Hyalgan®) on Osteoarthritis Progression in a Rabbit Model D

International Cartilage Repair Society

Long-term effect of sodium hyaluronate (Hyalgan®) on osteoarthritis progression in a rabbit model D. Amiel Ph.D.*, T. Toyoguchi M.D., Ph.D., K. Kobayashi M.D., Ph.D., K. Bowden A.A., M. E. Amiel M.A. and R. M. Healey B.S., M.B.A. Department of Orthopaedics, University of California San Diego, 9500 Gilman Dr, La Jolla, CA 92093-0630, USA

Summary

Background: Intra-articular (IA) hyaluronan (HA) injections are approved for the treatment of knee osteoarthritis (OA) pain. One of the currently available products is approved for repeat treatment courses. While HA is classed as a symptom-modifying agent, there is substantial evidence that this therapeutic modality also possesses disease-modifying activity. Objective: A rabbit model of OA, anterior cruciate ligament transection (ACLT), was used to investigate the long-term effects of single and sequential courses of HA therapy on OA progression. Design: One or two courses of five weekly IA injections of sodium hyaluronate (Hyalgan®) average molecular weight, MW, of 500–730 kDa, or vehicle were administered to rabbits (Nϭ10 per group), initiated 4 and 13 weeks (for groups that received a second course) after ACLT. Gross morphological and histomorphometric evaluations were performed on harvested knee joints following sacrifice at 26 weeks after surgery. Results: All the rabbits exhibited the characteristic pathologic changes of OA. Rabbits that received one or two courses of HA injections showed less disease progression than rabbits treated with ACLT alone or with 10 vehicle injections. However, rabbits that received five vehicle injections also showed improved morphology compared with those given no injections. Rabbits that received 10 HA injections showed significantly less surface roughness of the femoral cartilage compared with rabbits treated with ACLT, 5 HA injections, or 10 vehicle injections, and showed significantly less surface roughness of the tibial plateau compared with all other treatment groups (P<0.05). Conclusions: Repeat courses of HA injections reduced the degree of articular degeneration in a rabbit ACLT model of OA. Sequential courses of HA therapy may be advantageous in the long-term management of OA. © 2003 Published by Elsevier Ltd on behalf of OsteoArthritis Research Society International. Key words: Osteoarthritis, Hyaluronans, Sodium hyaluronate, Hyalgan®, Structure modification, Anterior cruciate ligament transection.

Introduction lishment of structure-modifying activity does not require demonstration of symptom-modifying activity, it is possible Intra-articular (IA) hyaluronan (HA) therapy is used for the that a single agent could possess both activities12. treatment of pain associated with osteoarthritis (OA) of There are growing preclinical and clinical data suggest- the knee. Clinical trials and clinical practice reports ing that IA HA has disease-modifying activity, in addition to have demonstrated that IA HA injections are effective and its proven efficacy and safety in treating the pain of OA. associated with an excellent safety and tolerability 1–11 With the use of in vitro, animal and human models, HA has profile . As with all other available non-surgical treat- been shown to exert a number of complex regulatory ments for OA, HA is currently viewed as treatment for only 12 effects on the environment of the synovium, the articular the symptoms of OA . However, development of pharma- cartilage, and the extracellular matrix of the knee joint13. cologic treatments with the potential for structure-modifying These effects include, but are not restricted to, influencing activity in the treatment of OA, also called chondro- the synthesis of endogenous HA by synoviocytes14, pre- protective disease-modifying drugs for OA (DMOADs), has venting the degradation of proteoglycan and collagen in become a major focus in the field of OA research. As the extracellular matrix15, enhancing chondrocyte defined by the Osteoarthritis Research Society (OARS) metabolism16, inhibiting chondrodegradation17, preventing and the European Group for the Respect of Ethics and apoptotic death of chondrocytes18 and inhibiting inflam- Excellence in Science (GREES), such compounds retard matory responses that are associated with cartilage or stabilize the progression of established OA by altering degradation13,19. the underlying pathologic processes12. Although the estab- A number of animal models have been developed to evaluate the underlying pathologic progress of OA. These Supported by a grant from Sanofi-Synthelabo Inc. and NIH/NIA models include meniscectomy and anterior cruciate AG07996. ligament transection (ACLT)16,20–26. Our laboratory has *Address correspondence and reprint requests to: David Amiel, previously shown that ACLT in rabbits initiates the Connective Tissue Biochemistry, 9500 Gilman Drive 0630, La 26 Jolla, CA 92093-0630, USA. Tel: 858-534-2676; Fax: 858-534- degenerative changes of OA . Using gross morphologic 5304; E-mail: [email protected] evaluation, the degenerative changes initiated by ACLT, Received 26 February 2003; revision accepted 6 May 2003. such as articular fibrillation and erosion, are progressive

636 Osteoarthritis and Cartilage Vol. 11, No. 9 637

Fig. 1. Experimental design. ACLT=no injections following anterior cruciate ligament transection (ACLT); 5 HA=five weekly injections of hyaluronan (HA) beginning on the 4th week after surgery; 5 PBS=five weekly PBS injections beginning on the 4th week after surgery; 10 HA=10 weekly HA injections, one course of five injections beginning on the 4th week after surgery and one course of five injections beginning on the 13th week after surgery; 10 PBS=10 weekly PBS injections, one course of five injections beginning on the 4th week after surgery and one course of five injections beginning on the 13th week after surgery. and easily observed. Histomorphometric analysis of the and xylazine (8 mg/kg). The left knee was shaved and joints subjected to ACLT allows for the changes disinfected with Betadine® solution. A medial parapatellar brought about by the breakdown of articular cartilage to be incision was made through the skin and an arthrotomy quantified26. performed. The patella was dislocated laterally and the In a previous study, we used the ACLT model to evaluate knee placed in full flexion. The ACL was visualized and the effects of one course of five weekly IA injections of HA transected with a no. 15 blade. The joint was then irrigated on the biochemical properties of the cartilage matrix during with sterile saline and closed. The capsule and the syn- the development of OA27. The results demonstrated a ovium were closed together with a running suture of 4–0 positive effect of HA on the preservation of articular carti- prolene. The skin was closed with a running 3–0 nylon lage and synovial tissue. Because the structure-modifying mattress suture supplemented with interrupted sutures. effects of treatment could be slow-acting, it is important to Postoperatively, the animals were permitted cage evaluate these effects over the long term, and with more activity (60 cm×60 cm×40 cm) without immobilization. The than one course of treatment. animals were closely monitored for infections and other The objective of the present study was to examine the complications. long-term effect of two courses of IA injections of sodium hyaluronate on the development of OA induced by transection of the anterior cruciate ligament in the rabbit knee joint. EXPERIMENTAL DESIGN AND HYALURONAN INJECTIONS Rabbits were divided into five treatment groups: Group 1 (ACLT only) underwent ACLT with no further treatment. Materials and methods Group 2 (5HA+ACLT) underwent ACLT, followed by a ANIMALS single course of five weekly sodium hyaluronate (Hyalgan®) injections initiated 4 weeks after surgery. Fifty New Zealand White (NZW) male and female rabbits Group 3 (5PBS+ACLT) underwent ACLT, followed by a (weighing 3.7 kg±200 g, 1 year of age with closed epi- course of five weekly injections of phosphate-buffered physes) were used in this study. The female and male saline (PBS) vehicle. Group 4 (10HA+ACLT) was treated rabbits were randomly divided into each group at a ratio of as Group 2 but also received a second course of five 50:50. The rabbits were divided into five groups of 10 each weekly sodium hyaluronate injections initiated at 13 weeks after surgical ACLT and simple arthrotomy of their left postsurgery. Group 5 (10PBS+ACLT) was treated as Group knees. All animals were sacrificed at week 26 of the study. 3 but also received a second course of five weekly PBS injections initiated at 13 weeks postsurgery. All animals were sacrificed at week 26 of the study, and both ACLT and SURGICAL PROCEDURE non-ACLT contralateral knees were harvested. The study The surgical protocol was approved by the UCSD design is summarized in Fig. 1. Institutional Review Board. All rabbits were anesthetized Rabbits receiving IA injections of sodium hyaluronate or by an intra-muscular injection of ketamine (100 mg/kg) vehicle were anesthetized with ketamine (30 mg/kg) and 638 D. Amiel et al.: Disease modification by repeat courses of hyaluronan therapy in OA xylazine (2 mg/kg). The ACL transected knees were shaved captured by high-definition video processing boards of the and disinfected with betadine solution. A 27-gauge needle image analysis system within the microcomputer (Everex on a tuberculin syringe was inserted beneath the patella 486/25, USA). Processed images were viewed on a high- in the region of the patellofemoral joint and 0.3 ml of resolution color monitor (Sony Trinitron, Japan). Custom- sodium hyaluronate (Hyalgan®, 10 mg/ml in PBS) or ized software was used to measure cartilage thickness and vehicle (PBS) was injected in the intercondylar notch of the area and surface roughness of cartilage, as well as thick- femur between the medial and lateral femoral injections. ness of the synovial lining cell layer28. A 7-mm weight- The injections were performed once weekly for 5 weeks, for bearing section of the femoral condyle (taken from the either one or two courses. region of greatest cartilage damage) was defined and analyzed. The thickness of cartilage from the surface to the tide mark was calculated from the mean of five measure- ASSESSMENTS ments made perpendicularly to the surface of each section All gross morphological, histological, and histo- at five equally spaced points. The area of the cartilage morphometric analyses were performed by treatment-blind present between the monitor screen edges was likewise personnel. calculated. The thickness and area of cartilage were com- puted using a mathematical algorithm with the coordinates of articular cartilage and tide mark. GROSS MORPHOLOGY The surface profile of each specimen was assessed as previously described26 by determining the surface rough- Gross morphological changes of the femoral condyles 26 ness. Calculation of cartilage roughness is based on the were assessed as previously described according to the deviations from an idealized smooth surface, which is following criteria after the application of India ink—grade 1 derived from shape parameters of normal cartilage outside (intact surface): surface appears normal and does not the region of degeneration. This parameter is expressed as retain any ink; grade 2 (minimal fibrillation): site appears the root mean square (RMS) surface roughness calculated normal before staining, but retains the India ink as from the following equation: elongated specks or light gray patches; grade 3 (overt fibrillation): the cartilage is velvety in appearance and 1/2 retains ink as intense black patches; grade 4 (erosion): loss 1 N 2 מ ϭ of cartilage exposing the underlying bone; grade 4a: RMS surface roughness ͚(Yidealized i Yreal i) SN ϭ D 0 mm

HISTOLOGICAL PREPARATION DNA CONCENTRATION (TISSUE CELLULARITY) The medial femoral condyles (ACLT and contralateral controls) were divided sagitally into two equal parts. The Relative cellularity in synovial tissue was assessed by measuring the concentration of DNA as previously tissue blocks were fixed in 10% neutral-buffered formalin 27,29 with cetylpyridinium chloride (CPC) for 7 days and decalci- described . Water-washed and lyophilized synovial fied with EDTA. After decalcification (confirmed by X-rays), tissue, 7–10 mg, were solubilized by incubation for 2 h in the two halves of the condyles were embedded in paraffin. 3.0 ml of 1 N NaOH at 65°C. Duplicate 1.0-ml aliquots were Five-micrometer sections, (three from each half) were cut combined with 0.04% indole–HCl reagent and extracted with a Reichert-Jung microtome and stained with Safranin with chloroform to remove interfering substances. The O/fast green. Synovia from each group were also pro- aqueous phase containing the DNA was removed, and the cessed, embedded in paraffin, and stained with hematoxy- absorbency was read at 490 nm. Standards of calf thymus lin and eosin (H&E) for cellular assessment by image DNA were also determined for comparison. Results are analysis. expressed as micrograms of DNA per milliliter of dry tissue.

HISTOMORPHOMETRY STATISTICAL ANALYSIS Six Safranin O-stained sections from each medial con- To establish differences between ACLT and contralateral dyle were assessed and averaged using a color image control joints for all parameters, a two-tailed t-test using analysis system (Oncor Instrument Systems, USA). Histo- paired sets of data was employed. To evaluate differences logic sections were visualized with a microscope (Nikon in ACLT data between treatment groups, analyses of vari- Microphot, Japan) and an attached high-resolution color ance using unpaired data sets were employed. The level of video camera (Hitachi HV-C10, Japan). Video images were signiﬁcance for each type of analysis was set at 0.05. Osteoarthritis and Cartilage Vol. 11, No. 9 639

Table I Distribution of OA grades by gross morphological assessment of the articular cartilage from the femoral condyle and tibial plateau Grade Group 1 Group 2 Group 3 Group 4 Group 5 ACLT only 5 HA injections+ACLT 5 PBS injections+ACLT 10 HA injections+ACLT 10 PBS injections+ACLT (N=10) (N=10) (N=10) (N=10) (N=10) Femoral condyle 10 0 0 0 0 20 1 1 0 0 31 2 2 2 0 4a 1 0 1 0 0 4b 2 3 1 2 5 4c 6 4 5 6 5 Tibial plateau 10 3 0 3 0 22 0 3 2 0 30 0 0 0 0 4a 0 1 0 0 0 4b 6 4 4 4 8 4c 2 2 3 1 2 Grade 1=intact surface; grade 2=minimal ﬁbrillation; grade 3=overt ﬁbrillation; grade 4a=0 mm

Results ACLT), Group 3 (5PBA+ACLT), and Group 5 (10PBS+ ACLT) were torn. In contrast, 2 of 10 medial menisci in Four rabbits were eliminated from the study and were Group 4 (10HA+ACLT) were not torn. Therefore, only the replaced to maintain 10 animals per study group. One group that received two courses of sodium hyaluronate rabbit died suddenly from an unknown reason, one was injections demonstrated any menisci that were not torn sacrificed because it had a toe infection, another because it (Fig. 2). had a swollen foot resulting from an injury in the cage, and one rabbit had a broken back in the postoperative course. No adverse reactions to sodium hyaluronate or vehicle DNA CONTENT IN SYNOVIAL TISSUE injections were noted in any of the treated animals. The measurement of DNA content in the synovial tissue is an index of inflammation and cellular proliferation. GROSS MORPHOLOGICAL ASSESSMENT Results from this study showed that the specimens from Group 2 (5HA+ACLT), Group 3 (5PBS+ACLT), and All specimens exhibited changes consistent with the Group 4 (10HA+ACLT) had significantly lower DNA concen- development of chronic OA and showed severe degrada- trations (P<0.05) in the tissues when compared with that in tion of the femoral condyle cartilage and tibial plateau. Group 1 (ACLT only) and Group 5 (10PBS+ACLT) (Fig. 3). Despite the finding that the general appearance of the There were no significant differences between Group 2, harvested knees was poor for all groups, the groups that Group 3, and Group 4 with regard to this parameter, nor received either one or two courses of HA injections follow- was there any difference between the joints that underwent ing ACLT exhibited slightly greater resistance to the ACLT and non-ACLT contralateral control joints within each changes of OA than did the groups that received two of these three groups. courses of vehicle or no treatment alone: 2 of 10 rabbits that received 10 HA injections (Group 4) exhibited only grade 3 morphology in the femoral condyle (Table I). HISTOMORPHOMETRY However, the group that received one course of vehicle injections also exhibited good morphological scoring in the Based on visual assessment of histologic staining, the femoral condyle. Three of 10 rabbits in Group 2 and Group general condition of the femoral condyles was poor in all of 3 (5HA+ACLT and 5PBS+ACLT, respectively) exhibited the groups (data not shown). Histomorphometrically, there only grades 2 or 3 morphology. All of the joints from the was a significant decrease in both measured height and group that received two courses of vehicle injections after measured area of cartilage in ACLT joints compared with ACLT were grade 4b or worse, which indicates lesions of at contralateral control joints in all of the groups [Fig. 4(a and least 2 mm in the cartilage exposing the bone. This finding b)]. No difference was detected between the cartilage was consistent for both the femoral condyles and tibial heights and areas of any of the experimental groups, which plateaus. exhibited mostly grade 4 cartilage as indicated by gross All rabbits exhibited some degree of mild to moderate morphological assessment (Table I). joint effusion, i.e. 200 µl of fluid compared with 100 µl of In contrast to the findings regarding cartilage height and normal synovial fluid in the rabbit knee. This amount was area, significant differences were identified between treat- not significant, however, for any group: 44 rabbits exhibited ment groups in the surface roughness characteristics. All of grade 1, and 6 rabbits exhibited grade 2. the femoral condyles demonstrated a significant increase In almost all of the animals at 26 weeks after ACLT, in RMS roughness over the contralateral control joints the medial menisci were completely torn and the lateral (P<0.05) (Fig. 4c). However, Group 4 (10HA+ACLT) exhib- menisci were either reversed or shifted medially because of ited significantly less RMS roughness than Group 1 (ACLT) the weight-bearing shift toward the posterior. All of the and Group 5 (10PBS+ACLT) for both the tibia and femur medial menisci in Group 1 (ACLT only), Group 2 (5HA+ (P<0.05). Roughness was also significantly lower in Group 640 D. Amiel et al.: Disease modification by repeat courses of hyaluronan therapy in OA

Fig. 2. Gross morphology of the meniscus in representative samples taken at 26 weeks after surgery from (a) Group 4 (10 HA+ACLT) and (b) Group 1 (ACLT only).

Fig. 3. DNA concentration in synovium from medial femoral condyles (mean±SD). Results are shown for ACLT and non-ACLT (untreated) control knees for each study group. Nϭ10 for each group. Osteoarthritis and Cartilage Vol. 11, No. 9 641

degradation in an animal model of OA. Administration of two courses of HA, totaling 10 injections, resulted in significantly less OA progression at 26 weeks after surgery than did no treatment, one or two treatment courses of vehicle, or only one course of five weekly HA injections. This beneficial effect was especially apparent with regard to prevention of degradation of cartilage of the tibial plateau. Gross morphological assessment of the menisci also indicated that repeat courses of HA injections may help to prevent changes associated with the progression of OA: only the group that had two courses of HA injections (10 injections total) produced any samples in which the medial menisci were not torn. Open-label and steroid-controlled studies used arthroscopy and electron microscopic assessment of paired biopsy samples to demonstrate that one course of five weekly Hyalgan® injections produced significant recon- stitution of the cartilage, improvement in chondrocyte metabolism, and reduction in inflammation of the synovial membrane30–32. Another recent study demonstrated that three courses of three weekly injections of Hyalgan® significantly reduced progression of joint space-narrowing in the subset of patients with the most severe narrowing at study entry33. It is likely that pharmacotherapeutic modification of OA disease progression will be a slower process than that of providing symptom relief. Therefore, meaningful clinical effects may require longer-term exposure to the therapeutic modality. Because of this likelihood, the OARS recom- mends that clinical studies designed to evaluate the structure-modifying effects of drugs for OA treatment should be performed over at least a 1-year period. The present study in an animal model of OA evaluated the effects of HA after a 26-week follow-up period, and revealed that two courses of HA injections provided a benefit compared with one course of injections for slowing OA progression. These results indicate that the long-term structure-modifying effects of repeat treatment courses of HA should be further evaluated in patients with OA of the knee. In conclusion, the results of this study indicate that repeat Fig. 4. Histomorphometric assessment of samples from the five courses of HA injections in an ACLT rabbit model of OA can treatment groups. (a) Cartilage height of the femoral condyles reduce the degree of articular degeneration observed over (mm) (mean±SD), (b) cartilage area of the femoral cartilage (mm2) a 26-week follow-up period compared with no treatment, (mean±SD), and (c) normalized cartilage roughness (RMS one course of HA injections, or one or two courses of roughness/cartilage thickness) of the femoral cartilage (mm) ϭ injections of vehicle only. Consistent with recent clinical (mean±SD). N 10 for each treatment group. findings of structure modification by HA, these results suggest the possibility that HA treatment may be able to modify the progression of OA of the knee, and it may have 4 (10HA+ACLT) than that seen in the tibia for Group 3 advantages in the long-term management of this disease. (5PBS+ACLT) and than that seen in the femur for Group 2 Hyalgan® is a sodium hyaluronate product that has been (5HA+ACLT) (P<0.05 for both comparisons). Group 3 found to be safe and effective in the ACLT rabbit model (5PBS+ACLT) exhibited significantly lower femoral rough- when given in repeat treatment courses, and this may ness than Group 1 (ACLT only), Group 2 (5HA+ACLT), and provide the benefit of mediating symptom relief and exert- Group 5 (10PBS+ACLT) (P<0.05 for all comparisons). Only ing a structure-modifying effect when used in repeat Group 4 (10HA+ACLT) had significantly less tibial rough- courses over the long term. ness than Group 2 (5PBS+ACLT) (P<0.05). These results indicate a significant benefit of two courses of HA injections with regard to surface roughness following ACLT as compared with no HA treatment, or treatment with only one References course of HA injections. 1. Adams ME, Lussier AJ, Peyron JG. A risk–benefit assessment of injections of hyaluronan and its Discussion derivatives in the treatment of osteoarthritis of the knee. Drug Safety 2000;23:115–30. The results of the present study indicate that repeat 2. Altman RD, Moskowitz R. Intraarticular sodium courses of HA injections reduced the degree of articular hyaluronate (Hyalgan®) in the treatment of patients 642 D. Amiel et al.: Disease modification by repeat courses of hyaluronan therapy in OA

with osteoarthritis of the knee: a randomized clinical investigation into mechanisms of action. Osteoarthri- trial. J Rheumatol 1998;25:2203–12. tis Cartilage 1994;2:133–40. 3. Brandt KD, Block JA, Michalski JP, Moreland LW, 18. Takahashi K, Hashimoto S, Kubo T, Hirasawa Y, Lotz Caldwell JR, Lavin PT. Efficacy and safety of intra- M, Amiel D. Effect of hyaluronan on chondrocyte articular sodium hyaluronate in knee osteoarthritis. apoptosis and nitric oxide production in experimen- ORTHOVISC Study Group. Clin Orthop 2001;385: tally induced osteoarthritis. J Rheumatol 2000; 130–43. 27:1713–20. 4. Dahlberg L, Lohmander LS, Ryd L. Intraarticular injec- 19. Corrado EM, Peluso GF, Gigliotti S, DeDurante C, tions of hyaluronan in patients with cartilage abnor- Palmieri D, Savoia M, et al. The effects of intra- malities and knee pain. Arthritis Rheum 1994;37: articular administration of hyaluronic acid on 521–8. osteoarthritis of the knee: a clinical study with 5. Dougados M, Nguyen M, Listrat V, Amor B. High immunological and biochemical evaluations. Eur J molecular weight sodium hyaluronate (hyalectin) in Rheumatol Inflamm 1995;15:47–56. osteoarthritis of the knee: a 1 year placebo-controlled 20. Sonoda M, Harwood FL, Wada Y, Moriya H, Amiel D. trial. Osteoarthritis Cartilage 1993;1:97–103. The effects of hyaluronan on the meniscus and on 6. Huskisson EC, Donnelly S. Hyaluronic acid in the the articular cartilage after partial meniscectomy. Am treatment of osteoarthritis of the knee. Rheumatology J Sports Med 1997;25:755–62. 1999;38:602–7. 21. Sonoda M, Harwood FL, Amiel ME, Moriya H, Temple 7. Lohmander LS, Daleń N, Englund G, Hamalainen M, M, Chang DG, et al. The effects of hyaluronan on Jensen EM, Karlsson K, et al. Intra-articular hyaluro- tissue healing after meniscus injury and repair in a nan injections in the treatment of osteoarthritis of the rabbit model. Am J Sports Med 2000;28:90–7. knee: a randomised, double blind, placebo controlled 22. Wenz W, Breusch SJ, Graf J, Stratmann U. Ultra- multicentre trial. Ann Rheum Dis 1996;55:424–31. structural findings after intraarticular application 8. Wobig M, Dickhut A, Maier R, Vetter G. Viscosupple- of hyaluronan in a canine model of arthropathy. J mentation with hylan G-F 20: a 26-week controlled Orthop Res 2000;18:604–12. trial of efficacy and safety in the osteoarthritic knee. 23. Wiig ME, Amiel D, VandeBerg J, Kitabayashi L, Clin Ther 1998;20:410–23. Harwood FL, Arfors KE. The early effect of high molecular weight hyaluronan (hyaluronic acid) on 9. Lussier A, Cividino AA, McFarlane CA, Olszynski WP, anterior cruciate ligament healing: an experimental Potasher WJ, De Medicis R. Viscosupplementation study in rabbits. J Orthop Res 1990;8:425–34. with hylan for the treatment of osteoarthritis: findings 24. Kikuchi T, Yamada H, Shimmei M. Effect of high from clinical practice in Canada. J Rheumatol 1996; molecular weight hyaluronan on cartilage degrada- 23:1579–85. tion in a rabbit model of osteoarthritis. Osteoarthritis 10. Barrett JP, Siviero P. Retrospective study of outcomes Cartilage 1996;4:99–110. in Hyalgan®-treated patients with osteoarthritis of the 25. Armstrong S, Read R, Ghosh P. The effects of intra- knee. Clin Drug Invest 2002;22:87–97. articular hyaluronan on cartilage and subchondral 11. Ueno Y, Kuramoto K, Konno N, Koizumi T, Hoshiba T, bone changes in an ovine model of early Nunmura T, et al. Investigation result of use after osteoarthritis. J Rheumatol 1994;21:680–8. launch of ARTZ and ARTZ dispo. Jpn Pharmacol 26. Yoshioka M, Coutts R, Amiel D, Hacker SA. Charac- Ther 1995;23:2151–70. terization of a model of osteoarthritis in the rabbit 12. Altman R, Brandt K, Hochberg M, Moskowitz R. knee. Osteoarthritis Cartilage 1996;4:87–98. Design and conduct of clinical trials in patients with 27. Yoshioka M, Shimizu C, Harwood FL, Coutts RD, osteoarthritis: recommendations from a task force of Amiel D. The effects of hyaluronan during the devel- the Osteoarthritis Research Society. Osteoarthritis opment of osteoarthritis. Osteoarthritis Cartilage Cartilage 1996;4:217–43. 1997;5:251–60. 13. Punzi L. The complexity of the mechanisms of action of 28. Hacker SA, Healey RM, Yoshioka M, Coutts R. A hyaluronan in joint diseases. Clin Exp Rheumatol methodology for the quantitative assessment of ar- 2001;19:242–6. ticular cartilage histomorphometry. Osteoarthritis 14. Smith MM, Ghosh P. The synthesis of hyaluronic acid Cartilage 1997;5:343–355. by human synovial fibroblasts is influenced by the 29. Amiel D, Frank CB, Harwood FL, Fronek J, Akoson nature of the hyaluronate in the extracellular WH. Tendons and ligaments: a morphological and environment. Rheumatol Int 1987;7:113–22. biochemical comparison. J Orthop Res 1984;1: 15. Takahashi K, Goomer RS, Harwood F, Kubo T, 257–65. Hirasawa Y, Amiel D. The effects of hyaluronan on 30. Frizziero L, Govoni E, Bacchini P. Intra-articular matrix metalloproteinase-3 (MMP-3), interleukin-1b, hyaluronic acid in the treatment of osteoarthritis of and tissue inhibitor of metalloproteinase-1 (TIMP-1) the knee: clinical and morphological study. Clin Exp gene expression during the development of Rheumatol 1998;16:441–9. osteoarthritis. Osteoarthritis Cartilage 1999;7: 31. Ronchetti IP, Guerra D, Taparelli F, Boraldi F, 182–90. Bergamini G, Mori G, et al. Morphological analysis of 16. Schiavinato A, Lini E, Guidolin D, Pezzoli G, Botti P, knee synovial membrane biopsies from a random- Martelli M, et al. Intraarticular sodium hyaluronate ized controlled clinical study comparing the effects of injections the Pond-Nuki experimental model of sodium hyaluronate (Hyalgan) and methyl- osteoarthritis in dogs. II. Morphological findings. Clin prednisolone acetate (Depomedrol) in osteoarthritis. Orthop 1989;241:286–99. Rheumatology 2001;40:158–69. 17. Creamer P, Sharif M, George E, Meadows K, 32. Guidolin DD, Ronchetti IP, Lini E, Guerra D, Frizziero Cushnaghan J, Shinmei M, et al. Intra-articular L. Morphological analysis of articular cartilage hyaluronic acid in osteoarthritis of the knee: an biopsies from a randomized, clinical study comparing Osteoarthritis and Cartilage Vol. 11, No. 9 643

the effects of 500–730 kDa sodium hyaluronate 33. Jubb RW, Piva S, Beinat I, Dacre P, Gishen P. (Hyalgan) and methylprednisolone acetate on pri- Structure modifying study of hyaluronan (500– mary osteoarthritis of the knee. Osteoarthritis 730 kDa, Hyalgan) on osteoarthritis of the knee. Cartilage 2001;9:371–81. Arthritis Rheum 2001;45:617 (Abstract).