Introduction

Osteoarthritis is a degenerative joint disease characterized by articular cartilage breakdown, joint space narrowing, osteophyte formation and subchondral sclerosis. It is more common among the elderly and is the leading cause for physical disability, low quality of life and health-care costs in industrialized societies [1]. Approximately 13.9% of adults more than 25 years old and 33.6% (12.4 million) of those >65 are affected with OA. The prevalence of symptomatic knee OA accounts for 6.3% of adults more than 30 years old and 9.5% in 63 to 93 years old. In 1997, the estimated cost of hip and knee replacement was 7.9 billion dollars and accounted for 7.1 million of all arthritis-related ambulatory care visits [2]. Osteoarthritis is a heterogenious disease with varied clinical presentations, clinical course and outcomes. Many efforts have been made to prevent or slow the progression of this condition without much long-term success.

Cartilage has been the target for most of the therapeutic approaches in osteoarthritis. It is composed of chondrocytes surrounded by an extra-cellular matrix containing collagen fibers and aggregates which include hyaluronic acid (hyaluronate) and aggrecans. The primary role of hyaluronate in the matrix is to maintain its viscoelasticity and provide loading protection to the cartilage. Hyaluronate was FDA approved in 1997 for intra-articular injection to provide viscosupplementation for the treatment of knee osteoarthritis. Intra-articular hyaluronic acid is indicated for patients who have failed conservative management, patients who cannot tolerate non-steroidal anti-inflammatory drugs because of renal and/or gastrointestinal intolerance, and patients who refuse or in whom surgery is contraindicated [3]. A recent review of randomized placebo-controlled trials suggests that intra-articular hyaluronic acid injections are safe, tolerable, and effective in reducing pain in patients with symptomatic osteoarthritis of the knee [4]. Other studies comparing intra-articular hyaluronic acid and intra-articular corticosteroids concluded that HA may provide superior treatment for chronic osteoarthritis [5]. In vitro studies have shown Fas-induced apoptosis of chondrocytes in patients with osteoarthritis is suppressed by intra-articular hyaluronic acid [6]. The anti-apoptotic effect of hyaluronic acid is mediated by its binding to specific receptors, CD44 and ICAM, suggesting that HA may decrease progression of chondrocyte apoptosis in OA.

Molecular markers have been identified that reflect joint damage and changes in cartilage metabolism. Christgau et al. first described an immunoassay for the measurement of urinary collagen type-II C-telopeptide (CTX-II) as an indicator of cartilage degradation; this study demonstrated a significant increase in urinary CTX-II in patients with RA and OA [7]. CTX-II values in patients with knee OA compared with healthy controls are significantly higher [8]. Hence, CTX-II has been considered to be a biomarker that monitors progression of OA [8–10]. A relationship between radiographic changes in OA and CTX-II levels has also been described [11, 12]. Furthermore, elevation of CTX-II has been associated with subchondral bone changes identified by MRI in patients with OA of the knee [13, 14].

Early disease detection in osteoarthritis remains a challenge. Symptoms of osteoarthritis correlate poorly with radiographic changes [15]. However, a recent study suggests that urinary CTX-II levels correlate with pain in patients with symptomatic knee osteoarthritis [16]. Though there are many studies evaluating pain after treatment with intra-articular hyaluronan, there are none that relate changes in biomarkers with this therapy. The objective of this pilot study was to evaluate the relationship between urinary CTX-II as a biomarker of collagen breakdown and intra-articular treatment with hyaluronic acid in patients with symptomatic knee OA. Given prior evidence of the relationship of CTX-II levels and disease progression, we postulated that the levels of urinary CTX-II would decrease after intra-articular HA injection.

Materials and methods

Patients

We recruited ten patients with symptomatic knee osteoarthritis who failed treatment with oral analgesics and prior steroid injections who were not considered sufficiently advance to warrant knee arthroplasty. Patients had a mean age of 59.4 and a mean BMI of 28.93 (Table 1). We also included one healthy control for internal validation purposes. Inclusion criteria included subjects more than 18 years old with radiographic evidence of osteoarthritis who were dissatisfied with prior attempts at non-surgical management, including non-steroidal anti-inflammatory drugs (NSAIDs), other oral analgesics, physical therapy, knee braces, and nutritional supplements. Patients who were taking NSAIDs were allowed to maintain their regular dose but not to change the dose for the duration of study, given that NSAIDs may affect OA biomarker measurements [17]. If during the study subjects experience knee pain, they were allowed to use Tramadol 50 to100 mg every 6–8 h as needed for pain control. Tramadol is not known to have anti-inflammatory effects that affect OA biomarkers.

Table 1 Patient characteristics and knees injected
Full size table

Patients were excluded if radiographs showed bone-on-bone arthritis, chondrocalcinosis, and physical examination demonstrated any insufficiency of the anterior or posterior ligaments, collateral ligaments, and history of crystalline, inflammatory or, neuropathic arthropathy or infection in the affected extremity. Patients who had intra-articular knee injection with corticosteroids or hyaluronic acid within the previous three months, had allergies or hypersensitivity to eggs, feathers, avian proteins or chicken, pregnant or lactating women, and patients who carried a diagnosis of diabetes mellitus or severe obesity defined as a body mass index more than 40 were also excluded.

Study protocol

This 6-month prospective pilot study was conducted at the Veterans Affair Hospital in Houston Texas. The study was approved by the institutional review board, all participating patients provided informed written consent and the study was performed in accordance to the Declaration of Helsinki.

We obtained weight-bearing anteroposterior, lateral and patellar skyline knee radiographs from all patients. Patients then received an intra-articular hyaluronic acid injection on the affected knee or knees weekly for three consecutive weeks. A total of 16 knees were injected. Patients 1, 4, 6, 7, 8, and 10 received intra-articular injections in both knees, patients 3, 5, and 9 on the left knee and patient 2 on the right knee. Each injection was given with the use of the manufacturer's prefilled 2.5-ml syringe, containing 25 mg of sodium hyaluronate (molecular weight of 620,000–1,170,000 daltons) dissolved in a physiological saline, with a pH of 6.8–7.8 [3]. If the patient had a significant knee effusion, the synovial fluid was aspirated prior to intra-articular injection of hyaluronan. No additional intra-articular injections were allowed during the 6 months study. All injections were performed under sterile conditions with anatomical guidance of an orthopedic physician, rheumatology fellow or an orthopedic-trained physician assistant. After the procedure patients were instructed to refrain from any strenuous activity for 24 h.

Urine for osteoarthritis biomarker CTX-II was collected at baseline, then weekly before each intra-articular injection and then every 2 weeks until completion of 6 months of study. Given the evidence of strongest correlation between radiographic severity, knee osteoarthritis and biomarkers when samples were collected in the first 4 h after getting up from bed, subjects were instructed to collect the urine samples 4–6 h after getting up in the morning [18]. Reminders for urine collection the day prior each urine collection was given to each patient by phone. Patients were instructed to store urine samples at temperatures lower than −5°C at their respective homes [19]. Aliquots of urine were frozen at −80°C within 24 h of collection. At the end of the 6 months of study, all urine samples were thawed and assayed for CTX-II by a specific competitive ELISA using a mouse monoclonal antibody, MAbF46, directed against a linear six amino acid epitope (EKGPDP) of human type-II collagen C-telopeptide (CartiLabs ©ELISA kit). CTX-II levels were then corrected for patients' creatinine and analysed using Kruskall–Wallis statistics for one-way repeated ANOVA, with the aid of Statistica Release 6 from Statsoft, Tulsa, OK, US.

Results

All patients were able to collect their urine samples in the first 4 hours after getting up from bed and store them at a temperature −5°C. After 6 months of urine collection frequent measurements of urinary CTX-II were obtained (Table 2). One urine collection from patient #7 at week 12 was missing. The majority of the patients experienced pain relief after the 3rd intra-articular injection. Only one patient reported no relief from the treatment.

Table 2 Results: urinary CTX-II telopeptide values in ten patients for the duration of 6 months
Full size table

CTX-II values were significantly increased from the 3rd week after injections began. Baseline measurement showed 2,233 ng/mmol ± 1,221; at week three the average level was 6,216 ng/mmol ± 4,428 (p = 0.014). This increase in CTX-II was sustained throughout the entire 6-month follow-up period (Fig. 1). Values were corrected for differences in plasma creatinine. Both analyses showed a statistically significant increase in urinary CTX-II levels that persisted through the 6th month of the study (p = 0.015) (Fig. 1).

Fig. 1
figure 1

Six months study repeated measures ANOVA (using Kruskall–Wallis statistics). Data for CTX-II levels (ng/mmol) corrected for plasma creatinine (mean ± SEM)

Full size image

Discussion

There is a large amount of clinical evidence showing symptomatic improvement with viscosupplementation in patients with knee OA [2]. However, we do not know whether this treatment slows the progression of knee OA. Urinary CTX-II had been shown to be an excellent marker of cartilage breakdown; levels correlate with OA disease severity. Repeated courses of viscosupplementation had been given only when pain from knee OA recurs; the treatment is effective in relieving pain for four to six months. Though viscosupplementation is not considered to be disease modifying in OA, there are some animal studies showing that it may also be “chondroprotective”. Since pain in knee OA does not correlate well with radiographic changes and pain does not arise from the articular cartilage, we postulated that cartilage breakdown may precede pain. Hence, when patients have recurrence of pain, significant amount of cartilage breakdown could have already occurred. We reasoned that if intra-articular HA injections protect cartilage we would find that OA biomarkers decrease after treatment and then, as the product is metabolized, they would begin to increase once more. We postulated that if we knew the kinetics of urinary CTX-II after intra-articular HA injections we could time injections appropriately to forestall further cartilage injury and prevent recurrence of pain. However, contrary to what we had expected, we found a highly significant increase in urinary CTX-II levels 3 weeks after the injection in all ten patients that was sustained throughout the remainder of the study. These observations suggest that intra-articular HA injections accelerate cartilage breakdown in patients with symptomatic knee osteoarthritis.

The main symptom in patients with osteoarthritis that brings them to a physician is pain. It is the stimulation of nociceptor nerve endings located at the joint capsule, synovium, and subchondral bone, not the cartilage, which causes pain in OA. Multiple animal studies support the idea that high molecular weight HA intra-articular injections reduce pain. This result is attributed to the decrease in concentration of inflammatory substances and reduction of transmission of mechanical forces to nociceptor nerve endings by increasing synovial fluid viscosity [20–22]. Even though pain is the main symptom in OA, it is also important to reduce cartilage breakdown. Our study suggests that HA provides temporary relief of pain at the expense of hastening OA progression.

One possible explanation for our findings is that even though the substance injected mainly contains high molecular weight (HMW) HA, it may also contain short HA fragments which displace HMW HA from CD44, the major chondrocyte cell receptor for HA. This may make the extra-cellular matrix unstable and cause chondrolysis. Knudson et al. demonstrated that exposure to hyaluronan hexasaccharides will cause chondrolysis and loss of proteoglycan-rich matrix [23]. Termeer et al. suggested that low molecular weight and not high molecular weight HA fragments stimulate dendritic cell maturation, pro-inflammatory cytokine production and T cell activation [24]. These may increase matrix metalloproteinase production and cartilage breakdown [25].

Another possibility is that insertion of a needle in the joint triggered a response that resulted in sustained matrix degradation. While acute joint injury may cause inflammation sufficient to increase urinary CTX-II levels [26, 27], it is unlikely that this increase in urinary CTX-II would be delayed until the 3rd week after the start of the treatments or that shedding of CTX-II would persist for more than 5 months after the treatments were stopped. If needle injury was the cause of CTX-II release, we would expect levels to increase after the first week not 3 weeks or more after the first intra-articular injection. In addition, if needle injury was the cause of CTX-II increase, we would expect CTX-II to decline after a couple of weeks, but in all patients these measurements remained elevated for 6 months.

Throughout the study period, our subjects had reported minimal changes in their physical activities after having IA HA injections. Even if the subjects were more active after the injections, it would be highly improbable that the observed 2–3-folds rise in urinary CTX-II from weeks 3 to 24 has resulted from increased physical activities. Though serum cartilage oligomeric matrix protein (COMP) has been noted to increase in subjects undergoing strenuous exercises, the rise is lower (estimated to be 1.6 folds in subjects running marathons and ultra-marathons) and short-lived (lasting 2 days) [28]. Urinary CTX-II has not been studied in the same way but Ganero et al. has reported that urinary CTX-II is highly correlated to serum COMP as biomarkers of cartilage degradation [29].

Since majority of the HA preparations are derived from rooster combs, there is concern that avian protein impurities may contain chicken type-II collagen. From the manufacturers’ data, HA preparations like Orthovisc® and Supartz® contain minute amount of protein impurities (less than 2 ppm) [30]. Therefore, only 150 ng of avian protein would have been introduced in each set of three injections given to the patient. Taking an average urinary output of 2 l/day per subject, the excess urinary CTX-II peptide that each subject in our study would have excreted from weeks 3 to 24 compared with baseline amounted to approximately 1 mg. Thus it is impossible that the excess excreted CTX-II observed in the study could have originated from avian protein impurities from the HA preparations

This study had several limitations. Only ten patients were treated and followed. We used only one biomarker (CTX-II); nevertheless this was sufficient to show sustained evidence of matrix degradation.

Studies evaluating the impact of different therapeutic modalities like NSAID’s, calcitonin and bisphosphonates on joint matrix integrity have used CTX-II measurements to assess the ability of these agents to prevent progression of OA [17, 31–34]. However, this is the first study that looks at the impact of intra-articular HA injections on the release of joint matrix collagen peptides. While the mechanisms responsible for this loss of matrix remain obscure, these results suggest that investigations to identify the cause of this outcome are indicated to mitigate this surprising and unwarranted side effect of HA injections.