Rapid Responses to:
|
|
Rapid Responses: Rapid Responses published:
-
![[Read Rapid Response]](/icons/misc/sectionDOWN.gif)
Antithrombin III in critically ill patients
- Arash Afshari, Jørn Wetterselv, Jesper Brok and Ann M Møller (19 December 2007)
|
|
|||
|
Arash Afshari, MD, senior resident Hvidovre hospital, Department of Anaesthesiology and intensive care, Kettegaard Allé 30, 2650, DK, Jørn Wetterselv, Jesper Brok and Ann M Møller
Send response to journal:
|
Authors reply:
We would like to thank Torossian A et al for their positive comments to our article “Antithrombin III in critically ill patients” in their recent editorial.(1-2) However, there seem to be some issues that need clarification regarding the interpretation and understanding of our systematic review.(2) Torossian et al criticised our definition of critically ill patients for being too loose defined and thus perhaps leading to a heterogeneous group of patients in our meta-analyses. Since there is no clear definition or gold standard of the latter in the literature, we chose to accept definitions as proposed by the authors of the included trials. Realizing the controversy surrounding this topic, we chose to perform 32 various subgroup and sensitivity analyses, and among them we performed a thorough analyses of all the relevant populations. Even when doing so, our results do not indicate any beneficial intervention effect at all. We were, however, unable to publish them all in the final version of our review due to BMJ’s obvious limitations on the number of published figures and tables. The justification of pooling these trials could partly rest on the fact that we found no detectable over-all intervention effect. Additionally the lack of heterogeneity when pooling the data in various meta-analyses in our review may also indicates lack of an intervention effect.(3) We agree, however, that the confidence interval for the heterogeneity, I2, is in fact wide with the upper value being 48% (according to the test-based method proposed by Higgins and Thompson(4)) for the mortality analysis, which may reveal that a point estimate of I2 of 0 may conceal considerable heterogeneity. On the other hand most meta-analyses have an upper 95% confidence of I2 of more than 50%.(5) It should not be overlooked that all of the included clinical conditions in our review may result in DIC and thus have similar inflammatory pathways, which potentially should be affected by AT III. The terminology of sepsis has been widely criticised throughout the years for being a “garbage” diagnosis since this terminology in itself engulfs such a heterogeneous population. (6-7) So the entire discussion of performing a trial with a homogenous population is often hindered and limited by our shortcomings in terms of finding an exact definition of the critically ill conditions.’Lumping or splitting’ in a meta-analysis is often debated. In our peer reviewed and published protocol we planned to pool all eligible randomised trials irrespective type of participants and AT III regime (lumping). This, of course, includes the risk of losing some details and pooling heterogeneous trials. However, we think that the generalisability, the external validity, and usefulness of meta-analyses are increased considerably if the individual trials cover different patient populations, settings, and treatment regimen. Further, a broad meta-analysis increases power, reduces the risk of erroneous conclusions by the play of chance, and facilitates exploratory analyses, which may generate hypotheses for future research. Usually it is also recommended to pool a broad range of studies.(8) Additionally, we chose to exclude trials with AT III in patients with cardiovascular diseases, because AT III in this field has been used to compare various simultaneous active interventions. For instance, AT III has been used in cardiovascular by-pass procedures in order to evaluate graft function versus conventional anti-coagulant therapy and in ischemic patients(9). AT III has also been compared to other anticoagulants in patients with cardiac surgery.(10-11) Thus, due to the comparison with various active interventions in these trials, we find it more relevant to examine AT III’s role in the cardiovascular setting in a future separate systematic review. Torossian et al advocate for the use of a validated measure such as the Jadad scale in order to evaluate the role of bias. However, the Jadad-score has not been recommended to score quality of trials in meta-analyses for some years. This is due to trials with different methodological deficits may receive the same score on this scale and relationship between such a score and the degree to which a study is free from bias is unclear. This score has been abandoned for this reason of the Cochrane Collaboration. However, since there is no 'gold standard' for the 'true' validity of a trial, the value of validating any proposed scoring system is limited. Thus, none of the currently available scales for measuring the validity or 'quality' of trial can be recommended without major reservation.(12-14) Instead it is recommended to compare each individual quality component e.g., adequate vs. inadequate blinding in a subgroup analysis. (1)4 We did not find any significant difference in our subgroup analyses of different quality components. Torossian et al indicate that our calculations using Trial sequential analysis (TSA) may be inaccurate. TSA is used to eliminate spurious statistical significance due to repeated testing of accumulating data in a cumulative meta-analysis when the meta-analysis is updated whenever a new trial is added. As described in our article this is the fifth time that the meta-analysis of AT III is up-dated so far which underline the necessity to adjust p-values for multiple updating. Considering that a meta-analysis should at least include the number of participants as a well-powered randomised trial addressing the same question and that the required information size needs adjustment for heterogeneity our calculations of the information size could not be more accurate. What the authors might have had in mind is that the estimate of the anticipated intervention effect relates to the control event rate, which may vary with the subgroup one chose to focus on. This is evidently correct, but again, if we include only sepsis trials the point estimate of the pooled intervention effect or control event rate of trials performed so far is not noticeably affected as sepsis trials already by far constitute the bulk evidence for the (absent) effect of ATIII. Afshari A, Wetterslev J, Brok J, Moller A References: 1) Torossian A, Graf J, Bauhofer A. Antithrombin III in critically ill patients. BMJ. 2007;335:1219-1220 2) Afshari A, Wetterslev J, Brok J, Moller A. Antithrombin III in critically ill patients: systematic review with meta-analysis and trial sequential analysis. BMJ. 2007;335:1248-1251. doi:10.1136/bmj.39398.682500.25 3) Ioannidis JP, Trikalinos TA, Zintzaras E. Extreme between-study homogeneity in meta-analyses could offer useful insights. J Clin Epidemiol. 2006;59(10):1023-32 4) Higgins JP, Thompson SG. Quantifying heterogeneity in meta-analysis. Stat Med. 2002;15(11):1539-1558 5) Ioannidis JPA, Patsopoulos NA, Evangelou E. Uncertainty in heterogeneity estimates in meta-analysis. BMJ. 2007;335:914 6) Opal SM. Clinical trial design and outcomes in patients with severe sepsis. Shock. 2003;20:4: 295-302 7) Abraham E, Matthay MA, Dinarello CA, Vincent JL, Opal SM, Glauser M et al Consensus conference definitions for sepsis, septic shock, acute lung injury, and acute respiratory distress syndrome: time for a reevaluation. Crit Care Med. 2000;28(1):232-5 8) Gotzsche P. Why we need a broad perspective on meta-analysis. It may be crucially important for patients. BMJ. 2000;321:585-6 9) Avidan MS, Levy JH, Scholz J, Delphin E, Rosseel PM, Howie MB, et al. A phase III, double-blind, placebo-controlled, multicenter study on the efficacy of recombinant human antithrombin in heparin-resistant patients scheduled to undergo cardiac surgery necessitating cardiopulmonary bypass. Anesthesiology. 2005;102(2):276-84 10) Lemmer JH Jr, Despotis GJ. Antithrombin III concentrate to treat heparin resistance in patients undergoing cardiac surgery. J Thorac Cardiovasc Surg. 2002;123(2):213-7 11) Avidan MS, Levy JH, van Aken H, Feneck RO, Latimer RD,et al. Recombinant human antithrombin III restores heparin responsiveness and decreases activation of coagulation in heparin-resistant patients during cardiopulmonary bypass. J Thorac Cardiovasc Surg. 2005;130(1):107-13 12) Juni P, Altman DG, Egger M. Assessing the quality of randomized controlled trials. In: Egger M, Smith GD, Altman DG, editors. Systematic reviews in health care. Meta-analysis in context. London: BMJ Publishing Group. 2001;87-109 13) Higgins JPT, Green S, editors. Assessment of study quality. Cochrane handbook for systematic reviews of interventions 4.2.5 [Updated May 2005]; Section 6. In: The Cochrane Library, Issue 3. Chichester: John Wiley & Sons, Ltd.. 2005 14) Higgins J, Green S, editors. Cochrane Handbook for Systematic Reviews of Interventions 4.2.6 [Updated Sep 2006]. http://www.cochrane.dk/cochrane/handbook/handbook.htm Competing interests: None declared |
|||