Synovial Iron Deposition in Rheumatoid Arthritis.

Synovial Iron Deposition in Rheumatoid Arthritis
Robert M. Bennett, E.D. Williams, S.M. Lewis and P.J.L. Holt
An isotope ferrokinetic study has been used to study the rate of iron deposition in 13 rheumatoid knees. Iron accumulation occurred only after
incorporation of labelled iron into circulating erythrocytes, suggesting
that intermittent intraarticular hemorrhages were the source of the deposits. The mean iron accumulation was 1.25 mg/knee/day representing a
mean intraarticular blood loss of 3.58 ml. There was no correlation between the degree of iron deposition and the hemoglobin, platelet count
or serum iron.
T h e presence of hemosiderin deposits in the
synovial membrane of rheumatoid joints has
long been recognized by pathologists (1). T h e
mean concentration of iron in rheumatoid synovial membrane was found to be 347 pg/g dry
tissue compared with 15.2 pg/g in normal synovium (2), and it has been suggested that these
deposits are unavailable for erythropoiesis possibly contributing to the anemia of rheumatoid
arthritis (3). Controversy has arisen over
whether these deposits exert a beneficial effect
on the course of the arthritis (4) or are deleterious (5), and theories on their origin include
a direct uptake from the plasma (6), a result of
chronic intraarticular bleeding (3) and phag-
ocytosis of leukocytes containing iron bound to
cytoplasmic lactoferrin (4). Using an isotope kinetic method we have studied the rate of iron
deposition in synovial membrane, its mechanism and whether there is any relationship to
development of anemia in patients with rheumatoid arthritis.
From the Departments of Medicine, Medical Physics and
Haematology, The Royal Postgraduate Medical School,
Hammenmith, London, W.12.
This work was supported by the Arthritis and Rheumatism Council for Research.
ROBERT M . BENNETT, MRCP: Registrar and Tutor in Medicine, Department of Medicine, Royal Postgraduate Medical School; E. D. WILLIAMS, MA, MSC: Senior Physicist, Department of Medical Physics, Royal Postgraduate Medical
School; s. M . LEWIS, MD, MRC (PATH): Reader in Haematology, Department of Haematology, Royal Postgraduate
Medical School; P. J. L. HOLT, MRCP: Consultant and Lecturer in Medicine, Department of Medicine, Royal Postgraduate Medical School.
Reprint requests should be addressed to: Dr. Robert M.
Bennett, The University of Chicago, Department of Medicine, Box 404,950 East 59th Street, Chicago 111 60637.
Submitted for publication July 11, 1972; accepted Dec 1,
1972.
Isotope Administration. Ten microcuries of 59Fein
the form of ferric citrate was incubated with the patients'
plasma for 30 minutes at 37" C in order to label the transferrin. An accurately measured volume was injected intravenously.
298
MATERIALS AND METHODS
Subjects. Patients with definite rheumatoid arthritis (7) involving the knees were studied. For comparison, 2
patients with osteoarthrosis of the knees were used as controls. All patients were ambulant within the hospital confines. All the rheumatoid patients had a full hematologic
profile including bone marrow aspiration; they had not
taken any iron preparation for at least 3 months.
Counting. Blood samples were taken every 10 minutes
for the first hour and every 3 to 10 days over the course of
the ensuing 50 to 90 days. Surface counting was performed
at similar intervals over the affected knees, the thighs and
sacrum by means of a collimated scintillation counter. Sufficient counts were obtained to give an error of less than 3%.
All counts were adjusted for background and decay.
Calculations. These are largely based on the assumption that iron accumulation in a rheumatoid knee has a
Arthritis and Rheumatism, Vol. 16, No. 3 (May-June 1973)
SYNOVIAL IRON DEPOSITION IN RA
150
SACRUM
100
I
60
1w
50
0
KNE
1%
1W
m
THIGH
60
Fig 1. Ferrokinetic study
in a rheumatoid patient
showing surface counting
and blood activity.
do
03060w120
1
10
20
30
Fig 2. lntraarticular accumulation of iron after an
intravenous
injection
of
59Fe-labelledtransferrin i n 2
patients-1 with rheumatoid
arthritis (W) and the other
with osteoarthrosis of the
knees ( 0 ) .
40
M
Days
Minutes
1
Arthritis and Rheumatism, Vol. 16, No. 3 (May-June 1973)
10
15
w
m
23
299
BENNETT ET AL
monoexponential increase of the type y = 1 - e-x. Reference to Figure 1 (data from a rheumatoid patient) shows
that iron rapidly leaves the plasma over the first 24 hours
and appears in the bone marrow (sacral trace). T h e counts
over the involved knee and thigh (representing blood flow)
also decrease over the first 24 hours. Over the subsequent
50 days, as iron leaves the marrow and becomes incorporated into circulating red blood cells, there is an increased
activity over the thigh and knee. However, with time the
knee trace continues to increase while the thigh trace falls
gradually. This is attributed to the deposition of iron within
the knee. By equating the thigh counts (representative of
blood flow) to the knee counts on Day 1 and using the same
coefficient for the thigh counts on subsequent days, a correcfed count is obtained. Substracting this from the knee
counts gives an excess count in the knee which is proportional to iron deposition within the knee. When plotted
against time the excess knee counts (Figure 2) are seen to
rise steeply and then level off to a plateau. By comparison, 2
osteoarthritic knees showed no accumulation of iron. A diagrammatic representation of the accumulation of iron in a
rheumatoid knee is seen in Figure 3, from which the following mathematical analysis is derived:
Ai = accumulated knee counts at time i days
P = plateaucounts
Ki = excess knee counts at i days (KO being 24
hours post isotope injection)
r = fractional rate of accumulation of final activity in knee
t
= timeindays
B = percentage of blood-borne iron accumulated per day
C = percentage of total blood-borne iron in
knee
Then
A;=Ki-Ko
also
~i = P (1 - e - n )
or
P - ~i = Pe-"
therefore
log 10 ( P - Ai) a -rt - - _
_ ___-- ---__-__ ---- 1
-ie, plotting log,,, (P - Ai) against t gives a line
slope -r, using the least squares fit of a straight
line.
This slope was derived graphically from
-,.
-
0.693
%t
knowing -r:
B -
r
100
P x c ----------2
KO
As 1 g/lOO ml of hemoglobin (Hb) contains 0.335
g of iron the total blood-borne iron
(T)
=
0.355 x blood volume x H b
100
----- 3
Thus absolute iron accumulation per day
(F) =
Xg
____________
4
100
Assuming that all this iron is derived from red
Excess
Counts
Fig 3. Diagramatic representation of iron accumulation in a rheumatoid knee. The curve is a
monoexponential function
of the type y = 1 = e-'.
Arthritis and Rheumatism, Vol. 16, No. 3 (May-June 1973)
SYNOVIAL IRON DEPOSITION IN RA
Fig 4. Semilogarithmic plot of
excess knee counts against
time, in a rheumatoid patient
(Case C).
rheumatoid knees have a similar monoexponential increase, in that when plotted on a
(I)=
F X 1 0 0 100
~
ml _ _ _ _ _ _ _ _ _ _ _ 5
logarithmic scale against time, the points lie on
Hb x 0.335
a straight line, well within the conventional levels of statistical significance. T h e three osT h e value “C” was obtained by injecting 20 ml of autologous blood (with 100% erythrocyte incorporation of 59Fe) teoarthritic knees show no such monointo the knee joints of 2 patients undergoing synovectomy exponential increase. T h e coefficient of “x” in
the next day. “B” was then directly calculated from the increase in count rate and thus, by substitution in Formula 2, the regression equations is the “r” of Formua value for “C” was obtained. T h e values for the two knees las 1 and 2. From this value the percentage of
were 0.43% and 0.2570, the mean value of 0.34% was used
blood-borne iron accumulated per day (B), the
in all calculations. Total blood volume was calculated from absolute iron accumulation per day (F) and the
the packed cell volume and plasma volume (derived from
intraarticular blood loss (I) are calculated from
the initial dilution of labelled transferrin).
Formulas 2, 4 and 5. T h e values are seen in
Table 2. T h e mean iron accumulation per day
RESULTS
for the 13 rheumatoid knees was 1.25 mg
Accumulation of iron in the affected knees (+ 0.7 2 SD). If this were all derived from red
only occurred after it had been completely blood cells, as suggested by the fact that accucleared from the plasma and had become incor- mulation only occurred after incorporation of
iron into erythrocytes, it represents a mean
porated into the circulating red cells (Figure 1).
T h e results of plotting log,c(P - Ai) against daily intraarticular blood loss of 3.58 ml (k9.0
“t” for a rheumatoid knee is seen in Figure 4. 2 SD). There was no obvious correlation beT h e regression equations of the 13 rheumatoid tween iron deposition and the degree of inflamknees are given in Table 1. It is seen that all the mation in the individual patients, but in those
blood cells the intraarticular blood loss per day
Arthritis and Rheumatism, Vol. 16, No. 3 (May-June 1973)
301
BENNETT ET AL
Table 1. RegressionEquationsfor iron Accumulation in 13 Rheumatoid
Knees and 3 OsteoarthriticKnees
Patient
Diagnosis"
RA
RA
RA
RA
E
RA
F
RA
RA
OA
OA
Regression
equation
Coefficient of
correlation ( I )
Significance
+
+
y =- 2.6~
+ 5272
y =- 3.9~
+ 4266
y = - 4 . 8 ~+ 2630
y = - 4.6~
+ 3467
v = - 3 . 2 ~+ 3890
y = -14.5~+ 4677
y =- 4.1~
+ 3388
y =- 3 3 +
~ 5248
y =- 2 6+
~ 3236
y =- 3 2+
~ 3890
y =- 2 6 +
~ 3631
y = - 0 . 7 ~+ 613
y=- 0.7~
+ 389
y =- 1.8~
+ 1445
-0.97
-0.89
o.ooooo1
0.0006
-0.96
-0.94
0.0000001
0.000007
-0.95
-0.97
0.000007
0.000002
-0.95
0.0000006
-0.89
-0.6
0.001
0.05
-0.85
-0.73
0.03
0.04
-0.96
-0.94
0.003
0.007
-0.34
-0.49
0.13
0.20
-0.36
0.24
y = -10.3~ 1866
y = - 3 . 5 ~ 2630
(0
"RA = Rheumatoid arthritis
OA = Osteoarthritis
Table 2. iron Accumulation and Bleeding in 13 RheumatoidKnees
Patient
Knee
Accumulation
of iron/day (TO)
A
R
L
R
L
R
L
L
R
L
R
L
R
L
0.054
0.014
0.025
0.022
0.026
0.023
0.024
0.342
0.074
1.034
0.523
0.76
0.57
B
C
D
E
F
G
302
Absolute
accumulation
of ironlday (mg)
0.32
0.08
0.18
0.16
0.18
0.16
0.12
1.61
0.35
4.60
2.30
3.5
2.7
Bleeding into
knee/day (ml)
0.91
0.23
0.43
0.38
0.44
0.39
0.36
3.96
0.86
13.80
6.90
10.1
7.9
Arthritis and Rheumatism, Vol. 16, No. 3 (May-June 1973)
SYNOVIAL IRON DEPOSITION IN RA
patients showing a marked difference between
the knees (Subjects E, F and G) the more inflamed knee showed the higher iron accumulation.
All the rheumatoid patients were anemic
with a low serum iron although total iron binding capacity was normal and all had plentiful
iron stores in the bone marrow. There was no
correlation between the degree of intraarticular
bleeding and the hemoglobin: y = - 0 . 1 8 ~ +
11.3 ( r = 0.49, P > 0.1) or the serum iron y =
- 0 . 0 5 ~ + 56 ( r = -0.2, P > 0.1). Similarly,
no correlation was found between the platelet
count and degree of intraarticular bleeding: y =
0.01~
+ 385,000 ( T = 0.41, P > 0.1).
DISCUSSION
T h e validity of these results is based on two
assumptions. First, that the vascularity of the
knee remains unchanged throughout the study:
in so far as these patients were all in hospital
having supervised therapy it is likely that the
vascularity would, if anything, regress slightly,
a change that would lead to an underestimation
of iron deposition in subsequent calculations.
Second, that the percentage of blood-borne iron
in the knee (0.34%),a figure derived from the
mean values for 2 rheumatoid knees from different patients, was the same for all the knees studied. Neither of these assumptions is likely to be
completely valid, although the errors involved
are of a small magnitude and as such our results
are probably a close approximation of intraarticular iron deposition.
As significant iron accumulation in rheumatoid knees only occurs after it has become incorporated into circulating erythrocytes, it is reasonable to suppose that the synovial iron
deposits accrue directly from hemoglobinbound iron rather than transferrin-bound iron.
Conceivably this could result from small hemorrhages within the synovial membrane with subsequent incorporation of the iron into subintimal macrophages, or from intraarticular
bleeding. We have previously shown in a
patient with a popliteal synovial fistula (8) that
intraarticular bleeding does occur and is directly related to exercise of the affected knee.
T h e average blood loss in that particular
patient, measured directly with 51Cr-labelled
red blood cells, was 2.5 ml/day, a figure agreeing closely with our results derived from ferrokinetic studies.
It is not surprising to find that there is no
relationship between the degree of intraarticular bleeding and the anemia of rheumatoid arthritis. Although quite large amounts of
iron may be sequestered in the synovial membrane of a rheumatoid patient, these patients
are seldom iron deficient as judged by their iron
marrow stores. Poor incorporation of iron into
developing erythroblasts despite the fact that
there are plentiful iron stores within the reticuloendothelial cells of the bone marrow has led to
the concept of a recticuloendothelial cellerythroblast block in the anemia of chronic inflammation (9- 1 1 ).
However, in rheumatoid patients with inadequate iron stores it is probable that intraarticular bleeding could tip the balance towards
overt iron deficiency. Whether under such circumstances, synovial iron deposits could be utilized for subsequent erythropoiesis remains to
be elucidated. T h e finding of a relationship between synovial iron deposits and the severity of
joint destruction (3) has led to a belief that the
accumulated iron has a cytotoxic effect in the
joint. An equally acceptable explanation would
be that the more severely involved joint will exhibit the most intraarticular bleeding and this,
compounded by a possible longer duration of
the disease in this situation, leads to the observed increase in iron deposits.
ACKNOWLEDGMENTS
We are grateful to Professor E.G.L. Bywaters for his advice and encouragement and Dr. J.M. Gumpel for allowing
us to study several of his patients.
REFERENCES
1 . Collins DH: Haemosiderosis and haemochro-
Arthritis and Rheumatism, Vol. 16, No. 3 (May-June 1973)
303
BENNETT ET AL
2.
3.
4.
5.
6.
matosis of synovial tissue. J Bone Joint Surg Br
33:436-441,1951
Senator GB, Muirden KD: Concentration of
iron in synovial membrane, synovial fluid and
serum in rheumatoid arthritis and other joint
diseases. Ann Rheum Dis 27:49-53, 1968
Muirden KD: The anaemia of rheumatoid arthritis: the significance of iron deposits in the
synovial membrane. Aust Ann Med 2:97-104,
1970
Mowat AG, Disney T F , Vaughan J H : Effect of
iron dextran, gold thiosulphate and hydrocortisone acetate on experimental synovitis in
the guinea pig. Ann Rheum Dis 30:187-193,
1971
Muirden KD, Senator CB: Iron in the synovial
membrane in rheumatoid arihritis and other
joint diseases. Ann Rheum Dis 27:38-48, 1968
Mowat AG, Hathersell T E : Nature of anaemia
in rheumatoid arthritis: VIII. Iron content of
304
synovial tissue in patients with rheumatoid arthritis and in normal individuals. Ann Rheum
Dis 27:345-350,1968
7. Ropes MW, Bennett GA, Cobb S, et al: Diagnostic criteria-1958 review. Ann Rheum Dis
18:49-53, 1959
8. Bennett RM, Williams ED, Lewis SM, et al:
Studies of a popliteal synovial fistula. Ann
Rheum Dis 31:482-486,1972
9. Freireich EJ, Miller A, Emerson EP, et al: The
effect of inflammation in the utilization of
erythrocyte and transferrin bound radioiron for
red cell production. Blood 12:972-983, 1957
10. Cartwright LE, Lee GR: The anaemia of
chronic disorders. Br J Haematol 2 1 :147- 152,
1971
11. Bennett RM, Holt PJL, Lewis SM: The kinetics of distribution, uptake and utilization of iron
dextran in rheumatoid arthritis. (Unpublished
data)
Arthritis and Rheumatism, Vol. 16, No. 3 (May-June 1973)