| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
From the American Academy of Neurology, St. Paul, MN.
Address correspondence and reprint requests to the Quality Standards Subcommittee of the American Academy of Neurology, 1080 Montreal Avenue, St. Paul, MN 55116.
 |
Introduction |
|---|
 Top. Introduction Process.Analysis.Conclusions.Practice recommendation.Recommendations for future...Disclaimer.Appendix 1Appendix 2References |
|---|
In 1939, Blalock et al.1 reported the remission of generalized MG in a 21-year-old woman after removal of a cystic thymic tumor. Subsequently, Blalock et al.2,3 performed thymectomy on MG patients without thymoma, found hyperplasia in the thymus glands, and reported improvement in at least half of their patients. Since these reports, thymectomy, with or without the presence of thymoma, has gained widespread acceptance as a form of treatment for MG.
Because a definitive study of the effectiveness of thymectomy has never been done,4,5 the role of thymectomy in the management of MG remains uncertain. Nevertheless, physicians have to advise their patients regarding the benefits of thymectomy based on the existing literature. Our goal is to develop evidence-based recommendations for clinicians considering thymectomy for patients with nonthymomatous autoimmune MG by performing a systematic review and analysis of the literature.
|
Process. |
|---|
|
Top.IntroductionProcess.Analysis.Conclusions.Practice recommendation.Recommendations for future...Disclaimer.Appendix 1Appendix 2References |
|---|
We classified the resulting articles into the following categories: Class II, controlled but nonrandomized studies describing outcomes in MG patients with and without thymectomy; and Class III, uncontrolled case series describing outcomes in patients with or without thymectomy.
Within a given controlled (Class II) study, we planned to describe the therapeutic effect of thymectomy by comparing the rate MG patients receiving thymectomy achieved desirable outcomes with the rate MG patients not receiving thymectomy achieved desirable outcomes. The nonrandomized nature of Class II studies often introduces differences in the baseline characteristics of the thymectomy and nonthymectomy patient groups that can effect MG outcomes independent of thymectomy. To the extent possible, we planned to adjust for these confounding differences in baseline prognostic characteristics. In the absence of randomized controlled trials, such adjustments are critical for estimating the actual therapeutic effect, if any, of thymectomy.6
Between uncontrolled (Class III) studies, comparisons of outcome rates in thymectomy and nonthymectomy MG patients introduce additional confounding baseline differences.6 These additional confounders include institutional, geographic, and historical variations that potentially impact MG patient outcomes independent of thymectomy. Because these additional confounders are difficult to identify and adjust for, outcome comparisons between uncontrolled studies are particularly unreliable.7 For this reason, we excluded Class III studies from further consideration.
Abstracted study characteristics. From the controlled studies describing outcomes we abstracted the following characteristics: method and setting of cohort assembly, years during which patients were enrolled in the cohort, number of subjects assembled, duration of follow-up, proportion of subjects lost to follow-up, and the thymectomy techniques employed.
Because of reported associations with MG outcomes, we also extracted the following study population characteristics: proportion of patients younger than 50 years at the time of diagnosis of MG, proportion of female patients, proportion of patients with more severe MG at the time of diagnosis (defined by Osserman’s grade 2b, 3, or 48), and the proportion of patients with strictly ocular myasthenia at the time of diagnosis.
Thymectomy in patients with thymoma has benefits other than MG treatment.1 Therefore, we excluded patients with thymomas from the analysis.
Measures of association between thymectomy and outcome. In reviewing each study we considered the following MG outcomes: survival (taking into account perioperative mortality), improvement since diagnosis, asymptomatic on or off medication, and asymptomatic off medication (i.e., medication-free remission).
Ideally, controlled studies of thymectomy in MG patients report outcomes using time-to-outcome (survival) techniques.9 Survival techniques account for the differences in the duration of follow-up and changes in outcomes over time. Unfortunately, most controlled studies of thymectomy in MG simply report the proportion, or crude rate, of patients achieving an outcome over the study’s follow-up period. Crude outcome rates vary with the duration of follow-up of the study. Within a given study, however, crude outcome rates of the thymectomy and nonthymectomy patient groups can be compared when the patient groups have similar follow-up durations.
To compare the crude rates within a given controlled study, we calculated the relative rate by dividing the thymectomy patient group’s crude rate of achieving the outcome by the nonthymectomy patient group’s crude rate of achieving the outcome using the formula presented in table 1.
|
In studies providing sufficient information, we recalculated the relative rates of outcomes after controlling for potential confounding variables of age, gender, and severity of myasthenia. We used Wilcoxon’s test to determine significance of the changes in relative rates measured after controlling for these variables.
|
Analysis. |
|---|
|
Top.IntroductionProcess.Analysis.Conclusions.Practice recommendation.Recommendations for future...Disclaimer.Appendix 1Appendix 2References |
|---|
Table 2 summarizes the characteristics of these studies. Publication years ranged from 1953 to 1998. Articles describing the same cohort are grouped in the same row. The article by Beghi et al.15 describes a subset of the cohort of Mantegazza et al.17 The matched cohort of Buckingham et al.27 likely included subjects described initially by Eaton and Clagett.35
|
As mentioned previously MG patients were not allocated randomly to thymectomy or nonthymectomy treatment groups. Additionally, no study described formal selection criteria for MG patients undergoing thymectomy (e.g., the presence or absence of thymic hyperplasia during preoperative evaluations.)
Many studies did not describe the thymectomy technique employed. The majority of the studies reporting the thymectomy technique used the transsternal approach. Subjects in the cohorts described by Beghi et al.15 Mantegazza et al.,17 Papatestas et al.,20 and Perlo et al.29 underwent either transsternal or transcervical thymectomies.
Follow-up, when described, varied from an average of 3 to 28 years. Within each cohort, investigators reported comparable follow-up periods in the thymectomy and nonthymectomy groups. Most studies failed to mention the number of patients lost to follow-up. Reported losses varied from 0 to 33%.
Authors obtained data, including outcome information, from periodic examinations, retrospective chart review, or mailed and telephone surveys. No author described blinded assessments of outcomes.
Authors used nonstandard definitions of remission. The majority required the absence of MG symptoms or signs. Some studies allowed mild restricted, fixed residual weakness. The cohorts described by Henson et al.33 and Simpson34 defined remission as the ability to do a full day’s work. Most investigators did not describe a minimal asymptomatic period necessary to define remission. Beghi et al.,15 Mantegazza et al.,17 and Evoli et al.11,14 required patients to have no symptoms or signs for at least 1 year to be considered in remission.
Positive associations between thymectomy and improved outcomes. The relative rates of attaining a desired outcome in MG patients receiving and not receiving thymectomy are listed in table 3 and are illustrated in the figure. Table 3 additionally lists for each study the percentage of patients who attained a desired outcome (crude rate).
|
|
The association between thymectomy and improved outcomes achieved significance in seven of 15 studies describing medication-free remission, in eight of 12 studies describing asymptomatic patients on or off medication, in eight of 13 studies describing improvement, and in four of 13 studies describing survival. No study described a significant negative association between thymectomy and any outcome.
Using the median relative outcome rates of the controlled studies as an estimate of the magnitude of benefit, MG patients undergoing thymectomy were twice as likely to attain medication-free remission, 1.6 times as likely to become asymptomatic, and 1.7 times as likely to improve.
Confounding variables. Although the published controlled studies consistently demonstrated a positive relationship between thymectomy and outcomes, they also consistently demonstrated confounding differences between surgical and nonsurgical groups. Thymectomy patients were often enrolled in their respective cohorts years after their nonthymectomy counterparts.21,25 Additionally, with few exceptions, MG patients undergoing thymectomy were younger, more often women, and were more likely to have generalized and severe myasthenia. Younger age and female gender are associated with better MG outcomes15. The younger age of the thymectomy patients could have spuriously increased the thymectomy–better outcome association. Generalized and severe MG, by contrast, are associated with worse outcomes.15 The worse MG severity patients undergoing thymectomy could spuriously suppress the magnitude of a thymectomy benefit.
Adjusting for confounders. To adjust for these potential confounding variables, when possible we recalculated the relative rates of outcomes within MG patient subgroups. These subgroup calculations allowed us to control for one confounder variable at a time. To control for multiple confounders simultaneously, we looked for studies using a matched design or multivariate techniques.
Controlling for single confounding variables. Several studies provided sufficient data to recalculate relative rates of outcomes within MG patient subgroups.12,17,20,21,25-27,29,32-34 Table 4 lists the median change in relative outcome rates within age, gender, and severity subgroups. The positive association between thymectomy and improved outcomes persisted after controlling for each of these variables. However, the magnitude of the thymectomy-improved outcome association changed significantly in several subgroups.
|
The studies described by Mantegazza et al.,17 Papatestas et al.,20 Rodriguez et al.,25 Perlo et al.,29 and Zeldowicz and Saxton32 allowed the calculation of relative rates within the severe disease subgroup. Relative outcome rates in severe MG patients undergoing thymectomy increased by a median of 54% (p < 0.01) from the relative rate of the entire cohort. For example, in the cohort described by Perlo et al.,29 severe MG patients undergoing thymectomy were approximately 3.7 times more likely to attain remission than severe MG patients not undergoing thymectomy—a 50% increase over the relative remission rate of 2.4 of the entire thymectomy cohort described by Perlo et al.29
The calculated relative outcome rates within the mild17,20,25,29,32 subgroup was not significantly different from the baseline rates.
The association between thymectomy and favorable outcomes was 11% stronger in the generalized MG patient subgroup (i.e., after excluding ocular patients from the analysis; p < 0.10).12,17,25,26,29,33
Age and gender. The studies described by Rodriguez et al.,25 Buckingham et al.,27 and Simpson34 allowed us to determine relative outcome rates in the male MG subgroup. With the exception of the pediatric myasthenia study by Rodriguez et al.,25 the male patients were older than the female patients. The relative outcome rate decreased by a median of 25% (p < 0.05) within the male MG subgroup.
The relative outcome rates of the female myasthenia patient subgroup25,27,33,34 were 7% more than the baseline rates (p < 0.10). The relative outcome rates within the younger12,27,34 subgroup was not significantly different from the baseline rates.
Year of enrollment. We were also able to control for the patients’ year of study enrollment in the studies by Grob et al.21 and Rodriguez et al.25 In the study by Rodriguez et al.,25 a 1.4 medication-free remission relative rate persisted in the subgroup seen before 1950, and increased to 1.7 in the subgroup seen after 1950. In the cohort by Grob et al.,21 the overall cohort’s thymectomy relative remission rate of 0.7 increased to 2.2 in the subgroup of patients seen from 1940 to 1957. Thymectomy relative outcome rates were unchanged in the subgroups seen from 1958 to 1965 and 1965 to 1985.
Medical therapy. None of the studies allowed us to control for the aggressiveness of medical therapies. Thus, we cannot exclude the possibility that MG patients undergoing thymectomy received more aggressive medical therapy than MG patients not receiving thymectomy. It is conceivable that we observed a positive thymectomy-favorable outcome association merely because thymectomy was a marker for more aggressive medical therapy.
Controlling for multiple confounders. The preceding subgroup analyses adjusted for differences in one confounding characteristic at a time. In most of the studies reviewed, MG patients undergoing thymectomy differed from patients not undergoing thymectomy by several baseline characteristics. Thus, we looked for studies that attempted to adjust for multiple confounding variables simultaneously. Common mechanisms to compensate for multiple confounders include matched study designs and multivariate analyses. Such adjustments are critical for measuring a benefit from thymectomy when surgical and nonsurgical groups differ in more than one prognostically significant baseline characteristic.6
Matched design. Four studies11,13,27,37 employed a matched design and attempted to pair thymectomy MG patients with nonthymectomy patients with similar prognostic characteristics including gender, age, severity, and duration of myasthenia. With the exception of the study by Werneck et al.,13 all of these studies reported positive associations between thymectomy and outcome. However, no study matched patients successfully for all confounding patient characteristics. The studies of both Evoli et al.11 and Werneck et al.13 failed to match for disease severity. The study of Eaton et al.37 did not describe the matching technique employed. In the study by Buckingham et al.,27 patients receiving thymectomy may have been enrolled many years after patients not receiving thymectomy. Thus, because of unsuccessful matching, these studies are inconclusive.
Multivariate analyses. The studies by Beghi et al.,15 Papatestas et al.,20 and Christensen et al.10 attempted to control simultaneously for multiple confounding variables by performing a multivariate analysis (Cox proportional hazard model). The study by Christensen et al.10 did not measure an independent association between thymectomy and survival after adjusting for baseline characteristics. Papatestas et al.20 observed a significant benefit of thymectomy after adjusting for confounding differences in baseline patient characteristics (adjusted relative rate not reported). The adjusted relative rate of attaining medication-free remission after thymectomy calculated by Beghi et al.15 did not attain significance (relative rate, 1.6; 95% CI, 0.84 to 3.2). Thus, a consistent benefit from thymectomy is not reported in these studies after adjusting for differences in baseline characteristics.
Temporal relationship to remission. Indirect evidence of a thymectomy benefit could be found in a large increase in the proportion of patients attaining medication remission per unit time shortly after thymectomy. Such evidence is provided by the study by Rodriguez et al.25 This study of pediatric MG patients reports a spontaneous remission rate of 22.4 per 1,000 person-years, increasing more than 10-fold to a rate of 260 per 1,000 person-years during the first year after thymectomy. A similar although less dramatic change in remission rates is apparent in the cohort described by Beghi et al.,15 consisting mostly of adult MG patients.
Alternate explanations for the post-thymectomy changes in remission rates exist, however. Patients selected for thymectomy often receive aggressive MG medical therapy38 in an attempt to prepare them for surgery. The aggressive medical therapy could account for the change in remission rates.
Timing of surgery. Some authors suggest that a thymectomy benefit is realized more readily the earlier thymectomy is performed.39 This suggestion is based on the observation of larger remission rates per unit of time in patients undergoing thymectomy earlier after the diagnosis of MG than in those undergoing thymectomy later. This observation can be potentially explained, however, solely on the basis of the nonlinear rate at which MG patients achieve remission after diagnosis. Regardless of thymectomy, for a given duration of time, MG patients are more likely to remit earlier than later.15 Under these circumstances any therapy, effective or not, would appear to be more effective if given earlier.6
Thymectomy technique. The operative techniques employed in the controlled trials reviewed here either were not identified or consisted of standard transsternal or basic transcervical thymectomies. Some authors assert more aggressive resections, such as transcervical–transsternal maximal thymectomy, are more likely to remove accessory thymic tissue in the neck and mediastinum and result in better outcomes.9 This assertion is based on comparisons of remission rates between uncontrolled studies. Outcome comparisons between uncontrolled studies are confounded by differences in patient baseline characteristics in a way similar to that of the nonrandomized controlled trials reviewed here. Additionally, such comparisons introduce other confounders including institutional, geographic, and historical differences in the aggressiveness of medical interventions and outcome assessments that are difficult to control for.7 For these reasons, although valuable for hypothesis generation, outcome comparison between uncontrolled studies does not provide conclusive evidence of the superiority of one technique over another.
The few controlled trials comparing outcomes in MG patients undergoing thymectomy with different surgical techniques demonstrate inconsistent results. The cohorts described by Mantegazza et al.17 and Papatestas et al.20,40 included patients undergoing transsternal or basic transcervical thymectomies. In both cohorts, subjects undergoing transcervical thymectomy had better outcomes than their transsternal counterparts. In a cohort survey comparing outcomes after simple transsternal, simple transcervical, and extended thymectomy, Masaoka and Monden41 observed MG patients who received extended thymectomies had better outcomes. In all of these studies, there were numerous confounding differences between MG patients in each technique group. Thus, controlled trials did not provide convincing evidence that one thymectomy technique was superior.
Thymectomy complications. Eleven of the studies reviewed here reported thymectomy-related mortality.12,20,21,24,27-29,32-35 Perioperative mortality rates were higher in patients undergoing thymectomy before 1970, with commonly reported rates being between 5% and 15%.24,27,28,35 After 1970, reported mortality rates in the studies reviewed here were consistently less than 1%. Similarly, the mortality of thymectomy in MG patients in most modern uncontrolled series is reported to be less than 1%.42 Common morbidities include acute respiratory failure from crisis in 6%, infection in 11%, and permanent nerve injury (usually the recurrent laryngeal or phrenic nerve) in 2% of patients.42 Lower complication rates are reported with newer thymectomy techniques such as video-assisted thoracic surgery.43
|
Conclusions. |
|---|
|
Top.IntroductionProcess.Analysis.Conclusions.Practice recommendation.Recommendations for future...Disclaimer.Appendix 1Appendix 2References |
|---|
|
Practice recommendation. |
|---|
|
Top.IntroductionProcess.Analysis.Conclusions.Practice recommendation.Recommendations for future...Disclaimer.Appendix 1Appendix 2References |
|---|
|
Recommendations for future research. |
|---|
|
Top.IntroductionProcess.Analysis.Conclusions.Practice recommendation.Recommendations for future...Disclaimer.Appendix 1Appendix 2References |
|---|
If a well-designed, prospective, controlled study demonstrates that thymectomy is beneficial, then future studies could address the timing of thymectomy, which patients benefit most, and the role of other therapies in relation to thymectomy.
|
Disclaimer. |
|---|
|
Top.IntroductionProcess.Analysis.Conclusions.Practice recommendation.Recommendations for future...Disclaimer.Appendix 1Appendix 2References |
|---|
|
Appendix 1 |
|---|
|
Top.IntroductionProcess.Analysis.Conclusions.Practice recommendation.Recommendations for future...Disclaimer.Appendix 1Appendix 2References |
|---|
|
Appendix 2 |
|---|
|
Top.IntroductionProcess.Analysis.Conclusions.Practice recommendation.Recommendations for future...Disclaimer.Appendix 1Appendix 2References |
|---|
Class I. Evidence provided by one or more well-designed randomized controlled trials, including overviews (meta-analyses) of such trials.
Class II. Evidence provided by well-designed observational studies with concurrent controls (e.g., case–control and cohort studies).
Class III. Evidence provided by expert opinion, case series, and studies with historical controls.
Definitions for strength of recommendations.
Standards. A principle for patient management that reflects a high degree of clinical certainty (usually this requires Class I evidence that directly addresses the clinical question or overwhelming Class II evidence when circumstances preclude randomized clinical trials).
Guidelines. A recommendation for patient management that reflects moderate clinical certainty (usually this requires Class II evidence or a strong consensus of Class III evidence).
Practice option. A strategy for patient management for which the clinical utility is uncertain (inconclusive or conflicting evidence or opinion).
Practice advisory. A practice recommendation for emerging and/or newly approved therapies or technologies based on evidence from at least one Class I study. The evidence may demonstrate only a modest statistical effect or limited (partial) clinical response, or notable cost–benefit questions may exist. Substantial (or potential) disagreement among practitioners or between payers and practitioners may exist.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Approved by the AAN Quality Standards Subcommittee October 9, 1999. Approved by the Practice Committee January 15, 2000. Approved by the AAN Board of Directors February 26, 2000.
|
References |
|---|
|
Top.IntroductionProcess.Analysis.Conclusions.Practice recommendation.Recommendations for future...Disclaimer.Appendix 1Appendix 2References |
|---|
This article has been cited by other articles:
|
|
 |
|
  R. Gold, R. Hohlfeld, and K. V. Toyka Review: Progress in the treatment of myasthenia gravis Therapeutic Advances in Neurological Disorders, September 1, 2008; 1(2): 99 - 114. [Abstract] [PDF]
|
|
|
|
 |
|
 E. Davenport and R. A. Malthaner The role of surgery in the management of thymoma: a systematic review. Ann. Thorac. Surg., August 1, 2008; 86(2): 673 - 684. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. K. Kim, M. S. Park, Y. S. Choi, K. Kim, Y. M. Shim, J. Han, B. J. Kim, and J. Kim Neurologic outcomes of thymectomy in myasthenia gravis: Comparative analysis of the effect of thymoma J. Thorac. Cardiovasc. Surg., September 1, 2007; 134(3): 601 - 607. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Gilboa-Geffen, P. P. Lacoste, L. Soreq, G. Cizeron-Clairac, R. Le Panse, F. Truffault, I. Shaked, H. Soreq, and S. Berrih-Aknin The thymic theme of acetylcholinesterase splice variants in myasthenia gravis Blood, May 15, 2007; 109(10): 4383 - 4391. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. Zinman, E. Ng, and V. Bril IV immunoglobulin in patients with myasthenia gravis: A randomized controlled trial Neurology, March 13, 2007; 68(11): 837 - 841. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 I. K. Park, S. S. Choi, J. G. Lee, D. J. Kim, and K. Y. Chung Complete stable remission after extended transsternal thymectomy in myasthenia gravis. Eur. J. Cardiothorac. Surg., September 1, 2006; 30(3): 525 - 528. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J F Owe, A K Daltveit, and N E Gilhus Causes of death among patients with myasthenia gravis in Norway between 1951 and 2001 J. Neurol. Neurosurg. Psychiatry, February 1, 2006; 77(2): 203 - 207. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Manlulu, T. W. Lee, I. Wan, C. Y. Law, C. Chang, J. C. Garzon, and A. Yim Video-Assisted Thoracic Surgery Thymectomy for Nonthymomatous Myasthenia Gravis Chest, November 1, 2005; 128(5): 3454 - 3460. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Y. Nagane, K. Utsugisawa, H. Akutsu, Y. Sato, and Y. Terayama Perivascular infiltrate of memory lymphocytes and mature dendritic cells in MG thymomas Neurology, September 13, 2005; 65(5): 770 - 772. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. Natarajan and R. Weinstein Therapeutic Apheresis in Neurology Critical Care J Intensive Care Med, July 1, 2005; 20(4): 212 - 225. [Abstract] [PDF]
|
|
|
|
 |
|
 P. Bernasconi, M. Barberis, F. Baggi, L. Passerini, M. Cannone, E. Arnoldi, L. Novellino, F. Cornelio, and R. Mantegazza Increased Toll-Like Receptor 4 Expression in Thymus of Myasthenic Patients with Thymitis and Thymic Involution Am. J. Pathol., July 1, 2005; 167(1): 129 - 139. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E Ciafaloni Mycophenolate mofetil and myasthenia gravis Lupus, March 1, 2005; 14(3_suppl): s46 - s49. [Abstract] [PDF]
|
|
|
|
|
|
S. Endo, T. Yamaguchi, N. Saito, S. Otani, T. Hasegawa, Y. Sato, and Y. Sohara Experience with programmed steroid treatment with thymectomy in nonthymomatous myasthenia gravis Ann. Thorac. Surg., May 1, 2004; 77(5): 1745 - 1750. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. P. Richman and M. A. Agius Treatment of autoimmune myasthenia gravis Neurology, December 23, 2003; 61(12): 1652 - 1661. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Hill THE NEUROMUSCULAR JUNCTION DISORDERS J. Neurol. Neurosurg. Psychiatry, June 1, 2003; 74(90002): ii32 - 37. [Full Text] [PDF]
|
|
|
|
|
|
P. Granone, S. Margaritora, A. Cesario, and D. Galetta Focus on cosmesis in thymectomy for myasthenia gravis Ann. Thorac. Surg., October 1, 2001; 72(4): 1441 - 1442. [Full Text] [PDF]
|
|
|
|
 |
|
 P. Hjelmström Lymphoid neogenesis: de novo formation of lymphoid tissue in chronic inflammation through expression of homing chemokines J. Leukoc. Biol., March 1, 2001; 69(3): 331 - 339. [Abstract] [Full Text]
|
|
|
|
 |
|
 Thymectomy in Myasthenia Gravis: Problems and Questions Remain Journal Watch Neurology, December 7, 2000; 2000(1207): 7 - 7. [Full Text]
|
|
