British Journal of Rheumatology 1997;36:444–448
BODY COMPOSITION IN RHEUMATOID ARTHRITIS
R. WESTHOVENS, J. NIJS, V. TAELMAN and J. DEQUEKER
Arthritis and Metabolic Bone Disease Research Unit, K.U. Leuven, U.Z. Pellenberg, B-3212 Pellenberg, Belgium
SUMMARY
The objectives were to assess bone mineral density (BMD) at different body sites in rheumatoid arthritis (RA) patients related
to disease severity parameters, disease duration and corticosteroid intake, and to look for body composition measurements as
lean body mass (LBM) and per cent fat as possible prognostic factors in RA. Body composition values were measured
cross-sectionally in 89 RA patients and compared with 157 controls. Patients were divided into males and postmenopausal
females, ever steroid treated and never steroid treated. BMD values of all body sites were significantly lower compared to normals
in all subgroups, except for the lumbar spine (L2–L4) in all postmenopausal women and males never treated with steroids. There
was also no clear BMD decrease in the arms of male RA patients. LBM was significantly lower in all body parts compared
to controls, whereas the fat distribution ratio (FDR) showed a clear shift to abdominal in all patients. These are parameters
of chronic illness and a predictor of cardiovascular disease, respectively. BMD data confirm our previous data in different patient
groups (low at appendicular sites, normal lumbar BMD), but no clear influence of disease severity and steroid intake could be
found. Body composition data, as LBM and FDR, are also altered in RA patients: decreased LMB and more central FDR.
Their usefulness as prognostic markers in early RA patients needs to be clarified prospectively in these patient groups.
K : Rheumatoid arthritis, Bone mineral density, Lean body mass, Fat distribution ratio.
T presence of generalized osteoporosis in rheumatoid arthritis (RA) patients is still debated [1]. Most
studies have shown a trend toward lower bone mass in
RA groups, especially at appendicular sites [2–4]. In
our centre [5], as in others [2], we could not find lower
bone mass in the axial skeleton. Disease activity and
severity, as well as disease duration, may be factors
leading to osteoporosis in RA. The effects of
glucocorticosteroids on bone turnover are well known,
but their effect in RA is more controversial [6], and
influences the results of most studies of bone
densitometry in RA patients.
Since the introduction of total body DEXA
measurements, it has been possible to evaluate bone
mass at different sites throughout the body as well as
soft-tissue body composition [7]. Both are possibly
influenced by cytokines released in active disease as
well as by immobilization and treatment. Depletion of
lean body mass in RA as a measure of rheumatoid
cachexia has not been well studied until now. Most of
the time, anthropometric body composition measures,
underwater weighing or tetrapolar whole-body bioelectrical impedance assessment are used [8].
Bone mass decrease, as well as a decrease in lean
body mass, could be markers of disease severity and
could have early prognostic value for bad outcome.
The present study focuses on 43 male RA patients and
46 postmenopausal women with RA, compared to 175
female and 81 male controls. Cross-sectional data on
bone density at different sites are compared to controls,
as well as lean body mass and total fat. Patients are
divided into ever steroid taking and never steroid
treated, and the importance of disease severity is
assessed.
PATIENTS AND METHODS
Patients
Eighty-nine patients (43 males and 46 postmenopausal females) with RA according to the 1987
ARA criteria [9], regularly seen at our rheumatology
department, had an evaluation of body composition.
Of the 43 male patients, 28 never took steroids, while
26 of the 46 postmenopausal female patients never took
steroids. Patient data concerning age, height, weight,
disease duration and years since menopause are
depicted in Table I. Steroid intake was documented
from the patients’ charts as months of steroid intake
and mean daily dose in milligrams of prednisone
equivalent: 6.7 2 1.9 mg/day for 64.8 2 88 months in
female patients and 8.5 2 3.0 mg/day for 24.1 2 14.6
months in male patients. Neither oral, intra-articular,
i.m. or i.v. therapy was ever used in the non-steroidtreated groups. In the same table, some direct and
indirect markers for disease severity at the time of body
composition measurements are shown: Steinbrocker’s
functional index [10], number of hospitalizations for
RA, number of orthopaedic interventions, number of
previous second-line drugs and rheumatoid factor
positivity, all derived from the patient charts.
The control group consisted of 175 females and 81
males of all ages starting from 20 yr, drawn from the
general population and also measured at our
department (not published). The mean age of normal
postmenopausal females q50 yr (n = 110) was
62.4 2 8.54; the mean age of normal males q50 yr
(n = 71) was 62.8 2 9.74. Normal body composition
values were determined per age-decade.
Submitted 21 May 1996; revised version accepted 24 September
1996.
Correspondence to: R. Westhovens, Division of Rheumatology,
U.Z. Pellenberg, Weligerveld 1, B-3212 Pellenberg, Belgium.
= 1997 British Society for Rheumatology
444
445
WESTHOVENS ET AL.: BODY COMPOSITION IN RA
TABLE I
Patient characteristics. Values are expressed as mean 2 ..
Postmenopausal females
Males
No corticosteroids
(n = 26)
Corticosteroids
(n = 20)
No corticosteroids
(n = 28)
Corticosteroids
(n = 15)
Age (yr)
Height (cm)
Weight (kg)
Disease duration (months)
Months since menopause
% rheumatoid factor-positive patients
Months of steroid intake
Daily dose of steroids (mg prednisone equivalent)
63.3 2 8
158 2 0.08
63.0 2 8
153 2 133
217 2 117
73.1
–
–
58.9 2 7
161 2 0.06
65.7 2 13
207 2 130
154 2 99
74
64.8 2 88
6.7 2 1.9
55.9 2 11.9
172.6 2 6.6
71.9 2 10.9
80.8 2 75
60.5 2 9.2
168.9 2 6.9
69.2 2 10.5
114.0 2 101
67.9
–
–
92.9
24.1 2 14.6
8.5 2 3.0
Severity of disease
No. of second-line drugs previously taken for RA
No. of hospital admissions for RA
No. of orthopaedic interventions
Steinbrocker’s functional index
2.4 2 1.5
1.7 2 1.3
0.59 2 1.3
2.38 2 0.8
3.7 2 1.5
2.6 2 1.8
0.97 2 1.5
2.76 2 0.6
1.9 2 1
2.0 2 2.1
0.29 2 0.6
2.17 2 0.8
(
((
((
°
°°
3.5 2 1.7
2.1 2 1.5
0.33 2 0.5
2.71 2 0.5
*P Q 0.002; **P Q 0.05.
°P Q 0.01; °°P Q 0.025.
Methods
Because of upgrading, two different densitometers
were used for each analysed region of interest. A DP-4
and DPX instrument from Lunar measured bone
mineral density (BMD), per cent fat, lean body
mass for total body, arms, legs and trunk. The fat
distribution ratio (FDR) was calculated as the fat mass
of the trunk divided by the sum of the fat mass of arms
and legs. For lumbar spine BMD, a Lab 22-A from
Novo and DPX from Lunar were used, while the
peripheral measurements (BMD radius 8 cm) were
obtained on the Norland Cameron and on the Lunar
SP-2 densitometer. Reproducibility for DP4 total body
measurements was published previously [11], as was
reproducibility for lumbar spine and femoral neck
acquisition on DP4 and DPX systems [12].
Evaluation was carried out on data obtained after
conversion with formulae constructed in our own
laboratory (not published), except for the total body
measurements [13]. Radius: BMDSP2 = 0.2467 + 0.649
BMDCameron; spine BMDDPX = 0.071 + 1.271 BMDNovo.
Only converted data were compared with normals
measured on the one densitometer, all data were
converted to and used as ‘the standard’.
Statistics
For all study variables, Z scores were evaluated
against zero by Student’s t-statistics. Differences in the
analysed variables for corticosteroid treatment were
evaluated by Student’s t-statistics with 5% probability
limits.
RESULTS
The characteristics of the 89 patients are shown in
Table I. Compared with never steroid users, there was
a trend to longer disease duration in the corticosteroidtreated groups, although this was not statistically
significant. There was no statistical difference between
corticosteroid-treated and never-corticosteroid-treated
patients concerning age, height, weight and rheumatoid
factor positivity.
As one would expect, patients ever treated with
corticosteroids had more severe disease compared to
the never corticosteroid users, measured by Steinbrocker’s functional index (P Q 0.05 in females and
P Q 0.025 in males), number of previously used
second-line drugs (P Q 0.002 in females and P Q 0.01
in males) and number of hospital admissions (P Q 0.05
in females). There was a trend to more orthopaedic
interventions in the steroid group.
BMD values of postmenopausal female RA patients
were significantly lower at all body sites, except for the
lumbar spine (L2–L4), compared to the age- and
TABLE II
Postmenopausal female RA patients: Z-BMD, Z-% fat, Z-lean body
mass. Values are expressed as mean Z scores 2 .. P values are
versus the age- and sex-matched controls. There were no statistically
significant differences between the two subgroups for any value
Never-steroidtreated cases
(n = 26)
Ever-steroidtreated patients
(n = 20)
Z-BMD
Total body
Arms
Legs
Trunk
Radius 8 cm
Spine
− 0.87 2 1.1*
− 2.50 2 2.2**
− 4.11 2 2.1**
− 0.82 2 1.0°
− 1.13 2 1.7°°
− 0.03 2 1.2 (NS)
− 0.86 2 1.7+
− 2.33 2 1.9**
− 3.84 2 2.5**
− 0.58 2 1.0++
− 1.05 2 2.1+
+ 0.03 2 1.2 (NS)
Z-% fat
Total body
Arms
Legs
Trunk
+ 0.14 2 0.5 (NS)
+ 1.14 2 0.5**
− 0.60 2 0.5**
+ 0.25 2 0.5+
+ 0.41 2 0.7++
+ 1.16 2 0.5**
− 0.52 2 0.3*
+ 0.38 2 0.6+
Z-lean body mass
Total body
Arms
Legs
Trunk
− 1.46 2 0.7**
− 0.36 2 0.7++
− 2.05 2 0.9**
− 0.76 2 0.6**
− 1.39 2 0.7**
− 0.67 2 0.7*
− 2.09 2 0.8**
− 0.60 2 0.3**
*P Q 0.001; **P Q 0.0001.
°P Q 0.0005; °°P Q 0.002.
P Q 0.05; + +P Q 0.02.
+
446
BRITISH JOURNAL OF RHUEMATOLOGY VOL. 36 NO. 4
TABLE III
Male RA patients: Z-BMD, Z-% fat, Z-lean body mass. Values are
expressed as mean Z scores 2 .. P values are versus the age- and
sex-matched controls
Never-steroidtreated cases
(n = 28)
Ever-steroidtreated patients
(n = 15)
Z-BMD
Total body
Arms
Legs
Trunk
Radius 8 cm
Spine
− 0.61 2 1.0°°
+ 1.05 2 2.3°
− 1.28 2 2.9°
− 0.04 2 0.8 (NS)
− 0.28 2 1.3 (NS)
− 0.34 2 1.0 (NS)
Z-% fat
Total body
Arms
Legs
Trunk
+ 1.22 2 1.5+
+ 2.42 2 2.1**
+ 1.13 2 1.4+
+ 1.05 2 1.2**
+ 0.64 2 1.3 (NS)
+ 2.20 2 2.2*
+ 0.68 2 1.1 (NS)
+ 1.08 2 1.2+++
Z-lean body mass
Total body
Arms
Legs
Trunk
− 2.61 2 2.6**
− 1.08 2 1.5**
− 2.75 2 2.3**
− 1.80 2 2.2**
− 1.99 2 1.8§
− 1.45 2 1.1*
− 2.44 2 1.7°°°
− 1.56 2 1.8+++
¶
− 0.77 2 0.9°°
+ 0.68 2 1.7 (NS)
− 1.15 2 2.2 (NS)
− 0.69 2 0.5++
− 0.17 2 1.0 (NS)
− 0.65 2 0.8°°
*P Q 0.001; **P Q 0.0001.
°P Q 0.05; °°P Q 0.005; °°°P Q 0.0005.
+
P Q 0.0003; ++P Q 0.0002; +++P Q 0.02.
§P Q 0.0025.
¶Statistically significant differences between the two subgroups
(P Q 0.01).
sex-matched controls. Z scores are depicted in Table II.
There were no statistically significant differences
between the steroid-treated and never-steroid-treated
patients. Per cent fat was higher in the patient group
at all body sites, except for the legs, in the
steroid-treated as well as in the non-steroid-treated
subgroups. Highest values were found in the arms.
Z-% fat scores are also given in Table II. There were
no differences between steroid-treated and never-steroid-treated patients. Lean body mass was lower at all
body sites, compared to normals at the same amount
in both subgroups. Values were least significant in the
arms (see Table II).
In male RA patients, BMD changes vs age- and
sex-matched controls were less prominent. BMD was
reduced at all body sites, except for the arms, where
there was a slight increase. Z-BMD values are depicted
in Table III. BMD of the spine (L2–L4) was only
significantly lower for the ever-steroid-treated subgroup (P Q 0.005).
The only statistically significant difference between
the steroid- and non-steroid-treated groups was found
for Z-BMD of the trunk (P Q 0.01).
Per cent body fat in male RA patients was
significantly higher at all body sites, even more
TABLE IV
Fat distribution ratio (Z-FDR mean 2 ..)
{
|
Never steroid treated
Ever steroid treated
P
1.45 2 0.43
1.26 2 0.26
1.67 2 0.17
1.33 2 0.26
NS
NS
prominent in the non-steroid-treated subgroup (no
statistically significant differences between the subgroups). See also Table III for Z-% fat values.
Lean body mass was significantly lower at all body
sites, compared to normal controls in both subgroups.
The FDR for females, as well as for males, showed
a clear shift to abdominal for RA patients compared
with age- and sex-matched controls. Z-FDR was
higher in steroid-treated patients, reaching some
statistical significance in males (P Q 0.02) (Table IV,
Fig. 1).
There were no clear statistically significant correlations between measured parameters of disease
severity, disease duration, cumulative dose of steroid
intake and Z-BMD, Z-% fat and Z-lean body mass.
DISCUSSION
This study confirms in RA patients a reduction in
BMD at all peripheral sites, but not in the lumbar
spine, except for ever-steroid-treated males, but this
should be interpreted with caution because of the small
size of the subgroup.
We were unable to show any influence of
low-dose-steroid taking on BMD values in men or
postmenopausal females, despite the fact that steroidtaking patients also had a longer disease duration and
had somewhat more severe disease. This is in line with
previous data from our centre in another patient group
[5].
One of the possible explanations is that we did not
divide the study group into current and previous
steroid users, so part of the bone loss could have
recovered, but the most likely reason could be that the
total amount of steroid intake, one of the most
important predictors of fracture risk [14], is not that
high in our patients. DEXA techniques in the lateral
position could be more sensitive at showing trabecular
bone loss, as shown by Reid et al. [15] and Buckley
et al. [16]. Laan et al. [17] found a decreased spinal
BMD in corticosteroid-treated patients, using dualenergy quantitative computerized tomography, but one
has to be aware of the limitations of this method [18].
The final outcome of the corticosteroid effect on
bone, namely fracture, was not carefully assessed in our
study, but patient height, as an indirect reflector of
fracture, did not differ between the subgroups.
Probably the most important limitation of this study is
the long disease duration and the cross-sectional
setting, while even in longitudinal studies often no
increased generalized bone loss was found in RA
patients [19]. Only in recent-onset RA patients could
the generalized loss of bone due to steroid treatment or
RA activity perhaps be determined in a prospective
way, as was shown by Gough et al. [4].
We could not demonstrate a correlation of BMD
values with indirect measures of disease severity. Other
centres could demonstrate that immobility and disease
severity were important factors influencing bone loss in
RA patients, especially as measured by BMD of the
femur [4, 20]. Like Gough et al. [4], we could speculate
that if steroids satisfactorily suppress inflammation,
WESTHOVENS ET AL.: BODY COMPOSITION IN RA
447
F. 1.—Normalized FDR ratios for both sexes.
this may outweigh the deleterious effect of these drugs
on bone metabolism. A difference in mobility, although
between ever-corticoid-treated and never-corticoidtreated patients, as measured by the Steinbrocker
Index, we would expect to be reflected in different bone
density measures.
Loss of lean body mass is a hallmark of acute and
chronic illness, but also of aging. Changes in growth
hormone production, physical activity and cytokine
production can play an important role in influencing
lean body mass [21]. Available DEXA techniques are
useful to determine soft-tissue composition [7, 22]. As
the loss of lean body mass, an important predictor of
survival in patients with AIDS [23] and cancer [24], has
been linked to the cytokine tumour necrosis factor
(TNF-a), also a pivotal cytokine in RA, it could be
interesting to look for those values in RA patients as
well. Preliminary data on small RA patient samples
showed that lean body mass was markedly decreased
and inversely associated with parameters of disease
activity and disability [8]. In a monozygotic co-twin
control study, total lean soft tissue and lower limb lean
soft tissue were significantly reduced [25] in the affected
sibs.
Using total body density techniques, we can confirm
in this larger RA patient group an important decrease
in lean body mass at all body sites compared to
normals, in male and female patients, as well as in
ever-steroid-treated and never-steroid-treated subgroups. No statistically significant difference could be
found between the steroid-treated and non-steroidtreated subgroups despite greater disease duration and
even more severe disease in the latter.
Inversely, fat mass, the other major mammalian
body compartment, is increased at all body sites, except
for the legs, resulting in a clear shift to abdominal for
the FDR in RA patients, most prominent in the
ever-steroid-taking subgroup, as one would expect. In
other non-RA patient groups, central fat distribution
correlated positively with a bad lipoprotein profile and
coronary heart disease [26].
All these body composition data could possibly be
important markers in RA, a disease that has a high
mortality rate [27, 28], with a particular excess of
cardiovascular deaths [29]. To determine their real
prognostic significance and correlation with disease
activity and severity parameters, prospective work has
to be carried out in patients with recent-onset RA
taken into longitudinal follow-up.
R
1. Dequeker J, Maenaut K, Verwilghen J, Westhovens R.
Osteoporosis in rheumatoid arthritis. Clin Exp Rheumatol 1995;13(suppl. 12):S21–6.
2. Hall GM, Spector TD, Griffin AS, Jawad ASM, Hall
ML, Doyle DV. The effect of rheumatoid arthritis and
steroid therapy on bone density in postmenopausal
women. Arthritis Rheum 1993;36:1510–6.
3. Garton MJ, Reid DM. Bone mineral density of the hip
and of the anteroposterior and lateral dimensions of the
spine in men with rheumatoid arthritis. Arthritis Rheum
1993;36:222–8.
4. Gough AKS, Lilley J, Eyre S, Holder RL, Emery P.
Generalised bone loss in patients with early rheumatoid
arthritis. Lancet 1994;344:23–7.
5. Verstraeten A, Dequeker J. Vertebral and peripheral
bone mineral content and fracture incidence of
postmenopausal patients with rheumatoid arthritis: effect
of low dose corticosteroids. Ann Rheum Dis
1986;45:852–7.
6. Dequeker J, Westhovens R. Low dose corticosteroid
osteoporosis in rheumatoid arthritis and its prophylaxis
and treatment: bones of contention. J Rheumatol
1995;22:1013–9.
7. Jensen MD, Kanaley JA, Roust LR et al. Assessment of
body composition with use of dual-energy X-ray
absorptiometry: evaluation and comparison with other
methods. Mayo Clin Proc 1993;68:867–73.
8. Roubenoff R, Roubenoff RA, Ward LM, Holland SM,
Hellmann DB. Rheumatoid cachexia: depletion of lean
body mass in rheumatoid arthritis. Possible association
with tumor necrosis factor. J Rheumatol 1992;19:1505–
10.
9. Arnett FC, Edworthy SM, Block DA et al. The
American Rheumatism Association 1987 revised criteria
for the classification of rheumatoid arthritis. Arthritis
Rheum 1988;31:315–24.
10. Steinbrocker O, Traeger CH, Batterman RC. Therapeutic criteria in rheumatoid arthritis. J Am Med Assoc
1949;140:659–62.
448
BRITISH JOURNAL OF RHEUMATOLOGY VOL. 36 NO. 4
11. Nijs J, Geusens P, Dumortier F, Borghs H, Dequeker J.
Comparison of DEXA and DPA in clinical situations. In:
Christiansen C, Overgaard K, eds. Osteoporosis 1990.
Copenhagen: Osteopress ApS, 1990:726–7.
12. Nijs J, Geusens P, Dequeker J, Verstraeten A.
Reproducibility and intercorrelations of total bone
mineral and dissected regional BMC measurements. In:
Dequeker J, Geusens P, Wahner HW, eds. Bone mineral
measurements by photon absorptiometry. Leuven:
Leuven University Press, 1988:451–3.
13. Russell-Aulet M, Wang J, Thornton J, Pierson RN.
Comparison of dual photon absorptiometry systems for
total body bone and soft tissue measurements: dual
energy X-rays versus guadolinium 153. J Bone Miner Res
1991;6:411–5.
14. Dykman TR, Gluck OS, Murphy WA, Hahn TJ, Hahn
BH. Evaluation of factors associated with glucocorticoid
induced osteopenia in patients with rheumatic diseases.
Arthritis Rheum 1985;28:361–8.
15. Reid IR, Evans MC, Stapleton J. Lateral spine
densitometry is a more sensitive indicator of glucocorticoid-induced bone loss. J Bone Miner Res 1992;7:1221–5.
16. Buckley LM, Leib ES, Cartularo KS, Vacek PM, Cooper
SM. Effects of low dose corticosteroids on the bone
mineral density of patients with rheumatoid arthritis. J
Rheumatol 1995;22:1055–9.
17. Laan RFJM, Van Riel PLCM, Erning LJThO, Lemmens
JAM, Ruys SHJ, Van de Putte LBA. Vertebral
osteoporosis in rheumatoid arthritis patients: effect of
low dose prednisone therapy. Br J Rheumatol
1992;31:91–6.
18. Karantanas AH, Kalef-Ezra J, Glaros D. Limitations of
quantitative CT in corticosteroid induced osteopenia.
Acta Radiol 1991;32:339–41.
19. Sambrook PM, Cohen ML, Eisman JA, Pocock NA,
Champion GD, Yeates MG. Effects of low dose
corticosteroids on bone mass in rheumatoid arthritis: a
longitudinal study. Ann Rheum Dis 1989;48:535–8.
20. Laan RFSM, Buys WCAM, Verbeek ALM et al. Bone
mineral density in patients with recent onset rheumatoid
arthritis: influence of disease activity and functional
capacity. Ann Rheum Dis 1993;52:21–6.
21. Roubenoff R. Hormones, cytokines and body composition: can lessons from illness be applied to aging? J Nutr
1993;123:469–73.
22. Lukaski HC. Soft tissue composition and bone mineral
status: evaluation by dual-energy X-ray absorptiometry.
J Nutr 1993;123:438–43.
23. Kotler DP, Thierney AR, Pierson RM. Magnitude of
body cell mass depletion and the timing of death from
wasting in AIDS. Am J Clin Nutr 1989;50:444–7.
24. De Wys WD, Begg C, Lavin PT et al. Prognostic effect
of weight loss prior to chemotherapy in cancer patients.
Am J Med 1980;69:491–7.
25. Sambrook PH, Spector TD, Seeman E, Bellamy N,
Buchanan RRC, Duffy DL et al. Osteoporosis in
rheumatoid arthritis. Arthritis Rheum 1995;38:806–9.
26. Haarbo J, Hassager C, Riis BJ, Christiansen C. Relation
of body fat distribution to serum lipids and lipoproteins
in elderly women. Atherosclerosis 1989;80:57–62.
27. Symmons DPM. Mortality in rheumatoid arthritis. Br J
Rheumatol 1988;27(suppl.):44–54.
28. Pincus T, Callahan LF. Mortality in rheumatoid arthritis
over 15 years. Arthritis Rheum 1990;33:S140.
29. Wolfe F, Mitchell DM, Sibley JT et al. The mortality of
rheumatoid arthritis. Arthritis Rheum 1994;37:481–94.