New Drugs for Rheumatoid Arthritis
http://www.100md.com
《新英格兰医药杂志》
To the Editor: Messori et al. (Aug. 26 issue)1 summarize the evidence supporting the use of a combination of methotrexate and tumor necrosis factor (TNF) inhibitor over methotrexate alone in patients with rheumatoid arthritis. However, their contentions that it is now unacceptable to use methotrexate alone in the control group in a randomized trial and that the combination of methotrexate and TNF inhibitor "might . . . be considered as standard therapy for patients with rheumatoid arthritis" are not supported by evidence.
In one third of untreated patients with rheumatoid arthritis, progressive joint destruction does not develop.2 In these patients, TNF inhibition is unnecessary and has not been studied. Randomized trials of TNF inhibitors have included only the half3 of all patients with rheumatoid arthritis who do not have a response to at least one disease-modifying antirheumatic drug (DMARD), usually methotrexate (Table 1).4 The results of these trials may not apply to groups of patients who have not yet been studied. Furthermore, in patients whose disease is resistant to a DMARD, combination DMARD treatment without TNF inhibitors is highly effective,6 is substantially cheaper, and has not been directly compared with methotrexate–TNF-inhibitor combinations. Therefore, current evidence suggests that TNF inhibitors provide a useful addition to our medical armamentarium, rather than a panacea, for the treatment of rheumatoid arthritis.
Table 1. Baseline Characteristics of Patients in Trials of Disease-Modifying Antirheumatic Drugs (DMARDs).
Lynden J. Roberts, M.B., B.S., Ph.D.
Royal Melbourne Hospital
Parkville 3050, Australia
References
Messori A, Santarlasci B, Vaiani M. New drugs for rheumatoid arthritis. N Engl J Med 2004;351:937-938.
Brook A, Corbett M. Radiographic changes in early rheumatoid disease. Ann Rheum Dis 1977;36:71-73.
O'Dell JR, Haire CE, Erikson N, et al. Treatment of rheumatoid arthritis with methotrexate alone, sulfasalazine and hydroxychloroquine, or a combination of all three medications. N Engl J Med 1996;334:1287-1291.
Olsen NJ, Stein CM. New drugs for rheumatoid arthritis. N Engl J Med 2004;350:2167-2179.
Klareskog L, van der Heijde D, de Jager JP, et al. Therapeutic effect of the combination of etanercept and methotrexate compared with each treatment alone in patients with rheumatoid arthritis: double-blind randomised controlled trial. Lancet 2004;363:675-681.
Grigor C, Capell H, Stirling A, et al. Effect of a treatment strategy of tight control for rheumatoid arthritis (the TICORA study): a single-blind randomised controlled trial. Lancet 2004;364:263-269.
To the Editor: Messori et al. conclude "that treatment with an anti-TNF drug plus methotrexate is more effective than methotrexate alone." Their methods are cited1 but not reported. Even in a letter to the editor, the authors of a meta-analysis should at least report whether the data were suitable for combination.2
By our analysis, there is statistically significant heterogeneity among the results of the studies analyzed (P<0.001, American College of Rheumatology criteria for an improvement of 20 percent [ACR20]; P=0.009, ACR50; and P=0.4, ACR70). Tests of heterogeneity are notorious for their low statistical power, which is further decreased when the number of studies is small.3 Moreover, the I-square analysis,4 which quantifies the variation that is due to heterogeneity rather than chance, shows a remarkable degree of variation among the results of the studies (I-square=82 percent, ACR20; I-square=74 percent, ACR50; and I-square=0 percent, ACR70). Hence, the use of a fixed-effects model seems inappropriate, since it assumes zero heterogeneity. We question whether these studies should be mathematically or informally pooled. Finally, Egger's method for regression analysis reveals statistically significant intercepts, clearly indicating publication bias or true heterogeneity. Thus, Messori and colleagues' conclusions are not justified by their methods.
Andre C. Kalil, M.D.
University of Nebraska
Omaha, NE 68198-5400
akalil@unmc.edu
References
Messori A, Trippoli S, Vaiani M, Gorini M, Corrado A. Bleeding and pneumonia in intensive care patients given ranitidine and sucralfate for prevention of stress ulcer: meta-analysis of randomised controlled trials. BMJ 2000;321:1103-1106.
Moher D, Cook DJ, Eastwood S, Olkin I, Rennie D, Stroup DF. Improving the quality of reports of meta-analyses of randomised controlled trials: the QUOROM statement. Onkologie 2000;23:597-602.
Hardy RJ, Thompson SG. Detecting and describing heterogeneity in meta-analysis. Stat Med 1998;17:841-856.
Higgins JP, Thompson SG, Deeks JJ, Altman DG. Measuring inconsistency in meta-analyses. BMJ 2003;327:557-560.
The authors reply: The observation by Kalil and by O'Dell1 that a random-effects model2 is better than a fixed-effects model for conducting meta-analyses that include potentially heterogeneous data is correct. However, after we repeated our analysis with use of a random-effects approach,2 the results remain virtually unchanged (e.g., an odds ratio for ACR20 of 4.63 [95 percent confidence interval, 2.0 to 10.8]; z=3.57; P<0.001), thus confirming that TNF inhibitors are more effective than methotrexate alone under these experimental conditions.
The controversies about new drugs for rheumatoid arthritis are evident in correspondence published recently in the Journal. For example, the evidence that TNF inhibitors are superior to methotrexate alone is limited3; the superiority of TNF inhibitors is supported by our meta-analysis, which showed a high level of statistical significance, even though the analysis was affected by heterogeneity among the patients studied; and "the evidence that TNF inhibition is an extremely effective strategy for a subgroup of patients with rheumatoid arthritis is overwhelming."1 Although they differ, each of these positions is based on reasonable considerations. Hence, the conclusion by Roberts that "TNF inhibitors provide a useful addition to our medical armamentarium, rather than a panacea" seems to be the best compromise in light of present knowledge.
One question that remains unsettled is whether, in future randomized trials involving patients with rheumatoid arthritis, the use of methotrexate will be ethically acceptable for control groups. The effectiveness of methotrexate alone consistently declined from 1999 to 2003 (Table 1), but this decline may reflect changes in enrollment criteria over time. When a patient's history includes a poor response to methotrexate or no response, the question is whether that patient can be randomly assigned to receive that drug alone. Under such circumstances, our opinion is that the use of methotrexate alone is unethical, and consequently, trials designed to make head-to-head comparisons between biologic agents (without a control group treated with methotrexate alone) are far more appropriate than trials based on a methotrexate-only control group.
Table 1. Clinical Results in Four Control Groups, Including Patients in Whom Methotrexate Had Failed, in Randomized Trials Evaluating Innovative Drugs for Rheumatoid Arthritis.
O'Dell JR. Rheumatoid arthritis. N Engl J Med 2004;351:1360-1361.
DerSimonian R, Laird N. Meta-analysis in clinical trials. Control Clin Trials 1986;7:177-188.
Burchini G, Orsi C. Rheumatoid arthritis. N Engl J Med 2004;351:1360-1360.
Olsen NJ, Stein CM. New drugs for rheumatoid arthritis. N Engl J Med 2004;350:2167-2179.
In one third of untreated patients with rheumatoid arthritis, progressive joint destruction does not develop.2 In these patients, TNF inhibition is unnecessary and has not been studied. Randomized trials of TNF inhibitors have included only the half3 of all patients with rheumatoid arthritis who do not have a response to at least one disease-modifying antirheumatic drug (DMARD), usually methotrexate (Table 1).4 The results of these trials may not apply to groups of patients who have not yet been studied. Furthermore, in patients whose disease is resistant to a DMARD, combination DMARD treatment without TNF inhibitors is highly effective,6 is substantially cheaper, and has not been directly compared with methotrexate–TNF-inhibitor combinations. Therefore, current evidence suggests that TNF inhibitors provide a useful addition to our medical armamentarium, rather than a panacea, for the treatment of rheumatoid arthritis.
Table 1. Baseline Characteristics of Patients in Trials of Disease-Modifying Antirheumatic Drugs (DMARDs).
Lynden J. Roberts, M.B., B.S., Ph.D.
Royal Melbourne Hospital
Parkville 3050, Australia
References
Messori A, Santarlasci B, Vaiani M. New drugs for rheumatoid arthritis. N Engl J Med 2004;351:937-938.
Brook A, Corbett M. Radiographic changes in early rheumatoid disease. Ann Rheum Dis 1977;36:71-73.
O'Dell JR, Haire CE, Erikson N, et al. Treatment of rheumatoid arthritis with methotrexate alone, sulfasalazine and hydroxychloroquine, or a combination of all three medications. N Engl J Med 1996;334:1287-1291.
Olsen NJ, Stein CM. New drugs for rheumatoid arthritis. N Engl J Med 2004;350:2167-2179.
Klareskog L, van der Heijde D, de Jager JP, et al. Therapeutic effect of the combination of etanercept and methotrexate compared with each treatment alone in patients with rheumatoid arthritis: double-blind randomised controlled trial. Lancet 2004;363:675-681.
Grigor C, Capell H, Stirling A, et al. Effect of a treatment strategy of tight control for rheumatoid arthritis (the TICORA study): a single-blind randomised controlled trial. Lancet 2004;364:263-269.
To the Editor: Messori et al. conclude "that treatment with an anti-TNF drug plus methotrexate is more effective than methotrexate alone." Their methods are cited1 but not reported. Even in a letter to the editor, the authors of a meta-analysis should at least report whether the data were suitable for combination.2
By our analysis, there is statistically significant heterogeneity among the results of the studies analyzed (P<0.001, American College of Rheumatology criteria for an improvement of 20 percent [ACR20]; P=0.009, ACR50; and P=0.4, ACR70). Tests of heterogeneity are notorious for their low statistical power, which is further decreased when the number of studies is small.3 Moreover, the I-square analysis,4 which quantifies the variation that is due to heterogeneity rather than chance, shows a remarkable degree of variation among the results of the studies (I-square=82 percent, ACR20; I-square=74 percent, ACR50; and I-square=0 percent, ACR70). Hence, the use of a fixed-effects model seems inappropriate, since it assumes zero heterogeneity. We question whether these studies should be mathematically or informally pooled. Finally, Egger's method for regression analysis reveals statistically significant intercepts, clearly indicating publication bias or true heterogeneity. Thus, Messori and colleagues' conclusions are not justified by their methods.
Andre C. Kalil, M.D.
University of Nebraska
Omaha, NE 68198-5400
akalil@unmc.edu
References
Messori A, Trippoli S, Vaiani M, Gorini M, Corrado A. Bleeding and pneumonia in intensive care patients given ranitidine and sucralfate for prevention of stress ulcer: meta-analysis of randomised controlled trials. BMJ 2000;321:1103-1106.
Moher D, Cook DJ, Eastwood S, Olkin I, Rennie D, Stroup DF. Improving the quality of reports of meta-analyses of randomised controlled trials: the QUOROM statement. Onkologie 2000;23:597-602.
Hardy RJ, Thompson SG. Detecting and describing heterogeneity in meta-analysis. Stat Med 1998;17:841-856.
Higgins JP, Thompson SG, Deeks JJ, Altman DG. Measuring inconsistency in meta-analyses. BMJ 2003;327:557-560.
The authors reply: The observation by Kalil and by O'Dell1 that a random-effects model2 is better than a fixed-effects model for conducting meta-analyses that include potentially heterogeneous data is correct. However, after we repeated our analysis with use of a random-effects approach,2 the results remain virtually unchanged (e.g., an odds ratio for ACR20 of 4.63 [95 percent confidence interval, 2.0 to 10.8]; z=3.57; P<0.001), thus confirming that TNF inhibitors are more effective than methotrexate alone under these experimental conditions.
The controversies about new drugs for rheumatoid arthritis are evident in correspondence published recently in the Journal. For example, the evidence that TNF inhibitors are superior to methotrexate alone is limited3; the superiority of TNF inhibitors is supported by our meta-analysis, which showed a high level of statistical significance, even though the analysis was affected by heterogeneity among the patients studied; and "the evidence that TNF inhibition is an extremely effective strategy for a subgroup of patients with rheumatoid arthritis is overwhelming."1 Although they differ, each of these positions is based on reasonable considerations. Hence, the conclusion by Roberts that "TNF inhibitors provide a useful addition to our medical armamentarium, rather than a panacea" seems to be the best compromise in light of present knowledge.
One question that remains unsettled is whether, in future randomized trials involving patients with rheumatoid arthritis, the use of methotrexate will be ethically acceptable for control groups. The effectiveness of methotrexate alone consistently declined from 1999 to 2003 (Table 1), but this decline may reflect changes in enrollment criteria over time. When a patient's history includes a poor response to methotrexate or no response, the question is whether that patient can be randomly assigned to receive that drug alone. Under such circumstances, our opinion is that the use of methotrexate alone is unethical, and consequently, trials designed to make head-to-head comparisons between biologic agents (without a control group treated with methotrexate alone) are far more appropriate than trials based on a methotrexate-only control group.
Table 1. Clinical Results in Four Control Groups, Including Patients in Whom Methotrexate Had Failed, in Randomized Trials Evaluating Innovative Drugs for Rheumatoid Arthritis.
O'Dell JR. Rheumatoid arthritis. N Engl J Med 2004;351:1360-1361.
DerSimonian R, Laird N. Meta-analysis in clinical trials. Control Clin Trials 1986;7:177-188.
Burchini G, Orsi C. Rheumatoid arthritis. N Engl J Med 2004;351:1360-1360.
Olsen NJ, Stein CM. New drugs for rheumatoid arthritis. N Engl J Med 2004;350:2167-2179.