Introduction
Osteoarthritis (OA) is a chronic condition affecting a large proportion of the older population. It has a considerable impact on quality of life and is the 10th leading cause of non-fatal disease burden worldwide.1 In Australia, it is estimated to affect 1.6 million people.1 To date, no disease-modifying agent has been reliably identified, and pharmacological treatment has largely targeted symptom control. Although various therapeutic interventions have been trialled, paracetamol and non-steroidal antiflammatory drugs (NSAIDs) continue to be the mainstay of treatment. Paracetamol is generally safe at appropriate doses but may not be satisfactory at maintaining pain relief, while the adverse cardiovascular, renal and gastric side effects of NSAIDs limit their use. Furthermore, their benefit is limited to symptomatic control as they do not slow the progression of OA. Clearly, a treatment that both slows the natural history of this disease and provides safe analgesia is needed.
Physiological role of glucosamine
Glucosamine has been long postulated as a safe and effective treatment for OA based on its physiological role in the synthesis of articular cartilage.2,3 Glucosamine sulphate (GS) is a naturally occurring component of aggrecan and other proteoglycans that form part of the connective tissue structures of the synovial joint, particularly the articular cartilage. The sulphate group appended to glucosamine provides the negative charge that allows proteoglycans such as aggrecan to trap water molecules in the cartilage matrix. Thus, they are necessary for the structural integrity and the resilience of articular cartilage. In vitro studies have demonstrated increased synthesis of aggrecan when glucosamine is added,4,5 and anti-inflammatory effects and repair of damaged cartilage have also been observed.6 Results of this research have suggested a promising role for glucosamine, but clinical trials of oral glucosamine for OA, however, have demonstrated conflicting results.7,8 One reason for this may be that in vitro studies have used much higher concentrations (50-5000 [mu]M)9 than those achieved in clinical studies of bioavailability in both humans and animals where concentrations measured are in a relatively tight range of 8-12 [mu]M.6,9,10
There has been a focus on the question of the preparation used as, in general terms, the sulphate salt of glucosamine has been used in the largely positive (European) studies whereas glucosamine hydrochloride has been used in the North American studies, which have been largely negative. When each preparation enters the stomach (normally pH 2.5), both salts dissociate completely and the aminosugar component is identical.9 An effect of sulphate has been postulated, but physicochemical considerations make this explanation improbable.9 This issue is also confused by the makers of GS referring to their product as DONA Glucosamine but has not released precise details. Dosing studies for both salts have shown virtually maximal serum concentrations (Cmax) of 10 [mu]M at 4 h.
Clinical trials
The failure to translate laboratory studies into a clinically useful treatment is not because there have been limited studies in the area. The 2005 Cochrane review identified 25 trials as being of suitable quality for inclusion,7 and independent meta-analyses have also been undertaken (Table 1). Results of these, too, have been discordant. Various editorials have proposed reasons for these discrepancies, including methodological flaws, quality of preparations used, publisher bias and placebo effect.10,21 It therefore remains difficult to gain any kind of consensus on the validity or otherwise of glucosamine as an intervention.
Despite the lack of strong evidence for the clinical efficacy of glucosamine, it is well accepted as a safe and 'natural' therapy by the wider public. In both Australia and North America, glucosamine preparations are registered as dietary supplements rather than a prescription medications, as in the European Union. Not being a prescription medication implies that marketers are held to less stringent standards.6 This is well demonstrated by the fact that there are 532 products listed with the Therapeutic Goods Administration containing glucosamine in Australia. Sales of glucosamine supplements are projected to reach $2.3 billion worldwide in 2013.5,22 Should glucosamine have been registered as a medication on the pharmaceutical benefit scheme (PBS), the estimated cost to the Australian Government was estimated to be $10-30 million in 2006. So should patients be encouraged to buy and use glucosamine supplements, or is this a safe but ultimately ineffectual treatment?
The Osteoarthritis Research Society International (OARSI), which is an international collaboration of basic and clinical researchers in OA, has published guidelines for the treatment of OA. They consider that the currently available research suggests that glucosamine may provide symptomatic relief and thus suggest a 6-month therapeutic trial for the individual patient. Disease-modifying potential is mentioned but the strength of evidence is tenuous.23 Though the first clinical trials of glucosamine were conducted in 1980,2,3 there still does not appear to be a satisfactory conclusion to this vital question.
Outcome measures
One difficulty in drawing conclusions from the available data lies in the number of different outcome measures used in the trials. Several validated scales, the Western Ontario and McMaster University (WOMAC) arthritis scale, Lequesne Index and various visual analogue scales, are utilised in numerous studies and are endorsed by OARSI.24 Some studies, however, also use percentage of reduction in knee pain, analgesic use, dichotomous measures, 'disease flare', radiological joint space narrowing and joint replacement surgery as outcome measures.11-13,16-19 Some of these may be inappropriate both from a design and reporting aspect. For example, joint replacement surgery is influenced by multiple patient-specific factors, including age, comorbidities and attitudes towards surgery of both the patient and the surgeon. There are too many variables for this to be considered a reliable outcome measure for OA treatment.
The Cochrane meta-analysis found a moderate clinical reduction (effect size (ES) = 0.47) in pain compared with placebo when several different methods of pain measurement were pooled (ESs are considered small (0.2), moderate (0.5) or large (0.8)8). This was not, however, supported by randomised controlled trials using the WOMAC pain subscale only.7 There was a moderate reduction in functional outcomes as measured by the Lequesne Index (ES = 0.47) but minimal reduction using total WOMAC scale (ES = 0.18). Joint space narrowing is the major outcome measured by available studies evaluating progression of disease over time. Reginster et al. reported in Lancet in 2001,18 in a randomised control trial over a 3-year period, a reduction in both mean joint space narrowing and joint space width using a digitising method on plain radiographs. This was followed shortly by a second very similar trial17 that, in fact, showed a trend towards increase in joint space width in the glucosamine supplemented group over 3 years. These initial reports have not been validated in follow-up studies and problems have been raised with the initial studies including small numbers (68 and 66 in the treated groups) with high drop-out rates. Plain radiographs are a relatively insensitive method of estimating cartilage volume and joint space narrowing is not tightly correlated with other x-ray changes of OA. Mazzuca and Brandt25 suggest that while this can be a useful measure of disease modification when performed to appropriate international standards, changes in both the patient (body weight and varus-valgus deformity) and radiography (technician variability and equipment) over several years can produce a markedly different result. They point out that a reduction in pain may alter stance and therefore the measurement of joint space narrowing on a plain radiograph. A meta-analysis of radiographic studies concluded no effect of GS supplementation on joint space narrowing.5 Furthermore, studies in hip OA have been similarly negative.23 We await the results of the Australian Long-Term Evaluation of Glucosamine Sulphate (LEGS) study of 400 subjects with mild to moderate knee OA supplemented with GS where cartilage volume will be measured by magnetic resonance imaging.
Independent versus industry funded research
When a sub-analysis was done of trials using the Rotta preparation (and indeed sponsored by the Rotta Research Group), a mild (ES = 0.25), statistically significant reduction in total WOMAC scores was found.7,14,17,18 Conversely, in studies using a non-Rotta glucosamine preparation (and not funded by Rottapharm), the reduction in both total WOMAC scores and individual WOMAC subscales was negligible and not statistically significant.12,13,15,19,26
The divergence in outcomes between Rotta-funded trials and those funded by independent groups is worth noting. Rottapharm is a pharmaceutical company based within the European Union, where glucosamine is a prescription medication. This could therefore suggest a superiority of the Rotta preparations compared with less regulated international formulations. Alternatively, one could argue that several studies included in the non-Rotta sub-analysis have smaller numbers and would have been less well powered to show a large ES compared with their Rotta-funded counterparts. It seems prudent, however, not to disclude the possibility of industry bias in the face of such a significant anomaly.
Recommendations from OARSI based on published literature to the end of 2009 articulate some of the discrepancies identified by the Cochrane meta-analysis. Their analysis showed a moderate ES (0.46) but reiterated the difference between GS (ES = 0.58) and glucosamine hydrochloride (-0.02), even though pharmacokinetically there should be no difference in the two formulations.9 Zhang et al.8 also comment on the heterogeneity of trials using GS and, furthermore, on the possibility of a significant publisher bias. When trials of low quality and one trial with an extremely large ES are excluded, GS group outcomes become more homogenous; the ES, however, is only 0.15, indicating minimal clinical benefit. Most studies struggle to show convincing long-term benefit in pain or function of any treatment.20
Placebo effect
Another aspect of the glucosamine argument is the placebo effect. The Glucosamine/Chrondroitin Arthritis Intervention Trial in 2006 was funded by the National Institute of Health in an attempt to provide rigorous independent evidence for the efficacy and safety of glucosamine.13 Clegg et al.13 tested glucosamine hydrochloride and chondroitin sulphate alone and in combination, celecoxib and placebo. The difference in clinical improvement between the glucosamine and placebo groups was disappointingly small although a sub-analysis showed significant pain relief in patients with moderate to severe pain at baseline who took a combination of glucosamine and chondroitin. Interestingly, though, an abnormally high placebo effect of 60.1% was noted, and the authors point out that the benefit of celecoxib, was smaller than that seen in other trials. This may, in part, obscure some of the actual benefit of the active group. It also raises the unusual idea of placebo as an effective intervention for OA.
A meta-analysis by Zhang et al.,21 looking at studies of a range of therapies, indicates that placebo is, in fact, effective in relieving pain, stiffness and self-reported function in OA. The overall estimate of ES for pain was 0.51 for all trials and 0.71 when rescue medications were not used. In trials using both placebo and no treatment, the ES for pain was 0.77 for placebo and -0.08 for the untreated group. In a disease where the natural history is not one of remission or chance regression to the mean, this represents an important clinical benefit. While effective for pain, placebo was not effective, however, for more objective outcomes.21 The evidence for glucosamine lies primarily in pain outcomes and less so in objective measures; this similarity in these two 'interventions' could be interpreted into two ways. On one hand, the abnormally high placebo response may serve to hide some of the true benefit of glucosamine. Conversely, it may be that some of the apparent 'efficacy' of glucosamine can, in fact, be attributed to placebo.
Adverse effects
One important attribute of glucosamine, though, is its excellent safety profile. Glucosamine has been consistently found to have few side effects and was found to be as safe as placebo in the results of the Cochrane meta-analysis.7 Studies that reported numbers of adverse events and serious adverse events are listed in Table 2. The occurrence of adverse events in these studies is inevitable given the age of the study populations and the duration of some of the studies, but there is a striking similarity in occurrence in the treatment and placebo groups. Concern that glucosamine might affect glycaemic control has not been substantiated.20 The ES for pain reduction in all trials included in the OARSI report8 for NSAIDs was 0.29, compared with 0.58 for glucosamine. This is significant, not just in the larger benefit seen in the glucosamine group, but when serious adverse effects and contraindications to NSAIDs are taken into account. Clearly, an intervention that is at least as effective as the best current therapy, but markedly safer, is desirable. It may also explain the popularity of glucosamine in the wider community, despite the lack of conclusive clinical evidence.9
Cost
The cost of glucosamine to the Australian consumer varies widely. A 100-day supply of the 1500-mg GS tablets can be obtained at the time of writing for A$52. From an online website, a discount chemist may, at times, be up to A$15 cheaper. The PBS subsidised price of Naproxen SR 1000 at the same time is A$70 and Panadol Osteo, the sustained release for of paracetamol commonly prescribed for OA, is A$65. The PBS quoted prices do not take into account discounts for pensioners and others entitled to discounts. Such discounts obviously do not apply for GS.
Conclusion
Hence, the place of glucosamine in the treatment of OA is confusing. While the biological effect of glucosamine has held promise, the failure to translate this into a clinically proven therapy remains disappointing, and it is difficult to determine whether this is due to methodological flaws or a true lack of clinical benefit. The LEGS study, which is a large, 2-year, industry-independent trial due for completion in late 2011, may provide some clear answers to these questions.27
At present, however, in the absence of robust clinical evidence, practitioners are faced with a dilemma. While its role as a 'natural', highly efficacious disease-modifying agent in the treatment of OA is far from established, glucosamine remains a safe treatment that provides some patients with a reasonable level of analgesia. Notably, this pain relief is not accompanied by the serious adverse effect profile of other therapies, such as NSAIDs. Even if for this reason only, it would seem reasonable to suggest a trial of glucosamine (as recommended by OARSI8) to patients with pain and reduced quality of life due to OA.
References