Bone, 6,21-27 (1985)

Printed in the USA. All rights reserved.

Copyright

87.56-3282185 $3.00 + .OO

0 1985 Pergamon Press Ltd.

Relationships
BetweenCalciumandPhosphorus
Homeostasis,
Parathyroid
HormoneLevels,BoneAluminum,andBone
Histomorphometry
in PatientsonMaintenance
Hemodialysis
P. NILSSON,

F. MELSEN,3 J. MALMAEUS,*

B.G. DANIELSON,

and L. MOSEKILDE4

Departments of Internal Medicine and 2Surgery, University Hospital, Uppsala, Sweden.

?Jniversity Institute of Pathology and 4Medical Department Ill, Aarhus Amtssygehus, Aarhus, Denmark.
Address for correspondence

and reprints:

Per Nilsson, Department of Internal Medicine, University Hospital, S-751 85, Uppsala, Sweden.

Abstract

1976; Nielsen et al., 1980a, 198Ob). In hemodialysis patients in

whom hypocalcemia is corrected by a sufficiently high dialysate calcium concentration (Johnson, 1976) several investigators have found largely normal bone mineralization rates at all
levels(Nielsen et al., 1980a; Teitelbaum et al., 1980; Weinstein,
1982) and normal relationships between mineralization rate,
bone formation rates at tissue and BMU levels, and serum PTH
have been demonstrated. However, a subpopulation of dialysis patients develops a disorder of bone characterized histomorphometrically
by low bone turnover and a severe
mineralization defect (Kanis, 1981). This type of bone disease
has been associated with the use of aluminum-contaminated
water for dialysis, suggesting possible toxic effects of aluminum on bone mineralization (Ward et al., 1979; Parkinson et al.,
1979; Walker et al., 1982). Direct effects of aluminum on the
mineralization process have also been demonstrated in the rat
(Ellis et al., 1979).
This study emanates from a center where, despite the use of
softened water with some aluminum content, there have been
very few clinical problems of dialysis bone disease. The aim of
this study was to determine the incidence and histomorpho
metric characteristics of bone disease in patients treated with
long-term hemodialysis without symptoms of musculoskeletal
disease and to evaluate relationships among bone dynamics
and indices of parathyroid function and aluminum accumulation in such patients.

Serum biochemistry related to calcium and phosphorus

homeostasis and parathyroid function was studied together
with bone histomorphometry after double-labeling with tetracycline and staining for aluminum in 17 patients without
symptoms of bone disease, treated with maintenance
hemodialysis for at least 6 months. A close correlation was
found between the serum level of parathyroid hormone
(PTH) and bone resorption surfaces and bone formation
rates, both at tissue and basic multicellular unit (EMU)
levels. The patients could be divided into a high turnover
group with a normal mineralization process and a low turnover group with markedly defective mineralization. The second group was further characterized by lower PTH and
higher fractional aluminum-stained trabecular bone surfaces. For the whole patient material, the fractional aluminum-stained surfaces related inversely to tetracyclinelabeled surfaces and to bone formation rates at both BMU
and tissue levels, but not to the time on dialysis or to the
cumulative ingested amount of aluminum hydroxide. The
data provide evidence that PTH or PTH-related factors,
besides activating bone remodeling, directlyenhance bone
formation in dialysis patients and that aluminum incorporation into bone is associated with a progressive disturbance
of bone mineralization.
Key words: Hemodialysis-Calcium-Phosphate-Parathyroid
Hormone-Bone Aluminum-Bone
Histomorphcmetry.

Patients and Methods

The study comprised 17 patients (13 males and 4 females) with a mean
age of 49 years (range 20-75 years), treated with maintenance
hemodialysis (HD) for at least 6 months (mean 19 months, range 6-62
months). The cause of renal failure was chronic glomerulonephritis in
11, arteriosclerotic kidney disease in 2. tuberculcus nephropathy in 2,
and polycystic kidney disease and hereditary nephropathy in 1 patient
each. The studied patients were gathered from our current population
of some 35 patients receiving hospital hemodialysis. All had been
treated at our unit exclusively. None had symptoms of musculoskeletal

Introduction
Hypocalcemia,
secondary to phosphate retention and deficiency of 1,25-dihydroxycholecalciferol,
is important in the
pathcgenesis of the defective bone mineralization regularly
observed in nondiafyzed uremic patients (Malluche et al.,
21

P. N~lsson et al.: PTH, alummum, and dialysis bone disease

22

disease or a history of recent exposure to treatment with vrtamin D,

calcium, or alkali supplements. One patient was bilaterally nephrecto
mized and 2 patients had been subjected to parathyroid surgery
because of hypercafcemia, the operations taking place 28 to 50
months before the present investigation. Dialysis treatment was given
for lo-15 h/week, using a single-pass system and standard hollow
fiber or parallel plate dialyzers. The dialysate contained 1.8 mmoill of
calcium and 0.75 mmolll of magnesium. Softened water was used with
an aluminum content, during recent years, of 10 to 20 pgll. Aluminum
hydroxide-containing phosphate binders were administered to all patients. The total duration of this treatment and the estimated cumulative
dose of aluminum hydroxide were recorded for each patient. Seventeen apparently healthy volunteers, matched for sex and age with the
patients, served as a control group in the evaluation of histomorphometric data (Melsen and Mosekilde, 1978a).

Fractional labeled formation surfaces (ShXtflab,&m2/~m2.

The
extent of single- and double-labeled surfaces expressed as a fraction of
the formation surfaces.
Appositional

rate (M/t)~mlday.
The mean distance between tetracycline lines In all double-labeled areas divided by the interval in days
(t) between the two labelings, uncorrected (u) for obliquity of the plane
of section.

Fractiona! a/um;num-stained surfaces (St,acttn,&m2/~m2. The

extent of trabecular bone surfaces revealing a positive reaction for
aluminum, expressed as a fraction of the total trabecular bone surface.
Calculated parameters follow:
Bone formation rate, tissue /eve/,surface referent (sV,)pm3/~m2per day
calculated as:

Laboratory Methods
Monthly determinations of serum calcium, phosphate, and alkalrne
phosphatase were performed by standard laboratory methods. Serum
calcium was adjusted for serum albumin concentration by 0.019
mmolll for each g/l that the individual albumin concentration deviated
from the normal mean of 46 g/l. Serum aluminum was determined at
least once in 10 patients, using atomic absorption photometry. As
estimated in healthy individuals, normal values are below 10 pgll.
lmmunoreactive serum parathyroid hormone (PTE& was measured
every 2 to 3 months by a radioimmunoassay using
l-labeled bovine
PTH (lnolex) and sheep antiserum (S478) against bovine and porcine
PTH. This assay measures intact PTH and the C-terminal two thirds of
the molecule (Hehrmann et al., 1980). The reference range, as estimated from 50 apparently healthy individuals, IS 0.40 to 1.20 arbitrary
units II (arbU/I). For each patient all biochemical data collected during
the B-month period before and at the time of the bone biopsy were
pooled. and the average values of the individual variables were used in
the statistical analysis.

Histomorphometric

V, = SlrtilabI x M/t
This represents theamount of new bone mineralized per unit of time per
unit of trabecular bone surface.
Bone formation rate, BMU level, surface referent (5V,rBMuJ~m3/~m2
per day as:

This vanable represents the amount of new bone mineralized per

unit of time per unit of osteoid surface. In a steady-state situation
with regard to osteoid thickness, this is numerically identical to the
appositional rate of osteoid formation.
Mineralization lag time (t,) days:

analysis.

Transcortical iliac crest biopsies

wereobtained with an8 mm trephine(Bordieretal.,
1964) after doublelabeling with tetracycline (600 mg of demethylchlortetracycline 14,13,
4, and 3 days before biopsy). Undecafcified specimens were embedded in methylmetacrylate, and sections (7-8 pm) were cut and stained
with Masson trichrcme and Goldner trichrcune for the measurement of
static parameters of bone resorption and formation (Melsen et al.,
1978). Unstained sections (20 am) were mounted for the quantification
of dynamic parameters by fluorescent microscopy. Sections (7-8 am)
were also stained for aluminum with aurine tricarboxylic acid (Lillie and
Fullmer. 1976).
The following measurements were performed in trabecular bone
using point counting and direct measurements:

This defines the average time lag between osteoid formation and
subsequent mineralization (Melsen and Mosekilde, 1978b).

Statistical methods
Statistical analysis of differences between group means was performed with the Wilcoxon rank sum test for independent samples.
Possible relationships between biochemical and histomorphometric variables were analyzed by calculation of the Spearman rank
correlation coefficient (Colton. 1974).

fractional trabecular bone volume (tVracttb))~m3/~m3, The fraction of a given volume of total trabecular bone tissue occupied by
mineralrzed and unmineralized trabecular bone

Results
Fractional formation surfaces (Sr,,t~rJ~m/~m.
teoid surface expressed
surface

The extent of osas a fraction of the total trabecular bone

Serum biochemistry
Serum calcium (normal 2.20-2.60 mmol/l) for the 6-month
period preceding the bone biopsy was 2.44 -+ 0.02 mmolll
(mean f SEM), and all patients were consistently normocalcemic. Hyperphosphatemia
was noted in 13/17 (mean
1.78 f 0.12 mmol/l, normal 0.76-l .44 mmol/l). Mean serum
alkaline phosphatases were slightly elevated at 6 + 2 rkat/l
(normal 0.8-4.8 &t/l). All patients had elevated PTH (5.5 +
1 .O arbU/I, normal 0.40-l .20 arbU/I). Serum aluminum was
elevated in 9/l 0 studied patients to a mean value of 46 f 14
fig/l (normal < 10 pg/l).

Fractional resorption

SUrfaCeS (St,act&m2/~m2. The extent of

Howshrp lacunae as a fraction of the total trabecular bone surface.

Mean width ofosteoidseams (Wr)cLm. The mean of four extreme,

equally spaced measurements in all surfaces covered by osteoid,
uncorrected (u) for obliquity of the plane of section.
Fractional labeled surfaces (Sract~,a,,))~m2m/~m2.The extent of
single- and double-labeled surfaces expressed as a fraction of the total
trabecular bone surface.

BMU, base multrcellularunit, the functional group of cells (osteoclasts arm osteoblasts) tha! are responsible for the turnover of a structural bone unit (I e., a packet in
trabecular bone)

P. Nilsson et al.: PTH, aluminum, and dialysis bone disease

Bone histomorphometry
The results of the histomorphometric

analysis are illustrated in

Table I for the whole group of patients and controls. The fractional trabecular bone volume was not different in patients and
controls, but the mean values of both fractional resorption surfaces and fractional formation surfaces were increased in the
patient group (P c 0.001). The mean width of osteoid seams
was slightly increased (P < 0.05). The analysis of the dynamic
variables related to tetracycline labeling yielded normal mean
values for fractional labeled surfaces. Thus the fractional labeled formation surfaces were reduced (P < 0.05) since the
fractional formation surfaces were increased. Appositional rate
and bone formation rates at both levels were normal, whereas
the mineralization lag time was prolonged (P c 0.05).
There was a close correlation between the indices of bone
resorption and formation (turnover), expressed as fractional
resorption surfaces and bone formation rate at tissue level,
respectively (r = 0.81, P c 0.0001). Although the mean bone
turnover rate (bone formation rate, tissue level) was similar in
patients and controls, the variance was much greater in the
patient group (F = 17.4, P c 0.001). It was therefore found
convenientto divide the material into two groups on the basis of
a bone turnover rateabove(groupA,n
= 9)or below(group B,
n = 8) the mean value in the control group. The two groups
were comparable regarding sex and age, but the patients in the
low turnover group (B) had spent a longer time on HD (group A
11 k 2 months, group B 29 f 8 months, mean f SEM,
P c 0.05). Histomorphometrically, there were no differences in
trabecular bone volume or in any of the static parameters of
bone formation (Table II). Fractional resorption surfaces were
higher in the high turnover group (A) (P < 0.01). The results of
Table I. Bone hisotmorphometry

23
the tetracycline double-labeling were clearly different; the law
turnover group exhibited lower fractional labeled surfaces and
highly depressed fractional labeled formation surfaces (P <
0.001). The mean appositional rate was the same in both
groups, but in the low turnover group the boneformation rate at
BMU level was reduced (P c 0.01) and the mineralization lag
time was prolonged (P < 0.01). Serum calcium, phosphate,
and alkaline phosphatase were similar in both groups. The
mean serum PTH was lower in the low turnover group (group A
7.8 + 1.5 arbU/I, group B 2.8 + 0.2 arbU/I, mean + SEM, P
c 0.01). The two parathyroidectomized
patients were found in
this group. They were not excluded from the analysis, since they
had maintained normocalcemia for a long time without the use
of vitamin D or calcium supplements, indicating residual functioning parathyroid tissue.
Fractional aluminum-stained
trabecular
surfaces were
higher in patients with low bone turnover (P < 0.05) (Table II).
This was not reflected in any differences in serum aluminum
concentration, in the duration of treatment with, or in the calculated total ingested dose of aluminum hydroxide. There was no
correlation between fractional aluminum-stained surfaces and
time on HD. In the whole patient material, fractional aluminumstained surfaces did not correlate with fractional formation
surfaces but were inversely related to fractional resorption
P < 0.05). The relationships
surfaces (r = -0.51,
between aluminum staining and the results of tetracycline labeling are illustrated in Figure 1. Inverse relationships were
found between fractional aluminum-stained surfaces and fractional labeled formation surfaces (r = - 0.52, P < 0.05) and
bone formation rates at tissue (r = - 0.55, P < 0.05) and BMU
levels (r = -0.70, P < O.Ol), whereas fractional aluminumstained surfaces were directly related to mineralization lag time

in patients and sex- and age-matched controls.

Patients

VI@,,
pm3/pm3
Sf=l(f)
pm/pm2

Controls

0.28 ? 0.02

0.21 * 0.01

(0.15 - 0.50)

(0.12 - 0.32)

0.49 f 0.05

0.20 f 0.01

(0.19 - 1.0)

(0.06 - 0.30)

Sf=W
~m/pm*

0.11 f 0.02

0.05 f 0.01

(0 - 0.24)

(0.03 - 0.07)

uW,

13.0 f

flm

1.7

(4.6 - 34.9)

9.6 + 0.53

0.20 f 0.04

0.15 f 0.02

(0.001 - 0.60)

(0.09 - 0.27)

Sf,Zt(lkJlrJ
fim2/pm2

(0.001 - 1 .OO)

0.71 f 0.07

0.64 k 0.05

0.60 f 0.02

(0.001 - 1.08)

(0.61 - 0.71)

0.13 f 0.04

0.09 + 0.01

pm31pm2 per day

(0.001 - 0.57)

(0.06 - 0.17)

sVf(EW,
pm3/~m per day

0.33 * 0.10

0.44 + 0.05

(0.001 - 1.04)

(0.25 - 0.68)

23 f 3

44

h,
days

(9 -

m)

Values are mean f SEM ( median). Ranges are indicated in parentheses. ns = not significant.

P < 0.001

P < 0.001

P < 0.05

ns

P < 0.05

(0.30 - 1 .OO)

am/day

M/t

ns

(6.5 - 12.1)

Sfrwab)
pmlfim

0.40 f 0.07

Patients vs controls

(11 - 37)

ns

ns

ns

P < 0.05

24

P. Nilsson et al.: PTH, aluminum, and dialysis bone disease

Table II. Bone histcmorphcmetry

in patients with high turnover (group A) and low turnover (group B).
Group B

Group A
(n = 9)

(n = 8)

A vs 0

0.28 + 0.03
(0.16 - 0.50)

.I 0.26 + 0.03

ns

Sr=WI
j.4m2/pmz

0.49 f 0.07
(0.19 - 0.78)

0.50 * 0.09
(0.27 - 1 .O)

S*a(r)
~m2/pm2

0.15 f 0.02
(0.06 - 0.24)

0.07 -c 0.02
(0.001 - 0.15)

Wit
rmlday

0.65 f 0.06
(0.40 - 0.96)

0.62 f 0.10
(0.10 - 1.08)

S(r=t(lab)
pml~m

0.30 f 0.05
(0.12 - 0.60)

0.07 +- 0.02
(0.0001 - 0.15)

P < 0.01

SrWMWnl
pm21~m2

0.64 -t 0.07
(0.40 - 1 .O)

0.15 -c 0.04
(0.0001 - 0.38)

P < 0.001

Vkt,b,

am3/pm3

(0.15 - 0.40)

14 f 3.2
(8.1 - 35)

12 f 1.7
(4.6 - 21)

WI
pm

ns
P < 0.01
ns

ns

0.21 + 0.05
(0.10 - 0.57)

0.04 f 0.01
(0.001 - 0.08)

P < 0.01

0.43 f 0.08
(0.23 - 1.04)

0.10 -c 0.03
(0.001 - 0.28)

P < 0.01

~m3/~m per day

33.6 + 5.7
(9.0 - 57.4)

115*
(36.4 - m)

P < 0.01

days

S*acf(All
pm*/~m

0.05 f 0.02
(0 - 0.16)

0.21 -c 0.06
(0 - 0.48)

P < 0.05

%
V

t,

pm3/c(m2 per day

YBMUI

Valuesare mean + SEM ( median). Rangesin parentheses.

ns = not significant.
(r = 0.64, P < 0.02). There was no correlation between alumi-

num-stained surfaces and appositional

rate.

Correlations between histormorphometric

data

and biochemical

An analysis of correlations between histomorphometric and

biochemical variables was performed for the whole patient
material. All calculations were based on the averages of biochemical values obtained in each patient during the 6-month
period preceding the bone biopsy. Serum calcium did not
relate to any of the histomorphometric variables. For serum
phosphate, there was a posititive correlation with appositional
rate (r = 0.57, P < 0.05). Serum alkaline phosphatase levels
were positively
related to fractional formation surfaces
(r = 0.54, P < 0.05) and to the width of osteoid seams
(r = 0.74, P < 0.001) butnottoanyofthedynamicvariables.
A
positive correlation was observed between serum PTH and
fractional resorption surfaces (r = 0.70, P c 0.01) but there
was a complete absence of correlation between PTH and fractional formation surfaces (r = 0.03). Interrelations between
PTH and variables related to tetracycline labeling are illustrated
in Figure 2. PTH was directly related to fractional labeled
formation surfaces (r = 0.78, P c 0.01) and to bone formation
rate at tissue level (r = 0.87, P < 0.001) and at BMU level (r =
0.78, P c 0.01). PTH was inversely related to mineralization
lag time (r = 0.64, P < 0.01). There was no correlation between PTH and appositional rate. No statistically significant

interrelations were noted between any of the determined biochemical variables nor between serum biochemistry and patient age, time on HD, or time between the diagnosis of renal
disease and the start of HD.

Discussion
The histomorphometric findings in this study of a normal bone
turnover rate (bone formation rate, tissue level) and increased
fractional formation and resorption surfaces are in accordance
with what has been reported by others for dialysis patients
without symptoms of musculoskeletal disease (Nielsen et al.,
1980a). However, even though the mean bone turnover rate
was normal, the patients showed a much greater variation of
this variable than the controls. The material was therefore subdivided into two groups with highly different and nonoverlapping bone turnover rates. The two groups had almost identical
mean values of the static parameters of bone formation. The
fractional resorption surfaces were lower in patients with low
bone turnover, but even in this group the mean value was
slightly above that found in the controls. Such a combination of
increased formation and increased resorption surfaces would
result either from an increased frequency of activation of trabecular bone remodeling foci (BMU) or from an increased
lifespan of the resorptive and formative processes (Frost,
1969).

P. Nilsson et al.: PTH, aluminum, and dialysis bone disease

fract(lab/f)

0.6

25

. .
1

svf
0 6 vm3/pm2/d

prnVprn2
l.O-

0.6-

,
0 0.1

0.3
Sfract(Al)

0.5

,=:

ph/vrn2/d

O.Sl

.
l

.
.-

,_

4
S-PTH

fract(Al)

12
20
arbU/I

S4PTH

svf(BfvlU)
vm5/pm2/d

svf(BMU)
pm3/pm2/d
1.2

1000

.
200 I

12
20
arbU/I

tm
days

1.2

.
0.8

0.8

200

1-4
0.4 i ;,

1..- l. * .
0t**
0.1
O
I Y3 25
l

SfractiAl)

100

I.

l*

0. .* .*
o??
0 0.1

100

.
I

0.3
Sfract(Al)

0.5

Fig. 1. Fractional labeled formation surfaces (S,rtiC,a&, bone formation rates at tissue (V,) and BMU (sV,(BMu,)
levels, and mrneralrzatron lag
time (t,,,) in relation to fractional aluminum-stained surfaces (Sfrac,& in
17 HD patients.

In the patients with high bone turnover, the first of these

mechanisms was evidently operative. In this group the process
of mineralization appeared largely normal, as shown by a normal appositional rate, a normal bone formation rate at BMU
level, and an only very moderately prolonged mineralization lag
time. The patients with low bone turnover, on the other hand,
showed evidence of defective mineralization and an increased
lifespan of the remodeling cycle. The mineralization defect was
characterized by a combination of very low fractional labeled
formation surfaces, a normal appositional rate, and a prolonged mineralization lag time. This indicates a low number of
formative foci being on, i.e., active in mineralization, but a
normal cellular (osteoblast) activity where mineralization is in
progress. In contrast to the situation in patients with high bone
turnover, the group with low bone turnover showed an abnormal ratio of fractional resorption surfacesfractional formation
surfaces (mean value 0.14 vs controls 0.26, P < 0.05). This
ratio can be used as an indicator of the presence of a steadystate situation regarding bone remodeling (Frost, 1969). The
low value in these patients could indicate that they were in a
transient period of still decreasing bone turnover. It could also
be explained by a deviation in the proportion of the lengths of
the resorptive to the formative periods. To solve this problem,
repeated biopsies would be necessary.
These two groups of patients with radically dissimilar bone
dynamics
were comparable
regarding serum calcium,
phosphate, and alkaline phosphatase. All patients were normocalcemic, and no correlations were found between serum calcium and histomorphometric variables. This is in contrast to the
situation in nondialyzed patients, in whom a strong influence
of serum calcium on bone mineralization has been reported

20
S!PTH

::bU/I

4.
:

..

S-PTH

12

20

arbU/I

Fig. 2. Fractional labeled formation surfaces (S,ractr,awrJ,

bone formation rates at tissue @r) and BMU (sVt(sM,, levels, and mineralization lag
time (t,) in relation to serum PTH in 17 D patients.

(Nielsen et al., 1980a; Mora Palma et al., 1983). Serum

phosphate correlated with appositional rate, in accordance
with a role for phosphate in osteoblast function. The association between histomorphometric signs of osteomalacia and low
serum phosphate reported by others (Kanis et al., 1977) was
not found, however. Alkaline phosphatase levels yielded no
information about bone dynamics in these patients. Analysis of
bone isoenzyme activities would possibly have been more
informative (Siede et al., 1980).
Serum PTH was significantly lower in patients with low bone
turnover, and for whole patient material, we found highly significant correlations between serum PTH and fractional resorption
surfaces and between PTH and the indices of bone formation
related to tetracycline labeling. This is in agreement with other
reports and with the action of PTH as an activator of the coupled
processes of bone resorption and formation, and thereby as a
regulator of bone turnover (Jaworski, 1981). Interestingly, there
was no relationship between PTH and fractional formation surfaces. This might support the concept that PTH or some PTHrelated factor, besides stimulating bone turnover, directly or
indirectly enhances bone mineralization in hemodialyzed patients. Thus, the absence of a relationship between PTH and
fractional formation surfaces would result from two opposite,
PTH-related effects: (1) increased activation of bone remodeling foci, which would lead to an increase in formation surfaces,
and (2) an increased mineralization rate (bone formation rate,
BMU level), which would reduce formation surfaces. That PTH
is indeed important for the maintenance of bone formation in
hemodialyzed patients is also suggested from reports of the
development of a severe mineralization defect after radical
parathyroidectomy in such patients (Teitelbaum et al., 1980;
Felsenfeld et al., 1981; Weinstein, 1982).

26

As in other studies (Pierides et al., 1980; Hodsman et al.,

1981, 1982), an association was found between a defective
mineralization
process and signs of aluminum accumulation
in bone. Thus, the patients with low bone turnover had
significantly higher fractional aluminum-stained
trabecular
bone surfaces, and in the whole patient material fractional
aluminum-stained surfaces were inversely related to tetracycline uptake and to bone formation rates at both levels. No
relationship between the rate of mineral apposition and the
surface extent of aluminum could be found, however. This is at
variance with an earlier report (Ott et al., 1982) stating a direct
relationship between surface extent of aluminum and appositional rate. The reason for this difference is not clear but may be
related to the fact that only asymptomatic patients with presumably less severe bone disease were included in the present
study. By the staining method used, aluminum is shown to be
deposited mainly along the interface of osteoid and mineralized bone. This locale has also been demonstrated by x-ray
microanalysis (Boyce et al., 1981; Cournot-Witmer et al.,
1981). Serial sectioning has suggested that the majority of
formative sites with positive staining for aluminum do not take
up tetracycline (Maloney et al., 1982). This would seem to fit
well with our findings of an aluminum-associated
mineralization defect characterized by a depression of the number of
formative foci being active in mineralization and could suggest
a blocking effect of aluminum on the initiation of the mineralization process.
No simple relationships could be found between histomorphometric variables and the time or amount of aluminum
exposure, via the dialysate, or in the form of phosphate-binding
therapy. Thus, some other factor probably governs the tendency of some dialysis patients to accumulate aluminum in
bone. This study confirms previous reports (Cournot-Witmer et
al., 1981; Hodsman et al., 1982) of low serum PTH in these
patients, These data and the reported appearance of a mineralization defect, together with signs of aluminum accumulation
after parathyroidectomy in dialysis patients (Ellis, 1982) may
indicate a role for PTH in the protection of bone against aluminum deposition. A low bone turnover rate secondary to low
PTH could be speculated to facilitate aluminum deposition.
Abnormal parathyroid function, in the form of a deficient PTH
response to an acute hypocalcemic stimulus, has recently
been demonstrated in HD patients with osteomalacia and aluminum accumulation (Kraut et al., 1983; Andress et al., 1983).
Whether aluminum accumulation in the parathyroid glands
(Cann et al., 1979) causes parathyroid dysfunction, as suggested by one in vitro study (Morrissey et al., 1983) or whether
a relative hypoparathyroidism is the primary event, remains to
be elucidated. The latter explanation is supported by the findings in this and previous studies (Nilsson et al., 1983) of constant serum PTH during long-term HD, making it probable that
the two subgroups of patients in this study had different levels
of parathyroid function already at the start of HD. The observations suggesting the presence of a nonsteady-state situation
of decreasing bone turnover in patients with extensive bone
aluminum accumulation
lend further support to a direct
pathogenetic role of aluminum in the development
of a
progressive mineralization defect in dialysis patients.

Mrs. S. Bergmann performed skillful laboratory

work in preparing specimens for the histomorphometric analysis, Mrs.
Iva Kulhanek provided excellent secretarial assistance. This proiect
was supported by grants from the Swedish Medical Research douncil
(Pro. No. B84-18X-02329-17A).
Acknowledgement:

P. Nilsson et al.: PTH, aluminum, and dialysis bone disease

References
Ancfress 0 , FelsenfeldA.J., Voigts A and Llach F.: Parathyroid hormone response to hypocalcemia in hemodialysis patrents with osteanalacia. Kidney
Int. 24:364-370, 1963.
Bordier P., Matraft H., Miravet L. and Hioco D.: Mesure histologique de la masse
et de la resorption des travees osseuses. Pathol. Viol. 12:1238-l 243, 1964.
Boyce B.F , Elder H.Y.. Fell G.S., Nicholson W.A.P., Smith G.D., Dempster D.W.,
Gray C C and Boyle 1.T: Quantikation and localization of aluminum in human
cancelious bone in renal osteodystrophy. Scann. fkfroff Microsc.3:329-

337,198l
Cann C.E.. Prussin S.G. and Gordan G.S.: Aluminum uptake by the parathyroid
glands, J C/in Endocrinol. Metab. 49543-545, 1979.
Colton T.: Statistics in Medicine. Little, Brown, & Co., Boston, 1974.
Cournot-Witmer G., Zingraff J., Plachot J J., Scaig F., Lefevre FL. Boumati P.,
Bcurcieau A., Garabedian M., Galla P., Bourdon R.. DrOekeT and BafsanS.:
Alumtnum localization in bone from hemodialyzed patients: relationship to
matrix mineralization. Kidney Inf. 20:375-385. 1981.
Ellis H.A.. Aluminum and ostecmalacia after parathyroidectomy. Ann. Intern.
Med. 96533-534, 1982.
Ellis H.A., McCarthy J.H. and Herrington J.: Bone aluminum in hemodfalysed
patients and in rats infected with aluminum chloride: relationship to impaired
bone mineralfzatron. J. C/in. Patho/. 32:832-844, 1979.
Felsenfeld A.J Harrelson J.M., Gutman R.A.. Wells S.A. and Drezner M.K.:
Ostecmafacia after parathyroidectcxny in patients with uremia. Ann. Intern.
Med. W:34-39, 1981.
Frost H.M.: Tetracycline based histolcgfcal analysis of bone remodeling Calcif.
fissue Res 3:21 l-237, 1969.
Hehrmann R., Nordmayer H., Mohr M. and Hesch R.-D.: Human parathyroid
hormone antibody characterization. J. immunoassay 1: 151-l 74, 1980.
Hodsman A.B.. Shenard D.J.. Alfrey A.C , Ott S., Brickman A.S.. Miller N.L..
Maloney N.A. and Coburn J.W: Bona aluminum and histomorphcmetrii
features of renal osteodystrophy J. C/in. EndoWno/. Meteb. 54:539-546.
1982
Hodsman A.B., Sharrard D.J., Wong E.C.G., Brickman AS.. Lee D.B.N., Alfrey
AC., Singer FE., Norman A.W. and Coburn J.W.: Vitamin D-resistant osteomalacia in hemodiafysis patients lacking secondary hyperparathyroidism.
Ann. Intern. Med. 94:629-637, 1981.
Jaworski Z.G.F.: Physiology and pathology of bone remodeling. Orthoped, C/in.
North Am. 12:485-512,1981.
Johnson W.J.- Optimum diafysate calcium concentration during maintenance
hemodiafysis. Nephron 17:241-258, 1976.
Kanis J.A.: Ostecmalacia and chronic renal failure. J. C/in. Patho/. 34:12951307, 1981.
Kanis J.A., Adams N.D.. Earnshaw M., Heynen G.. Ledingharn J.G.G.. Oliver
D.O.. Russell R.G.G. andWoodsC.G.: Vitamin D. osteanafaciaarxf chronic
renal failure. In: Vitamin D. Biochemical, Chemical and Clinical Aspects
Reiated to Calcium Metabolism. A.W. Norman, K. Schaefer, J.W. Coburn,
H.F. deluca, D. Fraser, H.-G. Grigoleit and D. v Herrath, eds. Walter da
Gruyter, Berlin, 1977, pp. 671-673.
Kraut J.A.. Shtnaberger J.H., Singer F.R., Sherrard D.J., Saxtcn J., Miller J.H..
Kurokawa K. ard Coburn J.W.: Parathymid gland responsiveness to acute
hypocalcemia in dialysis ostecmalacia. Kidney lnt. 23:725-730, 1963.
Lil1ieR.D. and Fullmer H.M.: Histopathologic TechnicandPractkxlHistochemistry. New York, McGraw-Hill Book Ccmpany, 1976, p. 534.
Mafluche H.H., Ritz E., Lange H.R.. Kutschera J.. Hodgson M., Seifferf U. and
Schoeppe W.: Bone histology in Incipient and advanced renal failure. Kidney
Inr. 9:355-362, 1976.
MaloneyN.A.,OttS.M.,AlfreyA.C..MillerN.L.,CoburnJ.W.andSherrardD.J.:
Histofogical quantitation of aluminum in iliac bone frcm patients with renal
failure. J. Lab. C/in. Med. 88206-216, 1982.
Melsen F. and Mosekflde L.: Tetracycline double-labeling of iliac trabecular bone
in 41 normal adults. Calcif. Tissue Res. 26:99-l 02. 1978a.
Melsen F. and Mosekilda L.: Dynamic studies of trabecular bore formation and
osteoid maturation in normal ano certain pathological conditions. Me&b.
Bone 01s. Rel. Res. 1:45-48, 1978b.
Melsan F.. Melsen B., Mosekilde L. and Bergmann S.: Histaorphcmatric
anafysis of normaf bone from the iliac crest. Acta Patho/. Microbiof. Scand. [A]
86:70-61, 1978.
Mora Palma F.J.. Ellis H.A.. Cook D.B., Dewar J.H., Ward M.K., Wilkinson R. and
Kerr D.N.S.: Ostecmalacia in patients with chronic renal failure before diafy
sis or transplantation. Quart. J. Med. 52:332-348. 1983.

P. Nilsson et al.: PTH, aluminum,

and dialysis

bone disease

Morrisey J., Rothstein M., Mayor G. and Slatopolsky E.: Suppression of parathyroid hormone secretion by aluminum. Kidney hf. 22699-704, 1983.
Nielsen H.E.. Melsen F and Christensen M.S.: Interrelationships between calcium-phosphorus metabolism, serum parathyroid hormone and bone histomorphcxnetry in non-dialyzed ano dialyzed patients with chronic renal
failure. Mineral Hectfo/yte Mefab. 4: 113-I 22, 1980a.
Nielsen H.E., Rdmer F.K., Melsen F.. Christensen MS. and Hansen J.E.: lahydroxyvitamin D, treatment of nondiafyzed patients with chronic renal failure Effectson bone, mineral metabolism and kidneyfunction. C/in. Nephrol.
13:103-108, 1980b.
Nilsson P., Mafmaeus J.. Wide L. and Danielson B.G.: An investigation of biochemical and radiographic signs of hyperparathyroidism in chronic renal
disease. To be published.
0ttS.M.. Maloney N.A., Coburn J.W., Alfrey AC. andSherrard D.J.: The prevalence of aluminum deposition in renal osteodystrophy and its relation to the
response to calcitriol therapy. N. Engl. J. Med. 307:709-713, 1982.
Parkinson I.S., Feest T.G., Ward M.K., Fawcett R.W.P. and Kerr D.N.S.: Fracturing dialysis osteodystrophy and dialysis encephalopathy: an epidemiological survey. Lancet i, 406.409, 1979.
Pierides A.M., Edwards W.G.. Jr., Cullu U.S., Jr., McCall J.T and Ellis H.A.:
Hemodialysis encephalopathy
with osteomalacic fractures and muscle
weakness. Kidney ht. 15:115-l 24, 1980.

27

Siede W.H., Seiffert U.B., Bundschuh F, Malluchs H.H. and Schoeppe W.:
Alkaline phosphatase bone isoenzyme activity in serum in variis degrees
of micmmorphcmetricafly assessed renal osteopathy. C/in. Nephrol. 13:277281,198O.
Teitelbaum S.L.. Eergfeld M.-A., Freitag J., Hruska K.A. and Slatopolsky E.: Do
parathyroid hormoneand 1,2Sdihydroxyvitamin
D modulatetoneformatii
in uremia? J. C/in. Endocrinol. Metab. 51:247-251, 1980.
Walker G.S., Aaron J.E.. Peacock M., Robinson P.J.A. and Davism AM.: Diafysate aluminum concentration and renal bone disease. Kidney ht. 21:41 l415.1982.
Ward M.K., Ellis H.A.. Feest T.G. and Parkinson IS.. Osteomafacic dialysis
osteodystrophy: Evidence for a water-borne etiological agent, probably aluminum. Lancet 1:841-845. 1978.
Weinstein R.S.: Decreased mineralization in hemodiafysis patients after subtotal
parathymidectomy. C&if. Tissue ht. 34: 16-20. 1982.

Received: August 3, 1983

Revised: January 31, 1984
Accepted: April 9, 1984

Lea parametrea biochimiquea refl&ant Ihom6oataale calcique et phoaphor6a Sinai quo la fonction parathyrdidlenne ant 6t6 6tudi6s, parall6tetnant b
lhlatomo@om6trle oaaeuae aprea double maquage par is t6tracycllne et coloration de ialumlnium, chez 17 maladea Mmodialydr depulaa molna
6 mola et na pr6aentant pas de aymptomea oaaeux. Une corrdlatlon 6troite a 6t6 trouv6e entra lea taux a6rlquea de la Parathormona (PTH) duna Part,
lea surfaces de &sorption et leavlteaaea doat6oformation dmma part, 6 la fola mt nlveau tlaaulalre et au nlveau de iunit6 baaale multlcebutalre (UgM)
de remodelege. Lea maladea ont pu 6tre dlvla6a en un groups 6 haut nlvaw de remodelaga ayant una mln6rallaatlon normale, et un grOUpe L bae r&ear
de remodelege ayant un Important d6faut de mln6rallaatlon. La deuxleme groupe 6tait en outre carect6rld par un taux plus baa de PTH et una plus
grand0 extenalon dea surfaces trab6culalrea relatives coior6ea pour Ialuminium. Pour Ienaemble dea maiadea, lea surfaces reiatlvea color&a pour
iaiuminlum 6talent invemement propottlonnellas aux surfaces marqu6ea par la t6tracycll~ et h la v&easedoat6oformatlon6valu6e au nlveaude IUBM
et au nlveau tlaaulalre, mala aana relation avec Ianclennet6 dea dlalyaeaou la doaecumui6e dhydroxyde daluminlum lngti.
Cea donn6eamontrent
quo la PTH ou lea facteura 116s6 la PTH, outre leur effet actlveteur aur le remodelage oaaeux, atlmulent directement la formatlon oaaeuaa cheziea
diaiya68, et qua Iincorporation daiumlnium dana 10seat aaaoci6e 6 un d6tart progreaaff de la min6rallaation oaSeua&