 |
|
|
 |
Editorial
Translating Research into Practice: Acetaminophen in Osteoarthritis Revisited
DAVID L. SCOTT King's College Hospital, Department of Rheumatology, Denmark Hill, London SE5 9RS, UK; GABRIELLE KINGSLEY, Department of Rheumatology, University Hospital Lewisham, London, England. Address reprint requests to Prof. D.L. Scott, Rheumatology Unit, Dulwich Hospital, East Dulwich Grove, London SE22 8PT, England. E-mail: david.l.scott@kcl.ac.uk Whether clinicians should first use acetaminophen or a nonsteroidal antiinflammatory drug (NSAID) to treat osteoarthritis (OA) has preoccupied rheumatologists several years. A further contribution to this debate is published in this issue. Wegman and colleagues1 report both a systematic review of some of the relevant randomized controlled trials and an overview of OA treatment guidelines. They conclude that, while NSAID are more effective, the first drug used to treat OA should be acetaminophen. They conclude that, while NSAID are more effective, the first drug used to treat OA should be acetaminophen. Although they favor acetaminophen because of the potential risks of NSAID treatment, they note that not all guidelines take the same approach. As treating the pain of OA is such a common clinical problem, it seems an obvious area in which evidence-based treatment decisions should be used to build strong clinical guidelines. The question of whether to use acetaminophen or NSAID when treating OA is not novel, with trials dating from the 1980s. So why are we still debating such a basic question? Providing an answer involves dealing with fundamental issues on the nature of clinical evidence, its translation into practice, and the role of clinical guidelines. Although such problems extend throughout medicine, exploring their impact on the narrow question of pain control in OA highlights the problems in developing evidence-based clinical care. EQUATING EFFICACY AND TOXICITY The overall assessment of Wegman and colleagues is supported by an earlier analysis from Courtney and Doherty2, who evaluated a wider range of current evidence. They recommended nonpharmacological interventions should be the core element in managing OA, that acetaminophen should retain a pivotal role in drug treatment of symptomatic OA, and that NSAID should be considered as additional treatments whose use would be defined by individual patient factors. One explanation for different groups of experts reaching dissimilar conclusions after evaluating virtually identical trial data is the way in which they restrict the use of information from non-trial sources such as longterm toxicity surveys. The analysis in the recent Cochrane review on acetaminophen3, which evaluated 6 trials involving 1689 patients, concluded NSAID are superior to acetaminophen. However, it focused exclusively on risks and benefits shown in the trials themselves, and ignored information from other sources, which may be more important in defining longterm toxic effects of treatment. There are clearly difficulties in balancing efficacy and toxicity. NSAID have significant toxicities, which can be quantified using standard methods4. There is also wide agreement on assessing efficacy in OA trials. However, there are no rules for balancing these 2 sides of the equation. Therefore some experts place more weight on toxicity and others on efficacy. The advent of the coxibs has reduced some, though not all, risks of serious adverse effects with NSAID, and this may swing the balance towards antiinflammatory drugs5. An associated problem is the continuing expansion of the clinical trial database. The conclusions in any metaanalysis must be tempered by the time it was undertaken, and new trials can modify the situation. Since Wegman and colleagues closed their metaanalysis in December 2001, 2 further randomized controlled trials have been published. Geba, et al6 compared rofecoxib and celecoxib with acetaminophen in 382 OA patients over 6 weeks. More patients treated with acetaminophen discontinued early due to lack of efficacy. Initial and late efficacy, especially pain relief, was greater with coxibs. Case and his colleagues7 compared diclofenac with acetaminophen in 82 OA patients over 12 weeks. There were early and sustained clinically and statistically significant improvements with diclofenac but not with acetaminophen. Thus the evidence of efficacy increasing favors NSAID. However, it remains likely that acetaminophen is still the safer treatment, leaving the problem of balancing risks and benefits essentially unchanged. THE LIMITS OF TRIALS OA trials, like many clinical trials, tend to exclude the elderly and patients with comorbid conditions like cardiac and renal disease, and enrol the relatively restricted healthy group of patients with OA who meet the American College of Rheumatology diagnostic criteria. Such strict exclusion and inclusion criteria raise substantial concerns about the generalizability of the findings, as OA patients who need treatment often have less clear-cut diagnostic features, are often very old, and frequently have multiple illnesses. As those patients who are excluded from trials are likely to be at greater risk of drug-induced toxicity, the design of trials pushes the balance away from lower risk treatments like acetaminophen, and this requires a balancing adjustment in an overview of the area. The clinical trial evidence base has several other negative influences. These include poor trial quality, particularly with non-drug treatments8, systematic bias from under-reporting of trials, especially if they are pharmaceutically funded9, and the potential irrelevance of trial outcome measures like WOMAC (Western Ontario and McMaster Universities OA Index) to routine practice. Discrepancies between standard patient-based measurements and patients' narrative accounts suggest an over-reliance on doctor-based assessments can prejudice the appraisal of whether treatments are successful10. Trials, because of their focus on tightly-defined patient populations, often ignore ethnic and cultural diversity. Such diversity can not only affect biological outcomes like toxicity, but also influence healthcare beliefs and thus responses to treatment. It is important not to confuse efficacy, shown in carefully controlled trials, with overall clinical effectiveness. However, whether this means we should undertake large studies to determine clinical effectiveness remains questionable. Judging the role of a treatment can only be partly evidence-based, and there must be some room for interpretation of results in the light of experience. PATIENTS' PREFERENCES Patients have central roles in determining their own care and have variable preferences when choosing treatment, especially in areas such as risk. Patients with very defined or idiosyncratic views may be under-represented in studies, since they will not wish to be randomized to a treatment they have not chosen. This may vary across cultural and ethnic backgrounds, relating to beliefs about healthcare in general and treatment in particular. Relevant examples include preferences for non-drug therapy, complementary therapy, or traditional medicine, and invasive treatments such as injections. Several studies have directly reported patients' preferences in this area, although their findings need to be interpreted with caution. In one large study Wolfe and colleagues11 sent postal questionnaires to 1799 patients with OA, rheumatoid arthritis, and fibromyalgia. Overall, patients preferred NSAID, although this preference was less in the elderly and in those with OA. Wolfe and colleagues concluded that if safety and costs were ignored, there would rarely be a reason to recommend acetaminophen over NSAID, since only 14% preferred acetaminophen. However, when safety and costs were entered into the equation, they believed acetaminophen should be tried first, as 38% of patients found it was as effective or more effective than NSAID, and it remains safer and cheaper. Pincus, et al12 reported the findings of a 15-minute telephone survey in 300 patients, including 172 with OA. Twenty-four percent of patients who took acetaminophen rated it "very helpful" compared to 31% for ibuprofen and 56% for diclofenac. However, drug continuation beyond 24 months was reported by 33% of patients with acetaminophen but only 19% for diclofenac. This was because acetaminophen was less often discontinued due to toxicity. This mirrors the findings of Wolfe and colleagues and shows the complexities of balancing short-term efficacy with longterm toxicity. Treatment withdrawals are related to adherence to treatment interventions. Adherence is another barrier to treatment success. Estimates suggest that adherence to any intervention in OA is between 50% and 95%, but as these estimates are mainly derived from clinical trials, the real levels in clinical practice are likely to be much lower13. Few trials focus on adherence as a key outcome measure, although there is some evidence that in OA, patient adherence is increased with coxibs compared to conventional NSAID14. Linked to questions of adherence is the fact that acetaminophen is freely available "over the counter." Many patients will have already taken it prior to seeking medical advice, and may not view it in the same way as a medically prescribed NSAID. Further, the most effective dose of acetaminophen, 2 tablets 4 times daily, is more intrusive than single or twice daily doses of most NSAID, which will be likely to reduce adherence to its use in routine clinical practice. CLINICIANS' VIEWS There is evidence in rheumatoid arthritis15 and OA16 and across a spectrum of rheumatic disorders17 that clinicians may have different views from their patients on the risks, benefits, and values of different treatment approaches. Despite such differences, patients' choices of treatment are strongly influenced by the health care professional caring for them. Despite, or perhaps because of, the profusion of information available, clinicians' views are only partially evidence-based. Professional background has a major influence. Therefore physiotherapists will invariably focus more on exercise, while physicians will focus more on drug treatments. Local healthcare systems provide widely different treatment options; for example, hydrotherapy is used far less in the UK than in Central Europe. Whatever the evidence base, clinicians will not advocate treatments that conflict with their own beliefs. This problem is compounded where clinicians are not involved in clinical trials, either because of personal choice or working environment, as they will feel little ownership of any data generated. GUIDELINES AND THEIR ALTERNATIVES Theoretically, guidelines provide explicit recommendations and influence practice through a formal process of disseminating advice on effective management. Ideally, they should identify and eliminate ineffective or unnecessary treatment and deliver high quality care and reproducible standards18. It is unlikely they actually achieve any of these rather grandiose aims and aspirations. In a critique of OA guidelines, Dieppe pointed out their many limitations19. First, they use a constrained or even distorted evidence base. Second, they are aimed at the average patient, and not the individual patient. Finally, the opinions in most guidelines are derived from the views of small numbers of physicians, and often minimize the assessments of other stakeholders like therapists and patients. As no one is average, each individual patient with OA will have different needs and priorities, making the application of guidelines developed for the average patient often irrelevant in the individual case. He recommended moving from protocols to statements of principles and moving from guidelines to the provision of "toolboxes" or options for both patients and professionals. It is over 10 years since we were first involved in producing OA guidelines20. The sentiments underlying these guidelines were revisited some years later21 in an attempt to ensure patients received high quality care. There have been many guidelines produced in the ensuing years, and their benefits and drawbacks are regularly reexamined22,23. It is almost certainly too naïve to ask simply whether or not they are useful. In some circumstances guidelines can have a significant role in defining clinical practice, although this is usually focused around a single treatment option in a single regional or national setting. For example, within the UK, guidance from the National Institute of Clinical Excellence has had a significant effect on the use of coxibs24. General guidelines have less significant effects, but they set the tenor for the pattern of medical care. In OA it seems eminently sensible for patients to try acetaminophen before they take an NSAID. However, most patients will already have tried this prior to seeking medical advice. Therefore a policy of clinicians always recommending acetaminophen seems bound to be impractical in routine practice. On the other hand, always giving NSAID appears equally futile. A balanced approach, treating pain with acetaminophen and NSAID singly or in combination, seems most appropriate, with the provisos that patients should not be exposed to NSAID for too long and that the elderly and unwell would be best advised to avoid them if possible. 1. Wegman A, der van Windt D, van Tulder M, Stalman W, De Vries T. Nonsteroidal antiinflammatory drugs or acetaminophen for osteoarthritis of the hip or knee? J Rheumatol 2004;31:344-54. 2. Courtney P, Doherty M. Key questions concerning paracetamol and NSAIDs for osteoarthritis. Ann Rheum Dis 2002;61:767-73. [MEDLINE] 3. Towheed TE, Judd MJ, Hochberg MC, Wells G. Acetaminophen for osteoarthritis. Cochrane Database Syst Rev 2003;2:CD004257. [MEDLINE] 4. Fries JF, Williams CA, Bloch DA. The relative toxicity of nonsteroidal antiinflammatory drugs. Arthritis Rheum 1991;34:1353-60. [MEDLINE] 5. Hochberg MC. New directions in symptomatic therapy for patients with osteoarthritis and rheumatoid arthritis. Semin Arthritis Rheum 2002;32 Suppl 1:4-14. 6. Geba GP, Weaver AL, Polis AB, Dixon ME, Schnitzer TJ; Vioxx, Acetaminophen, Celecoxib Trial (VACT) Group. Efficacy of rofecoxib, celecoxib, and acetaminophen in osteoarthritis of the knee: a randomized trial. JAMA 2002;287:64-71. [MEDLINE] 7. Case JP, Baliunas AJ, Block JA. Lack of efficacy of acetaminophen in treating symptomatic knee osteoarthritis: a randomized, double-blind, placebo-controlled comparison trial with diclofenac sodium. Arch Intern Med 2003;163:169-78. [MEDLINE] 8. Boutron I, Tubach F, Giraudeau B, Ravaud P. Methodological differences in clinical trials evaluating nonpharmacological and pharmacological treatments of hip and knee osteoarthritis. JAMA 2003;290:1062-70. [MEDLINE] 9. Gotzsche PC. Methodology and overt and hidden bias in reports of 196 double-blind trials of nonsteroidal antiinflammatory drugs in rheumatoid arthritis. Control Clin Trials 1989;10:31-56. [MEDLINE] 10. Campbell R, Quilty B, Dieppe P. Discrepancies between patients' assessments of outcome: qualitative study nested within a randomised controlled trial. BMJ 2003;326:252-3. [MEDLINE] 11. Wolfe F, Zhao S, Lane N. Preference for nonsteroidal antiinflammatory drugs over acetaminophen by rheumatic disease patients: a survey of 1,799 patients with osteoarthritis, rheumatoid arthritis, and fibromyalgia. Arthritis Rheum 2000;43:378-85. [MEDLINE] 12. Pincus T, Swearingen C, Cummins P, Callahan LF. Preference for nonsteroidal antiinflammatory drugs versus acetaminophen and concomitant use of both types of drugs in patients with osteoarthritis. J Rheumatol 2000;27:1020-7. [MEDLINE] 13. Carr A. Barriers to the effectiveness of any intervention in OA. Best Pract Res Clin Rheumatol 2001;15:645-56. [MEDLINE] 14. Arboleya LR, de la Figuera E, Soledad Garcia M, Aragon B; VICOXX Study Group. Experience of rofecoxib in patients with osteoarthritis previously treated with traditional non-steroidal anti-inflammatory drugs in Spain: results of phase 2 of the VICOXX study. Curr Med Res Opin 2003;19:288-97. [MEDLINE] 15. Hewlett SA. Patients and clinicians have different perspectives on outcomes in arthritis. J Rheumatol 2003;30:877-9. [MEDLINE] 16. Chard J, Dickson J, Tallon D, Dieppe P. A comparison of the views of rheumatologists, general practitioners and patients on the treatment of osteoarthritis. Rheumatology Oxford 2002;41:1208-10. [MEDLINE] 17. Lambert BL, Butin DN, Moran D, et al. Arthritis care: comparison of physicians' and patients' views. Semin Arthritis Rheum 2000;30:100-10. [MEDLINE] 18. Scott DL. Clinical guidelines for management. Baillieres Clin Rheumatol 1997;11:157-79. [MEDLINE] 19. Dieppe P. From protocols to principles, from guidelines to toolboxes: aids to good management of osteoarthritis. Rheumatology Oxford 2001;40:841-2. [MEDLINE] 20. Guidelines for the diagnosis, investigation and management of osteoarthritis of the hip and knee. Report of a Joint Working Group of the British Society for Rheumatology and the Research Unit of the Royal College of Physicians. J R Coll Physicians Lond 1993;27:391-6. [MEDLINE] 21. Scott DL, Shipley M, Dawson A, Edwards S, Symmons DP, Woolf AD. The clinical management of rheumatoid arthritis and osteoarthritis: strategies for improving clinical effectiveness. Br J Rheumatol 1998;37:546-54. [MEDLINE] 22. Homik JE, Suarez-Almazor ME. Clinical guidelines: a must for rheumatology? Baillieres Best Pract Res Clin Rheumatol 2000;14:649-61. [MEDLINE] 23. Pencharz JN, Grigoriadis E, Jansz GF, Bombardier C. A critical appraisal of clinical practice guidelines for the treatment of lower-limb osteoarthritis. Arthritis Res 2002;4:36-44. [MEDLINE] 24. Simister K. The role of national agencies in the managed introduction of new drugs for arthritis. Ann Rheum Dis 2000;59 Suppl 1:i38-40. |
