacing. J Arthroplasty (Suppl 1) 2008; 6: 36-43.
Australian Orthopaedic Association, National Joint Replacement Registry,
Annual Report 2007. www.aoa.org.au/jointregistry_pub.asp.
Campbell P, Beaulé P E, Ebramzadeh E, LeDuff M, De Smet K, Lu Z, Amstutz
H C. The John Charnley Award: a study of implant failure in metal-on-
metal surface arthroplasties. Clin Orthop 2006; (453): 35-46.
Forrest N, Welch A, Murray A D, Schweiger L, Hutchison J, Ascroft G P.
Femoral head viability after Birmingham hip arthroplasty: Assessment
with use of (18F) fluoride positron emission tomography. J Bone Joint Sur-
gery (Am) (Suppl 3) 2006: 88: 84-9.
Grant F D, Fahey F H, Packard A B, Davis R T, Alavi A, Treves S T. Skeletal
PET with 18F-fluoride: applying new technology to an old tracer. J Nucl
Med 2008; 49 (1): 68-78.

Page 5

674 Acta Orthopaedica 2009; 80 (6): 670–674
Khan A, Yates P, Lovering A, Bannister G C, Spencer R F J. The effect of
surgical approach on blood flow to the femoral head during resurfacing.
Bone Joint Surg (Br) 2007; 89; 21-5.
Khan M, Kuiper J-H, Edwards D, Robinson E, Richardson J B. Birmingham
hip arthroplasty. Five to eight years of prospective multicenter results. J
Arthroplasty 2008 (in press).
Little C P, Ruiz A L, Harding I J, McLardy-Smith P, Gundle R, Murray D W,
Athanasou N A. Osteonecrosis in retrieved femoral heads after failed resur-
facing arthroplasty in the hip. J Bone Joint Surg (Br) 2005; 87: 302-3.
Messa C, Goodman W G, Hoh C K, Choi Y, Nissenson A R, Salusky I B,
Phelps M E, Hawkins R A. Bone metabolic activity measured with posi-
tron emission tomography and (18F)fluoride ion in renal osteodystrophy:
correlation with bone histomorphometry. J Clin Endocrinol Metab 1993;
77 (4): 949-55.
Morlock M M, Bishop N, Zustin J, Hahn M, Rüther W, Amling M. Modes
of implant failure after hip resurfacing: morphological and wear analysis
of 267 retrieval specimens. J Bone Joint Surg (Am) (Suppl 3) 2008; 90:
89-95.
Nelson C L, Walz B H, Gruenwald J M. Resurfacing of only the femoral
head for osteonecrosis. Long-term follow-up study. J Arthroplasty 1997;
12: 736-40.
Nötzli H P, Siebenroch K A, Hempfing A, Ramseier L E, Ganz R. Perfusion
of the femoral head during surgical dislocation of the hip. J Bone Joint Surg
(Br) 2002; 84: 300-4.
Piert M, Tilman T, Zittel G A, Becker M J, Stahlsmidt A, Maier G, Machulla
H-J, Bares R. Assessment of porcine bone metabolism by dynamic 18F-
fluoride ion PET: correlation with bone histomorphometry. J Nucl Med
2001; 42: 1091-100.
Steffen R T, Smith S R, Urban J P G, McLardy-Smith P, Beard D J, Murray
D W. The effect of hip resurfacing on oxygen concentration in the femoral
head. J Bone Joint Surg (Br) 2005; 87: 1468-74.
Stephen J, McMahon S J, Young D, Ballok Z, Badaruddin BS , Larbpaiboon-
pong V, Hawdon G. Vascularity of the femoral head after Birmingham hip
resurfacing. A technetium Tc 99m bone scan/single photon emission com-
puted tomography study. J Arthroplasty 2006; 21: 514-21.
Steffen R T, O’Rourke K, Gill H S, Murray D W. The anterolateral approach
leads to less disruption of the femoral head-neck blood supply than the
posterior approach during hip resurfacing. J Bone Joint Surg (Br) 2007;
89: 1293-8.
Sörensen J, Ullmark G, Långström B, Nilsson O. Rapid bone and blood flow
formation in impacted morcelized allografts. Acta Orthop Scand Award
2003; 74 (6): 633-43.
Ullmark G, Sörensen J, Långström B, Nilsson O. Bone regeneration six years
after impaction bone grafting: A PET analysis. Acta Orthop 2007; 78 (2):
201-6.