Written by Kenneth D. Brandt, MD, Professor of Medicine and Head, Rheumatology Division, Indiana University School of Medicine

Osteoarthritis (OA) is the most common joint disease, and OA of the knee, in particular, is the major cause of chronic disability among the elderly in this country. It is clear that nonpharmacologic measures, such as patient education; exercises to strengthen periarticular muscles and maintain or improve range of motion and cardiovascular fitness; weight reduction (if the patient is obese); and provision of social support can all significantly benefit the management of the patient with OA. The 1995 American College of Rheumatology (ACR) Guidelines for medical management of OA of the knee and OA of the hip1,2 emphasized that nonpharmacologic measures are the keystone of symptomatic therapy for this disease. The recently published update of the ACR recommendations3 reinforces the importance of such measures in the overall treatment program.

Limited Use by Family Physicians of Nonpharmacologic Measures For Management of OA

Despite the acknowledged importance of non-medicinal measures in the management of OA, such measures are often not employed by physicians. For example, a Canadian survey4 of 292 OA patients cared for by family physicians-the results of which were presented at the 2000 ACR meeting-revealed that many of the physicians did not discuss such treatment options with their patients. Less than 38% of these patients said their doctors recommended exercise; even fewer reported that their doctors recommended application of heat or cold to the joints (29.8%) or the use of assistive devices (15.8%).

The Canadian survey data are consistent with those of a recent report5 that only a minority of specialists in general internal medicine, family medicine, or rheumatology in the state of Indiana utilized nonpharmacologic modalities in the management of patients with OA, although patients seen by rheumatologists were more likely to report instruction on the use of isometric and range of motion exercises than patients of the other specialists. Notably, the survey revealed also that family practice specialists in Indiana prescribed nonsteroidal anti-inflammatory drugs (NSAIDs) for their OA patients in doses substantially higher than those prescribed by internists or rheumatologists-a finding of some importance, since serious adverse events associated with NSAID use are dose-dependent.

Several studies have emphasized the benefits of patient education in management of OA. An abstract presented at the 2000 ACR meeting described preliminary findings of a controlled evaluation of a self-care intervention employed by patients with knee OA within a health maintenance organization (HMO).6 The results were consistent with those of a previous study by the same investigators,7 in which pain control and maintenance of physical function by patients who were taught self-care measures by a rheumatology nurse educator were significantly improved compared to control subjects.

It is important to recognize that most patients with significant OA symptoms cannot be effectively managed over the long term with drug treatment alone. Furthermore, any medication used for the treatment of OA pain is likely to be more effective if employed in the context of an individualized nonmedicinal program. This has the obvious advantage of permitting a lower dose of analgesic or NSAID, reducing the risk of adverse events. Furthermore, lower doses mean lower cost of treatment.

Why, then, are nonpharmacologic measures not employed more often in management of OA symptoms? A large part of the problem may be the limited time available to primary care physicians for implementation of such measures. In a managed care setting, physicians may be able to spend not more than 10 minutes on a follow-up visit-during which time they must deal not only with the painful knee but, often, also with the patient's diabetes, heart failure, constipation, etc. It is understandable that physicians rely almost exclusively on drugs to manage patients with OA. However, satisfaction with NSAIDs for treatment of OA is low, due to their modest efficacy and adverse effects. Concerns exists not only about catastrophic upper GI adverse events, but also nonspecific GI toxicities and adverse effects in other organs resulting, e.g., in salt and water retention, hypertension, precipitation of congestive heart failure and renal insufficiency. The risk of these side effects exists with the use of COX-1 sparing NSAIDs, as well as with nonselective COX inhibitors.8,9,10 It is not surprising, therefore, that only about 15% of patients with hip or knee OA in whom NSAID treatment was initiated were found to be taking the same NSAID 12 months later.11 Adverse Effects of Drugs Commonly Used for the Management of OA Pain

NSAIDs

Although serious GI adverse events associated with NSAID use (e.g., gastroduodenal ulceration, bleeding, perforation, gastric outlet obstruction) are relatively uncommon, nonspecific GI complaints (e.g., dyspepsia, nausea, epigastric pain), while not life-threatening, are common and important. GI complaints affect compliance and lead to changes in medication and increased cost of care due to use of concomitant gastroprotective therapy and investigations into the cause of the symptoms.

In a recently published meta-analysis, Langman et al.12 found that the incidence of non-specific GI complaints in patients treated with rofecoxib (Vioxx(r)) was similar to that in patients treated with nonselective NSAIDs. However, Bombardier et al.,13 discussing the results of the long-term rofecoxib safety study (Vioxx(r) Gastrointestinal Outcomes Research (VIGOR) study, see below) at the Philadelphia ACR meeting, reported that rofecoxib-which significantly decreased the incidence of upper GI perforation, ulcers and bleeding in comparison with naproxen-also decreased the frequency with which new gastroprotective agents were prescribed and GI procedures performed.

For patients at high risk for a serious GI complication from a nonselective NSAID, the coxibs hold an important advantage over nonselective NSAIDs. In an abstract presented at the Philadelphia ACR meeting, describing the results of the Celecoxib Long-term Arthritis Safety Study (CLASS)-a randomized, double-blinded, multicenter trial in which nearly 8,000 patients were treated with a supratherapeutic dose of celecoxib (400 mg bid) or with standard doses of ibuprofen (800 mg tid) or diclofenac sodium (75 mg bid) and followed for 6 months or more-Silverstein et al.14 reported that, among patients who were not taking concomitant low-dose aspirin for cardiovascular prophylaxis, the annualized incidence of symptomatic ulcers and ulcer complications was significantly lower in the celecoxib group than among those taking nonselective NSAIDs. However, although celecoxib treatment resulted in a statistically significant decrease in the combined incidence of symptomatic ulcers and ulcer complications relative to the nonselective NSAIDS, the ulcer complication rate alone (i.e., excluding symptomatic ulcers) was not significantly lower than with the nonselective NSAIDs.

Furthermore, although celecoxib use in the CLASS study was associated with a lower incidence of clinically significant decreases in hemoglobin and hematocrit and depletion of iron stores than the non-selective NSAID comparators-even among patients receiving low-dose aspirin15-the recently published peer-reviewed full-length report of the CLASS study16 indicates that celecoxib offered no advantage-relative to nonselective NSAIDs-with respect to the incidence of serious GI adverse events among patients who were also taking low-dose aspirin (approximately 20% of the subjects enrolled). This is relevant because the prevalence of low-dose aspirin use among patients 65 or older-an age group in which the prevalence of symptomatic OA is high-may be 50% or more.

On the other hand, in the long-term GI outcome study of rofecoxib (the VIGOR study), in which-in contrast to the CLASS study-the use of low-dose aspirin was an exclusion criterion, the incidence of myocardial infarction was some four-fold greater in the rofecoxib group than in subjects treated with the nonselective NSAID naproxen.17 Whether COX-1 sparing NSAIDs such as rofecoxib and celecoxib (Celebrex(r))-perhaps by virtue of their lack of effect on platelet thromboxane synthesis with concurrent inhibition of endothelial prostacyclin production-increase the risk of thrombosis in predisposed individuals remains to be determined.18

These findings are important: the benefits of low-dose aspirin for prophylaxis of cardiovascular disease are clear. No one would argue that the prevention of potentially fatal thromboembolic disease should not take precedence over treatment of joint pain. If, as the results of the CLASS study suggest, coxibs are not more effective than non-selective NSAIDs in preventing GI catastrophes among OA patients taking low-dose aspirin, this is a significant limitation.

Acetaminophen

What about safety issues relative to acetaminophen, the drug most widely recommended as the initial pharmacologic agent for the symptomatic management of OA? The new ACR recommendations for the management of OA3 state, "although it is one of the safest analgesics, acetaminophen can be associated with clinically important adverse events," and advise " ...increased awareness of potential toxicity..." in patients taking warfarin sodium (Coumadin(r)), and in those with existing liver disease or chronic alcohol abuse. What are the data with respect to these issues?

In therapeutic doses (1,000 mg every 4 to 6 hours up to 4,000 mg per day or 1,300 mg of an extended release formulation given every 8 hours up to 3,900 mg per day) acetaminophen is the safest analgesic drug available. Although concern has been raised recently about the possibility that it may prolong the prothombin time in some patients taking warfarin sodium,19, 20 the evidence is contradictory with regard to this point.21, 22 It is reasonable, nonetheless, to re-evaluate the prothombin time in patients taking warfarin sodium if chronic dosing with acetaminophen is initiated. On the other hand, it should be noted that the use of warfarin sodium is a relative contraindication to the prescription of a nonselective NSAID and that administration of celecoxib or rofecoxib to patients receiving warfarin sodium has been associated with clinically significant bleeding episodes.23, 24

With respect to use of acetaminophen in patients who consume alcohol regularly, careful examination of the literature raises doubt about the validity of the widely held premise that the use of therapeutic doses of acetaminophen increase the risk of developing hepatotoxicity in chronic alcoholics at a time when they theoretically would be most susceptible to any acetaminophen effect on the liver. In a prospective placebo-controlled trial, acetaminophen (4,000 mg/d) was given at a detoxification center to chronic alcoholics, hepatic function was measured at baseline, after 2 days of dosing, and 2 days after dosing was terminated. No difference in hepatic function test measurements was observed between the acetaminophen treatment group and those who received placebo.25

It is recognized that NSAIDs in recommended doses may cause hepatotoxicity.26, 27 On the other hand, although acetaminophen overdose may cause hepatic necrosis-and even hepatic failure-in recommended doses, acetaminophen has not been shown to adversely affect liver function in patients with chronic liver disease,28 suggesting that it is preferable to NSAIDs as an analgesic in such patients.

Furthermore, the Scientific Advisory Committee of the National Kidney Foundation has recommended acetaminophen as the non-narcotic analgesic of choice for episodic use in patients with renal disease.29

Efficacy of NSAIDs for the Management of OA Pain

The magnitude of reduction of OA pain achieved with NSAIDs is only modest. NSAIDs decrease OA pain, on average, by only 20-25%, relative to the baseline value.30, 31 Lack of efficacy is a major factor underlying the lack of satisfaction with NSAID therapy. And, in head-to-head comparisons, rofecoxib and celecoxib have been no more effective than nonselective NSAIDs with respect to the magnitude of improvement of OA pain.

Selection of the Initial Drug for Treatment of OA Pain

The recent approval by the FDA of COX-1 sparing NSAIDs has had a profound effect on the management of OA-more than 50% of all NSAID prescriptions written for patients over the age of 65 who have not had a NSAID prescription filled in the past 12 months are now written for a coxib. Is this warranted?

The revised ACR recommendations for the management of OA3 suggest that NSAIDs receive consideration as an alternative to acetaminophen as initial drug therapy for OA patients with more severe joint pain or clinical signs of joint inflammation. However, these recommendations stand in contrast not only to the 1995 ACR recommendations,1,2 but to the new recommendations for the management of OA published by the European League Against Rheumatism (EULAR);32 recommendations derived from an evidenced-based analysis of the results of OA treatment with analgesics and NSAIDs,33 and conclusions reached in a recent British Medical Journal publication based on an evaluation of the published data.34

Indeed, few data exist with respect to the question whether evidence of joint inflammation predicts a better response to an NSAID than to a simple analgesic in patients with OA. In a post hoc analysis, Bradley et al.35 found that the presence of clinical signs of joint inflammation did not indicate a superior response to an anti-inflammatory dose of ibuprofen (2400 mg/d), in comparison to acetaminophen (4000 mg/d) in patients with knee OA. While the retrospective nature of this analysis is a limitation, and the study was not sufficiently powered to answer the question, the results do not support the ACR recommendation of NSAIDs as alternative first-line therapy for the management of OA. Rather, they emphasize the need for a prospective clinical trial, with a sufficient number of participants and the use objective and sensitive assays of inflammation, to help determine whether the severity of joint inflammation predicts the response to NSAIDs and analgesics in patients with OA.

Similarly, how useful is the severity of joint pain as a predictor of the response to an anti-inflammatory versus analgesic therapy of OA? In a study presented at the previous year's ACR meeting in Boston, in which patients with knee OA were treated for 6 days with ibuprofen (1,200 mg/d-i.e., only an analgesic dose), acetaminophen (4000 mg/d), or placebo, among those subjects who had more severe knee pain at baseline, improvement in joint pain was significantly greater in the ibuprofen treatment group than among those who received acetaminophen.36

However, in short-term studies such as the above, differences between drugs with respect to pharmacodynamics and pharmacokinetics may determine outcomes.37 Therefore, a pain study of only a few days duration may not indicate how drugs will compare in a longer study of chronic pain in OA. In support of this caveat, results of a retrospective study38 in which the treatment period was longer (i.e., 30 days), the severity of general arthritis pain, rest pain or walking pain at baseline did not indicate a better clinical response to an NSAID over an analgesic.

The abstract presented by Geba et al.39 at the 2000 ACR meeting provides a direct comparison of rofecoxib, celecoxib, and acetaminophen in a 6-week double-blind randomized study of patients with knee OA. Rofecoxib (25 mg/d) was significantly superior to the other treatments, as judged by Western Ontario and McMaster Universities (WOMAC) pain and stiffness scores, and superior to acetaminophen (4000 mg/d) with respect to functional improvement. Furthermore, the patients' global assessment of response to therapy over 6 weeks favored rofecoxib (either 12.5 mg or 25 mg/d) over acetaminophen, and rofecoxib (25 mg/d) over celecoxib (100 mg/d).

In considering the above results, however, it is important to differentiate statistical significance from clinical significance. For example, at the 1999 Boston ACR meeting-in which results of a double-blind crossover comparison of a diclofenac/misoprostol formulation and acetaminophen (4000 mg/d) were presented,40 a statistically significant difference favoring the NSAID over acetaminophen with respect to magnitude of improvement in pain, function, and quality of life was reported. However, when the subjects and investigators-who were blinded to the treatments-were asked which treatment they preferred, nearly 50% of both favored acetaminophen.

It would appear however, that acetaminophen is underutilized as first-line drug therapy for OA. Based on the results of a survey of family physicians, Li et al.4 reported at the 2000 ACR meeting that "while most patients were advised to try anti-inflammatory medications, less than two-thirds were told to take acetaminophen, despite it being recommended as first-line therapy according to the ACR recommendations for managing osteoarthritis." Nearly 83% of the patients in her survey had been advised to take NSAIDs, while only 172 (58.9%) reported that they had been told to take acetaminophen. Her findings are consistent with those of Mazzuca et al.,5 who noted that family physicians in Indiana often prescribed NSAIDs as initial therapy for their OA patients, and often in high doses.

Conclusion

Nonpharmacologic modalities-the cornerstone of the management of OA-remain underutilized by physicians. Acetaminophen remains the preferred drug for initial pharmacologic treatment of OA pain but also appears to be underutilized. In 1995, based on considerations of safety, efficacy, and cost, the ACR recommendations cite acetaminophen as the initial drug of choice for the management of OA pain.1,2 The evidence to date provides no basis for altering or qualifying this recommendation. While COX-1 sparing NSAIDs are a significant therapeutic advance and monumental achievement in medicinal chemistry, they should not be considered the standard of care for initial pharmacotherapy of OA. If, however, when used in conjunction with appropriate nonpharmacological measures, acetaminophen is ineffective after two or three weeks, the use of an NSAID is then reasonable in most cases.

References

1. Hochberg MC, Altman RD, Brandt KD, Clark BM, Dieppe PA, Griffin MR, et al. Guidelines for the medical management of osteoarthritis. Part I. Osteoarthritis of the hip. Arthritis Rheum. 1995;38:1535-40.

2. Hochberg MC, Altman RD, Brandt KD, Clark BM, Dieppe PA, Griffin MR, et al. Guidelines for the medical management of osteoarthritis. Part II. Osteoarthritis of the knee. Arthritis Rheum. 1995;38:1541-6.

3. Altman RD, Hochberg MC, Moskowitz RW, Schnitzer TJ. Recommendations for the medical management of osteoarthritis of the hip and knee: 2000 Update. Arthritis Rheum. 2000;43:1905-1915.

4. Li LC, Maetzel A, Maguire L, Jansz G, Nair B, Bombardier C, The CHAP team Toronto, ON, Canada. Utilization of Medications and Non-medical Remedies by People with Osteoarthritis. Arthritis Rheum. 2000;43 (Suppl 9) 391:S132.

5. Mazzuca SA, Brandt KD, Katz BP, Dittus RS, Freund DA, Lubitz R, Hawker G and Eckert G: Comparison of general internists, family physicians, and rheumatologists managing patients with symptoms of osteoarthritis of the knee. Arthritis Care Res. 10:289-299, 1997.

6. Mazzuca SA, Brandt KD, Katz BP, Olin PM, Ragozzino LR, Caffrey HM. Controlled evaluation of a self-care intervention for patients with knee osteoarthritis (OA) in a health maintenance organization (HMO). Arthritis Rheum. 2000;43(Suppl 9)1589:S331.

7. Mazzuca SA, Brandt KD, Katz BP, Hanna MP, Melfi CA: Reduced utilization and cost of primary care clinic visits resulting from self-care education for patients with osteoarthritis of the knee. Arthritis Rheum. 42:1267-73, 1999.

8. Whelton A. Renal and related cardiovascular effects of conventional and COX-2 specific NSAIDs and non-NSAID analgesics. Am J Ther. 2000;7:63-74.

9. Swan SK, Rudy DW, Lasseter KC, Ryan CF, Buechel KL, Lambrecht LJ, Pinto MB, Dilzer SC, Obrda O, Sundblad KJ, Gumbs CP, Ebel DL, Quan H, Larson PJ, Schwartz JI, Musliner TA, Gertz BJ, Brater DC, Yao SL. Effect of cyclooxygenase-2 inhibition on renal function in elderly persons receiving a low-salt diet. Ann Intern Med. 2000;133:1-9.

10. Brater DC. Effects of nonsteroidal anti-inflammatory drugs on renal function: focus on cyclooxygenase -2 - selective inhibition. Am J Med. 1999;107(6A):65S-71S.

11. Scholes D, Stergachis A, Penna PM, Normand EH, Hansten PD. Nonsteriodal anti-inflammatory drug discontinuation in patients with osteoarthritis. J Rheum. 1995;22:708-712

12. Langman MJ, Jenson DM, Watson DJ, Harper SE, Zhao P-l, Quan H, et al. Adverse upper gastrointestinal effects of rofecoxib compared with NSAIDS. JAMA. 1999;282-1929-1933.

13. Bombardier C, Lane L, Reicin A, Watson D, Ramey DR, Reagan P. Fewer gastrointestinal protective agents, procedures, and hospitalizations with rofecoxib vs. naproxen in the VIGOR (Vioxx GI Outcomes Research) study. Arthritis Rheum. 2000;43(Suppl 9)S225:949.

14. Silverstein FE, Simon L. Celexocib is associated with lower incidence of serious GI toxicity relative to nonsteroidal anti-inflammatory drugs (NSAIDs). Arthritis Rheum. 2000;49(Suppl 9)1907:S384.

15. Goldstein JL. Celecoxib is associated with reduced chronic gastrointestinal (GI) blood loss relative to nonsteroidal anti-inflammatory drugs (NSAIDs). Arthritis & Rheum. 2000;43(Suppl 9)S147:482.

16. Silverstein G, Faich G, Goldstein JL, Simon LS, Pincus T, Whelton A, et al. Gastrointestinal toxicity and celecoxib vs nonsteroidal anti-inflammatory drugs for osteoarthritis and rheumatoid arthritis. JAMA. 2000;284:1247-1255.

17. Bombardier C, Laine L, Reicin A, Shapiro D, Burgos-Vargas R, Davis B, Day R, Feraz MB, Hawkey CJ, Hochberg MC, Kvien TK, Schnitzer TJ. Comparison of upper gastrointestinal toxicity of rofecoxib and naproxen in patients with rheumatoid arthritis. NEJM. 2000;343:1520-1528.

18. Crofford LJ, Oates JC, McCune WJ, Gupta S, Kaplan MJ, Catella-Lawson F, Morrow JD, McDonagh KT, Schmaier AH. Thrombosis in patients with connective tissue diseases treated with specific cyclooxygenase 2 inhibitors. Arthritis Rheum. 2000;43:1891-1896.

19. Hylek EM, Heiman H, Skates SJ, Sheehan MA, Singer DE. Acetaminophen and other risk factors for excessive warfarin anticoagulation. JAMA. 1998;279:657-662.

20. Fitzmaurice DA, Murray JA. Potentiation of anticoagulant effect of warfarin. Postgrad Med J. 1997;73:439-440.

21. Antlitz Am, Awalt LF. A double blind study of acetaminophen used in conjunction with oral anticoagulant therapy. Curr Ther Res. 1969;360-361.

22. Udall JA. Drug interference with warfarin therapy. Clin Med. 1970;20-25.

23. Package insert. Vioxx(r), West Point PA, Merck and Co., Inc. 1999.

24. Bello A. Letter to Health Care Professional, Searle, Skokie IL and Pfizer, New York NY, May, 1999.

25. Kuffner E, Dart RC, Bogdon GM et al. Evaluation of hepatotoxicity in alcoholic patients from therapeutic dosing of acetaminophen. J Toxicol, Clin Tox. 1999;37(5):641.

26. Physician's Desk Reference. 54th ed. Montvale, NJ: Medical Economics Co. 2000:1685.

27. Riley TR, Smith JP. Ibuprofen-induced hepatotoxicity in patients with chronic hepatitis C: A case series. 1998;93:1563-1565.

28. Forrest JA, Adriaenssens P, Finlayson ND, Prescott LF. Paracetamol metabolism in chronic liver disease. Eur J Clin Pharmacol. 1979 ;15 :427-431.

29. Henrich WL, Agodaoa LE, Barret B, Bennett WM, Blantz RC, Buckalew VM, et al. Analgesics and the kidney: summary and recommendations to the Scientific Advisory Board of the National Kidney Foundation from an Ad Hoc Committee of the National Kidney Foundation. Am J Kidney Dis. 1996;27:162-165.

30. Tyson VCH, Glynne A. A comparative study of benoxaprofen and ibuprofen in osteoarthritis in general practice. J Rheum. 1980;7 (Suppl 6):132-138.

31. Levinson DJ, Rubinstein HM. Double-blind comparison of fenoprofen calcium and ibuprofen in osteoarthritis of large joints. Curr Ther Res. 1983;34:280-284.

32. Pendleton A, Arden N, Dougados M, Doherty M, Bannwarth B, Bijlsma J, et al. EULAR recommendations for the management of knee osteoarthritis. Report of a task force of the Standing Committee for International Clinical Studies Including Therapeutic Trails. Ann Rheum Dis. 2000;59:936-944.

33. Eccles M, Freemantle N, Mason J. North of England evidence based guideline development project: summary guideline for non-steroidal anti-inflammatory drugs versus basic analgesia in treating the pain of degenerative arthritis. BMJ. 1998;317:526-530.

34. Osteoarthritis. In: Clinical Evidence, BMJ Publishing Group, London, 2000;3:530.

35. Bradley JD, Brandt KD, Katz BP, Kalasinski LA, Ryan SI. Treatment of knee osteoarthritis : relationship of clinical features of joint inflammation to the response to a nonsteroidal anti-inflammatory drug or pure analgesic. J Rheum. 1992;19:1950-1954.

36. Altman RD. Ibuprofen, acetaminophen and placebo in osteoarthritis of the knee: a six-day double-blind study. Arthritis Rheum. 1999;42:S403.

37. Collins SL, Moore RA, McQuay JH, Wiffen PJ, Edwards JE. Single dose ibuprofen and diclofenac for postoperative pain. The Cochrane Database of Systematic Reviews, vol. 4, 1999.

38. Bradley J, Katz BP, Brandt KD: Severity of knee pain does not predict a better response to an anti-inflammatory dose of ibuprofen than to analgesic therapy in patients with osteoarthritis. J Rheum. In press.

39. Geba CP, Weaver AL, Schnizter TJ, Fleischmann RM, Polis AB, Matzura-Wolfe DM, Dixon ME, Gormley GJ. A comparison of rofecoxib to celecoxib and acetaminophen in the treatment of osteoarthritis. Arthritis & Rheum. 2000;43(Supp 9)S384:1906.

40. Pincus T, Callahan LF, Wolfe F, Cummins P, Weaver A, Caldwell J, Lautzenheiser R, Schnitzer T, Jordan J, Fort J, Drower E, Osterhaus J, Zhao S, Lefkowith J. Arthrotec compared to acetaminophen (ACTA): A clinical trial in patients with osteoarthritis (OA) of the hip or knee. Arthritis Rheum. 1999;9:42(Suppl 9) S404.