Precision of the measurements of
periprosthetic bone mineral density in hips
with a custom-made femoral stem
Franz Martini, Carmen Lebherz, Frank Mayer, Ulf Leichtle, Elisabeth
Kremling, Stefan Sell
From the University Hospital of Tübingen, Germany
ur aim was to determine the precision of the
measurements of bone mineral density (BMD) by
dual-energy x-ray absorptiometry in the proximal
femur before and after implantation of an uncemented
implant, with particular regard to the significance of
retro- and prospective studies.
We examined 60 patients to determine the
difference in preoperative BMD between osteoarthritic
and healthy hips. The results showed a preoperative
BMD of the affected hip which was lower by a mean
of 4% and by a maximum of 9% compared with the
opposite side. In addition, measurements were made in
the operated hip before and at ten days after
operation to determine the effect of the implantation
of an uncemented custom-made femoral stem. The
mean increase in the BMD was 8% and the maximum
was 24%. Previous retrospective studies have reported
a marked loss of BMD on the operated side.
The precision of double measurements using a
special foot jig showed a modified coefficient of
variation of 0.6% for the non-operated side in 15
patients and of 0.6% for the operated femur in 20
patients.
The effect of rotation on the precision of the
measurements after implantation of an uncemented
femoral stem was determined in ten explanted femora
and for the operated side in ten patients at 10°
rotation and in 20 patients at 30° rotation. Rotation
within 30° influenced the precision in studies in vivo
and in vitro by a mean of 3% and in single cases in
up to 60%.
O
F. Martini, MD, Registrar
C. Lebherz, Research Student
U. Leichtle, Research Student
S. Sell, PhD, Registrar
Department of Orthopaedics
F. Mayer, PhD, Registrar
Department of Sports Medicine
University Hospital of Tübingen, Hoppe-Seyler-Strasse 3, D-72076 Tübingen, Germany.
E. Kremling, MD, Orthopaedic Surgeon
Department of Hand Surgery, Rhoen-Klinikum, Salzburger Leite 1, D97616 Bad Neustadt a.d.S., Germany.
Correspondence should be sent to Dr F. Martini.
©2000 British Editorial Society of Bone and Joint Surgery
0301-620X/00/79791 $2.00
VOL. 82-B, NO. 7, SEPTEMBER 2000
Precise prediction of the degree of loss of BMD is
thus only possible in prospective cross-sectional
measurements, since the effect of the difference in
preoperative BMD, as well as the apparent increase in
BMD after implantation of an uncemented stem, is not
known from retrospective studies. The DEXA method
is a reliable procedure for determining periprosthetic
BMD when positioning and rotation are strictly
controlled.
J Bone Joint Surg [Br] 2000;82-B:1065-71.
Received 8 January 1999; Accepted after revision 27 August 1999
Measurements of the bone mineral density (BMD) of the
proximal femur after arthroplasty of the hip have been
made for a number of years, mainly in retrospective stud1-5
ies in which a reduction of up to 50% has been observed
compared with the non-operated side. Prospective studies
could not predict a reduction in BMD by this amount. The
maximum reduction of BMD two years after arthroplasty
was 38% compared with the immediate postoperative den6-10
sity of the operated femur.
Direct comparison of retroand prospective studies is not possible since the implant
time of the prostheses is markedly different and it has not
been possible to identify the degree to which the BMD of
both femora differ because of individual differences, and
11
the effect of implantation of an uncemented stem.
We have compared the preoperative with the immediate
postoperative BMD and examined the effect of implantation. The method of measurement and the influence of
rotation in vitro and in vivo after implantation of an
uncemented custom-made femoral stem were compared
with those in standard uncemented stems.
Materials and Methods
Measurements were made by dual-energy x-ray absorptiometry (DEXA) using the Lunar DPX-L instrument (Lunar
Corporation, Madison, Wisconsin) and ‘orthopaedic hip
software’. This was the ‘fast scan mode’ with 66 kV, 3 mA
and a resolution of 0.6 * 1.2 mm. The scans began 1 cm
distal to the tip of the prosthesis and were continued to
3 cm proximal to the tip of the greater trochanter. A bag
equivalent to soft tissue was positioned lateral to the upper
1065
1066
F. MARTINI, C. LEBHERZ, F. MAYER, U. LEICHTLE, E. KREMLING, S. SELL
Fig. 1
Seven regions of interest
(ROIs) accord12
ing to Gruen et al.
thigh to prevent scanning of air which would give a false
measurement.
The software could identify and subtract the metal prosthesis by differentation between soft tissues, bones and the
stem so that medial and lateral cortical bone could be
measured alone. Measurement of the BMD was made using
the ‘orthopaedic hip program’ in the zones of Gruen,
12
McNeice and Amstutz which divide the periprosthetic
bone into distal, lateral and medial regions of interest
2
(ROIs) (Fig. 1). The BMD (g/cm ) of the whole periprosthetic bone was expressed by ROIALL which is the mean of
ROIs 1 to 7.
There were 60 patients, 27 women and 33 men, with a
median age of 56 years (36 to 66). All had unilateral
osteoarthritis of the hip. Patients with systemic disease or
chronic inflammatory joint disease were excluded, as were
those taking medication which influenced bone metabolism. The measurements for precision and of the effect of
rotation could not be performed on all patients of the initial
group, since they did not consent to all the examinations
because of recent surgery and the times required for measurement which were often longer than one hour.
The patients lay in the supine position for the measurements. The leg to be measured was immobilised in a
special foot jig, which allowed continuous fixation between
30° of internal rotation and 30° of external rotation. Thus it
was possible to calculate the external rotation contracture
which is often present before operation and to obtain a
reproducible measurement. The leg was positioned and
fixed in 10° of internal rotation or at the maximum preoperative internal rotation. Further measurements were performed at the identical rotational position to obtain
reproducible results.
The mode ‘scan comparison’ was selected for each hip to
compare the individual measurements so that the size and
position of the ROIs were as nearly identical as possible.
All examinations and evaluations were performed by the
same examiner (CL).
Preoperative difference in BMD. Both femora were
measured before operation in the 60 patients to determine
the influence of unilateral osteoarthritis. Both legs were
fixed at an identical rotation in the foot jig to eliminate the
effect of rotation.
Using the Gruen analysis, we calculated the mean of the
percentage deviation (d%) for each individual of the side to
be operated on (op) compared with the non-affected side
(non-op) using the following equation:
where BMD1 = the non-operated femur and BMD2 = the
operated femur.
Implantation effect on BMD. In order to determine the
effect of implantation on the periprosthetic BMD, measurements were made on the operated femur of the 60 patients
one day before (preop) and ten days after surgery (postop).
An uncemented custom-made femoral stem (Adaptiva;
Endopro Corporation, Dinslaken, Germany) and an uncemented cup (Bivalent; Fehling Corporation, Karlstein,
Germany) with ceramic inlay and head were implanted in
all patients. The mean of the percentage deviation (d%) of
postoperative to preoperative measurements was calculated
using equation 1.
Precision of measurements in the operated and nonoperated femora. In 15 patients (seven women and eight
men) with a median age of 56 years (44 to 64) the nonoperated femur was measured twice within one hour. The
patients were repositioned between measurements. In 20
patients (eight women and 12 men) with a median age of
56 years (44 to 66) the operated femur was also measured
twice with repositioning.
Influence of rotation in vivo. In 20 patients (nine women
and 11 men) with a median age of 55 years (40 to 64)
rotation was measured on the operated femur at 15° internal
rotation (15° INT), 5° internal rotation (5° INT) and 15°
external rotation (15° EXT) three months after implantation
of an Adaptiva stem. In another ten patients (three women
and seven men) with a median age of 58 years (41 to 65)
measurements were made on the operated femur at 5° INT
and 5° EXT according to equation 1.
Influence of rotation in vitro. Ten explanted cadaver
femora were fixed at the distal femur in a special apparatus
which allowed continuous rotation from 60° external to 60°
internal rotation.
In five femora CT was performed and an uncemented
custom-made femoral stem of the Adaptiva type implanted
in each. An uncemented Zweymüller stem (Allopro Corporation, Gelsenkirchen, Germany) was introduced into the
other five femora. The missing soft-tissue mantle was
compensated for by the use of a so-called ‘dry-water
insert’. Measurements were made at 15° INT, neutral posTHE JOURNAL OF BONE AND JOINT SURGERY
PRECISION OF THE MEASUREMENTS OF PERIPROSTHETIC BONE MINERAL DENSITY IN HIPS
itioning (0°) and 15° EXT according to equation 1.
Statistical analysis. We calculated the means of the
percentage differences (d%) between the second (BMD2)
and the first measurement (BMD1) for each individual
according to equation 1.
Most studies use the coefficient of variation (CV%) as
7,10,13,14
In
the measurement of precision of the method.
order to compare the precision of our measurements with
that of other series, we also used this method, but called it
a modified coefficient of variation (mCV%), since, to be
mathematically correct, the CV% is calculated from the
standard deviation and the mean. The mCV% was calculated according to the following equation:
1067
2
Table I. Mean (± SD) preoperative values for the BMD (g/cm ) in
operated (op) and non-operated (non-op) femora for all the ROIs
Non-op
Op
Mean
percentage
deviation
(d %)
ROI 1
0.870 (0.174)
0.799 (0.202)
-8.6
<0.0001
ROI 2
1.440 (0.231)
1.375 (0.271)
-4.6
0.0008
ROI 3
1.868 (0.261)
1.805 (0.312)
-3.5
0.0030
ROI 4
1.938 (0.267)
1.857 (0.287)
-4.1
<0.0001
ROI 5
1.892 (0.229)
1.825 (0.284)
-3.7
0.0006
ROI 6
1.475 (0.229)
1.370 (0.275)
-7.2
<0.0001
ROI 7
1.187 (0.217)
1.205 (0.302)
+1.6
0.5239
ROIALL*
1.524 (0.197)
1.462 (0.240)
-4.3
<0.0002
p value
* mean values of ROI 1 to 7
where d = individual difference in double measurements,
n = number of patients and m1 and m2 = the mean value
of the sum of the first and second measurement.
Statistical analysis was made by the paired t-test after
determining the normal distribution for all groups. The
level of significance was set at p < 0.05. We used the
statistics program JMP for Windows (version 3.1.6.2; SAS
Inc, Cary, North Carolina).
Results
Preoperative difference in BMD. There was a significantly lower preoperative BMD in the affected femur compared with the opposite side, ranging from -3.5% to -8.6%
(Table I). A higher BMD of +1.6% was observed only in
the region of the calcar (ROI 7).
Implantation effect on BMD. After implantation of the
femoral stem, the BMD increased from +5.2% (ROI 5) to
+24.2% (ROI 1) with a mean of ROIALL of +7.7% (Table
II). ROI 4, used as a reference value, shows an increase in
BMD of only +1.6%, which is within the range of error of
the measurement. If the medial and lateral periprosthetic
regions (ROISTEM) only are considered the increase in
BMD is +9% (Table II).
Precision of measurements of the operated and nonoperated femur. Double measurements show an mCV%
for the non-operated side of 0.9% (ROI 5) to 1.6% (ROI 1
and 7) (Table III). The operated side shows an mCV% of
0.8% (ROI 4) to 3.3% (ROI 1) (Table IV). The mCV% for
ROIALL shows a value of 0.6% for both femora (Tables III
and IV).
Influence of rotation in vivo. The results of the measurements in vivo (15° INT, 5° INT, 15° EXT) likewise vary
with rotation from -10.5% to +2.8%, with a mean of -2.8%
(Table V). Small changes in rotation can thus cause differences in BMD of more than 10%, and in individual cases
up to 60% (data not shown). The measurements at 5° INT
and 5° EXT show similar results, with mean differences of
-5.1% to +2.1%, with up to 23% in individual cases (Table
VI).
VOL. 82-B, NO. 7, SEPTEMBER 2000
2
Table II. Mean (± SD) values for the BMD (g/cm ) pre- (preop) and
postoperatively (postop) for all the ROIs
Preop
Postop
Mean
percentage
deviation
(d %)
ROI 1
0.743 (0.197)
0.887 (0.199)
+24.2
<0.0001
ROI 2
1.535 (0.282)
1.685 (0.313)
+10.6
<0.0001
ROI 3
1.852 (0.284)
1.965 (0.248)
+6.7
<0.0001
ROI 4
1.897 (0.267)
1.924 (0.269)
+1.6
0.0893
ROI 5
1.844 (0.259)
1.931 (0.229)
+5.2
<0.0001
ROI 6
1.502 (0.281)
1.612 (0.301)
+7.6
<0.0001
ROI 7
1.367 (0.309)
1.511 (0.297)
+12.7
<0.0001
ROISTEM*
ROIALL†
1.474 (0.227)
1.534 (0.226)
1.598 (0.215)
1.645 (0.210)
+9.0
+7.7
<0.0001
<0.0001
p value
* mean values of ROI 1 to 3 and 5 to 7
† mean values of ROI 1 to 7
Influence of rotation in vitro. The percentage difference
in BMD in the explanted cadaver femora at different
positions of rotation varies considerably from -8.2% to
+10.4% and up to 64% in ROI 7 in some cases (Table VII).
On average, allowance should be made for differences of
-2.9% with changing positions of rotation.
Comparison of the two types of prosthesis used shows
differences of -3.3% on average for the custom-made stem
(Table VIII) and of -2.5% for the uncemented stem (Table
IX).
Discussion
Retrospective studies have reported considerable loss of
BMD after implantation of an uncemented total hip arthro1-5
plasty compared with the opposite side. These results
have often been interpreted without taking into account
possible errors of measurement and changes related to
5
positioning. According to Niinimäki and Jalovaara, for
example, the precision was reduced by 0.6% if two experienced examiners analysed the measurements. The fact that
1068
F. MARTINI, C. LEBHERZ, F. MAYER, U. LEICHTLE, E. KREMLING, S. SELL
Table III. Mean (±
femur in all ROIs
SD)
2
values for the BMD (g/cm ) for double measurements (m1 and m2) for the non-operated
Mean
percentage
deviation
(d %)
m2
m1
p value
Modified
coefficient of
variation
(mCV%)
ROI 1
0.936 (0.208)
0.937 (0.200)
+0.7
0.8971
1.6
ROI 2
1.570 (0.281)
1.558 (0.285)
-0.8
0.1498
1.4
ROI 3
1.961 (0.300)
1.953 (0.304)
-0.4
0.3865
1.2
ROI 4
2.040 (0.293)
2.021 (0.286)
-0.8
0.0537
1.3
ROI 5
1.969 (0.265)
1.982 (0.265)
+0.7
0.0609
0.9
ROI 6
1.559 (0.287)
1.553 (0.282)
-0.3
0.4380
1.2
ROI 7
1.253 (0.253)
1.260 (0.243)
+0.8
0.3413
1.6
ROIALL*
1.612 (0.254)
1.609 (0.249)
-0.1
0.3423
0.6
* mean values of ROI 1 to 7
Table IV. Mean (±
in all ROIs
SD)
2
value for the BMD (g/cm ) for double measurements (m1 and m2) for the operated femur
m1
m2
Mean
percentage
deviation
(d %)
p value
Modified
coefficient of
variation
(mCV%)
ROI 1
0.895 (0.185)
0.891 (0.186)
-0.3
0.6879
3.3
ROI 2
1.619 (0.365)
1.619 (0.366)
+0.2
0.9577
1.6
ROI 3
1.939 (0.262)
1.938 (0.255)
+0.04
0.9699
1.5
ROI 4
1.957 (0.291)
1.953 (0.299)
-0.2
0.5499
0.8
ROI 5
1.951 (0.226)
1.958 (0.228)
+0.4
0.5204
1.6
ROI 6
1.574 (0.364)
1.567 (0.357)
-0.3
0.3289
1.2
ROI 7
1.457 (0.291)
1.461 (0.292)
+0.4
0.5108
1.5
ROIALL*
1.6273 (0.232)
1.6270 (0.228)
+0.02
0.9287
0.6
* mean values of ROI 1 to 7
Table V. Mean (±
SD)
2
value for the BMD (g/cm ) in rotation measurements in 15° INT, 5° INT and 15° EXT rotation for all the ROIs
15° INT
5° INT
15° EXT
da%*
p value
db%†
p value
dc%‡
p value
ROI 1
0.848 (0.199)
0.849 (0.201)
0.836 (0.207)
+0.4
0.8851
-1.2
0.4605
-1.4
0.4193
ROI 2
1.644 (0.278)
1.617 (0.298)
1.566 (0.365)
-1.8
0.1559
-5.5
0.0348
-3.8
0.0884
ROI 3
1.858 (0.236)
1.869 (0.250)
1.887 (0.285)
+0.5
0.5492
+1.3
0.1939
+0.8
0.2941
ROI 4
1.817 (0.209)
1.822 (0.216)
1.804 (0.198)
+0.2
0.5481
-0.7
0.2201
-0.9
0.1758
ROI 5
1.828 (0.192)
1.861 (0.217)
1.880 (0.223)
+1.7
0.0632
+2.8
0.0619
+1.1
0.4266
ROI 6
1.659 (0.231)
1.550 (0.282)
1.479 (0.279)
-6.3
0.0417
-10.5
0.0044
-4.6
0.0025
ROI 7
1.337 (0.244)
1.310 (0.240)
1.250 (0.202)
-1.8
0.0986
-5.8
0.0037
-4.1
0.0027
ROIALL§
1.570 (0.187)
1.554 (0.198)
1.529 (0.212)
-1.1
0.0819
-2.8
0.0032
-1.7
0.0115
*
†
‡
§
percentage difference from 5° INT (BMD2) to 15° INT (BMD1)
percentage difference from 15° EXT (BMD2) to 15° INT (BMD1)
percentage difference from 15° EXT (BMD2) to 5° INT (BMD1)
mean values of ROI 1 to 7
the BMD may differ according to hip pathology and many
years of reduced activity by up to 20% in some cases has
11,15,16
also been ignored.
In general, we were able to confirm the differences
previously reported although not to the degree described.
15
While Hall et al found an increase in BMD in the femoral
neck of up to 10% in arthritic patients, we found an
17
increase of +1.6% in ROI 7 (Table I). Masuhara et al also
found an increase in the BMD of the femoral neck of 13%
in degenerative hips. We noted a mean preoperative reduction of BMD in periprosthetic bone of -4.3% compared
with the opposite side with a maximum in cancellous bone
18
(ROI 1) of -8.6% (Table I). Adolphson et al also described a preoperative reduction in the middle femur of -1%, in
the distal femur of -11% and in the proximal tibia of -14%.
17
Masuhara et al reported a reduction in BMD of -25% in
THE JOURNAL OF BONE AND JOINT SURGERY
PRECISION OF THE MEASUREMENTS OF PERIPROSTHETIC BONE MINERAL DENSITY IN HIPS
2
Table VI. Mean (± SD) for the BMD (g/cm ) in rotation measurements in
5° INT and 5° EXT rotation for all the ROIs
5° INT
5° EXT
Mean
percentage
deviation
(d%)
ROI 1
0.776 (0.141)
0.784 (0.111)
+2.1
0.7235
ROI 2
1.677 (0.220)
1.642 (0.171)
-1.6
0.3305
ROI 3
1.909 (0.262)
1.947 (0.277)
+2.0
0.2027
ROI 4
1.923 (0.309)
1.911 (0.288)
-0.5
0.5636
ROI 5
1.877 (0.230)
1.890 (0.215)
+0.8
0.5928
ROI 6
1.616 (0.248)
1.441 (0.253)
-5.1
0.0118
ROI 7
1.330 (0.221)
1.271 (0.241)
-4.7
0.0278
ROIALL*
1.573 (0.190)
1.555 (0.183)
-1.0
0.1085
p value
* mean values of ROI 1 to 7; 5° INT = BMD1 5° EXT= BMD2
the proximal tibia. These results show that there is a
marked decrease in the BMD in the affected femur as a
result of reduced activity. An increase is seen only in the
region of the femoral neck or calcar (ROI 7) due to the
formation of osteophytes.
There was, however, an ‘apparent’ increase in BMD
averaging +7.7% (+1.6 to +24.2) with a maximum of
+12.7% in ROI 7 and +24.2% in ROI 1 after implantation
of a femoral stem (Table II). If only the medial and lateral
Table VII. Mean (±
SD)
1069
periprosthetic bone bed (ROI 1 to 3 and 5 to 7) is considered, the apparent increase in BMD is +9.0% (Table II).
19
Checovich and McBeath also found an increase in BMD
20
of +6 to +11% immediately after surgery. Markel et al
showed a mean increase of +11% (+4.7 to +23.1) in their
canine model. They interpret this increase as reflecting a
19
problem with software, while Checovich and McBeath
describe it as being due to ‘compacting of trabeculae’.
However, it is probably due to the former since the same
values were found after explantation of uncemented pressfit stems in experiments in vitro when compared with the
measurements before implantation (unpublished results).
5
However, Niinimäki and Jalovaara found a mean reduction
in the BMD immediately after surgery of -1.7% (-3.8 to
21
+4.4) and Kröger et al of -0.2% (-19.2 to + 7.7).
Thus, in spite of a preoperative reduction of BMD on the
affected side, there is an apparent postoperative mean
increase compared with the opposite side of 3%, especially
in ROI 1 and ROI 7, and possibly of more than 10%. This
phenomenon must be taken into account in retrospective
studies, since in the past reductions of BMD of at least 35%
were seen on the operated side compared with the opposite
2,10,22
side, especially in ROI 1 and 7.
We found that the precision of the measurements varied
from 0.8% to 3.3% for the operated and 0.9% to 1.6% for
2
values for BMD (g/cm ) in rotation measurements in 15° INT, 0° and 15° EXT rotation for all the ROIs
15° INT
0°
15° EXT
da%*
p value
db%†
p value
dc%‡
p value
ROI 1
0.589 (0.252)
0.637 (0.250)
0.601 (0.284)
+10.4
0.0112
+0.5
0.6942
-8.2
0.2568
ROI 2
1.325 (0.344)
1.350 (0.411)
1.241 (0.452)
+1.0
0.5817
-6.5
0.3438
-8.0
0.0818
ROI 3
1.534 (0.389)
1.540 (0.401)
1.541 (0.462)
+0.2
0.7854
-0.4
0.8696
-0.5
0.9888
ROI 4
1.279 (0.438)
1.297 (0.453)
1.320 (0.456)
+1.2
0.0741
+3.3
0.0269
+2.0
0.0132
ROI 5
1.433 (0.395)
1.480 (0.394)
1.452 (0.433)
+4.2
0.2779
+1.7
0.6756
-2.0
0.5813
ROI 6
1.211 (0.397)
1.176 (0.322)
1.136 (0.316)
-0.8
0.4438
-3.5
0.3327
-3.4
0.3287
ROI 7
1.014 (0.354)
0.963 (0.381)
0.956 (0.460)
-4.3
0.3974
-6.9
0.5270
-2.0
0.9230
ROIALL§
1.198 (0.345)
1.206 (0.346)
1.178 (0.373)
+0.8
0.5498
-2.1
0.3831
-2.9
0.0998
*
†
‡
§
percentage difference from 0° (BMD2) to 15° INT (BMD1)
percentage difference from 15° EXT (BMD2) to 15° INT (BMD1)
percentage difference from 15° EXT (BMD2) to 0° (BMD1)
mean values of ROI 1 to 7
Table VIII. Mean (±
ROI 1
SD)
2
values of the BMD (g/cm ) in 15° INT, 0° and 15° EXT rotation in the Adaptiva stem for all ROIs
15° INT
0°
15° EXT
da%*
p value
db%†
p value
dc%‡
p value
0.559 (0.196)
0.613 (0.213)
0.604 (0.202)
+11.3
0.1123
+10.4
0.3081
+0.3
0.8532
ROI 2
1.393 (0.256)
1.397 (0.350)
1.187 (0.477)
-0.2
0.9654
-16.4
0.2208
-17.3
0.0460
ROI 3
1.506 (0.186)
1.504 (0.177)
1.481 (0.212)
+0.01
0.9522
-1.8
0.4885
-1.7
0.5548
ROI 4
1.191 (0.116)
1.202 (0.109)
1.215 (0.108)
+1.0
0.4094
+2.2
0.3475
+1.1
0.2954
ROI 5
1.434 (0.080)
1.458 (0.215)
1.421 (0.215)
+1.5
0.7768
-1.3
0.8414
-2.2
0.6558
ROI 6
1.132 (0.257)
1.171 (0.190)
1.195 (0.216)
+6.0
0.5712
+9.1
0.5748
+2.4
0.7267
ROI 7
0.958 (0.288)
0.819 (0.235)
0.814 (0.361)
-11.8
0.2470
-12.7
0.4462
-0.3
0.9724
ROIALL§
1.167 (0.157)
1.166 (0.156)
1.131 (0.187)
+0.05
0.9640
-3.2
0.3453
-3.3
0.0851
*
†
‡
§
percentage difference from 0° (BMD2) to 15° INT (BMD1)
percentage difference from 15° EXT (BMD2) to 15° INT (BMD1)
percentage difference from 15° EXT (BMD2) to 0° (BMD1)
mean values of ROI 1 to 7
VOL. 82-B, NO. 7, SEPTEMBER 2000
1070
F. MARTINI, C. LEBHERZ, F. MAYER, U. LEICHTLE, E. KREMLING, S. SELL
Table IX. Mean (±
SD)
2
value for the BMD (g/cm ) in 15° INT, 0° and 15° EXT rotation in the Zweymüller stem for all ROIs
15° INT
0°
15° EXT
da%*
p value
db%†
p value
dc%‡
p value
ROI 1
0.619 (0.319)
0.660 (0.305)
0.598 (0.375)
+9.6
0.0725
-9.5
0.6636
-16.8
0.1955
ROI 2
1.257 (0.435)
1.303 (0.502)
1.296 (0.475)
+2.3
0.2278
+3.4
0.5601
+1.3
0.9022
ROI 3
1.563 (0.552)
1.577 (0.572)
1.600 (0.652)
+0.4
0.6304
+0.9
0.6176
+0.7
0.7266
ROI 4
1.367 (0.631)
1.393 (0.653)
1.425 (0.654)
+1.4
0.1447
+4.3
0.0493
+2.9
0.0226
ROI 5
1.432 (0.587)
1.502 (0.549)
1.483 (0.610)
+7.0
0.0986
+4.8
0.4595
-1.8
0.7974
ROI 6
1.290 (0.522)
1.181 (0.444)
1.076 (0.412)
-7.6
0.0502
-16.1
0.0225
-9.3
0.0102
ROI 7
1.070 (0.436)
1.106 (0.469)
1.097 (0.543)
+3.1
0.2668
-1.0
0.6219
-3.6
0.8460
ROIALL§
1.228 (0.491)
1.246 (0.492)
1.225 (0.522)
+1.6
0.0638
-1.0
0.9179
-2.5
0.4965
*
†
‡
§
percentage difference from 0° (BMD2) to 15° INT (BMD1)
percentage difference from 15° EXT (BMD2) to 15° INT (BMD1)
percentage difference from 15° EXT (BMD2) to 0° (BMD1)
mean values of ROI 1 to 7
Table X. Precision of periprosthetic measurement in the literature
Author/s
21
Kröger et al
13
Mortimer et al
8
Nakamura
Checovich and
19
McBeath
23
Cohen and Rushton
2
Kilgus et al
14
Kiratli et al
7
Massari et al
25
Marchetti et al
9
Nishii et al
26
Petersen et al
10
Sabo et al
Modified coefficient
of variation
(mCV%)
Densitometer
2.3
Lunar DPX
1.7
Lunar DPX
1.74
Lunar DPX
3.0
Lunar DPX
2.4 to 3.4
Hologic
3.8
Lunar DPX
2.68 to 4.5
Lunar DPX
2.61 to 8.46
Hologic
3.2 to 4.9
Hologic
2 to 6.7
Lunar DPX-L
2.2 to 4.9
Hologic
2.6
Hologic
the non-operated side using the Lunar DPX-L instrument.
The largest difference in the cancellous greater trochanter
area (ROI 1) produced an mCV% of 3.3% for the operated
and 1.6% for the non-operated sides (Tables III and IV).
21
13
8
Kröger et al, Mortimer et al and Nakamura had similar
results, whereas other authors have reported a higher
mCV% (Table X). Whether the type of equipment had any
influence on the precision of the recordings is not certain.
In those studies in which the mCV% was less than 2% (our
13
8
results, Mortimer et al and Nakamura ) the data were
obtained using the DEXA units from Lunar, but a more
9
recent study obtained much worse results with the same
equipment.
Analysis of the literature also shows that the methods of
ensuring the precision of the measurements have not been
14
7
uniform. While, like us Kiratli et al, Massari et al,
13
10
Mortimer et al, and Sabo et al have used the same
equation (2) for calculation of the mCV% in double meas19
urements, Checovich and McBeath and Cohen and Rush23
ton used a different approach. Other authors have not
described the methods used in the calculations.
7
Massari et al reported a precision of 2.61 to 8.46% with
a total mCV% of 1.34%. We calculated a mean precision in
all seven regions of interest (ROIALL) in both femora of
0.6% (Tables III and IV). The calculation of the total
mCV% (ROIALL) is mathematically correct, but it reflects
an mCV% which is too low, since the individual fluctuations are balanced by combining the seven regions. These
results will reflect the precision of the method only if the
total mCV% is considered.
22
The results of Korovessis et al, who report a coefficient
of correlation of 0.8 to 0.9, also cannot be compared with
our findings due to different methods of calculation.
Reproducible positioning is required for accurate meas24
urements. Goh, Low and Bose using a standard foot
block, obtained a mCV% of 1.83% for the femoral neck.
Using a specially designed jig, which can reproduce rotation accurately, they obtained an CV% of 0.97%. We also
used a specially produced positioning jig, which allowed
continuous adjustment of rotation and showed an mCV% of
1.5% in ROI 7, compared with measurements on the
femoral neck for the operated femur (Table IV).
We found a mean difference related to positioning of up
to 2.8% with a very high scatter in individual measurements (Tables V and VI). In some ROIs the mean difference was more than 10% (ROI 6) (Table V) with individual
23
differences of up to 60% (ROI 6). Cohen and Rushton
also reported a variation of up to 24% in individual cases.
21
Kröger et al showed a mean variation for rotation of
3.5%, with the greatest variation being 5.1% in ROI 7.
13
Mortimer et al showed a mean difference of 5% between
15° INT and 15° EXT rotation in phantom measurements.
In our phantom measurements, we obtained similar results
with a mean of 2.9%, with maximum variations of 10%
seen in the same position of rotation (ROI 1) (Table VII).
There was, however, no significant difference between the
two types of uncemented stem used in the in vitro studies
with 3.3% for the Adaptiva and 2.5% for the Zweymüller
stems (Tables VIII and IX).
Our study shows that the custom-made femoral stem
gives good results for the precision of measurement
(mCV%) when compared with other series. Imprecise
measurements, such as those taken without correct and
THE JOURNAL OF BONE AND JOINT SURGERY
PRECISION OF THE MEASUREMENTS OF PERIPROSTHETIC BONE MINERAL DENSITY IN HIPS
identical rotation, result in deviations in individual measurements of up to 60%. These can be reduced to less than
15% by exact rotational positioning. When positioning,
rotation and assessment are precisely maintained, measurements of periprosthetic BMD with the DEXA method is an
accurate and reproducible procedure.
The authors dedicate this article to their teacher and paternal friend
Professor Dr W. Küsswetter (Medical Director of the Department of
Orthopaedics, Tübingen), who has been taken from their midst much too
early by a tragic accident.
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. Dickob M, Schittko A, Pfandl S, Wetzel R, Puhl W. Die DXAMethode zur Quantifizierung von Knochendichte-änderungen des
femurs nach zementfreier Hüftendoprothetik.
Osteologie
1993;2:206-12.
2. Kilgus DJ, Shimaoka EE, Tipton JS, Eberle RW. Dual-energy x-ray
absorptiometry measurement of bone mineral density around porouscoated cementless femoral implants. J Bone Joint Surg [Br]
1993;75-B:279-87.
3. Martini F, Sell S, Kremling E, Küsswetter W. Determination of
periprosthetic bone density with the DEXA method after implantation
of custom-made uncemented femoral stems. Int Orthop
1996;20:218-21.
4. McCarthy CK, Steinberg GG, Agren M, et al. Quantifying bone
loss from the proximal femur after total hip arthroplasty. J Bone Joint
Surg [Br] 1991;73-B:774-8.
5. Niinimäki T, Jalovaara P. Bone loss from the proximal femur after
arthroplasty with an isoelastic femoral stem: BMD measurements in
25 patients after 9 years. Acta Orthop Scand 1995;66:347-51.
6. Marchetti ME, Steinberg GG, Greene JM, Jenis LG, Baran DT. A
prospective study of proximal femur bone mass following cemented
and uncemented hip arthroplasty. J Bone Miner Res 1996;11:1033-9.
7. Massari L, Bagni B, Biscione R, Traina GC. Periprosthetic bone
density in uncemented femoral hip implants with proximal hydroxylapatite coating. Hosp Joint Dis 1996;54:206-10.
8. Nakamura K. Measurement of periprosthetic bone mineral density
after cementless hip arthroplasty by dual energy X-ray absorptiometry:
longitudinal and cross-sectional evaluation. J Orthop Sci
1996;1:113-22.
9. Nishii TN, Sugano N, Masuhara K, et al. Longitudinal evaluation of
time related bone remodelling after cementless total hip arthroplasty.
Clin Orthop 1997;339:121-31.
10. Sabo D, Reiter A, Simanek HG, et al. Periprosthetic mineralization
around cementless total hip endoprosthesis: longitudinal study and
cross-sectional study on titanium threaded acetabular cup and cementless Spotorno stem with DEXA. Calcif Tissue Int 1998;62:177-82.
VOL. 82-B, NO. 7, SEPTEMBER 2000
1071
11. Martini F, Schmidt B, Sell S. Validity and reproducibility of osteodensitometric DEXA-measurements following total hip endoprosthesis. Z Orthop Ihre Grenzgeb 1997;135:35-9.
12. Gruen TA, McNeice GM, Amstutz HC. “Modes of failure” of
cemented stem-type femoral components: a radiographic analysis of
loosening. Clin Orthop 1979;141:17-27.
13. Mortimer ES, Rosenthall L, Paterson I, Bobyn JD. Effect of
rotation on periprosthetic bone mineral measurements in a hip phantom. Clin Orthop 1996;324:269-74.
14. Kiratli BJ, Heiner JP, McBeath AA, Wilson MA. Determination of
bone mineral density by dual x-ray absorptiometry in patients with
uncemented total hip arthroplasty. J Orthop Res 1992;10:836-44.
15. Hall ML, Heavens J, Ell PJ. Variation between femurs as measured
by dual energy X-ray absorptiometry (DEXA). Eur J Nucl Med
1991;18:38-40.
16. Trevisan C, Bigoni M, Cherubini R, et al. Dual X-ray absorptiometry for the evaluation of bone density from the proximal femur after
total hip arthroplasty: analysis protocols and reproducibility. Calcif
Tissue Int 1993;53:158-61.
17. Masuhara K, Kato Y, Ejima Y, Fuji T, Hamada H. Bone mineral
assessment by dual-energy X-ray absorptiometry in patients with
coaxarthrosis. Int Orthop 1994;18:215-9.
18. Adolphson P, von Sivers K, Dalén N, Jonsson U, Dahlborn M.
Bone and muscle mass after hip arthroplasty. Acta Orthop Scand
1993;64:181-4.
19. Checovich MM, McBeath AA. Our experience with dual energy Xray absorptiometry (DEXA) and the Wisconsin hip. The Iowa Orthopaedic Journal 1991;15:62-5.
20. Markel MD, Gottsauner-Wolf F, Bogdanske JJ, Wahner HW,
Chao EYS. Dual energy X-ray absorptiometry of implanted femora
after cemented and press-fit hip arthroplasty in a canine model. J
Orthop Res 1993;11:452-6.
21. Kröger H, Miettinen H, Arnala I, et al. Evaluation of periprosthetic
bone using dual-energy X-ray absorptiometry: precision of the method
and effect of operation on bone mineral density. J Bone Miner Res
1996;11:1526-30.
22. Korovessis P, Piperos G, Michael A. Periprosthetic bone mineral
density after Mueller and Zweymueller total hip arthroplasties. Clin
Orthop 1994;309:214-21.
23. Cohen B, Rushton N. Accuracy of DEXA measurement of bone
mineral density after total hip arthroplasty. J Bone Joint Surg [Br]
1995;77-B:479-83.
24. Goh JC, Low SL, Bose K. Effect of femoral rotation on bone mineral
density measurements with dual energy X-ray absorptiometry. Calcif
Tissue Int 1995;57:340-3.
25. Marchetti ME, Steinberg GG, Greene JM, Jenis LG, Baran DT. A
prospective study of proximal femur bone mass following cemented
and uncemented hip arthroplasty. J Bone Miner Res 1996;11:1033-9.
26. Petersen MB, Kolthoff N, Eiken P. Bone mineral density around
femoral stems: DEXA measurements in 22 porous-coated implants
after 5 years. Acta Orthop Scand 1995;66:432-4.