Skip to main content

Development of musculoskeletal toxicity without clear benefit after administration of PG-116800, a matrix metalloproteinase inhibitor, to patients with knee osteoarthritis: a randomized, 12-month, double-blind, placebo-controlled study



We performed a randomized, double-blind, placebo-controlled, multicenter, parallel-group, dose-response study of the efficacy and safety of the oral administration of PG-116800, a matrix metalloproteinase (MMP) inhibitor, in patients with mild to moderate knee osteoarthritis. The primary efficacy endpoints included the progression of joint space narrowing in the osteoarthritic knee, as measured by microfocal radiography with fluoroscopic positioning, and the reduction of symptoms (pain and stiffness) and/or the improvement of function, as measured by the Western Ontario and McMaster Universities osteoarthritis index (WOMAC). Four hundred and one patients were randomly assigned to either placebo (n = 80) or one of fourdoses of PG-116800: 25 mg (n = 81), 50 mg (n = 80), 100 mg (n = 80), or 200 mg (n = 80) taken twice daily for 12 months. During the study, the 200-mg dose was discontinued based on an increased frequency of musculoskeletal adverse effects. After 1 year of treatment, no statistically significant difference was observed between placebo and PG-116800 with regard to mean changes in minimum joint space width of the knee or to WOMAC scores. The most frequent adverse effect was arthralgia (35%). Twenty-three percent of evaluable patients had at least a 30% decrease from baseline of at least onerange-of-motion measurement of either shoulder at a follow-up visit. The percentage of patients with reduction in range of motion was significantly greater in the twohighest dose groups relative to placebo. Thirteen percent of patients, half of whom were in the 200-mg group, reported hand adverse events (oedema, palmar fibrosis, Dupuytren contracture, or persistent tendon thickness or nodules). The threemost frequent shoulder adverse events were reversible arthralgia, stiffness, and myalgia, which mostly affected the twohighest dose groups. The unfavorable risk-benefit balance of the MMP inhibitor PG-116800 in patients with knee osteoarthritis precludes further development of the compound for this indication. This study adds to the weight of evidence suggesting that side effect profiles of MMP inhibitors in general make them unsuitable for use in osteoarthritis.

Trial Registration NCT00041756.


Osteoarthritis (OA) is a chronic, progressive disorder of the synovial joints, characterized by focal loss of cartilage and changes in subchondral and marginal bone, synovium, and periarticular structures [1]. The disorder commonly affects weight-bearing joints and results in pain and loss of function [2]. Current therapies (analgesics and nonsteroidal anti-inflammatory drugs [NSAIDs]) are mostly symptomatic and include adjuvant interventions such as weight loss and physical therapy to improve physical function. As the understanding of the pathogenesis of joint destruction in OA increases, new therapeutic approaches are targeting the tissue degradation process.

Matrix metalloproteinases (MMPs) are a group of approximately 30 proteolytic enzymes, which collectively degrade all the components of the extracellular matrix during tissue formation and remodelling [3, 4]. As degradation of the extracellular matrix is essential for growth and progression of malignant tumors, MMP inhibitors have been extensively studied as potential anticancer agents [4]. MMPs have also long been implicated in the joint destruction process that occurs in arthritis, and MMP inhibitors have been studied in the treatment of both rheumatoid arthritis and OA [57]. Increased levels of MMP-1 and MMP-3 have been observed in the cartilage and synovium of patients with OA and have been correlated with the severity of the condition [8]. Animal models have shown that MMPs were good therapeutic targets [9]. Despite promising preclinical data [1013], however, development of MMP inhibitors in arthritis seldom continued into humans, hampered by safety issues or an unfavorable efficacy profile in other indications [14]. To our knowledge, no controlled long-term studies with MMP inhibitors in OA have been performed to date.

In long-term trials in oncology, the use of MMP inhibitors has been associated with musculoskeletal toxicity, which was the primary reason for program termination for some compounds of this class [15, 16]. Signs and symptoms described in the literature generally consisted of musculoskeletal pain and inflammation, usually originating in the upper shoulder girdle or hands, and also arthralgia, myalgia, joint stiffness, limb pain, Dupuytren contracture, and/or peripheral pain and oedema [4, 1723]. In published studies, musculoskeletal symptoms tended to occur after 2 to 3 months of treatment, to be dose-dependent, and to be generally reversible within 1 to 3 weeks of treatment discontinuation [4, 1719, 23].

PG-116800 is a member of the hydroxyproline-based hydroxamic acid class of MMP inhibitors which has been shown to inhibit the joint damage caused by iodoacetate injection into rat knees, an experimental model for OA [24]. PG-116800 has high affinity for MMP-2, -3, -8, -9, -13, and -14, the therapeutic targets, while having substantially lower affinity for MMP-1 and -7, both of which were thought to be implicated in the pathogenesis of musculoskeletal toxicity [25].

This randomized, double-blind, placebo-controlled, multicenter, parallel-group, dose-response study was conducted to evaluate the relative efficacy of PG-116800 versus placebo in slowing the progression of joint space narrowing in knee OA (structural objective) and in reducing symptoms and/or improving function (symptomatic objective) over the course of 12 months and to assess the safety of its administration. The minimum joint space width (JSW) in the medial compartment of the tibiofemoral joint was measured from microfocal knee radiographs obtained by fluoroscopic positioning. Microfocal radiography was chosen for this study since investigators had shown that quantitative evaluation of measurements obtained from macroradiographs detected JSW changes within a shorter time frame than was possible using standard radiography [26, 27]. Here, we report one of the largest studies of MMP inhibitors and describe the adverse musculoskeletal effects of the drug.

Materials and methods

Design and patients

Men and postmenopausal and/or hysterectomized women, 40 to 80 years of age, with primary OA of the knee according to the American College of Rheumatology criteria [28] were eligible to participate in the study. For the purpose of the radiographic assessment, patients had to have at least one knee that could be designated a 'signal knee' on microfocal x-ray, as defined by at least oneosteophyte in either the medial or the lateral compartment of the tibiofemoral joint (tibial spine osteophytes could be included), a JSW in the medial tibiofemoral compartment greater than or equal to 2 mm and less than 4.5 mm in the semiflexed view, and a medial compartment JSW that was narrower than the lateral compartment JSW for the same knee. Exclusion criteria included secondary OA, nonosteoarthritic causes of knee pain, significant medical or psychiatric conditions, chronic shoulder disease, severe obesity, previous intra-articular injection of either knee or of any other joint with steroids, previous intra-articular injection of the signal knee with hyaluronic acid, diagnostic arthroscopy, use of systemic steroids or knee injury within 3 months of study entry, history of surgery or surgical arthroscopy of the signal knee, and recent therapy with a drug of the tetracycline class, with calcitonin, systemic fluoride, bisphosphonate, glucosamine, digoxin, warfarin, inhibitors, or inducers of cytochrome P450 3A4, 1,25 (OH)2 D3 or with more than 400 IU/day of vitamin D2 or D3.

The study was conducted in 16 centers in the UK and 8 centers in Hungary between July 2002 and February 2004 in compliance with International Conference on Harmonization guidelines, the US Code of Federal Regulations, European Community guidelines, and the Declaration of Helsinki. Each study site's ethics committee approved the protocol, and all patients provided written informed consent. An Independent Data Monitoring Committee (IDMC) monitored unblinded data for safety.

Treatments administered

Eligible patients were randomly assigned to receive orally twice daily for 1 year onecapsule containing either placebo or one of fourdoses of PG-116800 (25, 50, 100, or 200 mg). The PG-116800 capsules were supplied by Procter & Gamble Pharmaceuticals (Mason, OH, USA) as white opaque gelatin shells containing PG-530742 (the dihydrated sodium salt PG-116800), lactose monohydrate (National Formulary, NF), and magnesium stearate NF. Placebo capsules were identical in composition and appearance to the capsules containing active drug but did not contain drug substance. During the in-life portion of the study (November 2003), the IDMC recommended discontinuing the 200-mg dose based on an increased frequency of musculoskeletal adverse effects.

Outcome and safety assessments

The progression of joint space narrowing in the osteoarthritic knee (structural primary endpoint) was evaluated by measuring the 1-year change from baseline in minimum JSW in the medial compartment of the tibiofemoral joint of the signal knee, using microfocal knee radiographs obtained in the semiflexed position. To ensure proper alignment of the tibial plateau and centering of the tibial spines relative to the femoral notch, fluoroscopic positioning was performed before the radiograph was acquired. Study radiographers in twocenters (one in Hungary and one in the UK) performing microfocal x-ray underwent specific training and testing before participating in the study. The reduction of symptoms and/or improvement of function of the knee (symptomatic primary endpoint) was measured by the Western Ontario and McMaster Universities osteoarthritis index (WOMAC) total score (version3.1) at 1 year [29]. Secondary endpoints included the consumption of analgesics and NSAIDs as pain medication for OA and the Patient Global Assessment ('Considering all the ways your OA affects you, how have you been in the last 48 hours?').

Tolerability and safety were evaluated at each visit through interviews by the site personnel. All treatment-emergent adverse events were recorded, and their severity (mild, moderate, or severe) and relationship to the study drug (doubtful, possible, or probable) were graded by the investigator. Clinical laboratory tests, electrocardiograms, and chest radiographs were performed regularly.

The occurrence of musculoskeletal effects was carefully monitored because of musculoskeletal toxicity reported with other MMP inhibitors in oncology trials [17]. Based on a literature review, a working definition of the MMP inhibitor-associated musculoskeletal syndrome (MSS) was developed and included painless loss of range of motion (ROM) in large joints (particularly in the shoulders), joint stiffness and joint swelling, soft tissue pain, and fibrosis of palmar tendons (Dupuytren contracture). To assess potential musculoskeletal toxicity, each follow-up visit included serial measurements of shoulder ROM (anterior flexion, abduction, internal rotation, and external rotation) at the site by an appropriately trained person using a goniometer; examination of palmar tendons for evidence of palmar tendon fibrosis, tendinitis, fasciitis, or inflammation; and elicitation of MSS symptoms through questioning. A diagnosis of MSS was considered likely if a patient (a) developed palmar fibrosis, (b) lost at least 30% of any shoulder ROM, as measured by goniometer, or (c) presented with musculoskeletal signs or symptoms the investigator considered to be significant and consistent with those previously reported in the literature. No specific blood biomarker of MMP inhibitor-related musculoskeletal toxicity was identified and collected. Patients with signs and symptoms consistent with MSS could undergo ultrasound of the hands and shoulders at the investigators' discretion and the sponsor's recommendation. The IDMC reviewed unblinded study data quarterly to enable the early discontinuation of doses that appeared to be associated with any unacceptable adverse events, with particular attention to any evidence of MSS.

Sample size determination

The study was designed to detect a 50% reduction in JSW change at 1 year with 80% power, using one-sided comparisons, each made at an α value of 0.10 (not adjusted). The sample size of 75 patients per group (for a total of 375 enrolled) assumed a yearly loss in JSW of 0.2 mm, with a standard deviation of 0.25 mm. Treatment groups were sized to accommodate a 20% dropout rate. In addition, without regard for multiple outcomes, the study had 80% power to detect a 17% effect of treatment versus placebo with respect to pain modification, as measured by the pain subscale of the WOMAC index (assuming one-sided comparisons, each made at an α value of 0.10, a placebo mean of 160 mm, and a standard deviation of 70 mm).

Randomization method

Patients were randomly assigned to fivetreatment groups using an Interactive Voice Response System (IVRS). Randomization of patients was balanced according to their current use of estrogens or selective estrogen-receptor modulators (SERMs) in which twostrata were formed. An adaptative randomization technique [30] was employed to better achieve treatment balance within each site, each estrogen/SERM stratum, and across the entire study. The pooled study center was treated as a stratification factor for all applicable efficacy endpoint analyses. The treatment codes were controlled by the clinical supplies department of Procter & Gamble Pharmaceuticals.


This was a double-blind study with limited access to the randomization code. Study drug was dispensed according to the instructions provided via the IVRS. The treatment each patient received was not disclosed to study personnel, participants, contractors (except for clinical supplies distributor and IVRS contractors), or the sponsor (except for select clinical supplies, pharmacovigilance, bioanalytical, and pharmacokinetics personnel).

Statistical analysis

An intent-to-treat (ITT) analysis was conducted on all patients who were randomly assigned and took at least one dose of study drug. Any missing data for these patients were not imputed in the primary analyses. Thus, the efficacy analyses were based on the ITT patients with observed data at the time point under consideration. Paired t tests were used to test change from baseline values.

The 1-year change from baseline in minimum JSW in the medial compartment of the tibiofemoral joint of the signal knee was determined from the 12-month radiograph back to baseline. The significance of the 1-year change from baseline in minimum JSW and WOMAC total score was estimated using an analysis of variance. Minimum JSW change estimates were adjusted for baseline JSW, pooled center, and baseline use of estrogen or SERM drug replacement therapy as covariates. WOMAC change estimates were adjusted for baseline total scores, pooled center, and baseline use of estrogen or SERM drug replacement therapy as covariates. Each dose group was compared with the placebo control group. Each of these comparisons was made using a one-sided test with an α value of 0.10. Secondary endpoints were analyzed in a similar fashion.

Safety analyses included the ITT population. ROM measurements were analyzed and adjusted for pooled study centers as covariates. Comparisons between treatment groups were made to placebo for each visit. Additionally, the proportion of patients with detectable decreases in ROM was summarized. The proportion of patients with palmar tendon fibrosis, palmar tendon tendonitis, or palmar tendon fasciitis was summarized. Adverse events were tabulated and summarized according to the Medical Dictionary for Regulatory Activities (MedDRA). Percentages of patients reporting a decrease in shoulder ROM were compared to placebo by means of the Fisher exact test. Changes in joint symptom severity were tabulated and summarized, and percentages of patients reporting an increase in joint symptoms were compared to placebo by means of the Cochran-Mantel-Haenszel test after adjusting for pooled centers. P values for safety analyses were provided as flags for further investigation and were not adjusted for multiple testing.


A summary of patient accountability is presented in Table 1. Of 401 patients randomly assigned into the study, 395 took at least one dose of study drug (ITT population). The per-protocol population comprised 296 patients. Twenty-one patients in the 200-mg dose group were withdrawn from the study for safety reasons per IDMC recommendation, which was classified as a major protocol deviation. All patients were analyzed according to group assignment. Overall, treatment groups were balanced with regard to demographics and baseline characteristics (Table 2). The majority of patients were postmenopausal females who were not using hormone replacement therapy.

Table 1 Patient accountability
Table 2 Demographic and baseline characteristics (intent-to-treat population)

Efficacy results

Overall, PG-116800 administration did not result in slower progression of joint space narrowing compared with that observed with placebo. There was no statistically significant difference between placebo and any of the PG-116800 groups in mean change in minimum JSW of the signal knee after 1 year of treatment (Table 3). The decrease from baseline in mean JSW of the signal knee was observed in all dose groups at month 12 and was statistically significant in all with the exception of the 25-mg group. Similarly, analysis of the symptomatic primary endpoint did not demonstrate statistically significant differences between placebo and each of the PG-116800 groups on total WOMAC scores (Table 4). Analysis of secondary endpoints showed that Patient Global Assessment scores improved significantly from baseline over the course of 12 months in all groups with the exception of the 200-mg group, but there were no differences between treatment groups (data not shown). At month 12, the consumption of pain medication for OA had decreased from baseline in all groups, the change from baseline reaching statistical significance in the placebo group only (data not shown).

Table 3 One-year change from baseline in joint space width of the signal knee (intent-to-treat population)
Table 4 One-year change from baseline in total WOMAC scores (intent-to-treat population)

Safety results

The proportion of patients reporting adverse events was greatest in the twohighest dose groups (Table 5). Serious adverse events and withdrawals due to adverse events occurred with the highest frequency in the 200-mg group, in which adverse events were the primary cause of withdrawal. The majority of adverse events, serious adverse events, and withdrawals due to adverse events were musculoskeletal and connective tissue disorders. Arthralgia was the single most frequent adverse event overall (35% of all patients) and was also the most frequent serious adverse event and reason for withdrawal. Two patients, one in the 100-mg group and one in the 200-mg group, died of pulmonary embolism and cardiac failure, respectively, during the study. The investigators considered their deaths to be unrelated to study drug.

Table 5 Summary of adverse events (intent-to-treat population)

Objective monitoring of musculoskeletal toxicity included goniometer assessment of shoulder ROM at each visit. Of the 391 patients who had at least onevalid shoulder ROM measurement during a follow-up visit, 90 (23%) had a decrease of at least 30% from baseline for at least one ROM measurement of either shoulder (Table 6). The percentage of patients with changes in ROM was statistically significantly greater in the twohighest dose groups relative to placebo. Twenty-nine patients (7%) had an ROM decrease of at least 30% in both shoulders. Reproducibility of the decrease in shoulder ROM (that is, decrease of 30% in a measurement at twoconsecutive visits) was used as a practical measure of specificity of the finding. Of 376 patients with valid measurements, 40 patients (11%) had a decrease of at least 30% in the same measurement for either shoulder at twoor more consecutive visits. The decrease mostly affected shoulder internal and external rotations; fewer patients had decreases in shoulder abduction and anterior flexion. Finally, 34 of 391 patients (9%) had decreases of at least 30% in two or more ROM measurements of either shoulder at the same visit. The percentage of patients with two or more ROM decreases in the twohighest dose groups was statistically significantly greater relative to placebo (Table 6).

Table 6 Changes from baseline in shoulder range of motion (intent-to-treat population)

Adverse events affecting the hand were reported in 52 of 395 patients (13%), nearly half of whom were in the 200-mg group (Table 7). Common signs included hand oedema (5 patients), palmar fibrosis (15 patients), Dupuytren contracture (5 patients), and tendon thickening/nodules (3 patients). Most patients with hand involvement (16 patients) had asymptomatic nodules without function loss (contracture), hand pain, or hand oedema. Of the 7 symptomatic patients, 4 were in the 200-mg group, 2 in the 100-mg group, and 1 in the 50-mg group. A mild function loss in hands (mild contracture) was reported in 5 symptomatic patients, 2 of whom had accompanying pain and localized oedema, which were not reported as separate adverse events. Two other symptomatic patients included 1 patient with mild hand pain (localized hypersensitivity) and another patient with pain and oedema in the hand without any functional loss. The first cases of palmar fibrosis appeared at month3, and the number of affected patients in the 200-mg dose group reached statistical significance versus placebo at month9 (18% versus 1%; p = 0.001). The finding of fibrosis on hand palpation was confirmed by hand ultrasound in 20 patients. Ultrasound was not a scheduled procedure in the study protocol. The usual appearance of fibrosis on ultrasound consisted of elongated or oblique bands along flexor tendons, in their proximity but not within the tendons. Their size varied depending on the examination technique (palpation versus ultrasound) and ranged from a few millimeters to approximately 2 cm in length. In one case, a thickening of fascia palmaris of 6 cm in length was described.

Table 7 Patients with hand adverse events (intent-to-treat population)

The threemost frequent shoulder adverse events were arthralgia, stiffness, and myalgia. The first cases of shoulder arthralgia, which affected either shoulder with similar frequency, appeared after 1 month of dosing. Over the course of the whole study, 13%, 16%, 13%, 21%, and 25% of patients, respectively, in the placebo, 25-mg, 50-mg, 100-mg, and 200-mg dose groups reported shoulder arthralgia. Shoulder stiffness, which was usually isolated but sometimes accompanied by stiffness in other joints, was reported starting at month6. Over the course of the study, 1%, 4%, 6%, 9%, and 6% of patients, respectively, in the placebo, 25-mg, 50-mg, 100-mg, and 200-mg groups reported shoulder stiffness. Shoulder myalgia generally affected the deltoid muscle and was reported in 1%, 1%, 4%, 3%, and 6% of patients, respectively, in the placebo, 25-mg, 50-mg, 100-mg, and 200-mg groups over the course of the study. The threemost frequently reported shoulder adverse events could represent different reporting of a similar symptomatology. An analysis combining the threemost frequent shoulder adverse events showed a dose-related response starting after 3 months of dosing, continuing through months6 and 9, and disappearing at month 12 (data not shown).

Other symptoms included neck pain (3%, 4%, 4%, 6%, and 10% of patients, respectively, in the placebo, 25-mg, 50-mg, 100-mg, and 200-mg groups), which could be due to referred pain originating in the shoulder, and shoulder periarthritis (1%, 3%, 1%, 4%, and 6% of patients, respectively, in the placebo, 25-mg, 50-mg, 100-mg, and 200-mg groups). Shoulder ultrasounds were performed at the investigator's discretion in some symptomatic patients. Of 13 symptomatic patients in whom rotator cuff rupture (partial or complete tear) was found on shoulder ultrasound during the study, 3 patients each were in the placebo, 50-mg, 100-mg, and 200-mg groups and 1 patient was in the 25-mg group. Few patients (7%) had a combination of shoulder and hand symptoms that appeared independent of each other.

Most patients withdrawn from the study because of development of musculoskeletal adverse events were followed until resolution or stabilization of symptoms. Cessation of MMP inhibitor administration seemed to bring symptomatic relief as well as improvement in ROM in patients with shoulder involvement. Among 19 patients followed up for clinically significant shoulder involvement, the improvement was complete in 10 patients after a mean period of 122 days and partial in 3 patients after a mean period of 107 days (partial means that the symptoms or signs decreased but never returned to baseline status) (Table 8). Additionally, 3 patients with shoulder involvement recovered completely while still exposed to active drug and 3 other patients recovered partially. Complete recovery was seemingly promoted by local glucocorticoid injections in 4 patients and physiotherapy in 5 patients. Partial recovery was promoted by local glucocorticoid injections in 1 patient and physiotherapy in another patient. Patients were also administered NSAIDs and paracetamol for the shoulder symptoms or index knee OA. Recovery from shoulder symptoms was faster in patients who were exposed to the MMP inhibitors for a shorter period of time. Six patients with shoulder involvement underwent temporary drug interruption followed by drug rechallenge. The symptoms recurred in 4 of these patients. Of the 21 patients followed for hand adverse events, only 2 recovered completely after a mean period of 224 days from their last dose of study drug. Eleven patients recovered partially after a mean period of 210 days and the hand fibrosis was unchanged in 7 patients after a mean period of 196 days after stopping the study drug (partial recovery means that either the symptoms decreased or the extent of fibrosis decreased in size in clinical or ultrasound assessment) (Table 8). Of the 3 patients followed for involvement of both shoulders and hands, 2 patients recovered partially after a mean period of 170 days from the last dose of study drug (partial recovery means that symptoms or signs in at least one site of involvement, usually the shoulder, decreased). Additionally, 1 patient with both shoulder and hand involvement recovered partially while still taking active drug. Initial symptoms of hand pain or oedema, which accompanied the diagnosis of hand fibrosis, tended to disappear after stopping study drug.

Table 8 Follow-up data after PG-116800 withdrawal in patients with clinically significant shoulder findings and hand adverse events


This proof-of-concept and dose-ranging study failed to demonstrate efficacy of PG-116800, an MMP inhibitor, in modifying the course of knee OA in patients with mild to moderate disease during a 12-month treatment period, as determined by microfocal x-ray assessment of JSW and analysis of WOMAC total scores, despite promising results of nonclinical studies [9, 10, 12].

The study confirmed that microfocal radiography is a sensitive tool in detecting a decrease in JSW in osteoarthritic patients over the course of a 12-month period. The progression of OA as evidenced by a decreasing JSW coexisted with symptomatic improvement from baseline in all treatment groups, including placebo. Significant symptomatic placebo response has been observed in OA studies of both intra-articular and oral medications and was noted in other recent studies [3134]. This study seems to confirm that patients are likely to improve symptomatically and that the placebo effect persists for at least 12 months. Progression of JSW narrowing accompanied by symptomatic improvement over the course of a 1-year time period calls for further research in surrogate markers of OA activity and progression.

Although the majority of the cases described in the literature have been associated with the use of marimastat in patients with cancer, musculoskeletal toxicity has been reported in association with most, if not all, other MMP inhibitors. Musculoskeletal effects have been observed in preclinical studies with other MMP inhibitors [35]. In preclinical studies conducted with PG-116800, swelling was observed around the joints of both rats and dogs during chronic toxicology studies (3 and 6 months in rats and 12 months in dogs;, Procter & Gamble Pharmaceuticals, unpublished data). In apparent relation to the swelling, rats and dogs had accumulation of collagen associated with the joint structures. In dogs, after 12 months of study, proliferation of periosteal fibrous tissue and resorption of bone were observed in the joint. The effects appeared to be reversible when observed at the end of a 3-month recovery period.

Because the potential musculoskeletal toxicity of MMP inhibitors was recognized before the study start, safety measures included shoulder ROM assessment and monitoring with an IDMC. Although a detailed hand examination was part of the baseline physical examination, investigators were sensitized to hand findings after cases of asymptomatic nodules were observed. IDMC-driven withdrawal of patients in the 200-mg dose group could result in cases of musculoskeletal toxicity being over-reported at withdrawal in this group following the unblinding. The hand findings had a striking clinical resemblance to those of early development of Dupuytren contracture.

One of the major challenges in identifying patients with drug-related musculoskeletal toxicity in the study population was the need to differentiate from the symptoms and signs of OA itself. Although upper girdle involvement in OA is less frequent than involvement of hips and knees, it is not uncommon for the two to coexist in the same patient. Additionally, in published series, up to 50% of asymptomatic patients more than 60 years of age had partial or full rotator cuff tears in shoulder imaging studies [36]. Since shoulder ultrasound was not part of the screening process in this study but was used for further workup in some patients reporting shoulder symptoms, the finding of a rotator cuff tear was not helpful in differentiating drug toxicity from age-related tissue degeneration.

The occurrence of the most frequently reported shoulder adverse events (arthralgia, myalgia, and joint stiffness) followed a time- and dose-response pattern. This pattern disappeared after 12 months. Early termination of the study by patients in the highest dose group as per the IDMC recommendation after most of the patients had been dosed for 6 months offers a plausible explanation together with early withdrawals of symptomatic patients.

A number of theories to explain the occurrence of the musculoskeletal effects observed with MMP inhibitors have been offered, but none of them has definitive proof to support it. Inhibition of sheddase activity attributed to hydroxamate-based MMP inhibitors might be responsible for musculoskeletal symptoms [4]. Sheddase activity converts membrane-bound cytokines and receptors to inactive forms. Therefore, inhibiting this conversion could result in inflammation. In support of this theory, patients treated with BMS-275291 (which has reduced activity against sheddases) initially did not experience MSS [4]. This finding was verified after longer treatment duration when similar adverse events were reported. Another theory was that musculoskeletal toxicity might be caused by MMP-1 inhibition [4]. In an attempt to minimize these effects, Bayer Corp. (West Haven, CT, USA) and Agouron Pharmaceuticals Inc. (La Jolla, CA, USA) developed tanomastat and prinomastat, which strongly inhibit MMP-2 and -9 and weakly inhibit MMP-1, -7, and -11. However, prinomastat continued to exhibit musculoskeletal toxicity, whereas tanomastat was abandoned for lack of efficacy in oncology indications [4, 37].


The study did not demonstrate efficacy of PG-116800, an MMP inhibitor, in the treatment of patients with knee OA. It indicated that musculoskeletal side effects compromise the safety of long-term (greater than 3 months) systemic administration of the compound. Shoulders were affected clinically, with a decrease in ROM and an increase in pain largely reversible upon drug discontinuation. A dose- and time-related focal accumulation of tissue consistent with palmar tendon fibrosis was observed after 2 to 3 months of treatment. These musculoskeletal side effects were similar to those reported for other MMP inhibitors. This is likely to reflect a class effect, as these adverse effects are now reported for most of the MMP inhibitors, and this further suggests that these agents are unlikely to be of practical clinical use for the treatment of OA.



= Independent Data Monitoring Committee


= intent-to-treat


= Interactive Voice Response System


= joint space width


= Medical Dictionary for Regulatory Activities


= matrix metalloproteinase


= musculoskeletal syndrome


= National Formulary


= nonsteroidal anti-inflammatory drug


= osteoarthritis


= range of motion


= selective estrogen-receptor modulator


= Western Ontario and McMaster Universities osteoarthritis index.


  1. 1.

    Keuttner K, Godberg VM, (Eds): Osteoarthritic Disorders. 1995, Rosemont, IL, USA: American Academy of Orthopaedic Surgeons, 21-25.

    Google Scholar 

  2. 2.

    Oliveria SA, Felson DT, Reed JI, Cirillo PA, Walker AM: Incidence of symptomatic hand, hip, and knee osteoarthritis among patients in a health maintenance organization. Arthritis Rheum. 1995, 38: 1134-1141. 10.1002/art.1780380817.

    CAS  Article  PubMed  Google Scholar 

  3. 3.

    Coussens LM, Fingleton B, Matrisian LM: Matrix metalloproteinase inhibitors and cancer: trials and tribulations. Science. 2002, 295: 2387-2392. 10.1126/science.1067100.

    CAS  Article  PubMed  Google Scholar 

  4. 4.

    Chau I, Rigg A, Cunningham D: Matrix metalloproteinase inhibitors – an emphasis on gastrointestinal malignancies. Crit Rev Oncol Hematol. 2003, 45: 151-176. 10.1016/S1040-8428(02)00015-X.

    Article  PubMed  Google Scholar 

  5. 5.

    Shaw T, Nixon JS, Bottomley KM: Metalloproteinase inhibitors: new opportunities for the treatment of rheumatoid arthritis and osteoarthritis. Expert Opin Investig Drugs. 2000, 9: 1469-1478. 10.1517/13543784.9.7.1469.

    CAS  Article  PubMed  Google Scholar 

  6. 6.

    Bigg HF, Rowan AD: The inhibition of metalloproteinases as a therapeutic target in rheumatoid arthritis and osteoarthritis. Curr Opin Pharmacol. 2001, 1: 314-320. 10.1016/S1471-4892(01)00055-8.

    CAS  Article  PubMed  Google Scholar 

  7. 7.

    Cawston TE: Metalloproteinase inhibitors and the prevention of connective tissue breakdown. Pharmacol Ther. 1996, 70: 163-182. 10.1016/0163-7258(96)00015-0.

    CAS  Article  PubMed  Google Scholar 

  8. 8.

    Lohmander LS, Hoerrner LA, Lark MW: Metalloproteinases, tissue inhibitor, and proteoglycan fragments in knee synovial fluid in human osteoarthritis. Arthritis Rheum. 1993, 36: 181-189. 10.1002/art.1780360906.

    CAS  Article  PubMed  Google Scholar 

  9. 9.

    Janusz MJ, Hookfin EB, Heitmeyer SA, Woessner JF, Freemont AJ, Hoyland JA, Brown KK, Hsieh LC, Almstead NG, De B, et al: Moderation of iodoacetate-induced experimental osteoarthritis in rats by matrix metalloproteinase inhibitors. Osteoarthritis Cartilage. 2001, 9: 751-760. 10.1053/joca.2001.0472.

    CAS  Article  PubMed  Google Scholar 

  10. 10.

    Fujisawa T, Igeta K, Odake S, Morita Y, Yasuda J, Morikawa T: Highly water-soluble matrix metalloproteinases inhibitors and their effects in a rat adjuvant-induced arthritis model. Bioorg Med Chem. 2002, 10: 2569-2581. 10.1016/S0968-0896(02)00109-8.

    CAS  Article  PubMed  Google Scholar 

  11. 11.

    Lewis EJ, Bishop J, Bottomley KM, Bradshaw D, Brewster M, Broadhurst MJ, Brown PA, Budd JM, Elliott L, Greenham AK, et al: Ro 32-3555, an orally active collagenase inhibitor, prevents cartilage breakdown in vitro and in vivo. Br J Pharmacol. 1997, 121: 540-546. 10.1038/sj.bjp.0701150.

    PubMed Central  CAS  Article  PubMed  Google Scholar 

  12. 12.

    Brewster M, Lewis EJ, Wilson KL, Greenham AK, Bottomley KM: Ro 32-3555, an orally active collagenase selective inhibitor, prevents structural damage in the STR/ORT mouse model of osteoarthritis. Arthritis Rheum. 1998, 41: 1639-1644. 10.1002/1529-0131(199809)41:9<1639::AID-ART15>3.0.CO;2-0.

    CAS  Article  PubMed  Google Scholar 

  13. 13.

    Close DR: Matrix metalloproteinase inhibitors in rheumatic diseases. Ann Rheum Dis. 2001, 60 (Suppl 3): iii62-67.

    PubMed Central  CAS  PubMed  Google Scholar 

  14. 14.

    Hirte H, Vergote IB, Jeffrey JR, Grimshaw RN, Coppieters S, Schwartz B, Tu D, Sadura A, Brundage M, Seymour L: A phase III randomized trial of BAY 12-9566 (tanomastat) as maintenance therapy in patients with advanced ovarian cancer responsive to primary surgery and paclitaxel/platinum containing chemotherapy: a National Cancer Institute of Canada Clinical Trials Group Study. Gynecol Oncol. 2006, 102: 300-308. 10.1016/j.ygyno.2005.12.020.

    CAS  Article  PubMed  Google Scholar 

  15. 15.

    Nemunaitis J, Poole C, Primrose J, Rosemurgy A, Malfetano J, Brown P, Berrington A, Cornish A, Lynch K, Rasmussen H, et al: Combined analysis of studies of the effects of the matrix metalloproteinase inhibitor marimastat on serum tumor markers in advanced cancer: selection of a biologically active and tolerable dose for longer-term studies. Clin Cancer Res. 1998, 4: 1101-1109.

    CAS  PubMed  Google Scholar 

  16. 16.

    Rasmussen HS, McCann PP: Matrix metalloproteinase inhibition as a novel anticancer strategy: a review with special focus on batimastat and marimastat. Pharmacol Ther. 1997, 75: 69-75. 10.1016/S0163-7258(97)00023-5.

    CAS  Article  PubMed  Google Scholar 

  17. 17.

    Bramhall SR, Hallissey MT, Whiting J, Scholefield J, Tierney G, Stuart RC, Hawkins RE, McCulloch P, Maughan T, Brown PD, et al: Marimastat as maintenance therapy for patients with advanced gastric cancer: a randomised trial. Br J Cancer. 2002, 86: 1864-1870. 10.1038/sj.bjc.6600310.

    PubMed Central  CAS  Article  PubMed  Google Scholar 

  18. 18.

    Tierney GM, Griffin NR, Stuart RC, Kasem H, Lynch KP, Lury JT, Brown PD, Millar AW, Steele RJ, Parsons SL: A pilot study of the safety and effects of the matrix metalloproteinase inhibitor marimastat in gastric cancer. Eur J Cancer. 1999, 35: 563-568. 10.1016/S0959-8049(99)00007-6.

    CAS  Article  PubMed  Google Scholar 

  19. 19.

    Groves MD, Puduvalli VK, Hess KR, Jaeckle KA, Peterson P, Yung WK, Levin VA: Phase II trial of temozolomide plus the matrix metalloproteinase inhibitor, marimastat, in recurrent and progressive glioblastoma multiforme. J Clin Oncol. 2002, 20: 1383-1388. 10.1200/JCO.20.5.1383.

    CAS  Article  PubMed  Google Scholar 

  20. 20.

    Rosemurgy A, Harris J, Langleben A, Casper E, Goode S, Rasmussen H: Marimastat in patients with advanced pancreatic cancer: a dose-finding study. Am J Clin Oncol. 1999, 22: 247-252. 10.1097/00000421-199906000-00007.

    CAS  Article  PubMed  Google Scholar 

  21. 21.

    Wojtowicz-Praga S, Torri J, Johnson M, Steen V, Marshall J, Ness E, Dickson R, Sale M, Rasmussen HS, Chiodi TA, et al: Phase I trial of Marimastat, a novel matrix metalloproteinase inhibitor, administered orally to patients with advanced lung cancer. J Clin Oncol. 1998, 16: 2150-2156.

    CAS  PubMed  Google Scholar 

  22. 22.

    Hutchinson JW, Tierney GM, Parsons SL, Davis TR: Dupuytren's disease and frozen shoulder induced by treatment with a matrix metalloproteinase inhibitor. J Bone Joint Surg Br. 1998, 80: 907-908. 10.1302/0301-620X.80B5.8464.

    CAS  Article  PubMed  Google Scholar 

  23. 23.

    Brown PD: Ongoing trials with matrix metalloproteinase inhibitors. Expert Opin Invest Drugs. 2000, 9: 2167-2177. 10.1517/13543784.9.9.2167.

    CAS  Article  Google Scholar 

  24. 24.

    Guingamp C, Gegout-Pottie P, Philippe L, Terlain B, Netter P, Gillet P: Mono-iodoacetate-induced experimental osteoarthritis: a dose-response study of loss of mobility, morphology, and biochemistry. Arthritis Rheum. 1997, 40: 1670-1679. 10.1002/art.1780400917.

    CAS  Article  PubMed  Google Scholar 

  25. 25.

    Holmbeck K, Bianco P, Caterina J, Yamada S, Kromer M, Kuznetsov SA, Mankani M, Robey PG, Poole AR, Pidoux I, Ward JM, et al: MT1-MMP-deficient mice develop dwarfism, osteopenia, arthritis and connective tissue disease due to inadequate collagen turnover. Cell. 1999, 99: 81-92. 10.1016/S0092-8674(00)80064-1.

    CAS  Article  PubMed  Google Scholar 

  26. 26.

    Buckland-Wright JC, Macfarlane DG, Williams SA, Ward RJ: Accuracy and precision of joint space width measurements in standard and macroradiographs of osteoarthritic knees. Ann Rheum Dis. 1995, 54: 872-880.

    PubMed Central  CAS  Article  PubMed  Google Scholar 

  27. 27.

    Buckland-Wright JC, Macfarlane DG, Lynch JA, Jasani MK: Quantitative microfocal radiography detects changes in OA knee joint space width in subjects in placebo-controlled trial of NSAID therapy. J Rheumatol. 1995, 22: 937-943.

    CAS  PubMed  Google Scholar 

  28. 28.

    Altman R, Asch E, Bloch D, Bole G, Borenstein D, Brandt K, Christy W, Cooke TD, Greenwald R, Hochberg M, et al: Development of criteria for the classification and reporting of osteoarthritis. Classification of osteoarthritis of the knee. Diagnostic and Therapeutic Criteria Committee of the American Rheumatism Association. Arthritis Rheum. 1986, 29: 1039-1049. 10.1002/art.1780290816.

    CAS  Article  PubMed  Google Scholar 

  29. 29.

    Bellamy N, Buchanan WW, Goldsmith CH, Campbell J, Stitt LW: Validation study of WOMAC: a health status instrument for measuring clinically important patient relevant outcomes to antirheumatic drug therapy in patients with osteoarthritis of the hip or knee. J Rheumatol. 1988, 15: 1833-1840.

    CAS  PubMed  Google Scholar 

  30. 30.

    Pocock SJ, Simon R: Sequential treatment assignment with balancing for prognostic factors in the controlled clinical trial. Biometrics. 1975, 31: 103-115. 10.2307/2529712.

    CAS  Article  PubMed  Google Scholar 

  31. 31.

    Clegg DO, Reda DJ, Harris CL, Klein MA, O'Dell JR, Hooper MM, Bradley JD, Bingham CO, Weisman MH, Jackson CG, et al: Glucosamine, chondroitin sulfate, and the two in combination for painful knee osteoarthritis. N Engl J Med. 2006, 354: 795-808. 10.1056/NEJMoa052771.

    CAS  Article  PubMed  Google Scholar 

  32. 32.

    Ernst E: Glucosamine and chondroitin sulfate for knee osteoarthritis. N Engl J Med. 2006, 354: 2184-2185. 10.1056/NEJMc060829.

    CAS  Article  PubMed  Google Scholar 

  33. 33.

    Altman RD, Moskowitz R: Intraarticular sodium hyaluronate (Hyalgan) in the treatment of patients with osteoarthritis of the knee: a randomized clinical trial. Hyalgan Study Group. J Rheumatol. 1998, 25: 2203-2212.

    CAS  PubMed  Google Scholar 

  34. 34.

    Lehmann R, Brzosko M, Kopsa P, Nischik R, Kresisse A, Thurston H, Litschig S, Sloan VS: Efficacy and tolerability of lumiracoxib 100 mg once daily in knee osteoarthritis: a 13-week, randomized, double-blind study vs. placebo and celecoxib. Curr Med Res Opin. 2005, 21: 517-526. 10.1185/030079905X38196.

    CAS  Article  PubMed  Google Scholar 

  35. 35.

    Wojtowicz-Praga S: Clinical potential of matrix metalloprotease inhibitors. Drugs R D. 1999, 1: 117-129. 10.2165/00126839-199901020-00001.

    CAS  Article  PubMed  Google Scholar 

  36. 36.

    Sher JS, Uribe JW, Posada A: Abnormal findings on magnetic resonance images of asymptomatic shoulders. J Bone Joint Surg. 1995, 77A: 10-15.

    Google Scholar 

  37. 37.

    Bissett D, O'Byrne KJ, von Pawel J, Gatzemeier U, Price A, Nicoloson M, Mercier R, Mazabel E, Penning C, Zhang MH, et al: Phase III study of matrix metalloproteinase inhibitor prinomastat in non-small-cell lung cancer. J Clin Oncol. 2005, 23: 842-849. 10.1200/JCO.2005.03.170.

    CAS  Article  PubMed  Google Scholar 

Download references


The authors thank Alain Moufarege for his contribution to the design and execution of the study, Carol Tonkin and Karen Maunder for their contributions to the setting up of microfocal radiography units and radiographic assessments, Wajanat Jan for performing ultrasound examinations and assisting in the interpretation of findings, and Pal N Kaposi for his contributions to the setting up of microfocal radiography units and radiographic assessments and for performing ultrasound examinations and assisting in the interpretation of findings. This study was funded by Procter & Gamble Pharmaceuticals. Isabelle Darnis-Wilhelm prepared the manuscript and provided editorial assistance on behalf of Procter & Gamble Pharmaceuticals.

Investigators involved in the study were as follows: Hungary: Geza Bálint, Budapest; Katalin Bors, Gizella telep; Dezso Dávid, Petöfi; Gyula Poór, Budapest; Laszlo Hodinka, Budapest; Ferenc Szanyó, Györ; Miklos Szendröi, Budapest; Laszlo Tamási, Miskolc; UK: David Doyle, London; Ken Evans, Tonbridge; Ian Farmer, Ashford; Jean Fraser, Awpley Bridge; Peter Harvey, Addlestone; Neil Jackson, East Crowley; Ronald Jubb, Birmingham; Richard Keen, Stanmore; Phil Marazzi, East Horsley; Mansur Salman, Edgbaston; James Scrivens, Stratford-upon-Avon; Tim Spector, London; HelenYates, Tunbridge Wells.

Author information



Corresponding author

Correspondence to Piotr Krzeski.

Additional information

Competing interests

PK, GAC, KS, RL, JB, and WSA were all employees of Procter & Gamble Pharmaceuticals when the study was conducted. TDS and CB-W were consultants to Procter & Gamble Pharmaceuticals at the time of this study. GB declares that he has no competing interests.

Authors' contributions

GB and TDS were investigators in the study. CB-W developed the microfocal x-ray technique and reviewed all films. PK and JB were the medical monitors. KS and RL were the project leaders, participating in the study design and coordination. GAC performed the statistical analysis. WSA was medical advisor. All authors were involved in data interpretation and read and approved the final version of the manuscript.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Krzeski, P., Buckland-Wright, C., Bálint, G. et al. Development of musculoskeletal toxicity without clear benefit after administration of PG-116800, a matrix metalloproteinase inhibitor, to patients with knee osteoarthritis: a randomized, 12-month, double-blind, placebo-controlled study. Arthritis Res Ther 9, R109 (2007).

Download citation


  • Joint Space Width
  • Interactive Voice Response System
  • Independent Data Monitoring Committee
  • Minimum Joint Space Width
  • Signal Knee