Rapid Responses to:
|
|
Rapid Responses: Rapid Responses published:
-
![[Read Rapid Response]](/icons/misc/sectionDOWN.gif)
Generic or Branded
- Matthew J Jackson (28 February 2008)
-
Anti-depressant Meta-Analysis Misleading to Public and of Limited Value to Clinicians
- Fiona Moir, Linda Bryant, Robert Shieff, Elizabeth Robinson (28 February 2008)
-
Differences between antidepressants and placebo
- Francis J Dunne (1 March 2008)
-
Telling the antidepressant story
- D B Double (1 March 2008)
-
What is clinical significance?
- Richard K Morriss (2 March 2008)
-
Overprescribing of medications
- Ruchi Thakur (6 March 2008)
-
Universal Human Subjects Commitment
- David O. Antonuccio, Ph.D., David Healy, M.D., Cardiff University (7 March 2008)
-
How the NASCIS studies gamed the peer-review and regulatory processes
- William P. Coleman (7 March 2008)
-
NASCIS publications abused the processes of peer-review, publication, regulatory approval, and clinical consensus
- Fred H. Geisler (12 March 2008)
-
Demand, don't just ask
- P Ken Gillman (19 March 2008)
-
Rumsfeld misquoted
- Doug Altman (25 March 2008)
-
Methyprednisolone and acute spinal cord injury: redux
- Michael B Bracken (14 May 2008)
-
Corticosteroids could harm patients with spinal cord injury
- Ian Roberts (28 May 2008)
-
Reply to Dr. Bracken
- Fred H. Geisler, William P. Coleman (10 June 2008)
-
Storm in an antidepressant cup
- Riccardo N Caniato (11 December 2008)
|
|
|||
|
Matthew J Jackson, FY2, General Practice Hazeldene Medical Centre, M40 3HD
Send response to journal:
|
Following a National Audit Office report, the media was quick to point out that doctors could save the NHS large sums of money if generic, as opposed to branded drugs were prescribed [1]. However, when reporting the recent meta-analysis of the effect of selective serotonin reuptake inhibitors published in PLoS-Medicine [2], branded as opposed to generic names were used [3]. [1] http://news.bbc.co.uk/1/hi/health/6666805.stm [2] http://medicine.plosjournals.org/perlserv/?request=get- document&doi=10.1371/journal.pmed.0050045 [3] http://news.bbc.co.uk/1/hi/health/7263494.stm Competing interests: None declared |
|||
|
|
|||
|
Fiona Moir, Senior Tutor in Mental Health and Communication Skills, Department of General Practice and Primary Health Care, University of Auckland, 1142, New Zealand., Linda Bryant, Robert Shieff, Elizabeth Robinson
Send response to journal:
|
The paper by Kirsch et al. has certainly generated some much needed discussion. Unfortunately this has been largely through the media and has generated anxiety amongst many people currently taking antidepressants, and those in the future who may require them. The excellent aspect of the meta-analysis is the inclusion of unpublished studies. This is very important. In our view it would however, have been more appropriate to use these pre-approval studies as the foundation, and then add post-approval randomised trials in order to obtain a more complete meta-analysis. Whilst acknowledging that the FDA pre-approval studies should have been strong enough to support registration and marketing of the medicines, post-approval studies should not have been neglected. Inclusion of post-approval studies could have provided data collected over periods of longer than eight weeks, a more clinically useful time-frame, given the nature of depressive illness. Other aspects of the study that we feel should have been more clearly acknowledged as limitations, were the exclusion of a quarter of the original FDA studies, and the high level of imputation of the standard deviation of the change score in a third of the thirty-five studies. By using Last Observation Carried Forward analysis, early drop outs, such as those who withdrew after two weeks, would have skewed the results in favour of placebo, as anti-depressants have a considerable lag-time to peak efficiency, which is generally not the case in placebo treatments. There is little argument that anti-depressants provide minimal benefit in the treatment of mild depression, whereas the skill-based talking therapies have evidentially- based superiority in this patient group. The STAR*D trial shows that moderate to severe depression is hard to treat. Even with appropriate doses utilised for reasonable lengths of time (8 weeks and over), only 36% of patients will show a complete response to the first antidepressant they take. This figure increases by a further 30% with progressively assertive medical treatment of depression. Although this tells us that there is still much to learn about pathways and indicators to best care, and that the antidepressants currently available lack potency for many people who are in great need, it seems premature to conclude, on the basis of Kirsch et als evidence, that antidepressants offer no benefit in the treatment of depression. Finally, the title about antidepressants is very misleading as there were only four antidepressants included in the analysis, of which nefazadone has had limited availability. Whilst fluoxetine, paroxetine, and venlafaxine represent a significant proportion of anti-depressants commonly used, the exclusion of tri-cyclic anti-depressants creates an unhelpful inaccuracy. Perhaps if the title had reflected that the research was about short- term anti-depressant benefits in a specific group of drugs, instead of consisting of inappropriate generalisations, it would not have been responsible for much of the media hype that has been created. Dr Fiona Moir, MBChB, MRCGP Senior Tutor in Mental Health and Communication Skills, Department of General Practice and Primary Health Care, University of Auckland, New Zealand. Dr Linda Bryant MClinPharm, PhD (General Practice), FNZHPA, FNZCP, FPSNZ, MCAPA Clinical Advisory Pharmacist, Department of General Practice and Primary Health Care, University of Auckland, New Zealand. Dr Robert Shieff MBChB, FRANZP Consultant Psychiatrist in Private Practice, Auckland, New Zealand Elizabeth Robinson MSc Section of Epidemiology and Biostatistics School of Population Health, University of Auckland, New Zealand Competing interests: None declared |
|||
|
|
|||
|
Francis J Dunne, Consultant Psychiatrist Mascalls Park, Mascalls Lane, Brentwood, Essex, CM14 5HQ
Send response to journal:
|
Differences between antidepressants and placebo Sir, It seems premature to extrapolate that antidepressants are useful for only severe cases of depression. The promotion of beneficial effects in research papers is of course not limited to antidepressants. Publication bias has been a contributing factor for many years, with the emphasis on positive results. The more reputable journals have sought to eliminate this bias and have openly stated so. The conclusion of a study based on full datasets for a subset of clinical trials from many thousands, should nonetheless raise some scepticism. An interesting paper by Klein1 emphasizes the need for caution in evaluating studies by meta-analysis particularly when the studies differ in many important aspects. Meta- analyses, according to Klein, do not necessarily produce reliable results merely because of the accumulated large sample size. While it is quite probable that antidepressants are over prescribed and not as often effective as claimed, the antithesis does not necessarily follow i.e. that talking therapies are better. Such therapies are even more difficult to research and yet are widely used in clinical settings and by the media (as witnessed by the abundance of 'pop' pychologists/psychiatrists and agony aunts all too ready to pronounce on anything psychological). It would be interesting to carry out an objective evaluation of studies on any one of the hundreds of talking therapies, complete with unpublished data, and compare with placebo. The problem of course is that such a 'placebo might prove meaningless, and furthermore, many benefits of therapy (as in medicine) are actually spontaneous remission effects. Francis J Dunne consultant psychiatrist, Mascalls Park, Mascalls Lane, Brentwood, Essex CM14 5HQ dunnefrancis@googlemail.com Competing interests: none declared 1 Klein DF. Flawed meta-analyses comparing psychotherapy with pharmacotherapy. Am J Psychiatry 2000; 157:1204-11. Competing interests: None declared |
|||
|
|
|||
|
D B Double, Consultant Psychiatrist Norfolk & Waveney Mental Health NHS Foundation Trust, Peddars Centre, Norwich NR6 5BE
Send response to journal:
|
What's new about the Kirsch et al study?1 It's already established that the medical literature represents a selective and biased subset of study outcomes.2 NICE has also already agreed that the size of the difference favouring SSRIs over placebo as measured by the Hamilton Depression Rating Scale is unlikely to be clinically significant.3 And yet the media have widely reported that the new study has found that antidepressants do not work. Doctors do not generally tell patients about the small effect size and substantial non-response rate of antidepressants for fear of undermining the effectiveness of medication. Discontinuing antidepressants should be expected to cause a nocebo effect because the placebo response to treatment is so large. This study raises again the question of whether the small difference between antidepressant and placebo in clinical trials could be totally explained by placebo amplification due to methodological bias in clinical trials, such as unblinding.4 If this were the case, antidepressants are pure placebo. Ignoring this issue reinforces the belief that antidepressants work, to the extent that it has been regarded as malpractice not to prescribe antidepressants in major depression. The debate about whether antidepressants work needs to be legitimised.
Competing interests: None declared |
|||
|
|
|||
|
Richard K Morriss, Professor of Psychiatry University of Nottingham Division of Psychiatry, Queens' Medical Centre, Nottingham, NG7 2UH, UK
Send response to journal:
|
The Kirsch et al (2008) meta-analysis raises questions concerning the definition of clinical significance. To my mind, the term clinical significance should apply only to treatment choices that are made in clinical practice. Placebo drugs are not given to treat depression in clinical practice; they are a tool for establishing efficacy in research but otherwise are irrelevant to questions of clinical and cost effectiveness. The correct design for a randomised controlled trial to establish the clinical effectiveness of antidepressants is to compare an antidepressant treatment given by a doctor with commonly practised treatment not involving antidepressants. Such a design has been employed in the HTA funded THREAD randomised controlled trial of SSRI antidepressant treatment given by general practitioners (GPs) for mild and moderate depression versus consultation by GPs without the initial prescription of antidepressants with a six month folow-up period(1). Trials that examine clinical significance should be of sufficient duration to examine the whole duration of clinical improvement (at least 12 weeks not 6 weeks), and they should consider variance as well as absolute change scores. The NICE definition based on the Hamilton Depression Rating Scale that Kirsch et al (2008) used considered only differences in absolute change scores. The Kirsch et al meta-analysis shows that despite concerns about the provenance of FDA approval and publication bias, there was sufficent evidence to license these four antidepressants. The meta-analysis provides no evidence that antidepressants are clinically ineffective and regretfully patients may have already prematurely stopped antidepressant treatment as a result of the publicity around this highly contentious study. It is time to make the definition of clinical significance as rigorous as the trials to which Kirsch et al applied the NICE definition of clinical significance. 1. Chatwin J, Kendrick T and the THREAD study Group. Protocol for the THREAD (THREshold for AntiDepressants) study: a randomised controlled trial to determine the clinical and cost-effectiveness of antidepressants plus supportive care, versus supportive care alone, for mild to moderate depression in UK general practice. BMC Fam Pract 2007 Jan 4;8:2. Competing interests: Funding received for non-promotional talks from Lilly, Bristol Squibb Myers and Astra Zeneca. Grant funding for research on antidepressants from NHS HTA programme. RM has served in various capacities for NICE. |
|||
|
|
|||
|
Ruchi Thakur, Specialist Registrar ,Old age Psychiatry London
Send response to journal:
|
I believe this new information which has become public is useful in some ways .A lot of people are becoming aware of common psychiatric problems and are running to their GP's for antidepressants expecting a magical wand to lift their mood. The new findings will make some patients think twice before resorting to medications. This will also probably discourage over prescribing of antidepressants. On the other hand depressed patients who are relying on these medicines to make them feel better may be disheartened. Psychologically, it may come as a big blow to them. This news has made every doctor think twice before prescribing antidepressant, at the moment. Its not just antidepressants but one needs to think twice before keeping patients on long term antipsychotic or mood stabilisers. It has been seen that once a patient is put on these medications its v difficult to stop them as it generates anxiety in the patient, the family and even amongst the Community Psychiatric Nurses and junior doctors. Its imperative that every doctor looks at the evidence base but till the information and guidance from Royal College becomes clear one needs to be observing caution in prescribing these medicines. Competing interests: None declared |
|||
|
|
|||
|
David O. Antonuccio, Ph.D., Professor of Psychiatry and Behavioral Sciences at the University of Nevada School of Medicine 401 W. 2nd St., Suite 216, Reno, NV 89503, David Healy, M.D., Cardiff University
Send response to journal:
|
Lenzer and Brownlee (2008) have hit the nail on the head regarding the important issue of data access. We owe all human subjects who volunteer for behavioral and medical research more than they have been getting. For years now, the top scientific journals have required that all clinical trials be publicly registered before data collection begins in order to be eligible for publication. This was an important step designed to reduce publication bias but it did not go far enough. The recent FDA Amendments Act mandating public access to data summaries is another step in the right direction but, as Lenzer and Brownlee (2008) suggest, this too may not go far enough. There have been several examples from the psychopharmacology literature, including some involving the treatment of children, showing that nothing short of total public access to raw human subject data on efficacy and safety will be enough to ensure that data are independently and thoroughly evaluated (Antonuccio, Danton, & McClanahan, 2003; Healy, 2004). Issues of distorted or selective publication (Turner et al., 2008; Kirsch et al., 2008) continue to corrupt our ostensible scientific database. We urge all IRBs to require that, in exchange for the privilege of doing human subject research, researchers make their raw data (not just summaries of the data) accessible (without identifying information) within a reasonable period of time via the internet or in some other suitable fashion. We believe scientists owe this unfettered access to all participants who have ever volunteered for a scientific study with the hope, belief, and promise that their sacrifices would help science advance. We offer the following brief universal commitment to human subjects that can be used by any institutional review board (IRB) in the world. "I agree, in exchange for the privilege of doing research with human subjects, to not only register the trial in a publicly accessible clinical trials database, but also to make summaries of the primary results and the actual raw data internet accessible (without identifying information) within 1 year of collecting data on the last human subject, or within 2 years after the start of the study, whichever is sooner. This is my commitment to all human subjects who volunteer with the hope, expectation and promise that their efforts and sacrifices will result in independently verifiable contributions to science. I recognize that failure to follow through on this commitment may jeopardize approval for any future research protocols in which I may participate." Antonuccio, D.O., Danton, W.O., & McClanahan, T.M. (2003). Psychology in the prescription era: Building a firewall between marketing and science. American Psychologist, 58, 10281043. Healy, D. (2004). Let Them Eat Prozac: The Unhealthy Relationship between the Pharmaceutical Industry and Depression. New York: New York University Press. Kirsch, I., Deacon, B.J., Huedo-Medina, T.B., Scoboria, A., Moore, T.J., & Johnson, B.T. (2008). Initial severity and antidepressant benefits: A meta-analysis of data submitted to the Food and Drug Administration. PLoS Medicine, 5, 260-268. Lenzer, J., & Brownlee, S. (2008). Antidepressants: An untold story? BMJ, doi:10.1136/bmj.39504.662685.0F (published 27 February 2008). Turner, E.H., Matthews, A.M., Linardatos, E., Tell, R.A., & Rosenthal, R. (2008). Selective publication of antidepressant trials and its influence on apparent efficacy. New England Journal of Medicine, 358, 252-260. Competing interests: Competing interests for David Antonuccio: University of Nevada School of Medicine; Reno V.A. Medical Center; author of smoking cessation book Butt Out; past recipient of NIDA and NCI funding for smoking cessation research; past recipient (1995) of funding from Marion Merrill Dow for research on the nicotine patch; workshops on the treatment of depression; expert witness on nicotine dependence, depression or PTSD; private practice. In the past 10 years David Healy has had consultancies with, been a principal investigator or clinical trialist for, been a chairman or speaker at international symposia for or been in receipt of support to attend meetings from Astra-Zeneca, Boots/Knoll Pharmaceuticals, Eli Lilly, Janssen-Cilag, Lorex-Synthelabo, Lundbeck, Organon, Pharmacia & Upjohn, Pierre-Fabre, Pfizer, Rhone-Poulenc, Roche, Sanofi, GlaxoSmithKline, and Solvay. In the past two years, DH has had lecture fees and support to attend meetings from Astra-Zeneca and Lundbeck. In the past ten years DH has been an expert witness for the plaintiff in 15 legal actions involving SSRIs and has been consulted on a number of attempted suicide, suicide and suicide-homicide cases following antidepressant medication, in most of which he has offered the view that the treatment was not involved. He has also been an expert witness in one patent case, and one securities case. |
|||
|
|
|||
|
William P. Coleman, Biostatistical consultant WPCMath, 703 West Ferry Street, C-20, Buffalo NY 14222-1674 USA
Send response to journal:
|
Lenzer and Brownlee's article shows what can happen (1-9) when excessive secrecy is allowed to major studies like NASCIS (10,11), which promotes an off-label use of a steroid for a widespread, yet nonstandard, possibly dangerous treatment for an unapproved indication: spinal cord injury. Normally, it is a bad idea to fabricate retrospective, ad hoc conjectures about properly designed prospective studies. However, NASCIS 2 and NASCIS 3 voided their warranties by finessing the negative results of their primary analyses. For NASCIS 2, efficacy was only in a subgroup (treated < and equal to 8 hours, including 62+67=129 patients out of 487 randomized). NASCIS 3 found efficacy only in the 3-8 hour subgroup, or 71+80=151 patients of 499 randomized. These analyses may (or may not) have been planned in advance--but, to preserve the .05 standard and avoid cherry-picking, only a single analysis for each trial can be "the primary efficacy analysis." Yet both found reviewers and editors of distinguished journals who were prepared to allow over-optimistic wording to be published, which disguised that the trials were basically negative, not positive. These worries are not theoretical. In NASCIS 2, the < and equal to 8 hour placebo group did worse, not only compared to the < and equal to 8 hour steroid group, but also compared to the >8 hour placebo group. So, unless placebo can be made effective by waiting >8 hours, the < and equal to 8 hour placebo group may have been artificially weak. NASCIS 3 had a similar pattern of implausibility due perhaps to a low-dose "control" group with more (24.7% versus 13.9%) inappropriate subjects who had "normal" motor function at baseline and thus were mathematically unable to improve. The hallmark of science is independent reproducibility. Busy reviewers of major journals are not always thoughtful, but other interested analysts could confirm or refute results--if data were openly available. REFERENCES: 1. Coleman, WP, Benzel, E, Cahill, DW, Ducker, T, Geisler, F, Green, B, Gropper, MR, Goffin, J, Madsen, PW, Maiman, DJ, Ondra, SL, Rosner, M, Sasso, R, Trost, G, and Zeidman, S: A Critical Appraisal of the Reporting of the NASCIS II and III Studies of MPSS in Acute Spinal Cord Injury. Journal of Spinal Disorders, 13(3), pp. 185-199, 2000. 2. Hurlbert RJ. Methylprednisolone for acute spinal cord injury: An inappropriate standard of care. J Neurosurg 2000;93:1C7. 3. Nesathurai S. Steroids and spinal cord injury: Revisiting the NASCIS 2 and NASCIS 3 trials. J Trauma Injury Infection Crit Care 1998;45:1088C93. 4. Short DJ, El Masry WS, Jones PW. High dose methylprednisolone in the management of acute spinal cord injury: A systematic review from a clinical perspective. Spinal Cord 2000;38:273C86. 5. Hurlbert RJ. The Role of Steroids in Acute Spinal Cord Injury: An Evidence-Based Analysis. Spine 26 (24S) 2001: S39-S46 6. Hadley MN, Walters BC. Pharmacological Therapy after Acute Cervical Spinal Cord Injury in GUIDELINES FOR THE MANAGEMENT OF ACUTE CERVICAL SPINE AND SPINAL CORD INJURIES. Neurosurgery 50:S63-S72, 2002. 7. Hugenholtz H: Methylprednisolone for acute spinal cord injury: not a standard of care. Cmaj 168:1145-1146, 2003. 8. Hurlbert RJ: Strategies of medical intervention in the management of acute spinal cord injury. Spine 31:S16-21, 2006. 9. Sayer FT, Kronvall E, Nilsson OG: Methylprednisolone treatment in acute spinal cord injury: the myth challenged through a structured analysis of published literature. Spine J 6:335-343, 2006. 10. Bracken MB, Shepard MJ, Collins WF, et al. A randomized, controlled trial of methylprednisolone or naloxone in the treatment of acute spinal-cord injury. Results of the Second National Acute Spinal Cord Injury Study. N Engl J Med 322:1405-1411, 1990. 11. Bracken MB, Shepard MJ, Holford TR, et. al. Administration of methylprednisolone for 24 or 48 hours or tirilazad mesylate for 48 hours in the treatment of acute spinal cord injury. Results of the Third National Acute Spinal Cord Injury Randomized Controlled Trial. National Acute Spinal Cord Injury Study. JAMA 277:1597-1604, 1997. Competing interests: None declared |
|||
|
|
|||
|
Fred H. Geisler, Director Illinois Neuro-Spine Center, 2020 Ogden Ave, Suite 335, Aurora, IL 60504
Send response to journal:
|
I would like to add to Lenzer and Brownlee's reporting of my comments on how excessively closed science can hurt physicians and patients. Dr. Michael Bracken (1) was PI of the NASCIS 2 and 3 studies of high- dose steroids in acute spinal cord injury. NINDS conducted a public campaign in advance of the scientific publication of NASCIS 2 on May 17, 1990. They sent a fax on April 13, 1990 to some 19,000 emergency room physicians and hospitals. This fax was after a NINDS press release had resulted in coverage by the New York Times and the Chicago Tribune on March 31, 1990, by Science News on April 7, 1990, by Newsweek on April 9, 1990. This led to widespread use of steroids, off label. No application for regulatory approval for this indication was completed, and no agency ever approved it. Surgeons report (2) that MP is administered from fear of litigation, not belief in efficacy. Dr. Bracken, a statistician, reinforced this fear by testifying against physicians; he was deposed on June 9, 1998 in Civil Action File No. 96A-7768-6, Superior Court of Fulton County, GA. We have criticized (3) NASCIS science. The later AANS/CNS “Guidelines for the Management of Acute Cervical Spine and Spinal Cord Injuries” rated (4) the NASCIS publications as Evidence Class III, citing flaws in study design, data presentation, interpretation, and analysis. They listed steroid treatment only as an "option." The lack of demonstrated benefit must be weighed against documented risks. The CRASH trial showed (5) a 3% greater mortality when cortico- steroids were given to a multi-trauma group with head injury. If this increased mortality rate held in SCI, then 5000 extra patients may have died in the US since 1990. Yet it's difficult to stop the momentum--especially when primary data are unavailable for independent review. REFERENCES: 1. Bracken MB, Shepard MJ, Collins WF, et al. A randomized, controlled trial of methylprednisolone or naloxone in the treatment of acute spinal-cord injury. Results of the Second National Acute Spinal Cord Injury Study. N Engl J Med 322:1405-1411, 1990. 2. Eck JC, Nachtigall D, Humphreys SC, Hodges SD. Questionnaire survey of spine surgeons on the use of methylprednisolone for acute spinal cord injury. Spine 31:E250-253, 2006. 3. Coleman, WP, Benzel, E, Cahill, DW, Ducker, T, Geisler, F, Green, B, Gropper, MR, Goffin, J, Madsen, PW, Maiman, DJ, Ondra, SL, Rosner, M, Sasso, R, Trost, G, and Zeidman, S: A Critical Appraisal of the Reporting of the NASCIS II and III Studies of MPSS in Acute Spinal Cord Injury. Journal of Spinal Disorders, 13(3), pp. 185-199, 2000. 4. Hadley MN, Walters BC. Pharmacological Therapy after Acute Cervical Spinal Cord Injury in GUIDELINES FOR THE MANAGEMENT OF ACUTE CERVICAL SPINE AND SPINAL CORD INJURIES. Neurosurgery 50:S63-S72, 2002. 5. Edwards P, Arango M, Balica L, et. al. Final results of MRC CRASH, a randomised placebo-controlled trial of intravenous corticosteroid in adults with head injury-outcomes at 6 months. Lancet 365:1957-1959, 2005. Competing interests: None declared |
|||
|
|
|||
|
P Ken Gillman, Retired Qld Aus 4750
Send response to journal:
|
'Trial participants, ... deserve nothing less than ...' Perhaps patient advocacy organisations, on behalf of trial participants, could consider how they can take the initiative and empower themselves. Patients could decline to participate in trials that do not fulfil appropriate criteria and conditions: the boot would soon be seen to be on the other foot. Competing interests: None declared |
|||
|
|
|||
|
Doug Altman, Director Centre for Statistics in Medicine, Oxford OX2 6UD
Send response to journal:
|
This is a comment about the drawing at the top of the article, not the actual content of this important article. For the record, Donald Rumsfeld did not refer to unknown knowns but to known unknowns. There may indeed be things we don't know that we know, but he was referring to things we know that we don't know (before going on to refer to those things we don't know that we don't know). For example, in general we know that we don't know which patients will benefit from a given treatment even when it has been shown to be 'effective'. Competing interests: None declared |
|||
|
|
|||
|
Michael B Bracken, Professor Yale University Medical School, New Haven, CT06510
Send response to journal:
|
By extrapolating from the results of the CRASH trial, Geisler uses your column (letter 22 March 2008) to perpetuate the canard that 5000 additional deaths among patients with acute spinal cord injury (ASCI) may be due to the use of methylprednisolone (MP) in the US since 1990. If the calculation were done correctly the number is <500 but, more importantly, the comparison is bogus. Because of uncertainty about the use of steroids in head injured patients, the NASCIS trials of MP specifically excluded patients with head injury. Any use of MP in ASCI patients with head injury is an unwarranted extension beyond the recommendations from the RCTs. Moreover, the CRASH trial documented no difference in mortality among patients with minor head injuries (1), thereby replicating evidence from the NASCIS trials, and from other therapeutic uses of high dose steroids (2), that also found no significant association of MP use with increased mortality. As to other issues raised in Geislers letter, in 1990 the New England Journal of Medicine lifted their press embargo on the NASCIS 2 paper (3) at the request of the National Institute of Neurological Diseases and Stroke so that physicians treating ASCI could be alerted to the trial result, which was that MP provided some benefit to some patients. The therapy was never claimed to be other than of modest benefit but in the absence of therapeutic alternatives was considered to be an important observation. The study also offered the first evidence of principle: ASCI might be amenable to pharmacologic therapy. Dozens of laboratories around the world continue to actively work on developing neuroprotective therapies for ASCI and other CNS injuries. Geisler states that no agency ever approved the use of MP for ASCI but the regulatory bodies of many countries did so. The critique of the NASCIS trials by Geisler and Coleman has itself been refuted (4) and many neurosurgeons consider the guideline prepared by Hadley at el to be fundamentally flawed in its consideration of MP (5). Contrary to the impression left by Geisler, many neurosurgeons do support the use of MP for ASCI. Some commentators have criticized the non-use of steroids as being a failure to practice evidence-based medicine (6) and others suggest it is unethical not to use MP and wonder why the bar is set so high for MP when so many neurosurgical interventions are devoid of any reliable empirical evidence of efficacy (7). Regrettably, there have been no additional trials for many years but the totality of randomized evidence continues to favor using MP under the parameters established by the trials (8). Recent work with high resolution imaging has found that patients administered MP have reduced spinal cord edema (9), offering support for one mechanism initially hypothesized for MP. Carefully conducted but non-randomized studies continue to offer evidence supporting clinical use of MP (10). The need for protocol registration of RCTs and the need for protocols prior to undertaking a systematic review of studies are both well recognized but Lenzer and Brownlee (BMJ March 8 2008) fail to understand that the same imperative applies to requests for trial data. To avoid post hoc fishing expeditions, secondary analyses of large trial data sets should follow a protocol that sets out the planned use of the data. Lenzer and Geisler were both refused access to NASCIS data sets because they failed to provide a protocol. Under conditions established by a Yale Medical School IRB, NASCIS data sets have been available for many years, and have been provided to independent researchers. Michael B. Bracken,
1. CRASH trial collaborators. Effect of intravenous corticosteroids on death within 14 days in 10008 adults with clinically significant head injury (MRC CRASH trial): randomized placebo-controlled trial. Lancet 2004; 364: 1321-1328 2. Sauerland S, Nagelschmidt M, Mallman P, Neugebauer EA. Risks and benefits of preoperative high dose methylprednisolone in surgical patients: a systematic review. Drug Safety 2000; 55: 452-61 3. Bracken MB, Shepard MJ, Collins WF, Holford TR, Young W, Baskin DS, Eisenberg HM, et al. A Randomized, controlled trial of methylprednisolone or naloxone in the treatment of acute spinal cord injury: Results of the Second National Acute Spinal Cord Injury Study. New Engl J Med 1990; 322:1405 11. 4. Bracken MB, Aldrich EF, Herr DL, Hitchon PW, Holford TR, Marshall LF, et al. Clinical measurement, statistical analysis and risk-benefit: controversies from trials of spinal injury. J Trauma Inj Infect Crit Care 2000; 48:558-66. 5. Bracken MB, Eisenberg H, Herr D, Hitchon PW, Holford TR, Marshall LF, et al. Commentary on: Guidelines for Management of Acute Cervical Spinal Injuries. Neurosurgery 2002; 50: S xiv-S xix, 6 Molloy S, Middleton F, Casey AT. Failure to administer methylprednisolone for acute traumatic spinal cord injury - a prospective audit of 100 patients from a regional spinal injuries unit. Injury 2002; 33: 575-8. 7. Rozet I. Methylprednisolone in acute spinal cord injury: is there any other ethical choice? J Neurosurg Anesthesiol 2008; 20:137-9. 8. Bracken MB. Steroids for acute spinal cord injury. Cochrane Database of Systematic Reviews 2002, Issue 2 9. Leypold BG, Flanders AE, Schwartz ED, Burns AS. The impact of methyl-prednisolone on lesion severity following spinal cord injury. Spine 2007; 32: 373-8 10. Tsutsumi S, Ueta T, Shiba K, Yamamoto S, Takagishi K. Effects of the Second National Acute Spinal Cord Injury Study of high-dose methylprednisolone therapy on acute cervical spinal cord injury-results in spinal injuries center. Spine 2006; 31: 2992-6 Competing interests: MBB has a consulting agreement with Pfizer on an unrelated matter, was Principal Investigator of the NIH funded NASCIS trials and on the planning committee for the MRC funded CRASH trial. |
|||
|
|
|||
|
Ian Roberts, Professor of Epidemiology and Clinical Co-ordinator of the CRASH Trial London School of Hygiene & Tropical Medicine
Send response to journal:
|
Editor - The results of the MRC CRASH Trial of corticosteroids in patients with traumatic brain injury, which showed evidence of harm rather than benefit, should call into question the use of corticosteroids in patients with spinal cord injury. Indeed, I believe that it is a shame and an injustice that pateints with spinal cord injury continue to be treated with corticosteroids despite this important uncertainty. I disagree with the view of Professor Bracken that corticosteroids showed no evidence of harm in patients with mild head injury. I believe that the correct point of reference for a sub-group analysis in a clinical trial is the main effect from the whole trial rather than the null. In the CRASH trial, the sub-group analysis of the effect of corticosteroids according to the severity of head injury provided no evidence that the subgroup effects differed from the main effect which showed harm - other than by chance. To protect current and future patients from the possibility of harm, corticosteroids should only be used in patients with spinal cord injury in the context of a rigorously conducted and adequately powered randomised controlled trial. Competing interests: None declared |
|||
|
|
|||
|
Fred H. Geisler, Director Illinois Neuro-Spine Center, 2020 Ogden Ave, Suite 335, Aurora, IL 60504, William P. Coleman
Send response to journal:
|
In his response to our letters, Dr. Bracken evades questions and criticisms by diverting to ones not mentioned, and he ignores important ones altogether. In this letter, we itemize specific instances where such problems can be factually checked against original sources. We did not ask whether the NIH and NASCIS technically violated the New England Journals embargo; rather, we questioned their using, before scientific publication and open debate and without FDA approval, an extra- scientific campaign that created public opinion and legal coercion for methylprednisolone (MP) in acute spinal cord injury (SCI). Dr. Bracken implies that the Malloy study (1) provides evidence that not using MP in SCI is malpractice. This is not true. Malloy et. al. assume this idea -- based on NASCIS -- without giving new evidence. In their own words, their objective was only to determine the actual use and correct implementation of the NASCIS protocols. Dr. Bracken cites Dr. Rozets commentary (2) that MP is the ethical choice in SCI, but the summary in his letter is only an approximate reflection of her conclusions. Does Dr. Bracken agree with her actual words? She states: Unfortunately, neither NASCIS II nor NASCIS III trials reveal level-1 evidence of the beneficial effect of methylprednisolone in SCI. This is the conclusion we ourselves have published (3) and that Dr. Bracken has refuted. It is the same in the AANS/CNS guidelines (4) that Dr. Bracken calls fundamentally flawed. Dr. Rozet does feel (2) that pragmatically, if not scientifically, MP should be used: Let us not forget that even minimal improvement in control of sphincters or motor function of fingers, reported by NASCIS trials, can have a crucial impact on the patients lifestyle. . . . The dilemma of choosing a potentially beneficial drug, versus nothing, is pretty straightforward in favor of the drug. This is a fair conclusion from the evidence for and against MP that she discusses. However, she does not consider two other issues. First, there are many treatments besides MP that also have evidence (4,5,6,7) we might now regard as similarly reasonable but similarly inconclusive and for which the same argument might thereforebe made. Second, her reasoning assumes MP is safe and that its benefits therefore outweigh its risks. We turn to Dr. Brackens comments on the claimed benefits and risks of MP in SCI. As far as benefits, Dr. Bracken simply ignores our criticisms that -- for both NASCIS II and NASCIS III -- his primary analysis was negative. Efficacy was only in a secondary analysis in a subgroup. Dr. Bracken defended (8) those secondary analyses as apriori, but even were this true in some unspecified sense it misses the point that, to maintain the integrity of the p-value, only one analysis can be primary. If multiple analyses are to be combined, then details should formalized in the pre-trial protocol, explained in the post - trial publication, and normally supervised by the FDA. In NASCIS II and III, the efficacy claims, in addition to being modest, are also suspect. As far as safety, there are three important issues with Dr. Brackens letter. First, there seems to be an inconsistency about the findings of the CRASH study. Dr. Bracken did write a paper (9) in which he states: patients with mild head injuries were not at significantly increased risk of death in CRASH (4.2% v 4.1%, a difference that could be due to chance.) How are we to reconcile this with the publication (10) by the CRASH trial collaborators, the one Dr. Bracken cites in his letter? That paper appears to contradict him: The relative increase in deaths due to corticosteroids did not differ by injury severity (p=0.22) . . . Second, Dr. Brackens letter states, [a]ny use of MP in ASCI patients with head injury is an unwarranted extension beyond the recommendations from the RCTs. So, what actual recommendations were published? a. The NIH NASCIS II faxes to emergency rooms make no reference to head injury, although one (April 13, 1990) does detail the studys exclusion criteria. b. The subsequent NASCIS II publication (11) was equally without reservations: We conclude that in patients with acute spinal-cord injury, treatment with methylprednisolone in the dose used in this study improves neurologic recovery when the medication is given in the first eight hours. c. The final paragraph of the NASCIS III publication (12) also has no shade of doubt, complexity, or qualification in prescribing MP regimens on which patients should be maintained, varying with time of treatment. d. In his Cochrane Review article (8), Dr. Bracken states categorically that MP has been shown to enhance sustained neurologic recovery in a phase three randomized trial, and to have been replicated in a second trial. These four references supply no caution against using MP in patients with head injury. They also give little hint of the position stated in Dr. Brackens current letter: [t]he therapy was never claimed to be other than of modest benefit. Third, Dr. Bracken argues that the potential number of deaths (5000) we suggested is too high; he estimates it may be more like 500. We are not convinced. However, we do point out that a loss of 500 patients is certainly very large and is questionable for the sake of a treatment whose benefits Dr. Bracken himself now describes as modest and the evidence for which is considered by many -- including the sources we have cited, including the AANS/CNS Guidelines panel, and including a source (2) that Dr. Bracken himself cites as supportive -- to be less than completely scientific. These reservations are reinforced by fact that MP has no FDA indication in SCI and by the existence of several other treatment options (4,5,6,7) that, like MP, are both promising and not completely proven. If the FDA had been involved, then the efficacy claims and the safety restrictions for MP in SCI would be clearer. There should have been public debate and consensus. Instead there is secrecy and inconsistency, covered by rhetoric from Dr. Bracken ("perpetuate the canard," "fundamentally flawed"). Dr. Bracken's letter warns us that post hoc analyses are unscientific, yet the publicized efficacy results from NASCIS II and III were themselves in subgroups and cannot have been the primary prospective analysis. Dr. Bracken's letter extrapolates safety results from a subgroup of CRASH, and as evidence of MP safety in SCI he cites an article (13) about elective surgery in which a word search reveals no occurrence of "head," "brain," or any form of "spine." Dr. Bracken, a statistician not trained in medicine, warns against "malpractice," yet the evidence he cites is primarily dependent on his own study -- and he has testified legally against physicians and has accepted money for doing so. We do not know whether 5000 deaths have occurred -- or how many. Our complaint was, and is, that there is no way to use the NASCIS publications or data reliably to guess the safety or the effect of MP -- because of the many conflicts and uncertainties in the published reports, and because of the secrecy. Our own clinical trial (of Sygen GM-1) required the NASCIS II regimen of MPSS, so we have no bias against it. We have no personal stake, one way or the other, in the efficacy or safety of MP -- except that, as members of the SCI research community, we want to see progress against this disease. (Along with several other groups, we contributed data and new analyses to a scientifically open international effort (14) to enable new SCI trials.) We ourselves, although we have published reviews and letters criticizing NASCIS, have never written that MPSS is either unsafe or ineffective; what we have consistently written is that we have not been given enough believable evidence to know one way or the other. And that is the subject of this correspondence: secrecy, and the private judgement of results. REFERENCES (1) Molloy S, Middleton F, Casey AT. Failure to administer methylprednisolone for acute traumatic spinal cord injury - a prospective audit of 100 patients from a regional spinal injuries unit. Injury 2002; 33: 575-8. (2) Rozet I. Methylprednisolone in acute spinal cord injury: is there any other ethical choice? J Neurosurg Anesthesiol 2008; 20:137-9. (3) Coleman, WP, Benzel, E, Cahill, DW, Ducker, T, Geisler, F, Green, B, Gropper, MR, Goffin, J, Madsen, PW, Maiman, DJ, Ondra, SL, Rosner, M, Sasso, R, Trost, G, and Zeidman, S: A Critical Appraisal of the Reporting of the NASCIS II and III Studies of MPSS in Acute Spinal Cord Injury. Journal of Spinal Disorders, 13(3), pp. 185-199, 2000. (4) Hadley MN, Walters BC. Pharmacological Therapy after Acute Cervical Spinal Cord Injury in GUIDELINES FOR THE MANAGEMENT OF ACUTE CERVICAL SPINE AND SPINAL CORD INJURIES. Neurosurgery 50:S63-S72, 2002. (5) Steeves, et. al. Experimental Treatments for Spinal Cord Injury. International Campaign for Cures of spinal cord injury Paralysis (ICCP), Vancouver, 2007. (6) Baptiste DC, Fehlings MG. Update on the treatment of spinal cord injury. Prog Brain Res 2007;161:217-33. (7) Geisler F. Past and current human spinal cord injury drug trials. In: Tator CH BE, editor. Contemporary Management of Spinal Cord Injury: From Impact to Rehabilitation. Park Ridge: American Association of Neurological Surgeons Publications Committee, 2000:317-333. (8) Bracken MB. Steroids for acute spinal cord injury. Cochrane Database of Systematic Reviews 2002, Issue 2 (9) Bracken MB. CRASH (CORTICOSTEROID RANDOMIZATION AFTER SIGNIFICANT HEAD INJURY TRIAL): Landmark and storm warning. Neurosurgery 57:1300-1302, 2005 (10) CRASH trial collaborators. Effect of intravenous corticosteroids on death within 14 days in 10008 adults with clinically significant head injury (MRC CRASH trial): randomized placebo-controlled trial. Lancet 2004; 364: 1321- 1328 (11) Bracken MB, Shepard MJ, Collins WF, et al. A randomized, controlled trial of methylprednisolone or naloxone in the treatment of acute spinal-cord injury. Results of the Second National Acute Spinal Cord Injury Study. N Engl J Med 322:1405-1411, 1990. (12) Bracken MB, Shepard MJ, Holford TR, et. al. Administration of methylprednisolone for 24 or 48 hours or tirilazad mesylate for 48 hours in the treatment of acute spinal cord injury. Results of the Third National Acute Spinal Cord Injury Randomized Controlled Trial. National Acute Spinal Cord Injury Study. JAMA 277:1597-1604, 1997. (13) Sauerland S, Nagelschmidt M, Mallman P, Neugebauer EA. Risks and benefits of preoperative high dose methylprednisolone in surgical patients: a systematic review. Drug Safety 2000; 55: 452-61 (14) James W. Fawcett, Armin Curt, William P. Coleman, John D. Steeves, Mark Tuszynski, Dan Lammertse, Perry Bartlett, Andrew Blight, Volker Dietz, John Ditunno, Bruce Dobkin, Leif Havton, Peter Ellaway, Michael Fehlings, Manuel Gaviria, Robert Grossman, James Guest, Leif Havton, Naomi Kleitman, Masaya Nakamura, Alain Privat. Guidelines for the conduct of clinical trials for spinal cord injury (SCI) as developed by the ICCP Panel: Spontaneous recovery after spinal cord injury and statistical power needed for therapeutic clinical trials. Spinal Cord 45, 190 - 205 (01 Mar 2007) Competing interests: None declared |
|||
|
|
|||
|
Riccardo N Caniato, Clinical director of Northern Area Secure & Forensic Mental Health Service Townsville, Queensland, Australia
Send response to journal:
|
Dear Sir, I was intrigued to read the paper by Kirsh et al and I could not help thinking that I'd read this all before. My mind reflected on the first time I read a paper by Dr Joanna Moncrieff [1]. Concerns about the validity of antidepressant efficacy date back to the early 60's [2], and since that time there have been a steady stream of such papers, which it appears to me, journals publish from time to time only to allow psychiatrists to have heated conversations at drug company sponsored dinners. The much maligned pharmaceutical industry has always taken these subversive criticisms with grace and good humour, and continued to do its excellent marketing work. Last year Britain alone spent over 3.3 billion pounds on antidepressants [3], an overwhelming vote of confidence in these superlative medications. I accept that there is no evidence that antidepressants, despite their extensive use, have made any impact on the prevalence, severity or natural history of depressive illness in any naturalistic setting. Rather than being a negative, this highlights the resilience and inginuity of pharmaceutical marketing, in not allowing scientific evidence to unduly impact on sales data. The furour over Kirsh et al's paper will quickly pass, the mainstream papers will return to publishing articles on Princess Diana, and antidepressants script will continue to be written by doctors who have more important things to worry about than scientific rigour. 1. Moncrieff, J., Wessely, S. & Hardy,R. (1998). Meta-analysis of trial comparing antidepressants with active placebos. British Journal of Psychiatry, 172, 227-295 2. Thompson, R. (1982). side effects and placebo amplification. British Journal of Psychiatry, 140, 64-68 3. Stratton, A. (2008).A bitter pill. Guardian.co.uk. Tuesday, Feb 26. Competing interests: Dr Caniato has recieved research funds from Novartis, the makers of clozapine. Dr Caniato has recieve honoraria for educational presentations from Jansen-Ciliag and Sanofi-Synthelabo. |
|||