Malcolmvandenburg.co.uk

Editorials represent the opinions of the authors and not necessarily those of the BMJ or BMA eDITORIaLs
For the full versions of these articles see bmj.com Risk of suicidal behaviour in adults taking antidepressants
Increased risk is probably restricted to younger people and varies greatly
between individual medicines
ReseaRch, p 431
Antidepressant drugs currently carry warnings of the it used individual patient data from the trials, and sui- possibility of increased suicidal ideation and behaviour cidal events were reclassified according to a common John Richard Geddes professor
during treatment, especially in younger patients. In system to increase the reliability of the results.7 8 the linked meta-analysis, Stone and colleagues report None the less, important limitations remain because on the possible link between the risk of suicide and of the characteristics of the primary trials. A standard antidepressants using data on individual patients from exclusion in placebo controlled trials of antidepressant corrado Barbui lecturer in
placebo controlled trials.1 This analysis of 372 placebo drugs is that severely ill patients, especially those who controlled antidepressant trials and nearly 100 000 are actively suicidal, are not enrolled. This probably patients found that the association between antide- leads to very low numbers of completed suicides in pressant drugs and the incidence of reported suicidal these trials. If such trials aim to provide evidence of andrea cipriani lecturer in
behaviour is strongly related to age. The risk was raised the clinical effects of the investigational drug, then this in people under 25, not affected in those aged 25-64, exclusion is as clinically illogical as excluding patients and reduced in those aged 65 and older. The analysis with a high risk of mortality in trials in oncology or car- also found differences in risk between drugs. diology. It makes it impossible to estimate the potential This analysis is not new—it was published fully on the benefits of a reduction in baseline suicidality.9 Food and Drug Administration (FDA) website more Furthermore, the low event rate of completed suicide than two years ago.2 It was widely covered at the time means that retrospective analyses have to broaden the in the international medical press and led to warnings definition of suicide beyond completed suicides to gain being included on datasheets.3-6 Because the analysis sufficient statistical power. Regardless of how much Cite this as: BMJ 2009;339:b3066doi: 10.1136/bmj.b3066 has not been updated since the initial publication and effort is put into developing standardised reclassifica- the present report selectively reports the full analysis, tions, the fundamental uncertainty about the validity it raises the question of why it is being published in and meaning of a composite outcome that was not the BMJ now, more than two years later. Its objective prespecified in the primary trials remains. is to make a summary of these important results more The review procedures showed some lack of trans- widely available in a way similar to the publication in parency, as sometimes happens in analyses conducted the BMJ of summaries of Cochrane reviews. by regulatory authorities.10 It is unclear why optimal Other meta-analyses had already been performed, methods of meta-analysis of systematic review—for but this analysis was a methodological advance because example, prior pre-review and publication of the pro-tocol, unselective reporting of the outcomes—were not used. One way of ensuring adherence to currently optimal guidelines for systematic reviews would have been to conduct the analysis under the auspices of the Cochrane Collaboration. Although meta-analyses of individual patient data could usefully look at important clinical outcomes other than suicidality, the Cochrane database still contains few meta-analyses of individual patient data. Could it be that companies are willing to release individual patient data only when required to by regu- latory agencies who grant the marketing authorisations of drugs? If that is the reality, then the true collabo- ration between regulators and other agencies, which seems to be the FDA’s new aim, could be a power- ful approach to synthesising clinical knowledge.11 In particular, the age related decrease in risk for suicide, which seems to be inversely paralleled by increasing efficacy with age, could be investigated further with individual patient data from these trials.5 eDITORIaLs
sis found sertraline and escitalopram to have the best balance of short term efficacy and tolerability. Taking the results of the analyses together reinforces the view that sertraline has a highly favourable profile in terms of efficacy, acceptability, and safety. Although differ-ent mechanisms might lead to clinical relief of symp-toms and increased suicidality (perhaps via increased agitation), a more likely mechanism for the effects of sertraline is that it is simply better tolerated and more likely to be effective, hence reducing both depressive symptoms and suicidality.
1 Stone MB, Laughren T, Jones ML, Levenson M, Holland PC, Hughes A, et al. Risk of suicidality in clinical trials of antidepressants in adults: analysis of proprietary data submitted to US Food and Drug Administration. BMJ 2009;339:b2880.
Laughren TP. Briefing document for December 13 meeting of Psychopharmacologic Drugs Advisory Committee. 2006. www.fda.
gov/ohrms/dockets/ac/06/briefing/2006-4272b1-01-fda.pdf.
Friedman RA, Leon AC. Expanding the black box—depression, antidepressants, and the risk of suicide. N Engl J Med Kuehn BM. FDA panel seeks to balance risks in warnings for antidepressants. JAMA 2007;297:573-4.
Finally, we should consider these results alongside other Barbui C, Cipriani A, Geddes JR. Antidepressants and suicide symptoms: compelling new insights from the FDA’s analysis of recent evidence on antidepressants in major depression. individual patient level data. Evid Based Ment Health 2008;11:34-5.
Although this report focuses on age related differences in Leon AC. The revised warning for antidepressants and suicidality: unveiling the black box of statistical analyses. Am J Psychiatry the risk of suicidal behaviour, individual drugs seem to show some important differences. The odds of suicidal Gunnell D, Saperia J, Ashby D. Selective serotonin reuptake inhibitors behaviour on sertraline, for example, is around half that (SSRIs) and suicide in adults: meta-analysis of drug company data from placebo controlled, randomised controlled trials submitted to on placebo. In comparison, citalopram and escitalopram the MHRA’s safety review. BMJ 2005;330:385.
seem to increase the risk of suicidal events. Fergusson D, Doucette S, Glass KC, Shapiro S, Healy D, Hebert P, et Unfortunately, the analysis did not include indirect al. Association between suicide attempts and selective serotonin reuptake inhibitors: systematic review of randomised controlled comparisons (which would have been possible by vir- tue of the common placebo comparator) when compar- Cipriani A, Barbui C, Geddes JR. Suicide, depression, and antidepressants. BMJ 2005;330:373-4.
ing drugs, so that any conclusions about the differential 10 Vitry A, Lexchin J, Sasich L, Dupin-Spriet T, Reed T, Bertele V, et al. effects of treatments must be made with caution. None Provision of information on regulatory authorities’ websites. Intern the less, it is becoming apparent that antidepressants 11 Hamburg MA, Sharfstein JM. The FDA as a public health agency. N vary in both their efficacy and adverse effects. A recent multiple treatments meta-analysis that compared the 12 Cipriani A, Furukawa TA, Salanti G, Geddes JR, Higgins JP, Churchill R, et al. Comparative efficacy and acceptability of 12 new-generation efficacy and acceptability of antidepressants showed antidepressants: a multiple-treatments meta-analysis. Lancet meaningful differences between drugs.12 That analy- Diagnosis of venous thromboembolism
D-dimer tests can help management but cannot replace clinical judgment
ReseaRch, p 450
Because the signs and symptoms of deep venous hour, and point of care tests produce results within thrombosis and pulmonary embolism are common Pierre-Marie Roy professor,
but non-specific, they often present a diagnostic chal- In the linked systematic review and meta- lenge. Both underdiagnosis and overdiagnosis are analysis, Geersing and colleagues analysed the associated with substantial morbidity and mortality.
diagnostic performances of several qualitative and D-dimers are fibrin degradation products result- quantitative D-dimer tests used at the point of care.1 ing from endogenous fibrinolysis associated with They found that quantitative tests perform better intravascular thrombosis. A non-specific increase in than qualitative ones, but that the number of studies D-dimer concentration is seen in many situations, was limited. Their results also confirmed the value precluding its use for diagnosing venous thromboem- of a negative D-dimer result in excluding a diagnosis bolism (VTE). However, a low D-dimer concentra- of VTE and pulmonary embolism, but they make tion is thought to rule out the presence of circulating fibrin and therefore VTE. Early enzyme linked Firstly, in point of care testing, as in the labora- immunosorbent assay D-dimer tests took a long tory, diagnostic performance depends on the assay time to do, limiting their usefulness in acute care. technology.2-4 Secondly, some tests are still imprecise. Cite this as: BMJ 2009;339:b2799doi: 10.1136/bmj.b2799 Second generation assays provide results within an In particular, quantitative tests used at the point of eDITORIaLs
care have been poorly evaluated in patients with meta-analysis.1 Moreover, in a French national obser- suspected pulmonary embolism. More importantly, vational study, doctors ruled out pulmonary embolism none of these tests reliably ruled out VTE without in 57% of cases on the basis of inappropriate criteria, taking into account the clinical probability of the exposing patients to a high risk of recurrent VTE. One disease. The clinician’s estimate of the pretest prob- of the most common reasons for inappropriate testing ability of a target disorder is a crucial determinant of was the lack of evaluation of clinical probability.6 the direction and extent of the diagnostic work-up.
So how do Geersing and colleagues’ results trans- The authors used Bayes’s theorem to calculate the late into current practice? We have to follow some probability of VTE, conditioned by the likelihood evidence based rules: to use tests with confirmed ratio as a function of the pretest probability. For this diagnostic performance; to consider different diag- purpose, they assumed a test threshold probability of noses and their clinical probabilities before perform- 2%, below which further testing was not warranted. ing any test; and to perform tests that will lead to a They found that for all tests apart from the Cardiac post-test probability low enough to rule out VTE D-dimer test, pretest probability had to be below if the result is negative or high enough to diagnose 8-10% to rule out VTE with confidence when point VTE if the result is positive. Several tools can help of care D-dimer testing was negative.
to achieve these aims, such as the PERC (pulmo- Point of care D-dimer tests are particularly use- nary embolism rule-out criteria) rule, which can help ful for doctors who need rapid information while decide who to test7; a clinical probability score that on the move. Negative results may eliminate the defines pretest probability more accurately8 9; and the need for further diagnostic testing in almost 30% of diagram of Fagan, which can use the likelihood ratios patients with suspected VTE. However, in day to of the tests to calculate post-test probabilities.10 day practice, such easy tests carry some risks too; Finally, the effect and cost of point of care D-dimer for example, D-dimer tests are sometimes ordered in tests need to be evaluated in a randomised controlled patients with an obvious explanation for their signs cluster trial in which primary care doctors or emer- gency departments are provided or not provided In the best case scenario the D-dimer test will be with point of care D-dimer test facilities. However, negative with just the loss of a little time and money, one of the key points will be how doctors will apply but in the worst case scenario, a positive D-dimer test Bayesian reasoning in day to day clinical practice. will prompt the doctor to order further testing, such Computer based clinical decision support systems as leg vein ultrasonography or computed tomography are a promising tool in such complex medical situ- (or both), which carry risks of iatrogenic events and ations, and they may become another useful device false positive results. The decreasing prevalence of for the point of care diagnosis of VTE.11 cases in the diagnostic studies published during the Geersing GJ, Janssen KJM, Oudega R, Bax L, Hoes AW, Reitsma past decades illustrates the evolution of the implicit JB, et al. Excluding venous thrombo-embolism using point of care D-dimer tests in outpatients: diagnostic meta-analysis. BMJ threshold used by doctors when ordering tests.5 Of note, this prevalence was as low as 3-4% in the more Stein PD, Hull RD, Patel KC, Olson RE, Ghali WA, Brant R, et recent studies included in Geersing and colleagues’ al. D-dimer for the exclusion of acute venous thrombosis and pulmonary embolism: a systematic review. Ann Intern Med 2004;140:589-602.
Roy PM, Colombet I, Durieux P, Chatellier G, Sors H, Meyer G. Systematic review and meta-analysis of strategies for the diagnosis of suspected pulmonary embolism. BMJ 2005;331:259.
Di Nisio M, Squizzato A, Rutjes AW, Buller HR, Zwinderman AH, Bossuyt PM. Diagnostic accuracy of D-dimer test for exclusion of venous thromboembolism: a systematic review. J Thromb Haemost 2007;5:296-304.
Le Gal G, Bounameaux H. Diagnosing pulmonary embolism: running after the decreasing prevalence of cases among suspected patients. J Thromb Haemost 2004;2:1244-6.
Roy PM, Meyer G, Vielle B, Le Gall C, Verschuren F, Carpentier F, et al. Appropriateness of diagnostic management and outcomes of suspected pulmonary embolism. Ann Intern Med 2006;144:157-64.
Kline JA, Courtney DM, Kabrhel C, Moore CL, Smithline HA, Plewa MC, et al. Prospective multicenter evaluation of the pulmonary embolism rule-out criteria. J Thromb Haemost 2008;6:772-80.
Wells PS, Anderson DR, Rodger M, Forgie M, Kearon C, Dreyer J, et al. Evaluation of D-dimer in the diagnosis of suspected deep-vein thrombosis. N Engl J Med 2003;349:1227-35.
Righini M, Le Gal G, Aujesky D, Roy PM, Sanchez O, Verschuren F, et al. Diagnosis of pulmonary embolism by multidetector CT alone or combined with venous ultrasonography of the leg: a randomised non-inferiority trial. Lancet 2008;371:1343-52.
10 Richardson S, Wilson M, Guyatt G. The process of diagnosis. In: Guyatt G, Rennie D, eds. Users’ guides to the medical literature. A manual for evidence-based clinical practice. Chicago: AMA Press; 2002.
11 Garg AX, Adhikari NK, McDonald H, Rosas-Arellano MP, Devereaux PJ, Beyene J, et al. Effects of computerized clinical decision support systems on practitioner performance and patient outcomes: a systematic review. JAMA 2005;293:1223-38.
eDITORIaLs
BMJ pico for original research in the print BMJ
We are now rolling this out for all research
This year we have been asking authors to use a new will continue to carry a reminder like this: “This is evidence abstract called BMJ pico to abridge their a summary of a paper that was published on bmj.
original research articles, with the aim of making com: BMJ 2009;000:b000.” BMJ pico will not affect research more readable and useful for print readers, citation or impact factor calculation either, because and particularly for busy clinicians.1 We are delighted both these activities relate only to online publication that so many authors have volunteered to pilot this new format successfully, and from now on we will be Some authors may be concerned that the research adopting it for all newly accepted research articles. By section of the print BMJ will seem less substantial once January 2010 the entire research section of the print it contains only BMJ picos, or that we will pad it out journal will comprise BMJ picos.
with additional commentaries occupying more space BMJ pico has gone down well with readers and than the original research. On the contrary, research authors during the pilot phase, as rapid responses, will remain one of the most important elements of the Trish Groves deputy editor, BMJ,
other feedback, and formal market research attest. print BMJ, and we will work hard to ensure that readers Cross sectional surveys of authors, including those can easily find the research and appreciate its impor- who took part in the pilot phase, show considerable tance. Meanwhile on bmj.com we will continue to sara schroter senior researcher,
support for the concept of BMJ pico and its ability to develop innovative ways to help authors explain their convey the key aspects of a research study to general research and maximise its breadth, depth, transpar- readers (see http://resources.bmj.com/bmj/authors/ ency, and accessibility, often with added features such article-submission/bmj-pico-of-pico-surveys). Around as podcasts, videos, and additional data. We may also two thirds of responders in both surveys said that this explore the possibility of using the pico abstract within publishing model would make them more likely to sub- the full online version of each research article instead mit to the BMJ or would at least make no difference to of the current style of structured abstract. We would welcome readers’ and authors’ views on this.
BMJ pico is a succinct format that clearly states the If you intend to submit your research to the BMJ research question and answer, sums up the key scien- please note that there is no need to prepare a BMJ tific points, and includes prominent statements about pico before submission. We will ask for your pico later funding and any competing interests. It is similar to the in the peer review process, if and when we provi- abstracts published in evidence based journals but adds sionally accept your full article and ask you to revise a small figure or a table that gives, when appropriate, it. You may want to check our resources for authors relative risks and odds ratios.2 Authors retain control of on bmj.com before submitting your work, not least their work, writing their own BMJ picos using a range of because there may be something new there—we con- templates that we have developed with the help of our tinuously develop and update our advice and policies statistics and clinical epidemiology editors (see http:// on research and our services to authors.
resources.bmj.com/bmj/authors/article-submission/bmj- We are very grateful to everyone who has helped us test BMJ pico and those who have told us what Abridging research for the print BMJ has no bearing they think of it. One “piconeer,” Tom Jefferson, clearly on what counts as publication. Authors who do not enjoyed the experience, “I wrote the abstract using wish to abridge their research articles using BMJ pico a mixture of cut and paste from the main text of the may opt instead for online only publication.
article, edits, and rewrites of sections. My personal The BMJ is a fully online journal with new content bias is that I love the discipline of summarising and published every day on bmj.com, and hence it is the abstracting, as it teaches you to identify what is vital full online articles that are the definitive publications: and what is not. The pico format seems OK. The table unlike some other journals with websites the BMJ summarising the main results was a joy to construct as does not use interim “epublication ahead of print.”3 it gave me the chance to tell the story straight.”4 And We select and abridge online content to make up a reader cheered us with this rapid response: “I was each weekly print issue, and when that is published amused to see that the ‘pico’ version of this article every Friday the PDF (portable document format) in the print journal was accompanied by an editorial files of the print sections appear on bmj.com linked to their full online versions. BMJ pico changes none Groves T, Godlee F. Innovations in publishing BMJ research. BMJ of this. All BMJ research articles will continue to be Schwartz LM, Woloshin S, Dvorin EL, Welch HG. Ratio measures published first on bmj.com with free open access to in leading medical journals: structured review of accessibility of the full text, no word limit, no charges to authors, underlying absolute risks. BMJ 2006;333:1248.
Godlee F, Delamothe T, Smith J. Continuous publication. BMJ and immediate transfer to PubMed Central, the US National Library of Medicine’s full text archive. Each Jefferson T. Reflections of a piconeer. 2009. http://bmj.com/cgi/ online article will still be cited by its unique identifier Brandon HA. Small things. 2009. http://bmj.com/cgi/eletters/338/ (BMJ [year];[volume]:[elocator]), and each BMJ pico

Source: http://malcolmvandenburg.co.uk/pdf/BMJ_Risk-of-suicidal-behaviour_August-2009.pdf