Clinical Science (1991) 80,319-324
319
Total body calcium in patients with inflammatory bowel
disease: a longitudinal study
S. J. S. RYDE', D. CEMENTS', W. D. EVANS3, R. MOTLEY', W. D. MORGAN'y4, C. EVANS',
J. MODES' AND J. E. COMPSTON'
Departments of 2Gastroenterology,3MedicalPhysics and 5Radi010gy,University Hospital of Wales, Cardiff, U.K., and 'Medical Physics,
Singleton Hospital, Swansea,U.K.
(Received 24 May/24 September 1990;accepted 8 November 1990)
SUMMARY
1. Serial measurements of total body calcium have
been made by prompt y-neutron activation analysis in 13
patients with inflammatory bowel disease over a mean
period of 23 months. Changes in spinal trabecular bone
mineral density and radial shaft bone mineral content
were also assessed by using quantitative computed tomography and single photon absorptiometry,respectively.
2. The mean annual decreases (95%confidence intervals) were: total body calcium, 7.8% ( - 12.0 to - 3.7%;
P< 0.001); spinal trabecular bone mineral density, 2.5%
( - 5.0 to +0.1%, 0.05 < P<O.l), radial bone mineral
content, 2.1% ( - 3.4 to - 0.8%; P<O.O1).
3. No significant correlations were found between
rates of change of the three variables. However, there
were significant positive correlations between the baseline
values for total body calcium and radial bone mineral
content (r=0.638, P<0.05), spinal bone mineral density
and radial bone mineral content (r= 0.854, P< 0.01), and
total body calcium and spinal bone mineral density
(r=0.876, P<O.OOl).
4. These results demonstrate rapid decreases in total
body calcium in patients with inflammatorybowel disease
which, in conjunction with the sigmficant decrease in
radial shaft bone mineral content, indicate increased rates
of cortical bone loss. Whilst values for bone mass at
different skeletal sites showed positive correlations within
individuals, no relationship was found between the rates
of change in bone mass at these sites.
5. The rapid bone loss observed in some subjects
emphasizes the importance of early detection of osteoporosis by bone densitometry and the need for effective
prophylactic measures to be established in this group of
patients.
4Present address: Department of Medical Physics, Queen
Elizabeth Hospital, Edgbaston, Birmingham B15 2TH, U.K.
Correspondence: Dr J. E. Compston, Department of
Medicine, Level 5, Addenbrookes Hospital, Hills Road,
Cambridge CB2 2QQ, U.K.
Key words: bone mineral content, inflammatory bowel
disease,osteoporosis,total body calcium.
Abbeviations: BMC, bone mineral content; BMD, bone
mineral density; BMI, body mass index; BW, bone width,
TBc,,total body calcium.
INTRODUCTION
Several groups have reported an increased prevalence of
osteoporosis in patients with inflammatorybowel disease
[l-31. Both cortical and trabecular bone loss may occur
and some patients develop severe osteoporosis with
fractures, loss of height and disability.Rapid decreases in
spinal trabecular bone mineral density (BMD)have also
been demonstrated in one study of 54 patients in whom
the mean annual rate of bone loss was 3% [4].
Total or partial body calcium can be measured by in
vivo neutron activation analysis [5] and predominantly
reflects peripheral cortical bone mass. Traditionally, a
delayed y-counting technique has been used in which the
activation of 48Caby neutron irradiation is followed by a
counting period in a low background whole body monitor.
More recently, the technique of prompt y-neutron activation analysis has been described in which y-radiation
emitted from the activation product of 40Cais counted
during the irradiation period [6, 71. This technique has
been used in the present study to investigate serial
changes in total body calcium (TB,,) in a group of
patients with inflammatory bowel disease over 9 mean
period of 2 years; changes in bone mass in the spine and
radial shaft have also been assessed over the same time
period.
METHODS
patients
Thirteen patients (eight males, five females) were recruited from the Origilld cohort Of 75 patients included in
an earlier study [3] on the basis of their willingness to
S. J. S. Ryde et al.
320
participate in the study, which involved travelling to
Swansea for TB,, measurements (Table 1).Four patients
(nos. 3, 5, 6 and 11) had initial spinal trabecular BMD
values > 2 SD below the normal age- and sex-matched
values: patient no. 3 also had multiple vertebral crush
fractures. One male had ulcerative colitis (patient no. 8)
and the remainder had Crohn’s disease. Patient nos. 1-5,
7 and 9-13 with small-bowel Crohn’s disease had undergone at least one intestinal resection. Patient no. 13 had
both colonic and small-bowel Crohn’s disease, whilst
patient no. 6 had disease limited to the sigmoid colon.
During the study one patient (no. 10) received sodium
fluoride (20 mg, three times daily) and calcium supplements, and one patient (no. 3) was given an oestradiol
implant (50 mg) 4 months before her last measurement.
All patients except two (nos. 4 and 11) received corticosteroids during the course of the study.
Measurements were performed over a mean period of
23 months (range 20-26 months). Three serial
measurements of TB,,, spinal BMD and radial bone
content (BMC) were made in seven patients, and two in
the remaining six patients; however, one patient (no. 5)
had only a single measurement of radial BMC. The study
was approved by the local ethical committee.
Measurement of TB,,
TB,, was measured by detecting the prompt
y-radiation emitted after neutron capture in 40Caduring
irradiation. The neutron activation instrument contained
a 4-5 GBq californium neutron source (252Cf).The
subject was scanned on a motorized couch between the
neutron source and two n-type hyperpure germanium yray detectors of 20% relative efficiency and high energy
resolution (1.9 keV at 1.33 MeV). A full description of
the instrument has been given previously [6]. Activation
parameters for the prompt technique are compared with
those of the delayed method in Table 2.
The suitability of the instrument for the measurement
of calcium was assessed by investigating: (i) the
measurement uniformity with width and depth of subject,
and (ii)the precision of repeated measurements [7].
Using a ten-piece anthropomorphic phantom (height
1700 mm, water-filled weight 70 kg) the
anterior-posterior measurement uniformity from a
bilateral irradiation was determined by irradiating a small
titanium plate at different positions within the phantom.
Titanium has an intense prompt capture y-ray at 6417
keV, only 3 keV below that of calcium. The uniformity
(defined as the root mean square deviation about the
mean) varied from k 17.1% to k6.7% for phantom
sections of anterior-posterior thickness 120 mm and 200
mm, respectively. Measurements were made to within 10
mm of the anterior and posterior surfaces. The uniformity
in the lateral direction was 23.7% measured to within
20 mm of the surface of a 300 mm wide section of the tenpiece phantom.
The combined uniformity and linearity of response was
assessed by measurement of a spine and rib phantom in
which the mass and position of calcium could be varied.
The phantom was immersed in a tank filled with water to
a depth of 200 mm. Linear regression analysis of calcium
content against calcium y-ray counts gave a high correlation coefficient ( r = 0.986, P < O.OOl), indicating good
agreement with a coefficient of variation about the regression line of 7.8%.
The measurement precision was determined from eight
scans (over a 2 week period) of a Bush-type body phantom filled with a homogeneous tissue equivalent solution
containing 1000 g of calcium. The precision (expressed as
the coefficient of variation) for a skin dose equivalent of
6.4 mSv (neutron quality factor= 10) was 2.6%. Because
of the radiation dose involved, measurement of precision
in vivo could not be undertaken and an estimated precision of about 5% was derived from spectra in vivo. This
value mainly reflects the uncertainty associated with
calculation of the net counts in the calcium y-ray peaks;
for example, for the spectrum shown in Fig. 1, the error in
the peak counts arising from counting statistics (1 SD) is
4.9%. Other possible factors include changes in body
*
Table 1. Details of patients and percentage changes per annum in TB,, BMCIBW and BMD
Patient
no.
Females
1
2
3
4
5
Males
6
7
8
9
10
11
12
13
Age
(years)
BMI (kg/m*)
29
40
40
45
52
19.4
26.3
20.6
22.8
24.4
25
26
42
45
45
48
50
60
25.1
19.0
26.3
19.2
20.4
23.2
20.8
21.9
Change per annum (%)
TB,,
Corticosteroid therapy
BMC/BW
BMD
- 1.1
-11.5
- 17.1
- 7.5
- 2.5
- 3.2
+ 1.6
+ 3.2
- 4.3
-
+ 5.2
- 6.3
+ 4.9
-2.0
- 12.5
Prednisolone 2.5 mg/day
Prednisolone 7 mg/day
Prednisolone 5 mg/day
None
Prednisolone 5 mg/day
- 6.6
- 6.5
- 11.5
- 13.7
- 5.4
- 15.5
+ 0.6
+ 0.4
+ 3.0
- 1.3
- 8.1
- 2.9
- 5.8
+ 3.8
- 2.1
- 1.2
- 6.6
Prednisolone 10 mg/day
Prednisolone 5 mg/day
Prednisolone 8 mg/day
Prednisolone 15 mg/day
Prednisolone 7.5 mg/day
None
Prednisolone 7.5 mg/day
Prednisolone 10 mg/day
- 4.0
- 1.7
- 1.8
- 3.5
- 2.4
- 6.2
- 1.4
Total body calcium in inflammatgrybowel disease
321
Table 2. Activation parameters for the two stable isotopes of calcium used in the measurementof
TBcn
Initial stable Natural abundance
loz8x Neutron
Product nuclide Half-life Principal y-radiation
("4 capture cross-section
isotope
(4
energy (MeV)
(m')
(intensities)
48Ca
0.187
40Ca
96.94
1.1
49Ca
0.41
41Ca*
528
1O-I5
3.084 (92%)
4.072 (8%)
1.943 (53%)
6.420 (28%)
*Intermediatecompound nucleus in excited state.
t* I
1300
1
=
*
.
I
.
.
-
-
1
- ..
-
1
1
:I
Quantitativecomputed tomography
1060
CI
v1
9
820
8
8
x
580
Ca
340
100
summed and corrected for background. The 1943 keV
y-ray was not included in the analysis.
i
5000
5500
6000
6500
y-Ray energy (keV)
Figure 1. A portion of the prompt y-ray spectrum
recorded by one detector from patient 4. The calcium yray peaks used in the analysis are labelled (Ca). Other
prominent peaks arise from chlorine (C1).
habitus and subject repositioning. However, the former
was not significant as tested by estimating the body mass
index [BMI, ~eight/(height)~].
Patients were scanned from shoulder-to-knee in the
prone and supine positions; the neutron beam extended
about 100 mm beyond these limits as determined from
the 50% thermal neutron fluence point. A scan speed of
0.5 mm/s was used so that the total measurement time for
most patients was about 1 h. The scan length was divided
into four equal segments and the data acquired from each
segment were automatically stored so that a longitudinal
profile of calcium content could be obtained. The
segment spectra (eight per detector) were then added to
give two bilateral spectra, one from each detector. A
portion of a representative spectrum from patient no. 4 is
shown in Fig. 1. Counts within the principal calcium yradiation peaks at 6420, 5909 and 5398 keV were
Spinal trabecular BMD was measured in the first three
lumbar vertebrae on a Philips 350 X-ray CT scanner
using a modification of the method of Cann & Genant [8].
Subjects were positioned with the lumbar spine over a
phantom made of a water equivalent plastic, which contained solutions of dipotassium hydrogen phosphate and
water. Scans were made through the middle of each
vertebra at right angles to the long axis using a slice
thickness of 6 mm and a field of view of 400 mm. A
special low-dose parameter set (tube voltage 120 kVp,
tube current 100 mA) was used and this gave weighted
whole-body dose equivalents of 0.10 and 0.17 mSv for
men and women, respectively.
An irregular region of interest was drawn in the
anterior part of each lumbar body avoiding obvious areas
of heterogeneity, and the mean computed tomography
number was corrected for deviation of the water value
from zero [9].The results were corrected for the contribution of non-mineral components to the attenuation coefficient, using published tissue composition data [lo].
Over 2 years precision in vitro was 1.1% (coefficient of
variation, n = 12) whilst the short-term precision in vivo,
cdculated from duplicate measurements in six patients
over a time period of 6 h or less, was 2.6%, giving a
combined precision of 2.8% [ l 11.
Single photon absorptiometry
BMC was measured at the junction of the distal onethird and proximal two-thirds of the left radius by using
an in-house single photon absorptiometer. This instrument was based on the original design of Sorensen &
Cameron [12];it incorporated an 12JIradiation source and
was calibrated against a commercial instrument (Novo
Densitometer GT35). BMC was expressed in units of
g/cm, and the ratio of BMC to bone width (BW)in g/cm2.
Over a period of 14 months, precision in vitro for BMC/
BW in an aluminium/Perspex phantom was 2.4%
( n= 128) and precision in vivo in a normal male subject
was also 2.4% ( n= 43).
S. J. S. Ryde et al.
322
nificant correlation was found between bone loss at any
site and the dose of corticosteroids.
Statistical analysis
A separate linear regression of each measured variable
against time was performed on each patient. For each
variable the mean annual rate of change was obtained
from pooled estimates of the slopes by analysis of covariance. An F test was used to determine whether the
slope obtained differed s i m c a n t l y from zero.
Correlations between the slopes and baseline values of
all three measured variables were examined by linear
regression analysis. The confounding effects of sex and
age were eliminated, in the case of sex by expressing
values as differences from sex-specific means, and for age
by using partial correlation.
DISCUSSION
Our results demonstrate rapid decreases in TB,, in some
patients with inflammatory bowel disease and are consistent with the increased risk of osteoporosis reported in
this group of patients [l-31. There are no published data
on changes in TB,, in such patients, although we have
recently shown rapid rates of spinal trabecular bone loss
in a larger group of patients with inflammatory bowel
disease. Because of the relatively large radiation dose
involved in TB,, measurements we were unable to
include a control group in this study; however, Aloia er al.
[13] reported annual loss rates of only 0.45% in a group
of normal postmenopausal women. The 95% confidence
limits for TB,, in the present study were wide and
probably reflect both the imprecision of the measurement
technique and the clinical heterogeneity of the patients
studied. Nevertheless, the decrease observed in TB,, in
the group as a whole was highly s i m c a n t , the mean
value of 7.8% suggesting that considerable bone loss
occurred from sites other than those assessed in this
study. Biochemical indices of calcium balance, such as
urinary calcium excretion, were not assessed in these
patients, but it is unlikely that there was any systematic
error in the TB,, measurement from one set of
measurements to the next, since no sigruficant changes in
body habitus occurred during the course of the study, and
in the absence of such changes variation in 4LCaproduction by neutrons should not occur. The correlations
between bone mass at different skeletal sites within
individuals are in agreement with other published data
[14], whilst the lack of relationship between rates of bone
loss at different sites is consistent with a number of studies
demonstrating differential rates of bone loss both between
cortical and trabecular bone and between different
skeletal sites [15-171.
RESULTS
Individual changes in TB,,, radial BMC/BW and spinal
trabecular BMD are shown in Table 1, expressed as the
percentage change per year. The mean annual rate of
change for each variable is shown in Table 3. TB,,
showed a mean annual decrease of 7.8% (-12.0 to
- 3.7%); the corresponding changes in radial BMC/BW
and spinal trabecular BMD were -2.1% (-3.4 to
-0.8%, P < O . O l ) and -2.5% (-5.0 to +0.1%,
0.05 < P<O.l), respectively. Of the three patients with an
increase in spinal BMD, one (patient no. 3) had received
an oestrogen implant 4 months before the final
measurement and one (patient no. 10) had been given
sodium fluoride, 20 mg three times daily, throughout the
course of the study. Patient no. 3 also showed an increase
in radial BMC/BW.
No significant correlations were found between the rate
of change in any of the three variables measured (Table 4).
However, when correlations were examined between the
baseline values sigmficant positive relationships were
found (TB,, versus radial BMC/BW, r = 0.638, P< 0.05;
spinal BMD versus radial BMC/BW, r=0.854, P<O.Ol;
TB,, versus spinal BMD, r=0.876, P<O.OOl). No sig-
Table 3. Mean rates of change in TB,, BMC/BW and BMD
"Bca
BMC/BW
BMD
Estimated slope
per m u m ("A)
95% Confidence
t
Significance
- 7.8
- 12.0 to - 3.7
- 3.4 to - 0.8
- 5.0 to + 0.1
- 3.90
- 3.41
P < 0.001
P < 0.01
0.05 < P < 0.1
-2.1
- 2.5
interval (YO)
- 2.03
Table 4. Matrix of partial correlation between slopes and intercepts for the three measurements
Statistical sigdicance: *P<O.O5, **P<O.Ol, ***P< 0.001.
%a
(slope)
TBca(mtercept)
BMD (slope)
BMD (intercept)
BMC/BW (slope)
BMC/BW
BMC/BW
BMD
(intercept)
(slope)
(intercept)
+ 0.227
+ 0.638:
- 0.050
+ 0.854**
-0.363
- 0.220
- 0.003
- 0.271
-0.518
-0.169
+ 0.876***
-0.138
BMD
TBC,
(slope)
(intercept)
+ 0.055
-0.149
-0.190.
Total body calcium in inflammatorybowel disease
The mean annual decrease of 2.1% in radial BMC/BW
was also statistically significant and higher than the
reported normal age-related rate of bone loss at this site
of < 1%[18].This increased loss at a site predominantly
composed of cortical bone is in keeping with the decrease
observed in TBc,,of which peripheral cortical bone is the
largest component. In the present study, changes in spinal
trabecular BMD did not achieve statistical significance,
although losses occurred in 10 of the 13 patients, and two
of those in whom spinal BMD increased were receiving
treatment for osteoporosis; moveover, the mean annual
loss of 2.5% was greater than would be expected in
healthy subjects of a similar sex and age distribution.
The pathogenesis of osteoporosis associated with
intestinal disease is multifactorial; corticosteroid therapy
is likely to play an important role in some patients [19,
201, whilst oestrogen deficiency and amenorrhoea are
major factors in women [21, 221. Other possible pathogenetic factors include calcium malabsorption [23, 241,
secondary hyperparathyroidism resulting from calcium
and/or vitamin D deficiency [25]and altered sex hormone
status in men [26]. In our small group of patients it was
not possible to assess the relative contributions of these
risk factors; many of those studied had relatively severe
inflammatory bowel disease and all but two were receiving corticosteroids.However, patients with severe disease,
intestinal resection and high dose corticosteroid therapy
are most likely to be at risk, especially in the presence of
secondary amenorrhoeaor premature menopause.
The recent development of safe, precise techniques
which measure BMD at potential fracture sites enables
identification of those individuals who are most likely to
be at risk from fracture. Fracture risk is closely related to
bone mass [27]and some prospective studies indicate that
future fracture risk can be predicted from a single
measurement of BMD [28,29].The increased prevalence
of osteoporosis associated with inflammatory bowel
disease together with the rapid rates of bone loss in some
patients indicate that, where resources are available,
BMD should be routinely monitored, particularly where
known major risk factors exist. Measurement of TB,, is
impracticablein large numbers of patients because of its
lack of availability and the relatively high radiation dose
associated with its measurement; dual energy X-ray
absorptiometry,which enables BMD measurements in the
spine, femur and whole body, is the most appropriate
technique presently available [30].In patients with a low
BMD or rapid rates of bone loss, prophylactic or therapeutic measures should be considered. Long-term
oestrogen replacement during and after the menopause
prevents bone loss and reduces fracture rate in normal
women [31, 321; more information is required about the
effectiveness of such therapy in women with malabsorption, particularly those receiving corticosteroid therapy.
The long-term effects on bone mass and fracture rate of
other therapeutic regimens, such as sodium fluoride,
calcitonin and bisphosphonates, also require further
study, both in primary osteoporosis and in bone loss
associated with intestinal disease. Finally, known risk
factors such as cigarette smoking and excessive alcohol
323
consumption should be avoided, whilst physical activity
and adequate dietary calcium intake should be encouraged; corticosteroid therapy should be kept to a
minimum with regular revision of the dose required.
REFERENCES
1. Genant, H.K., Mall, J.C., Wagonfeld, J.B., Horst, J.V. &
Lanzl, L.V. Skeletal demineralisation and growth retardation in inflammatory bowel disease. Invest. Radiol. 1976;
11,541-9.
2. Hylander, E., Ladefoged, K. & Madsen, S. Calcium balance
and bone mineral content following small intestinal resection. Scand. J. Gastroenterol. 1981;16,167-76.
3. Compston, J.E., Judd, D., Crawley, E.O. et al. Osteoporosis
in patients with innammatory bowel disease. Gut 1986;28,
410-5.
4. Motley, R.J., Crawley, E.O., Evans, C., Rhodes, J. &
ComDston. J.E. Increased rates of ~ ~ i n trabecular
al
bone
loss h patients with inflammatory bo\;el disease. Gut 1988;
29,1332-6.
5. Cohn, S.H. The present state of an in vivo neutron activation analysis in clinical diagnosis and therapy. At. Energ.
Rev. 1980;18,599-660.
6. Ryde, SJ.S., Morgan, W.D., Sivyer, A., Evans, CJ. &
Dutton, J. A clinical instrument for multi-element in vivo
analysis by prompt, delayed and cyclic neutron activation
using 252Cf.Phys. Med. Biol. 1987;32,1257-71.
7. Ryde, SJ.S., Morgan, W.D., Cobbold, S. et al. Measurement
of total body calcium by prompt gamma neutron activation
analysis: instrument performance and preliminary clinical
experience. In:Ring,E.F.J., Evans, W.D. & Dixon, AS., eds.
Osteoporosis and bone mineral measurement. York
Institute of Physical Sciencesin Medicine, 1989 105-15.
8. Cam, C.E. & Genant, H.K. Precise measurement of
vertebral mineral content using computed tomography. J.
Comput. Assist. Tomogr. 1980;4,493-500.
9. Brooks, R.A. A quantitative theory of the Hounsfield unit
and its application to dual energy scanning. J. Comput.
Assist. Tomogr. 1977;1,487-97.
10. Crawley, E.O., Evans, W.D. & Owen, G.M. The accuracy of
single energy CT measurements. Phys. Med. Biol. 1988;33,
1113-27.
11. Christiansen, C. & Rodbro, P. Long-term reproducibility of
bone mineral content measurements. Scand. J. Clin.Lab.
Invest. 1977;37,321-3.
12. Sorenson, J.A. & Cameron, J.R. A reliable in vim
measurement of bone mineral content. J. Bone Joint Surg.
1967;49A,481-97.
13. Aloia, J.F., Ross, P., Vaswani, A., Zanzi, I. & Cohn, S.H.
Rate of bone loss in postmenopausal and osteoporotic
women. Am. J. Physiol. (Endocrinol. Metab. 5) 1982;242,
E82-6.
14. Mazess, R.B. The non-invasive measurement of skeletal
mass. Bone Miner. Res. 1983;1,223-79.
15. Eggs, B.L., Wahner, H.W., Dunn, W.L., Mazess, R.B.,
Offord, K.P. & Melton, L J . Differential changes in bone
mineral density of the appendicular skeleton with aging:
relationship to spinal osteoporosis. J. Clin. Invest. 1981;77,
328-5.
16. Stevenson, J.C., Banks, L.M., Sphks, TJ. et al. Regional and
total skeletal measurements in the early menopause. J. Clin.
Invest. 1987;80,258-62.
17. Schaadt, 0.& Bohr, H. Different trends of age-related
diminution of bone mineral content in the lumbar spine,
femoral neck and femoral shaft in women. Calcif. Tissue Int.
1988;42,71-6.
18. Mazess, R.B. On ageing bone loss. Clin. Orthop. 1982;165,
239-52.
324
S. J. S. Ryde et al.
19. Hahn, T.J., Boisseau, C. & Avioli, L.V. Effect of chronic
corticosteroid administration on diaphyseal and metaphyseal bone mass. J. Clin. Endocrinol. Metab. 1974; 39,
274-82.
20. Bressot, C., Meunier, PJ., Chapuy, M.C., Lejeune, E.,
Edouard, C. & Darby, A J . Histomorphometric profile,
pathophysiology and reversibility of corticosteroid induced
osteoporosis. Metab. Bone Dis. Relat. Res. 1979; 1,
303-11.
21. Aitken, J.M. Osteoporosis and its relation to oestrogen
deficiency. In: Campbell, S., ed. The management of the
menopause and postmenopausal years. Lancaster: MTP
Press, 1976: 225-36.
22. Drinkwater, B.L., Nilson, K., Chestnut, C.H., Bremmer,
W.J., Shainholtlz, S. & Southworth, M.B. Bone mineral
content of amenorrhoeic and eumenorrhoeic athletes. N.
Engl. J. Med. 1984; 311,277-81.
23. Hylander, E., Ladefoged, K. & Jarnum, S. The importance
of the colon in calcium absorption following small intestinal
resection. Scand. J. Gastroenterol. 1980; 15,55-60.
24. Southgate, D.A.T., Widdowson, E.M., Smits, B.J., Cooke,
W.T., Walker, C.H.M. & Mathers, N.P. Absorption and
excretion of fat by young infants. Lancet 1969; i, 487-9.
25. Parfitt, A.M., Gallagher, J.C., Heaney, R.P., Johnston, C.C.,
26.
27.
28.
29.
30.
31.
32.
Neer, R. & Whedon, G.D. Vitamin D and bone health in the
elderly. Am. J. Clin. Nutr. 1982; 36,1014-31.
Farthing, M.J.G. & Dawson, A.M. Impaired semen quality in
Crohn’s disease - drugs, ill health or undemutrition?
Scand. J. Gastroenterol. 1983; 18,57-60.
Jensen, G.F., Christansen, C., Boesen, J., Hegedus, V. &
Transbel, I. Relationship between bone mineral content and
frequency of postmenopausal fractures. Acta Med. Scand.
1983; 213,61-3.
Wasnich, R.D., Ross, P.D., Heilbrun, L.K. & Vogel, J.M.
Prediction of postmenopausal fracture risk with bone
mineral measurements. Am. J. Obstet. Gynecol. 1985; 153,
745-51.
Hui, S.L., Slemenda, C.S. & Johnston, C.C. Baseline
measurement of bone mass predicts fracture in white
women. AM. Intern. Med. 1989; 111,355-61.
Mazess, R.B., Collick, B., Trempe, J., Barden, H. & Hanson,
J. Performance evaluation of a dual energy X-ray densitometer. Calcif. Tissue Int. 1989; 44,228-32.
Lindsay, R., Hart, D.M., Forrest, C. & Baird, C. Prevention
of spinal osteoporosis in oophorectomised women. Lancet
1980;ii, 1151-3.
Ettinger, B., Genant, H.K. & Cann, C.E. Long-term
oestrogen replacement therapy prevents bone loss and
fractures. AM. Intern. Med. 1985; 102,319-24.