National Institute for Health and Clinical Excellence. CG 59 Review of Osteoarthritis Guideline Review Consultation Comments Table

Transcription

1 National Institute for Health and Clinical Excellence CG 59 Review of Osteoarthritis Guideline Review Consultation Table Type SH SH GDG Member British Society for Rheumatolog y Y N There is no need for an update at this point in time Main grounds: 1) Paracetamol is the first line recommended pharmacological agent in CG59 (after core non-pharmacological interventions). Recent data summarized in the OARSI evidence review (Zhang et al, 2010) has shown that paracetamol is both less effective: The ES for pain relief acetaminophen diminished numerically to 0.14 (95%CI 0.05, 0.22) and was no longer significant when analysis was restricted to high quality trials (ES 0.10, 95% CI 0.0, 0.23) and more toxic than previously suspected: New evidence for increased risks of hospitalisation due to perforation, peptic ulceration and bleeding acetaminophen >3 g/day have been published (HR 1.20, 95% CI 1.03, 1.40). on areas excluded No Decision for arthroplasty surgery & role of arthroscopy might usefully be included if a review is undertaken. on equality No None 1 of 77

2 HFA Healthcare This recommendation needs to change. Crystalline glucosamine sulphate oral powder (Glusartel) is now a licensed agent for knee osteoarthritis in the UK. The positive evidence for this preparation provided the basis for the recommendation of a 3 month trial of glucosamine sulphate in the Appraisal Consultation Document for CG59. However the licensing of Alateris (Glucosamine Hydrochloride) led to the controversial omission of this recommendation from the final guideline. Crystalline glucosamine sulphate oral powder has higher bioavailability than other preparations, which may explain the observed differences in the published studies. SIGN is currently considering a new health economic model, which shows that this preparation will save money against current prescribed preparations and is cost effective an ICER of 18k or less per QALY. The current clinical guideline (CG59) is out of date in relation to the availability and use of crystalline glucosamine sulphate oral powder in clinical practice. This product was approved and classified as a medicinal product only after the publication of CG59 and not considered during the development of the guideline. In the absence of any specific NICE guidance on this on areas excluded Yes - the prescribing of unlicensed glucosamine The scope is appropriate but in this the current clinical guidelines require updating in relation to glucosamine in order to reflect current clinical on equality As stated above the vast majority of PCT s are using the current clinical guideline to actively encourage patients to purchase unlicensed preparations directly 2 of 77

3 licensed preparation, healthcare professionals are applying the current guideline recommendations to all glucosamines irrespective of licence status and patient treatment quality is being compromised. In addition to updating the guideline to review crystalline glucosamine sulphate oral powder we would also highlight that the guideline requires updating in order to: 1. Correctly classify licensed and unlicensed glucosamines. on areas excluded practice and the availability of the first licensed crystalline glucosamine sulphate proven efficacy and safety on equality over the counter. These patients in addition to having to pay for their treatment are being denied access to the only licensed crystalline glucosamine sulphate 1500mg o.d. oral powder proven efficacy and safety. 2. Improve patient treatment quality and ensure patients have access to effective and safe licensed treatments. 3. Provide national prescribing guidance to health care professionals. 4. Clarify the clinical data to avoid misinterpretation Further information on each of these points is provided below 3 of 77

4 on areas excluded on equality 1. Availability of licensed crystalline glucosamine sulphate As stated above the current OA guideline was developed and published in 2008 prior to the availability of licensed crystalline glucosamine sulphate and as consequence the recommendations are based on unlicensed products which do not have any proven efficacy compared placebo. Crystalline glucosamine sulphate 1500mg o.d. oral powder is approved by the UK regulatory authorities (MHRA) as a medicinal product and was subsequently made available under prescription under the trade name Glusartel in July Glusartel has proven efficacy compared placebo. 2. Incorrect classification of all glucosamines as nutraceuticals. The current guideline classifies glucosamines under the heading of nutraceuticals. This is now incorrect and does not reflect licensed crystalline glucosamine sulphate 1500mg oral powder which is classified as a medicinal product and is available by prescription only. 3. Improving patient treatment quality. 4 of 77

5 Based on the recommendations of the current clinical guideline an increasing number of PCT s are recommending the complete discontinuation of all glucosamine prescribing and endorsing the purchase of unlicensed glucosamine by patients over the counter. on areas excluded on equality With the guideline having not considered the licensed preparation this preparation is being incorrectly grouped the unlicensed preparations and as a consequence patients are being denied access to a safe and effective treatment. In addition having undergone a thorough regulatory review the quality of licensed glucosamine can be assured; the same cannot be said of the unlicensed preparations which can put patients at unnecessary risk regard to the quality of the preparation they receive. Both the NHS white paper of July 2010 and the QIPP agenda have a focus on quality. Guidance given by NHS white paper in July 2010 The July 2010 NHS White Paper by the Coalition Government places quality and productivity at the heart of the NHS. 5 of 77

6 Consecutive governments have committed to improving patient safety and care by setting out a vision for an NHS quality at the heart of everything it does. This recent White Paper recognizes the importance of and commits to building on Lord Darzi s work in putting a stronger emphasis on quality. on areas excluded on equality QIPP agenda The productivity challenge is to improve quality out the need for additional resource. Prescribing of unlicensed (i.e. over the counter OTC) glucosamine is a real quality issue. Over 1 million prescriptions were issued in 2009* for glucosamine products whose quality has not been approved by the UK regulatory authorities (MHRA). The MHRA uses a specialist quality team to audit and approve the manufacturing facilities to ensure adherence to its own MHRA internal quality standards and processes therefore ensuring patients get what is actually stated on the pack. This is not a requirement for unlicensed (i.e. OTC) glucosamines in the UK. In 2009, the NHS in England alone spent 25.5M* on unlicensed (i.e. OTC) glucosamine. 6 of 77

7 81% of this NHS spend came from 5 products* whose reimbursement ranged from to per months treatment*. For all patients receiving treatment an unlicensed product quality cannot be guaranteed. on areas excluded on equality The quality solution would be to replace all unlicensed (i.e. OTC) glucosamine a licensed prescription only medicinal glucosamine product. Patient treatment quality would be immediately raised. In addition patients would receive a treatment proven efficacy. Productivity would also be raised out the need for additional resource*. If this quality solution was implemented in England it could result in potential annual savings to the NHS of 10.4M* based on 2009 data and using crystalline glucosamine sulphate 1500mg o.d. oral powder as the licensed prescription only glucosamine. Both NHS productivity and quality would be raised if this quality solution was implemented. * (Data source: NHS prescription services for England; accessed march 2010; all data relate to the calendar year 2009 for England only; cost data are based on that stated in Drug Tariff Part II Clause 8 and include all FP10 prescribing 7 of 77

8 apart from items prescribed in hospitals). 10.4M saving based on replacing top 5 unlicensed glucosamine product usage in 2009 Glusartel. Comment on lack of additional resource is based on a reduction in direct drug costs when moving from unlicensed to licensed prescribing. on areas excluded on equality PCT guidelines Currently, a vast majority of PCT s are using the current clinical guideline to stop clinicians from prescribing (unlicensed and licensed) glucosamines and patients who require such treatment are required to purchase unlicensed preparations directly over the counter. These patients in addition to having to pay for their treatment are actively being denied access to the only licensed and effective crystalline glucosamine sulphate. 4. Request by health care professionals Over the last 12 months, HFA Healthcare have been asked on numerous occasions when a recognised body such as NICE will reassess the glucosamine market and issue fresh guidance and especially where it sees the licensed crystalline glucosamine sulphate 1500mg o.d. 8 of 77

9 oral powder being prescribed. on areas excluded on equality In the absence of national guidance and in order to assist the NHS in developing prescribing guidance HFA have supported concerned healthcare professionals through the development of a consensus statement for glucosamine prescribing. This is summarised below and we would recommend that an update of the current NICE OA guideline should consider this statement. Consensus Statement on the appropriate prescribing of GLUSARTEL (crystalline glucosamine sulphate 1500mg oral powder) for patients Knee Osteoarthritis. The Panel will be made public once the Consensus statement below is submitted for publication Background: This consensus statement was developed to provide advice to prescribers and medicine management teams on the effective and cost-effective use of glucosamine sulphate on the NHS. There is inconsistent advice for NHS professionals 9 of 77

10 some PCOs recommending that when prescribed, only the sulphate form should be used whilst others advise against the prescribing of all forms of glucosamine. There is good evidence for the effectiveness and safety of a brand of glucosamine sulphate made by Rottapharm in a recent Cochrane review 1.The current cost of an NHS prescription of the five most widely prescribed unlicensed glucosamine products varies between to per month 2 and Glusartel [the UK licensed Rottapharm product] costs per month. Most of the current prescriptions for glucosamine appear to be for unlicensed preparations. on areas excluded on equality Methods: A panel of NHS health professionals from various relevant backgrounds was convened. A review of the published data and an economic analysis by BresMed Health Solutions 3 was presented. After discussion and debate the panel arrived at a consensus statement that was considered relevant to guide NHS prescribing of glucosamine sulphate. Results: The 2009 Cochrane review provides good evidence for the effectiveness 10 of 77

11 and safety of Glusartel (as the UK licensed Rottapharm product). Follow-up of the longterm studies suggests that the use of Glusartel may reduce the need for knee replacements 4. Clinical evidence only supports the use of crystalline glucosamine sulphate 1500mg once a day oral powder (Glusartel ): Other glucosamine salts, doses or formulations have a lower total bioavailability (by 75%) and peak plasma levels (by 50%) and do not allow the achievement of effective concentrations 5. The addition of chondroitin further decreases plasma concentrations. on areas excluded on equality NICE: Although CG59 Osteoarthritis (February 2008) did not support the use of glucosamine preparations, this was based on the fact that at the time of the guideline development, only glucosamine hydrochloride was licensed. This guidance has been superseded by the availability of a licensed glucosamine sulphate (Glusartel ). Drafts of CG59 did support glucosamine sulphate as being effective but removed the recommendation to provide it as an option before ratification as the only licensed product was the hydrochloride salt which had 11 of 77

12 been shown to be ineffective in an RCT. on areas excluded on equality Health economic analysis: data submitted to SMC 6 (April 2011) shows Glusartel is cost effective a cost per QALY of 12,402, well below the threshold of 20,000 per QALY. As the cost of glucosamine products currently supplied on the NHS is greater than Glusartel, switching these prescriptions to Glusartel will save the NHS money and provide patients a licensed, evidence-based product. Conclusions: For patients knee osteoarthritis, Glusartel is recommended as a safe, effective and cost-effective step 2 option [along paracetamol and topical NSAIDs] and should be included in local formularies and prescribing guidelines. A 3- month trial is suggested to assess response. Switching to this product will save money in these financially difficult times. References: 1. Towheed T, Maxwell L, Anastassiades TP, Shea B, Houpt JB,Welch V, HochbergMC,Wells GA. The Cochrane Library, Chichester, UK: John Wiley 12 of 77

13 & Sons 2. Data source: NHS prescription services for England; accessed march 2010; all data relate to the calendar year 2009 for England only; cost data are based on that stated in Drug Tariff Part II Clause 8 and include all FP10 prescribing apart from items prescribed in hospitals 3. BresMed Health Solutions is an independent health economics consultancy 4. Pavelka K et al, Arch Intern Med Persiani S Osteoarthritis Cart This new data was presented to the panel by BresMed Health Solutions and represented a bespoke pharmacoeconomic model for Scotland and specifically for SMC review. on areas excluded on equality 5. Clarification of the GDG s interpretation of the Cochrane Review, the OARSI guidelines and further evidence. Firstly, we would like to comment specifically on page 11 of the Review Consultation Document; the Cochrane Review 2009 (Review 13 of 77

14 Consultation Document Reference no. 153) and the 2009/2010 update of the OARSI guidelines (Review Consultation Document Reference no.155). In addition, we will also comment on the rest of the evidence that has been recently generated on the use of glucosamine in OA, including arguments and articles that have not been considered in the NICE Review Consultation Document. on areas excluded on equality We have provided below an executive summary followed by a more comprehensive analysis of each specific issue in Addendum 1 (see the last row of this proforma). Executive Summary Cochrane Review 2009 The conclusion in the NICE Review Consultation Document relating to the Cochrane Review is misleading and contrary to what the authors stated in their conclusion, that reads: Pooled results from studies using a non-rotta preparation or adequate allocation concealment failed to show benefit in pain and WOMAC function while those studies evaluating the Rotta 14 of 77

15 preparation showed that glucosamine was superior to placebo in the treatment of pain and functional impairment resulting from symptomatic OA. on areas excluded on equality In summary, trials of glucosamine adequate allocation concealment showed important benefit on the validated algo-functional index of Lequesne, but altogether (i.e. including glucosamine hydrochloride and uncontrolled formulations of glucosamine sulphate) actually failed to show significance on the WOMAC pain and functions subscales. However, when high quality, pivotal trials of glucosamine sulphate performed the Rottapharm preparation (namely Glusartel 1500 mg o.d. oral powder, that was made available in the UK following the publication of CG59 under the tradename Glusartel ) and adequate allocation concealment were separately analyzed (including the new study by Herrero-Beaumont 2007, Review Consultation Document reference no. 107), the effect sizes on both WOMAC pain and function are significant and clinically relevant. In addition to the Cochrane Review, this is also well described in another meta-analysis 15 of 77

16 (Reginster JY, Arthritis Rheum 2007; 56: ) that has not been considered in the NICE Review Consultation Document. on areas excluded on equality OARSI Guidelines The conclusion from the NICE Review Consultation Document about the OARSI guidelines is also misleading in quoting only a sentence from the abstract of the 2009 update of the guidelines and disregarding the full document. In summary, current OARSI guidelines recommend the use of glucosamine sulphate (but not glucosamine hydrochloride) in knee OA. When the OARSI analysis of trials, performed over a 27 years period, was restricted to high quality trials performed in the last decade, the effects of glucosamine sulphate are at least comparable and indirectly superior to paracetamol, out heterogeneity or evidence of publication bias. Efficacy increases when only homogeneous, high quality trials performed Glusartel would be considered (as also witnessed by the results of the recent pivotal trial by Herrero-Beaumont 2007, Review Consultation Document reference no.107, comparing 1500 mg o.d. oral powder 16 of 77

17 glucosamine sulphate placebo and paracetamol). on areas excluded on equality Other Evidence The vast majority of systematic reviews and meta-analyses indicate that glucosamine sulphate is effective and safe in OA, including the UK Health Technology Assessment by Black et al (Review Consultation Document reference no. 101) that highlights the clinical relevance of the symptom results of the longterm trials performed Glusartel. This should be interpreted also in view of a publication that has not been considered in the article and in the NICE Review Consultation Document (Bruyere, et al. Arthritis Rheum 2006; 54: ), in which the authors of the two long-term trials show the significant impact of glucosamine sulphate on clinically relevant parameters such as the Minimum Clinically Important Improvement (MCII) and the Patient Acceptable Symptom State (PASS). The only exception is represented by the metaanalysis by Wandel et al. published on the BMJ in late 2010 (Review Consultation Document reference no. 154). However, this study has been severely criticized for important 17 of 77

18 methodological (BMJ 2010; 341:C6328- C6338-C6335), forcing the BMJ Editors to post a note stating not to consider the authors conclusions respect to cost coverage and prescription of glucosamine products because the conclusions were not supported by the data. ( reply#bmj_el_247719). on areas excluded on equality The cost-effectiveness analysis performed by Black et al is also favourable for glucosamine sulphate 1500 mg o.d. This is was also supported by another study not considered in the NICE Review Consultation Document, i.e. the 8-year follow-up data of the two Glusartel long-term trials, that showed significantly fewer knee arthroplasties in the former glucosamine sulphate group than in the placebo group, a significant RR = 0.47 (Bruyere et al, Osteoarthritis Cartilage 2008; 16:254-60). Indeed, Black and colleagues found, in agreement the Cochrane Review, a significant effect of glucosamine sulphate on radiological joint space width. Together the 8-year follow-up data, this is a remarkable finding that, although not included in the current indications of Glusartel, cannot be ignored when assessing the potential of the drug. 18 of 77

19 on areas excluded on equality Another new and not considered costeffectiveness analysis is that published by Scholtissen and colleagues (Int J Clin Pract 2010; 64: ) based on the 6-month study of Herrero-Beaumont 2007 (Review Consultation Document reference no. 107). Scholtissen et al. found that glucosamine sulphate was highly cost-effective when compared to both paracetamol and placebo. Finally, new pharmacokinetics and pharmacodynamics data support the biological plausibility of the use of this formulation of glucosamine sulphate, as recently reviewed in another study that was not considered in the Review Consultation Document (Altman RD, Expert Rev Clin Pharmacol 2009; 2:359-71). Further details on which the above executive summary has been derived are provided in Addendum 1. Glucosamine sulphate and NICE guidelines The final draft of CG59 did support glucosamine sulphate as being effective but removed the 19 of 77

20 recommendation to provide it as an option before ratification as the only licensed product was the hydrochloride salt and this had been shown to be ineffective in a RCT. This was also evident from the excerpt of a recent presentation at the 2008 ACR meeting by Prof. Philip Conaghan, Chairman of the NICE OA guideline committee, and when commenting on the different aspects of the NICE guidelines stated: on areas excluded on equality For glucosamine it has been complex. I think it is fair to say that it looks like glucosamine sulphate 1500 mg-a-day is the only product, the Rottapharm product, showing that it has got benefit in trials. In the UK that product is not licensed. So we were not able to recommend this treatment, but we were able to say that if you want to use glucosamine we advice a trial of the glucosamine sulphate 1500 mg. * (accessible at: / a_Rheumatic_Disease_Update_What_New. mp4. * This statement does not represent a formal or any endorsement from Prof Conaghan towards 20 of 77

21 this stakeholder document but highlights views from the previous Chairman of the NICE OA guideline committee that have been made in public and made readily available in the public domain by the ACR on areas excluded on equality Since Glusartel is now licensed and available in UK, there is no reason not to recommend the product based on the further evidence presented. ADDENDUM 1 Detailed Review of the Evidence Cochrane Review 2009 The conclusion in the Review Proposal Consultation Document (page 11) relating to the Cochrane Review is misleading and contrary to what the authors stated in their conclusion. The updated Cochrane review (last assessed as up to date Nov 2008 and published in 2009) considered 25 studies 4963 patients from which the authors concluded: Pooled results from studies using a non-rotta preparation or adequate allocation concealment 21 of 77

22 failed to show benefit in pain and WOMAC function while those studies evaluating the Rotta preparation showed that glucosamine was superior to placebo in the treatment of pain and functional impairment resulting from symptomatic OA. on areas excluded on equality Important point: The Rotta preparation that the author is referring to is the very same patented crystalline glucosamine sulphate preparation originally developed by the company Rottapharm and in particular the 1500mg o.d. oral powder that was made available in the UK following the publication of CG59 under the tradename Glusartel. Further information on why the statement in the review consultation document is misleading is provided below: 1) WOMAC score First of all, while trials adequate allocation concealment (but performed any glucosamine formulation regardless of quality) did not show benefit for WOMAC pain and function subscales (they did show significant efficacy on the WOMAC Total index: see Analysis 3.7 in the Cochrane 22 of 77

23 Review), the medication was better than placebo when these trials used the Lequesne index, which is another validated outcome for osteoarthritis symptoms, as acknowledged in the European Medicines Agency guidelines ((CHMP/EWP/784/97), representing a combined pain and function measure. on areas excluded on equality The effect size, represented by the standardized mean difference, is 0.54 and thus of moderate size, which is not common for symptomatic treatments in OA, that usually have much smaller effect sizes compared placebo. Most importantly, inspection of Analysis 3.2 in the Cochrane Review where these data are reported, points out that all 4 trials considered for this analysis were performed the Rotta preparation, 3 of which the 1500 mg o.d. oral powder Glusartel formulation (the fourth one, Noack 1994, was performed a tablet formulation of the same Rotta crystalline glucosamine sulphate preparation). All of these trials are high quality studies, meeting all quality assessment parameters described in Table 1 23 of 77

24 of the Cochrane Review in addition to adequate allocation concealment, and being appropriately sized (3 out of 4 studies >100 patients/arm and one >75 patients/arm). The fact that this analysis of trials adequate allocation concealment could include only trials performed the Rotta preparation (including Glusartel), introduces the second, very important aspect. on areas excluded on equality 2) The Review Consultation Document actually fails to acknowledge the second conclusion by the authors of the Cochrane Review, i.e. that those studies evaluating the Rotta preparation showed that glucosamine given as crystalline glucosamine sulphate was superior to placebo in the treatment of pain and functional impairment resulting from symptomatic OA, as assessed by the WOMAC pain and function scores. Indeed, the observation that studies using a non-rotta preparation or adequate allocation concealment fail to show benefit in pain and WOMAC function are not mutually exclusive and studies adequate allocation concealment and that used the Rotta 24 of 77

25 preparation indeed showed efficacy in WOMAC pain and function. on areas excluded on equality This is clearly shown by looking at Analysis 3.4 on the WOMAC Pain Subscale, where no benefit of glucosamine vs. placebo could be described when all studies adequate allocation concealment are pooled. It appears that this analysis pools together studies the Rotta preparation and studies other, various and different preparations. When the 3 studies in this analysis that meet both criteria of having adequate allocation concealment and being performed the Rotta preparation (Herrero-Beaumont 2007, Pavelka 2002 and Reginster 2001) are separately analyzed, the benefit is significant. This is shown in Analysis 4.4, in the context of the assessment of the Rotta preparation only, where it is also evident that there is absence of heterogeneity (I 2 = 0.0%). All these 3 trials are high-quality trials, appropriately sized (all>100 patients/arm) and, very importantly, of long-term duration: this is most valuable in trials of OA that very often study symptomatic agents for short 25 of 77

26 treatment periods only: the Herrero- Beaumont 2007 study (Review Consultation Document reference no.107) treats patients for 6 months, while the Reginster 2001 and the Pavelka 2002 studies are even for 3 years. The latter two studies were already considered in the 2008 guidelines and are published in recognised journals (Reginster et al, Lancet 2001; 357:251-6, and Pavelka et al, Arch Intern Med 2002; 162: ). on areas excluded on equality These 3 trials were performed the Glusartel 1500 mg o.d. oral powder formulation and indeed are identified as the 3 pivotal trials for efficacy and safety in the Glusartel European regulatory marketing authorization application. They unequivocally show efficacy on knee OA pain as assessed by the WOMAC subscale. One may argue that the effect size shown in Analysis 4.4 of the Cochrane Review (standardized mean difference 0.17 [0.01, 0.32]) is statistically significant, but below the threshold for clinical relevance (which is usually set at 0.20). This conclusion would not be correct, as it is biased by a common drawback in Cochrane Reviews. In fact, 26 of 77

27 since they have frequently to deal low quality trials in publications that do not report all trial data, Cochrane Reviews have to rely on absolute final values of the chosen parameters that can only approximate the true effect size when calculating the difference baseline values, if an even tiny imbalance between groups is present at baseline. Conversely, the gold standard method by Hedges (Hedges LV and Olkin L. Statistical methods for meta-analysis. Orlando (FL): Academic Press;1985) prescribes that the effect size should be calculated based on the change from baseline, which is a more precise measure. Since these are high quality trials, the change from baseline was actually the primary endpoint in the 3 pivotal studies. While, curiously, the Cochrane Review artificially used the change from baseline to approximate the final absolute values and thus derive a wrong effect size, the appropriate methodology was used in another meta-analysis that considered the 3 pivotal trials of Glusartel (Reginster JY, Arthritis Rheum 2007; 56: ). The correct effect size on WOMAC pain was calculated to be 0.27 [0.12, 0.43]. on areas excluded on equality 27 of 77

28 While clinically relevant, this might be regarded as a small effect size, but it is comparable to the effect size of common symptomatic medications in OA (e.g. NSAIDs, as described e.g. in the OARSI guidelines). In addition, while symptomatic medications are more commonly studied over short-term treatment periods 12 weeks (and they have seldom been shown to be effective long-term), the 3 Glusartel pivotal trials are long-term (6 months to 3 years). on areas excluded on equality The same discussion applies to the WOMAC function assessment. While Analysis 3.6 performed on all studies adequate allocation concealment does not show benefit vs. placebo, this is significant when the 3 high quality trials of Glusartel included in the Analysis, are assessed separately in Analysis 4.6: the correct effect size calculated by Reginster (Arthritis Rheum 2007;56: ) is 0.33 [0.17, 0.48], again in the complete absence of heterogeneity. The meta-analysis by Reginster (Reginster JY, Arthritis Rheum 2007; 56: ) is also interesting because it highlights major 28 of 77

29 quality defects in glucosamine studies that are reported as having adequate allocation concealment in the Cochrane Review, but have many other problems in their design: these include the studies by Cibere 2004, Hughes 2002 and McAlindon 2004, although the latter was performed mainly a formulation of glucosamine hydrochloride. In this respect, other 2 studies are included in the Cochrane Review as having adequate allocation concealment, but have been performed glucosamine hydrochloride (Houpt 1999 and Clegg 2006): it is important e that glucosamine hydrochloride is not considered effective in any scientific document and e.g. in the OARSI guidelines or in the 2009 Black et al HTA document (see further on). The other two trials adequate allocation concealment are the Rozendaal 2008 study performed an unknown quality glucosamine sulphate preparation in hip OA (and that therefore cannot be pooled all the others, that are knee OA studies) and the Zenk 2002 study, that is an extremely small study of 23 patients. on areas excluded on equality There is little doubt that the careful assessment 29 of 77

30 of the Cochrane Review 2009 outlines a significant benefit of the Rotta preparation of glucosamine sulphate and especially of the Glusartel 1500 mg o.d. oral powder formulation, that emerges as unique among the many uncontrolled glucosamine products. on areas excluded on equality OARSI Guidelines The conclusion from the NICE Review Consultation Document about the OARSI guidelines is misleading too in quoting only a sentence from the abstract of the 2009 update of the guidelines and disregarding the full document. First of all, it is wrong to combine the evidence and conclusions from glucosamine sulphate and chondroitin sulphate clinical trials. In fact, chondroitin the evidence of pain relief detected in the early studies is biased by high heterogeneity and evidence of publication bias: when the analysis is restricted to recent and high quality trials, efficacy indeed disappears and there is no heterogeneity. The case of glucosamine sulphate is completely different as described in the article text and summarised below. 30 of 77

31 While it is correct that cumulative meta-analysis of studies from 1981 to 2008 (a 27 years period) shows a progressive diminution of the effect size and evidence of publication bias, this is largely explained by the authors the fact that a number of trials were published before adoption of the CONSORT reporting guidelines in When analysis is restricted to trials published after 1998, the outcomes are homogeneous and there is no evidence of publication bias: efficacy is indeed diminished (to a nevertheless significant effect size of 0.13 [0.02, 0.25] ), but it increases again to 0.29 [0.003, 0.57] and out publication bias, when the analysis is limited to high quality trials (text and Figure 4 in the OARSI guideline update), Unfortunately, this latter approach heterogeneity remains considerable when all high quality trials are considered. But heterogeneity in meta-analysis is due to major differences in the design of the single trials that, in the case of glucosamine sulphate in OA is due to major hidden quality in several purported high quality studies (as critically reviewed by Reginster, Arthritis Rheum 2007; 56: ), trial duration (it is hard to compare short-term trials of few weeks long-term trials up to 3 years) and, especially, the use of different glucosamine on areas excluded on equality 31 of 77

32 sulphate preparations administered variable dose regimens resulting in a different pharmacokinetic pattern (see below). We have already described above that when the analysis is restricted to the 3 pivotal high quality trials of the Rotta preparation, namely the 1500 mg o.d. oral powder formulation, the effect size is 0.27 [0.12, 0.43] on WOMAC pain and 0.33 [0.17, 0.48] on WOMAC function, for long-term treatment durations of 6 months to 3 years and in the complete absence of heterogeneity. on areas excluded on equality The authors of the OARSI guidelines suggest another, more conservative approach to eliminate heterogeneity simply excluding the one trial exceptionally large effect. With this approach (that would maintain all other high quality glucosamine sulphate trials, regardless of the study design described by Reginster 2007, treatment duration and use of different preparations at varying dose regimens), the effect size on pain relief is indeed decreased, but it is still significant, i.e [0.03, 0.27] and still superior to that described in the OARSI guidelines themselves for paracetamol, that is 0.14 (0.005, 0.23) in all trials and a non-significant 0.10 (-0.03, 0.23) in high quality trials: it is important e that the 32 of 77

33 current NICE guideline does recommend offering regular dosing of paracetamol in addition to core treatment and ahead of oral NSAIDs or COX-2 inhibitors, and that the NICE Review Consultation Document suggests not to change this recommendation, despite new suggestions for a risk of toxicity. Even the most conservative approach, glucosamine sulphate appears to be at least as effective (as also evidenced by the Herrero-Beaumont trial (Review Consultation Document reference no. 107) than paracetamol, at least similar safety. on areas excluded on equality In conclusion, the OARSI guidelines do recommend using glucosamine sulphate for symptomatic relief in knee OA and, potentially, for structure modification (Zhang W et al, Osteoarthritis Cartilage 2008; 16: ). It is unlikely that the 2009 update of the evidence (Zhang 2010, Review Consultation Document reference no. 155) will change this recommendation based on the thorough analysis performed by OARSI and summarized above that strongly supports the efficacy of the drug. Beside the Cochrane Review and the OARSI 33 of 77

34 guidelines, there is additional evidence that has not been appropriately considered in the NICE Review Consultation Document, including evidence that has not been considered at all. This is addressed in the next three sections. on areas excluded on equality Other new studies, systematic reviews and meta-analyses As discussed above, both the OARSI guidelines and the Cochrane Review strongly support the use of glucosamine sulphate in knee OA, contrary to the misleading view offered in the Review Consultation Document. This is especially true for the Glusartel preparation, as clearly highlighted by the Cochrane Review. The NICE Review Consultation Document contends that other systematic reviews found no beneficial effect. This is incorrect as the review by Lee et al (Review Consultation Document reference no. 113) also favoured glucosamine sulphate. We would also highlight that the meta-analysis published on the BMJ by Wandel et al (Review Consultation Document reference no. 154) which stated there was no clinically relevant benefit of glucosamine (and chondroitin) vs. 34 of 77

35 placebo on OA pain, despite statistical significance, has been strongly criticized by OA experts, including members of the OARSI guidelines committee. The most important of these comments have been published as BMJ 2010; 341:C6328-C6338-C6335 (full text of these and all comments available online at eply#bmj ). on areas excluded on equality Basically, the validity of the study has been questioned because of inappropriate trial selection (mixing studies performed glucosamine hydrochloride and glucosamine sulphate of various origin and even confusing the medications actually used in individual trials; pooling studies ranging from 1 month to 3 years in duration and regardless of the joint involved), resulting in high heterogeneity (I 2 = 63% by standard methodology), not appropriately controlled by a complex and not correctly applied Bayesian methodology. In addition, clinical relevance was derived from artificial back transformation of the effect size into a 10 cm VAS, surprisingly raising the bar for clinical relevance to effect sizes seldom achieved by pharmacological treatments of symptomatic OA. 35 of 77

36 Such criticisms were not appropriately addressed by the authors (BMJ 2010, 341: C6340). This obliged one of the journal senior research editor to post a report from BMJ Post Publication Review Meetings ( reply#bmj_el_247719, posted on January 10 th, 2011), whose members acknowledged the criticisms raised by experts and stated that the following conclusions in the article: on areas excluded on equality *at the end of the discussion section: Coverage of costs by health authorities or health insurers for these preparations and novel prescriptions to patients who have not received other treatments should be discouraged. *in the abstract: Health authorities and health insurers should not cover the costs of these preparations, and new prescriptions to patients who have not received treatment should be discouraged were not directly supported by the data. Such actions are seldom seen in Journals like the BMJ and should be carefully considered when referring to the paper of Wandel and 36 of 77

37 colleagues. on areas excluded on equality Black et al (Health Technol Assess 2009) This HTA publication does not appear to be appropriately considered in the Review Consultation Document. This report was published before the prescription only medicine glucosamine sulphate product developed by Rotta was available in the UK (page 15 of the HTA publication) as Glusartel. Deep analysis of this large report allows to conclude that while the results of the systematic reviews analyzed were inconsistent, there was a clear evidence of significant results for trials that used the Rotta preparation of glucosamine sulphate (namely Glusartel) in agreement the results of the Cochrane Review. A summary of the results is provided below: Since the main aim of this HTA was to establish the effectiveness of glucosamine (and chondroitin) in slowing or arresting progression of knee OA, the authors went on to separately analyze trials of at least 12 months follow-up, 37 of 77

38 identifying 3 trials of glucosamine, one of which performed glucosamine hydrochloride (showing no clinical benefit) and two the glucosamine sulphate Rotta preparation (Reginster 2001 and Pavelka 2002): the results are summarized on page of the report. They found that both trials outlined statistically significant results in favour of Glusartel for almost all pain and function parameters. While Black and colleagues suggest that clinical significance was uncertain, they found a clinically relevant effect size of 0.25 (0.05, 0.46) on WOMAC pain and 0.37 (0.15, 0.59) on WOMAC function (Figures 4 and 5 in the report), despite inclusion of the negative trial using glucosamine hydrochloride. While acknowledging effectiveness of the Rotta preparation, they concluded that changes were less than suggested by Tubach and colleagues, but they probably overlooked the article by Bruyere et al (Arthritis Rheum 2006; 54: ) in which the authors of the two long-term trials show the significant impact of glucosamine sulphate on the clinically relevant parameters set by Tubach and colleagues, namely the Minimum Clinically Important Improvement (MCII) and the Patient Acceptable Symptom State (PASS), in the two studies. on areas excluded on equality 38 of 77

39 on areas excluded on equality Importantly, and although this is not approved as an indication for Glusartel given the peculiarity of this assessment and the ongoing scientific discussion, Black and colleagues found, in agreement the Cochrane Review, a significant effect of glucosamine sulphate on radiological joint space width, an effect size of 0.26 (0.17, 0.35). In this respect, they stated that clinical significance was supported by the 8- year follow-up data that showed significantly fewer knee arthroplasties in the glucosamine sulphate group than in the placebo group, a significant RR = 0.47 (Bruyere et al, Osteoarthritis Cartilage 2008; 16:254-60). This is a remarkable finding that, although not included in the current indications of Glusartel, cannot be ignored when assessing the potential of the drug. Black et al. also performed their costeffectiveness assessment only on glucosamine sulphate, namely on Glusartel since the data were generated this preparation. Their data resulted in a cost per QALY estimate of approximately 21,335, i.e. a moderate magnitude of potential gains in relation to the additional costs. This appears to be particularly 39 of 77

40 relevant in the context of the sound analysis performed, over a lifetime horizon. At a cost per QALY gained threshold of 20,000, the likelihood that glucosamine sulphate is more cost-effective than current care is 0.43, while at a threshold of 30,000 the probability is on areas excluded on equality Another new cost-effectiveness analysis, not considered in the NICE Review Consultation Document, was in the meantime published by Scholtissen and colleagues (Int J Clin Pract 2010; 64: ) based on the 6-month study of Herrero-Beaumont 2007 (Review Consultation Document reference no. 107). They found that on a shorter, 6-month time horizon, glucosamine sulphate was highly costeffective when compared to both paracetamol and placebo. Biological relevance and pharmacokinetics. Black and colleagues (Review Consultation Document reference no. 101) concluded that several biological studies support the potential clinical impact of glucosamine sulphate, while the evidence is less consistent for chondroitin and even absent for glucosamine hydrochloride. The latter seem also to have a different 40 of 77

41 pharmacokinetic pattern compared to glucosamine sulphate, as recently reported by Altman (Expert Rev Clin Pharmacol 2009; 2:359-71). It appears that after administration of glucosamine sulphate 1500 mg o.d. oral powder, glucosamine is well bioavailable in blood and knee joint synovial fluid (as previously shown by Persiani et al, Osteoarthritis Cartilage 2005; 13: and 2007;15:764-72). Altman reports that change of glucosamine salt (sulphate to hydrochloride) and/or dose regimen (o.d. to t.i.d., to get the 1500 mg daily dose) and/or formulation (oral powder to other solid formulations) decrease glucosamine bioavailability by 75% and peak plasma concentrations by 50%, to levels that are probably not effective based on PK/in vitro PD correlation reported in the article. These data confer strong biological plausibility to support the clinical findings of the use of Glusartel as the most relevant glucosamine sulphate product. on areas excluded on equality SH British Association for Surgery of the Knee N We are strongly in favour of the that the review of this guideline be undertaken, but feel this needs to be done urgently. We are strongly of the view that the surgical treatment of osteoarthritis, which was an important but largely omitted aspect of the original terms of The surgical practitioners have not been afforded proper and equal representation in the 41 of 77

42 on areas excluded reference, should be a fundamental and important aspect of any review of this guideline. The original guideline failed to address many of the areas of surgical intervention which can allow patients osteoarthritis to receive conservative management which falls short of the need for such major surgery as arthroplasty. We applaud the belated recognition that the original guideline did not adequately address the surgical of osteoarthritis treatment but feel the problems created by this have been so significant that the review now needs to be accelerated substantially from the normal NICE timetable. on equality development of previous guidelines on this subject and this has in turn resulted in the disadvantaging of the patients who would be best treated by suitable surgical remedies. SH Merck Sharp and Dohme Ltd No We believe minor modification is required to one of the recommendations made in CG59; namely, recommendation Modification is required due to this recommendation being None None 42 of 77

43 misinterpreted by NHS bodies (including NICE and the NPC). The recommendation is also not entirely in line current SmPCs; these discrepancies may be the cause of any misinterpretation. on areas excluded on equality Current wording of Recommendation : "When offering treatment an oral NSAID/COX2 inhibitor, the first choice should be either a standard NSAID or a COX-2 inhibitor (other than etoricoxib 60 mg). In either case, these should be co-prescribed a PPI [proton pump inhibitor], choosing the one the lowest acquisition cost." Suggested modification to : Suggestion #1: We believe that clarity would be most greatly improved through minor modification of recommendation ; specifically the removal of the text "(other than etoricoxib 60 mg)". We respectfully suggest an alternative wording for as: "When offering treatment an oral NSAID/COX2 inhibitor, the first choice should be either a standard NSAID or a COX-2 inhibitor, prescribed at the recommended starting dose. In either case, these should be 43 of 77

44 co-prescribed a PPI, choosing the one the lowest acquisition cost " on areas excluded on equality Such a modification will place recommendations for etoricoxib in line those for other NSAIDs and COX-2 inhibitors, and in line marketing authorisations/mhra guidance for the starting dose of these treatments. Suggestion #2: As an alternative, we suggest that consistency in the recommendation could be provided by adding a similar level of detail for higher (non-starting) doses of other COX-2's, and all standard NSAIDs. If the guideline were to incorporate such advice, the exclusion in would be extended to non-starting doses of all NSAIDs, including celecoxib 400 mg. Rationale for : i) Exclusion of etoricoxib 60 mg is irrelevant, given it is not the authorised starting dose in OA We believe it is inappropriate for the guideline to specifically recommend against the use of etoricoxib 60 mg, as a starting dose for osteoarthritis. Based on current product SmPCs, such a recommendation is inappropriate because: The initial dose for etoricoxib in OA is 44 of 77

45 30 mg. The marketing authorisation states that: "The recommended dose is 30 mg once daily. In some patients insufficient relief from symptoms, an increased dose of 60 mg once daily may increase efficacy..." It is, therefore, not appropriate to consider and subsequently bar etoricoxib 60 mg as an initial treatment option for OA. The guideline, and supporting economic model upon which this statement is made, only make recommendations on the initial treatment for OA, after the use of paracetamol and topical NSAIDs. In the scenarios where initial treatment does not provide sufficient pain relief, or the treatment is not tolerated, no recommendations are made. The guideline does not include a similar 'initial dose' recommendation for celecoxib 400 mg. The initial dose for celecoxib in OA is 200 mg, although a 400 mg dose was also considered in the 2008 guideline. The marketing authorisation states that: "The usual recommended daily dose is 200 mg taken once daily or in two divided doses. In some patients, insufficient relief from on areas excluded on equality 45 of 77

46 symptoms, an increased dose of 200 mg twice daily may increase efficacy " Celecoxib 400 mg was also not included in the economic modelling supporting CG59. Based on the assumptions made in the modelling, celecoxib 400 mg would have been consistently dominated (i.e. giving a poorer health outcome at a higher cost). on areas excluded on equality Please refer to the 'Historical background' section below for further discussion on how CG59 was produced, and detail of the assumptions behind the health economic modelling, which was used in the development of the recommendations. ii) Exclusion of etoricoxib 60 mg is misleading and has been misinterpreted: Due to the inconsistency in the selective discussion of etoricoxib 60 mg in , the recommendation has been misinterpreted. It has been interpreted by a number of NHS bodies (including NICE, in other workflows), as a recommendation against using etoricoxib at any dose, which we understand was not the intention of the committee when drafting the guidance. It may also lead some prescribers to 46 of 77