Dexamethasone for Tuberculous Meningitis
http://www.100md.com
《新英格兰医药杂志》
To the Editor: The results of the placebo-controlled clinical trial of adjunctive dexamethasone in the treatment of tuberculous meningitis reported by Thwaites et al. (Oct. 21 issue)1 clearly indicate improved survival in the age group studied — patients more than 14 years of age. Since the percentage of dexamethasone-treated patients who died or had severe disability was not statistically different from that in the control group, one suspects that the patients who survived because of dexamethasone therapy tended to be left with severe disability. One factor not commented on either in the article or in the accompanying editorial by Quagliarello2 is the subgroup analysis, in which it appears that patients who had a grade I disability at the onset of treatment (indicated by a Glasgow Coma Scale score of 15, the best possible score) had a lower rate of death or severe disability with dexamethasone treatment than with placebo (21.1 percent vs. 34.9 percent). Those with an initial rating of grade II or III were not helped. These observations suggest that dexamethasone therapy is useful primarily for patients who start out with the least amount of difficulty.
Stephen J. Seligman, M.D.
New York Medical College
Valhalla, NY 10595
stephen_seligman@nymc.edu
References
Thwaites GE, Nguyen DB, Nguyen HD, et al. Dexamethasone for the treatment of tuberculous meningitis in adolescents and adults. N Engl J Med 2004;351:1741-1751.
Quagliarello V. Adjunctive steroids for tuberculous meningitis -- more evidence, more questions. N Engl J Med 2004;351:1792-1794.
To the Editor: There is a serious oversight in the report by Thwaites et al. The effects of corticosteroids in patients treated with rifampicin are greatly attenuated, owing to induction of P-450 enzymes in the liver, which inactivate dexamethasone and other corticosteroids. Patients with Addison's disease must double or triple their replacement dose of corticosteroids when they are receiving concomitant treatment with rifampicin in order to prevent addisonian crisis.1 The potency of dexamethasone in suppressing serum cortisol in normal volunteers is practically dissipated by the coadministration of rifampicin; even 4 mg of dexamethasone does not suppress serum cortisol.2 It appears, therefore, that the "effective" dose of dexamethasone received by the patients in the study by Thwaites et al. was much lower than the daily dose administered (a 70-kg man would have received 7 to 28 mg daily; the dosage varied according to the time period). Could the attenuation of dexamethasone's action by rifampicin have affected the outcomes? Perhaps higher doses of dexamethasone would have prevented severe disability.
Apostolos G. Vagenakis, M.D.
Venetsana Kyriazopoulou, M.D.
University of Patras Medical School
26500 Patras, Greece
vag.inmd@med.upatras.gr
References
Kyriazopoulou V, Parparousi O, Vagenakis AG. Rifampicin-induced adrenal crisis in addisonian patients receiving corticosteroid replacement therapy. J Clin Endocrinol Metab 1984;59:1204-1206.
Kyriazopoulou V, Vagenakis AG. Abnormal overnight dexamethasone suppression test in subjects receiving rifampicin therapy. J Clin Endocrinol Metab 1992;75:315-317.
To the Editor: The study by Thwaites et al. showed a substantial decrease in mortality among adolescents and adults with tuberculous meningitis treated with adjunctive systemic corticosteroids. I was surprised by Dr. Quagliarello's recommendation that further research be conducted to assess whether this therapy decreases morbidity. Although Thwaites et al. did not observe a benefit in terms of morbidity, it would seem difficult to justify additional placebo-controlled trials in this situation, when good evidence supports the reduction in mortality, with an impressive number needed to treat — approximately 10 — to save one life. The lack of a significant reduction in mortality in certain subgroups probably resulted from inadequate statistical power. This assumption is supported by the point estimates for the relative risk of death in all subgroups analyzed. Unless there is concern regarding increased morbidity among surviving patients treated with systemic corticosteroids, I would question the ethical basis for new placebo-controlled studies in this area.
Theodore K. Marras, M.D.
Toronto Western Hospital
Toronto, ON M5T 2S8, Canada
ted.marras@utoronto.ca
The authors reply: Vagenakis and Kyriazopoulou suggest that the induction of dexamethasone metabolism by rifampicin may have influenced the results of our trial. We were aware of this interaction and chose dexamethasone doses on the basis of regimens used in previous controlled trials in which adjustment was made for this effect; these trials suggested that the doses were sufficient to influence the outcome despite concomitant rifampicin therapy.1,2 Whether the interaction accounted for our observation that dexamethasone improved survival but not the incidence of severe disability among survivors is uncertain. Schoeman et al. reported a similar finding in children2 and suggested that corticosteroids improved survival by reducing basal meningeal inflammation and brain-stem encephalopathy but did not attenuate the periarteritis responsible for infarction and disability in survivors. A better understanding of how corticosteroids influence the pathophysiology and outcome of tuberculous meningitis is required and may suggest novel adjunctive therapies.
We agree with Dr. Marras's contention that further placebo-controlled trials of dexamethasone for tuberculous meningitis in persons without human immunodeficiency virus (HIV) infection would be unethical. However, we believe such trials would still be justified in certain circumstances — for instance, when restricted to patients coinfected with HIV. Moreover, randomized trials designed with clinically important end points are still required to determine the optimal dose and duration of adjunctive dexamethasone therapy.
Dr. Seligman suggests that dexamethasone may be of greatest benefit in those who present without coma or focal neurologic deficits (grade I disease). This is an important issue, since previous trials3 and many authorities4 have suggested that adjunctive corticosteroids should be given only to patients with reduced consciousness or focal neurologic signs (grades II and III). We would caution against overinterpretation of subgroup-specific analysis: the small numbers in each group reduce the statistical power to discriminate true effects from those occurring by chance. For this reason, we emphasized the stratified subgroup analysis, the results of which suggest that the beneficial effect of dexamethasone on survival is uniform across the spectrum of disease severity and that all patients with tuberculous meningitis who are over 14 years of age should be treated with adjunctive dexamethasone, regardless of the severity of disease. The prompt diagnosis and treatment of tuberculous meningitis, before the onset of coma, will have the greatest impact on the outcome.
Guy E. Thwaites, M.R.C.P.
John Radcliffe Hospital
Oxford OX3 9DU, United Kingdom
guy.thwaites@btinternet.com
Hoang Thi Quy, M.D.
Pham Ngoc Thach Hospital for Tuberculosis and Lung Disease
Ho Chi Minh City, Vietnam
Jeremy J. Farrar, F.R.C.P.
John Radcliffe Hospital
Oxford OX3 9DU, United Kingdom
References
Girgis NI, Farid Z, Kilpatrick ME, Sultan Y, Mikhail IA. Dexamethasone adjunctive treatment for tuberculous meningitis. Pediatr Infect Dis J 1991;10:179-183.
Schoeman JF, Van Zyl LE, Laubscher JA, Donald PR. Effect of corticosteroids on intracranial pressure, computed tomographic findings, and clinical outcome in young children with tuberculous meningitis. Pediatrics 1997;99:226-231.
Prasad K, Volmink J, Menon GR. Steroids for treating tuberculous meningitis. Cochrane Database Syst Rev 2000;3:CD00224-CD00224.
Treatment of tuberculosis. MMWR Recomm Rep 2003;52:1-77.
The editorialist replies: Dr. Seligman points out that in the subgroup analyses with patients stratified according to disease-severity grade, those with grade I disability at the onset of treatment had a decrease in the rate of death or severe disability from 34.9 percent to 21.1 percent with corticosteroid treatment, as compared with placebo. However, it is important to recognize the hazards of overinterpreting subgroup analyses. The result of the authors' test of heterogeneity was nonsignificant for the combined end point of death or severe disability. This suggests that there was no stochastic evidence of a heterogeneous effect among subgroups in their analyses. I do agree, however, that further exploration of the effects of corticosteroids among those with grade I disease would be interesting in future studies.
Dr. Marras feels it is difficult to justify an additional placebo-controlled trial to study the effect of corticosteroids on morbidity. I completely agree. However, the need for additional investigation is not equivalent to the need for additional placebo-controlled, randomized trials. In settings in which randomized trials are either unethical or not feasible, observational studies have been invaluable and quite justifiable. With the use of multivariable modeling and propensity analyses to control for baseline prognostic heterogeneity and treatment-selection bias, studies have shown that observational cohort studies can generate estimates of the effects of therapeutic interventions that are quite similar to the results of randomized trials.1,2
Vincent Quagliarello, M.D.
Yale University School of Medicine
New Haven, CT 06520
References
Concato J, Shah N, Horwitz RI. Randomized, controlled trials, observational studies, and the hierarchy of research designs. N Engl J Med 2000;342:1887-1892.
Benson K, Hartz AJ. A comparison of observational studies and randomized, controlled trials. N Engl J Med 2000;342:1878-1886.
Stephen J. Seligman, M.D.
New York Medical College
Valhalla, NY 10595
stephen_seligman@nymc.edu
References
Thwaites GE, Nguyen DB, Nguyen HD, et al. Dexamethasone for the treatment of tuberculous meningitis in adolescents and adults. N Engl J Med 2004;351:1741-1751.
Quagliarello V. Adjunctive steroids for tuberculous meningitis -- more evidence, more questions. N Engl J Med 2004;351:1792-1794.
To the Editor: There is a serious oversight in the report by Thwaites et al. The effects of corticosteroids in patients treated with rifampicin are greatly attenuated, owing to induction of P-450 enzymes in the liver, which inactivate dexamethasone and other corticosteroids. Patients with Addison's disease must double or triple their replacement dose of corticosteroids when they are receiving concomitant treatment with rifampicin in order to prevent addisonian crisis.1 The potency of dexamethasone in suppressing serum cortisol in normal volunteers is practically dissipated by the coadministration of rifampicin; even 4 mg of dexamethasone does not suppress serum cortisol.2 It appears, therefore, that the "effective" dose of dexamethasone received by the patients in the study by Thwaites et al. was much lower than the daily dose administered (a 70-kg man would have received 7 to 28 mg daily; the dosage varied according to the time period). Could the attenuation of dexamethasone's action by rifampicin have affected the outcomes? Perhaps higher doses of dexamethasone would have prevented severe disability.
Apostolos G. Vagenakis, M.D.
Venetsana Kyriazopoulou, M.D.
University of Patras Medical School
26500 Patras, Greece
vag.inmd@med.upatras.gr
References
Kyriazopoulou V, Parparousi O, Vagenakis AG. Rifampicin-induced adrenal crisis in addisonian patients receiving corticosteroid replacement therapy. J Clin Endocrinol Metab 1984;59:1204-1206.
Kyriazopoulou V, Vagenakis AG. Abnormal overnight dexamethasone suppression test in subjects receiving rifampicin therapy. J Clin Endocrinol Metab 1992;75:315-317.
To the Editor: The study by Thwaites et al. showed a substantial decrease in mortality among adolescents and adults with tuberculous meningitis treated with adjunctive systemic corticosteroids. I was surprised by Dr. Quagliarello's recommendation that further research be conducted to assess whether this therapy decreases morbidity. Although Thwaites et al. did not observe a benefit in terms of morbidity, it would seem difficult to justify additional placebo-controlled trials in this situation, when good evidence supports the reduction in mortality, with an impressive number needed to treat — approximately 10 — to save one life. The lack of a significant reduction in mortality in certain subgroups probably resulted from inadequate statistical power. This assumption is supported by the point estimates for the relative risk of death in all subgroups analyzed. Unless there is concern regarding increased morbidity among surviving patients treated with systemic corticosteroids, I would question the ethical basis for new placebo-controlled studies in this area.
Theodore K. Marras, M.D.
Toronto Western Hospital
Toronto, ON M5T 2S8, Canada
ted.marras@utoronto.ca
The authors reply: Vagenakis and Kyriazopoulou suggest that the induction of dexamethasone metabolism by rifampicin may have influenced the results of our trial. We were aware of this interaction and chose dexamethasone doses on the basis of regimens used in previous controlled trials in which adjustment was made for this effect; these trials suggested that the doses were sufficient to influence the outcome despite concomitant rifampicin therapy.1,2 Whether the interaction accounted for our observation that dexamethasone improved survival but not the incidence of severe disability among survivors is uncertain. Schoeman et al. reported a similar finding in children2 and suggested that corticosteroids improved survival by reducing basal meningeal inflammation and brain-stem encephalopathy but did not attenuate the periarteritis responsible for infarction and disability in survivors. A better understanding of how corticosteroids influence the pathophysiology and outcome of tuberculous meningitis is required and may suggest novel adjunctive therapies.
We agree with Dr. Marras's contention that further placebo-controlled trials of dexamethasone for tuberculous meningitis in persons without human immunodeficiency virus (HIV) infection would be unethical. However, we believe such trials would still be justified in certain circumstances — for instance, when restricted to patients coinfected with HIV. Moreover, randomized trials designed with clinically important end points are still required to determine the optimal dose and duration of adjunctive dexamethasone therapy.
Dr. Seligman suggests that dexamethasone may be of greatest benefit in those who present without coma or focal neurologic deficits (grade I disease). This is an important issue, since previous trials3 and many authorities4 have suggested that adjunctive corticosteroids should be given only to patients with reduced consciousness or focal neurologic signs (grades II and III). We would caution against overinterpretation of subgroup-specific analysis: the small numbers in each group reduce the statistical power to discriminate true effects from those occurring by chance. For this reason, we emphasized the stratified subgroup analysis, the results of which suggest that the beneficial effect of dexamethasone on survival is uniform across the spectrum of disease severity and that all patients with tuberculous meningitis who are over 14 years of age should be treated with adjunctive dexamethasone, regardless of the severity of disease. The prompt diagnosis and treatment of tuberculous meningitis, before the onset of coma, will have the greatest impact on the outcome.
Guy E. Thwaites, M.R.C.P.
John Radcliffe Hospital
Oxford OX3 9DU, United Kingdom
guy.thwaites@btinternet.com
Hoang Thi Quy, M.D.
Pham Ngoc Thach Hospital for Tuberculosis and Lung Disease
Ho Chi Minh City, Vietnam
Jeremy J. Farrar, F.R.C.P.
John Radcliffe Hospital
Oxford OX3 9DU, United Kingdom
References
Girgis NI, Farid Z, Kilpatrick ME, Sultan Y, Mikhail IA. Dexamethasone adjunctive treatment for tuberculous meningitis. Pediatr Infect Dis J 1991;10:179-183.
Schoeman JF, Van Zyl LE, Laubscher JA, Donald PR. Effect of corticosteroids on intracranial pressure, computed tomographic findings, and clinical outcome in young children with tuberculous meningitis. Pediatrics 1997;99:226-231.
Prasad K, Volmink J, Menon GR. Steroids for treating tuberculous meningitis. Cochrane Database Syst Rev 2000;3:CD00224-CD00224.
Treatment of tuberculosis. MMWR Recomm Rep 2003;52:1-77.
The editorialist replies: Dr. Seligman points out that in the subgroup analyses with patients stratified according to disease-severity grade, those with grade I disability at the onset of treatment had a decrease in the rate of death or severe disability from 34.9 percent to 21.1 percent with corticosteroid treatment, as compared with placebo. However, it is important to recognize the hazards of overinterpreting subgroup analyses. The result of the authors' test of heterogeneity was nonsignificant for the combined end point of death or severe disability. This suggests that there was no stochastic evidence of a heterogeneous effect among subgroups in their analyses. I do agree, however, that further exploration of the effects of corticosteroids among those with grade I disease would be interesting in future studies.
Dr. Marras feels it is difficult to justify an additional placebo-controlled trial to study the effect of corticosteroids on morbidity. I completely agree. However, the need for additional investigation is not equivalent to the need for additional placebo-controlled, randomized trials. In settings in which randomized trials are either unethical or not feasible, observational studies have been invaluable and quite justifiable. With the use of multivariable modeling and propensity analyses to control for baseline prognostic heterogeneity and treatment-selection bias, studies have shown that observational cohort studies can generate estimates of the effects of therapeutic interventions that are quite similar to the results of randomized trials.1,2
Vincent Quagliarello, M.D.
Yale University School of Medicine
New Haven, CT 06520
References
Concato J, Shah N, Horwitz RI. Randomized, controlled trials, observational studies, and the hierarchy of research designs. N Engl J Med 2000;342:1887-1892.
Benson K, Hartz AJ. A comparison of observational studies and randomized, controlled trials. N Engl J Med 2000;342:1878-1886.