Osmotic Pressure Gradients and Joint Effusions By RICKARD L. LIPSON,EDWARD J. BALDES, JOSEPH A. ANDERSON AND HOWARD F. POLLEY This study reveals that in rheumatoid joint disease the total osmotic pressure of synovial fluid tends to be lower than, or equal to, that of plasma. With local articular or general systemic improvement the total osmotic pressures of synovial fluid and plasma tend to increase, and that of synovial fluid may exceed that of plasma. The colloid osmotic pressure of synovial fluid from joints of patients with rheumatoid arthritis was lower than that of plasma. L e presente studio revela que in morbos rheumatic del articulationes le total pression osmotic del liquido synovial tende a esser inferior o equal a ill0 de plasma. Con local melioration articular o con melioration constitutional, le total pressiones osmotic del liquidos synovial e del plasma tende a montar, e ill0 del liquido synovial pote exceder ill0 del plasma. Le colloide pression osmotic de liquido synovial a b articulationes de patientes con arthritis rheumatoide esseva inferior a ill0 del plasma. T HE MECHANISMS RESPONSIBLE for the abnormal accumulation of fluid in the joint space are complex and poorly understood. Increase in the colloid osmotic pressure of the synovial fluid is usually cited as the principal factor in the formation and maintenance of joint eff~sions.l-~ However, the colloid osmotic pressures of synovial fluid and plasma and the gradient between them appear to require further study before their role can be fully understood. In an attempt to clarify the role of the osmotic pressure of synovial fluid and plasma and the resultant gradient” and its role in the pathophysiology of joint effusions, a study was made of the colloid and total osmotic pressures of fresh paired samples of human synovial fluid and plasma. MATERIALAND METHODS Samples of plasma and synovial fluid were obtained from 20 patients while they were in the fasting state. Thirteen patients were classified according to the diagnostic criteria of the American Rheumatism Association4 as having “definite” and two as having “possible” rheumatoid arthritis; two had rheumatoid spondylitis with rheumatoid involvement of the peripheral joints, and three patients had one or another of the following conditions: Reiter’s syndrome, chondromalacia of the patella and metastatic carcinoma in the lung (primary site unknown) without apparent articular involvement. Venous blood was drawn into a 5-ml. syringe that had been rinsed with an aqueous Read at the interim session of the American Rheumatism As.yociation, Boston, Massachu5 and 6, 1963. This investigation was supportad in part b y Research Grant ASP-9880-C1 from the National Institutes of Health, Public Health Service. *The term “osmotic gradient” is suggested because “gradient” indicates a dynamic system whereas “difference” in the terminology of physics implies a steady state and the system being studied in this report is not a static one. setts, December 29 ARTHRITIS AND RHEUMATISM,VOL. 8, No. 1 (FEBRUARY), 1985 30 LIPSON, BALDES, ANDERSON, POLLEY solution of heparin sodium. The sample was centrifuged to separate the cells and plasma within 1 hour after the specimen had been obtained, and the plasma was then drawn into a 2-ml. syringe where it was kept until studies were done. The synovial fluid was obtained by aspiration of one or both knee joints of each patient immediately after collection of the blood. The usual procedure for aspiration was employed except that special care was taken to avoid the introduction of any local anesthetic into the joint space. A sample of synovial fluid was first collected from each joint in a 5-ml. syringe that had been rinsed with an aqueous solution of heparin SOdium and sent to the laboratory for determination of the osmotic pressure. Preliminary studies demonstrated that the amount of heparin sodium used to rinse the syringes did not alter the total osmotic pressure, and recently Losowsky, Alltree and Atkinson5 reported that heparin did not affect the plasma colloid osmotic pressure. After the collection of the first specimen, the remainder of the synovial fluid was removed from the joint for determination of the specific gravity (refractometer method) and total protein content (biuret method). If steroids were to be given intra-articularly, 37.5 mg. of hydrocortisone acetate was instilled at this time. Osnwtic Pressure Studies. The total osmotic pressure was determined on 36 fresh, paired specimens from the 20 patients, and the colloid osmotic pressure was determined on 14 paired specimens from 10 of the patients. Studies were repeated once in the course of observation and therapy of six patients and a second time on two patients. Determinations also were repeated on several specimens that had been stored for weeks to months in the refrigerator (5" C.) and on three of them after they had been frozen (-70" C.) for 1 week. The osmotic pressure studies were done within 24 hours after the samples were obtained except when studies were repeated weeks or months later to determine the effect of storage or of freezing. If the initial determination was not done immediately, the samples were stored under refrigeration in a sealed syringe during the short (not more than 12 hours) delay. The total osmotic pressure was measured by vapor pressure osmometer as described by Baldes and Johnson.0 The osmometer was arranged so that two samples of the same specimen could be evaluated simultaneously. The rcference solution used was 0.9 per cent sodium chloride and the results are expressed in terms of per cent of sodium chloride. Double galvanometric deflections and repeated determinations with reversal of the position of the drops of samples and reference solutions were used to increase the reliability of the determinations so that the accuracy attained was ,005 per cent NaCl. The colloid osmotic pressure was measured by an osmometer patterned after the one described by Reiff and Yiengst.7 A Statham P23De pressure transducer* was used as the pressure sensing device, and the membrane was a Bac-T-Flex membrane filter no. B l 8 t which has an average pore diameter of 10 to 20 millimicrons. The estimations were performed at 37" C. with distilled water as the dialyzing fluid. The results were expressed in millimeters of water. Definition and Grading of Improvement. General systemic improvement was considered present when there was a recognizable decrease in fever, fatigue and morning stiffness. Local articular improvement was considered present when there was a decrease in at least three of the four following criteria: (1) tenderness to palpation, ( 2 ) local heat, ( 3 ) synovial swelling, and ( 4 ) limitation of motion. General articular improvement was indicated by local improvement of these criteria in at least three fourths of the affected joints. Systemic or local articular improvement or both were judged to be mild, moderate, marked or complete depending on the degree to which the criteria were modified in comparison to the pre-aspiration status. Thus, an * *Statham Instruments, Inc., Los Angeles, California. tschleicher & Schuell Co., Keene, New Hampshire. 31 OSMOTIC PRESSURE GRADIENTS estimated 25 per cent reduction was considered mild improvement; 50 per cent reduction, moderate improvement; 75 per cent reduction, marked improvement; 100 per cent reduction, complete remission. RESULTS The total osmotic pressure of synovial fluid from the 20 patients varied from 0.816 to 0.896 per cent NaCl with a mean of 0.874 per cent NaCI, while that of plasma ranged from 0.825 to 0.894 with a mean of 0.876. At the height of the effusion the total osmotic pressure of the synovial fluid tended to be less than, or equal to, that of the plasma; the difference was significant at p = .01 ( t = 2.9356). However, in patients with rheumatoid arthritis with local articular or general systemic improvement, the total osmotic pressures of both synovial fluid and plasma were likely to increase, but the total osmotic pressure of the synovial fluid tended to become greater than that of plasma (table 1). This table shows the correlation of only certain findings but statistical analysis revealed a significant difference in the total osmotic pressure of the synovial fluid and plasma before treatment, as just cited. After treatment the total osmotic pressure of both the synovial fluid and plasma was increased. In the synovial fluid the pressure increased from a mean of 0.869 to 0.883 ( p < .001; t = 8.1267). In the plasma the total osmotic pressure increased from 0.870 to 0.878 ( p < .01; t = 3.1319). After treatment, the total osmotic pressure of synovial fluid was greater than that of plasma, the difference being significant at p = 0.5 ( t = 2.3943). The colloid osmotic pressure of synovial fluid was consistently less than the colloid osmotic pressure of plasma. The pressure varied from 145 to 285 mm. H 2 0 for rheumatoid synovial fluid with a mean value of 197 mm. and from 220 to 430 mm. for plasma, with a mean value of 328 mm. In one patient without evident articular disease, the colloid osmotic pressures of the synovial fluid and plasma were 62.5 mm. and 245 mm. HzO, respectively. No correlation was evident between the colloid and total osmotic pressures of either the plasma or synovial fluid in the 10 patients whose fluids were available for determinations of colloid osmotic pressure. The rheumatoid arthritis in eight of these patients was classified as “definite,” and in one as “possible,” and there was no essential difference in the data of the two categories. Also, neither the total nor colloid osmotic pressure of the synovial fluid showed a direct correlation with the total protein content or specific gravity (table 2). A comparison of the total osmotic pressures of synovial fluid from both knees in four patients and the colloid osmotic pressures of fluids from one patient revealed no significant difference between the fluids from both knees in the same patient (table 3). Storage of the specimens at 5” C. had a varied effect on the total osmotic pressure but usually resulted in an increase. Freezing (-70” C.) the specimens had little effect on the total osmotic pressure but did reduce the colloid osmotic pressure (table 4) which was associated with an apparent decrease in viscosity. 32 LIPSON, BALDES, ANDERSON, POLLEY Table 1.-Effect Case Specimen from Knee’ Days after Previous Deterniination of Improvement on Total Osmotic Pressure in Rheumatoid Arthritis Total Osmotic Pressure, per cent NaCl Synovial fluid Plasma Comments : Hydrocortisone Acetate Injection into Knee?; Local Art.cular Improvement$ Unless Otherwise Stated 5 0.874 0.884 0.880 0.881 Hydrocortisone into L. Marked improvement in L. 0 6 0 6 0.859 0.879 0.858 0.882 0.866 0.877 0.866 0.877 Hydrocortisone injected into L. Moderate improvement in R. Hydrocortisone into L. Marked improvement in L. 0 60 0.872 0.886 0.871 0.893 No hydrocortisone 0 5 0.882 0.896 0.889 0.889 Hydrocortisone into L. Marked improvement in L. 0 11 0.862 0.872 0.857 0.867 0 0.867 0.867 49 0.884 0.876 Hydrocortisone into R. Mild improvement; hydrocortisone injected into R. Hydrocortisone into L.; injection coincides with second injection into R. Marked general systemic and articular improvement 0 6 0.864 0.878 0.877 0.879 0.889 0.862 0.875 0.872 0.875 0.872 0 5 0 5 Marked general systemic and articular improvement No hydrocortisone Hydrocortisone into L. Moderate improvement in R. Hydrocortisone into L. Marked improvement in L. “L = left knee; R = right knee; specimens were from affected knees. f37.5 Mg. of hydrocortisone acetate was injected in each instance indicated. *Definitions of improvement are described in the text. COMMENT The total osmotic pressure of any solution depends on the number, and for large molecules, the type of particles of solute per unit volume of solvent. An isolated solution has no osmotic pressure, for in order to develop osmotic pressure, the solution must be separated from either a pure solvent or a more dilute solution by a semipermeable membrane. Plasma and interstitial fluid have equal total osmotic pressures ( isosmotic) which are approximately 5100 mm. Hgs and almost entirely the result of the crystalloidal solutes. The capillary walls are freely permeable to the crystalloidal solutes which, therefore, exert no osmotic force across the endothelium. The capillary, on the other hand, is not as permeable to proteins as to crystalloids, therefore, the concentration of protein in the interstitial fluid differs from that in plasma. This difference in concentration of protein exerts an osmotic effect known as the “colloid osmotic pressure” which for the plasma, when compared with inter- 33 OSMOTIC PRESSURE GRADIENTS Table 2.-Osmotic Pressure, Specific Gravity and Total Protein of Synovial Fluid in 16 Patients Osmotic Pressure Case 1 3 7 8 9 10 11 12 13 14 15 16 17 18 19 Diagnosis Volume Aspirated, ml. Specific Gravity Total Protein, Gm./100 ml. Total,per 30 45 18 28 35 1.030 1.034 1.018 1.027 1.028 4.8 5.4 3.15 3.9 4.5 0.874 0.866 0.862 0.877 0.864 180 190 155 145 185 40 35 30+ 40 3 1.025 1.029 1.031 1.025 1.015 3.35 4.87 5.1 3.35 1.6 0.816 0.866 0.862 0.816 0.893 155 255 285 155 62.5 20 10 10f 5 1.016 1.016 1.027 1.014 4.26 2.77 4.5 3.2 0.872 0.896 0.869 0.876 - 30+ 1.031 6.8 0.867 - Rheumatoid arthritis Rheumatoid arthritis Rheumatoid arthritis Rheumatoid arthritis* Rheumatoid spondylitis Rheumatoid arthritis Rheumatoid arthritis Rheumatoid arthritis Rheumatoid arthritis Metastatic cancer in lung, primary unknown, no joint involvement Rheumatoid arthritis Rheumatoid arthritis Rheumatoid arthritis* Chondromalacia patellae Rheumatoid spondylitis Reiter’s syndrome Colloid, cent NaCl mm. HLO - - 1 0.882 *-Possible” rheumatoid arthritis; all other cases of rheumatoid arthritis referred to in 20 this table were classified as “definite.” stitial fluid free of protein, is of the order of 25 to 30 mm. Hg.* Despite the fact that the colloid osmotic pressure is such a small part of the total, it is physiologically the most significant part. It is known as the “effective” osmotic pressure because it is the osmotic force that actually moves fluid across the capillary wall. Articular effusion can be described as a form of localized edema in which the excessive amount of fluid is retained in a relati\,ely large interstitial space or connective tissue cleft, the joint cavity. A feature common to all forms of edema, systemic or localized, is a disturbance in Starling’sg equilibrium, and, therefore, one can analyze the pathophysiologic aspects of articular effusions in terms of the colloid osmotic pressure of the plasma, the hydrostatic pressure in the capillary, the colloid osmotic pressure of the interstitial fluid and the tissue tension. Inflammatory edema like that of articular effusion in rheumatoid arthritis is characterized by dilatation of the terminal vascular bed and marked increase in capillary permeability. The increased permeability and vasodilatation with associated increase in hydrostatic pressure result in the passage of large amounts of fluid and protein including fibrinogen into the interstitial space. Thus in terms of Starling’s equilibrium there is an increase in the 34 LIPSON, BALDES, ANDERSON, POLLEY Table 3.-Comparison of Osmotic Pressure in Synovial Fluid in Rhezrmatoid Arthritis Aflecting Both Knees Osmotic Pressure Case Knee Total, per cent NsCl Colloid, mm. H20 1 Right Left 0.879 0.874 Not done 2 Right Left 0.859 0.858 Not done 6 Right Left 0.878 0.879 162 160 15 Right Left 0.872 0.872 Not done Table 4.-Effect of Freezing (-70" C . ) on Colloid Osmotic Pressure Colloid Osmotic Pressure, mm. HzO Synovial fluid Plasma Sample Refore freezine After freezinr Before freezina After freezing 1 2 3 180 155 195 75 Not done 85 220 235 285 110 85 60 colloid osmotic pressure of the interstitial fluid due to the increased proltein content and the increase in the interstitial fluid pressure secondary to excessive volume of interstitial fluid. The specific role played by any of the forces of Starling's equilibrium in a particular type of localized edema is difficult to assess. Nevertheless, an increase in the colloid osmotic pressure of the synovial fluid is frequently cited as the principal etiologic factor in the formation and maintenance of joint effusions.1-3 Relatively little data are available concerning the total or colloid osmotic pressure of normal and abnormal synovial fluids. The average colloid osmotic pressure of serum of cattle was found by Bauer, Ropes and Wainel to be 365 mm. HzO and that of bo\Tine synovial fluid to vary from 127 to 170 mm. HzO with a mean of 150 mm. The colloid osmotic pressures of normal synovial fluid from two humans were reported by Jensen and ZachariaelO to be 140 and 121 mm. H20, respectively. The colloid osmotic pressure of rheumatoid synovial fluid before and after intra-articular injection of hydrocortisone was studied by Makinen3 who reported a mean value of the colloid osmotic pressure before treatment was 424 mm. H 2 0 in 24 patients, whereas after treatment it decreased to 357 mm. H20. An increase in the colloid osmotic pressure of synovial fluid was considered the principal causative agent in the formation of inflammatory efiusions in the synovial cavity. In the one patient with degenerative arthritis of the knees studied by Jensen and Zachariae,lo before and after the intra-articular administration of OSMOTIC PRESSURE GRADIENTS 35 hydrocortisone, the colloid osmotic pressure of degenerative synovial fluid increased from 113 mm. before the injection of hydrocortisone to 120 mm. H,O 5 days following it. In the one paticnt with rheumatoid arthritis evaluated in a similar manner in the studies reported here (case 6, tables 1 and 3 ) , the colloid osmotic pressure of the synovial fluid in the left knee increased from 160 to 210 mm. H 2 0 following steroid instillation. Also during the 5-day period following the injection of hydrocortisone into the left knee, the colloid osmotic pressure of fluid from the right knee increased from 162 to 207 mm. H20, that of plasma increased from 317 to 430 mm. H20 and there was an associated local articular (both knees) and general systemic improvement. The results of our study suggest that the osmotic gradient is not the principal causative factor in the formation of effusions or the resorption of fluid from the synovial cavity. Furthermore it appears that the osmotic gradient between the plasma and synovial fluid would favor exit rather than entry of fluid into the joint space. Recent studies have shown that the hydrostatic pressure in a distended joint frequently exceeds the capillary hydrostatic pressure.ll Thus, according to Starling’s equilibrium both the osmotic gradient and tissue tension favor reduction rather than accumulation and maintenance of excessive amounts of interstitial fluid. However, as pointed out by Bland,12 it is probable that Starling’s hypothesis is an oversimplification of what occurs in a complex situation. Pappenheimer13 calculated that if diffusion were limited to the filtering and absorbing areas as set forth in Starling’s hypothesis, transport would be inadequate to supply metabolic needs. He further showed that water and solutes actually move rapidly in both directions across the capillary wall. Early in the pathogenesis of inflammatory edema other factors than those outlined by StarlingQappear to play a role. The blockage of lymph vessels by fibrin may be one of the most important. The lymph vessels are extremely important in the removal of interstitial fluid and colloidal material from an area of localized edema because lymph vessels do not collapse despite a marked increase in tissue tension. This is the result of their connections to collagen fibers14 which are separated by the accumulating edema fluid and thereby tend to keep the lymph vessels open. However, if there is blockage of the lymph vessels by fibrin plugs, this important exit may cease to function.15 The synovial cavity is supplied by superficial and deep lymphatic plexuses.l6 Partial or complete blockage of this important resorption system would favor the entrance of fluid and solute at a faster rate than it can leave. The result is effusion. A possible explanation of our findings is that the osmotic gradient is in reality being used as a part of a homeostatic mechanism to move fluid and solutes out of the joint space in the presence of a plugged lymph system. In support of this is the fact that inflammation of the synovial membrane is associated with increased permeability which is subsequently reduced by 36 LIPSON, BALDES, ANDERSON, POLLEY local intra-articular therapy.l7 More direct evidence has been provided by isotope clearance studies by which Scholer, Lee and Polleyl* have shown that the exit of D20 from the joint space was reduced in two of three rheumatoid joints after intra-articular administration of hydrocortisone. Studies of the clearance rates of injected r a d i o ~ o d i u r nand ~ ~ radioacti\Te iodine-labeled albuminZ0also demonstrated that rheumatoid knees had a higher clearance rate than normal knees, whereas intra-articular injection of hydrocortisone and clinical improvement were associated with a reduction of the clearance rate. Thus, as a result of a variety of clearance studies and the osmotic pressure observations reported here, the proposition that the osmotic gradient is working as a compensatory mechanism to mor7e fluid out of the inflamed joint becomes more tenable. SUMMARY The results of this study reveal that in rheumatoid joint disease the total osmotic pressure of the synovial fluid tends to be lower than, or equal to, the total osmotic pressure of the plasma. However, with local articular or general systemic improvement in rheumatoid arthritis, the total osmotic pressure of the synovial fluid and plasma tends to increase, and contrary to the sitnation prior to improvement the total osmo'tic pressure of the synovial fluid may exceed that of the plasma. The colloid osmotic pressure of the synovial fluid from joints of patients with rheumatoid arthritis was consistently lower than the colloid osmotic pressure of the plasma. The range was 145 to 285 mm. HzO for the synovial fluid with a mean pressure of 197 mm., and from 220 to 430 mm. H,O for the colloid osmotic pressure of the plasma with a mean value of 328 mm. Freezing ( -70" C . ) the synovial fluid produced a decrease in the colloid osmotic pressure. These results indicate that intra-articular fluid accumulates despite the existence of an osmotic gradient that should favor the reduction, rather than the formation, of effusion. The possibility is suggested that the existing osmotic gradient is a compensatory mechanism to move fluid out of the joint in the face of blockage of the lymphatic system. REFERENCES 1. Bauer, W., Ropes, M. W. and Waine, H.: The physiology of articular structures. Physiol. Rev. 20:272, 1940. 2. Freyberg, R. H.: The joints. In Sodeman, W. A.: Pathologic Physiology: Mechanisms of Disease, Ed. 3. Philadelphia, Saunders, 1961, p. 1033. 3. Makinen, P.: Synovial fluid in rheumatoid arthritis with special reference to intra-articularly applied hydrocortisone. Ann. med. exper. et biol. Fenniae. (Suppl. 7 ) 36:1, 1958. 4. Ropes, M. W., Bennett, G. A., Cobb, S., Jacox, R. and Jessar, R. A.: 1958 Revision of diagnostic criteria for rheumatoid arthritis. Bull. Rheumat. Dis. 9:175, 1958. 5. Losowsky, M. S., Alltree, E. M. and Atkinson, M.: Plasma colloid osmotic pressure and its relation to protein fractions. Clin. Sc. 22:249, 1962. 6. Baldes, E. J. and Johnson, A. F.: The thermo-electric osmometer: its construction and use. Biodynamica. No. 47:1, 1939. 7 . Reiff, T. R. and Yiengst, M. J.: A rapid automatic semimicro colloid osmometer. J. Lab. & Clin. Med. 53:291, 37 OSMOTIC PRESSURE GRADIENTS 1959. 8. Pitts, R. F.: Physiology of the Kidney and Body Fluids. Chicago, Yr. Bk. Pub., 1963. 9. Starling, E. H.: On the absorption of fluids from the connective tissue spaces. J. Physiol. 19:312, 1896. 10. Jensen, C. E. and Zachariae, L.: The contributions from hyaluronic acid and from protein to the colloid osmotic pressure of human synovial fluid. Acta rheum. scandinav. 5: 18, 1959. 11. Caughey, D. E. and Bywaters, E. G.: Joint fluid pressure in chronic knee effusions. Ann. Rheumat. Dis. 22: 106, 1963. 12. Bland, J. H.: Clinical Metabolism of Body Water and Electrolytes. Philadelphia, Saunders, 1963. 13. Pappenheimer, J. R.: Passage of molecules through capillary walls. Physiol. Rev. 33:387, 1953. 14. Greep, R. 0.: Histology. New York, Blakiston, 1954. 15. Menkin, V.: The significance of lym- phatic blockade in immunity. Ann. New York A d . Sc. 46:789, 1946. 16. Bloom, W. and Fawcett, D. W.: A textbook of histology, Ed. 8. Philadelphia, Saunders, 1962. 17. Sharp, G. W. G.: Effect of certain antiarthritic compounds on the permeability of synovial membrane in the rabbit. Ann. Rheumat. Dis. 22:50, 1963. 18. Scholer, J. F., Lee, P. R. and Polley, H. F.: The absorption of heavy water and radioactive sodium from the knee joint of normal persons and patients with rheumatoid arthritis. Arth. & Rheumat. 2:426, 1959. 19. Harris, R., Millard, J. B. and Banerjee, S. K.: Radio-sodium clearance from the knee joint in rheumatoid arthritis. Ann. Rheumat. Dis. 17:189, 1958. 20. Ahlstrom, S., Gedda, P. 0. and Hedberg, H.: Disappearance of radioactive serum albumin from joints in rheumatoid arthritis. Acta rheumat. scandinav. 2: 129, 1956. Richard L. Lipson, M.D., Fellow i n Biophysics, Mayo Foundation, Graduate School, University of Minnesota; Rochester, Minnesota. Present address: Assistant Professor of Medicine, Rheumatism Research Center, Unizjersity of Vermont College of Medicine, Burlington, Vermont. Edward 1. Baldes, Ph.D., Emeritus Member, Section of Biophysics, Mayo Clinic; Emeritus Professor of Biophysics, Mayo Foundation, Graduate School, University of Minnesota; Rochester, Minnesota. Joseph A. Anderson, B.S., Technical Assistant, Section of Biophysics, Mayo Clinic; Rochester, Minnesota. Howard F. Polley, M.D., Consultant, Section of Medicine, Mayo Clinic; Professor of Medicine, Mayo Founhtion, Graduate School, University of Minnesota; Rochester, Minnesota.