вход по аккаунту


Hydrostatic and oncotic determinants of microvascular fluid balance in normal canine joints.

код для вставкиСкачать
Water moves between plasma (p) and synovial
fluid (SF) in response to gradients in the balance of
opposing hydrostatic pressures (HP) and oncotic pressures (OP). At the vascular site where proximal filtration ceases and distal reabsorption begins, all forces are
theoretically in balance. At this point, the transitional
microvascular pressure (TMP) may be estimated from
the equation TMP = HPsF OP, - OPsF. We measured these forces in the shoulders, wrists, and knees of
10 normal dogs, ages 2-10 years. The mean HPsF in the
knee was lower than that in the shoulders and significantly lower than that in the wrist. Conversely, the
OPsF in the wrist was significantly lower than that in the
shoulder or the knee. These factors combined indicate
that the microvascular bed in the normal knee has a
remarkably low mean TMP (7.9 mm Hg). We also found
a strong positive correlation between the age of each dog
and the mean oncotic pressure of its SF.
According to classic Starling-Landis concepts,
fluid exchange between plasma and the interstitial
space is driven by the balance of hydrostatic and
oncotic pressures interacting across a semipermeable
Presented in part at the 49th Annual Meeting of the American Rheumatism Association, Anaheim, CA, June 1985.
From the Department of Medicine, University of Washington, Seattle.
Suppoped in part by NIH g a n t s AM-3281 1 and CA-31787,
an Arthritis Research Center grant from the Arthritis Foundation,
and a grant from Merck and Company.
Peter A. Simkin, MD: Professor of Medicine; Richard S - %
Benedict, BS: Research Technologist.
Address reprint requests to Peter A. Simkin, MD, Division
of Rheumatology, RG-28, University of Washington, Seattle, WA
Submitted for publication March 17, 1989; accepted in
revised form August 7, 1989.
Arthritis and Rheumatism, Vol. 33, No. 1 (January 1990)
capillary wall (1,2). Injoints, the interstitial forces may
be considered to be those measurable in synovial fluid
(SF). This holds true because synovial cells lack tight
junctions and have no underlying basement membrane
(3). As a result, SF is continuous with, and in most
respects an extension of, the interstitial fluid in synovial tissue (4). Starling forces are readily accessible for
measurement in normal joints, where the hydrostatic
pressure can be measured in situ (5-7),and the oncotic
pressure is easily determined in a synovial aspirate (8).
Interjoint variation in these factors within the same
subject implies that joints differ in their normal microvascular physiology.
Vessels remain patent because their internal
hydrostatic pressure (HP,, where p represents plasma)
is greater than that of the investing interstitium
(HPSF), although the magnitude of the gradient (HP, HP,,) decreases progressively from the arteriole to
the venule. Throughout the length of the vessel, however, a hydrostatic pressure gradient drives filtration
of water from the plasma into the surrounding tissue.
In opposition, a corresponding oncotic pressure (OP)
gradient (OP, - OP,,) draws interstitial fluid back into
the plasma. Relatively high HP, drives arteriolar filtration, i.e., (HP, - HPsF) > (OP, - OP,,), whereas
reduced HP, may lead to venular reabsorption, i.e.,
(HP, - HPSF) < (OP, - OPSF).
At the site of transition from filtration to reabsorption, all forces are in balance, i.e., (HP, - HPsF)
= (OP, - OP,,), and there is no net flow across the
vessel wall. At this one point in the microcirculation,
microvascular pressure can be calculated from the
other 3 Starling forces. Although this value has been
referred to by a number of different names, we prefer
the term transitional microvascular pressure (TMP),
because it is the intravascular pressure that pertains at
the theoretical point of transition from filtration to
reabsorption. The formula for calculation of the TMP
is as follows:
OP, - OPsF + HP,,
In the present study, we determined the hydrostatic pressure of synovial fluid within wrists, shoulders, and knees (stiflejoints) of 10 normal adult dogs.
We also assessed the oncotic pressure of plasma and
of synovial fluid aspirates. We then used these values
to calculate the TMP in each joint. Significant differences were found between the wrists and knees in
HPSF, OPsF, and TMP. In addition, there was a strong
positive correlation between the age of the animal and
the mean oncotic pressure of its synovial fluid.
Dogs used in this study were obtained from the
kennels of the Fred Hutchinson Cancer Research Center,
which houses a large, stable population of healthy adult
dogs. The ages, pedigrees, and transplantation antigens
(DLA) (9) of the dogs are known. This population provided
the opportunity to evaluate a range of ages and of tissue
types when a pilot study of synovial oncotic pressures in
normal pound animals revealed significant differences between dogs, and between the wrist and knee joints of the
same dogs (10). Ten dogs were selected: 3 littermate hounds,
1 hound/German shorthair cross, and 6 unrelated beagles.
There were 6 male dogs and 4 female dogs, ranging in age
from 2.5 years to 10.6 years.
The first 3 animals were sedated with intravenous
fentanyl citrate/droperidol (Innovar; Parke-Davis, Moms
Plains, NJ) and anesthetized over each joint with 1%
lidocaine. Since these dogs showed occasional restlessness,
subsequent animals were anesthetized with intravenous sodium pentobarbital to ensure effective relaxation. All specimens were obtained within 15 minutes of the induction of
anesthesia. Each dog was placed in a side-lying position and
shaved over the knee, wrist, and shoulder, and the area was
appropriately cleaned with Betadine solution (Purdue Frederick, Norwalk, CT). After the procedures were completed
on that side, the animal was rolled over and the contralateral
joints were similarly prepared for study.
A sterile, saline-filled, 21-gauge needle was introduced into each joint, and the resting pressure was recorded
through bubble-free pressure lines connected to a Bell &
Howell transducer. Special care was taken to record the
atmospheric pressure at the mid-level of each joint before
the needle was inserted and to subtract this value from the
measured resting pressure. Each determination included an
initial observation period of -30 seconds, which was occasionally extended when necessary to ensure a stable result.
The wrist joints were positioned in 30" of flexion with a
goniometer. As in the human knee ( I I ) , readings taken using
this position of slight flexion produced stable results under
low pressure conditions. Shoulders and knees were studied
with the joints in the relaxed position of slight flexion that
these animals naturally assume under anesthesia.
To ensure accurate determinations of intraarticular
pressure, the entire measuring apparatus was filled with
saline at the time of needle insertion. To avoid contamination by the saline, SF specimens were aspirated through a
separate sampling needle inserted after the pressure needle
was withdrawn. Thus, we could not confirm the position of
the pressure needle, and a few outlying values may reflect
the pressure within intraarticular soft tissue, rather than that
of SF. Although accurate positioning could almost always be
confirmed by palpation and by subsequent aspiration of SF
from the same site, this potential problem must be acknowledged in view of the very small volumes of fluid within
normal joints.
On completion of each reading, the pressure needle
was withdrawn, and a second 21-gauge needle was inserted
and used to aspirate synovial fluid into a tuberculin syringe.
The volume and appearance of each sample were recorded.
Specimens showing slight evidence of contamination by
blood were transferred to hematocrit tubes and centrifuged.
Grossly bloody samples were discarded. SF specimens and
concurrently obtained samples of heparinized plasma were
then assayed for oncotic pressure at 37°C by the method of
Wiederhielm et al, which permits accurate determinations
with very small fluid samples (12).
Articular hydrostatic pressure readings were
obtained from both right and left joints of all 10
animals. An aberrant value (-35 mm Hg) in the right
wrist of the first dog was not used in our analyses. The
mean + SEM of all other individual values was - 1.4 f
0.9 mm Hg, with a range of + 11 to - 15. Readings for
the individual dogs represent the mean of findings from
the left and right sides (Table 1). Mean values (?SEM)
for the entire group were then calculated from these
lefthight averages, and were as follows: 2.9 f 1.3 mm
Hg in the wrist, -2.8 -+ 1.7 mm Hg in the shoulder,
and -4.5 2 1.1 mm Hg in the knee (Figure 1). The
difference between wrist and knee pressures was significant by paired t-test ( t = 7.36, P < 0.001). The
hydrostatic pressure readings showed no significant
correlation with the age, sex, or size of the animals.
Usable samples were recovered from 49 of the
60 joints aspirated. Normal, highly viscous, crystalclear synovial fluid was recovered from 29 joints.
Eleven other fluids were essentially normal, although
blood tinged (hematocrit < 1%); 9 modestly contaminated fluids were also used, although their hematocrit
values ranged between 1% and 4%. Four specimens
were discarded because they were frankly bloody
(hematocrit >4%), and SF could not be aspirated from
7 joints.
Table 1. Data on the 10 normal dogs studied*
Mean 2 SEM
Synovial fluid OP (mm Hg)
Synovial fluid HP (mm Hg)
(mm Hg)
5.7 f 1.0
14.4 2 2.8
20.5 f 0.5
6.2 f 0.8$
7.9 2 0.7
8.0 rt 0.6
* OP = oncotic (colloid osmotic) pressure;
HP = hydrostatic presure; ND
- 13.0
-2.8 f 1.7
- 10.0
1 .O
2 1.18
not determined.
t Value for a single joint. Other synovial fluid OP and HP values are the mean of the left and right joints.
P < 0.001 versus the mean value in knees; P < 0.05 versus the mean value in shoulders.
8 P < 0.001 versus the mean value in wrists.
The small specimens of recovered SF were not
measured precisely, but their volumes were estimated
at the time of aspiration. Mean volumes from usable
joints were 0.10 ml in wrist samples, 0.12 ml in
shoulder samples, and 0.09 ml in knee samples. These
values did not differ significantly from each other.
They were smaller than those previously obtained
from dog cadavers, because of the gentle technique
required to minimize the risk of bleeding in these living
animals (10). The largest single specimen was 0.4 ml,
from an apparent effusion in the left wrist of dog 6, a
10.6-year-old beagle. In general, the largest volumes
were from large, young hounds and the smallest samples were from small, older beagles. More animals will
be required, however, to analyze the possible role of
weight, age, or breed as determinants of synovial fluid
volume in normal dogs.
Mean values for oncotic pressure of SF were
derived from the mean of right-side and left-side
values in 20 joints and from single determinations in 9.
No usable sample could be obtained from either knee
of dog 6 (Table 1). The mean of all SF oncotic pressure
values was 7.5 ? 0.4 mm Hg, while individual values
ranged from a low of 3.2 mm Hg in the left wrist of dog
8 to a high of 15.9 mm Hg in the apparent effusion
within the left wrist of dog 6. As in a previous study
from this laboratory (lo), oncotic pressures in wrist
synovial fluids were consistently lower than those in
SF from knees (t = 5.50, P < 0.001) or shoulders (t =
2.44, P < 0.05) (Figure 2).
When the oncotic pressure values of all SF
Figure 1. Hydrostatic pressure in synovial joints of 10 dogs. Except
for 1 wrist value in 1 dog (see Results), values represent the mean
pressure in the left and right joints of each animal. Synovial fluid
(SF) hydrostatic pressure was higher in the wrist than in the
shoulder or knee (stifle) joints.
Figure 2. Oncotic pressure in synovial fluids (SF) from 10 dogs.
With some exceptions (see Results), values represent the mean
pressure in the left and right joints of each animal. In 1 animal, a
usable sample could not be obtained from either knee. SF oncotic
pressure was higher in the knee and shoulder joints than in the
The oncotic pressure of any biologic fluid reflects both the quantity and the quality of its dissolved
macromolecules (13). In synovial fluid, the responsible
solutes are plasma proteins and hyaluronic acid. One
or both of these factors must vary systematically in
order to produce age-dependent and joint-dependent
differences in oncotic pressure. Other work from our
laboratory indicates that variation in protein, rather
than hyaluronate, content is responsible for the interarticular patterns we have found (Weinberger A,
Chunn L, Simkin PA; unpublished observations). We
don't yet know, however, which macromolecules
cause the OP of synovial fluid to increase with advancing age (Figure 3). We think it unlikely that this
relationship simply reflects the higher SF protein content that could be expected with a higher prevalence of
acquired joint disease in older animals. The chevron
appearance of the data in Figure 2 reflects the fact that
the OP in each of the 3 studied joints varies in concert
with the other 2. Older animals have high oncotic
pressures in all 3 joints, while low values are found
with corresponding consistency in younger dogs.
In addition to finding an age-dependent difference between dogs, we confirmed the presence of
consistent differences in oncotic pressure between the
joints of the same normal dog. In each of the 9 animals
from which we have data, the OPsF in the wrist was
lower than that in the knee. Furthermore, the HPsF in
the wrist exceeded that in the knee with comparable
consistency. Each of these findings is statistically
specimens from each animal were averaged and plotted against age, a significant positive correlation was
found (r = 0.93, P < 0.001). There was no apparent
relationship, however, between the sex of the animal
and the oncotic pressure of its synovial fluid (Figure 3).
Oncotic pressure of plasma samples averaged
20.5 0.5 mm Hg and did not correlate with the age of
the animal. There were also no clear relationships
between the oncotic or hydrostatic pressure of synovial fluid and the breed or DLA type of the animals.
Sample volume permitted 3 or more oncotic
pressure determinations in 9 plasmas and in 27 SFs.
The mean coefficient of variation was 2.8% for the
plasma determinations and 3.2% for the S F determinations, indicating the good precision of membrane
osmometry .
z "1
A S (vn)
Figure 3. Synovial fluid (SF) oncotic pressure as a function of
increasing age. Age, but not sex, correlated highly with mean SF
oncotic pressure in the 10 dogs studied.
significant and is interesting in its own right. Taken
together, they suggest a marked interarticular difference in the mean intravascular pressure at the theoretical point of transition from microvascular filtration
to reabsorption.
Our calculations rest upon a longitudinal conception of the synovial microvasculature. We have
assumed that the OP of plasma and interstitium remain
essentially constant throughout the length of the vessel
from arteriole to venule. We have also assumed that an
additional variable, the osmotic reflection coefficient,
is a constant 1.0 (8). Although none of these assumptions will be entirely true, the magnitude of the differences should be sufficiently small that they need not be
considered in the present discussion.
Given these assumptions, the conventional formula for microvascular exchange of water may be
written as follows:
Q = k[(HP,
HPsF) - (OP, - OP,F)I
where Q is the net flow of water, and k, the coefficient
of capillary filtration, reflects the product of hydraulic
conductivity and surface area of the microvascular
wall (7). This equation may be applied to address
filtration of water at any point in the microcirculation.
It can be taken as axiomatic that the intravascular
pressure (HP,) will decrease progressively throughout
the length of the microvessel. In response, classic
Starling-Landis concepts suggest that Q will also
change over distance, making a transition from positive values at the filtering (arteriolar) end of the
microvasculature to negative values at the reabsorptive (venular) end. Substantial evidence indicates that
k also varies longitudinally in other tissues (14), and
similar changes seem certain in the synovium, where
the continuous endothelium of the arteriole gives way
to a fenestrated endothelium within the sometimes
“glomerular” or tufted pattern of reabsorptive microvessels (15,16). These considerations imply that
each segment of the microvasculature will have its
own values for Q, k, and HP,.
In this report, we have focused on the theoretical point of transition from filtration to reabsorption.
At this one site in the microvasculature, we may write
the filtration equation:
Figure 4. Starling forces in the normal canine knee. In this longitudinal model, the transitional microvascular pressure marks the site
where the intravascular hydrostatic pressure crosses the line determined by the other 3 Starling forces (HP, + OP, - 0Psf),where HP
= hydrostatic pressure, sf = synovial fluid, OP = oncotic pressure,
and p = plasma. This is the theoretical point of transition from
microvascular filtration to reabsorption. The horizontal axis reflects
not the lengths, but the areas of microvascular wall available for
filtration and for reabsorption. Exchange of small solutes may occur
primarily by bidirectional diffusion across the distal (reabsorptive)
microvascular bed, where the pressure gradient is small. Filtration
and reabsorption may be relatively unimportant to the exchange of
small solutes, but remains critical to tissue hydration and lymph flow
(filtered - reabsorbed = lymph flow).
This equation yielded mean 2 SEM transitional microvascular pressures of 7.9 2 1.1 mm Hg in the knee and
17.2 2 1.4 mm Hg in the wrist of normal adult dogs.
The difference between these values was significant by
paired t-test (t = 8.3, P < 0.001; n = 9). TMP results
in the shoulder, 9.7 2 1.7 mm Hg, tended to be
intermediate between those of the other 2 joints, but
were closer to the values found in the knee. Figure 4
presents a model of how these forces might balance in
the knee.
The differences in Starling forces in the wrist
and the knee are entirely consistent with the findings
of our companion study on the kinetics of articular
fluid exchange in a different group of normal dogs (17).
In that investigation, the net filtration of water from
wrist microvessels was consistently greater than that
from microvessels of the knee. More filtration may
lead to a higher hydrostatic pressure in the wrists,
while a resultant, pressure-driven increase in lymph
flow would “wash out” synovial fluid proteins and
thereby lower the colloid osmotic pressure. The most
plausible explanation for this constellation of findings
is a higher venous pressure in the wrist than in the
knee. Gravity provides a simple reason for such a
pressure d i f f e r e n c e t h e wrist lies farther below the
heart. However, though attractive in its simplicity, the
gravity hypothesis may be only a part of the explanation. A closer analysis of the data from canine knees
suggests that articular motion, and the system of
valves in local veins and lymphatic vessels, may be
even more important determinants of synovial microvascular function.
For the synovial microvasculature of the knee,
our calculations yield a TMP of only 7.9 mm Hg.
Moreover, this low value is an overestimate since the
osmotic reflection coefficient, which must be < 1, has
not been considered in the calculations. If this factor
were known for living canine joints, its inclusion
would result in an even lower value for the TMP in
knee synoviurn.
Could such a pressure be plausible within a
Starling-Landis context of proximal filtration and distal reabsorption, bearing in mind that all reabsorption
must occur at still lower intravascular pressures distal
to the transition point, and venous blood must then
flow back to the heart? The easiest answer to the
question, a simple no, is based largely on the assumption that pressures in the adjacent venous system are
too high to permit this to occur. This, in fact, is the
conclusion reached by Knox et al in their interpretation of similar data from rabbit knees (18). Their view
is consistent with the concept raised by Michel and
Phillips, that such microvascular beds do not sustain a
transition point, but instead, usually filter fluid
throughout their length (19).
Such a process of continuous filtration seems
likely t o characterize the synovial rnicrovasculature at
rest. Rest, however, is not the normal articular condition. Because the knee is seldom long at rest, even
during sleep, physical activity may play a major role in
sustaining its hydrostatic environment. It is reasonable
to expect that synovial tissue is regularly compressed
and released as joints flex and extend. If, as reported,
venous “valves are frequent in all the veins, even in
the most superficial plexus” of synovial vessels (20),
then cyclic compression may effectively pump venous
return and thus sustain an articular venous pressure
lower than that in the adjacent collecting system.
Analogous pumping of lymphatic vessels, which also
have valves within synovial tissue (21), would clear
extravascular plasma proteins from the synovium and
should thereby lower the oncotic pressure of the
interstitium. Thus, joint motion may concurrently promote distal reabsorption of filtered fluid by two dif-
ferent but complementary mechanisms: effective
pumping of venous blood and augmented clearance ,of
interstitial proteins.
In the absence of effective pumping, the pressure in the distal microvasculature must rise, and
thereby lead to continuing filtration. The net result of
such filtration would be an increase in the rate of
lymphatic flow, an increase in the intraarticular hydrostatic pressure, a fall in the oncotic pressure of
synovial fluid, and an accompanying rise in the calculated TMP at a new steady state. These, in fact, are the
features that distinguish the microvascular parameters
of the canine wrist from those of the knee. All of
the differences we have found between the wrist and
the knee may thus result from a higher venous pressure in the wrist, brought on, perhaps, by less effective
venous pumping a s well a s its more dependent
In this study of Starling-Landis forces in normal
canine joints, we have found that the oncotic pressure
of SF increases with age. We have also observed
consistent interarticular differences both in hydrostatic and in oncotic pressure. Analysis of these patterns suggests a remarkably low microvascular pressure in the knee, which is likely to be sustained by
articular motion.
We thank John Bassett and J. Rodney Levick for
helpful reviews and Linny Simkin for typing the manuscript.
We are deeply indebted to Rainer Storb for permitting us to
study normal animals in the kennels of the Fred Hutchinson
Cancer Research Center. We also appreciate the technical
assistance of Ted Graham.
Exchange of substances through capillary
walls, Circulatory and Respiratory Mass Transport.
Edited by GEW Wolstenholme, J Knight. Boston, Little, Brown & Co., 1969
2. Taylor AE: Capillary fluid filtration: Starling forces and
1. Renkin EM:
lymph flow. Circ Res 49557-575, 1981
3. Ghadially FN: Fine structure of joints, The Joints and
Synovial Fluid. Vol. 1. Edited by L Sokoloff. New
York, Academic Press, 1978
4. Bauer W, Ropes MW, Waine H: The physiology of
articular structures. Physiol Rev 20:272-312, 1940
5. Miiller W: Uber den negativen Luftdriick in Gelenkraum. Dtsch Z Chir 218:395-401, 1929
6. Levick JR: An investigation into the validity of subat-
mospheric synovial pressure recordings and their dependence on joint angle. J Physiol 28955-68, 1979
Wood L , Ferrell WR, Baxendale RH: Pressures in
normal and acutely distended human knee joints and
effects on quadriceps maximal voluntary contractions. Q
J Exp Physiol 73:305-314, 1988
Palmer DG, Myers DB: Some observations of joint
effusions. Arthritis Rheum 11:745-755, 1968
RaffRF, Deeg HJ, Farewell VT, deRose S, Storb R: The
canine major histocompatibility complex: a population
study of DLA-D alleles using a panel of homozygous
typing cells. Tissue Antigens 21:360-373, 1983
Simkin PA, Pickerell CC: Interarticular differences in
oncotic pressure of canine synovial fluid. J Rheumatol
11:14-16, 1984
11. Jayson MIV, Dixon ASJ: Intra-articular pressure in
rheumatoid arthritis of the knee. I. Pressure changes
during passive joint distension. Ann Rheum Dis 29:
26 1-265, 1970
12. Wiederhielm CA, Lee DR, Stromberg DD: A membrane
osmometer for microliter samples. J Appl Physiol 35:
432-435, 1973
13. Webster HL: Colloid osmotic pressure: theoretical aspects and background. Clin Perinatol 9505-521, 1982
14. Zweifach BW, Silberberg A: The interstitial-lymphatic
flow system, International Reviews of Physiology: Car-
diovascular Physiology 111. Vol. 18. Edited by AC
Guyton, DB Young. Baltimore, University Park Press,
15. Schumacher HR Jr: The microvasculature of the syno-
vial membrane of the monkey: ultrastructural studies.
Arthritis Rheum 12:387-404, 1969
16. Goldie IF: Synovectomy in rheumatoid arthritis: a general review and an eight-year follow-up of synovectomy
in 50 rheumatoid knee joints. Semin Arthritis Rheum
3:219-251, 1974
17. Simkin PA, Benedict RS: Iodide and albumin kinetics in
normal canine wrists and knees. Arthritis Rheum 33:
73-79, 1990
18. Knox P, Levick JR, McDonald JN: Synovial fluid: its
mass, macromolecular content and pressure in major
limb joints of the rabbit. Q J Exp Physiol73:3345, 1988
19. Michel CC, Phillips ME: Steady-state fluid filtration at
different capillary pressures in perfused frog mesenteric
capillaries. J Physiol 388:421435, 1987
20. Davies DV, Edwards DAW: The blood supply of the
synovial membrane and intra-articular structures. Ann R
Coll Surg Engl 2:142-156, 1948
21. Rovensk E, Hiittl S: Mikrocirkul cia synovi lnej blanyultrastruktura lymfatickej mikrovaskulatfiry. Bratisl Lek
Listy 87:262-274, 1987
Без категории
Размер файла
580 Кб
microvascular, balances, joint, determinants, norman, oncotic, canine, fluid, hydrostatic
Пожаловаться на содержимое документа